1 \input texinfo @c -*-texinfo-*-
4 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
6 @c GNAT DOCUMENTATION o
10 @c GNAT is maintained by Ada Core Technologies Inc (http://www.gnat.com). o
12 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
14 @setfilename gnat_ugn.info
17 Copyright @copyright{} 1995-2009 Free Software Foundation,
20 Permission is granted to copy, distribute and/or modify this document
21 under the terms of the GNU Free Documentation License, Version 1.2 or
22 any later version published by the Free Software Foundation; with no
23 Invariant Sections, with no Front-Cover Texts and with no Back-Cover
24 Texts. A copy of the license is included in the section entitled
25 ``GNU Free Documentation License''.
28 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
30 @c GNAT_UGN Style Guide
32 @c 1. Always put a @noindent on the line before the first paragraph
33 @c after any of these commands:
45 @c 2. DO NOT use @example. Use @smallexample instead.
46 @c a) DO NOT use highlighting commands (@b{}, @i{}) inside an @smallexample
47 @c context. These can interfere with the readability of the texi
48 @c source file. Instead, use one of the following annotated
49 @c @smallexample commands, and preprocess the texi file with the
50 @c ada2texi tool (which generates appropriate highlighting):
51 @c @smallexample @c ada
52 @c @smallexample @c adanocomment
53 @c @smallexample @c projectfile
54 @c b) The "@c ada" markup will result in boldface for reserved words
55 @c and italics for comments
56 @c c) The "@c adanocomment" markup will result only in boldface for
57 @c reserved words (comments are left alone)
58 @c d) The "@c projectfile" markup is like "@c ada" except that the set
59 @c of reserved words include the new reserved words for project files
61 @c 3. Each @chapter, @section, @subsection, @subsubsection, etc.
62 @c command must be preceded by two empty lines
64 @c 4. The @item command should be on a line of its own if it is in an
65 @c @itemize or @enumerate command.
67 @c 5. When talking about ALI files use "ALI" (all uppercase), not "Ali"
70 @c 6. DO NOT put trailing spaces at the end of a line. Such spaces will
71 @c cause the document build to fail.
73 @c 7. DO NOT use @cartouche for examples that are longer than around 10 lines.
74 @c This command inhibits page breaks, so long examples in a @cartouche can
75 @c lead to large, ugly patches of empty space on a page.
77 @c NOTE: This file should be submitted to xgnatugn with either the vms flag
78 @c or the unw flag set. The unw flag covers topics for both Unix and
81 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
84 @c This flag is used where the text refers to conditions that exist when the
85 @c text was entered into the document but which may change over time.
86 @c Update the setting for the flag, and (if necessary) the text surrounding,
87 @c the references to the flag, on future doc revisions:
88 @c search for @value{NOW}.
92 @set DEFAULTLANGUAGEVERSION Ada 2005
93 @set NONDEFAULTLANGUAGEVERSION Ada 95
100 @set PLATFORM OpenVMS
105 @c The ARG is an optional argument. To be used for macro arguments in
106 @c their documentation (@defmac).
108 @r{[}@var{\varname\}@r{]}@c
111 @settitle @value{EDITION} User's Guide @value{PLATFORM}
112 @dircategory GNU Ada tools
114 * @value{EDITION} User's Guide: (gnat_ugn). @value{PLATFORM}
117 @include gcc-common.texi
119 @setchapternewpage odd
124 @title @value{EDITION} User's Guide
128 @titlefont{@i{@value{PLATFORM}}}
134 @subtitle GNAT, The GNU Ada Compiler
139 @vskip 0pt plus 1filll
146 @node Top, About This Guide, (dir), (dir)
147 @top @value{EDITION} User's Guide
150 @value{EDITION} User's Guide @value{PLATFORM}
153 GNAT, The GNU Ada Compiler@*
154 GCC version @value{version-GCC}@*
161 * Getting Started with GNAT::
162 * The GNAT Compilation Model::
163 * Compiling Using gcc::
164 * Binding Using gnatbind::
165 * Linking Using gnatlink::
166 * The GNAT Make Program gnatmake::
167 * Improving Performance::
168 * Renaming Files Using gnatchop::
169 * Configuration Pragmas::
170 * Handling Arbitrary File Naming Conventions Using gnatname::
171 * GNAT Project Manager::
172 * The Cross-Referencing Tools gnatxref and gnatfind::
173 * The GNAT Pretty-Printer gnatpp::
174 * The GNAT Metric Tool gnatmetric::
175 * File Name Krunching Using gnatkr::
176 * Preprocessing Using gnatprep::
178 * The GNAT Run-Time Library Builder gnatlbr::
180 * The GNAT Library Browser gnatls::
181 * Cleaning Up Using gnatclean::
183 * GNAT and Libraries::
184 * Using the GNU make Utility::
186 * Memory Management Issues::
187 * Stack Related Facilities::
188 * Verifying Properties Using gnatcheck::
189 * Creating Sample Bodies Using gnatstub::
190 * Generating Ada Bindings for C and C++ headers::
191 * Other Utility Programs::
192 * Running and Debugging Ada Programs::
194 * Code Coverage and Profiling::
197 * Compatibility with HP Ada::
199 * Platform-Specific Information for the Run-Time Libraries::
200 * Example of Binder Output File::
201 * Elaboration Order Handling in GNAT::
202 * Conditional Compilation::
204 * Compatibility and Porting Guide::
206 * Microsoft Windows Topics::
208 * GNU Free Documentation License::
211 --- The Detailed Node Listing ---
215 * What This Guide Contains::
216 * What You Should Know before Reading This Guide::
217 * Related Information::
220 Getting Started with GNAT
223 * Running a Simple Ada Program::
224 * Running a Program with Multiple Units::
225 * Using the gnatmake Utility::
227 * Editing with Emacs::
230 * Introduction to GPS::
233 The GNAT Compilation Model
235 * Source Representation::
236 * Foreign Language Representation::
237 * File Naming Rules::
238 * Using Other File Names::
239 * Alternative File Naming Schemes::
240 * Generating Object Files::
241 * Source Dependencies::
242 * The Ada Library Information Files::
243 * Binding an Ada Program::
244 * Mixed Language Programming::
246 * Building Mixed Ada & C++ Programs::
247 * Comparison between GNAT and C/C++ Compilation Models::
249 * Comparison between GNAT and Conventional Ada Library Models::
251 * Placement of temporary files::
254 Foreign Language Representation
257 * Other 8-Bit Codes::
258 * Wide Character Encodings::
260 Compiling Ada Programs With gcc
262 * Compiling Programs::
264 * Search Paths and the Run-Time Library (RTL)::
265 * Order of Compilation Issues::
270 * Output and Error Message Control::
271 * Warning Message Control::
272 * Debugging and Assertion Control::
273 * Validity Checking::
276 * Using gcc for Syntax Checking::
277 * Using gcc for Semantic Checking::
278 * Compiling Different Versions of Ada::
279 * Character Set Control::
280 * File Naming Control::
281 * Subprogram Inlining Control::
282 * Auxiliary Output Control::
283 * Debugging Control::
284 * Exception Handling Control::
285 * Units to Sources Mapping Files::
286 * Integrated Preprocessing::
291 Binding Ada Programs With gnatbind
294 * Switches for gnatbind::
295 * Command-Line Access::
296 * Search Paths for gnatbind::
297 * Examples of gnatbind Usage::
299 Switches for gnatbind
301 * Consistency-Checking Modes::
302 * Binder Error Message Control::
303 * Elaboration Control::
305 * Binding with Non-Ada Main Programs::
306 * Binding Programs with No Main Subprogram::
308 Linking Using gnatlink
311 * Switches for gnatlink::
313 The GNAT Make Program gnatmake
316 * Switches for gnatmake::
317 * Mode Switches for gnatmake::
318 * Notes on the Command Line::
319 * How gnatmake Works::
320 * Examples of gnatmake Usage::
322 Improving Performance
323 * Performance Considerations::
324 * Text_IO Suggestions::
325 * Reducing Size of Ada Executables with gnatelim::
326 * Reducing Size of Executables with unused subprogram/data elimination::
328 Performance Considerations
329 * Controlling Run-Time Checks::
330 * Use of Restrictions::
331 * Optimization Levels::
332 * Debugging Optimized Code::
333 * Inlining of Subprograms::
334 * Other Optimization Switches::
335 * Optimization and Strict Aliasing::
337 * Coverage Analysis::
340 Reducing Size of Ada Executables with gnatelim
343 * Correcting the List of Eliminate Pragmas::
344 * Making Your Executables Smaller::
345 * Summary of the gnatelim Usage Cycle::
347 Reducing Size of Executables with unused subprogram/data elimination
348 * About unused subprogram/data elimination::
349 * Compilation options::
351 Renaming Files Using gnatchop
353 * Handling Files with Multiple Units::
354 * Operating gnatchop in Compilation Mode::
355 * Command Line for gnatchop::
356 * Switches for gnatchop::
357 * Examples of gnatchop Usage::
359 Configuration Pragmas
361 * Handling of Configuration Pragmas::
362 * The Configuration Pragmas Files::
364 Handling Arbitrary File Naming Conventions Using gnatname
366 * Arbitrary File Naming Conventions::
368 * Switches for gnatname::
369 * Examples of gnatname Usage::
374 * Examples of Project Files::
375 * Project File Syntax::
376 * Objects and Sources in Project Files::
377 * Importing Projects::
378 * Project Extension::
379 * Project Hierarchy Extension::
380 * External References in Project Files::
381 * Packages in Project Files::
382 * Variables from Imported Projects::
385 * Stand-alone Library Projects::
386 * Switches Related to Project Files::
387 * Tools Supporting Project Files::
388 * An Extended Example::
389 * Project File Complete Syntax::
391 The Cross-Referencing Tools gnatxref and gnatfind
393 * gnatxref Switches::
394 * gnatfind Switches::
395 * Project Files for gnatxref and gnatfind::
396 * Regular Expressions in gnatfind and gnatxref::
397 * Examples of gnatxref Usage::
398 * Examples of gnatfind Usage::
400 The GNAT Pretty-Printer gnatpp
402 * Switches for gnatpp::
405 The GNAT Metrics Tool gnatmetric
407 * Switches for gnatmetric::
409 File Name Krunching Using gnatkr
414 * Examples of gnatkr Usage::
416 Preprocessing Using gnatprep
417 * Preprocessing Symbols::
419 * Switches for gnatprep::
420 * Form of Definitions File::
421 * Form of Input Text for gnatprep::
424 The GNAT Run-Time Library Builder gnatlbr
427 * Switches for gnatlbr::
428 * Examples of gnatlbr Usage::
431 The GNAT Library Browser gnatls
434 * Switches for gnatls::
435 * Examples of gnatls Usage::
437 Cleaning Up Using gnatclean
439 * Running gnatclean::
440 * Switches for gnatclean::
441 @c * Examples of gnatclean Usage::
447 * Introduction to Libraries in GNAT::
448 * General Ada Libraries::
449 * Stand-alone Ada Libraries::
450 * Rebuilding the GNAT Run-Time Library::
452 Using the GNU make Utility
454 * Using gnatmake in a Makefile::
455 * Automatically Creating a List of Directories::
456 * Generating the Command Line Switches::
457 * Overcoming Command Line Length Limits::
460 Memory Management Issues
462 * Some Useful Memory Pools::
463 * The GNAT Debug Pool Facility::
468 Stack Related Facilities
470 * Stack Overflow Checking::
471 * Static Stack Usage Analysis::
472 * Dynamic Stack Usage Analysis::
474 Some Useful Memory Pools
476 The GNAT Debug Pool Facility
482 * Switches for gnatmem::
483 * Example of gnatmem Usage::
486 Verifying Properties Using gnatcheck
488 * Format of the Report File::
489 * General gnatcheck Switches::
490 * gnatcheck Rule Options::
491 * Adding the Results of Compiler Checks to gnatcheck Output::
492 * Project-Wide Checks::
496 Sample Bodies Using gnatstub
499 * Switches for gnatstub::
501 Other Utility Programs
503 * Using Other Utility Programs with GNAT::
504 * The External Symbol Naming Scheme of GNAT::
505 * Converting Ada Files to html with gnathtml::
508 Code Coverage and Profiling
510 * Code Coverage of Ada Programs using gcov::
511 * Profiling an Ada Program using gprof::
514 Running and Debugging Ada Programs
516 * The GNAT Debugger GDB::
518 * Introduction to GDB Commands::
519 * Using Ada Expressions::
520 * Calling User-Defined Subprograms::
521 * Using the Next Command in a Function::
524 * Debugging Generic Units::
525 * GNAT Abnormal Termination or Failure to Terminate::
526 * Naming Conventions for GNAT Source Files::
527 * Getting Internal Debugging Information::
535 Compatibility with HP Ada
537 * Ada Language Compatibility::
538 * Differences in the Definition of Package System::
539 * Language-Related Features::
540 * The Package STANDARD::
541 * The Package SYSTEM::
542 * Tasking and Task-Related Features::
543 * Pragmas and Pragma-Related Features::
544 * Library of Predefined Units::
546 * Main Program Definition::
547 * Implementation-Defined Attributes::
548 * Compiler and Run-Time Interfacing::
549 * Program Compilation and Library Management::
551 * Implementation Limits::
552 * Tools and Utilities::
554 Language-Related Features
556 * Integer Types and Representations::
557 * Floating-Point Types and Representations::
558 * Pragmas Float_Representation and Long_Float::
559 * Fixed-Point Types and Representations::
560 * Record and Array Component Alignment::
562 * Other Representation Clauses::
564 Tasking and Task-Related Features
566 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
567 * Assigning Task IDs::
568 * Task IDs and Delays::
569 * Task-Related Pragmas::
570 * Scheduling and Task Priority::
572 * External Interrupts::
574 Pragmas and Pragma-Related Features
576 * Restrictions on the Pragma INLINE::
577 * Restrictions on the Pragma INTERFACE::
578 * Restrictions on the Pragma SYSTEM_NAME::
580 Library of Predefined Units
582 * Changes to DECLIB::
586 * Shared Libraries and Options Files::
590 Platform-Specific Information for the Run-Time Libraries
592 * Summary of Run-Time Configurations::
593 * Specifying a Run-Time Library::
594 * Choosing the Scheduling Policy::
595 * Solaris-Specific Considerations::
596 * Linux-Specific Considerations::
597 * AIX-Specific Considerations::
598 * Irix-Specific Considerations::
600 Example of Binder Output File
602 Elaboration Order Handling in GNAT
605 * Checking the Elaboration Order::
606 * Controlling the Elaboration Order::
607 * Controlling Elaboration in GNAT - Internal Calls::
608 * Controlling Elaboration in GNAT - External Calls::
609 * Default Behavior in GNAT - Ensuring Safety::
610 * Treatment of Pragma Elaborate::
611 * Elaboration Issues for Library Tasks::
612 * Mixing Elaboration Models::
613 * What to Do If the Default Elaboration Behavior Fails::
614 * Elaboration for Access-to-Subprogram Values::
615 * Summary of Procedures for Elaboration Control::
616 * Other Elaboration Order Considerations::
618 Conditional Compilation
619 * Use of Boolean Constants::
620 * Debugging - A Special Case::
621 * Conditionalizing Declarations::
622 * Use of Alternative Implementations::
627 * Basic Assembler Syntax::
628 * A Simple Example of Inline Assembler::
629 * Output Variables in Inline Assembler::
630 * Input Variables in Inline Assembler::
631 * Inlining Inline Assembler Code::
632 * Other Asm Functionality::
634 Compatibility and Porting Guide
636 * Compatibility with Ada 83::
637 * Compatibility between Ada 95 and Ada 2005::
638 * Implementation-dependent characteristics::
640 @c This brief section is only in the non-VMS version
641 @c The complete chapter on HP Ada issues is in the VMS version
642 * Compatibility with HP Ada 83::
644 * Compatibility with Other Ada Systems::
645 * Representation Clauses::
647 * Transitioning to 64-Bit GNAT for OpenVMS::
651 Microsoft Windows Topics
653 * Using GNAT on Windows::
654 * CONSOLE and WINDOWS subsystems::
656 * Mixed-Language Programming on Windows::
657 * Windows Calling Conventions::
658 * Introduction to Dynamic Link Libraries (DLLs)::
659 * Using DLLs with GNAT::
660 * Building DLLs with GNAT::
661 * GNAT and Windows Resources::
663 * Setting Stack Size from gnatlink::
664 * Setting Heap Size from gnatlink::
671 @node About This Guide
672 @unnumbered About This Guide
676 This guide describes the use of @value{EDITION},
677 a compiler and software development toolset for the full Ada
678 programming language, implemented on OpenVMS for HP's Alpha and
679 Integrity server (I64) platforms.
682 This guide describes the use of @value{EDITION},
683 a compiler and software development
684 toolset for the full Ada programming language.
686 It documents the features of the compiler and tools, and explains
687 how to use them to build Ada applications.
689 @value{EDITION} implements Ada 95 and Ada 2005, and it may also be invoked in
690 Ada 83 compatibility mode.
691 By default, @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
692 but you can override with a compiler switch
693 (@pxref{Compiling Different Versions of Ada})
694 to explicitly specify the language version.
695 Throughout this manual, references to ``Ada'' without a year suffix
696 apply to both the Ada 95 and Ada 2005 versions of the language.
700 For ease of exposition, ``@value{EDITION}'' will be referred to simply as
701 ``GNAT'' in the remainder of this document.
708 * What This Guide Contains::
709 * What You Should Know before Reading This Guide::
710 * Related Information::
714 @node What This Guide Contains
715 @unnumberedsec What This Guide Contains
718 This guide contains the following chapters:
722 @ref{Getting Started with GNAT}, describes how to get started compiling
723 and running Ada programs with the GNAT Ada programming environment.
725 @ref{The GNAT Compilation Model}, describes the compilation model used
729 @ref{Compiling Using gcc}, describes how to compile
730 Ada programs with @command{gcc}, the Ada compiler.
733 @ref{Binding Using gnatbind}, describes how to
734 perform binding of Ada programs with @code{gnatbind}, the GNAT binding
738 @ref{Linking Using gnatlink},
739 describes @command{gnatlink}, a
740 program that provides for linking using the GNAT run-time library to
741 construct a program. @command{gnatlink} can also incorporate foreign language
742 object units into the executable.
745 @ref{The GNAT Make Program gnatmake}, describes @command{gnatmake}, a
746 utility that automatically determines the set of sources
747 needed by an Ada compilation unit, and executes the necessary compilations
751 @ref{Improving Performance}, shows various techniques for making your
752 Ada program run faster or take less space.
753 It discusses the effect of the compiler's optimization switch and
754 also describes the @command{gnatelim} tool and unused subprogram/data
758 @ref{Renaming Files Using gnatchop}, describes
759 @code{gnatchop}, a utility that allows you to preprocess a file that
760 contains Ada source code, and split it into one or more new files, one
761 for each compilation unit.
764 @ref{Configuration Pragmas}, describes the configuration pragmas
768 @ref{Handling Arbitrary File Naming Conventions Using gnatname},
769 shows how to override the default GNAT file naming conventions,
770 either for an individual unit or globally.
773 @ref{GNAT Project Manager}, describes how to use project files
774 to organize large projects.
777 @ref{The Cross-Referencing Tools gnatxref and gnatfind}, discusses
778 @code{gnatxref} and @code{gnatfind}, two tools that provide an easy
779 way to navigate through sources.
782 @ref{The GNAT Pretty-Printer gnatpp}, shows how to produce a reformatted
783 version of an Ada source file with control over casing, indentation,
784 comment placement, and other elements of program presentation style.
787 @ref{The GNAT Metric Tool gnatmetric}, shows how to compute various
788 metrics for an Ada source file, such as the number of types and subprograms,
789 and assorted complexity measures.
792 @ref{File Name Krunching Using gnatkr}, describes the @code{gnatkr}
793 file name krunching utility, used to handle shortened
794 file names on operating systems with a limit on the length of names.
797 @ref{Preprocessing Using gnatprep}, describes @code{gnatprep}, a
798 preprocessor utility that allows a single source file to be used to
799 generate multiple or parameterized source files by means of macro
804 @ref{The GNAT Run-Time Library Builder gnatlbr}, describes @command{gnatlbr},
805 a tool for rebuilding the GNAT run time with user-supplied
806 configuration pragmas.
810 @ref{The GNAT Library Browser gnatls}, describes @code{gnatls}, a
811 utility that displays information about compiled units, including dependences
812 on the corresponding sources files, and consistency of compilations.
815 @ref{Cleaning Up Using gnatclean}, describes @code{gnatclean}, a utility
816 to delete files that are produced by the compiler, binder and linker.
820 @ref{GNAT and Libraries}, describes the process of creating and using
821 Libraries with GNAT. It also describes how to recompile the GNAT run-time
825 @ref{Using the GNU make Utility}, describes some techniques for using
826 the GNAT toolset in Makefiles.
830 @ref{Memory Management Issues}, describes some useful predefined storage pools
831 and in particular the GNAT Debug Pool facility, which helps detect incorrect
834 It also describes @command{gnatmem}, a utility that monitors dynamic
835 allocation and deallocation and helps detect ``memory leaks''.
839 @ref{Stack Related Facilities}, describes some useful tools associated with
840 stack checking and analysis.
843 @ref{Verifying Properties Using gnatcheck}, discusses @code{gnatcheck},
844 a utility that checks Ada code against a set of rules.
847 @ref{Creating Sample Bodies Using gnatstub}, discusses @code{gnatstub},
848 a utility that generates empty but compilable bodies for library units.
851 @ref{Generating Ada Bindings for C and C++ headers}, describes how to
852 generate automatically Ada bindings from C and C++ headers.
855 @ref{Other Utility Programs}, discusses several other GNAT utilities,
856 including @code{gnathtml}.
860 @ref{Code Coverage and Profiling}, describes how to perform a structural
861 coverage and profile the execution of Ada programs.
865 @ref{Running and Debugging Ada Programs}, describes how to run and debug
870 @ref{Compatibility with HP Ada}, details the compatibility of GNAT with
871 HP Ada 83 @footnote{``HP Ada'' refers to the legacy product originally
872 developed by Digital Equipment Corporation and currently supported by HP.}
873 for OpenVMS Alpha. This product was formerly known as DEC Ada,
876 historical compatibility reasons, the relevant libraries still use the
881 @ref{Platform-Specific Information for the Run-Time Libraries},
882 describes the various run-time
883 libraries supported by GNAT on various platforms and explains how to
884 choose a particular library.
887 @ref{Example of Binder Output File}, shows the source code for the binder
888 output file for a sample program.
891 @ref{Elaboration Order Handling in GNAT}, describes how GNAT helps
892 you deal with elaboration order issues.
895 @ref{Conditional Compilation}, describes how to model conditional compilation,
896 both with Ada in general and with GNAT facilities in particular.
899 @ref{Inline Assembler}, shows how to use the inline assembly facility
903 @ref{Compatibility and Porting Guide}, contains sections on compatibility
904 of GNAT with other Ada development environments (including Ada 83 systems),
905 to assist in porting code from those environments.
909 @ref{Microsoft Windows Topics}, presents information relevant to the
910 Microsoft Windows platform.
914 @c *************************************************
915 @node What You Should Know before Reading This Guide
916 @c *************************************************
917 @unnumberedsec What You Should Know before Reading This Guide
919 @cindex Ada 95 Language Reference Manual
920 @cindex Ada 2005 Language Reference Manual
922 This guide assumes a basic familiarity with the Ada 95 language, as
923 described in the International Standard ANSI/ISO/IEC-8652:1995, January
925 It does not require knowledge of the new features introduced by Ada 2005,
926 (officially known as ISO/IEC 8652:1995 with Technical Corrigendum 1
928 Both reference manuals are included in the GNAT documentation
931 @node Related Information
932 @unnumberedsec Related Information
935 For further information about related tools, refer to the following
940 @xref{Top, GNAT Reference Manual, About This Guide, gnat_rm, GNAT
941 Reference Manual}, which contains all reference material for the GNAT
942 implementation of Ada.
946 @cite{Using the GNAT Programming Studio}, which describes the GPS
947 Integrated Development Environment.
950 @cite{GNAT Programming Studio Tutorial}, which introduces the
951 main GPS features through examples.
955 @cite{Ada 95 Reference Manual}, which contains reference
956 material for the Ada 95 programming language.
959 @cite{Ada 2005 Reference Manual}, which contains reference
960 material for the Ada 2005 programming language.
963 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
965 in the GNU:[DOCS] directory,
967 for all details on the use of the GNU source-level debugger.
970 @xref{Top,, The extensible self-documenting text editor, emacs,
973 located in the GNU:[DOCS] directory if the EMACS kit is installed,
975 for full information on the extensible editor and programming
982 @unnumberedsec Conventions
984 @cindex Typographical conventions
987 Following are examples of the typographical and graphic conventions used
992 @code{Functions}, @command{utility program names}, @code{standard names},
996 @option{Option flags}
999 @file{File names}, @samp{button names}, and @samp{field names}.
1002 @code{Variables}, @env{environment variables}, and @var{metasyntactic
1009 @r{[}optional information or parameters@r{]}
1012 Examples are described by text
1014 and then shown this way.
1019 Commands that are entered by the user are preceded in this manual by the
1020 characters @w{``@code{$ }''} (dollar sign followed by space). If your system
1021 uses this sequence as a prompt, then the commands will appear exactly as
1022 you see them in the manual. If your system uses some other prompt, then
1023 the command will appear with the @code{$} replaced by whatever prompt
1024 character you are using.
1027 Full file names are shown with the ``@code{/}'' character
1028 as the directory separator; e.g., @file{parent-dir/subdir/myfile.adb}.
1029 If you are using GNAT on a Windows platform, please note that
1030 the ``@code{\}'' character should be used instead.
1033 @c ****************************
1034 @node Getting Started with GNAT
1035 @chapter Getting Started with GNAT
1038 This chapter describes some simple ways of using GNAT to build
1039 executable Ada programs.
1041 @ref{Running GNAT}, through @ref{Using the gnatmake Utility},
1042 show how to use the command line environment.
1043 @ref{Introduction to GPS}, provides a brief
1044 introduction to the GNAT Programming Studio, a visually-oriented
1045 Integrated Development Environment for GNAT.
1046 GPS offers a graphical ``look and feel'', support for development in
1047 other programming languages, comprehensive browsing features, and
1048 many other capabilities.
1049 For information on GPS please refer to
1050 @cite{Using the GNAT Programming Studio}.
1055 * Running a Simple Ada Program::
1056 * Running a Program with Multiple Units::
1057 * Using the gnatmake Utility::
1059 * Editing with Emacs::
1062 * Introduction to GPS::
1067 @section Running GNAT
1070 Three steps are needed to create an executable file from an Ada source
1075 The source file(s) must be compiled.
1077 The file(s) must be bound using the GNAT binder.
1079 All appropriate object files must be linked to produce an executable.
1083 All three steps are most commonly handled by using the @command{gnatmake}
1084 utility program that, given the name of the main program, automatically
1085 performs the necessary compilation, binding and linking steps.
1087 @node Running a Simple Ada Program
1088 @section Running a Simple Ada Program
1091 Any text editor may be used to prepare an Ada program.
1093 used, the optional Ada mode may be helpful in laying out the program.)
1095 program text is a normal text file. We will assume in our initial
1096 example that you have used your editor to prepare the following
1097 standard format text file:
1099 @smallexample @c ada
1101 with Ada.Text_IO; use Ada.Text_IO;
1104 Put_Line ("Hello WORLD!");
1110 This file should be named @file{hello.adb}.
1111 With the normal default file naming conventions, GNAT requires
1113 contain a single compilation unit whose file name is the
1115 with periods replaced by hyphens; the
1116 extension is @file{ads} for a
1117 spec and @file{adb} for a body.
1118 You can override this default file naming convention by use of the
1119 special pragma @code{Source_File_Name} (@pxref{Using Other File Names}).
1120 Alternatively, if you want to rename your files according to this default
1121 convention, which is probably more convenient if you will be using GNAT
1122 for all your compilations, then the @code{gnatchop} utility
1123 can be used to generate correctly-named source files
1124 (@pxref{Renaming Files Using gnatchop}).
1126 You can compile the program using the following command (@code{$} is used
1127 as the command prompt in the examples in this document):
1134 @command{gcc} is the command used to run the compiler. This compiler is
1135 capable of compiling programs in several languages, including Ada and
1136 C. It assumes that you have given it an Ada program if the file extension is
1137 either @file{.ads} or @file{.adb}, and it will then call
1138 the GNAT compiler to compile the specified file.
1141 The @option{-c} switch is required. It tells @command{gcc} to only do a
1142 compilation. (For C programs, @command{gcc} can also do linking, but this
1143 capability is not used directly for Ada programs, so the @option{-c}
1144 switch must always be present.)
1147 This compile command generates a file
1148 @file{hello.o}, which is the object
1149 file corresponding to your Ada program. It also generates
1150 an ``Ada Library Information'' file @file{hello.ali},
1151 which contains additional information used to check
1152 that an Ada program is consistent.
1153 To build an executable file,
1154 use @code{gnatbind} to bind the program
1155 and @command{gnatlink} to link it. The
1156 argument to both @code{gnatbind} and @command{gnatlink} is the name of the
1157 @file{ALI} file, but the default extension of @file{.ali} can
1158 be omitted. This means that in the most common case, the argument
1159 is simply the name of the main program:
1167 A simpler method of carrying out these steps is to use
1169 a master program that invokes all the required
1170 compilation, binding and linking tools in the correct order. In particular,
1171 @command{gnatmake} automatically recompiles any sources that have been
1172 modified since they were last compiled, or sources that depend
1173 on such modified sources, so that ``version skew'' is avoided.
1174 @cindex Version skew (avoided by @command{gnatmake})
1177 $ gnatmake hello.adb
1181 The result is an executable program called @file{hello}, which can be
1189 assuming that the current directory is on the search path
1190 for executable programs.
1193 and, if all has gone well, you will see
1200 appear in response to this command.
1202 @c ****************************************
1203 @node Running a Program with Multiple Units
1204 @section Running a Program with Multiple Units
1207 Consider a slightly more complicated example that has three files: a
1208 main program, and the spec and body of a package:
1210 @smallexample @c ada
1213 package Greetings is
1218 with Ada.Text_IO; use Ada.Text_IO;
1219 package body Greetings is
1222 Put_Line ("Hello WORLD!");
1225 procedure Goodbye is
1227 Put_Line ("Goodbye WORLD!");
1244 Following the one-unit-per-file rule, place this program in the
1245 following three separate files:
1249 spec of package @code{Greetings}
1252 body of package @code{Greetings}
1255 body of main program
1259 To build an executable version of
1260 this program, we could use four separate steps to compile, bind, and link
1261 the program, as follows:
1265 $ gcc -c greetings.adb
1271 Note that there is no required order of compilation when using GNAT.
1272 In particular it is perfectly fine to compile the main program first.
1273 Also, it is not necessary to compile package specs in the case where
1274 there is an accompanying body; you only need to compile the body. If you want
1275 to submit these files to the compiler for semantic checking and not code
1276 generation, then use the
1277 @option{-gnatc} switch:
1280 $ gcc -c greetings.ads -gnatc
1284 Although the compilation can be done in separate steps as in the
1285 above example, in practice it is almost always more convenient
1286 to use the @command{gnatmake} tool. All you need to know in this case
1287 is the name of the main program's source file. The effect of the above four
1288 commands can be achieved with a single one:
1291 $ gnatmake gmain.adb
1295 In the next section we discuss the advantages of using @command{gnatmake} in
1298 @c *****************************
1299 @node Using the gnatmake Utility
1300 @section Using the @command{gnatmake} Utility
1303 If you work on a program by compiling single components at a time using
1304 @command{gcc}, you typically keep track of the units you modify. In order to
1305 build a consistent system, you compile not only these units, but also any
1306 units that depend on the units you have modified.
1307 For example, in the preceding case,
1308 if you edit @file{gmain.adb}, you only need to recompile that file. But if
1309 you edit @file{greetings.ads}, you must recompile both
1310 @file{greetings.adb} and @file{gmain.adb}, because both files contain
1311 units that depend on @file{greetings.ads}.
1313 @code{gnatbind} will warn you if you forget one of these compilation
1314 steps, so that it is impossible to generate an inconsistent program as a
1315 result of forgetting to do a compilation. Nevertheless it is tedious and
1316 error-prone to keep track of dependencies among units.
1317 One approach to handle the dependency-bookkeeping is to use a
1318 makefile. However, makefiles present maintenance problems of their own:
1319 if the dependencies change as you change the program, you must make
1320 sure that the makefile is kept up-to-date manually, which is also an
1321 error-prone process.
1323 The @command{gnatmake} utility takes care of these details automatically.
1324 Invoke it using either one of the following forms:
1327 $ gnatmake gmain.adb
1328 $ gnatmake ^gmain^GMAIN^
1332 The argument is the name of the file containing the main program;
1333 you may omit the extension. @command{gnatmake}
1334 examines the environment, automatically recompiles any files that need
1335 recompiling, and binds and links the resulting set of object files,
1336 generating the executable file, @file{^gmain^GMAIN.EXE^}.
1337 In a large program, it
1338 can be extremely helpful to use @command{gnatmake}, because working out by hand
1339 what needs to be recompiled can be difficult.
1341 Note that @command{gnatmake}
1342 takes into account all the Ada rules that
1343 establish dependencies among units. These include dependencies that result
1344 from inlining subprogram bodies, and from
1345 generic instantiation. Unlike some other
1346 Ada make tools, @command{gnatmake} does not rely on the dependencies that were
1347 found by the compiler on a previous compilation, which may possibly
1348 be wrong when sources change. @command{gnatmake} determines the exact set of
1349 dependencies from scratch each time it is run.
1352 @node Editing with Emacs
1353 @section Editing with Emacs
1357 Emacs is an extensible self-documenting text editor that is available in a
1358 separate VMSINSTAL kit.
1360 Invoke Emacs by typing @kbd{Emacs} at the command prompt. To get started,
1361 click on the Emacs Help menu and run the Emacs Tutorial.
1362 In a character cell terminal, Emacs help is invoked with @kbd{Ctrl-h} (also
1363 written as @kbd{C-h}), and the tutorial by @kbd{C-h t}.
1365 Documentation on Emacs and other tools is available in Emacs under the
1366 pull-down menu button: @code{Help - Info}. After selecting @code{Info},
1367 use the middle mouse button to select a topic (e.g.@: Emacs).
1369 In a character cell terminal, do @kbd{C-h i} to invoke info, and then @kbd{m}
1370 (stands for menu) followed by the menu item desired, as in @kbd{m Emacs}, to
1371 get to the Emacs manual.
1372 Help on Emacs is also available by typing @kbd{HELP EMACS} at the DCL command
1375 The tutorial is highly recommended in order to learn the intricacies of Emacs,
1376 which is sufficiently extensible to provide for a complete programming
1377 environment and shell for the sophisticated user.
1381 @node Introduction to GPS
1382 @section Introduction to GPS
1383 @cindex GPS (GNAT Programming Studio)
1384 @cindex GNAT Programming Studio (GPS)
1386 Although the command line interface (@command{gnatmake}, etc.) alone
1387 is sufficient, a graphical Interactive Development
1388 Environment can make it easier for you to compose, navigate, and debug
1389 programs. This section describes the main features of GPS
1390 (``GNAT Programming Studio''), the GNAT graphical IDE.
1391 You will see how to use GPS to build and debug an executable, and
1392 you will also learn some of the basics of the GNAT ``project'' facility.
1394 GPS enables you to do much more than is presented here;
1395 e.g., you can produce a call graph, interface to a third-party
1396 Version Control System, and inspect the generated assembly language
1398 Indeed, GPS also supports languages other than Ada.
1399 Such additional information, and an explanation of all of the GPS menu
1400 items. may be found in the on-line help, which includes
1401 a user's guide and a tutorial (these are also accessible from the GNAT
1405 * Building a New Program with GPS::
1406 * Simple Debugging with GPS::
1409 @node Building a New Program with GPS
1410 @subsection Building a New Program with GPS
1412 GPS invokes the GNAT compilation tools using information
1413 contained in a @emph{project} (also known as a @emph{project file}):
1414 a collection of properties such
1415 as source directories, identities of main subprograms, tool switches, etc.,
1416 and their associated values.
1417 See @ref{GNAT Project Manager} for details.
1418 In order to run GPS, you will need to either create a new project
1419 or else open an existing one.
1421 This section will explain how you can use GPS to create a project,
1422 to associate Ada source files with a project, and to build and run
1426 @item @emph{Creating a project}
1428 Invoke GPS, either from the command line or the platform's IDE.
1429 After it starts, GPS will display a ``Welcome'' screen with three
1434 @code{Start with default project in directory}
1437 @code{Create new project with wizard}
1440 @code{Open existing project}
1444 Select @code{Create new project with wizard} and press @code{OK}.
1445 A new window will appear. In the text box labeled with
1446 @code{Enter the name of the project to create}, type @file{sample}
1447 as the project name.
1448 In the next box, browse to choose the directory in which you
1449 would like to create the project file.
1450 After selecting an appropriate directory, press @code{Forward}.
1452 A window will appear with the title
1453 @code{Version Control System Configuration}.
1454 Simply press @code{Forward}.
1456 A window will appear with the title
1457 @code{Please select the source directories for this project}.
1458 The directory that you specified for the project file will be selected
1459 by default as the one to use for sources; simply press @code{Forward}.
1461 A window will appear with the title
1462 @code{Please select the build directory for this project}.
1463 The directory that you specified for the project file will be selected
1464 by default for object files and executables;
1465 simply press @code{Forward}.
1467 A window will appear with the title
1468 @code{Please select the main units for this project}.
1469 You will supply this information later, after creating the source file.
1470 Simply press @code{Forward} for now.
1472 A window will appear with the title
1473 @code{Please select the switches to build the project}.
1474 Press @code{Apply}. This will create a project file named
1475 @file{sample.prj} in the directory that you had specified.
1477 @item @emph{Creating and saving the source file}
1479 After you create the new project, a GPS window will appear, which is
1480 partitioned into two main sections:
1484 A @emph{Workspace area}, initially greyed out, which you will use for
1485 creating and editing source files
1488 Directly below, a @emph{Messages area}, which initially displays a
1489 ``Welcome'' message.
1490 (If the Messages area is not visible, drag its border upward to expand it.)
1494 Select @code{File} on the menu bar, and then the @code{New} command.
1495 The Workspace area will become white, and you can now
1496 enter the source program explicitly.
1497 Type the following text
1499 @smallexample @c ada
1501 with Ada.Text_IO; use Ada.Text_IO;
1504 Put_Line("Hello from GPS!");
1510 Select @code{File}, then @code{Save As}, and enter the source file name
1512 The file will be saved in the same directory you specified as the
1513 location of the default project file.
1515 @item @emph{Updating the project file}
1517 You need to add the new source file to the project.
1519 the @code{Project} menu and then @code{Edit project properties}.
1520 Click the @code{Main files} tab on the left, and then the
1522 Choose @file{hello.adb} from the list, and press @code{Open}.
1523 The project settings window will reflect this action.
1526 @item @emph{Building and running the program}
1528 In the main GPS window, now choose the @code{Build} menu, then @code{Make},
1529 and select @file{hello.adb}.
1530 The Messages window will display the resulting invocations of @command{gcc},
1531 @command{gnatbind}, and @command{gnatlink}
1532 (reflecting the default switch settings from the
1533 project file that you created) and then a ``successful compilation/build''
1536 To run the program, choose the @code{Build} menu, then @code{Run}, and
1537 select @command{hello}.
1538 An @emph{Arguments Selection} window will appear.
1539 There are no command line arguments, so just click @code{OK}.
1541 The Messages window will now display the program's output (the string
1542 @code{Hello from GPS}), and at the bottom of the GPS window a status
1543 update is displayed (@code{Run: hello}).
1544 Close the GPS window (or select @code{File}, then @code{Exit}) to
1545 terminate this GPS session.
1548 @node Simple Debugging with GPS
1549 @subsection Simple Debugging with GPS
1551 This section illustrates basic debugging techniques (setting breakpoints,
1552 examining/modifying variables, single stepping).
1555 @item @emph{Opening a project}
1557 Start GPS and select @code{Open existing project}; browse to
1558 specify the project file @file{sample.prj} that you had created in the
1561 @item @emph{Creating a source file}
1563 Select @code{File}, then @code{New}, and type in the following program:
1565 @smallexample @c ada
1567 with Ada.Text_IO; use Ada.Text_IO;
1568 procedure Example is
1569 Line : String (1..80);
1572 Put_Line("Type a line of text at each prompt; an empty line to exit");
1576 Put_Line (Line (1..N) );
1584 Select @code{File}, then @code{Save as}, and enter the file name
1587 @item @emph{Updating the project file}
1589 Add @code{Example} as a new main unit for the project:
1592 Select @code{Project}, then @code{Edit Project Properties}.
1595 Select the @code{Main files} tab, click @code{Add}, then
1596 select the file @file{example.adb} from the list, and
1598 You will see the file name appear in the list of main units
1604 @item @emph{Building/running the executable}
1606 To build the executable
1607 select @code{Build}, then @code{Make}, and then choose @file{example.adb}.
1609 Run the program to see its effect (in the Messages area).
1610 Each line that you enter is displayed; an empty line will
1611 cause the loop to exit and the program to terminate.
1613 @item @emph{Debugging the program}
1615 Note that the @option{-g} switches to @command{gcc} and @command{gnatlink},
1616 which are required for debugging, are on by default when you create
1618 Thus unless you intentionally remove these settings, you will be able
1619 to debug any program that you develop using GPS.
1622 @item @emph{Initializing}
1624 Select @code{Debug}, then @code{Initialize}, then @file{example}
1626 @item @emph{Setting a breakpoint}
1628 After performing the initialization step, you will observe a small
1629 icon to the right of each line number.
1630 This serves as a toggle for breakpoints; clicking the icon will
1631 set a breakpoint at the corresponding line (the icon will change to
1632 a red circle with an ``x''), and clicking it again
1633 will remove the breakpoint / reset the icon.
1635 For purposes of this example, set a breakpoint at line 10 (the
1636 statement @code{Put_Line@ (Line@ (1..N));}
1638 @item @emph{Starting program execution}
1640 Select @code{Debug}, then @code{Run}. When the
1641 @code{Program Arguments} window appears, click @code{OK}.
1642 A console window will appear; enter some line of text,
1643 e.g.@: @code{abcde}, at the prompt.
1644 The program will pause execution when it gets to the
1645 breakpoint, and the corresponding line is highlighted.
1647 @item @emph{Examining a variable}
1649 Move the mouse over one of the occurrences of the variable @code{N}.
1650 You will see the value (5) displayed, in ``tool tip'' fashion.
1651 Right click on @code{N}, select @code{Debug}, then select @code{Display N}.
1652 You will see information about @code{N} appear in the @code{Debugger Data}
1653 pane, showing the value as 5.
1655 @item @emph{Assigning a new value to a variable}
1657 Right click on the @code{N} in the @code{Debugger Data} pane, and
1658 select @code{Set value of N}.
1659 When the input window appears, enter the value @code{4} and click
1661 This value does not automatically appear in the @code{Debugger Data}
1662 pane; to see it, right click again on the @code{N} in the
1663 @code{Debugger Data} pane and select @code{Update value}.
1664 The new value, 4, will appear in red.
1666 @item @emph{Single stepping}
1668 Select @code{Debug}, then @code{Next}.
1669 This will cause the next statement to be executed, in this case the
1670 call of @code{Put_Line} with the string slice.
1671 Notice in the console window that the displayed string is simply
1672 @code{abcd} and not @code{abcde} which you had entered.
1673 This is because the upper bound of the slice is now 4 rather than 5.
1675 @item @emph{Removing a breakpoint}
1677 Toggle the breakpoint icon at line 10.
1679 @item @emph{Resuming execution from a breakpoint}
1681 Select @code{Debug}, then @code{Continue}.
1682 The program will reach the next iteration of the loop, and
1683 wait for input after displaying the prompt.
1684 This time, just hit the @kbd{Enter} key.
1685 The value of @code{N} will be 0, and the program will terminate.
1686 The console window will disappear.
1691 @node The GNAT Compilation Model
1692 @chapter The GNAT Compilation Model
1693 @cindex GNAT compilation model
1694 @cindex Compilation model
1697 * Source Representation::
1698 * Foreign Language Representation::
1699 * File Naming Rules::
1700 * Using Other File Names::
1701 * Alternative File Naming Schemes::
1702 * Generating Object Files::
1703 * Source Dependencies::
1704 * The Ada Library Information Files::
1705 * Binding an Ada Program::
1706 * Mixed Language Programming::
1708 * Building Mixed Ada & C++ Programs::
1709 * Comparison between GNAT and C/C++ Compilation Models::
1711 * Comparison between GNAT and Conventional Ada Library Models::
1713 * Placement of temporary files::
1718 This chapter describes the compilation model used by GNAT. Although
1719 similar to that used by other languages, such as C and C++, this model
1720 is substantially different from the traditional Ada compilation models,
1721 which are based on a library. The model is initially described without
1722 reference to the library-based model. If you have not previously used an
1723 Ada compiler, you need only read the first part of this chapter. The
1724 last section describes and discusses the differences between the GNAT
1725 model and the traditional Ada compiler models. If you have used other
1726 Ada compilers, this section will help you to understand those
1727 differences, and the advantages of the GNAT model.
1729 @node Source Representation
1730 @section Source Representation
1734 Ada source programs are represented in standard text files, using
1735 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1736 7-bit ASCII set, plus additional characters used for
1737 representing foreign languages (@pxref{Foreign Language Representation}
1738 for support of non-USA character sets). The format effector characters
1739 are represented using their standard ASCII encodings, as follows:
1744 Vertical tab, @code{16#0B#}
1748 Horizontal tab, @code{16#09#}
1752 Carriage return, @code{16#0D#}
1756 Line feed, @code{16#0A#}
1760 Form feed, @code{16#0C#}
1764 Source files are in standard text file format. In addition, GNAT will
1765 recognize a wide variety of stream formats, in which the end of
1766 physical lines is marked by any of the following sequences:
1767 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1768 in accommodating files that are imported from other operating systems.
1770 @cindex End of source file
1771 @cindex Source file, end
1773 The end of a source file is normally represented by the physical end of
1774 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1775 recognized as signalling the end of the source file. Again, this is
1776 provided for compatibility with other operating systems where this
1777 code is used to represent the end of file.
1779 Each file contains a single Ada compilation unit, including any pragmas
1780 associated with the unit. For example, this means you must place a
1781 package declaration (a package @dfn{spec}) and the corresponding body in
1782 separate files. An Ada @dfn{compilation} (which is a sequence of
1783 compilation units) is represented using a sequence of files. Similarly,
1784 you will place each subunit or child unit in a separate file.
1786 @node Foreign Language Representation
1787 @section Foreign Language Representation
1790 GNAT supports the standard character sets defined in Ada as well as
1791 several other non-standard character sets for use in localized versions
1792 of the compiler (@pxref{Character Set Control}).
1795 * Other 8-Bit Codes::
1796 * Wide Character Encodings::
1804 The basic character set is Latin-1. This character set is defined by ISO
1805 standard 8859, part 1. The lower half (character codes @code{16#00#}
1806 @dots{} @code{16#7F#)} is identical to standard ASCII coding, but the upper half
1807 is used to represent additional characters. These include extended letters
1808 used by European languages, such as French accents, the vowels with umlauts
1809 used in German, and the extra letter A-ring used in Swedish.
1811 @findex Ada.Characters.Latin_1
1812 For a complete list of Latin-1 codes and their encodings, see the source
1813 file of library unit @code{Ada.Characters.Latin_1} in file
1814 @file{a-chlat1.ads}.
1815 You may use any of these extended characters freely in character or
1816 string literals. In addition, the extended characters that represent
1817 letters can be used in identifiers.
1819 @node Other 8-Bit Codes
1820 @subsection Other 8-Bit Codes
1823 GNAT also supports several other 8-bit coding schemes:
1826 @item ISO 8859-2 (Latin-2)
1829 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1832 @item ISO 8859-3 (Latin-3)
1835 Latin-3 letters allowed in identifiers, with uppercase and lowercase
1838 @item ISO 8859-4 (Latin-4)
1841 Latin-4 letters allowed in identifiers, with uppercase and lowercase
1844 @item ISO 8859-5 (Cyrillic)
1847 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
1848 lowercase equivalence.
1850 @item ISO 8859-15 (Latin-9)
1853 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
1854 lowercase equivalence
1856 @item IBM PC (code page 437)
1857 @cindex code page 437
1858 This code page is the normal default for PCs in the U.S. It corresponds
1859 to the original IBM PC character set. This set has some, but not all, of
1860 the extended Latin-1 letters, but these letters do not have the same
1861 encoding as Latin-1. In this mode, these letters are allowed in
1862 identifiers with uppercase and lowercase equivalence.
1864 @item IBM PC (code page 850)
1865 @cindex code page 850
1866 This code page is a modification of 437 extended to include all the
1867 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
1868 mode, all these letters are allowed in identifiers with uppercase and
1869 lowercase equivalence.
1871 @item Full Upper 8-bit
1872 Any character in the range 80-FF allowed in identifiers, and all are
1873 considered distinct. In other words, there are no uppercase and lowercase
1874 equivalences in this range. This is useful in conjunction with
1875 certain encoding schemes used for some foreign character sets (e.g.,
1876 the typical method of representing Chinese characters on the PC).
1879 No upper-half characters in the range 80-FF are allowed in identifiers.
1880 This gives Ada 83 compatibility for identifier names.
1884 For precise data on the encodings permitted, and the uppercase and lowercase
1885 equivalences that are recognized, see the file @file{csets.adb} in
1886 the GNAT compiler sources. You will need to obtain a full source release
1887 of GNAT to obtain this file.
1889 @node Wide Character Encodings
1890 @subsection Wide Character Encodings
1893 GNAT allows wide character codes to appear in character and string
1894 literals, and also optionally in identifiers, by means of the following
1895 possible encoding schemes:
1900 In this encoding, a wide character is represented by the following five
1908 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1909 characters (using uppercase letters) of the wide character code. For
1910 example, ESC A345 is used to represent the wide character with code
1912 This scheme is compatible with use of the full Wide_Character set.
1914 @item Upper-Half Coding
1915 @cindex Upper-Half Coding
1916 The wide character with encoding @code{16#abcd#} where the upper bit is on
1917 (in other words, ``a'' is in the range 8-F) is represented as two bytes,
1918 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
1919 character, but is not required to be in the upper half. This method can
1920 be also used for shift-JIS or EUC, where the internal coding matches the
1923 @item Shift JIS Coding
1924 @cindex Shift JIS Coding
1925 A wide character is represented by a two-character sequence,
1927 @code{16#cd#}, with the restrictions described for upper-half encoding as
1928 described above. The internal character code is the corresponding JIS
1929 character according to the standard algorithm for Shift-JIS
1930 conversion. Only characters defined in the JIS code set table can be
1931 used with this encoding method.
1935 A wide character is represented by a two-character sequence
1937 @code{16#cd#}, with both characters being in the upper half. The internal
1938 character code is the corresponding JIS character according to the EUC
1939 encoding algorithm. Only characters defined in the JIS code set table
1940 can be used with this encoding method.
1943 A wide character is represented using
1944 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1945 10646-1/Am.2. Depending on the character value, the representation
1946 is a one, two, or three byte sequence:
1951 16#0000#-16#007f#: 2#0@var{xxxxxxx}#
1952 16#0080#-16#07ff#: 2#110@var{xxxxx}# 2#10@var{xxxxxx}#
1953 16#0800#-16#ffff#: 2#1110@var{xxxx}# 2#10@var{xxxxxx}# 2#10@var{xxxxxx}#
1958 where the @var{xxx} bits correspond to the left-padded bits of the
1959 16-bit character value. Note that all lower half ASCII characters
1960 are represented as ASCII bytes and all upper half characters and
1961 other wide characters are represented as sequences of upper-half
1962 (The full UTF-8 scheme allows for encoding 31-bit characters as
1963 6-byte sequences, but in this implementation, all UTF-8 sequences
1964 of four or more bytes length will be treated as illegal).
1965 @item Brackets Coding
1966 In this encoding, a wide character is represented by the following eight
1974 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1975 characters (using uppercase letters) of the wide character code. For
1976 example, [``A345''] is used to represent the wide character with code
1977 @code{16#A345#}. It is also possible (though not required) to use the
1978 Brackets coding for upper half characters. For example, the code
1979 @code{16#A3#} can be represented as @code{[``A3'']}.
1981 This scheme is compatible with use of the full Wide_Character set,
1982 and is also the method used for wide character encoding in the standard
1983 ACVC (Ada Compiler Validation Capability) test suite distributions.
1988 Note: Some of these coding schemes do not permit the full use of the
1989 Ada character set. For example, neither Shift JIS, nor EUC allow the
1990 use of the upper half of the Latin-1 set.
1992 @node File Naming Rules
1993 @section File Naming Rules
1996 The default file name is determined by the name of the unit that the
1997 file contains. The name is formed by taking the full expanded name of
1998 the unit and replacing the separating dots with hyphens and using
1999 ^lowercase^uppercase^ for all letters.
2001 An exception arises if the file name generated by the above rules starts
2002 with one of the characters
2004 @samp{A}, @samp{G}, @samp{I}, or @samp{S},
2007 @samp{a}, @samp{g}, @samp{i}, or @samp{s},
2009 and the second character is a
2010 minus. In this case, the character ^tilde^dollar sign^ is used in place
2011 of the minus. The reason for this special rule is to avoid clashes with
2012 the standard names for child units of the packages System, Ada,
2013 Interfaces, and GNAT, which use the prefixes
2015 @samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},
2018 @samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},
2022 The file extension is @file{.ads} for a spec and
2023 @file{.adb} for a body. The following list shows some
2024 examples of these rules.
2031 @item arith_functions.ads
2032 Arith_Functions (package spec)
2033 @item arith_functions.adb
2034 Arith_Functions (package body)
2036 Func.Spec (child package spec)
2038 Func.Spec (child package body)
2040 Sub (subunit of Main)
2041 @item ^a~bad.adb^A$BAD.ADB^
2042 A.Bad (child package body)
2046 Following these rules can result in excessively long
2047 file names if corresponding
2048 unit names are long (for example, if child units or subunits are
2049 heavily nested). An option is available to shorten such long file names
2050 (called file name ``krunching''). This may be particularly useful when
2051 programs being developed with GNAT are to be used on operating systems
2052 with limited file name lengths. @xref{Using gnatkr}.
2054 Of course, no file shortening algorithm can guarantee uniqueness over
2055 all possible unit names; if file name krunching is used, it is your
2056 responsibility to ensure no name clashes occur. Alternatively you
2057 can specify the exact file names that you want used, as described
2058 in the next section. Finally, if your Ada programs are migrating from a
2059 compiler with a different naming convention, you can use the gnatchop
2060 utility to produce source files that follow the GNAT naming conventions.
2061 (For details @pxref{Renaming Files Using gnatchop}.)
2063 Note: in the case of @code{Windows NT/XP} or @code{OpenVMS} operating
2064 systems, case is not significant. So for example on @code{Windows XP}
2065 if the canonical name is @code{main-sub.adb}, you can use the file name
2066 @code{Main-Sub.adb} instead. However, case is significant for other
2067 operating systems, so for example, if you want to use other than
2068 canonically cased file names on a Unix system, you need to follow
2069 the procedures described in the next section.
2071 @node Using Other File Names
2072 @section Using Other File Names
2076 In the previous section, we have described the default rules used by
2077 GNAT to determine the file name in which a given unit resides. It is
2078 often convenient to follow these default rules, and if you follow them,
2079 the compiler knows without being explicitly told where to find all
2082 However, in some cases, particularly when a program is imported from
2083 another Ada compiler environment, it may be more convenient for the
2084 programmer to specify which file names contain which units. GNAT allows
2085 arbitrary file names to be used by means of the Source_File_Name pragma.
2086 The form of this pragma is as shown in the following examples:
2087 @cindex Source_File_Name pragma
2089 @smallexample @c ada
2091 pragma Source_File_Name (My_Utilities.Stacks,
2092 Spec_File_Name => "myutilst_a.ada");
2093 pragma Source_File_name (My_Utilities.Stacks,
2094 Body_File_Name => "myutilst.ada");
2099 As shown in this example, the first argument for the pragma is the unit
2100 name (in this example a child unit). The second argument has the form
2101 of a named association. The identifier
2102 indicates whether the file name is for a spec or a body;
2103 the file name itself is given by a string literal.
2105 The source file name pragma is a configuration pragma, which means that
2106 normally it will be placed in the @file{gnat.adc}
2107 file used to hold configuration
2108 pragmas that apply to a complete compilation environment.
2109 For more details on how the @file{gnat.adc} file is created and used
2110 see @ref{Handling of Configuration Pragmas}.
2111 @cindex @file{gnat.adc}
2114 GNAT allows completely arbitrary file names to be specified using the
2115 source file name pragma. However, if the file name specified has an
2116 extension other than @file{.ads} or @file{.adb} it is necessary to use
2117 a special syntax when compiling the file. The name in this case must be
2118 preceded by the special sequence @option{-x} followed by a space and the name
2119 of the language, here @code{ada}, as in:
2122 $ gcc -c -x ada peculiar_file_name.sim
2127 @command{gnatmake} handles non-standard file names in the usual manner (the
2128 non-standard file name for the main program is simply used as the
2129 argument to gnatmake). Note that if the extension is also non-standard,
2130 then it must be included in the @command{gnatmake} command, it may not
2133 @node Alternative File Naming Schemes
2134 @section Alternative File Naming Schemes
2135 @cindex File naming schemes, alternative
2138 In the previous section, we described the use of the @code{Source_File_Name}
2139 pragma to allow arbitrary names to be assigned to individual source files.
2140 However, this approach requires one pragma for each file, and especially in
2141 large systems can result in very long @file{gnat.adc} files, and also create
2142 a maintenance problem.
2144 GNAT also provides a facility for specifying systematic file naming schemes
2145 other than the standard default naming scheme previously described. An
2146 alternative scheme for naming is specified by the use of
2147 @code{Source_File_Name} pragmas having the following format:
2148 @cindex Source_File_Name pragma
2150 @smallexample @c ada
2151 pragma Source_File_Name (
2152 Spec_File_Name => FILE_NAME_PATTERN
2153 @r{[},Casing => CASING_SPEC@r{]}
2154 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2156 pragma Source_File_Name (
2157 Body_File_Name => FILE_NAME_PATTERN
2158 @r{[},Casing => CASING_SPEC@r{]}
2159 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2161 pragma Source_File_Name (
2162 Subunit_File_Name => FILE_NAME_PATTERN
2163 @r{[},Casing => CASING_SPEC@r{]}
2164 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2166 FILE_NAME_PATTERN ::= STRING_LITERAL
2167 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
2171 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
2172 It contains a single asterisk character, and the unit name is substituted
2173 systematically for this asterisk. The optional parameter
2174 @code{Casing} indicates
2175 whether the unit name is to be all upper-case letters, all lower-case letters,
2176 or mixed-case. If no
2177 @code{Casing} parameter is used, then the default is all
2178 ^lower-case^upper-case^.
2180 The optional @code{Dot_Replacement} string is used to replace any periods
2181 that occur in subunit or child unit names. If no @code{Dot_Replacement}
2182 argument is used then separating dots appear unchanged in the resulting
2184 Although the above syntax indicates that the
2185 @code{Casing} argument must appear
2186 before the @code{Dot_Replacement} argument, but it
2187 is also permissible to write these arguments in the opposite order.
2189 As indicated, it is possible to specify different naming schemes for
2190 bodies, specs, and subunits. Quite often the rule for subunits is the
2191 same as the rule for bodies, in which case, there is no need to give
2192 a separate @code{Subunit_File_Name} rule, and in this case the
2193 @code{Body_File_name} rule is used for subunits as well.
2195 The separate rule for subunits can also be used to implement the rather
2196 unusual case of a compilation environment (e.g.@: a single directory) which
2197 contains a subunit and a child unit with the same unit name. Although
2198 both units cannot appear in the same partition, the Ada Reference Manual
2199 allows (but does not require) the possibility of the two units coexisting
2200 in the same environment.
2202 The file name translation works in the following steps:
2207 If there is a specific @code{Source_File_Name} pragma for the given unit,
2208 then this is always used, and any general pattern rules are ignored.
2211 If there is a pattern type @code{Source_File_Name} pragma that applies to
2212 the unit, then the resulting file name will be used if the file exists. If
2213 more than one pattern matches, the latest one will be tried first, and the
2214 first attempt resulting in a reference to a file that exists will be used.
2217 If no pattern type @code{Source_File_Name} pragma that applies to the unit
2218 for which the corresponding file exists, then the standard GNAT default
2219 naming rules are used.
2224 As an example of the use of this mechanism, consider a commonly used scheme
2225 in which file names are all lower case, with separating periods copied
2226 unchanged to the resulting file name, and specs end with @file{.1.ada}, and
2227 bodies end with @file{.2.ada}. GNAT will follow this scheme if the following
2230 @smallexample @c ada
2231 pragma Source_File_Name
2232 (Spec_File_Name => "*.1.ada");
2233 pragma Source_File_Name
2234 (Body_File_Name => "*.2.ada");
2238 The default GNAT scheme is actually implemented by providing the following
2239 default pragmas internally:
2241 @smallexample @c ada
2242 pragma Source_File_Name
2243 (Spec_File_Name => "*.ads", Dot_Replacement => "-");
2244 pragma Source_File_Name
2245 (Body_File_Name => "*.adb", Dot_Replacement => "-");
2249 Our final example implements a scheme typically used with one of the
2250 Ada 83 compilers, where the separator character for subunits was ``__''
2251 (two underscores), specs were identified by adding @file{_.ADA}, bodies
2252 by adding @file{.ADA}, and subunits by
2253 adding @file{.SEP}. All file names were
2254 upper case. Child units were not present of course since this was an
2255 Ada 83 compiler, but it seems reasonable to extend this scheme to use
2256 the same double underscore separator for child units.
2258 @smallexample @c ada
2259 pragma Source_File_Name
2260 (Spec_File_Name => "*_.ADA",
2261 Dot_Replacement => "__",
2262 Casing = Uppercase);
2263 pragma Source_File_Name
2264 (Body_File_Name => "*.ADA",
2265 Dot_Replacement => "__",
2266 Casing = Uppercase);
2267 pragma Source_File_Name
2268 (Subunit_File_Name => "*.SEP",
2269 Dot_Replacement => "__",
2270 Casing = Uppercase);
2273 @node Generating Object Files
2274 @section Generating Object Files
2277 An Ada program consists of a set of source files, and the first step in
2278 compiling the program is to generate the corresponding object files.
2279 These are generated by compiling a subset of these source files.
2280 The files you need to compile are the following:
2284 If a package spec has no body, compile the package spec to produce the
2285 object file for the package.
2288 If a package has both a spec and a body, compile the body to produce the
2289 object file for the package. The source file for the package spec need
2290 not be compiled in this case because there is only one object file, which
2291 contains the code for both the spec and body of the package.
2294 For a subprogram, compile the subprogram body to produce the object file
2295 for the subprogram. The spec, if one is present, is as usual in a
2296 separate file, and need not be compiled.
2300 In the case of subunits, only compile the parent unit. A single object
2301 file is generated for the entire subunit tree, which includes all the
2305 Compile child units independently of their parent units
2306 (though, of course, the spec of all the ancestor unit must be present in order
2307 to compile a child unit).
2311 Compile generic units in the same manner as any other units. The object
2312 files in this case are small dummy files that contain at most the
2313 flag used for elaboration checking. This is because GNAT always handles generic
2314 instantiation by means of macro expansion. However, it is still necessary to
2315 compile generic units, for dependency checking and elaboration purposes.
2319 The preceding rules describe the set of files that must be compiled to
2320 generate the object files for a program. Each object file has the same
2321 name as the corresponding source file, except that the extension is
2324 You may wish to compile other files for the purpose of checking their
2325 syntactic and semantic correctness. For example, in the case where a
2326 package has a separate spec and body, you would not normally compile the
2327 spec. However, it is convenient in practice to compile the spec to make
2328 sure it is error-free before compiling clients of this spec, because such
2329 compilations will fail if there is an error in the spec.
2331 GNAT provides an option for compiling such files purely for the
2332 purposes of checking correctness; such compilations are not required as
2333 part of the process of building a program. To compile a file in this
2334 checking mode, use the @option{-gnatc} switch.
2336 @node Source Dependencies
2337 @section Source Dependencies
2340 A given object file clearly depends on the source file which is compiled
2341 to produce it. Here we are using @dfn{depends} in the sense of a typical
2342 @code{make} utility; in other words, an object file depends on a source
2343 file if changes to the source file require the object file to be
2345 In addition to this basic dependency, a given object may depend on
2346 additional source files as follows:
2350 If a file being compiled @code{with}'s a unit @var{X}, the object file
2351 depends on the file containing the spec of unit @var{X}. This includes
2352 files that are @code{with}'ed implicitly either because they are parents
2353 of @code{with}'ed child units or they are run-time units required by the
2354 language constructs used in a particular unit.
2357 If a file being compiled instantiates a library level generic unit, the
2358 object file depends on both the spec and body files for this generic
2362 If a file being compiled instantiates a generic unit defined within a
2363 package, the object file depends on the body file for the package as
2364 well as the spec file.
2368 @cindex @option{-gnatn} switch
2369 If a file being compiled contains a call to a subprogram for which
2370 pragma @code{Inline} applies and inlining is activated with the
2371 @option{-gnatn} switch, the object file depends on the file containing the
2372 body of this subprogram as well as on the file containing the spec. Note
2373 that for inlining to actually occur as a result of the use of this switch,
2374 it is necessary to compile in optimizing mode.
2376 @cindex @option{-gnatN} switch
2377 The use of @option{-gnatN} activates inlining optimization
2378 that is performed by the front end of the compiler. This inlining does
2379 not require that the code generation be optimized. Like @option{-gnatn},
2380 the use of this switch generates additional dependencies.
2382 When using a gcc-based back end (in practice this means using any version
2383 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
2384 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
2385 Historically front end inlining was more extensive than the gcc back end
2386 inlining, but that is no longer the case.
2389 If an object file @file{O} depends on the proper body of a subunit through
2390 inlining or instantiation, it depends on the parent unit of the subunit.
2391 This means that any modification of the parent unit or one of its subunits
2392 affects the compilation of @file{O}.
2395 The object file for a parent unit depends on all its subunit body files.
2398 The previous two rules meant that for purposes of computing dependencies and
2399 recompilation, a body and all its subunits are treated as an indivisible whole.
2402 These rules are applied transitively: if unit @code{A} @code{with}'s
2403 unit @code{B}, whose elaboration calls an inlined procedure in package
2404 @code{C}, the object file for unit @code{A} will depend on the body of
2405 @code{C}, in file @file{c.adb}.
2407 The set of dependent files described by these rules includes all the
2408 files on which the unit is semantically dependent, as dictated by the
2409 Ada language standard. However, it is a superset of what the
2410 standard describes, because it includes generic, inline, and subunit
2413 An object file must be recreated by recompiling the corresponding source
2414 file if any of the source files on which it depends are modified. For
2415 example, if the @code{make} utility is used to control compilation,
2416 the rule for an Ada object file must mention all the source files on
2417 which the object file depends, according to the above definition.
2418 The determination of the necessary
2419 recompilations is done automatically when one uses @command{gnatmake}.
2422 @node The Ada Library Information Files
2423 @section The Ada Library Information Files
2424 @cindex Ada Library Information files
2425 @cindex @file{ALI} files
2428 Each compilation actually generates two output files. The first of these
2429 is the normal object file that has a @file{.o} extension. The second is a
2430 text file containing full dependency information. It has the same
2431 name as the source file, but an @file{.ali} extension.
2432 This file is known as the Ada Library Information (@file{ALI}) file.
2433 The following information is contained in the @file{ALI} file.
2437 Version information (indicates which version of GNAT was used to compile
2438 the unit(s) in question)
2441 Main program information (including priority and time slice settings,
2442 as well as the wide character encoding used during compilation).
2445 List of arguments used in the @command{gcc} command for the compilation
2448 Attributes of the unit, including configuration pragmas used, an indication
2449 of whether the compilation was successful, exception model used etc.
2452 A list of relevant restrictions applying to the unit (used for consistency)
2456 Categorization information (e.g.@: use of pragma @code{Pure}).
2459 Information on all @code{with}'ed units, including presence of
2460 @code{Elaborate} or @code{Elaborate_All} pragmas.
2463 Information from any @code{Linker_Options} pragmas used in the unit
2466 Information on the use of @code{Body_Version} or @code{Version}
2467 attributes in the unit.
2470 Dependency information. This is a list of files, together with
2471 time stamp and checksum information. These are files on which
2472 the unit depends in the sense that recompilation is required
2473 if any of these units are modified.
2476 Cross-reference data. Contains information on all entities referenced
2477 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
2478 provide cross-reference information.
2483 For a full detailed description of the format of the @file{ALI} file,
2484 see the source of the body of unit @code{Lib.Writ}, contained in file
2485 @file{lib-writ.adb} in the GNAT compiler sources.
2487 @node Binding an Ada Program
2488 @section Binding an Ada Program
2491 When using languages such as C and C++, once the source files have been
2492 compiled the only remaining step in building an executable program
2493 is linking the object modules together. This means that it is possible to
2494 link an inconsistent version of a program, in which two units have
2495 included different versions of the same header.
2497 The rules of Ada do not permit such an inconsistent program to be built.
2498 For example, if two clients have different versions of the same package,
2499 it is illegal to build a program containing these two clients.
2500 These rules are enforced by the GNAT binder, which also determines an
2501 elaboration order consistent with the Ada rules.
2503 The GNAT binder is run after all the object files for a program have
2504 been created. It is given the name of the main program unit, and from
2505 this it determines the set of units required by the program, by reading the
2506 corresponding ALI files. It generates error messages if the program is
2507 inconsistent or if no valid order of elaboration exists.
2509 If no errors are detected, the binder produces a main program, in Ada by
2510 default, that contains calls to the elaboration procedures of those
2511 compilation unit that require them, followed by
2512 a call to the main program. This Ada program is compiled to generate the
2513 object file for the main program. The name of
2514 the Ada file is @file{b~@var{xxx}.adb} (with the corresponding spec
2515 @file{b~@var{xxx}.ads}) where @var{xxx} is the name of the
2518 Finally, the linker is used to build the resulting executable program,
2519 using the object from the main program from the bind step as well as the
2520 object files for the Ada units of the program.
2522 @node Mixed Language Programming
2523 @section Mixed Language Programming
2524 @cindex Mixed Language Programming
2527 This section describes how to develop a mixed-language program,
2528 specifically one that comprises units in both Ada and C.
2531 * Interfacing to C::
2532 * Calling Conventions::
2535 @node Interfacing to C
2536 @subsection Interfacing to C
2538 Interfacing Ada with a foreign language such as C involves using
2539 compiler directives to import and/or export entity definitions in each
2540 language---using @code{extern} statements in C, for instance, and the
2541 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada.
2542 A full treatment of these topics is provided in Appendix B, section 1
2543 of the Ada Reference Manual.
2545 There are two ways to build a program using GNAT that contains some Ada
2546 sources and some foreign language sources, depending on whether or not
2547 the main subprogram is written in Ada. Here is a source example with
2548 the main subprogram in Ada:
2554 void print_num (int num)
2556 printf ("num is %d.\n", num);
2562 /* num_from_Ada is declared in my_main.adb */
2563 extern int num_from_Ada;
2567 return num_from_Ada;
2571 @smallexample @c ada
2573 procedure My_Main is
2575 -- Declare then export an Integer entity called num_from_Ada
2576 My_Num : Integer := 10;
2577 pragma Export (C, My_Num, "num_from_Ada");
2579 -- Declare an Ada function spec for Get_Num, then use
2580 -- C function get_num for the implementation.
2581 function Get_Num return Integer;
2582 pragma Import (C, Get_Num, "get_num");
2584 -- Declare an Ada procedure spec for Print_Num, then use
2585 -- C function print_num for the implementation.
2586 procedure Print_Num (Num : Integer);
2587 pragma Import (C, Print_Num, "print_num");
2590 Print_Num (Get_Num);
2596 To build this example, first compile the foreign language files to
2597 generate object files:
2599 ^gcc -c file1.c^gcc -c FILE1.C^
2600 ^gcc -c file2.c^gcc -c FILE2.C^
2604 Then, compile the Ada units to produce a set of object files and ALI
2607 gnatmake ^-c^/ACTIONS=COMPILE^ my_main.adb
2611 Run the Ada binder on the Ada main program:
2613 gnatbind my_main.ali
2617 Link the Ada main program, the Ada objects and the other language
2620 gnatlink my_main.ali file1.o file2.o
2624 The last three steps can be grouped in a single command:
2626 gnatmake my_main.adb -largs file1.o file2.o
2629 @cindex Binder output file
2631 If the main program is in a language other than Ada, then you may have
2632 more than one entry point into the Ada subsystem. You must use a special
2633 binder option to generate callable routines that initialize and
2634 finalize the Ada units (@pxref{Binding with Non-Ada Main Programs}).
2635 Calls to the initialization and finalization routines must be inserted
2636 in the main program, or some other appropriate point in the code. The
2637 call to initialize the Ada units must occur before the first Ada
2638 subprogram is called, and the call to finalize the Ada units must occur
2639 after the last Ada subprogram returns. The binder will place the
2640 initialization and finalization subprograms into the
2641 @file{b~@var{xxx}.adb} file where they can be accessed by your C
2642 sources. To illustrate, we have the following example:
2646 extern void adainit (void);
2647 extern void adafinal (void);
2648 extern int add (int, int);
2649 extern int sub (int, int);
2651 int main (int argc, char *argv[])
2657 /* Should print "21 + 7 = 28" */
2658 printf ("%d + %d = %d\n", a, b, add (a, b));
2659 /* Should print "21 - 7 = 14" */
2660 printf ("%d - %d = %d\n", a, b, sub (a, b));
2666 @smallexample @c ada
2669 function Add (A, B : Integer) return Integer;
2670 pragma Export (C, Add, "add");
2674 package body Unit1 is
2675 function Add (A, B : Integer) return Integer is
2683 function Sub (A, B : Integer) return Integer;
2684 pragma Export (C, Sub, "sub");
2688 package body Unit2 is
2689 function Sub (A, B : Integer) return Integer is
2698 The build procedure for this application is similar to the last
2699 example's. First, compile the foreign language files to generate object
2702 ^gcc -c main.c^gcc -c main.c^
2706 Next, compile the Ada units to produce a set of object files and ALI
2709 gnatmake ^-c^/ACTIONS=COMPILE^ unit1.adb
2710 gnatmake ^-c^/ACTIONS=COMPILE^ unit2.adb
2714 Run the Ada binder on every generated ALI file. Make sure to use the
2715 @option{-n} option to specify a foreign main program:
2717 gnatbind ^-n^/NOMAIN^ unit1.ali unit2.ali
2721 Link the Ada main program, the Ada objects and the foreign language
2722 objects. You need only list the last ALI file here:
2724 gnatlink unit2.ali main.o -o exec_file
2727 This procedure yields a binary executable called @file{exec_file}.
2731 Depending on the circumstances (for example when your non-Ada main object
2732 does not provide symbol @code{main}), you may also need to instruct the
2733 GNAT linker not to include the standard startup objects by passing the
2734 @option{^-nostartfiles^/NOSTART_FILES^} switch to @command{gnatlink}.
2736 @node Calling Conventions
2737 @subsection Calling Conventions
2738 @cindex Foreign Languages
2739 @cindex Calling Conventions
2740 GNAT follows standard calling sequence conventions and will thus interface
2741 to any other language that also follows these conventions. The following
2742 Convention identifiers are recognized by GNAT:
2745 @cindex Interfacing to Ada
2746 @cindex Other Ada compilers
2747 @cindex Convention Ada
2749 This indicates that the standard Ada calling sequence will be
2750 used and all Ada data items may be passed without any limitations in the
2751 case where GNAT is used to generate both the caller and callee. It is also
2752 possible to mix GNAT generated code and code generated by another Ada
2753 compiler. In this case, the data types should be restricted to simple
2754 cases, including primitive types. Whether complex data types can be passed
2755 depends on the situation. Probably it is safe to pass simple arrays, such
2756 as arrays of integers or floats. Records may or may not work, depending
2757 on whether both compilers lay them out identically. Complex structures
2758 involving variant records, access parameters, tasks, or protected types,
2759 are unlikely to be able to be passed.
2761 Note that in the case of GNAT running
2762 on a platform that supports HP Ada 83, a higher degree of compatibility
2763 can be guaranteed, and in particular records are layed out in an identical
2764 manner in the two compilers. Note also that if output from two different
2765 compilers is mixed, the program is responsible for dealing with elaboration
2766 issues. Probably the safest approach is to write the main program in the
2767 version of Ada other than GNAT, so that it takes care of its own elaboration
2768 requirements, and then call the GNAT-generated adainit procedure to ensure
2769 elaboration of the GNAT components. Consult the documentation of the other
2770 Ada compiler for further details on elaboration.
2772 However, it is not possible to mix the tasking run time of GNAT and
2773 HP Ada 83, All the tasking operations must either be entirely within
2774 GNAT compiled sections of the program, or entirely within HP Ada 83
2775 compiled sections of the program.
2777 @cindex Interfacing to Assembly
2778 @cindex Convention Assembler
2780 Specifies assembler as the convention. In practice this has the
2781 same effect as convention Ada (but is not equivalent in the sense of being
2782 considered the same convention).
2784 @cindex Convention Asm
2787 Equivalent to Assembler.
2789 @cindex Interfacing to COBOL
2790 @cindex Convention COBOL
2793 Data will be passed according to the conventions described
2794 in section B.4 of the Ada Reference Manual.
2797 @cindex Interfacing to C
2798 @cindex Convention C
2800 Data will be passed according to the conventions described
2801 in section B.3 of the Ada Reference Manual.
2803 A note on interfacing to a C ``varargs'' function:
2804 @findex C varargs function
2805 @cindex Interfacing to C varargs function
2806 @cindex varargs function interfaces
2810 In C, @code{varargs} allows a function to take a variable number of
2811 arguments. There is no direct equivalent in this to Ada. One
2812 approach that can be used is to create a C wrapper for each
2813 different profile and then interface to this C wrapper. For
2814 example, to print an @code{int} value using @code{printf},
2815 create a C function @code{printfi} that takes two arguments, a
2816 pointer to a string and an int, and calls @code{printf}.
2817 Then in the Ada program, use pragma @code{Import} to
2818 interface to @code{printfi}.
2821 It may work on some platforms to directly interface to
2822 a @code{varargs} function by providing a specific Ada profile
2823 for a particular call. However, this does not work on
2824 all platforms, since there is no guarantee that the
2825 calling sequence for a two argument normal C function
2826 is the same as for calling a @code{varargs} C function with
2827 the same two arguments.
2830 @cindex Convention Default
2835 @cindex Convention External
2842 @cindex Interfacing to C++
2843 @cindex Convention C++
2844 @item C_Plus_Plus (or CPP)
2845 This stands for C++. For most purposes this is identical to C.
2846 See the separate description of the specialized GNAT pragmas relating to
2847 C++ interfacing for further details.
2851 @cindex Interfacing to Fortran
2852 @cindex Convention Fortran
2854 Data will be passed according to the conventions described
2855 in section B.5 of the Ada Reference Manual.
2858 This applies to an intrinsic operation, as defined in the Ada
2859 Reference Manual. If a pragma Import (Intrinsic) applies to a subprogram,
2860 this means that the body of the subprogram is provided by the compiler itself,
2861 usually by means of an efficient code sequence, and that the user does not
2862 supply an explicit body for it. In an application program, the pragma may
2863 be applied to the following sets of names:
2867 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right,
2868 Shift_Right_Arithmetic. The corresponding subprogram declaration must have
2869 two formal parameters. The
2870 first one must be a signed integer type or a modular type with a binary
2871 modulus, and the second parameter must be of type Natural.
2872 The return type must be the same as the type of the first argument. The size
2873 of this type can only be 8, 16, 32, or 64.
2876 Binary arithmetic operators: ``+'', ``-'', ``*'', ``/''
2877 The corresponding operator declaration must have parameters and result type
2878 that have the same root numeric type (for example, all three are long_float
2879 types). This simplifies the definition of operations that use type checking
2880 to perform dimensional checks:
2882 @smallexample @c ada
2883 type Distance is new Long_Float;
2884 type Time is new Long_Float;
2885 type Velocity is new Long_Float;
2886 function "/" (D : Distance; T : Time)
2888 pragma Import (Intrinsic, "/");
2892 This common idiom is often programmed with a generic definition and an
2893 explicit body. The pragma makes it simpler to introduce such declarations.
2894 It incurs no overhead in compilation time or code size, because it is
2895 implemented as a single machine instruction.
2898 General subprogram entities, to bind an Ada subprogram declaration to
2899 a compiler builtin by name with back-ends where such interfaces are
2900 available. A typical example is the set of ``__builtin'' functions
2901 exposed by the GCC back-end, as in the following example:
2903 @smallexample @c ada
2904 function builtin_sqrt (F : Float) return Float;
2905 pragma Import (Intrinsic, builtin_sqrt, "__builtin_sqrtf");
2908 Most of the GCC builtins are accessible this way, and as for other
2909 import conventions (e.g. C), it is the user's responsibility to ensure
2910 that the Ada subprogram profile matches the underlying builtin
2918 @cindex Convention Stdcall
2920 This is relevant only to Windows XP/2000/NT implementations of GNAT,
2921 and specifies that the @code{Stdcall} calling sequence will be used,
2922 as defined by the NT API. Nevertheless, to ease building
2923 cross-platform bindings this convention will be handled as a @code{C} calling
2924 convention on non-Windows platforms.
2927 @cindex Convention DLL
2929 This is equivalent to @code{Stdcall}.
2932 @cindex Convention Win32
2934 This is equivalent to @code{Stdcall}.
2938 @cindex Convention Stubbed
2940 This is a special convention that indicates that the compiler
2941 should provide a stub body that raises @code{Program_Error}.
2945 GNAT additionally provides a useful pragma @code{Convention_Identifier}
2946 that can be used to parametrize conventions and allow additional synonyms
2947 to be specified. For example if you have legacy code in which the convention
2948 identifier Fortran77 was used for Fortran, you can use the configuration
2951 @smallexample @c ada
2952 pragma Convention_Identifier (Fortran77, Fortran);
2956 And from now on the identifier Fortran77 may be used as a convention
2957 identifier (for example in an @code{Import} pragma) with the same
2961 @node Building Mixed Ada & C++ Programs
2962 @section Building Mixed Ada and C++ Programs
2965 A programmer inexperienced with mixed-language development may find that
2966 building an application containing both Ada and C++ code can be a
2967 challenge. This section gives a few
2968 hints that should make this task easier. The first section addresses
2969 the differences between interfacing with C and interfacing with C++.
2971 looks into the delicate problem of linking the complete application from
2972 its Ada and C++ parts. The last section gives some hints on how the GNAT
2973 run-time library can be adapted in order to allow inter-language dispatching
2974 with a new C++ compiler.
2977 * Interfacing to C++::
2978 * Linking a Mixed C++ & Ada Program::
2979 * A Simple Example::
2980 * Interfacing with C++ constructors::
2981 * Interfacing with C++ at the Class Level::
2984 @node Interfacing to C++
2985 @subsection Interfacing to C++
2988 GNAT supports interfacing with the G++ compiler (or any C++ compiler
2989 generating code that is compatible with the G++ Application Binary
2990 Interface ---see http://www.codesourcery.com/archives/cxx-abi).
2993 Interfacing can be done at 3 levels: simple data, subprograms, and
2994 classes. In the first two cases, GNAT offers a specific @code{Convention
2995 C_Plus_Plus} (or @code{CPP}) that behaves exactly like @code{Convention C}.
2996 Usually, C++ mangles the names of subprograms. To generate proper mangled
2997 names automatically, see @ref{Generating Ada Bindings for C and C++ headers}).
2998 This problem can also be addressed manually in two ways:
3002 by modifying the C++ code in order to force a C convention using
3003 the @code{extern "C"} syntax.
3006 by figuring out the mangled name (using e.g. @command{nm}) and using it as the
3007 Link_Name argument of the pragma import.
3011 Interfacing at the class level can be achieved by using the GNAT specific
3012 pragmas such as @code{CPP_Constructor}. @xref{Interfacing to C++,,,
3013 gnat_rm, GNAT Reference Manual}, for additional information.
3015 @node Linking a Mixed C++ & Ada Program
3016 @subsection Linking a Mixed C++ & Ada Program
3019 Usually the linker of the C++ development system must be used to link
3020 mixed applications because most C++ systems will resolve elaboration
3021 issues (such as calling constructors on global class instances)
3022 transparently during the link phase. GNAT has been adapted to ease the
3023 use of a foreign linker for the last phase. Three cases can be
3028 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
3029 The C++ linker can simply be called by using the C++ specific driver
3032 Note that if the C++ code uses inline functions, you will need to
3033 compile your C++ code with the @code{-fkeep-inline-functions} switch in
3034 order to provide an existing function implementation that the Ada code can
3038 $ g++ -c -fkeep-inline-functions file1.C
3039 $ g++ -c -fkeep-inline-functions file2.C
3040 $ gnatmake ada_unit -largs file1.o file2.o --LINK=g++
3044 Using GNAT and G++ from two different GCC installations: If both
3045 compilers are on the @env{PATH}, the previous method may be used. It is
3046 important to note that environment variables such as
3047 @env{C_INCLUDE_PATH}, @env{GCC_EXEC_PREFIX}, @env{BINUTILS_ROOT}, and
3048 @env{GCC_ROOT} will affect both compilers
3049 at the same time and may make one of the two compilers operate
3050 improperly if set during invocation of the wrong compiler. It is also
3051 very important that the linker uses the proper @file{libgcc.a} GCC
3052 library -- that is, the one from the C++ compiler installation. The
3053 implicit link command as suggested in the @command{gnatmake} command
3054 from the former example can be replaced by an explicit link command with
3055 the full-verbosity option in order to verify which library is used:
3058 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
3060 If there is a problem due to interfering environment variables, it can
3061 be worked around by using an intermediate script. The following example
3062 shows the proper script to use when GNAT has not been installed at its
3063 default location and g++ has been installed at its default location:
3071 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
3075 Using a non-GNU C++ compiler: The commands previously described can be
3076 used to insure that the C++ linker is used. Nonetheless, you need to add
3077 a few more parameters to the link command line, depending on the exception
3080 If the @code{setjmp/longjmp} exception mechanism is used, only the paths
3081 to the libgcc libraries are required:
3086 CC $* `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a`
3087 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3090 Where CC is the name of the non-GNU C++ compiler.
3092 If the @code{zero cost} exception mechanism is used, and the platform
3093 supports automatic registration of exception tables (e.g.@: Solaris or IRIX),
3094 paths to more objects are required:
3099 CC `gcc -print-file-name=crtbegin.o` $* \
3100 `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a` \
3101 `gcc -print-file-name=crtend.o`
3102 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3105 If the @code{zero cost} exception mechanism is used, and the platform
3106 doesn't support automatic registration of exception tables (e.g.@: HP-UX,
3107 Tru64 or AIX), the simple approach described above will not work and
3108 a pre-linking phase using GNAT will be necessary.
3112 Another alternative is to use the @command{gprbuild} multi-language builder
3113 which has a large knowledge base and knows how to link Ada and C++ code
3114 together automatically in most cases.
3116 @node A Simple Example
3117 @subsection A Simple Example
3119 The following example, provided as part of the GNAT examples, shows how
3120 to achieve procedural interfacing between Ada and C++ in both
3121 directions. The C++ class A has two methods. The first method is exported
3122 to Ada by the means of an extern C wrapper function. The second method
3123 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
3124 a limited record with a layout comparable to the C++ class. The Ada
3125 subprogram, in turn, calls the C++ method. So, starting from the C++
3126 main program, the process passes back and forth between the two
3130 Here are the compilation commands:
3132 $ gnatmake -c simple_cpp_interface
3135 $ gnatbind -n simple_cpp_interface
3136 $ gnatlink simple_cpp_interface -o cpp_main --LINK=g++
3137 -lstdc++ ex7.o cpp_main.o
3141 Here are the corresponding sources:
3149 void adainit (void);
3150 void adafinal (void);
3151 void method1 (A *t);
3173 class A : public Origin @{
3175 void method1 (void);
3176 void method2 (int v);
3186 extern "C" @{ void ada_method2 (A *t, int v);@}
3188 void A::method1 (void)
3191 printf ("in A::method1, a_value = %d \n",a_value);
3195 void A::method2 (int v)
3197 ada_method2 (this, v);
3198 printf ("in A::method2, a_value = %d \n",a_value);
3205 printf ("in A::A, a_value = %d \n",a_value);
3209 @smallexample @c ada
3211 package body Simple_Cpp_Interface is
3213 procedure Ada_Method2 (This : in out A; V : Integer) is
3219 end Simple_Cpp_Interface;
3222 package Simple_Cpp_Interface is
3225 Vptr : System.Address;
3229 pragma Convention (C, A);
3231 procedure Method1 (This : in out A);
3232 pragma Import (C, Method1);
3234 procedure Ada_Method2 (This : in out A; V : Integer);
3235 pragma Export (C, Ada_Method2);
3237 end Simple_Cpp_Interface;
3240 @node Interfacing with C++ constructors
3241 @subsection Interfacing with C++ constructors
3244 In order to interface with C++ constructors GNAT provides the
3245 @code{pragma CPP_Constructor} (@xref{Interfacing to C++,,,
3246 gnat_rm, GNAT Reference Manual}, for additional information).
3247 In this section we present some common uses of C++ constructors
3248 in mixed-languages programs in GNAT.
3250 Let us assume that we need to interface with the following
3258 @b{virtual} int Get_Value ();
3259 Root(); // Default constructor
3260 Root(int v); // 1st non-default constructor
3261 Root(int v, int w); // 2nd non-default constructor
3265 For this purpose we can write the following package spec (further
3266 information on how to build this spec is available in
3267 @ref{Interfacing with C++ at the Class Level} and
3268 @ref{Generating Ada Bindings for C and C++ headers}).
3270 @smallexample @c ada
3271 with Interfaces.C; use Interfaces.C;
3273 type Root is tagged limited record
3277 pragma Import (CPP, Root);
3279 function Get_Value (Obj : Root) return int;
3280 pragma Import (CPP, Get_Value);
3282 function Constructor return Root;
3283 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ev");
3285 function Constructor (v : Integer) return Root;
3286 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ei");
3288 function Constructor (v, w : Integer) return Root;
3289 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Eii");
3293 On the Ada side the constructor is represented by a function (whose
3294 name is arbitrary) that returns the classwide type corresponding to
3295 the imported C++ class. Although the constructor is described as a
3296 function, it is typically a procedure with an extra implicit argument
3297 (the object being initialized) at the implementation level. GNAT
3298 issues the appropriate call, whatever it is, to get the object
3299 properly initialized.
3301 Constructors can only appear in the following contexts:
3305 On the right side of an initialization of an object of type @var{T}.
3307 On the right side of an initialization of a record component of type @var{T}.
3309 In an Ada 2005 limited aggregate.
3311 In an Ada 2005 nested limited aggregate.
3313 In an Ada 2005 limited aggregate that initializes an object built in
3314 place by an extended return statement.
3318 In a declaration of an object whose type is a class imported from C++,
3319 either the default C++ constructor is implicitly called by GNAT, or
3320 else the required C++ constructor must be explicitly called in the
3321 expression that initializes the object. For example:
3323 @smallexample @c ada
3325 Obj2 : Root := Constructor;
3326 Obj3 : Root := Constructor (v => 10);
3327 Obj4 : Root := Constructor (30, 40);
3330 The first two declarations are equivalent: in both cases the default C++
3331 constructor is invoked (in the former case the call to the constructor is
3332 implicit, and in the latter case the call is explicit in the object
3333 declaration). @code{Obj3} is initialized by the C++ non-default constructor
3334 that takes an integer argument, and @code{Obj4} is initialized by the
3335 non-default C++ constructor that takes two integers.
3337 Let us derive the imported C++ class in the Ada side. For example:
3339 @smallexample @c ada
3340 type DT is new Root with record
3341 C_Value : Natural := 2009;
3345 In this case the components DT inherited from the C++ side must be
3346 initialized by a C++ constructor, and the additional Ada components
3347 of type DT are initialized by GNAT. The initialization of such an
3348 object is done either by default, or by means of a function returning
3349 an aggregate of type DT, or by means of an extension aggregate.
3351 @smallexample @c ada
3353 Obj6 : DT := Function_Returning_DT (50);
3354 Obj7 : DT := (Constructor (30,40) with C_Value => 50);
3357 The declaration of @code{Obj5} invokes the default constructors: the
3358 C++ default constructor of the parent type takes care of the initialization
3359 of the components inherited from Root, and GNAT takes care of the default
3360 initialization of the additional Ada components of type DT (that is,
3361 @code{C_Value} is initialized to value 2009). The order of invocation of
3362 the constructors is consistent with the order of elaboration required by
3363 Ada and C++. That is, the constructor of the parent type is always called
3364 before the constructor of the derived type.
3366 Let us now consider a record that has components whose type is imported
3367 from C++. For example:
3369 @smallexample @c ada
3370 type Rec1 is limited record
3371 Data1 : Root := Constructor (10);
3372 Value : Natural := 1000;
3375 type Rec2 (D : Integer := 20) is limited record
3377 Data2 : Root := Constructor (D, 30);
3381 The initialization of an object of type @code{Rec2} will call the
3382 non-default C++ constructors specified for the imported components.
3385 @smallexample @c ada
3389 Using Ada 2005 we can use limited aggregates to initialize an object
3390 invoking C++ constructors that differ from those specified in the type
3391 declarations. For example:
3393 @smallexample @c ada
3394 Obj9 : Rec2 := (Rec => (Data1 => Constructor (15, 16),
3399 The above declaration uses an Ada 2005 limited aggregate to
3400 initialize @code{Obj9}, and the C++ constructor that has two integer
3401 arguments is invoked to initialize the @code{Data1} component instead
3402 of the constructor specified in the declaration of type @code{Rec1}. In
3403 Ada 2005 the box in the aggregate indicates that unspecified components
3404 are initialized using the expression (if any) available in the component
3405 declaration. That is, in this case discriminant @code{D} is initialized
3406 to value @code{20}, @code{Value} is initialized to value 1000, and the
3407 non-default C++ constructor that handles two integers takes care of
3408 initializing component @code{Data2} with values @code{20,30}.
3410 In Ada 2005 we can use the extended return statement to build the Ada
3411 equivalent to C++ non-default constructors. For example:
3413 @smallexample @c ada
3414 function Constructor (V : Integer) return Rec2 is
3416 return Obj : Rec2 := (Rec => (Data1 => Constructor (V, 20),
3419 -- Further actions required for construction of
3420 -- objects of type Rec2
3426 In this example the extended return statement construct is used to
3427 build in place the returned object whose components are initialized
3428 by means of a limited aggregate. Any further action associated with
3429 the constructor can be placed inside the construct.
3431 @node Interfacing with C++ at the Class Level
3432 @subsection Interfacing with C++ at the Class Level
3434 In this section we demonstrate the GNAT features for interfacing with
3435 C++ by means of an example making use of Ada 2005 abstract interface
3436 types. This example consists of a classification of animals; classes
3437 have been used to model our main classification of animals, and
3438 interfaces provide support for the management of secondary
3439 classifications. We first demonstrate a case in which the types and
3440 constructors are defined on the C++ side and imported from the Ada
3441 side, and latter the reverse case.
3443 The root of our derivation will be the @code{Animal} class, with a
3444 single private attribute (the @code{Age} of the animal) and two public
3445 primitives to set and get the value of this attribute.
3450 @b{virtual} void Set_Age (int New_Age);
3451 @b{virtual} int Age ();
3457 Abstract interface types are defined in C++ by means of classes with pure
3458 virtual functions and no data members. In our example we will use two
3459 interfaces that provide support for the common management of @code{Carnivore}
3460 and @code{Domestic} animals:
3463 @b{class} Carnivore @{
3465 @b{virtual} int Number_Of_Teeth () = 0;
3468 @b{class} Domestic @{
3470 @b{virtual void} Set_Owner (char* Name) = 0;
3474 Using these declarations, we can now say that a @code{Dog} is an animal that is
3475 both Carnivore and Domestic, that is:
3478 @b{class} Dog : Animal, Carnivore, Domestic @{
3480 @b{virtual} int Number_Of_Teeth ();
3481 @b{virtual} void Set_Owner (char* Name);
3483 Dog(); // Constructor
3490 In the following examples we will assume that the previous declarations are
3491 located in a file named @code{animals.h}. The following package demonstrates
3492 how to import these C++ declarations from the Ada side:
3494 @smallexample @c ada
3495 with Interfaces.C.Strings; use Interfaces.C.Strings;
3497 type Carnivore is interface;
3498 pragma Convention (C_Plus_Plus, Carnivore);
3499 function Number_Of_Teeth (X : Carnivore)
3500 return Natural is abstract;
3502 type Domestic is interface;
3503 pragma Convention (C_Plus_Plus, Set_Owner);
3505 (X : in out Domestic;
3506 Name : Chars_Ptr) is abstract;
3508 type Animal is tagged record
3511 pragma Import (C_Plus_Plus, Animal);
3513 procedure Set_Age (X : in out Animal; Age : Integer);
3514 pragma Import (C_Plus_Plus, Set_Age);
3516 function Age (X : Animal) return Integer;
3517 pragma Import (C_Plus_Plus, Age);
3519 type Dog is new Animal and Carnivore and Domestic with record
3520 Tooth_Count : Natural;
3521 Owner : String (1 .. 30);
3523 pragma Import (C_Plus_Plus, Dog);
3525 function Number_Of_Teeth (A : Dog) return Integer;
3526 pragma Import (C_Plus_Plus, Number_Of_Teeth);
3528 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3529 pragma Import (C_Plus_Plus, Set_Owner);
3531 function New_Dog return Dog;
3532 pragma CPP_Constructor (New_Dog);
3533 pragma Import (CPP, New_Dog, "_ZN3DogC2Ev");
3537 Thanks to the compatibility between GNAT run-time structures and the C++ ABI,
3538 interfacing with these C++ classes is easy. The only requirement is that all
3539 the primitives and components must be declared exactly in the same order in
3542 Regarding the abstract interfaces, we must indicate to the GNAT compiler by
3543 means of a @code{pragma Convention (C_Plus_Plus)}, the convention used to pass
3544 the arguments to the called primitives will be the same as for C++. For the
3545 imported classes we use @code{pragma Import} with convention @code{C_Plus_Plus}
3546 to indicate that they have been defined on the C++ side; this is required
3547 because the dispatch table associated with these tagged types will be built
3548 in the C++ side and therefore will not contain the predefined Ada primitives
3549 which Ada would otherwise expect.
3551 As the reader can see there is no need to indicate the C++ mangled names
3552 associated with each subprogram because it is assumed that all the calls to
3553 these primitives will be dispatching calls. The only exception is the
3554 constructor, which must be registered with the compiler by means of
3555 @code{pragma CPP_Constructor} and needs to provide its associated C++
3556 mangled name because the Ada compiler generates direct calls to it.
3558 With the above packages we can now declare objects of type Dog on the Ada side
3559 and dispatch calls to the corresponding subprograms on the C++ side. We can
3560 also extend the tagged type Dog with further fields and primitives, and
3561 override some of its C++ primitives on the Ada side. For example, here we have
3562 a type derivation defined on the Ada side that inherits all the dispatching
3563 primitives of the ancestor from the C++ side.
3566 @b{with} Animals; @b{use} Animals;
3567 @b{package} Vaccinated_Animals @b{is}
3568 @b{type} Vaccinated_Dog @b{is new} Dog @b{with null record};
3569 @b{function} Vaccination_Expired (A : Vaccinated_Dog) @b{return} Boolean;
3570 @b{end} Vaccinated_Animals;
3573 It is important to note that, because of the ABI compatibility, the programmer
3574 does not need to add any further information to indicate either the object
3575 layout or the dispatch table entry associated with each dispatching operation.
3577 Now let us define all the types and constructors on the Ada side and export
3578 them to C++, using the same hierarchy of our previous example:
3580 @smallexample @c ada
3581 with Interfaces.C.Strings;
3582 use Interfaces.C.Strings;
3584 type Carnivore is interface;
3585 pragma Convention (C_Plus_Plus, Carnivore);
3586 function Number_Of_Teeth (X : Carnivore)
3587 return Natural is abstract;
3589 type Domestic is interface;
3590 pragma Convention (C_Plus_Plus, Set_Owner);
3592 (X : in out Domestic;
3593 Name : Chars_Ptr) is abstract;
3595 type Animal is tagged record
3598 pragma Convention (C_Plus_Plus, Animal);
3600 procedure Set_Age (X : in out Animal; Age : Integer);
3601 pragma Export (C_Plus_Plus, Set_Age);
3603 function Age (X : Animal) return Integer;
3604 pragma Export (C_Plus_Plus, Age);
3606 type Dog is new Animal and Carnivore and Domestic with record
3607 Tooth_Count : Natural;
3608 Owner : String (1 .. 30);
3610 pragma Convention (C_Plus_Plus, Dog);
3612 function Number_Of_Teeth (A : Dog) return Integer;
3613 pragma Export (C_Plus_Plus, Number_Of_Teeth);
3615 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3616 pragma Export (C_Plus_Plus, Set_Owner);
3618 function New_Dog return Dog'Class;
3619 pragma Export (C_Plus_Plus, New_Dog);
3623 Compared with our previous example the only difference is the use of
3624 @code{pragma Export} to indicate to the GNAT compiler that the primitives will
3625 be available to C++. Thanks to the ABI compatibility, on the C++ side there is
3626 nothing else to be done; as explained above, the only requirement is that all
3627 the primitives and components are declared in exactly the same order.
3629 For completeness, let us see a brief C++ main program that uses the
3630 declarations available in @code{animals.h} (presented in our first example) to
3631 import and use the declarations from the Ada side, properly initializing and
3632 finalizing the Ada run-time system along the way:
3635 @b{#include} "animals.h"
3636 @b{#include} <iostream>
3637 @b{using namespace} std;
3639 void Check_Carnivore (Carnivore *obj) @{@dots{}@}
3640 void Check_Domestic (Domestic *obj) @{@dots{}@}
3641 void Check_Animal (Animal *obj) @{@dots{}@}
3642 void Check_Dog (Dog *obj) @{@dots{}@}
3645 void adainit (void);
3646 void adafinal (void);
3652 Dog *obj = new_dog(); // Ada constructor
3653 Check_Carnivore (obj); // Check secondary DT
3654 Check_Domestic (obj); // Check secondary DT
3655 Check_Animal (obj); // Check primary DT
3656 Check_Dog (obj); // Check primary DT
3661 adainit (); test(); adafinal ();
3666 @node Comparison between GNAT and C/C++ Compilation Models
3667 @section Comparison between GNAT and C/C++ Compilation Models
3670 The GNAT model of compilation is close to the C and C++ models. You can
3671 think of Ada specs as corresponding to header files in C. As in C, you
3672 don't need to compile specs; they are compiled when they are used. The
3673 Ada @code{with} is similar in effect to the @code{#include} of a C
3676 One notable difference is that, in Ada, you may compile specs separately
3677 to check them for semantic and syntactic accuracy. This is not always
3678 possible with C headers because they are fragments of programs that have
3679 less specific syntactic or semantic rules.
3681 The other major difference is the requirement for running the binder,
3682 which performs two important functions. First, it checks for
3683 consistency. In C or C++, the only defense against assembling
3684 inconsistent programs lies outside the compiler, in a makefile, for
3685 example. The binder satisfies the Ada requirement that it be impossible
3686 to construct an inconsistent program when the compiler is used in normal
3689 @cindex Elaboration order control
3690 The other important function of the binder is to deal with elaboration
3691 issues. There are also elaboration issues in C++ that are handled
3692 automatically. This automatic handling has the advantage of being
3693 simpler to use, but the C++ programmer has no control over elaboration.
3694 Where @code{gnatbind} might complain there was no valid order of
3695 elaboration, a C++ compiler would simply construct a program that
3696 malfunctioned at run time.
3699 @node Comparison between GNAT and Conventional Ada Library Models
3700 @section Comparison between GNAT and Conventional Ada Library Models
3703 This section is intended for Ada programmers who have
3704 used an Ada compiler implementing the traditional Ada library
3705 model, as described in the Ada Reference Manual.
3707 @cindex GNAT library
3708 In GNAT, there is no ``library'' in the normal sense. Instead, the set of
3709 source files themselves acts as the library. Compiling Ada programs does
3710 not generate any centralized information, but rather an object file and
3711 a ALI file, which are of interest only to the binder and linker.
3712 In a traditional system, the compiler reads information not only from
3713 the source file being compiled, but also from the centralized library.
3714 This means that the effect of a compilation depends on what has been
3715 previously compiled. In particular:
3719 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3720 to the version of the unit most recently compiled into the library.
3723 Inlining is effective only if the necessary body has already been
3724 compiled into the library.
3727 Compiling a unit may obsolete other units in the library.
3731 In GNAT, compiling one unit never affects the compilation of any other
3732 units because the compiler reads only source files. Only changes to source
3733 files can affect the results of a compilation. In particular:
3737 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3738 to the source version of the unit that is currently accessible to the
3743 Inlining requires the appropriate source files for the package or
3744 subprogram bodies to be available to the compiler. Inlining is always
3745 effective, independent of the order in which units are complied.
3748 Compiling a unit never affects any other compilations. The editing of
3749 sources may cause previous compilations to be out of date if they
3750 depended on the source file being modified.
3754 The most important result of these differences is that order of compilation
3755 is never significant in GNAT. There is no situation in which one is
3756 required to do one compilation before another. What shows up as order of
3757 compilation requirements in the traditional Ada library becomes, in
3758 GNAT, simple source dependencies; in other words, there is only a set
3759 of rules saying what source files must be present when a file is
3763 @node Placement of temporary files
3764 @section Placement of temporary files
3765 @cindex Temporary files (user control over placement)
3768 GNAT creates temporary files in the directory designated by the environment
3769 variable @env{TMPDIR}.
3770 (See the HP @emph{C RTL Reference Manual} on the function @code{getenv()}
3771 for detailed information on how environment variables are resolved.
3772 For most users the easiest way to make use of this feature is to simply
3773 define @env{TMPDIR} as a job level logical name).
3774 For example, if you wish to use a Ramdisk (assuming DECRAM is installed)
3775 for compiler temporary files, then you can include something like the
3776 following command in your @file{LOGIN.COM} file:
3779 $ define/job TMPDIR "/disk$scratchram/000000/temp/"
3783 If @env{TMPDIR} is not defined, then GNAT uses the directory designated by
3784 @env{TMP}; if @env{TMP} is not defined, then GNAT uses the directory
3785 designated by @env{TEMP}.
3786 If none of these environment variables are defined then GNAT uses the
3787 directory designated by the logical name @code{SYS$SCRATCH:}
3788 (by default the user's home directory). If all else fails
3789 GNAT uses the current directory for temporary files.
3792 @c *************************
3793 @node Compiling Using gcc
3794 @chapter Compiling Using @command{gcc}
3797 This chapter discusses how to compile Ada programs using the @command{gcc}
3798 command. It also describes the set of switches
3799 that can be used to control the behavior of the compiler.
3801 * Compiling Programs::
3802 * Switches for gcc::
3803 * Search Paths and the Run-Time Library (RTL)::
3804 * Order of Compilation Issues::
3808 @node Compiling Programs
3809 @section Compiling Programs
3812 The first step in creating an executable program is to compile the units
3813 of the program using the @command{gcc} command. You must compile the
3818 the body file (@file{.adb}) for a library level subprogram or generic
3822 the spec file (@file{.ads}) for a library level package or generic
3823 package that has no body
3826 the body file (@file{.adb}) for a library level package
3827 or generic package that has a body
3832 You need @emph{not} compile the following files
3837 the spec of a library unit which has a body
3844 because they are compiled as part of compiling related units. GNAT
3846 when the corresponding body is compiled, and subunits when the parent is
3849 @cindex cannot generate code
3850 If you attempt to compile any of these files, you will get one of the
3851 following error messages (where @var{fff} is the name of the file you compiled):
3854 cannot generate code for file @var{fff} (package spec)
3855 to check package spec, use -gnatc
3857 cannot generate code for file @var{fff} (missing subunits)
3858 to check parent unit, use -gnatc
3860 cannot generate code for file @var{fff} (subprogram spec)
3861 to check subprogram spec, use -gnatc
3863 cannot generate code for file @var{fff} (subunit)
3864 to check subunit, use -gnatc
3868 As indicated by the above error messages, if you want to submit
3869 one of these files to the compiler to check for correct semantics
3870 without generating code, then use the @option{-gnatc} switch.
3872 The basic command for compiling a file containing an Ada unit is
3875 $ gcc -c @ovar{switches} @file{file name}
3879 where @var{file name} is the name of the Ada file (usually
3881 @file{.ads} for a spec or @file{.adb} for a body).
3884 @option{-c} switch to tell @command{gcc} to compile, but not link, the file.
3886 The result of a successful compilation is an object file, which has the
3887 same name as the source file but an extension of @file{.o} and an Ada
3888 Library Information (ALI) file, which also has the same name as the
3889 source file, but with @file{.ali} as the extension. GNAT creates these
3890 two output files in the current directory, but you may specify a source
3891 file in any directory using an absolute or relative path specification
3892 containing the directory information.
3895 @command{gcc} is actually a driver program that looks at the extensions of
3896 the file arguments and loads the appropriate compiler. For example, the
3897 GNU C compiler is @file{cc1}, and the Ada compiler is @file{gnat1}.
3898 These programs are in directories known to the driver program (in some
3899 configurations via environment variables you set), but need not be in
3900 your path. The @command{gcc} driver also calls the assembler and any other
3901 utilities needed to complete the generation of the required object
3904 It is possible to supply several file names on the same @command{gcc}
3905 command. This causes @command{gcc} to call the appropriate compiler for
3906 each file. For example, the following command lists three separate
3907 files to be compiled:
3910 $ gcc -c x.adb y.adb z.c
3914 calls @code{gnat1} (the Ada compiler) twice to compile @file{x.adb} and
3915 @file{y.adb}, and @code{cc1} (the C compiler) once to compile @file{z.c}.
3916 The compiler generates three object files @file{x.o}, @file{y.o} and
3917 @file{z.o} and the two ALI files @file{x.ali} and @file{y.ali} from the
3918 Ada compilations. Any switches apply to all the files ^listed,^listed.^
3921 @option{-gnat@var{x}} switches, which apply only to Ada compilations.
3924 @node Switches for gcc
3925 @section Switches for @command{gcc}
3928 The @command{gcc} command accepts switches that control the
3929 compilation process. These switches are fully described in this section.
3930 First we briefly list all the switches, in alphabetical order, then we
3931 describe the switches in more detail in functionally grouped sections.
3933 More switches exist for GCC than those documented here, especially
3934 for specific targets. However, their use is not recommended as
3935 they may change code generation in ways that are incompatible with
3936 the Ada run-time library, or can cause inconsistencies between
3940 * Output and Error Message Control::
3941 * Warning Message Control::
3942 * Debugging and Assertion Control::
3943 * Validity Checking::
3946 * Using gcc for Syntax Checking::
3947 * Using gcc for Semantic Checking::
3948 * Compiling Different Versions of Ada::
3949 * Character Set Control::
3950 * File Naming Control::
3951 * Subprogram Inlining Control::
3952 * Auxiliary Output Control::
3953 * Debugging Control::
3954 * Exception Handling Control::
3955 * Units to Sources Mapping Files::
3956 * Integrated Preprocessing::
3957 * Code Generation Control::
3966 @cindex @option{-b} (@command{gcc})
3967 @item -b @var{target}
3968 Compile your program to run on @var{target}, which is the name of a
3969 system configuration. You must have a GNAT cross-compiler built if
3970 @var{target} is not the same as your host system.
3973 @cindex @option{-B} (@command{gcc})
3974 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
3975 from @var{dir} instead of the default location. Only use this switch
3976 when multiple versions of the GNAT compiler are available.
3977 @xref{Directory Options,, Options for Directory Search, gcc, Using the
3978 GNU Compiler Collection (GCC)}, for further details. You would normally
3979 use the @option{-b} or @option{-V} switch instead.
3982 @cindex @option{-c} (@command{gcc})
3983 Compile. Always use this switch when compiling Ada programs.
3985 Note: for some other languages when using @command{gcc}, notably in
3986 the case of C and C++, it is possible to use
3987 use @command{gcc} without a @option{-c} switch to
3988 compile and link in one step. In the case of GNAT, you
3989 cannot use this approach, because the binder must be run
3990 and @command{gcc} cannot be used to run the GNAT binder.
3994 @cindex @option{-fno-inline} (@command{gcc})
3995 Suppresses all back-end inlining, even if other optimization or inlining
3997 This includes suppression of inlining that results
3998 from the use of the pragma @code{Inline_Always}.
3999 Any occurrences of pragma @code{Inline} or @code{Inline_Always}
4000 are ignored, and @option{-gnatn} and @option{-gnatN} have no
4001 effect if this switch is present.
4003 @item -fno-inline-functions
4004 @cindex @option{-fno-inline-functions} (@command{gcc})
4005 Suppresses automatic inlining of simple subprograms, which is enabled
4006 if @option{-O3} is used.
4008 @item -fno-inline-small-functions
4009 @cindex @option{-fno-inline-small-functions} (@command{gcc})
4010 Suppresses automatic inlining of small subprograms, which is enabled
4011 if @option{-O2} is used.
4013 @item -fno-inline-functions-called-once
4014 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
4015 Suppresses inlining of subprograms local to the unit and called once
4016 from within it, which is enabled if @option{-O1} is used.
4019 @cindex @option{-fno-ivopts} (@command{gcc})
4020 Suppresses high-level loop induction variable optimizations, which are
4021 enabled if @option{-O1} is used. These optimizations are generally
4022 profitable but, for some specific cases of loops with numerous uses
4023 of the iteration variable that follow a common pattern, they may end
4024 up destroying the regularity that could be exploited at a lower level
4025 and thus producing inferior code.
4027 @item -fno-strict-aliasing
4028 @cindex @option{-fno-strict-aliasing} (@command{gcc})
4029 Causes the compiler to avoid assumptions regarding non-aliasing
4030 of objects of different types. See
4031 @ref{Optimization and Strict Aliasing} for details.
4034 @cindex @option{-fstack-check} (@command{gcc})
4035 Activates stack checking.
4036 See @ref{Stack Overflow Checking} for details.
4039 @cindex @option{-fstack-usage} (@command{gcc})
4040 Makes the compiler output stack usage information for the program, on a
4041 per-function basis. See @ref{Static Stack Usage Analysis} for details.
4043 @item -fcallgraph-info@r{[}=su@r{]}
4044 @cindex @option{-fcallgraph-info} (@command{gcc})
4045 Makes the compiler output callgraph information for the program, on a
4046 per-file basis. The information is generated in the VCG format. It can
4047 be decorated with stack-usage per-node information.
4050 @cindex @option{^-g^/DEBUG^} (@command{gcc})
4051 Generate debugging information. This information is stored in the object
4052 file and copied from there to the final executable file by the linker,
4053 where it can be read by the debugger. You must use the
4054 @option{^-g^/DEBUG^} switch if you plan on using the debugger.
4057 @cindex @option{-gnat83} (@command{gcc})
4058 Enforce Ada 83 restrictions.
4061 @cindex @option{-gnat95} (@command{gcc})
4062 Enforce Ada 95 restrictions.
4065 @cindex @option{-gnat05} (@command{gcc})
4066 Allow full Ada 2005 features.
4069 @cindex @option{-gnata} (@command{gcc})
4070 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
4071 activated. Note that these pragmas can also be controlled using the
4072 configuration pragmas @code{Assertion_Policy} and @code{Debug_Policy}.
4073 It also activates pragmas @code{Check}, @code{Precondition}, and
4074 @code{Postcondition}. Note that these pragmas can also be controlled
4075 using the configuration pragma @code{Check_Policy}.
4078 @cindex @option{-gnatA} (@command{gcc})
4079 Avoid processing @file{gnat.adc}. If a @file{gnat.adc} file is present,
4083 @cindex @option{-gnatb} (@command{gcc})
4084 Generate brief messages to @file{stderr} even if verbose mode set.
4087 @cindex @option{-gnatB} (@command{gcc})
4088 Assume no invalid (bad) values except for 'Valid attribute use
4089 (@pxref{Validity Checking}).
4092 @cindex @option{-gnatc} (@command{gcc})
4093 Check syntax and semantics only (no code generation attempted).
4096 @cindex @option{-gnatC} (@command{gcc})
4097 Generate CodePeer information (no code generation attempted).
4098 This switch will generate an intermediate representation suitable for
4099 use by CodePeer (@file{.scil} files). This switch is not compatible with
4100 code generation (it will, among other things, disable some switches such
4101 as -gnatn, and enable others such as -gnata).
4104 @cindex @option{-gnatd} (@command{gcc})
4105 Specify debug options for the compiler. The string of characters after
4106 the @option{-gnatd} specify the specific debug options. The possible
4107 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
4108 compiler source file @file{debug.adb} for details of the implemented
4109 debug options. Certain debug options are relevant to applications
4110 programmers, and these are documented at appropriate points in this
4115 @cindex @option{-gnatD[nn]} (@command{gcc})
4118 @item /XDEBUG /LXDEBUG=nnn
4120 Create expanded source files for source level debugging. This switch
4121 also suppress generation of cross-reference information
4122 (see @option{-gnatx}).
4124 @item -gnatec=@var{path}
4125 @cindex @option{-gnatec} (@command{gcc})
4126 Specify a configuration pragma file
4128 (the equal sign is optional)
4130 (@pxref{The Configuration Pragmas Files}).
4132 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=@var{value}@r{]}
4133 @cindex @option{-gnateD} (@command{gcc})
4134 Defines a symbol, associated with @var{value}, for preprocessing.
4135 (@pxref{Integrated Preprocessing}).
4138 @cindex @option{-gnatef} (@command{gcc})
4139 Display full source path name in brief error messages.
4142 @cindex @option{-gnateG} (@command{gcc})
4143 Save result of preprocessing in a text file.
4145 @item -gnatem=@var{path}
4146 @cindex @option{-gnatem} (@command{gcc})
4147 Specify a mapping file
4149 (the equal sign is optional)
4151 (@pxref{Units to Sources Mapping Files}).
4153 @item -gnatep=@var{file}
4154 @cindex @option{-gnatep} (@command{gcc})
4155 Specify a preprocessing data file
4157 (the equal sign is optional)
4159 (@pxref{Integrated Preprocessing}).
4162 @cindex @option{-gnateS} (@command{gcc})
4163 Generate SCO (Source Coverage Obligation) information in the ALI
4164 file. This information is used by advanced coverage tools. See
4165 unit @file{SCOs} in the compiler sources for details in files
4166 @file{scos.ads} and @file{scos.adb}.
4169 @cindex @option{-gnatE} (@command{gcc})
4170 Full dynamic elaboration checks.
4173 @cindex @option{-gnatf} (@command{gcc})
4174 Full errors. Multiple errors per line, all undefined references, do not
4175 attempt to suppress cascaded errors.
4178 @cindex @option{-gnatF} (@command{gcc})
4179 Externals names are folded to all uppercase.
4181 @item ^-gnatg^/GNAT_INTERNAL^
4182 @cindex @option{^-gnatg^/GNAT_INTERNAL^} (@command{gcc})
4183 Internal GNAT implementation mode. This should not be used for
4184 applications programs, it is intended only for use by the compiler
4185 and its run-time library. For documentation, see the GNAT sources.
4186 Note that @option{^-gnatg^/GNAT_INTERNAL^} implies
4187 @option{^-gnatwae^/WARNINGS=ALL,ERRORS^} and
4188 @option{^-gnatyg^/STYLE_CHECKS=GNAT^}
4189 so that all standard warnings and all standard style options are turned on.
4190 All warnings and style error messages are treated as errors.
4194 @cindex @option{-gnatG[nn]} (@command{gcc})
4197 @item /EXPAND_SOURCE, /LEXPAND_SOURCE=nnn
4199 List generated expanded code in source form.
4201 @item ^-gnath^/HELP^
4202 @cindex @option{^-gnath^/HELP^} (@command{gcc})
4203 Output usage information. The output is written to @file{stdout}.
4205 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
4206 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
4207 Identifier character set
4209 (@var{c}=1/2/3/4/8/9/p/f/n/w).
4211 For details of the possible selections for @var{c},
4212 see @ref{Character Set Control}.
4214 @item ^-gnatI^/IGNORE_REP_CLAUSES^
4215 @cindex @option{^-gnatI^IGNORE_REP_CLAUSES^} (@command{gcc})
4216 Ignore representation clauses. When this switch is used,
4217 representation clauses are treated as comments. This is useful
4218 when initially porting code where you want to ignore rep clause
4219 problems, and also for compiling foreign code (particularly
4220 for use with ASIS). The representation clauses that are ignored
4221 are: enumeration_representation_clause, record_representation_clause,
4222 and attribute_definition_clause for the following attributes:
4223 Address, Alignment, Bit_Order, Component_Size, Machine_Radix,
4224 Object_Size, Size, Small, Stream_Size, and Value_Size.
4225 Note that this option should be used only for compiling -- the
4226 code is likely to malfunction at run time.
4229 @cindex @option{-gnatjnn} (@command{gcc})
4230 Reformat error messages to fit on nn character lines
4232 @item -gnatk=@var{n}
4233 @cindex @option{-gnatk} (@command{gcc})
4234 Limit file names to @var{n} (1-999) characters ^(@code{k} = krunch)^^.
4237 @cindex @option{-gnatl} (@command{gcc})
4238 Output full source listing with embedded error messages.
4241 @cindex @option{-gnatL} (@command{gcc})
4242 Used in conjunction with -gnatG or -gnatD to intersperse original
4243 source lines (as comment lines with line numbers) in the expanded
4246 @item -gnatm=@var{n}
4247 @cindex @option{-gnatm} (@command{gcc})
4248 Limit number of detected error or warning messages to @var{n}
4249 where @var{n} is in the range 1..999999. The default setting if
4250 no switch is given is 9999. If the number of warnings reaches this
4251 limit, then a message is output and further warnings are suppressed,
4252 but the compilation is continued. If the number of error messages
4253 reaches this limit, then a message is output and the compilation
4254 is abandoned. The equal sign here is optional. A value of zero
4255 means that no limit applies.
4258 @cindex @option{-gnatn} (@command{gcc})
4259 Activate inlining for subprograms for which
4260 pragma @code{inline} is specified. This inlining is performed
4261 by the GCC back-end.
4264 @cindex @option{-gnatN} (@command{gcc})
4265 Activate front end inlining for subprograms for which
4266 pragma @code{Inline} is specified. This inlining is performed
4267 by the front end and will be visible in the
4268 @option{-gnatG} output.
4270 When using a gcc-based back end (in practice this means using any version
4271 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
4272 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
4273 Historically front end inlining was more extensive than the gcc back end
4274 inlining, but that is no longer the case.
4277 @cindex @option{-gnato} (@command{gcc})
4278 Enable numeric overflow checking (which is not normally enabled by
4279 default). Note that division by zero is a separate check that is not
4280 controlled by this switch (division by zero checking is on by default).
4283 @cindex @option{-gnatp} (@command{gcc})
4284 Suppress all checks. See @ref{Run-Time Checks} for details.
4287 @cindex @option{-gnatP} (@command{gcc})
4288 Enable polling. This is required on some systems (notably Windows NT) to
4289 obtain asynchronous abort and asynchronous transfer of control capability.
4290 @xref{Pragma Polling,,, gnat_rm, GNAT Reference Manual}, for full
4294 @cindex @option{-gnatq} (@command{gcc})
4295 Don't quit. Try semantics, even if parse errors.
4298 @cindex @option{-gnatQ} (@command{gcc})
4299 Don't quit. Generate @file{ALI} and tree files even if illegalities.
4302 @cindex @option{-gnatr} (@command{gcc})
4303 Treat pragma Restrictions as Restriction_Warnings.
4305 @item ^-gnatR@r{[}0@r{/}1@r{/}2@r{/}3@r{[}s@r{]]}^/REPRESENTATION_INFO^
4306 @cindex @option{-gnatR} (@command{gcc})
4307 Output representation information for declared types and objects.
4310 @cindex @option{-gnats} (@command{gcc})
4314 @cindex @option{-gnatS} (@command{gcc})
4315 Print package Standard.
4318 @cindex @option{-gnatt} (@command{gcc})
4319 Generate tree output file.
4321 @item ^-gnatT^/TABLE_MULTIPLIER=^@var{nnn}
4322 @cindex @option{^-gnatT^/TABLE_MULTIPLIER^} (@command{gcc})
4323 All compiler tables start at @var{nnn} times usual starting size.
4326 @cindex @option{-gnatu} (@command{gcc})
4327 List units for this compilation.
4330 @cindex @option{-gnatU} (@command{gcc})
4331 Tag all error messages with the unique string ``error:''
4334 @cindex @option{-gnatv} (@command{gcc})
4335 Verbose mode. Full error output with source lines to @file{stdout}.
4338 @cindex @option{-gnatV} (@command{gcc})
4339 Control level of validity checking (@pxref{Validity Checking}).
4341 @item ^-gnatw@var{xxx}^/WARNINGS=(@var{option}@r{[},@dots{}@r{]})^
4342 @cindex @option{^-gnatw^/WARNINGS^} (@command{gcc})
4344 ^@var{xxx} is a string of option letters that^the list of options^ denotes
4345 the exact warnings that
4346 are enabled or disabled (@pxref{Warning Message Control}).
4348 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
4349 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
4350 Wide character encoding method
4352 (@var{e}=n/h/u/s/e/8).
4355 (@var{e}=@code{BRACKETS, NONE, HEX, UPPER, SHIFT_JIS, EUC, UTF8})
4359 @cindex @option{-gnatx} (@command{gcc})
4360 Suppress generation of cross-reference information.
4362 @item ^-gnaty^/STYLE_CHECKS=(option,option@dots{})^
4363 @cindex @option{^-gnaty^/STYLE_CHECKS^} (@command{gcc})
4364 Enable built-in style checks (@pxref{Style Checking}).
4366 @item ^-gnatz^/DISTRIBUTION_STUBS=^@var{m}
4367 @cindex @option{^-gnatz^/DISTRIBUTION_STUBS^} (@command{gcc})
4368 Distribution stub generation and compilation
4370 (@var{m}=r/c for receiver/caller stubs).
4373 (@var{m}=@code{RECEIVER} or @code{CALLER} to specify the type of stubs
4374 to be generated and compiled).
4377 @item ^-I^/SEARCH=^@var{dir}
4378 @cindex @option{^-I^/SEARCH^} (@command{gcc})
4380 Direct GNAT to search the @var{dir} directory for source files needed by
4381 the current compilation
4382 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4384 @item ^-I-^/NOCURRENT_DIRECTORY^
4385 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gcc})
4387 Except for the source file named in the command line, do not look for source
4388 files in the directory containing the source file named in the command line
4389 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4393 @cindex @option{-mbig-switch} (@command{gcc})
4394 @cindex @code{case} statement (effect of @option{-mbig-switch} option)
4395 This standard gcc switch causes the compiler to use larger offsets in its
4396 jump table representation for @code{case} statements.
4397 This may result in less efficient code, but is sometimes necessary
4398 (for example on HP-UX targets)
4399 @cindex HP-UX and @option{-mbig-switch} option
4400 in order to compile large and/or nested @code{case} statements.
4403 @cindex @option{-o} (@command{gcc})
4404 This switch is used in @command{gcc} to redirect the generated object file
4405 and its associated ALI file. Beware of this switch with GNAT, because it may
4406 cause the object file and ALI file to have different names which in turn
4407 may confuse the binder and the linker.
4411 @cindex @option{-nostdinc} (@command{gcc})
4412 Inhibit the search of the default location for the GNAT Run Time
4413 Library (RTL) source files.
4416 @cindex @option{-nostdlib} (@command{gcc})
4417 Inhibit the search of the default location for the GNAT Run Time
4418 Library (RTL) ALI files.
4422 @cindex @option{-O} (@command{gcc})
4423 @var{n} controls the optimization level.
4427 No optimization, the default setting if no @option{-O} appears
4430 Normal optimization, the default if you specify @option{-O} without
4431 an operand. A good compromise between code quality and compilation
4435 Extensive optimization, may improve execution time, possibly at the cost of
4436 substantially increased compilation time.
4439 Same as @option{-O2}, and also includes inline expansion for small subprograms
4443 Optimize space usage
4447 See also @ref{Optimization Levels}.
4452 @cindex @option{/NOOPTIMIZE} (@code{GNAT COMPILE})
4453 Equivalent to @option{/OPTIMIZE=NONE}.
4454 This is the default behavior in the absence of an @option{/OPTIMIZE}
4457 @item /OPTIMIZE@r{[}=(keyword@r{[},@dots{}@r{]})@r{]}
4458 @cindex @option{/OPTIMIZE} (@code{GNAT COMPILE})
4459 Selects the level of optimization for your program. The supported
4460 keywords are as follows:
4463 Perform most optimizations, including those that
4465 This is the default if the @option{/OPTIMIZE} qualifier is supplied
4466 without keyword options.
4469 Do not do any optimizations. Same as @code{/NOOPTIMIZE}.
4472 Perform some optimizations, but omit ones that are costly.
4475 Same as @code{SOME}.
4478 Full optimization as in @option{/OPTIMIZE=ALL}, and also attempts
4479 automatic inlining of small subprograms within a unit
4482 Try to unroll loops. This keyword may be specified together with
4483 any keyword above other than @code{NONE}. Loop unrolling
4484 usually, but not always, improves the performance of programs.
4487 Optimize space usage
4491 See also @ref{Optimization Levels}.
4495 @item -pass-exit-codes
4496 @cindex @option{-pass-exit-codes} (@command{gcc})
4497 Catch exit codes from the compiler and use the most meaningful as
4501 @item --RTS=@var{rts-path}
4502 @cindex @option{--RTS} (@command{gcc})
4503 Specifies the default location of the runtime library. Same meaning as the
4504 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
4507 @cindex @option{^-S^/ASM^} (@command{gcc})
4508 ^Used in place of @option{-c} to^Used to^
4509 cause the assembler source file to be
4510 generated, using @file{^.s^.S^} as the extension,
4511 instead of the object file.
4512 This may be useful if you need to examine the generated assembly code.
4514 @item ^-fverbose-asm^/VERBOSE_ASM^
4515 @cindex @option{^-fverbose-asm^/VERBOSE_ASM^} (@command{gcc})
4516 ^Used in conjunction with @option{-S}^Used in place of @option{/ASM}^
4517 to cause the generated assembly code file to be annotated with variable
4518 names, making it significantly easier to follow.
4521 @cindex @option{^-v^/VERBOSE^} (@command{gcc})
4522 Show commands generated by the @command{gcc} driver. Normally used only for
4523 debugging purposes or if you need to be sure what version of the
4524 compiler you are executing.
4528 @cindex @option{-V} (@command{gcc})
4529 Execute @var{ver} version of the compiler. This is the @command{gcc}
4530 version, not the GNAT version.
4533 @item ^-w^/NO_BACK_END_WARNINGS^
4534 @cindex @option{-w} (@command{gcc})
4535 Turn off warnings generated by the back end of the compiler. Use of
4536 this switch also causes the default for front end warnings to be set
4537 to suppress (as though @option{-gnatws} had appeared at the start of
4543 @c Combining qualifiers does not work on VMS
4544 You may combine a sequence of GNAT switches into a single switch. For
4545 example, the combined switch
4547 @cindex Combining GNAT switches
4553 is equivalent to specifying the following sequence of switches:
4556 -gnato -gnatf -gnati3
4561 The following restrictions apply to the combination of switches
4566 The switch @option{-gnatc} if combined with other switches must come
4567 first in the string.
4570 The switch @option{-gnats} if combined with other switches must come
4571 first in the string.
4575 @option{^-gnatz^/DISTRIBUTION_STUBS^}, @option{-gnatzc}, and @option{-gnatzr}
4576 may not be combined with any other switches.
4580 Once a ``y'' appears in the string (that is a use of the @option{-gnaty}
4581 switch), then all further characters in the switch are interpreted
4582 as style modifiers (see description of @option{-gnaty}).
4585 Once a ``d'' appears in the string (that is a use of the @option{-gnatd}
4586 switch), then all further characters in the switch are interpreted
4587 as debug flags (see description of @option{-gnatd}).
4590 Once a ``w'' appears in the string (that is a use of the @option{-gnatw}
4591 switch), then all further characters in the switch are interpreted
4592 as warning mode modifiers (see description of @option{-gnatw}).
4595 Once a ``V'' appears in the string (that is a use of the @option{-gnatV}
4596 switch), then all further characters in the switch are interpreted
4597 as validity checking options (@pxref{Validity Checking}).
4601 @node Output and Error Message Control
4602 @subsection Output and Error Message Control
4606 The standard default format for error messages is called ``brief format''.
4607 Brief format messages are written to @file{stderr} (the standard error
4608 file) and have the following form:
4611 e.adb:3:04: Incorrect spelling of keyword "function"
4612 e.adb:4:20: ";" should be "is"
4616 The first integer after the file name is the line number in the file,
4617 and the second integer is the column number within the line.
4619 @code{GPS} can parse the error messages
4620 and point to the referenced character.
4622 The following switches provide control over the error message
4628 @cindex @option{-gnatv} (@command{gcc})
4631 The v stands for verbose.
4633 The effect of this setting is to write long-format error
4634 messages to @file{stdout} (the standard output file.
4635 The same program compiled with the
4636 @option{-gnatv} switch would generate:
4640 3. funcion X (Q : Integer)
4642 >>> Incorrect spelling of keyword "function"
4645 >>> ";" should be "is"
4650 The vertical bar indicates the location of the error, and the @samp{>>>}
4651 prefix can be used to search for error messages. When this switch is
4652 used the only source lines output are those with errors.
4655 @cindex @option{-gnatl} (@command{gcc})
4657 The @code{l} stands for list.
4659 This switch causes a full listing of
4660 the file to be generated. In the case where a body is
4661 compiled, the corresponding spec is also listed, along
4662 with any subunits. Typical output from compiling a package
4663 body @file{p.adb} might look like:
4665 @smallexample @c ada
4669 1. package body p is
4671 3. procedure a is separate;
4682 2. pragma Elaborate_Body
4706 When you specify the @option{-gnatv} or @option{-gnatl} switches and
4707 standard output is redirected, a brief summary is written to
4708 @file{stderr} (standard error) giving the number of error messages and
4709 warning messages generated.
4711 @item -^gnatl^OUTPUT_FILE^=file
4712 @cindex @option{^-gnatl^OUTPUT_FILE^=fname} (@command{gcc})
4713 This has the same effect as @option{-gnatl} except that the output is
4714 written to a file instead of to standard output. If the given name
4715 @file{fname} does not start with a period, then it is the full name
4716 of the file to be written. If @file{fname} is an extension, it is
4717 appended to the name of the file being compiled. For example, if
4718 file @file{xyz.adb} is compiled with @option{^-gnatl^OUTPUT_FILE^=.lst},
4719 then the output is written to file ^xyz.adb.lst^xyz.adb_lst^.
4722 @cindex @option{-gnatU} (@command{gcc})
4723 This switch forces all error messages to be preceded by the unique
4724 string ``error:''. This means that error messages take a few more
4725 characters in space, but allows easy searching for and identification
4729 @cindex @option{-gnatb} (@command{gcc})
4731 The @code{b} stands for brief.
4733 This switch causes GNAT to generate the
4734 brief format error messages to @file{stderr} (the standard error
4735 file) as well as the verbose
4736 format message or full listing (which as usual is written to
4737 @file{stdout} (the standard output file).
4739 @item -gnatm=@var{n}
4740 @cindex @option{-gnatm} (@command{gcc})
4742 The @code{m} stands for maximum.
4744 @var{n} is a decimal integer in the
4745 range of 1 to 999999 and limits the number of error or warning
4746 messages to be generated. For example, using
4747 @option{-gnatm2} might yield
4750 e.adb:3:04: Incorrect spelling of keyword "function"
4751 e.adb:5:35: missing ".."
4752 fatal error: maximum number of errors detected
4753 compilation abandoned
4757 The default setting if
4758 no switch is given is 9999. If the number of warnings reaches this
4759 limit, then a message is output and further warnings are suppressed,
4760 but the compilation is continued. If the number of error messages
4761 reaches this limit, then a message is output and the compilation
4762 is abandoned. A value of zero means that no limit applies.
4765 Note that the equal sign is optional, so the switches
4766 @option{-gnatm2} and @option{-gnatm=2} are equivalent.
4769 @cindex @option{-gnatf} (@command{gcc})
4770 @cindex Error messages, suppressing
4772 The @code{f} stands for full.
4774 Normally, the compiler suppresses error messages that are likely to be
4775 redundant. This switch causes all error
4776 messages to be generated. In particular, in the case of
4777 references to undefined variables. If a given variable is referenced
4778 several times, the normal format of messages is
4780 e.adb:7:07: "V" is undefined (more references follow)
4784 where the parenthetical comment warns that there are additional
4785 references to the variable @code{V}. Compiling the same program with the
4786 @option{-gnatf} switch yields
4789 e.adb:7:07: "V" is undefined
4790 e.adb:8:07: "V" is undefined
4791 e.adb:8:12: "V" is undefined
4792 e.adb:8:16: "V" is undefined
4793 e.adb:9:07: "V" is undefined
4794 e.adb:9:12: "V" is undefined
4798 The @option{-gnatf} switch also generates additional information for
4799 some error messages. Some examples are:
4803 Details on possibly non-portable unchecked conversion
4805 List possible interpretations for ambiguous calls
4807 Additional details on incorrect parameters
4811 @cindex @option{-gnatjnn} (@command{gcc})
4812 In normal operation mode (or if @option{-gnatj0} is used, then error messages
4813 with continuation lines are treated as though the continuation lines were
4814 separate messages (and so a warning with two continuation lines counts as
4815 three warnings, and is listed as three separate messages).
4817 If the @option{-gnatjnn} switch is used with a positive value for nn, then
4818 messages are output in a different manner. A message and all its continuation
4819 lines are treated as a unit, and count as only one warning or message in the
4820 statistics totals. Furthermore, the message is reformatted so that no line
4821 is longer than nn characters.
4824 @cindex @option{-gnatq} (@command{gcc})
4826 The @code{q} stands for quit (really ``don't quit'').
4828 In normal operation mode, the compiler first parses the program and
4829 determines if there are any syntax errors. If there are, appropriate
4830 error messages are generated and compilation is immediately terminated.
4832 GNAT to continue with semantic analysis even if syntax errors have been
4833 found. This may enable the detection of more errors in a single run. On
4834 the other hand, the semantic analyzer is more likely to encounter some
4835 internal fatal error when given a syntactically invalid tree.
4838 @cindex @option{-gnatQ} (@command{gcc})
4839 In normal operation mode, the @file{ALI} file is not generated if any
4840 illegalities are detected in the program. The use of @option{-gnatQ} forces
4841 generation of the @file{ALI} file. This file is marked as being in
4842 error, so it cannot be used for binding purposes, but it does contain
4843 reasonably complete cross-reference information, and thus may be useful
4844 for use by tools (e.g., semantic browsing tools or integrated development
4845 environments) that are driven from the @file{ALI} file. This switch
4846 implies @option{-gnatq}, since the semantic phase must be run to get a
4847 meaningful ALI file.
4849 In addition, if @option{-gnatt} is also specified, then the tree file is
4850 generated even if there are illegalities. It may be useful in this case
4851 to also specify @option{-gnatq} to ensure that full semantic processing
4852 occurs. The resulting tree file can be processed by ASIS, for the purpose
4853 of providing partial information about illegal units, but if the error
4854 causes the tree to be badly malformed, then ASIS may crash during the
4857 When @option{-gnatQ} is used and the generated @file{ALI} file is marked as
4858 being in error, @command{gnatmake} will attempt to recompile the source when it
4859 finds such an @file{ALI} file, including with switch @option{-gnatc}.
4861 Note that @option{-gnatQ} has no effect if @option{-gnats} is specified,
4862 since ALI files are never generated if @option{-gnats} is set.
4866 @node Warning Message Control
4867 @subsection Warning Message Control
4868 @cindex Warning messages
4870 In addition to error messages, which correspond to illegalities as defined
4871 in the Ada Reference Manual, the compiler detects two kinds of warning
4874 First, the compiler considers some constructs suspicious and generates a
4875 warning message to alert you to a possible error. Second, if the
4876 compiler detects a situation that is sure to raise an exception at
4877 run time, it generates a warning message. The following shows an example
4878 of warning messages:
4880 e.adb:4:24: warning: creation of object may raise Storage_Error
4881 e.adb:10:17: warning: static value out of range
4882 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
4886 GNAT considers a large number of situations as appropriate
4887 for the generation of warning messages. As always, warnings are not
4888 definite indications of errors. For example, if you do an out-of-range
4889 assignment with the deliberate intention of raising a
4890 @code{Constraint_Error} exception, then the warning that may be
4891 issued does not indicate an error. Some of the situations for which GNAT
4892 issues warnings (at least some of the time) are given in the following
4893 list. This list is not complete, and new warnings are often added to
4894 subsequent versions of GNAT. The list is intended to give a general idea
4895 of the kinds of warnings that are generated.
4899 Possible infinitely recursive calls
4902 Out-of-range values being assigned
4905 Possible order of elaboration problems
4908 Assertions (pragma Assert) that are sure to fail
4914 Address clauses with possibly unaligned values, or where an attempt is
4915 made to overlay a smaller variable with a larger one.
4918 Fixed-point type declarations with a null range
4921 Direct_IO or Sequential_IO instantiated with a type that has access values
4924 Variables that are never assigned a value
4927 Variables that are referenced before being initialized
4930 Task entries with no corresponding @code{accept} statement
4933 Duplicate accepts for the same task entry in a @code{select}
4936 Objects that take too much storage
4939 Unchecked conversion between types of differing sizes
4942 Missing @code{return} statement along some execution path in a function
4945 Incorrect (unrecognized) pragmas
4948 Incorrect external names
4951 Allocation from empty storage pool
4954 Potentially blocking operation in protected type
4957 Suspicious parenthesization of expressions
4960 Mismatching bounds in an aggregate
4963 Attempt to return local value by reference
4966 Premature instantiation of a generic body
4969 Attempt to pack aliased components
4972 Out of bounds array subscripts
4975 Wrong length on string assignment
4978 Violations of style rules if style checking is enabled
4981 Unused @code{with} clauses
4984 @code{Bit_Order} usage that does not have any effect
4987 @code{Standard.Duration} used to resolve universal fixed expression
4990 Dereference of possibly null value
4993 Declaration that is likely to cause storage error
4996 Internal GNAT unit @code{with}'ed by application unit
4999 Values known to be out of range at compile time
5002 Unreferenced labels and variables
5005 Address overlays that could clobber memory
5008 Unexpected initialization when address clause present
5011 Bad alignment for address clause
5014 Useless type conversions
5017 Redundant assignment statements and other redundant constructs
5020 Useless exception handlers
5023 Accidental hiding of name by child unit
5026 Access before elaboration detected at compile time
5029 A range in a @code{for} loop that is known to be null or might be null
5034 The following section lists compiler switches that are available
5035 to control the handling of warning messages. It is also possible
5036 to exercise much finer control over what warnings are issued and
5037 suppressed using the GNAT pragma Warnings, @xref{Pragma Warnings,,,
5038 gnat_rm, GNAT Reference manual}.
5043 @emph{Activate all optional errors.}
5044 @cindex @option{-gnatwa} (@command{gcc})
5045 This switch activates most optional warning messages, see remaining list
5046 in this section for details on optional warning messages that can be
5047 individually controlled. The warnings that are not turned on by this
5049 @option{-gnatwd} (implicit dereferencing),
5050 @option{-gnatwh} (hiding),
5051 @option{-gnatwl} (elaboration warnings),
5052 @option{-gnatw.o} (warn on values set by out parameters ignored)
5053 and @option{-gnatwt} (tracking of deleted conditional code).
5054 All other optional warnings are turned on.
5057 @emph{Suppress all optional errors.}
5058 @cindex @option{-gnatwA} (@command{gcc})
5059 This switch suppresses all optional warning messages, see remaining list
5060 in this section for details on optional warning messages that can be
5061 individually controlled.
5064 @emph{Activate warnings on failing assertions.}
5065 @cindex @option{-gnatw.a} (@command{gcc})
5066 @cindex Assert failures
5067 This switch activates warnings for assertions where the compiler can tell at
5068 compile time that the assertion will fail. Note that this warning is given
5069 even if assertions are disabled. The default is that such warnings are
5073 @emph{Suppress warnings on failing assertions.}
5074 @cindex @option{-gnatw.A} (@command{gcc})
5075 @cindex Assert failures
5076 This switch suppresses warnings for assertions where the compiler can tell at
5077 compile time that the assertion will fail.
5080 @emph{Activate warnings on bad fixed values.}
5081 @cindex @option{-gnatwb} (@command{gcc})
5082 @cindex Bad fixed values
5083 @cindex Fixed-point Small value
5085 This switch activates warnings for static fixed-point expressions whose
5086 value is not an exact multiple of Small. Such values are implementation
5087 dependent, since an implementation is free to choose either of the multiples
5088 that surround the value. GNAT always chooses the closer one, but this is not
5089 required behavior, and it is better to specify a value that is an exact
5090 multiple, ensuring predictable execution. The default is that such warnings
5094 @emph{Suppress warnings on bad fixed values.}
5095 @cindex @option{-gnatwB} (@command{gcc})
5096 This switch suppresses warnings for static fixed-point expressions whose
5097 value is not an exact multiple of Small.
5100 @emph{Activate warnings on biased representation.}
5101 @cindex @option{-gnatw.b} (@command{gcc})
5102 @cindex Biased representation
5103 This switch activates warnings when a size clause, value size clause, component
5104 clause, or component size clause forces the use of biased representation for an
5105 integer type (e.g. representing a range of 10..11 in a single bit by using 0/1
5106 to represent 10/11). The default is that such warnings are generated.
5109 @emph{Suppress warnings on biased representation.}
5110 @cindex @option{-gnatwB} (@command{gcc})
5111 This switch suppresses warnings for representation clauses that force the use
5112 of biased representation.
5115 @emph{Activate warnings on conditionals.}
5116 @cindex @option{-gnatwc} (@command{gcc})
5117 @cindex Conditionals, constant
5118 This switch activates warnings for conditional expressions used in
5119 tests that are known to be True or False at compile time. The default
5120 is that such warnings are not generated.
5121 Note that this warning does
5122 not get issued for the use of boolean variables or constants whose
5123 values are known at compile time, since this is a standard technique
5124 for conditional compilation in Ada, and this would generate too many
5125 false positive warnings.
5127 This warning option also activates a special test for comparisons using
5128 the operators ``>='' and`` <=''.
5129 If the compiler can tell that only the equality condition is possible,
5130 then it will warn that the ``>'' or ``<'' part of the test
5131 is useless and that the operator could be replaced by ``=''.
5132 An example would be comparing a @code{Natural} variable <= 0.
5134 This warning option also generates warnings if
5135 one or both tests is optimized away in a membership test for integer
5136 values if the result can be determined at compile time. Range tests on
5137 enumeration types are not included, since it is common for such tests
5138 to include an end point.
5140 This warning can also be turned on using @option{-gnatwa}.
5143 @emph{Suppress warnings on conditionals.}
5144 @cindex @option{-gnatwC} (@command{gcc})
5145 This switch suppresses warnings for conditional expressions used in
5146 tests that are known to be True or False at compile time.
5149 @emph{Activate warnings on missing component clauses.}
5150 @cindex @option{-gnatw.c} (@command{gcc})
5151 @cindex Component clause, missing
5152 This switch activates warnings for record components where a record
5153 representation clause is present and has component clauses for the
5154 majority, but not all, of the components. A warning is given for each
5155 component for which no component clause is present.
5157 This warning can also be turned on using @option{-gnatwa}.
5160 @emph{Suppress warnings on missing component clauses.}
5161 @cindex @option{-gnatwC} (@command{gcc})
5162 This switch suppresses warnings for record components that are
5163 missing a component clause in the situation described above.
5166 @emph{Activate warnings on implicit dereferencing.}
5167 @cindex @option{-gnatwd} (@command{gcc})
5168 If this switch is set, then the use of a prefix of an access type
5169 in an indexed component, slice, or selected component without an
5170 explicit @code{.all} will generate a warning. With this warning
5171 enabled, access checks occur only at points where an explicit
5172 @code{.all} appears in the source code (assuming no warnings are
5173 generated as a result of this switch). The default is that such
5174 warnings are not generated.
5175 Note that @option{-gnatwa} does not affect the setting of
5176 this warning option.
5179 @emph{Suppress warnings on implicit dereferencing.}
5180 @cindex @option{-gnatwD} (@command{gcc})
5181 @cindex Implicit dereferencing
5182 @cindex Dereferencing, implicit
5183 This switch suppresses warnings for implicit dereferences in
5184 indexed components, slices, and selected components.
5187 @emph{Treat warnings as errors.}
5188 @cindex @option{-gnatwe} (@command{gcc})
5189 @cindex Warnings, treat as error
5190 This switch causes warning messages to be treated as errors.
5191 The warning string still appears, but the warning messages are counted
5192 as errors, and prevent the generation of an object file.
5195 @emph{Activate every optional warning}
5196 @cindex @option{-gnatw.e} (@command{gcc})
5197 @cindex Warnings, activate every optional warning
5198 This switch activates all optional warnings, including those which
5199 are not activated by @code{-gnatwa}.
5202 @emph{Activate warnings on unreferenced formals.}
5203 @cindex @option{-gnatwf} (@command{gcc})
5204 @cindex Formals, unreferenced
5205 This switch causes a warning to be generated if a formal parameter
5206 is not referenced in the body of the subprogram. This warning can
5207 also be turned on using @option{-gnatwa} or @option{-gnatwu}. The
5208 default is that these warnings are not generated.
5211 @emph{Suppress warnings on unreferenced formals.}
5212 @cindex @option{-gnatwF} (@command{gcc})
5213 This switch suppresses warnings for unreferenced formal
5214 parameters. Note that the
5215 combination @option{-gnatwu} followed by @option{-gnatwF} has the
5216 effect of warning on unreferenced entities other than subprogram
5220 @emph{Activate warnings on unrecognized pragmas.}
5221 @cindex @option{-gnatwg} (@command{gcc})
5222 @cindex Pragmas, unrecognized
5223 This switch causes a warning to be generated if an unrecognized
5224 pragma is encountered. Apart from issuing this warning, the
5225 pragma is ignored and has no effect. This warning can
5226 also be turned on using @option{-gnatwa}. The default
5227 is that such warnings are issued (satisfying the Ada Reference
5228 Manual requirement that such warnings appear).
5231 @emph{Suppress warnings on unrecognized pragmas.}
5232 @cindex @option{-gnatwG} (@command{gcc})
5233 This switch suppresses warnings for unrecognized pragmas.
5236 @emph{Activate warnings on hiding.}
5237 @cindex @option{-gnatwh} (@command{gcc})
5238 @cindex Hiding of Declarations
5239 This switch activates warnings on hiding declarations.
5240 A declaration is considered hiding
5241 if it is for a non-overloadable entity, and it declares an entity with the
5242 same name as some other entity that is directly or use-visible. The default
5243 is that such warnings are not generated.
5244 Note that @option{-gnatwa} does not affect the setting of this warning option.
5247 @emph{Suppress warnings on hiding.}
5248 @cindex @option{-gnatwH} (@command{gcc})
5249 This switch suppresses warnings on hiding declarations.
5252 @emph{Activate warnings on implementation units.}
5253 @cindex @option{-gnatwi} (@command{gcc})
5254 This switch activates warnings for a @code{with} of an internal GNAT
5255 implementation unit, defined as any unit from the @code{Ada},
5256 @code{Interfaces}, @code{GNAT},
5257 ^^@code{DEC},^ or @code{System}
5258 hierarchies that is not
5259 documented in either the Ada Reference Manual or the GNAT
5260 Programmer's Reference Manual. Such units are intended only
5261 for internal implementation purposes and should not be @code{with}'ed
5262 by user programs. The default is that such warnings are generated
5263 This warning can also be turned on using @option{-gnatwa}.
5266 @emph{Disable warnings on implementation units.}
5267 @cindex @option{-gnatwI} (@command{gcc})
5268 This switch disables warnings for a @code{with} of an internal GNAT
5269 implementation unit.
5272 @emph{Activate warnings on overlapping actuals.}
5273 @cindex @option{-gnatw.i} (@command{gcc})
5274 This switch enables a warning on statically detectable overlapping actuals in
5275 a subprogram call, when one of the actuals is an in-out parameter, and the
5276 types of the actuals are not by-copy types. The warning is off by default,
5277 and is not included under -gnatwa.
5280 @emph{Disable warnings on overlapping actuals.}
5281 @cindex @option{-gnatw.I} (@command{gcc})
5282 This switch disables warnings on overlapping actuals in a call..
5285 @emph{Activate warnings on obsolescent features (Annex J).}
5286 @cindex @option{-gnatwj} (@command{gcc})
5287 @cindex Features, obsolescent
5288 @cindex Obsolescent features
5289 If this warning option is activated, then warnings are generated for
5290 calls to subprograms marked with @code{pragma Obsolescent} and
5291 for use of features in Annex J of the Ada Reference Manual. In the
5292 case of Annex J, not all features are flagged. In particular use
5293 of the renamed packages (like @code{Text_IO}) and use of package
5294 @code{ASCII} are not flagged, since these are very common and
5295 would generate many annoying positive warnings. The default is that
5296 such warnings are not generated. This warning is also turned on by
5297 the use of @option{-gnatwa}.
5299 In addition to the above cases, warnings are also generated for
5300 GNAT features that have been provided in past versions but which
5301 have been superseded (typically by features in the new Ada standard).
5302 For example, @code{pragma Ravenscar} will be flagged since its
5303 function is replaced by @code{pragma Profile(Ravenscar)}.
5305 Note that this warning option functions differently from the
5306 restriction @code{No_Obsolescent_Features} in two respects.
5307 First, the restriction applies only to annex J features.
5308 Second, the restriction does flag uses of package @code{ASCII}.
5311 @emph{Suppress warnings on obsolescent features (Annex J).}
5312 @cindex @option{-gnatwJ} (@command{gcc})
5313 This switch disables warnings on use of obsolescent features.
5316 @emph{Activate warnings on variables that could be constants.}
5317 @cindex @option{-gnatwk} (@command{gcc})
5318 This switch activates warnings for variables that are initialized but
5319 never modified, and then could be declared constants. The default is that
5320 such warnings are not given.
5321 This warning can also be turned on using @option{-gnatwa}.
5324 @emph{Suppress warnings on variables that could be constants.}
5325 @cindex @option{-gnatwK} (@command{gcc})
5326 This switch disables warnings on variables that could be declared constants.
5329 @emph{Activate warnings for elaboration pragmas.}
5330 @cindex @option{-gnatwl} (@command{gcc})
5331 @cindex Elaboration, warnings
5332 This switch activates warnings on missing
5333 @code{Elaborate_All} and @code{Elaborate} pragmas.
5334 See the section in this guide on elaboration checking for details on
5335 when such pragmas should be used. In dynamic elaboration mode, this switch
5336 generations warnings about the need to add elaboration pragmas. Note however,
5337 that if you blindly follow these warnings, and add @code{Elaborate_All}
5338 warnings wherever they are recommended, you basically end up with the
5339 equivalent of the static elaboration model, which may not be what you want for
5340 legacy code for which the static model does not work.
5342 For the static model, the messages generated are labeled "info:" (for
5343 information messages). They are not warnings to add elaboration pragmas,
5344 merely informational messages showing what implicit elaboration pragmas
5345 have been added, for use in analyzing elaboration circularity problems.
5347 Warnings are also generated if you
5348 are using the static mode of elaboration, and a @code{pragma Elaborate}
5349 is encountered. The default is that such warnings
5351 This warning is not automatically turned on by the use of @option{-gnatwa}.
5354 @emph{Suppress warnings for elaboration pragmas.}
5355 @cindex @option{-gnatwL} (@command{gcc})
5356 This switch suppresses warnings on missing Elaborate and Elaborate_All pragmas.
5357 See the section in this guide on elaboration checking for details on
5358 when such pragmas should be used.
5361 @emph{Activate warnings on modified but unreferenced variables.}
5362 @cindex @option{-gnatwm} (@command{gcc})
5363 This switch activates warnings for variables that are assigned (using
5364 an initialization value or with one or more assignment statements) but
5365 whose value is never read. The warning is suppressed for volatile
5366 variables and also for variables that are renamings of other variables
5367 or for which an address clause is given.
5368 This warning can also be turned on using @option{-gnatwa}.
5369 The default is that these warnings are not given.
5372 @emph{Disable warnings on modified but unreferenced variables.}
5373 @cindex @option{-gnatwM} (@command{gcc})
5374 This switch disables warnings for variables that are assigned or
5375 initialized, but never read.
5378 @emph{Activate warnings on suspicious modulus values.}
5379 @cindex @option{-gnatw.m} (@command{gcc})
5380 This switch activates warnings for modulus values that seem suspicious.
5381 The cases caught are where the size is the same as the modulus (e.g.
5382 a modulus of 7 with a size of 7 bits), and modulus values of 32 or 64
5383 with no size clause. The guess in both cases is that 2**x was intended
5384 rather than x. The default is that these warnings are given.
5387 @emph{Disable warnings on suspicious modulus values.}
5388 @cindex @option{-gnatw.M} (@command{gcc})
5389 This switch disables warnings for suspicious modulus values.
5392 @emph{Set normal warnings mode.}
5393 @cindex @option{-gnatwn} (@command{gcc})
5394 This switch sets normal warning mode, in which enabled warnings are
5395 issued and treated as warnings rather than errors. This is the default
5396 mode. the switch @option{-gnatwn} can be used to cancel the effect of
5397 an explicit @option{-gnatws} or
5398 @option{-gnatwe}. It also cancels the effect of the
5399 implicit @option{-gnatwe} that is activated by the
5400 use of @option{-gnatg}.
5403 @emph{Activate warnings on address clause overlays.}
5404 @cindex @option{-gnatwo} (@command{gcc})
5405 @cindex Address Clauses, warnings
5406 This switch activates warnings for possibly unintended initialization
5407 effects of defining address clauses that cause one variable to overlap
5408 another. The default is that such warnings are generated.
5409 This warning can also be turned on using @option{-gnatwa}.
5412 @emph{Suppress warnings on address clause overlays.}
5413 @cindex @option{-gnatwO} (@command{gcc})
5414 This switch suppresses warnings on possibly unintended initialization
5415 effects of defining address clauses that cause one variable to overlap
5419 @emph{Activate warnings on modified but unreferenced out parameters.}
5420 @cindex @option{-gnatw.o} (@command{gcc})
5421 This switch activates warnings for variables that are modified by using
5422 them as actuals for a call to a procedure with an out mode formal, where
5423 the resulting assigned value is never read. It is applicable in the case
5424 where there is more than one out mode formal. If there is only one out
5425 mode formal, the warning is issued by default (controlled by -gnatwu).
5426 The warning is suppressed for volatile
5427 variables and also for variables that are renamings of other variables
5428 or for which an address clause is given.
5429 The default is that these warnings are not given. Note that this warning
5430 is not included in -gnatwa, it must be activated explicitly.
5433 @emph{Disable warnings on modified but unreferenced out parameters.}
5434 @cindex @option{-gnatw.O} (@command{gcc})
5435 This switch suppresses warnings for variables that are modified by using
5436 them as actuals for a call to a procedure with an out mode formal, where
5437 the resulting assigned value is never read.
5440 @emph{Activate warnings on ineffective pragma Inlines.}
5441 @cindex @option{-gnatwp} (@command{gcc})
5442 @cindex Inlining, warnings
5443 This switch activates warnings for failure of front end inlining
5444 (activated by @option{-gnatN}) to inline a particular call. There are
5445 many reasons for not being able to inline a call, including most
5446 commonly that the call is too complex to inline. The default is
5447 that such warnings are not given.
5448 This warning can also be turned on using @option{-gnatwa}.
5449 Warnings on ineffective inlining by the gcc back-end can be activated
5450 separately, using the gcc switch -Winline.
5453 @emph{Suppress warnings on ineffective pragma Inlines.}
5454 @cindex @option{-gnatwP} (@command{gcc})
5455 This switch suppresses warnings on ineffective pragma Inlines. If the
5456 inlining mechanism cannot inline a call, it will simply ignore the
5460 @emph{Activate warnings on parameter ordering.}
5461 @cindex @option{-gnatw.p} (@command{gcc})
5462 @cindex Parameter order, warnings
5463 This switch activates warnings for cases of suspicious parameter
5464 ordering when the list of arguments are all simple identifiers that
5465 match the names of the formals, but are in a different order. The
5466 warning is suppressed if any use of named parameter notation is used,
5467 so this is the appropriate way to suppress a false positive (and
5468 serves to emphasize that the "misordering" is deliberate). The
5470 that such warnings are not given.
5471 This warning can also be turned on using @option{-gnatwa}.
5474 @emph{Suppress warnings on parameter ordering.}
5475 @cindex @option{-gnatw.P} (@command{gcc})
5476 This switch suppresses warnings on cases of suspicious parameter
5480 @emph{Activate warnings on questionable missing parentheses.}
5481 @cindex @option{-gnatwq} (@command{gcc})
5482 @cindex Parentheses, warnings
5483 This switch activates warnings for cases where parentheses are not used and
5484 the result is potential ambiguity from a readers point of view. For example
5485 (not a > b) when a and b are modular means ((not a) > b) and very likely the
5486 programmer intended (not (a > b)). Similarly (-x mod 5) means (-(x mod 5)) and
5487 quite likely ((-x) mod 5) was intended. In such situations it seems best to
5488 follow the rule of always parenthesizing to make the association clear, and
5489 this warning switch warns if such parentheses are not present. The default
5490 is that these warnings are given.
5491 This warning can also be turned on using @option{-gnatwa}.
5494 @emph{Suppress warnings on questionable missing parentheses.}
5495 @cindex @option{-gnatwQ} (@command{gcc})
5496 This switch suppresses warnings for cases where the association is not
5497 clear and the use of parentheses is preferred.
5500 @emph{Activate warnings on redundant constructs.}
5501 @cindex @option{-gnatwr} (@command{gcc})
5502 This switch activates warnings for redundant constructs. The following
5503 is the current list of constructs regarded as redundant:
5507 Assignment of an item to itself.
5509 Type conversion that converts an expression to its own type.
5511 Use of the attribute @code{Base} where @code{typ'Base} is the same
5514 Use of pragma @code{Pack} when all components are placed by a record
5515 representation clause.
5517 Exception handler containing only a reraise statement (raise with no
5518 operand) which has no effect.
5520 Use of the operator abs on an operand that is known at compile time
5523 Comparison of boolean expressions to an explicit True value.
5526 This warning can also be turned on using @option{-gnatwa}.
5527 The default is that warnings for redundant constructs are not given.
5530 @emph{Suppress warnings on redundant constructs.}
5531 @cindex @option{-gnatwR} (@command{gcc})
5532 This switch suppresses warnings for redundant constructs.
5535 @emph{Activate warnings for object renaming function.}
5536 @cindex @option{-gnatw.r} (@command{gcc})
5537 This switch activates warnings for an object renaming that renames a
5538 function call, which is equivalent to a constant declaration (as
5539 opposed to renaming the function itself). The default is that these
5540 warnings are given. This warning can also be turned on using
5544 @emph{Suppress warnings for object renaming function.}
5545 @cindex @option{-gnatwT} (@command{gcc})
5546 This switch suppresses warnings for object renaming function.
5549 @emph{Suppress all warnings.}
5550 @cindex @option{-gnatws} (@command{gcc})
5551 This switch completely suppresses the
5552 output of all warning messages from the GNAT front end.
5553 Note that it does not suppress warnings from the @command{gcc} back end.
5554 To suppress these back end warnings as well, use the switch @option{-w}
5555 in addition to @option{-gnatws}.
5558 @emph{Activate warnings for tracking of deleted conditional code.}
5559 @cindex @option{-gnatwt} (@command{gcc})
5560 @cindex Deactivated code, warnings
5561 @cindex Deleted code, warnings
5562 This switch activates warnings for tracking of code in conditionals (IF and
5563 CASE statements) that is detected to be dead code which cannot be executed, and
5564 which is removed by the front end. This warning is off by default, and is not
5565 turned on by @option{-gnatwa}, it has to be turned on explicitly. This may be
5566 useful for detecting deactivated code in certified applications.
5569 @emph{Suppress warnings for tracking of deleted conditional code.}
5570 @cindex @option{-gnatwT} (@command{gcc})
5571 This switch suppresses warnings for tracking of deleted conditional code.
5574 @emph{Activate warnings on unused entities.}
5575 @cindex @option{-gnatwu} (@command{gcc})
5576 This switch activates warnings to be generated for entities that
5577 are declared but not referenced, and for units that are @code{with}'ed
5579 referenced. In the case of packages, a warning is also generated if
5580 no entities in the package are referenced. This means that if the package
5581 is referenced but the only references are in @code{use}
5582 clauses or @code{renames}
5583 declarations, a warning is still generated. A warning is also generated
5584 for a generic package that is @code{with}'ed but never instantiated.
5585 In the case where a package or subprogram body is compiled, and there
5586 is a @code{with} on the corresponding spec
5587 that is only referenced in the body,
5588 a warning is also generated, noting that the
5589 @code{with} can be moved to the body. The default is that
5590 such warnings are not generated.
5591 This switch also activates warnings on unreferenced formals
5592 (it includes the effect of @option{-gnatwf}).
5593 This warning can also be turned on using @option{-gnatwa}.
5596 @emph{Suppress warnings on unused entities.}
5597 @cindex @option{-gnatwU} (@command{gcc})
5598 This switch suppresses warnings for unused entities and packages.
5599 It also turns off warnings on unreferenced formals (and thus includes
5600 the effect of @option{-gnatwF}).
5603 @emph{Activate warnings on unassigned variables.}
5604 @cindex @option{-gnatwv} (@command{gcc})
5605 @cindex Unassigned variable warnings
5606 This switch activates warnings for access to variables which
5607 may not be properly initialized. The default is that
5608 such warnings are generated.
5609 This warning can also be turned on using @option{-gnatwa}.
5612 @emph{Suppress warnings on unassigned variables.}
5613 @cindex @option{-gnatwV} (@command{gcc})
5614 This switch suppresses warnings for access to variables which
5615 may not be properly initialized.
5616 For variables of a composite type, the warning can also be suppressed in
5617 Ada 2005 by using a default initialization with a box. For example, if
5618 Table is an array of records whose components are only partially uninitialized,
5619 then the following code:
5621 @smallexample @c ada
5622 Tab : Table := (others => <>);
5625 will suppress warnings on subsequent statements that access components
5629 @emph{Activate warnings on wrong low bound assumption.}
5630 @cindex @option{-gnatww} (@command{gcc})
5631 @cindex String indexing warnings
5632 This switch activates warnings for indexing an unconstrained string parameter
5633 with a literal or S'Length. This is a case where the code is assuming that the
5634 low bound is one, which is in general not true (for example when a slice is
5635 passed). The default is that such warnings are generated.
5636 This warning can also be turned on using @option{-gnatwa}.
5639 @emph{Suppress warnings on wrong low bound assumption.}
5640 @cindex @option{-gnatwW} (@command{gcc})
5641 This switch suppresses warnings for indexing an unconstrained string parameter
5642 with a literal or S'Length. Note that this warning can also be suppressed
5643 in a particular case by adding an
5644 assertion that the lower bound is 1,
5645 as shown in the following example.
5647 @smallexample @c ada
5648 procedure K (S : String) is
5649 pragma Assert (S'First = 1);
5654 @emph{Activate warnings on unnecessary Warnings Off pragmas}
5655 @cindex @option{-gnatw.w} (@command{gcc})
5656 @cindex Warnings Off control
5657 This switch activates warnings for use of @code{pragma Warnings (Off, entity}
5658 where either the pragma is entirely useless (because it suppresses no
5659 warnings), or it could be replaced by @code{pragma Unreferenced} or
5660 @code{pragma Unmodified}.The default is that these warnings are not given.
5661 Note that this warning is not included in -gnatwa, it must be
5662 activated explicitly.
5665 @emph{Suppress warnings on unnecessary Warnings Off pragmas}
5666 @cindex @option{-gnatw.W} (@command{gcc})
5667 This switch suppresses warnings for use of @code{pragma Warnings (Off, entity}.
5670 @emph{Activate warnings on Export/Import pragmas.}
5671 @cindex @option{-gnatwx} (@command{gcc})
5672 @cindex Export/Import pragma warnings
5673 This switch activates warnings on Export/Import pragmas when
5674 the compiler detects a possible conflict between the Ada and
5675 foreign language calling sequences. For example, the use of
5676 default parameters in a convention C procedure is dubious
5677 because the C compiler cannot supply the proper default, so
5678 a warning is issued. The default is that such warnings are
5680 This warning can also be turned on using @option{-gnatwa}.
5683 @emph{Suppress warnings on Export/Import pragmas.}
5684 @cindex @option{-gnatwX} (@command{gcc})
5685 This switch suppresses warnings on Export/Import pragmas.
5686 The sense of this is that you are telling the compiler that
5687 you know what you are doing in writing the pragma, and it
5688 should not complain at you.
5691 @emph{Activate warnings for No_Exception_Propagation mode.}
5692 @cindex @option{-gnatwm} (@command{gcc})
5693 This switch activates warnings for exception usage when pragma Restrictions
5694 (No_Exception_Propagation) is in effect. Warnings are given for implicit or
5695 explicit exception raises which are not covered by a local handler, and for
5696 exception handlers which do not cover a local raise. The default is that these
5697 warnings are not given.
5700 @emph{Disable warnings for No_Exception_Propagation mode.}
5701 This switch disables warnings for exception usage when pragma Restrictions
5702 (No_Exception_Propagation) is in effect.
5705 @emph{Activate warnings for Ada 2005 compatibility issues.}
5706 @cindex @option{-gnatwy} (@command{gcc})
5707 @cindex Ada 2005 compatibility issues warnings
5708 For the most part Ada 2005 is upwards compatible with Ada 95,
5709 but there are some exceptions (for example the fact that
5710 @code{interface} is now a reserved word in Ada 2005). This
5711 switch activates several warnings to help in identifying
5712 and correcting such incompatibilities. The default is that
5713 these warnings are generated. Note that at one point Ada 2005
5714 was called Ada 0Y, hence the choice of character.
5715 This warning can also be turned on using @option{-gnatwa}.
5718 @emph{Disable warnings for Ada 2005 compatibility issues.}
5719 @cindex @option{-gnatwY} (@command{gcc})
5720 @cindex Ada 2005 compatibility issues warnings
5721 This switch suppresses several warnings intended to help in identifying
5722 incompatibilities between Ada 95 and Ada 2005.
5725 @emph{Activate warnings on unchecked conversions.}
5726 @cindex @option{-gnatwz} (@command{gcc})
5727 @cindex Unchecked_Conversion warnings
5728 This switch activates warnings for unchecked conversions
5729 where the types are known at compile time to have different
5731 is that such warnings are generated. Warnings are also
5732 generated for subprogram pointers with different conventions,
5733 and, on VMS only, for data pointers with different conventions.
5734 This warning can also be turned on using @option{-gnatwa}.
5737 @emph{Suppress warnings on unchecked conversions.}
5738 @cindex @option{-gnatwZ} (@command{gcc})
5739 This switch suppresses warnings for unchecked conversions
5740 where the types are known at compile time to have different
5741 sizes or conventions.
5743 @item ^-Wunused^WARNINGS=UNUSED^
5744 @cindex @option{-Wunused}
5745 The warnings controlled by the @option{-gnatw} switch are generated by
5746 the front end of the compiler. The @option{GCC} back end can provide
5747 additional warnings and they are controlled by the @option{-W} switch.
5748 For example, @option{^-Wunused^WARNINGS=UNUSED^} activates back end
5749 warnings for entities that are declared but not referenced.
5751 @item ^-Wuninitialized^WARNINGS=UNINITIALIZED^
5752 @cindex @option{-Wuninitialized}
5753 Similarly, @option{^-Wuninitialized^WARNINGS=UNINITIALIZED^} activates
5754 the back end warning for uninitialized variables. This switch must be
5755 used in conjunction with an optimization level greater than zero.
5757 @item ^-Wall^/ALL_BACK_END_WARNINGS^
5758 @cindex @option{-Wall}
5759 This switch enables all the above warnings from the @option{GCC} back end.
5760 The code generator detects a number of warning situations that are missed
5761 by the @option{GNAT} front end, and this switch can be used to activate them.
5762 The use of this switch also sets the default front end warning mode to
5763 @option{-gnatwa}, that is, most front end warnings activated as well.
5765 @item ^-w^/NO_BACK_END_WARNINGS^
5767 Conversely, this switch suppresses warnings from the @option{GCC} back end.
5768 The use of this switch also sets the default front end warning mode to
5769 @option{-gnatws}, that is, front end warnings suppressed as well.
5775 A string of warning parameters can be used in the same parameter. For example:
5782 will turn on all optional warnings except for elaboration pragma warnings,
5783 and also specify that warnings should be treated as errors.
5785 When no switch @option{^-gnatw^/WARNINGS^} is used, this is equivalent to:
5810 @node Debugging and Assertion Control
5811 @subsection Debugging and Assertion Control
5815 @cindex @option{-gnata} (@command{gcc})
5821 The pragmas @code{Assert} and @code{Debug} normally have no effect and
5822 are ignored. This switch, where @samp{a} stands for assert, causes
5823 @code{Assert} and @code{Debug} pragmas to be activated.
5825 The pragmas have the form:
5829 @b{pragma} Assert (@var{Boolean-expression} @r{[},
5830 @var{static-string-expression}@r{]})
5831 @b{pragma} Debug (@var{procedure call})
5836 The @code{Assert} pragma causes @var{Boolean-expression} to be tested.
5837 If the result is @code{True}, the pragma has no effect (other than
5838 possible side effects from evaluating the expression). If the result is
5839 @code{False}, the exception @code{Assert_Failure} declared in the package
5840 @code{System.Assertions} is
5841 raised (passing @var{static-string-expression}, if present, as the
5842 message associated with the exception). If no string expression is
5843 given the default is a string giving the file name and line number
5846 The @code{Debug} pragma causes @var{procedure} to be called. Note that
5847 @code{pragma Debug} may appear within a declaration sequence, allowing
5848 debugging procedures to be called between declarations.
5851 @item /DEBUG@r{[}=debug-level@r{]}
5853 Specifies how much debugging information is to be included in
5854 the resulting object file where 'debug-level' is one of the following:
5857 Include both debugger symbol records and traceback
5859 This is the default setting.
5861 Include both debugger symbol records and traceback in
5864 Excludes both debugger symbol records and traceback
5865 the object file. Same as /NODEBUG.
5867 Includes only debugger symbol records in the object
5868 file. Note that this doesn't include traceback information.
5873 @node Validity Checking
5874 @subsection Validity Checking
5875 @findex Validity Checking
5878 The Ada Reference Manual defines the concept of invalid values (see
5879 RM 13.9.1). The primary source of invalid values is uninitialized
5880 variables. A scalar variable that is left uninitialized may contain
5881 an invalid value; the concept of invalid does not apply to access or
5884 It is an error to read an invalid value, but the RM does not require
5885 run-time checks to detect such errors, except for some minimal
5886 checking to prevent erroneous execution (i.e. unpredictable
5887 behavior). This corresponds to the @option{-gnatVd} switch below,
5888 which is the default. For example, by default, if the expression of a
5889 case statement is invalid, it will raise Constraint_Error rather than
5890 causing a wild jump, and if an array index on the left-hand side of an
5891 assignment is invalid, it will raise Constraint_Error rather than
5892 overwriting an arbitrary memory location.
5894 The @option{-gnatVa} may be used to enable additional validity checks,
5895 which are not required by the RM. These checks are often very
5896 expensive (which is why the RM does not require them). These checks
5897 are useful in tracking down uninitialized variables, but they are
5898 not usually recommended for production builds.
5900 The other @option{-gnatV^@var{x}^^} switches below allow finer-grained
5901 control; you can enable whichever validity checks you desire. However,
5902 for most debugging purposes, @option{-gnatVa} is sufficient, and the
5903 default @option{-gnatVd} (i.e. standard Ada behavior) is usually
5904 sufficient for non-debugging use.
5906 The @option{-gnatB} switch tells the compiler to assume that all
5907 values are valid (that is, within their declared subtype range)
5908 except in the context of a use of the Valid attribute. This means
5909 the compiler can generate more efficient code, since the range
5910 of values is better known at compile time. However, an uninitialized
5911 variable can cause wild jumps and memory corruption in this mode.
5913 The @option{-gnatV^@var{x}^^} switch allows control over the validity
5914 checking mode as described below.
5916 The @code{x} argument is a string of letters that
5917 indicate validity checks that are performed or not performed in addition
5918 to the default checks required by Ada as described above.
5921 The options allowed for this qualifier
5922 indicate validity checks that are performed or not performed in addition
5923 to the default checks required by Ada as described above.
5929 @emph{All validity checks.}
5930 @cindex @option{-gnatVa} (@command{gcc})
5931 All validity checks are turned on.
5933 That is, @option{-gnatVa} is
5934 equivalent to @option{gnatVcdfimorst}.
5938 @emph{Validity checks for copies.}
5939 @cindex @option{-gnatVc} (@command{gcc})
5940 The right hand side of assignments, and the initializing values of
5941 object declarations are validity checked.
5944 @emph{Default (RM) validity checks.}
5945 @cindex @option{-gnatVd} (@command{gcc})
5946 Some validity checks are done by default following normal Ada semantics
5948 A check is done in case statements that the expression is within the range
5949 of the subtype. If it is not, Constraint_Error is raised.
5950 For assignments to array components, a check is done that the expression used
5951 as index is within the range. If it is not, Constraint_Error is raised.
5952 Both these validity checks may be turned off using switch @option{-gnatVD}.
5953 They are turned on by default. If @option{-gnatVD} is specified, a subsequent
5954 switch @option{-gnatVd} will leave the checks turned on.
5955 Switch @option{-gnatVD} should be used only if you are sure that all such
5956 expressions have valid values. If you use this switch and invalid values
5957 are present, then the program is erroneous, and wild jumps or memory
5958 overwriting may occur.
5961 @emph{Validity checks for elementary components.}
5962 @cindex @option{-gnatVe} (@command{gcc})
5963 In the absence of this switch, assignments to record or array components are
5964 not validity checked, even if validity checks for assignments generally
5965 (@option{-gnatVc}) are turned on. In Ada, assignment of composite values do not
5966 require valid data, but assignment of individual components does. So for
5967 example, there is a difference between copying the elements of an array with a
5968 slice assignment, compared to assigning element by element in a loop. This
5969 switch allows you to turn off validity checking for components, even when they
5970 are assigned component by component.
5973 @emph{Validity checks for floating-point values.}
5974 @cindex @option{-gnatVf} (@command{gcc})
5975 In the absence of this switch, validity checking occurs only for discrete
5976 values. If @option{-gnatVf} is specified, then validity checking also applies
5977 for floating-point values, and NaNs and infinities are considered invalid,
5978 as well as out of range values for constrained types. Note that this means
5979 that standard IEEE infinity mode is not allowed. The exact contexts
5980 in which floating-point values are checked depends on the setting of other
5981 options. For example,
5982 @option{^-gnatVif^VALIDITY_CHECKING=(IN_PARAMS,FLOATS)^} or
5983 @option{^-gnatVfi^VALIDITY_CHECKING=(FLOATS,IN_PARAMS)^}
5984 (the order does not matter) specifies that floating-point parameters of mode
5985 @code{in} should be validity checked.
5988 @emph{Validity checks for @code{in} mode parameters}
5989 @cindex @option{-gnatVi} (@command{gcc})
5990 Arguments for parameters of mode @code{in} are validity checked in function
5991 and procedure calls at the point of call.
5994 @emph{Validity checks for @code{in out} mode parameters.}
5995 @cindex @option{-gnatVm} (@command{gcc})
5996 Arguments for parameters of mode @code{in out} are validity checked in
5997 procedure calls at the point of call. The @code{'m'} here stands for
5998 modify, since this concerns parameters that can be modified by the call.
5999 Note that there is no specific option to test @code{out} parameters,
6000 but any reference within the subprogram will be tested in the usual
6001 manner, and if an invalid value is copied back, any reference to it
6002 will be subject to validity checking.
6005 @emph{No validity checks.}
6006 @cindex @option{-gnatVn} (@command{gcc})
6007 This switch turns off all validity checking, including the default checking
6008 for case statements and left hand side subscripts. Note that the use of
6009 the switch @option{-gnatp} suppresses all run-time checks, including
6010 validity checks, and thus implies @option{-gnatVn}. When this switch
6011 is used, it cancels any other @option{-gnatV} previously issued.
6014 @emph{Validity checks for operator and attribute operands.}
6015 @cindex @option{-gnatVo} (@command{gcc})
6016 Arguments for predefined operators and attributes are validity checked.
6017 This includes all operators in package @code{Standard},
6018 the shift operators defined as intrinsic in package @code{Interfaces}
6019 and operands for attributes such as @code{Pos}. Checks are also made
6020 on individual component values for composite comparisons, and on the
6021 expressions in type conversions and qualified expressions. Checks are
6022 also made on explicit ranges using @samp{..} (e.g.@: slices, loops etc).
6025 @emph{Validity checks for parameters.}
6026 @cindex @option{-gnatVp} (@command{gcc})
6027 This controls the treatment of parameters within a subprogram (as opposed
6028 to @option{-gnatVi} and @option{-gnatVm} which control validity testing
6029 of parameters on a call. If either of these call options is used, then
6030 normally an assumption is made within a subprogram that the input arguments
6031 have been validity checking at the point of call, and do not need checking
6032 again within a subprogram). If @option{-gnatVp} is set, then this assumption
6033 is not made, and parameters are not assumed to be valid, so their validity
6034 will be checked (or rechecked) within the subprogram.
6037 @emph{Validity checks for function returns.}
6038 @cindex @option{-gnatVr} (@command{gcc})
6039 The expression in @code{return} statements in functions is validity
6043 @emph{Validity checks for subscripts.}
6044 @cindex @option{-gnatVs} (@command{gcc})
6045 All subscripts expressions are checked for validity, whether they appear
6046 on the right side or left side (in default mode only left side subscripts
6047 are validity checked).
6050 @emph{Validity checks for tests.}
6051 @cindex @option{-gnatVt} (@command{gcc})
6052 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
6053 statements are checked, as well as guard expressions in entry calls.
6058 The @option{-gnatV} switch may be followed by
6059 ^a string of letters^a list of options^
6060 to turn on a series of validity checking options.
6062 @option{^-gnatVcr^/VALIDITY_CHECKING=(COPIES, RETURNS)^}
6063 specifies that in addition to the default validity checking, copies and
6064 function return expressions are to be validity checked.
6065 In order to make it easier
6066 to specify the desired combination of effects,
6068 the upper case letters @code{CDFIMORST} may
6069 be used to turn off the corresponding lower case option.
6072 the prefix @code{NO} on an option turns off the corresponding validity
6075 @item @code{NOCOPIES}
6076 @item @code{NODEFAULT}
6077 @item @code{NOFLOATS}
6078 @item @code{NOIN_PARAMS}
6079 @item @code{NOMOD_PARAMS}
6080 @item @code{NOOPERANDS}
6081 @item @code{NORETURNS}
6082 @item @code{NOSUBSCRIPTS}
6083 @item @code{NOTESTS}
6087 @option{^-gnatVaM^/VALIDITY_CHECKING=(ALL, NOMOD_PARAMS)^}
6088 turns on all validity checking options except for
6089 checking of @code{@b{in out}} procedure arguments.
6091 The specification of additional validity checking generates extra code (and
6092 in the case of @option{-gnatVa} the code expansion can be substantial).
6093 However, these additional checks can be very useful in detecting
6094 uninitialized variables, incorrect use of unchecked conversion, and other
6095 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
6096 is useful in conjunction with the extra validity checking, since this
6097 ensures that wherever possible uninitialized variables have invalid values.
6099 See also the pragma @code{Validity_Checks} which allows modification of
6100 the validity checking mode at the program source level, and also allows for
6101 temporary disabling of validity checks.
6103 @node Style Checking
6104 @subsection Style Checking
6105 @findex Style checking
6108 The @option{-gnaty^x^(option,option,@dots{})^} switch
6109 @cindex @option{-gnaty} (@command{gcc})
6110 causes the compiler to
6111 enforce specified style rules. A limited set of style rules has been used
6112 in writing the GNAT sources themselves. This switch allows user programs
6113 to activate all or some of these checks. If the source program fails a
6114 specified style check, an appropriate warning message is given, preceded by
6115 the character sequence ``(style)''.
6117 @code{(option,option,@dots{})} is a sequence of keywords
6120 The string @var{x} is a sequence of letters or digits
6122 indicating the particular style
6123 checks to be performed. The following checks are defined:
6128 @emph{Specify indentation level.}
6129 If a digit from 1-9 appears
6130 ^in the string after @option{-gnaty}^as an option for /STYLE_CHECKS^
6131 then proper indentation is checked, with the digit indicating the
6132 indentation level required. A value of zero turns off this style check.
6133 The general style of required indentation is as specified by
6134 the examples in the Ada Reference Manual. Full line comments must be
6135 aligned with the @code{--} starting on a column that is a multiple of
6136 the alignment level, or they may be aligned the same way as the following
6137 non-blank line (this is useful when full line comments appear in the middle
6141 @emph{Check attribute casing.}
6142 Attribute names, including the case of keywords such as @code{digits}
6143 used as attributes names, must be written in mixed case, that is, the
6144 initial letter and any letter following an underscore must be uppercase.
6145 All other letters must be lowercase.
6147 @item ^A^ARRAY_INDEXES^
6148 @emph{Use of array index numbers in array attributes.}
6149 When using the array attributes First, Last, Range,
6150 or Length, the index number must be omitted for one-dimensional arrays
6151 and is required for multi-dimensional arrays.
6154 @emph{Blanks not allowed at statement end.}
6155 Trailing blanks are not allowed at the end of statements. The purpose of this
6156 rule, together with h (no horizontal tabs), is to enforce a canonical format
6157 for the use of blanks to separate source tokens.
6159 @item ^B^BOOLEAN_OPERATORS^
6160 @emph{Check Boolean operators.}
6161 The use of AND/OR operators is not permitted except in the cases of modular
6162 operands, array operands, and simple stand-alone boolean variables or
6163 boolean constants. In all other cases AND THEN/OR ELSE are required.
6166 @emph{Check comments.}
6167 Comments must meet the following set of rules:
6172 The ``@code{--}'' that starts the column must either start in column one,
6173 or else at least one blank must precede this sequence.
6176 Comments that follow other tokens on a line must have at least one blank
6177 following the ``@code{--}'' at the start of the comment.
6180 Full line comments must have two blanks following the ``@code{--}'' that
6181 starts the comment, with the following exceptions.
6184 A line consisting only of the ``@code{--}'' characters, possibly preceded
6185 by blanks is permitted.
6188 A comment starting with ``@code{--x}'' where @code{x} is a special character
6190 This allows proper processing of the output generated by specialized tools
6191 including @command{gnatprep} (where ``@code{--!}'' is used) and the SPARK
6193 language (where ``@code{--#}'' is used). For the purposes of this rule, a
6194 special character is defined as being in one of the ASCII ranges
6195 @code{16#21#@dots{}16#2F#} or @code{16#3A#@dots{}16#3F#}.
6196 Note that this usage is not permitted
6197 in GNAT implementation units (i.e., when @option{-gnatg} is used).
6200 A line consisting entirely of minus signs, possibly preceded by blanks, is
6201 permitted. This allows the construction of box comments where lines of minus
6202 signs are used to form the top and bottom of the box.
6205 A comment that starts and ends with ``@code{--}'' is permitted as long as at
6206 least one blank follows the initial ``@code{--}''. Together with the preceding
6207 rule, this allows the construction of box comments, as shown in the following
6210 ---------------------------
6211 -- This is a box comment --
6212 -- with two text lines. --
6213 ---------------------------
6217 @item ^d^DOS_LINE_ENDINGS^
6218 @emph{Check no DOS line terminators present.}
6219 All lines must be terminated by a single ASCII.LF
6220 character (in particular the DOS line terminator sequence CR/LF is not
6224 @emph{Check end/exit labels.}
6225 Optional labels on @code{end} statements ending subprograms and on
6226 @code{exit} statements exiting named loops, are required to be present.
6229 @emph{No form feeds or vertical tabs.}
6230 Neither form feeds nor vertical tab characters are permitted
6234 @emph{GNAT style mode}
6235 The set of style check switches is set to match that used by the GNAT sources.
6236 This may be useful when developing code that is eventually intended to be
6237 incorporated into GNAT. For further details, see GNAT sources.
6240 @emph{No horizontal tabs.}
6241 Horizontal tab characters are not permitted in the source text.
6242 Together with the b (no blanks at end of line) check, this
6243 enforces a canonical form for the use of blanks to separate
6247 @emph{Check if-then layout.}
6248 The keyword @code{then} must appear either on the same
6249 line as corresponding @code{if}, or on a line on its own, lined
6250 up under the @code{if} with at least one non-blank line in between
6251 containing all or part of the condition to be tested.
6254 @emph{check mode IN keywords}
6255 Mode @code{in} (the default mode) is not
6256 allowed to be given explicitly. @code{in out} is fine,
6257 but not @code{in} on its own.
6260 @emph{Check keyword casing.}
6261 All keywords must be in lower case (with the exception of keywords
6262 such as @code{digits} used as attribute names to which this check
6266 @emph{Check layout.}
6267 Layout of statement and declaration constructs must follow the
6268 recommendations in the Ada Reference Manual, as indicated by the
6269 form of the syntax rules. For example an @code{else} keyword must
6270 be lined up with the corresponding @code{if} keyword.
6272 There are two respects in which the style rule enforced by this check
6273 option are more liberal than those in the Ada Reference Manual. First
6274 in the case of record declarations, it is permissible to put the
6275 @code{record} keyword on the same line as the @code{type} keyword, and
6276 then the @code{end} in @code{end record} must line up under @code{type}.
6277 This is also permitted when the type declaration is split on two lines.
6278 For example, any of the following three layouts is acceptable:
6280 @smallexample @c ada
6303 Second, in the case of a block statement, a permitted alternative
6304 is to put the block label on the same line as the @code{declare} or
6305 @code{begin} keyword, and then line the @code{end} keyword up under
6306 the block label. For example both the following are permitted:
6308 @smallexample @c ada
6326 The same alternative format is allowed for loops. For example, both of
6327 the following are permitted:
6329 @smallexample @c ada
6331 Clear : while J < 10 loop
6342 @item ^Lnnn^MAX_NESTING=nnn^
6343 @emph{Set maximum nesting level}
6344 The maximum level of nesting of constructs (including subprograms, loops,
6345 blocks, packages, and conditionals) may not exceed the given value
6346 @option{nnn}. A value of zero disconnects this style check.
6348 @item ^m^LINE_LENGTH^
6349 @emph{Check maximum line length.}
6350 The length of source lines must not exceed 79 characters, including
6351 any trailing blanks. The value of 79 allows convenient display on an
6352 80 character wide device or window, allowing for possible special
6353 treatment of 80 character lines. Note that this count is of
6354 characters in the source text. This means that a tab character counts
6355 as one character in this count but a wide character sequence counts as
6356 a single character (however many bytes are needed in the encoding).
6358 @item ^Mnnn^MAX_LENGTH=nnn^
6359 @emph{Set maximum line length.}
6360 The length of lines must not exceed the
6361 given value @option{nnn}. The maximum value that can be specified is 32767.
6363 @item ^n^STANDARD_CASING^
6364 @emph{Check casing of entities in Standard.}
6365 Any identifier from Standard must be cased
6366 to match the presentation in the Ada Reference Manual (for example,
6367 @code{Integer} and @code{ASCII.NUL}).
6370 @emph{Turn off all style checks}
6371 All style check options are turned off.
6373 @item ^o^ORDERED_SUBPROGRAMS^
6374 @emph{Check order of subprogram bodies.}
6375 All subprogram bodies in a given scope
6376 (e.g.@: a package body) must be in alphabetical order. The ordering
6377 rule uses normal Ada rules for comparing strings, ignoring casing
6378 of letters, except that if there is a trailing numeric suffix, then
6379 the value of this suffix is used in the ordering (e.g.@: Junk2 comes
6382 @item ^O^OVERRIDING_INDICATORS^
6383 @emph{Check that overriding subprograms are explicitly marked as such.}
6384 The declaration of a primitive operation of a type extension that overrides
6385 an inherited operation must carry an overriding indicator.
6388 @emph{Check pragma casing.}
6389 Pragma names must be written in mixed case, that is, the
6390 initial letter and any letter following an underscore must be uppercase.
6391 All other letters must be lowercase.
6393 @item ^r^REFERENCES^
6394 @emph{Check references.}
6395 All identifier references must be cased in the same way as the
6396 corresponding declaration. No specific casing style is imposed on
6397 identifiers. The only requirement is for consistency of references
6400 @item ^S^STATEMENTS_AFTER_THEN_ELSE^
6401 @emph{Check no statements after THEN/ELSE.}
6402 No statements are allowed
6403 on the same line as a THEN or ELSE keyword following the
6404 keyword in an IF statement. OR ELSE and AND THEN are not affected,
6405 and a special exception allows a pragma to appear after ELSE.
6408 @emph{Check separate specs.}
6409 Separate declarations (``specs'') are required for subprograms (a
6410 body is not allowed to serve as its own declaration). The only
6411 exception is that parameterless library level procedures are
6412 not required to have a separate declaration. This exception covers
6413 the most frequent form of main program procedures.
6416 @emph{Check token spacing.}
6417 The following token spacing rules are enforced:
6422 The keywords @code{@b{abs}} and @code{@b{not}} must be followed by a space.
6425 The token @code{=>} must be surrounded by spaces.
6428 The token @code{<>} must be preceded by a space or a left parenthesis.
6431 Binary operators other than @code{**} must be surrounded by spaces.
6432 There is no restriction on the layout of the @code{**} binary operator.
6435 Colon must be surrounded by spaces.
6438 Colon-equal (assignment, initialization) must be surrounded by spaces.
6441 Comma must be the first non-blank character on the line, or be
6442 immediately preceded by a non-blank character, and must be followed
6446 If the token preceding a left parenthesis ends with a letter or digit, then
6447 a space must separate the two tokens.
6450 if the token following a right parenthesis starts with a letter or digit, then
6451 a space must separate the two tokens.
6454 A right parenthesis must either be the first non-blank character on
6455 a line, or it must be preceded by a non-blank character.
6458 A semicolon must not be preceded by a space, and must not be followed by
6459 a non-blank character.
6462 A unary plus or minus may not be followed by a space.
6465 A vertical bar must be surrounded by spaces.
6468 @item ^u^UNNECESSARY_BLANK_LINES^
6469 @emph{Check unnecessary blank lines.}
6470 Unnecessary blank lines are not allowed. A blank line is considered
6471 unnecessary if it appears at the end of the file, or if more than
6472 one blank line occurs in sequence.
6474 @item ^x^XTRA_PARENS^
6475 @emph{Check extra parentheses.}
6476 Unnecessary extra level of parentheses (C-style) are not allowed
6477 around conditions in @code{if} statements, @code{while} statements and
6478 @code{exit} statements.
6480 @item ^y^ALL_BUILTIN^
6481 @emph{Set all standard style check options}
6482 This is equivalent to @code{gnaty3aAbcefhiklmnprst}, that is all checking
6483 options enabled with the exception of @option{-gnatyo}, @option{-gnatyI},
6484 @option{-gnatyS}, @option{-gnatyLnnn},
6485 @option{-gnatyd}, @option{-gnatyu}, and @option{-gnatyx}.
6489 @emph{Remove style check options}
6490 This causes any subsequent options in the string to act as canceling the
6491 corresponding style check option. To cancel maximum nesting level control,
6492 use @option{L} parameter witout any integer value after that, because any
6493 digit following @option{-} in the parameter string of the @option{-gnaty}
6494 option will be threated as canceling indentation check. The same is true
6495 for @option{M} parameter. @option{y} and @option{N} parameters are not
6496 allowed after @option{-}.
6499 This causes any subsequent options in the string to enable the corresponding
6500 style check option. That is, it cancels the effect of a previous ^-^REMOVE^,
6506 @emph{Removing style check options}
6507 If the name of a style check is preceded by @option{NO} then the corresponding
6508 style check is turned off. For example @option{NOCOMMENTS} turns off style
6509 checking for comments.
6514 In the above rules, appearing in column one is always permitted, that is,
6515 counts as meeting either a requirement for a required preceding space,
6516 or as meeting a requirement for no preceding space.
6518 Appearing at the end of a line is also always permitted, that is, counts
6519 as meeting either a requirement for a following space, or as meeting
6520 a requirement for no following space.
6523 If any of these style rules is violated, a message is generated giving
6524 details on the violation. The initial characters of such messages are
6525 always ``@code{(style)}''. Note that these messages are treated as warning
6526 messages, so they normally do not prevent the generation of an object
6527 file. The @option{-gnatwe} switch can be used to treat warning messages,
6528 including style messages, as fatal errors.
6532 @option{-gnaty} on its own (that is not
6533 followed by any letters or digits), then the effect is equivalent
6534 to the use of @option{-gnatyy}, as described above, that is all
6535 built-in standard style check options are enabled.
6539 /STYLE_CHECKS=ALL_BUILTIN enables all checking options with
6540 the exception of ORDERED_SUBPROGRAMS, UNNECESSARY_BLANK_LINES,
6541 XTRA_PARENS, and DOS_LINE_ENDINGS. In addition
6551 clears any previously set style checks.
6553 @node Run-Time Checks
6554 @subsection Run-Time Checks
6555 @cindex Division by zero
6556 @cindex Access before elaboration
6557 @cindex Checks, division by zero
6558 @cindex Checks, access before elaboration
6559 @cindex Checks, stack overflow checking
6562 By default, the following checks are suppressed: integer overflow
6563 checks, stack overflow checks, and checks for access before
6564 elaboration on subprogram calls. All other checks, including range
6565 checks and array bounds checks, are turned on by default. The
6566 following @command{gcc} switches refine this default behavior.
6571 @cindex @option{-gnatp} (@command{gcc})
6572 @cindex Suppressing checks
6573 @cindex Checks, suppressing
6575 This switch causes the unit to be compiled
6576 as though @code{pragma Suppress (All_checks)}
6577 had been present in the source. Validity checks are also eliminated (in
6578 other words @option{-gnatp} also implies @option{-gnatVn}.
6579 Use this switch to improve the performance
6580 of the code at the expense of safety in the presence of invalid data or
6583 Note that when checks are suppressed, the compiler is allowed, but not
6584 required, to omit the checking code. If the run-time cost of the
6585 checking code is zero or near-zero, the compiler will generate it even
6586 if checks are suppressed. In particular, if the compiler can prove
6587 that a certain check will necessarily fail, it will generate code to
6588 do an unconditional ``raise'', even if checks are suppressed. The
6589 compiler warns in this case. Another case in which checks may not be
6590 eliminated is when they are embedded in certain run time routines such
6591 as math library routines.
6593 Of course, run-time checks are omitted whenever the compiler can prove
6594 that they will not fail, whether or not checks are suppressed.
6596 Note that if you suppress a check that would have failed, program
6597 execution is erroneous, which means the behavior is totally
6598 unpredictable. The program might crash, or print wrong answers, or
6599 do anything else. It might even do exactly what you wanted it to do
6600 (and then it might start failing mysteriously next week or next
6601 year). The compiler will generate code based on the assumption that
6602 the condition being checked is true, which can result in disaster if
6603 that assumption is wrong.
6606 @cindex @option{-gnato} (@command{gcc})
6607 @cindex Overflow checks
6608 @cindex Check, overflow
6609 Enables overflow checking for integer operations.
6610 This causes GNAT to generate slower and larger executable
6611 programs by adding code to check for overflow (resulting in raising
6612 @code{Constraint_Error} as required by standard Ada
6613 semantics). These overflow checks correspond to situations in which
6614 the true value of the result of an operation may be outside the base
6615 range of the result type. The following example shows the distinction:
6617 @smallexample @c ada
6618 X1 : Integer := "Integer'Last";
6619 X2 : Integer range 1 .. 5 := "5";
6620 X3 : Integer := "Integer'Last";
6621 X4 : Integer range 1 .. 5 := "5";
6622 F : Float := "2.0E+20";
6631 Note that if explicit values are assigned at compile time, the
6632 compiler may be able to detect overflow at compile time, in which case
6633 no actual run-time checking code is required, and Constraint_Error
6634 will be raised unconditionally, with or without
6635 @option{-gnato}. That's why the assigned values in the above fragment
6636 are in quotes, the meaning is "assign a value not known to the
6637 compiler that happens to be equal to ...". The remaining discussion
6638 assumes that the compiler cannot detect the values at compile time.
6640 Here the first addition results in a value that is outside the base range
6641 of Integer, and hence requires an overflow check for detection of the
6642 constraint error. Thus the first assignment to @code{X1} raises a
6643 @code{Constraint_Error} exception only if @option{-gnato} is set.
6645 The second increment operation results in a violation of the explicit
6646 range constraint; such range checks are performed by default, and are
6647 unaffected by @option{-gnato}.
6649 The two conversions of @code{F} both result in values that are outside
6650 the base range of type @code{Integer} and thus will raise
6651 @code{Constraint_Error} exceptions only if @option{-gnato} is used.
6652 The fact that the result of the second conversion is assigned to
6653 variable @code{X4} with a restricted range is irrelevant, since the problem
6654 is in the conversion, not the assignment.
6656 Basically the rule is that in the default mode (@option{-gnato} not
6657 used), the generated code assures that all integer variables stay
6658 within their declared ranges, or within the base range if there is
6659 no declared range. This prevents any serious problems like indexes
6660 out of range for array operations.
6662 What is not checked in default mode is an overflow that results in
6663 an in-range, but incorrect value. In the above example, the assignments
6664 to @code{X1}, @code{X2}, @code{X3} all give results that are within the
6665 range of the target variable, but the result is wrong in the sense that
6666 it is too large to be represented correctly. Typically the assignment
6667 to @code{X1} will result in wrap around to the largest negative number.
6668 The conversions of @code{F} will result in some @code{Integer} value
6669 and if that integer value is out of the @code{X4} range then the
6670 subsequent assignment would generate an exception.
6672 @findex Machine_Overflows
6673 Note that the @option{-gnato} switch does not affect the code generated
6674 for any floating-point operations; it applies only to integer
6676 For floating-point, GNAT has the @code{Machine_Overflows}
6677 attribute set to @code{False} and the normal mode of operation is to
6678 generate IEEE NaN and infinite values on overflow or invalid operations
6679 (such as dividing 0.0 by 0.0).
6681 The reason that we distinguish overflow checking from other kinds of
6682 range constraint checking is that a failure of an overflow check, unlike
6683 for example the failure of a range check, can result in an incorrect
6684 value, but cannot cause random memory destruction (like an out of range
6685 subscript), or a wild jump (from an out of range case value). Overflow
6686 checking is also quite expensive in time and space, since in general it
6687 requires the use of double length arithmetic.
6689 Note again that @option{-gnato} is off by default, so overflow checking is
6690 not performed in default mode. This means that out of the box, with the
6691 default settings, GNAT does not do all the checks expected from the
6692 language description in the Ada Reference Manual. If you want all constraint
6693 checks to be performed, as described in this Manual, then you must
6694 explicitly use the -gnato switch either on the @command{gnatmake} or
6695 @command{gcc} command.
6698 @cindex @option{-gnatE} (@command{gcc})
6699 @cindex Elaboration checks
6700 @cindex Check, elaboration
6701 Enables dynamic checks for access-before-elaboration
6702 on subprogram calls and generic instantiations.
6703 Note that @option{-gnatE} is not necessary for safety, because in the
6704 default mode, GNAT ensures statically that the checks would not fail.
6705 For full details of the effect and use of this switch,
6706 @xref{Compiling Using gcc}.
6709 @cindex @option{-fstack-check} (@command{gcc})
6710 @cindex Stack Overflow Checking
6711 @cindex Checks, stack overflow checking
6712 Activates stack overflow checking. For full details of the effect and use of
6713 this switch see @ref{Stack Overflow Checking}.
6718 The setting of these switches only controls the default setting of the
6719 checks. You may modify them using either @code{Suppress} (to remove
6720 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
6723 @node Using gcc for Syntax Checking
6724 @subsection Using @command{gcc} for Syntax Checking
6727 @cindex @option{-gnats} (@command{gcc})
6731 The @code{s} stands for ``syntax''.
6734 Run GNAT in syntax checking only mode. For
6735 example, the command
6738 $ gcc -c -gnats x.adb
6742 compiles file @file{x.adb} in syntax-check-only mode. You can check a
6743 series of files in a single command
6745 , and can use wild cards to specify such a group of files.
6746 Note that you must specify the @option{-c} (compile
6747 only) flag in addition to the @option{-gnats} flag.
6750 You may use other switches in conjunction with @option{-gnats}. In
6751 particular, @option{-gnatl} and @option{-gnatv} are useful to control the
6752 format of any generated error messages.
6754 When the source file is empty or contains only empty lines and/or comments,
6755 the output is a warning:
6758 $ gcc -c -gnats -x ada toto.txt
6759 toto.txt:1:01: warning: empty file, contains no compilation units
6763 Otherwise, the output is simply the error messages, if any. No object file or
6764 ALI file is generated by a syntax-only compilation. Also, no units other
6765 than the one specified are accessed. For example, if a unit @code{X}
6766 @code{with}'s a unit @code{Y}, compiling unit @code{X} in syntax
6767 check only mode does not access the source file containing unit
6770 @cindex Multiple units, syntax checking
6771 Normally, GNAT allows only a single unit in a source file. However, this
6772 restriction does not apply in syntax-check-only mode, and it is possible
6773 to check a file containing multiple compilation units concatenated
6774 together. This is primarily used by the @code{gnatchop} utility
6775 (@pxref{Renaming Files Using gnatchop}).
6778 @node Using gcc for Semantic Checking
6779 @subsection Using @command{gcc} for Semantic Checking
6782 @cindex @option{-gnatc} (@command{gcc})
6786 The @code{c} stands for ``check''.
6788 Causes the compiler to operate in semantic check mode,
6789 with full checking for all illegalities specified in the
6790 Ada Reference Manual, but without generation of any object code
6791 (no object file is generated).
6793 Because dependent files must be accessed, you must follow the GNAT
6794 semantic restrictions on file structuring to operate in this mode:
6798 The needed source files must be accessible
6799 (@pxref{Search Paths and the Run-Time Library (RTL)}).
6802 Each file must contain only one compilation unit.
6805 The file name and unit name must match (@pxref{File Naming Rules}).
6808 The output consists of error messages as appropriate. No object file is
6809 generated. An @file{ALI} file is generated for use in the context of
6810 cross-reference tools, but this file is marked as not being suitable
6811 for binding (since no object file is generated).
6812 The checking corresponds exactly to the notion of
6813 legality in the Ada Reference Manual.
6815 Any unit can be compiled in semantics-checking-only mode, including
6816 units that would not normally be compiled (subunits,
6817 and specifications where a separate body is present).
6820 @node Compiling Different Versions of Ada
6821 @subsection Compiling Different Versions of Ada
6824 The switches described in this section allow you to explicitly specify
6825 the version of the Ada language that your programs are written in.
6826 By default @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
6827 but you can also specify @value{NONDEFAULTLANGUAGEVERSION} or
6828 indicate Ada 83 compatibility mode.
6831 @cindex Compatibility with Ada 83
6833 @item -gnat83 (Ada 83 Compatibility Mode)
6834 @cindex @option{-gnat83} (@command{gcc})
6835 @cindex ACVC, Ada 83 tests
6839 Although GNAT is primarily an Ada 95 / Ada 2005 compiler, this switch
6840 specifies that the program is to be compiled in Ada 83 mode. With
6841 @option{-gnat83}, GNAT rejects most post-Ada 83 extensions and applies Ada 83
6842 semantics where this can be done easily.
6843 It is not possible to guarantee this switch does a perfect
6844 job; some subtle tests, such as are
6845 found in earlier ACVC tests (and that have been removed from the ACATS suite
6846 for Ada 95), might not compile correctly.
6847 Nevertheless, this switch may be useful in some circumstances, for example
6848 where, due to contractual reasons, existing code needs to be maintained
6849 using only Ada 83 features.
6851 With few exceptions (most notably the need to use @code{<>} on
6852 @cindex Generic formal parameters
6853 unconstrained generic formal parameters, the use of the new Ada 95 / Ada 2005
6854 reserved words, and the use of packages
6855 with optional bodies), it is not necessary to specify the
6856 @option{-gnat83} switch when compiling Ada 83 programs, because, with rare
6857 exceptions, Ada 95 and Ada 2005 are upwardly compatible with Ada 83. Thus
6858 a correct Ada 83 program is usually also a correct program
6859 in these later versions of the language standard.
6860 For further information, please refer to @ref{Compatibility and Porting Guide}.
6862 @item -gnat95 (Ada 95 mode)
6863 @cindex @option{-gnat95} (@command{gcc})
6867 This switch directs the compiler to implement the Ada 95 version of the
6869 Since Ada 95 is almost completely upwards
6870 compatible with Ada 83, Ada 83 programs may generally be compiled using
6871 this switch (see the description of the @option{-gnat83} switch for further
6872 information about Ada 83 mode).
6873 If an Ada 2005 program is compiled in Ada 95 mode,
6874 uses of the new Ada 2005 features will cause error
6875 messages or warnings.
6877 This switch also can be used to cancel the effect of a previous
6878 @option{-gnat83} or @option{-gnat05} switch earlier in the command line.
6880 @item -gnat05 (Ada 2005 mode)
6881 @cindex @option{-gnat05} (@command{gcc})
6882 @cindex Ada 2005 mode
6885 This switch directs the compiler to implement the Ada 2005 version of the
6887 Since Ada 2005 is almost completely upwards
6888 compatible with Ada 95 (and thus also with Ada 83), Ada 83 and Ada 95 programs
6889 may generally be compiled using this switch (see the description of the
6890 @option{-gnat83} and @option{-gnat95} switches for further
6893 For information about the approved ``Ada Issues'' that have been incorporated
6894 into Ada 2005, see @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs}.
6895 Included with GNAT releases is a file @file{features-ada0y} that describes
6896 the set of implemented Ada 2005 features.
6900 @node Character Set Control
6901 @subsection Character Set Control
6903 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
6904 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
6907 Normally GNAT recognizes the Latin-1 character set in source program
6908 identifiers, as described in the Ada Reference Manual.
6910 GNAT to recognize alternate character sets in identifiers. @var{c} is a
6911 single character ^^or word^ indicating the character set, as follows:
6915 ISO 8859-1 (Latin-1) identifiers
6918 ISO 8859-2 (Latin-2) letters allowed in identifiers
6921 ISO 8859-3 (Latin-3) letters allowed in identifiers
6924 ISO 8859-4 (Latin-4) letters allowed in identifiers
6927 ISO 8859-5 (Cyrillic) letters allowed in identifiers
6930 ISO 8859-15 (Latin-9) letters allowed in identifiers
6933 IBM PC letters (code page 437) allowed in identifiers
6936 IBM PC letters (code page 850) allowed in identifiers
6938 @item ^f^FULL_UPPER^
6939 Full upper-half codes allowed in identifiers
6942 No upper-half codes allowed in identifiers
6945 Wide-character codes (that is, codes greater than 255)
6946 allowed in identifiers
6949 @xref{Foreign Language Representation}, for full details on the
6950 implementation of these character sets.
6952 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
6953 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
6954 Specify the method of encoding for wide characters.
6955 @var{e} is one of the following:
6960 Hex encoding (brackets coding also recognized)
6963 Upper half encoding (brackets encoding also recognized)
6966 Shift/JIS encoding (brackets encoding also recognized)
6969 EUC encoding (brackets encoding also recognized)
6972 UTF-8 encoding (brackets encoding also recognized)
6975 Brackets encoding only (default value)
6977 For full details on these encoding
6978 methods see @ref{Wide Character Encodings}.
6979 Note that brackets coding is always accepted, even if one of the other
6980 options is specified, so for example @option{-gnatW8} specifies that both
6981 brackets and UTF-8 encodings will be recognized. The units that are
6982 with'ed directly or indirectly will be scanned using the specified
6983 representation scheme, and so if one of the non-brackets scheme is
6984 used, it must be used consistently throughout the program. However,
6985 since brackets encoding is always recognized, it may be conveniently
6986 used in standard libraries, allowing these libraries to be used with
6987 any of the available coding schemes.
6990 If no @option{-gnatW?} parameter is present, then the default
6991 representation is normally Brackets encoding only. However, if the
6992 first three characters of the file are 16#EF# 16#BB# 16#BF# (the standard
6993 byte order mark or BOM for UTF-8), then these three characters are
6994 skipped and the default representation for the file is set to UTF-8.
6996 Note that the wide character representation that is specified (explicitly
6997 or by default) for the main program also acts as the default encoding used
6998 for Wide_Text_IO files if not specifically overridden by a WCEM form
7002 @node File Naming Control
7003 @subsection File Naming Control
7006 @item ^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{n}
7007 @cindex @option{-gnatk} (@command{gcc})
7008 Activates file name ``krunching''. @var{n}, a decimal integer in the range
7009 1-999, indicates the maximum allowable length of a file name (not
7010 including the @file{.ads} or @file{.adb} extension). The default is not
7011 to enable file name krunching.
7013 For the source file naming rules, @xref{File Naming Rules}.
7016 @node Subprogram Inlining Control
7017 @subsection Subprogram Inlining Control
7022 @cindex @option{-gnatn} (@command{gcc})
7024 The @code{n} here is intended to suggest the first syllable of the
7027 GNAT recognizes and processes @code{Inline} pragmas. However, for the
7028 inlining to actually occur, optimization must be enabled. To enable
7029 inlining of subprograms specified by pragma @code{Inline},
7030 you must also specify this switch.
7031 In the absence of this switch, GNAT does not attempt
7032 inlining and does not need to access the bodies of
7033 subprograms for which @code{pragma Inline} is specified if they are not
7034 in the current unit.
7036 If you specify this switch the compiler will access these bodies,
7037 creating an extra source dependency for the resulting object file, and
7038 where possible, the call will be inlined.
7039 For further details on when inlining is possible
7040 see @ref{Inlining of Subprograms}.
7043 @cindex @option{-gnatN} (@command{gcc})
7044 This switch activates front-end inlining which also
7045 generates additional dependencies.
7047 When using a gcc-based back end (in practice this means using any version
7048 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
7049 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
7050 Historically front end inlining was more extensive than the gcc back end
7051 inlining, but that is no longer the case.
7054 @node Auxiliary Output Control
7055 @subsection Auxiliary Output Control
7059 @cindex @option{-gnatt} (@command{gcc})
7060 @cindex Writing internal trees
7061 @cindex Internal trees, writing to file
7062 Causes GNAT to write the internal tree for a unit to a file (with the
7063 extension @file{.adt}.
7064 This not normally required, but is used by separate analysis tools.
7066 these tools do the necessary compilations automatically, so you should
7067 not have to specify this switch in normal operation.
7068 Note that the combination of switches @option{-gnatct}
7069 generates a tree in the form required by ASIS applications.
7072 @cindex @option{-gnatu} (@command{gcc})
7073 Print a list of units required by this compilation on @file{stdout}.
7074 The listing includes all units on which the unit being compiled depends
7075 either directly or indirectly.
7078 @item -pass-exit-codes
7079 @cindex @option{-pass-exit-codes} (@command{gcc})
7080 If this switch is not used, the exit code returned by @command{gcc} when
7081 compiling multiple files indicates whether all source files have
7082 been successfully used to generate object files or not.
7084 When @option{-pass-exit-codes} is used, @command{gcc} exits with an extended
7085 exit status and allows an integrated development environment to better
7086 react to a compilation failure. Those exit status are:
7090 There was an error in at least one source file.
7092 At least one source file did not generate an object file.
7094 The compiler died unexpectedly (internal error for example).
7096 An object file has been generated for every source file.
7101 @node Debugging Control
7102 @subsection Debugging Control
7106 @cindex Debugging options
7109 @cindex @option{-gnatd} (@command{gcc})
7110 Activate internal debugging switches. @var{x} is a letter or digit, or
7111 string of letters or digits, which specifies the type of debugging
7112 outputs desired. Normally these are used only for internal development
7113 or system debugging purposes. You can find full documentation for these
7114 switches in the body of the @code{Debug} unit in the compiler source
7115 file @file{debug.adb}.
7119 @cindex @option{-gnatG} (@command{gcc})
7120 This switch causes the compiler to generate auxiliary output containing
7121 a pseudo-source listing of the generated expanded code. Like most Ada
7122 compilers, GNAT works by first transforming the high level Ada code into
7123 lower level constructs. For example, tasking operations are transformed
7124 into calls to the tasking run-time routines. A unique capability of GNAT
7125 is to list this expanded code in a form very close to normal Ada source.
7126 This is very useful in understanding the implications of various Ada
7127 usage on the efficiency of the generated code. There are many cases in
7128 Ada (e.g.@: the use of controlled types), where simple Ada statements can
7129 generate a lot of run-time code. By using @option{-gnatG} you can identify
7130 these cases, and consider whether it may be desirable to modify the coding
7131 approach to improve efficiency.
7133 The optional parameter @code{nn} if present after -gnatG specifies an
7134 alternative maximum line length that overrides the normal default of 72.
7135 This value is in the range 40-999999, values less than 40 being silently
7136 reset to 40. The equal sign is optional.
7138 The format of the output is very similar to standard Ada source, and is
7139 easily understood by an Ada programmer. The following special syntactic
7140 additions correspond to low level features used in the generated code that
7141 do not have any exact analogies in pure Ada source form. The following
7142 is a partial list of these special constructions. See the spec
7143 of package @code{Sprint} in file @file{sprint.ads} for a full list.
7145 If the switch @option{-gnatL} is used in conjunction with
7146 @cindex @option{-gnatL} (@command{gcc})
7147 @option{-gnatG}, then the original source lines are interspersed
7148 in the expanded source (as comment lines with the original line number).
7151 @item new @var{xxx} @r{[}storage_pool = @var{yyy}@r{]}
7152 Shows the storage pool being used for an allocator.
7154 @item at end @var{procedure-name};
7155 Shows the finalization (cleanup) procedure for a scope.
7157 @item (if @var{expr} then @var{expr} else @var{expr})
7158 Conditional expression equivalent to the @code{x?y:z} construction in C.
7160 @item @var{target}^^^(@var{source})
7161 A conversion with floating-point truncation instead of rounding.
7163 @item @var{target}?(@var{source})
7164 A conversion that bypasses normal Ada semantic checking. In particular
7165 enumeration types and fixed-point types are treated simply as integers.
7167 @item @var{target}?^^^(@var{source})
7168 Combines the above two cases.
7170 @item @var{x} #/ @var{y}
7171 @itemx @var{x} #mod @var{y}
7172 @itemx @var{x} #* @var{y}
7173 @itemx @var{x} #rem @var{y}
7174 A division or multiplication of fixed-point values which are treated as
7175 integers without any kind of scaling.
7177 @item free @var{expr} @r{[}storage_pool = @var{xxx}@r{]}
7178 Shows the storage pool associated with a @code{free} statement.
7180 @item [subtype or type declaration]
7181 Used to list an equivalent declaration for an internally generated
7182 type that is referenced elsewhere in the listing.
7184 @item freeze @var{type-name} @ovar{actions}
7185 Shows the point at which @var{type-name} is frozen, with possible
7186 associated actions to be performed at the freeze point.
7188 @item reference @var{itype}
7189 Reference (and hence definition) to internal type @var{itype}.
7191 @item @var{function-name}! (@var{arg}, @var{arg}, @var{arg})
7192 Intrinsic function call.
7194 @item @var{label-name} : label
7195 Declaration of label @var{labelname}.
7197 @item #$ @var{subprogram-name}
7198 An implicit call to a run-time support routine
7199 (to meet the requirement of H.3.1(9) in a
7202 @item @var{expr} && @var{expr} && @var{expr} @dots{} && @var{expr}
7203 A multiple concatenation (same effect as @var{expr} & @var{expr} &
7204 @var{expr}, but handled more efficiently).
7206 @item [constraint_error]
7207 Raise the @code{Constraint_Error} exception.
7209 @item @var{expression}'reference
7210 A pointer to the result of evaluating @var{expression}.
7212 @item @var{target-type}!(@var{source-expression})
7213 An unchecked conversion of @var{source-expression} to @var{target-type}.
7215 @item [@var{numerator}/@var{denominator}]
7216 Used to represent internal real literals (that) have no exact
7217 representation in base 2-16 (for example, the result of compile time
7218 evaluation of the expression 1.0/27.0).
7222 @cindex @option{-gnatD} (@command{gcc})
7223 When used in conjunction with @option{-gnatG}, this switch causes
7224 the expanded source, as described above for
7225 @option{-gnatG} to be written to files with names
7226 @file{^xxx.dg^XXX_DG^}, where @file{xxx} is the normal file name,
7227 instead of to the standard output file. For
7228 example, if the source file name is @file{hello.adb}, then a file
7229 @file{^hello.adb.dg^HELLO.ADB_DG^} will be written. The debugging
7230 information generated by the @command{gcc} @option{^-g^/DEBUG^} switch
7231 will refer to the generated @file{^xxx.dg^XXX_DG^} file. This allows
7232 you to do source level debugging using the generated code which is
7233 sometimes useful for complex code, for example to find out exactly
7234 which part of a complex construction raised an exception. This switch
7235 also suppress generation of cross-reference information (see
7236 @option{-gnatx}) since otherwise the cross-reference information
7237 would refer to the @file{^.dg^.DG^} file, which would cause
7238 confusion since this is not the original source file.
7240 Note that @option{-gnatD} actually implies @option{-gnatG}
7241 automatically, so it is not necessary to give both options.
7242 In other words @option{-gnatD} is equivalent to @option{-gnatDG}).
7244 If the switch @option{-gnatL} is used in conjunction with
7245 @cindex @option{-gnatL} (@command{gcc})
7246 @option{-gnatDG}, then the original source lines are interspersed
7247 in the expanded source (as comment lines with the original line number).
7249 The optional parameter @code{nn} if present after -gnatD specifies an
7250 alternative maximum line length that overrides the normal default of 72.
7251 This value is in the range 40-999999, values less than 40 being silently
7252 reset to 40. The equal sign is optional.
7255 @cindex @option{-gnatr} (@command{gcc})
7256 @cindex pragma Restrictions
7257 This switch causes pragma Restrictions to be treated as Restriction_Warnings
7258 so that violation of restrictions causes warnings rather than illegalities.
7259 This is useful during the development process when new restrictions are added
7260 or investigated. The switch also causes pragma Profile to be treated as
7261 Profile_Warnings, and pragma Restricted_Run_Time and pragma Ravenscar set
7262 restriction warnings rather than restrictions.
7265 @item -gnatR@r{[}0@r{|}1@r{|}2@r{|}3@r{[}s@r{]]}
7266 @cindex @option{-gnatR} (@command{gcc})
7267 This switch controls output from the compiler of a listing showing
7268 representation information for declared types and objects. For
7269 @option{-gnatR0}, no information is output (equivalent to omitting
7270 the @option{-gnatR} switch). For @option{-gnatR1} (which is the default,
7271 so @option{-gnatR} with no parameter has the same effect), size and alignment
7272 information is listed for declared array and record types. For
7273 @option{-gnatR2}, size and alignment information is listed for all
7274 declared types and objects. Finally @option{-gnatR3} includes symbolic
7275 expressions for values that are computed at run time for
7276 variant records. These symbolic expressions have a mostly obvious
7277 format with #n being used to represent the value of the n'th
7278 discriminant. See source files @file{repinfo.ads/adb} in the
7279 @code{GNAT} sources for full details on the format of @option{-gnatR3}
7280 output. If the switch is followed by an s (e.g.@: @option{-gnatR2s}), then
7281 the output is to a file with the name @file{^file.rep^file_REP^} where
7282 file is the name of the corresponding source file.
7285 @item /REPRESENTATION_INFO
7286 @cindex @option{/REPRESENTATION_INFO} (@command{gcc})
7287 This qualifier controls output from the compiler of a listing showing
7288 representation information for declared types and objects. For
7289 @option{/REPRESENTATION_INFO=NONE}, no information is output
7290 (equivalent to omitting the @option{/REPRESENTATION_INFO} qualifier).
7291 @option{/REPRESENTATION_INFO} without option is equivalent to
7292 @option{/REPRESENTATION_INFO=ARRAYS}.
7293 For @option{/REPRESENTATION_INFO=ARRAYS}, size and alignment
7294 information is listed for declared array and record types. For
7295 @option{/REPRESENTATION_INFO=OBJECTS}, size and alignment information
7296 is listed for all expression information for values that are computed
7297 at run time for variant records. These symbolic expressions have a mostly
7298 obvious format with #n being used to represent the value of the n'th
7299 discriminant. See source files @file{REPINFO.ADS/ADB} in the
7300 @code{GNAT} sources for full details on the format of
7301 @option{/REPRESENTATION_INFO=SYMBOLIC} output.
7302 If _FILE is added at the end of an option
7303 (e.g.@: @option{/REPRESENTATION_INFO=ARRAYS_FILE}),
7304 then the output is to a file with the name @file{file_REP} where
7305 file is the name of the corresponding source file.
7307 Note that it is possible for record components to have zero size. In
7308 this case, the component clause uses an obvious extension of permitted
7309 Ada syntax, for example @code{at 0 range 0 .. -1}.
7311 Representation information requires that code be generated (since it is the
7312 code generator that lays out complex data structures). If an attempt is made
7313 to output representation information when no code is generated, for example
7314 when a subunit is compiled on its own, then no information can be generated
7315 and the compiler outputs a message to this effect.
7318 @cindex @option{-gnatS} (@command{gcc})
7319 The use of the switch @option{-gnatS} for an
7320 Ada compilation will cause the compiler to output a
7321 representation of package Standard in a form very
7322 close to standard Ada. It is not quite possible to
7323 do this entirely in standard Ada (since new
7324 numeric base types cannot be created in standard
7325 Ada), but the output is easily
7326 readable to any Ada programmer, and is useful to
7327 determine the characteristics of target dependent
7328 types in package Standard.
7331 @cindex @option{-gnatx} (@command{gcc})
7332 Normally the compiler generates full cross-referencing information in
7333 the @file{ALI} file. This information is used by a number of tools,
7334 including @code{gnatfind} and @code{gnatxref}. The @option{-gnatx} switch
7335 suppresses this information. This saves some space and may slightly
7336 speed up compilation, but means that these tools cannot be used.
7339 @node Exception Handling Control
7340 @subsection Exception Handling Control
7343 GNAT uses two methods for handling exceptions at run-time. The
7344 @code{setjmp/longjmp} method saves the context when entering
7345 a frame with an exception handler. Then when an exception is
7346 raised, the context can be restored immediately, without the
7347 need for tracing stack frames. This method provides very fast
7348 exception propagation, but introduces significant overhead for
7349 the use of exception handlers, even if no exception is raised.
7351 The other approach is called ``zero cost'' exception handling.
7352 With this method, the compiler builds static tables to describe
7353 the exception ranges. No dynamic code is required when entering
7354 a frame containing an exception handler. When an exception is
7355 raised, the tables are used to control a back trace of the
7356 subprogram invocation stack to locate the required exception
7357 handler. This method has considerably poorer performance for
7358 the propagation of exceptions, but there is no overhead for
7359 exception handlers if no exception is raised. Note that in this
7360 mode and in the context of mixed Ada and C/C++ programming,
7361 to propagate an exception through a C/C++ code, the C/C++ code
7362 must be compiled with the @option{-funwind-tables} GCC's
7365 The following switches may be used to control which of the
7366 two exception handling methods is used.
7372 @cindex @option{--RTS=sjlj} (@command{gnatmake})
7373 This switch causes the setjmp/longjmp run-time (when available) to be used
7374 for exception handling. If the default
7375 mechanism for the target is zero cost exceptions, then
7376 this switch can be used to modify this default, and must be
7377 used for all units in the partition.
7378 This option is rarely used. One case in which it may be
7379 advantageous is if you have an application where exception
7380 raising is common and the overall performance of the
7381 application is improved by favoring exception propagation.
7384 @cindex @option{--RTS=zcx} (@command{gnatmake})
7385 @cindex Zero Cost Exceptions
7386 This switch causes the zero cost approach to be used
7387 for exception handling. If this is the default mechanism for the
7388 target (see below), then this switch is unneeded. If the default
7389 mechanism for the target is setjmp/longjmp exceptions, then
7390 this switch can be used to modify this default, and must be
7391 used for all units in the partition.
7392 This option can only be used if the zero cost approach
7393 is available for the target in use, otherwise it will generate an error.
7397 The same option @option{--RTS} must be used both for @command{gcc}
7398 and @command{gnatbind}. Passing this option to @command{gnatmake}
7399 (@pxref{Switches for gnatmake}) will ensure the required consistency
7400 through the compilation and binding steps.
7402 @node Units to Sources Mapping Files
7403 @subsection Units to Sources Mapping Files
7407 @item -gnatem=@var{path}
7408 @cindex @option{-gnatem} (@command{gcc})
7409 A mapping file is a way to communicate to the compiler two mappings:
7410 from unit names to file names (without any directory information) and from
7411 file names to path names (with full directory information). These mappings
7412 are used by the compiler to short-circuit the path search.
7414 The use of mapping files is not required for correct operation of the
7415 compiler, but mapping files can improve efficiency, particularly when
7416 sources are read over a slow network connection. In normal operation,
7417 you need not be concerned with the format or use of mapping files,
7418 and the @option{-gnatem} switch is not a switch that you would use
7419 explicitly. It is intended primarily for use by automatic tools such as
7420 @command{gnatmake} running under the project file facility. The
7421 description here of the format of mapping files is provided
7422 for completeness and for possible use by other tools.
7424 A mapping file is a sequence of sets of three lines. In each set, the
7425 first line is the unit name, in lower case, with @code{%s} appended
7426 for specs and @code{%b} appended for bodies; the second line is the
7427 file name; and the third line is the path name.
7433 /gnat/project1/sources/main.2.ada
7436 When the switch @option{-gnatem} is specified, the compiler will
7437 create in memory the two mappings from the specified file. If there is
7438 any problem (nonexistent file, truncated file or duplicate entries),
7439 no mapping will be created.
7441 Several @option{-gnatem} switches may be specified; however, only the
7442 last one on the command line will be taken into account.
7444 When using a project file, @command{gnatmake} creates a temporary
7445 mapping file and communicates it to the compiler using this switch.
7449 @node Integrated Preprocessing
7450 @subsection Integrated Preprocessing
7453 GNAT sources may be preprocessed immediately before compilation.
7454 In this case, the actual
7455 text of the source is not the text of the source file, but is derived from it
7456 through a process called preprocessing. Integrated preprocessing is specified
7457 through switches @option{-gnatep} and/or @option{-gnateD}. @option{-gnatep}
7458 indicates, through a text file, the preprocessing data to be used.
7459 @option{-gnateD} specifies or modifies the values of preprocessing symbol.
7462 Note that when integrated preprocessing is used, the output from the
7463 preprocessor is not written to any external file. Instead it is passed
7464 internally to the compiler. If you need to preserve the result of
7465 preprocessing in a file, then you should use @command{gnatprep}
7466 to perform the desired preprocessing in stand-alone mode.
7469 It is recommended that @command{gnatmake} switch ^-s^/SWITCH_CHECK^ should be
7470 used when Integrated Preprocessing is used. The reason is that preprocessing
7471 with another Preprocessing Data file without changing the sources will
7472 not trigger recompilation without this switch.
7475 Note that @command{gnatmake} switch ^-m^/MINIMAL_RECOMPILATION^ will almost
7476 always trigger recompilation for sources that are preprocessed,
7477 because @command{gnatmake} cannot compute the checksum of the source after
7481 The actual preprocessing function is described in details in section
7482 @ref{Preprocessing Using gnatprep}. This section only describes how integrated
7483 preprocessing is triggered and parameterized.
7487 @item -gnatep=@var{file}
7488 @cindex @option{-gnatep} (@command{gcc})
7489 This switch indicates to the compiler the file name (without directory
7490 information) of the preprocessor data file to use. The preprocessor data file
7491 should be found in the source directories.
7494 A preprocessing data file is a text file with significant lines indicating
7495 how should be preprocessed either a specific source or all sources not
7496 mentioned in other lines. A significant line is a nonempty, non-comment line.
7497 Comments are similar to Ada comments.
7500 Each significant line starts with either a literal string or the character '*'.
7501 A literal string is the file name (without directory information) of the source
7502 to preprocess. A character '*' indicates the preprocessing for all the sources
7503 that are not specified explicitly on other lines (order of the lines is not
7504 significant). It is an error to have two lines with the same file name or two
7505 lines starting with the character '*'.
7508 After the file name or the character '*', another optional literal string
7509 indicating the file name of the definition file to be used for preprocessing
7510 (@pxref{Form of Definitions File}). The definition files are found by the
7511 compiler in one of the source directories. In some cases, when compiling
7512 a source in a directory other than the current directory, if the definition
7513 file is in the current directory, it may be necessary to add the current
7514 directory as a source directory through switch ^-I.^/SEARCH=[]^, otherwise
7515 the compiler would not find the definition file.
7518 Then, optionally, ^switches^switches^ similar to those of @code{gnatprep} may
7519 be found. Those ^switches^switches^ are:
7524 Causes both preprocessor lines and the lines deleted by
7525 preprocessing to be replaced by blank lines, preserving the line number.
7526 This ^switch^switch^ is always implied; however, if specified after @option{-c}
7527 it cancels the effect of @option{-c}.
7530 Causes both preprocessor lines and the lines deleted
7531 by preprocessing to be retained as comments marked
7532 with the special string ``@code{--! }''.
7534 @item -Dsymbol=value
7535 Define or redefine a symbol, associated with value. A symbol is an Ada
7536 identifier, or an Ada reserved word, with the exception of @code{if},
7537 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7538 @code{value} is either a literal string, an Ada identifier or any Ada reserved
7539 word. A symbol declared with this ^switch^switch^ replaces a symbol with the
7540 same name defined in a definition file.
7543 Causes a sorted list of symbol names and values to be
7544 listed on the standard output file.
7547 Causes undefined symbols to be treated as having the value @code{FALSE}
7549 of a preprocessor test. In the absence of this option, an undefined symbol in
7550 a @code{#if} or @code{#elsif} test will be treated as an error.
7555 Examples of valid lines in a preprocessor data file:
7558 "toto.adb" "prep.def" -u
7559 -- preprocess "toto.adb", using definition file "prep.def",
7560 -- undefined symbol are False.
7563 -- preprocess all other sources without a definition file;
7564 -- suppressed lined are commented; symbol VERSION has the value V101.
7566 "titi.adb" "prep2.def" -s
7567 -- preprocess "titi.adb", using definition file "prep2.def";
7568 -- list all symbols with their values.
7571 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=value@r{]}
7572 @cindex @option{-gnateD} (@command{gcc})
7573 Define or redefine a preprocessing symbol, associated with value. If no value
7574 is given on the command line, then the value of the symbol is @code{True}.
7575 A symbol is an identifier, following normal Ada (case-insensitive)
7576 rules for its syntax, and value is any sequence (including an empty sequence)
7577 of characters from the set (letters, digits, period, underline).
7578 Ada reserved words may be used as symbols, with the exceptions of @code{if},
7579 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7582 A symbol declared with this ^switch^switch^ on the command line replaces a
7583 symbol with the same name either in a definition file or specified with a
7584 ^switch^switch^ -D in the preprocessor data file.
7587 This switch is similar to switch @option{^-D^/ASSOCIATE^} of @code{gnatprep}.
7590 When integrated preprocessing is performed and the preprocessor modifies
7591 the source text, write the result of this preprocessing into a file
7592 <source>^.prep^_prep^.
7596 @node Code Generation Control
7597 @subsection Code Generation Control
7601 The GCC technology provides a wide range of target dependent
7602 @option{-m} switches for controlling
7603 details of code generation with respect to different versions of
7604 architectures. This includes variations in instruction sets (e.g.@:
7605 different members of the power pc family), and different requirements
7606 for optimal arrangement of instructions (e.g.@: different members of
7607 the x86 family). The list of available @option{-m} switches may be
7608 found in the GCC documentation.
7610 Use of these @option{-m} switches may in some cases result in improved
7613 The GNAT Pro technology is tested and qualified without any
7614 @option{-m} switches,
7615 so generally the most reliable approach is to avoid the use of these
7616 switches. However, we generally expect most of these switches to work
7617 successfully with GNAT Pro, and many customers have reported successful
7618 use of these options.
7620 Our general advice is to avoid the use of @option{-m} switches unless
7621 special needs lead to requirements in this area. In particular,
7622 there is no point in using @option{-m} switches to improve performance
7623 unless you actually see a performance improvement.
7627 @subsection Return Codes
7628 @cindex Return Codes
7629 @cindex @option{/RETURN_CODES=VMS}
7632 On VMS, GNAT compiled programs return POSIX-style codes by default,
7633 e.g.@: @option{/RETURN_CODES=POSIX}.
7635 To enable VMS style return codes, use GNAT BIND and LINK with the option
7636 @option{/RETURN_CODES=VMS}. For example:
7639 GNAT BIND MYMAIN.ALI /RETURN_CODES=VMS
7640 GNAT LINK MYMAIN.ALI /RETURN_CODES=VMS
7644 Programs built with /RETURN_CODES=VMS are suitable to be called in
7645 VMS DCL scripts. Programs compiled with the default /RETURN_CODES=POSIX
7646 are suitable for spawning with appropriate GNAT RTL routines.
7650 @node Search Paths and the Run-Time Library (RTL)
7651 @section Search Paths and the Run-Time Library (RTL)
7654 With the GNAT source-based library system, the compiler must be able to
7655 find source files for units that are needed by the unit being compiled.
7656 Search paths are used to guide this process.
7658 The compiler compiles one source file whose name must be given
7659 explicitly on the command line. In other words, no searching is done
7660 for this file. To find all other source files that are needed (the most
7661 common being the specs of units), the compiler examines the following
7662 directories, in the following order:
7666 The directory containing the source file of the main unit being compiled
7667 (the file name on the command line).
7670 Each directory named by an @option{^-I^/SOURCE_SEARCH^} switch given on the
7671 @command{gcc} command line, in the order given.
7674 @findex ADA_PRJ_INCLUDE_FILE
7675 Each of the directories listed in the text file whose name is given
7676 by the @env{ADA_PRJ_INCLUDE_FILE} ^environment variable^logical name^.
7679 @env{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
7680 driver when project files are used. It should not normally be set
7684 @findex ADA_INCLUDE_PATH
7685 Each of the directories listed in the value of the
7686 @env{ADA_INCLUDE_PATH} ^environment variable^logical name^.
7688 Construct this value
7689 exactly as the @env{PATH} environment variable: a list of directory
7690 names separated by colons (semicolons when working with the NT version).
7693 Normally, define this value as a logical name containing a comma separated
7694 list of directory names.
7696 This variable can also be defined by means of an environment string
7697 (an argument to the HP C exec* set of functions).
7701 DEFINE ANOTHER_PATH FOO:[BAG]
7702 DEFINE ADA_INCLUDE_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
7705 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
7706 first, followed by the standard Ada
7707 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADAINCLUDE].
7708 If this is not redefined, the user will obtain the HP Ada 83 IO packages
7709 (Text_IO, Sequential_IO, etc)
7710 instead of the standard Ada packages. Thus, in order to get the standard Ada
7711 packages by default, ADA_INCLUDE_PATH must be redefined.
7715 The content of the @file{ada_source_path} file which is part of the GNAT
7716 installation tree and is used to store standard libraries such as the
7717 GNAT Run Time Library (RTL) source files.
7719 @ref{Installing a library}
7724 Specifying the switch @option{^-I-^/NOCURRENT_DIRECTORY^}
7725 inhibits the use of the directory
7726 containing the source file named in the command line. You can still
7727 have this directory on your search path, but in this case it must be
7728 explicitly requested with a @option{^-I^/SOURCE_SEARCH^} switch.
7730 Specifying the switch @option{-nostdinc}
7731 inhibits the search of the default location for the GNAT Run Time
7732 Library (RTL) source files.
7734 The compiler outputs its object files and ALI files in the current
7737 Caution: The object file can be redirected with the @option{-o} switch;
7738 however, @command{gcc} and @code{gnat1} have not been coordinated on this
7739 so the @file{ALI} file will not go to the right place. Therefore, you should
7740 avoid using the @option{-o} switch.
7744 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
7745 children make up the GNAT RTL, together with the simple @code{System.IO}
7746 package used in the @code{"Hello World"} example. The sources for these units
7747 are needed by the compiler and are kept together in one directory. Not
7748 all of the bodies are needed, but all of the sources are kept together
7749 anyway. In a normal installation, you need not specify these directory
7750 names when compiling or binding. Either the environment variables or
7751 the built-in defaults cause these files to be found.
7753 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
7754 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
7755 consisting of child units of @code{GNAT}. This is a collection of generally
7756 useful types, subprograms, etc. @xref{Top, GNAT Reference Manual, About
7757 This Guid, gnat_rm, GNAT Reference Manual}, for further details.
7759 Besides simplifying access to the RTL, a major use of search paths is
7760 in compiling sources from multiple directories. This can make
7761 development environments much more flexible.
7763 @node Order of Compilation Issues
7764 @section Order of Compilation Issues
7767 If, in our earlier example, there was a spec for the @code{hello}
7768 procedure, it would be contained in the file @file{hello.ads}; yet this
7769 file would not have to be explicitly compiled. This is the result of the
7770 model we chose to implement library management. Some of the consequences
7771 of this model are as follows:
7775 There is no point in compiling specs (except for package
7776 specs with no bodies) because these are compiled as needed by clients. If
7777 you attempt a useless compilation, you will receive an error message.
7778 It is also useless to compile subunits because they are compiled as needed
7782 There are no order of compilation requirements: performing a
7783 compilation never obsoletes anything. The only way you can obsolete
7784 something and require recompilations is to modify one of the
7785 source files on which it depends.
7788 There is no library as such, apart from the ALI files
7789 (@pxref{The Ada Library Information Files}, for information on the format
7790 of these files). For now we find it convenient to create separate ALI files,
7791 but eventually the information therein may be incorporated into the object
7795 When you compile a unit, the source files for the specs of all units
7796 that it @code{with}'s, all its subunits, and the bodies of any generics it
7797 instantiates must be available (reachable by the search-paths mechanism
7798 described above), or you will receive a fatal error message.
7805 The following are some typical Ada compilation command line examples:
7808 @item $ gcc -c xyz.adb
7809 Compile body in file @file{xyz.adb} with all default options.
7812 @item $ gcc -c -O2 -gnata xyz-def.adb
7815 @item $ GNAT COMPILE /OPTIMIZE=ALL -gnata xyz-def.adb
7818 Compile the child unit package in file @file{xyz-def.adb} with extensive
7819 optimizations, and pragma @code{Assert}/@code{Debug} statements
7822 @item $ gcc -c -gnatc abc-def.adb
7823 Compile the subunit in file @file{abc-def.adb} in semantic-checking-only
7827 @node Binding Using gnatbind
7828 @chapter Binding Using @code{gnatbind}
7832 * Running gnatbind::
7833 * Switches for gnatbind::
7834 * Command-Line Access::
7835 * Search Paths for gnatbind::
7836 * Examples of gnatbind Usage::
7840 This chapter describes the GNAT binder, @code{gnatbind}, which is used
7841 to bind compiled GNAT objects.
7843 Note: to invoke @code{gnatbind} with a project file, use the @code{gnat}
7844 driver (see @ref{The GNAT Driver and Project Files}).
7846 The @code{gnatbind} program performs four separate functions:
7850 Checks that a program is consistent, in accordance with the rules in
7851 Chapter 10 of the Ada Reference Manual. In particular, error
7852 messages are generated if a program uses inconsistent versions of a
7856 Checks that an acceptable order of elaboration exists for the program
7857 and issues an error message if it cannot find an order of elaboration
7858 that satisfies the rules in Chapter 10 of the Ada Language Manual.
7861 Generates a main program incorporating the given elaboration order.
7862 This program is a small Ada package (body and spec) that
7863 must be subsequently compiled
7864 using the GNAT compiler. The necessary compilation step is usually
7865 performed automatically by @command{gnatlink}. The two most important
7866 functions of this program
7867 are to call the elaboration routines of units in an appropriate order
7868 and to call the main program.
7871 Determines the set of object files required by the given main program.
7872 This information is output in the forms of comments in the generated program,
7873 to be read by the @command{gnatlink} utility used to link the Ada application.
7876 @node Running gnatbind
7877 @section Running @code{gnatbind}
7880 The form of the @code{gnatbind} command is
7883 $ gnatbind @ovar{switches} @var{mainprog}@r{[}.ali@r{]} @ovar{switches}
7887 where @file{@var{mainprog}.adb} is the Ada file containing the main program
7888 unit body. If no switches are specified, @code{gnatbind} constructs an Ada
7889 package in two files whose names are
7890 @file{b~@var{mainprog}.ads}, and @file{b~@var{mainprog}.adb}.
7891 For example, if given the
7892 parameter @file{hello.ali}, for a main program contained in file
7893 @file{hello.adb}, the binder output files would be @file{b~hello.ads}
7894 and @file{b~hello.adb}.
7896 When doing consistency checking, the binder takes into consideration
7897 any source files it can locate. For example, if the binder determines
7898 that the given main program requires the package @code{Pack}, whose
7900 file is @file{pack.ali} and whose corresponding source spec file is
7901 @file{pack.ads}, it attempts to locate the source file @file{pack.ads}
7902 (using the same search path conventions as previously described for the
7903 @command{gcc} command). If it can locate this source file, it checks that
7905 or source checksums of the source and its references to in @file{ALI} files
7906 match. In other words, any @file{ALI} files that mentions this spec must have
7907 resulted from compiling this version of the source file (or in the case
7908 where the source checksums match, a version close enough that the
7909 difference does not matter).
7911 @cindex Source files, use by binder
7912 The effect of this consistency checking, which includes source files, is
7913 that the binder ensures that the program is consistent with the latest
7914 version of the source files that can be located at bind time. Editing a
7915 source file without compiling files that depend on the source file cause
7916 error messages to be generated by the binder.
7918 For example, suppose you have a main program @file{hello.adb} and a
7919 package @code{P}, from file @file{p.ads} and you perform the following
7924 Enter @code{gcc -c hello.adb} to compile the main program.
7927 Enter @code{gcc -c p.ads} to compile package @code{P}.
7930 Edit file @file{p.ads}.
7933 Enter @code{gnatbind hello}.
7937 At this point, the file @file{p.ali} contains an out-of-date time stamp
7938 because the file @file{p.ads} has been edited. The attempt at binding
7939 fails, and the binder generates the following error messages:
7942 error: "hello.adb" must be recompiled ("p.ads" has been modified)
7943 error: "p.ads" has been modified and must be recompiled
7947 Now both files must be recompiled as indicated, and then the bind can
7948 succeed, generating a main program. You need not normally be concerned
7949 with the contents of this file, but for reference purposes a sample
7950 binder output file is given in @ref{Example of Binder Output File}.
7952 In most normal usage, the default mode of @command{gnatbind} which is to
7953 generate the main package in Ada, as described in the previous section.
7954 In particular, this means that any Ada programmer can read and understand
7955 the generated main program. It can also be debugged just like any other
7956 Ada code provided the @option{^-g^/DEBUG^} switch is used for
7957 @command{gnatbind} and @command{gnatlink}.
7959 However for some purposes it may be convenient to generate the main
7960 program in C rather than Ada. This may for example be helpful when you
7961 are generating a mixed language program with the main program in C. The
7962 GNAT compiler itself is an example.
7963 The use of the @option{^-C^/BIND_FILE=C^} switch
7964 for both @code{gnatbind} and @command{gnatlink} will cause the program to
7965 be generated in C (and compiled using the gnu C compiler).
7967 @node Switches for gnatbind
7968 @section Switches for @command{gnatbind}
7971 The following switches are available with @code{gnatbind}; details will
7972 be presented in subsequent sections.
7975 * Consistency-Checking Modes::
7976 * Binder Error Message Control::
7977 * Elaboration Control::
7979 * Binding with Non-Ada Main Programs::
7980 * Binding Programs with No Main Subprogram::
7987 @cindex @option{--version} @command{gnatbind}
7988 Display Copyright and version, then exit disregarding all other options.
7991 @cindex @option{--help} @command{gnatbind}
7992 If @option{--version} was not used, display usage, then exit disregarding
7996 @cindex @option{-a} @command{gnatbind}
7997 Indicates that, if supported by the platform, the adainit procedure should
7998 be treated as an initialisation routine by the linker (a constructor). This
7999 is intended to be used by the Project Manager to automatically initialize
8000 shared Stand-Alone Libraries.
8002 @item ^-aO^/OBJECT_SEARCH^
8003 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatbind})
8004 Specify directory to be searched for ALI files.
8006 @item ^-aI^/SOURCE_SEARCH^
8007 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8008 Specify directory to be searched for source file.
8010 @item ^-A^/BIND_FILE=ADA^
8011 @cindex @option{^-A^/BIND_FILE=ADA^} (@command{gnatbind})
8012 Generate binder program in Ada (default)
8014 @item ^-b^/REPORT_ERRORS=BRIEF^
8015 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@command{gnatbind})
8016 Generate brief messages to @file{stderr} even if verbose mode set.
8018 @item ^-c^/NOOUTPUT^
8019 @cindex @option{^-c^/NOOUTPUT^} (@command{gnatbind})
8020 Check only, no generation of binder output file.
8022 @item ^-C^/BIND_FILE=C^
8023 @cindex @option{^-C^/BIND_FILE=C^} (@command{gnatbind})
8024 Generate binder program in C
8026 @item ^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8027 @cindex @option{^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}} (@command{gnatbind})
8028 This switch can be used to change the default task stack size value
8029 to a specified size @var{nn}, which is expressed in bytes by default, or
8030 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8032 In the absence of a @samp{@r{[}k@r{|}m@r{]}} suffix, this switch is equivalent,
8033 in effect, to completing all task specs with
8034 @smallexample @c ada
8035 pragma Storage_Size (nn);
8037 When they do not already have such a pragma.
8039 @item ^-D^/DEFAULT_SECONDARY_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8040 @cindex @option{^-D^/DEFAULT_SECONDARY_STACK_SIZE=nnnnn^} (@command{gnatbind})
8041 This switch can be used to change the default secondary stack size value
8042 to a specified size @var{nn}, which is expressed in bytes by default, or
8043 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8046 The secondary stack is used to deal with functions that return a variable
8047 sized result, for example a function returning an unconstrained
8048 String. There are two ways in which this secondary stack is allocated.
8050 For most targets, the secondary stack is growing on demand and is allocated
8051 as a chain of blocks in the heap. The -D option is not very
8052 relevant. It only give some control over the size of the allocated
8053 blocks (whose size is the minimum of the default secondary stack size value,
8054 and the actual size needed for the current allocation request).
8056 For certain targets, notably VxWorks 653,
8057 the secondary stack is allocated by carving off a fixed ratio chunk of the
8058 primary task stack. The -D option is used to define the
8059 size of the environment task's secondary stack.
8061 @item ^-e^/ELABORATION_DEPENDENCIES^
8062 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@command{gnatbind})
8063 Output complete list of elaboration-order dependencies.
8065 @item ^-E^/STORE_TRACEBACKS^
8066 @cindex @option{^-E^/STORE_TRACEBACKS^} (@command{gnatbind})
8067 Store tracebacks in exception occurrences when the target supports it.
8068 This is the default with the zero cost exception mechanism.
8070 @c The following may get moved to an appendix
8071 This option is currently supported on the following targets:
8072 all x86 ports, Solaris, Windows, HP-UX, AIX, PowerPC VxWorks and Alpha VxWorks.
8074 See also the packages @code{GNAT.Traceback} and
8075 @code{GNAT.Traceback.Symbolic} for more information.
8077 Note that on x86 ports, you must not use @option{-fomit-frame-pointer}
8078 @command{gcc} option.
8081 @item ^-F^/FORCE_ELABS_FLAGS^
8082 @cindex @option{^-F^/FORCE_ELABS_FLAGS^} (@command{gnatbind})
8083 Force the checks of elaboration flags. @command{gnatbind} does not normally
8084 generate checks of elaboration flags for the main executable, except when
8085 a Stand-Alone Library is used. However, there are cases when this cannot be
8086 detected by gnatbind. An example is importing an interface of a Stand-Alone
8087 Library through a pragma Import and only specifying through a linker switch
8088 this Stand-Alone Library. This switch is used to guarantee that elaboration
8089 flag checks are generated.
8092 @cindex @option{^-h^/HELP^} (@command{gnatbind})
8093 Output usage (help) information
8096 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8097 Specify directory to be searched for source and ALI files.
8099 @item ^-I-^/NOCURRENT_DIRECTORY^
8100 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatbind})
8101 Do not look for sources in the current directory where @code{gnatbind} was
8102 invoked, and do not look for ALI files in the directory containing the
8103 ALI file named in the @code{gnatbind} command line.
8105 @item ^-l^/ORDER_OF_ELABORATION^
8106 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@command{gnatbind})
8107 Output chosen elaboration order.
8109 @item ^-L@var{xxx}^/BUILD_LIBRARY=@var{xxx}^
8110 @cindex @option{^-L^/BUILD_LIBRARY^} (@command{gnatbind})
8111 Bind the units for library building. In this case the adainit and
8112 adafinal procedures (@pxref{Binding with Non-Ada Main Programs})
8113 are renamed to ^@var{xxx}init^@var{XXX}INIT^ and
8114 ^@var{xxx}final^@var{XXX}FINAL^.
8115 Implies ^-n^/NOCOMPILE^.
8117 (@xref{GNAT and Libraries}, for more details.)
8120 On OpenVMS, these init and final procedures are exported in uppercase
8121 letters. For example if /BUILD_LIBRARY=toto is used, the exported name of
8122 the init procedure will be "TOTOINIT" and the exported name of the final
8123 procedure will be "TOTOFINAL".
8126 @item ^-Mxyz^/RENAME_MAIN=xyz^
8127 @cindex @option{^-M^/RENAME_MAIN^} (@command{gnatbind})
8128 Rename generated main program from main to xyz. This option is
8129 supported on cross environments only.
8131 @item ^-m^/ERROR_LIMIT=^@var{n}
8132 @cindex @option{^-m^/ERROR_LIMIT^} (@command{gnatbind})
8133 Limit number of detected errors or warnings to @var{n}, where @var{n} is
8134 in the range 1..999999. The default value if no switch is
8135 given is 9999. If the number of warnings reaches this limit, then a
8136 message is output and further warnings are suppressed, the bind
8137 continues in this case. If the number of errors reaches this
8138 limit, then a message is output and the bind is abandoned.
8139 A value of zero means that no limit is enforced. The equal
8143 Furthermore, under Windows, the sources pointed to by the libraries path
8144 set in the registry are not searched for.
8148 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8152 @cindex @option{-nostdinc} (@command{gnatbind})
8153 Do not look for sources in the system default directory.
8156 @cindex @option{-nostdlib} (@command{gnatbind})
8157 Do not look for library files in the system default directory.
8159 @item --RTS=@var{rts-path}
8160 @cindex @option{--RTS} (@code{gnatbind})
8161 Specifies the default location of the runtime library. Same meaning as the
8162 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
8164 @item ^-o ^/OUTPUT=^@var{file}
8165 @cindex @option{^-o ^/OUTPUT^} (@command{gnatbind})
8166 Name the output file @var{file} (default is @file{b~@var{xxx}.adb}).
8167 Note that if this option is used, then linking must be done manually,
8168 gnatlink cannot be used.
8170 @item ^-O^/OBJECT_LIST^
8171 @cindex @option{^-O^/OBJECT_LIST^} (@command{gnatbind})
8174 @item ^-p^/PESSIMISTIC_ELABORATION^
8175 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@command{gnatbind})
8176 Pessimistic (worst-case) elaboration order
8179 @cindex @option{^-R^-R^} (@command{gnatbind})
8180 Output closure source list.
8182 @item ^-s^/READ_SOURCES=ALL^
8183 @cindex @option{^-s^/READ_SOURCES=ALL^} (@command{gnatbind})
8184 Require all source files to be present.
8186 @item ^-S@var{xxx}^/INITIALIZE_SCALARS=@var{xxx}^
8187 @cindex @option{^-S^/INITIALIZE_SCALARS^} (@command{gnatbind})
8188 Specifies the value to be used when detecting uninitialized scalar
8189 objects with pragma Initialize_Scalars.
8190 The @var{xxx} ^string specified with the switch^option^ may be either
8192 @item ``@option{^in^INVALID^}'' requesting an invalid value where possible
8193 @item ``@option{^lo^LOW^}'' for the lowest possible value
8194 @item ``@option{^hi^HIGH^}'' for the highest possible value
8195 @item ``@option{@var{xx}}'' for a value consisting of repeated bytes with the
8196 value @code{16#@var{xx}#} (i.e., @var{xx} is a string of two hexadecimal digits).
8199 In addition, you can specify @option{-Sev} to indicate that the value is
8200 to be set at run time. In this case, the program will look for an environment
8201 @cindex GNAT_INIT_SCALARS
8202 variable of the form @env{GNAT_INIT_SCALARS=@var{xx}}, where @var{xx} is one
8203 of @option{in/lo/hi/@var{xx}} with the same meanings as above.
8204 If no environment variable is found, or if it does not have a valid value,
8205 then the default is @option{in} (invalid values).
8209 @cindex @option{-static} (@code{gnatbind})
8210 Link against a static GNAT run time.
8213 @cindex @option{-shared} (@code{gnatbind})
8214 Link against a shared GNAT run time when available.
8217 @item ^-t^/NOTIME_STAMP_CHECK^
8218 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8219 Tolerate time stamp and other consistency errors
8221 @item ^-T@var{n}^/TIME_SLICE=@var{n}^
8222 @cindex @option{^-T^/TIME_SLICE^} (@code{gnatbind})
8223 Set the time slice value to @var{n} milliseconds. If the system supports
8224 the specification of a specific time slice value, then the indicated value
8225 is used. If the system does not support specific time slice values, but
8226 does support some general notion of round-robin scheduling, then any
8227 nonzero value will activate round-robin scheduling.
8229 A value of zero is treated specially. It turns off time
8230 slicing, and in addition, indicates to the tasking run time that the
8231 semantics should match as closely as possible the Annex D
8232 requirements of the Ada RM, and in particular sets the default
8233 scheduling policy to @code{FIFO_Within_Priorities}.
8235 @item ^-u@var{n}^/DYNAMIC_STACK_USAGE=@var{n}^
8236 @cindex @option{^-u^/DYNAMIC_STACK_USAGE^} (@code{gnatbind})
8237 Enable dynamic stack usage, with @var{n} results stored and displayed
8238 at program termination. A result is generated when a task
8239 terminates. Results that can't be stored are displayed on the fly, at
8240 task termination. This option is currently not supported on Itanium
8241 platforms. (See @ref{Dynamic Stack Usage Analysis} for details.)
8243 @item ^-v^/REPORT_ERRORS=VERBOSE^
8244 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8245 Verbose mode. Write error messages, header, summary output to
8250 @cindex @option{-w} (@code{gnatbind})
8251 Warning mode (@var{x}=s/e for suppress/treat as error)
8255 @item /WARNINGS=NORMAL
8256 @cindex @option{/WARNINGS} (@code{gnatbind})
8257 Normal warnings mode. Warnings are issued but ignored
8259 @item /WARNINGS=SUPPRESS
8260 @cindex @option{/WARNINGS} (@code{gnatbind})
8261 All warning messages are suppressed
8263 @item /WARNINGS=ERROR
8264 @cindex @option{/WARNINGS} (@code{gnatbind})
8265 Warning messages are treated as fatal errors
8268 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8269 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8270 Override default wide character encoding for standard Text_IO files.
8272 @item ^-x^/READ_SOURCES=NONE^
8273 @cindex @option{^-x^/READ_SOURCES^} (@code{gnatbind})
8274 Exclude source files (check object consistency only).
8277 @item /READ_SOURCES=AVAILABLE
8278 @cindex @option{/READ_SOURCES} (@code{gnatbind})
8279 Default mode, in which sources are checked for consistency only if
8283 @item ^-y^/ENABLE_LEAP_SECONDS^
8284 @cindex @option{^-y^/ENABLE_LEAP_SECONDS^} (@code{gnatbind})
8285 Enable leap seconds support in @code{Ada.Calendar} and its children.
8287 @item ^-z^/ZERO_MAIN^
8288 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8294 You may obtain this listing of switches by running @code{gnatbind} with
8298 @node Consistency-Checking Modes
8299 @subsection Consistency-Checking Modes
8302 As described earlier, by default @code{gnatbind} checks
8303 that object files are consistent with one another and are consistent
8304 with any source files it can locate. The following switches control binder
8309 @item ^-s^/READ_SOURCES=ALL^
8310 @cindex @option{^-s^/READ_SOURCES=ALL^} (@code{gnatbind})
8311 Require source files to be present. In this mode, the binder must be
8312 able to locate all source files that are referenced, in order to check
8313 their consistency. In normal mode, if a source file cannot be located it
8314 is simply ignored. If you specify this switch, a missing source
8317 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8318 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8319 Override default wide character encoding for standard Text_IO files.
8320 Normally the default wide character encoding method used for standard
8321 [Wide_[Wide_]]Text_IO files is taken from the encoding specified for
8322 the main source input (see description of switch
8323 @option{^-gnatWx^/WIDE_CHARACTER_ENCODING^} for the compiler). The
8324 use of this switch for the binder (which has the same set of
8325 possible arguments) overrides this default as specified.
8327 @item ^-x^/READ_SOURCES=NONE^
8328 @cindex @option{^-x^/READ_SOURCES=NONE^} (@code{gnatbind})
8329 Exclude source files. In this mode, the binder only checks that ALI
8330 files are consistent with one another. Source files are not accessed.
8331 The binder runs faster in this mode, and there is still a guarantee that
8332 the resulting program is self-consistent.
8333 If a source file has been edited since it was last compiled, and you
8334 specify this switch, the binder will not detect that the object
8335 file is out of date with respect to the source file. Note that this is the
8336 mode that is automatically used by @command{gnatmake} because in this
8337 case the checking against sources has already been performed by
8338 @command{gnatmake} in the course of compilation (i.e.@: before binding).
8341 @item /READ_SOURCES=AVAILABLE
8342 @cindex @code{/READ_SOURCES=AVAILABLE} (@code{gnatbind})
8343 This is the default mode in which source files are checked if they are
8344 available, and ignored if they are not available.
8348 @node Binder Error Message Control
8349 @subsection Binder Error Message Control
8352 The following switches provide control over the generation of error
8353 messages from the binder:
8357 @item ^-v^/REPORT_ERRORS=VERBOSE^
8358 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8359 Verbose mode. In the normal mode, brief error messages are generated to
8360 @file{stderr}. If this switch is present, a header is written
8361 to @file{stdout} and any error messages are directed to @file{stdout}.
8362 All that is written to @file{stderr} is a brief summary message.
8364 @item ^-b^/REPORT_ERRORS=BRIEF^
8365 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@code{gnatbind})
8366 Generate brief error messages to @file{stderr} even if verbose mode is
8367 specified. This is relevant only when used with the
8368 @option{^-v^/REPORT_ERRORS=VERBOSE^} switch.
8372 @cindex @option{-m} (@code{gnatbind})
8373 Limits the number of error messages to @var{n}, a decimal integer in the
8374 range 1-999. The binder terminates immediately if this limit is reached.
8377 @cindex @option{-M} (@code{gnatbind})
8378 Renames the generated main program from @code{main} to @code{xxx}.
8379 This is useful in the case of some cross-building environments, where
8380 the actual main program is separate from the one generated
8384 @item ^-ws^/WARNINGS=SUPPRESS^
8385 @cindex @option{^-ws^/WARNINGS=SUPPRESS^} (@code{gnatbind})
8387 Suppress all warning messages.
8389 @item ^-we^/WARNINGS=ERROR^
8390 @cindex @option{^-we^/WARNINGS=ERROR^} (@code{gnatbind})
8391 Treat any warning messages as fatal errors.
8394 @item /WARNINGS=NORMAL
8395 Standard mode with warnings generated, but warnings do not get treated
8399 @item ^-t^/NOTIME_STAMP_CHECK^
8400 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8401 @cindex Time stamp checks, in binder
8402 @cindex Binder consistency checks
8403 @cindex Consistency checks, in binder
8404 The binder performs a number of consistency checks including:
8408 Check that time stamps of a given source unit are consistent
8410 Check that checksums of a given source unit are consistent
8412 Check that consistent versions of @code{GNAT} were used for compilation
8414 Check consistency of configuration pragmas as required
8418 Normally failure of such checks, in accordance with the consistency
8419 requirements of the Ada Reference Manual, causes error messages to be
8420 generated which abort the binder and prevent the output of a binder
8421 file and subsequent link to obtain an executable.
8423 The @option{^-t^/NOTIME_STAMP_CHECK^} switch converts these error messages
8424 into warnings, so that
8425 binding and linking can continue to completion even in the presence of such
8426 errors. The result may be a failed link (due to missing symbols), or a
8427 non-functional executable which has undefined semantics.
8428 @emph{This means that
8429 @option{^-t^/NOTIME_STAMP_CHECK^} should be used only in unusual situations,
8433 @node Elaboration Control
8434 @subsection Elaboration Control
8437 The following switches provide additional control over the elaboration
8438 order. For full details see @ref{Elaboration Order Handling in GNAT}.
8441 @item ^-p^/PESSIMISTIC_ELABORATION^
8442 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@code{gnatbind})
8443 Normally the binder attempts to choose an elaboration order that is
8444 likely to minimize the likelihood of an elaboration order error resulting
8445 in raising a @code{Program_Error} exception. This switch reverses the
8446 action of the binder, and requests that it deliberately choose an order
8447 that is likely to maximize the likelihood of an elaboration error.
8448 This is useful in ensuring portability and avoiding dependence on
8449 accidental fortuitous elaboration ordering.
8451 Normally it only makes sense to use the @option{^-p^/PESSIMISTIC_ELABORATION^}
8453 elaboration checking is used (@option{-gnatE} switch used for compilation).
8454 This is because in the default static elaboration mode, all necessary
8455 @code{Elaborate} and @code{Elaborate_All} pragmas are implicitly inserted.
8456 These implicit pragmas are still respected by the binder in
8457 @option{^-p^/PESSIMISTIC_ELABORATION^} mode, so a
8458 safe elaboration order is assured.
8461 @node Output Control
8462 @subsection Output Control
8465 The following switches allow additional control over the output
8466 generated by the binder.
8471 @item ^-A^/BIND_FILE=ADA^
8472 @cindex @option{^-A^/BIND_FILE=ADA^} (@code{gnatbind})
8473 Generate binder program in Ada (default). The binder program is named
8474 @file{b~@var{mainprog}.adb} by default. This can be changed with
8475 @option{^-o^/OUTPUT^} @code{gnatbind} option.
8477 @item ^-c^/NOOUTPUT^
8478 @cindex @option{^-c^/NOOUTPUT^} (@code{gnatbind})
8479 Check only. Do not generate the binder output file. In this mode the
8480 binder performs all error checks but does not generate an output file.
8482 @item ^-C^/BIND_FILE=C^
8483 @cindex @option{^-C^/BIND_FILE=C^} (@code{gnatbind})
8484 Generate binder program in C. The binder program is named
8485 @file{b_@var{mainprog}.c}.
8486 This can be changed with @option{^-o^/OUTPUT^} @code{gnatbind}
8489 @item ^-e^/ELABORATION_DEPENDENCIES^
8490 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@code{gnatbind})
8491 Output complete list of elaboration-order dependencies, showing the
8492 reason for each dependency. This output can be rather extensive but may
8493 be useful in diagnosing problems with elaboration order. The output is
8494 written to @file{stdout}.
8497 @cindex @option{^-h^/HELP^} (@code{gnatbind})
8498 Output usage information. The output is written to @file{stdout}.
8500 @item ^-K^/LINKER_OPTION_LIST^
8501 @cindex @option{^-K^/LINKER_OPTION_LIST^} (@code{gnatbind})
8502 Output linker options to @file{stdout}. Includes library search paths,
8503 contents of pragmas Ident and Linker_Options, and libraries added
8506 @item ^-l^/ORDER_OF_ELABORATION^
8507 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@code{gnatbind})
8508 Output chosen elaboration order. The output is written to @file{stdout}.
8510 @item ^-O^/OBJECT_LIST^
8511 @cindex @option{^-O^/OBJECT_LIST^} (@code{gnatbind})
8512 Output full names of all the object files that must be linked to provide
8513 the Ada component of the program. The output is written to @file{stdout}.
8514 This list includes the files explicitly supplied and referenced by the user
8515 as well as implicitly referenced run-time unit files. The latter are
8516 omitted if the corresponding units reside in shared libraries. The
8517 directory names for the run-time units depend on the system configuration.
8519 @item ^-o ^/OUTPUT=^@var{file}
8520 @cindex @option{^-o^/OUTPUT^} (@code{gnatbind})
8521 Set name of output file to @var{file} instead of the normal
8522 @file{b~@var{mainprog}.adb} default. Note that @var{file} denote the Ada
8523 binder generated body filename. In C mode you would normally give
8524 @var{file} an extension of @file{.c} because it will be a C source program.
8525 Note that if this option is used, then linking must be done manually.
8526 It is not possible to use gnatlink in this case, since it cannot locate
8529 @item ^-r^/RESTRICTION_LIST^
8530 @cindex @option{^-r^/RESTRICTION_LIST^} (@code{gnatbind})
8531 Generate list of @code{pragma Restrictions} that could be applied to
8532 the current unit. This is useful for code audit purposes, and also may
8533 be used to improve code generation in some cases.
8537 @node Binding with Non-Ada Main Programs
8538 @subsection Binding with Non-Ada Main Programs
8541 In our description so far we have assumed that the main
8542 program is in Ada, and that the task of the binder is to generate a
8543 corresponding function @code{main} that invokes this Ada main
8544 program. GNAT also supports the building of executable programs where
8545 the main program is not in Ada, but some of the called routines are
8546 written in Ada and compiled using GNAT (@pxref{Mixed Language Programming}).
8547 The following switch is used in this situation:
8551 @cindex @option{^-n^/NOMAIN^} (@code{gnatbind})
8552 No main program. The main program is not in Ada.
8556 In this case, most of the functions of the binder are still required,
8557 but instead of generating a main program, the binder generates a file
8558 containing the following callable routines:
8563 You must call this routine to initialize the Ada part of the program by
8564 calling the necessary elaboration routines. A call to @code{adainit} is
8565 required before the first call to an Ada subprogram.
8567 Note that it is assumed that the basic execution environment must be setup
8568 to be appropriate for Ada execution at the point where the first Ada
8569 subprogram is called. In particular, if the Ada code will do any
8570 floating-point operations, then the FPU must be setup in an appropriate
8571 manner. For the case of the x86, for example, full precision mode is
8572 required. The procedure GNAT.Float_Control.Reset may be used to ensure
8573 that the FPU is in the right state.
8577 You must call this routine to perform any library-level finalization
8578 required by the Ada subprograms. A call to @code{adafinal} is required
8579 after the last call to an Ada subprogram, and before the program
8584 If the @option{^-n^/NOMAIN^} switch
8585 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8586 @cindex Binder, multiple input files
8587 is given, more than one ALI file may appear on
8588 the command line for @code{gnatbind}. The normal @dfn{closure}
8589 calculation is performed for each of the specified units. Calculating
8590 the closure means finding out the set of units involved by tracing
8591 @code{with} references. The reason it is necessary to be able to
8592 specify more than one ALI file is that a given program may invoke two or
8593 more quite separate groups of Ada units.
8595 The binder takes the name of its output file from the last specified ALI
8596 file, unless overridden by the use of the @option{^-o file^/OUTPUT=file^}.
8597 @cindex @option{^-o^/OUTPUT^} (@command{gnatbind})
8598 The output is an Ada unit in source form that can
8599 be compiled with GNAT unless the -C switch is used in which case the
8600 output is a C source file, which must be compiled using the C compiler.
8601 This compilation occurs automatically as part of the @command{gnatlink}
8604 Currently the GNAT run time requires a FPU using 80 bits mode
8605 precision. Under targets where this is not the default it is required to
8606 call GNAT.Float_Control.Reset before using floating point numbers (this
8607 include float computation, float input and output) in the Ada code. A
8608 side effect is that this could be the wrong mode for the foreign code
8609 where floating point computation could be broken after this call.
8611 @node Binding Programs with No Main Subprogram
8612 @subsection Binding Programs with No Main Subprogram
8615 It is possible to have an Ada program which does not have a main
8616 subprogram. This program will call the elaboration routines of all the
8617 packages, then the finalization routines.
8619 The following switch is used to bind programs organized in this manner:
8622 @item ^-z^/ZERO_MAIN^
8623 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8624 Normally the binder checks that the unit name given on the command line
8625 corresponds to a suitable main subprogram. When this switch is used,
8626 a list of ALI files can be given, and the execution of the program
8627 consists of elaboration of these units in an appropriate order. Note
8628 that the default wide character encoding method for standard Text_IO
8629 files is always set to Brackets if this switch is set (you can use
8631 @option{^-Wx^WIDE_CHARACTER_ENCODING^} to override this default).
8634 @node Command-Line Access
8635 @section Command-Line Access
8638 The package @code{Ada.Command_Line} provides access to the command-line
8639 arguments and program name. In order for this interface to operate
8640 correctly, the two variables
8652 are declared in one of the GNAT library routines. These variables must
8653 be set from the actual @code{argc} and @code{argv} values passed to the
8654 main program. With no @option{^n^/NOMAIN^} present, @code{gnatbind}
8655 generates the C main program to automatically set these variables.
8656 If the @option{^n^/NOMAIN^} switch is used, there is no automatic way to
8657 set these variables. If they are not set, the procedures in
8658 @code{Ada.Command_Line} will not be available, and any attempt to use
8659 them will raise @code{Constraint_Error}. If command line access is
8660 required, your main program must set @code{gnat_argc} and
8661 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
8664 @node Search Paths for gnatbind
8665 @section Search Paths for @code{gnatbind}
8668 The binder takes the name of an ALI file as its argument and needs to
8669 locate source files as well as other ALI files to verify object consistency.
8671 For source files, it follows exactly the same search rules as @command{gcc}
8672 (@pxref{Search Paths and the Run-Time Library (RTL)}). For ALI files the
8673 directories searched are:
8677 The directory containing the ALI file named in the command line, unless
8678 the switch @option{^-I-^/NOCURRENT_DIRECTORY^} is specified.
8681 All directories specified by @option{^-I^/SEARCH^}
8682 switches on the @code{gnatbind}
8683 command line, in the order given.
8686 @findex ADA_PRJ_OBJECTS_FILE
8687 Each of the directories listed in the text file whose name is given
8688 by the @env{ADA_PRJ_OBJECTS_FILE} ^environment variable^logical name^.
8691 @env{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
8692 driver when project files are used. It should not normally be set
8696 @findex ADA_OBJECTS_PATH
8697 Each of the directories listed in the value of the
8698 @env{ADA_OBJECTS_PATH} ^environment variable^logical name^.
8700 Construct this value
8701 exactly as the @env{PATH} environment variable: a list of directory
8702 names separated by colons (semicolons when working with the NT version
8706 Normally, define this value as a logical name containing a comma separated
8707 list of directory names.
8709 This variable can also be defined by means of an environment string
8710 (an argument to the HP C exec* set of functions).
8714 DEFINE ANOTHER_PATH FOO:[BAG]
8715 DEFINE ADA_OBJECTS_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
8718 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
8719 first, followed by the standard Ada
8720 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB].
8721 If this is not redefined, the user will obtain the HP Ada 83 IO packages
8722 (Text_IO, Sequential_IO, etc)
8723 instead of the standard Ada packages. Thus, in order to get the standard Ada
8724 packages by default, ADA_OBJECTS_PATH must be redefined.
8728 The content of the @file{ada_object_path} file which is part of the GNAT
8729 installation tree and is used to store standard libraries such as the
8730 GNAT Run Time Library (RTL) unless the switch @option{-nostdlib} is
8733 @ref{Installing a library}
8738 In the binder the switch @option{^-I^/SEARCH^}
8739 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8740 is used to specify both source and
8741 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
8742 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8743 instead if you want to specify
8744 source paths only, and @option{^-aO^/LIBRARY_SEARCH^}
8745 @cindex @option{^-aO^/LIBRARY_SEARCH^} (@command{gnatbind})
8746 if you want to specify library paths
8747 only. This means that for the binder
8748 @option{^-I^/SEARCH=^}@var{dir} is equivalent to
8749 @option{^-aI^/SOURCE_SEARCH=^}@var{dir}
8750 @option{^-aO^/OBJECT_SEARCH=^}@var{dir}.
8751 The binder generates the bind file (a C language source file) in the
8752 current working directory.
8758 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
8759 children make up the GNAT Run-Time Library, together with the package
8760 GNAT and its children, which contain a set of useful additional
8761 library functions provided by GNAT. The sources for these units are
8762 needed by the compiler and are kept together in one directory. The ALI
8763 files and object files generated by compiling the RTL are needed by the
8764 binder and the linker and are kept together in one directory, typically
8765 different from the directory containing the sources. In a normal
8766 installation, you need not specify these directory names when compiling
8767 or binding. Either the environment variables or the built-in defaults
8768 cause these files to be found.
8770 Besides simplifying access to the RTL, a major use of search paths is
8771 in compiling sources from multiple directories. This can make
8772 development environments much more flexible.
8774 @node Examples of gnatbind Usage
8775 @section Examples of @code{gnatbind} Usage
8778 This section contains a number of examples of using the GNAT binding
8779 utility @code{gnatbind}.
8782 @item gnatbind hello
8783 The main program @code{Hello} (source program in @file{hello.adb}) is
8784 bound using the standard switch settings. The generated main program is
8785 @file{b~hello.adb}. This is the normal, default use of the binder.
8788 @item gnatbind hello -o mainprog.adb
8791 @item gnatbind HELLO.ALI /OUTPUT=Mainprog.ADB
8793 The main program @code{Hello} (source program in @file{hello.adb}) is
8794 bound using the standard switch settings. The generated main program is
8795 @file{mainprog.adb} with the associated spec in
8796 @file{mainprog.ads}. Note that you must specify the body here not the
8797 spec, in the case where the output is in Ada. Note that if this option
8798 is used, then linking must be done manually, since gnatlink will not
8799 be able to find the generated file.
8802 @item gnatbind main -C -o mainprog.c -x
8805 @item gnatbind MAIN.ALI /BIND_FILE=C /OUTPUT=Mainprog.C /READ_SOURCES=NONE
8807 The main program @code{Main} (source program in
8808 @file{main.adb}) is bound, excluding source files from the
8809 consistency checking, generating
8810 the file @file{mainprog.c}.
8813 @item gnatbind -x main_program -C -o mainprog.c
8814 This command is exactly the same as the previous example. Switches may
8815 appear anywhere in the command line, and single letter switches may be
8816 combined into a single switch.
8820 @item gnatbind -n math dbase -C -o ada-control.c
8823 @item gnatbind /NOMAIN math dbase /BIND_FILE=C /OUTPUT=ada-control.c
8825 The main program is in a language other than Ada, but calls to
8826 subprograms in packages @code{Math} and @code{Dbase} appear. This call
8827 to @code{gnatbind} generates the file @file{ada-control.c} containing
8828 the @code{adainit} and @code{adafinal} routines to be called before and
8829 after accessing the Ada units.
8832 @c ------------------------------------
8833 @node Linking Using gnatlink
8834 @chapter Linking Using @command{gnatlink}
8835 @c ------------------------------------
8839 This chapter discusses @command{gnatlink}, a tool that links
8840 an Ada program and builds an executable file. This utility
8841 invokes the system linker ^(via the @command{gcc} command)^^
8842 with a correct list of object files and library references.
8843 @command{gnatlink} automatically determines the list of files and
8844 references for the Ada part of a program. It uses the binder file
8845 generated by the @command{gnatbind} to determine this list.
8847 Note: to invoke @code{gnatlink} with a project file, use the @code{gnat}
8848 driver (see @ref{The GNAT Driver and Project Files}).
8851 * Running gnatlink::
8852 * Switches for gnatlink::
8855 @node Running gnatlink
8856 @section Running @command{gnatlink}
8859 The form of the @command{gnatlink} command is
8862 $ gnatlink @ovar{switches} @var{mainprog}@r{[}.ali@r{]}
8863 @ovar{non-Ada objects} @ovar{linker options}
8867 The arguments of @command{gnatlink} (switches, main @file{ALI} file,
8869 or linker options) may be in any order, provided that no non-Ada object may
8870 be mistaken for a main @file{ALI} file.
8871 Any file name @file{F} without the @file{.ali}
8872 extension will be taken as the main @file{ALI} file if a file exists
8873 whose name is the concatenation of @file{F} and @file{.ali}.
8876 @file{@var{mainprog}.ali} references the ALI file of the main program.
8877 The @file{.ali} extension of this file can be omitted. From this
8878 reference, @command{gnatlink} locates the corresponding binder file
8879 @file{b~@var{mainprog}.adb} and, using the information in this file along
8880 with the list of non-Ada objects and linker options, constructs a
8881 linker command file to create the executable.
8883 The arguments other than the @command{gnatlink} switches and the main
8884 @file{ALI} file are passed to the linker uninterpreted.
8885 They typically include the names of
8886 object files for units written in other languages than Ada and any library
8887 references required to resolve references in any of these foreign language
8888 units, or in @code{Import} pragmas in any Ada units.
8890 @var{linker options} is an optional list of linker specific
8892 The default linker called by gnatlink is @command{gcc} which in
8893 turn calls the appropriate system linker.
8894 Standard options for the linker such as @option{-lmy_lib} or
8895 @option{-Ldir} can be added as is.
8896 For options that are not recognized by
8897 @command{gcc} as linker options, use the @command{gcc} switches
8898 @option{-Xlinker} or @option{-Wl,}.
8899 Refer to the GCC documentation for
8900 details. Here is an example showing how to generate a linker map:
8903 $ ^gnatlink my_prog -Wl,-Map,MAPFILE^GNAT LINK my_prog.ali /MAP^
8906 Using @var{linker options} it is possible to set the program stack and
8909 See @ref{Setting Stack Size from gnatlink} and
8910 @ref{Setting Heap Size from gnatlink}.
8913 @command{gnatlink} determines the list of objects required by the Ada
8914 program and prepends them to the list of objects passed to the linker.
8915 @command{gnatlink} also gathers any arguments set by the use of
8916 @code{pragma Linker_Options} and adds them to the list of arguments
8917 presented to the linker.
8920 @command{gnatlink} accepts the following types of extra files on the command
8921 line: objects (@file{.OBJ}), libraries (@file{.OLB}), sharable images
8922 (@file{.EXE}), and options files (@file{.OPT}). These are recognized and
8923 handled according to their extension.
8926 @node Switches for gnatlink
8927 @section Switches for @command{gnatlink}
8930 The following switches are available with the @command{gnatlink} utility:
8936 @cindex @option{--version} @command{gnatlink}
8937 Display Copyright and version, then exit disregarding all other options.
8940 @cindex @option{--help} @command{gnatlink}
8941 If @option{--version} was not used, display usage, then exit disregarding
8944 @item ^-A^/BIND_FILE=ADA^
8945 @cindex @option{^-A^/BIND_FILE=ADA^} (@command{gnatlink})
8946 The binder has generated code in Ada. This is the default.
8948 @item ^-C^/BIND_FILE=C^
8949 @cindex @option{^-C^/BIND_FILE=C^} (@command{gnatlink})
8950 If instead of generating a file in Ada, the binder has generated one in
8951 C, then the linker needs to know about it. Use this switch to signal
8952 to @command{gnatlink} that the binder has generated C code rather than
8955 @item ^-f^/FORCE_OBJECT_FILE_LIST^
8956 @cindex Command line length
8957 @cindex @option{^-f^/FORCE_OBJECT_FILE_LIST^} (@command{gnatlink})
8958 On some targets, the command line length is limited, and @command{gnatlink}
8959 will generate a separate file for the linker if the list of object files
8961 The @option{^-f^/FORCE_OBJECT_FILE_LIST^} switch forces this file
8962 to be generated even if
8963 the limit is not exceeded. This is useful in some cases to deal with
8964 special situations where the command line length is exceeded.
8967 @cindex Debugging information, including
8968 @cindex @option{^-g^/DEBUG^} (@command{gnatlink})
8969 The option to include debugging information causes the Ada bind file (in
8970 other words, @file{b~@var{mainprog}.adb}) to be compiled with
8971 @option{^-g^/DEBUG^}.
8972 In addition, the binder does not delete the @file{b~@var{mainprog}.adb},
8973 @file{b~@var{mainprog}.o} and @file{b~@var{mainprog}.ali} files.
8974 Without @option{^-g^/DEBUG^}, the binder removes these files by
8975 default. The same procedure apply if a C bind file was generated using
8976 @option{^-C^/BIND_FILE=C^} @code{gnatbind} option, in this case the filenames
8977 are @file{b_@var{mainprog}.c} and @file{b_@var{mainprog}.o}.
8979 @item ^-n^/NOCOMPILE^
8980 @cindex @option{^-n^/NOCOMPILE^} (@command{gnatlink})
8981 Do not compile the file generated by the binder. This may be used when
8982 a link is rerun with different options, but there is no need to recompile
8986 @cindex @option{^-v^/VERBOSE^} (@command{gnatlink})
8987 Causes additional information to be output, including a full list of the
8988 included object files. This switch option is most useful when you want
8989 to see what set of object files are being used in the link step.
8991 @item ^-v -v^/VERBOSE/VERBOSE^
8992 @cindex @option{^-v -v^/VERBOSE/VERBOSE^} (@command{gnatlink})
8993 Very verbose mode. Requests that the compiler operate in verbose mode when
8994 it compiles the binder file, and that the system linker run in verbose mode.
8996 @item ^-o ^/EXECUTABLE=^@var{exec-name}
8997 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatlink})
8998 @var{exec-name} specifies an alternate name for the generated
8999 executable program. If this switch is omitted, the executable has the same
9000 name as the main unit. For example, @code{gnatlink try.ali} creates
9001 an executable called @file{^try^TRY.EXE^}.
9004 @item -b @var{target}
9005 @cindex @option{-b} (@command{gnatlink})
9006 Compile your program to run on @var{target}, which is the name of a
9007 system configuration. You must have a GNAT cross-compiler built if
9008 @var{target} is not the same as your host system.
9011 @cindex @option{-B} (@command{gnatlink})
9012 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
9013 from @var{dir} instead of the default location. Only use this switch
9014 when multiple versions of the GNAT compiler are available.
9015 @xref{Directory Options,,, gcc, The GNU Compiler Collection},
9016 for further details. You would normally use the @option{-b} or
9017 @option{-V} switch instead.
9019 @item --GCC=@var{compiler_name}
9020 @cindex @option{--GCC=compiler_name} (@command{gnatlink})
9021 Program used for compiling the binder file. The default is
9022 @command{gcc}. You need to use quotes around @var{compiler_name} if
9023 @code{compiler_name} contains spaces or other separator characters.
9024 As an example @option{--GCC="foo -x -y"} will instruct @command{gnatlink} to
9025 use @code{foo -x -y} as your compiler. Note that switch @option{-c} is always
9026 inserted after your command name. Thus in the above example the compiler
9027 command that will be used by @command{gnatlink} will be @code{foo -c -x -y}.
9028 A limitation of this syntax is that the name and path name of the executable
9029 itself must not include any embedded spaces. If the compiler executable is
9030 different from the default one (gcc or <prefix>-gcc), then the back-end
9031 switches in the ALI file are not used to compile the binder generated source.
9032 For example, this is the case with @option{--GCC="foo -x -y"}. But the back end
9033 switches will be used for @option{--GCC="gcc -gnatv"}. If several
9034 @option{--GCC=compiler_name} are used, only the last @var{compiler_name}
9035 is taken into account. However, all the additional switches are also taken
9037 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9038 @option{--GCC="bar -x -y -z -t"}.
9040 @item --LINK=@var{name}
9041 @cindex @option{--LINK=} (@command{gnatlink})
9042 @var{name} is the name of the linker to be invoked. This is especially
9043 useful in mixed language programs since languages such as C++ require
9044 their own linker to be used. When this switch is omitted, the default
9045 name for the linker is @command{gcc}. When this switch is used, the
9046 specified linker is called instead of @command{gcc} with exactly the same
9047 parameters that would have been passed to @command{gcc} so if the desired
9048 linker requires different parameters it is necessary to use a wrapper
9049 script that massages the parameters before invoking the real linker. It
9050 may be useful to control the exact invocation by using the verbose
9056 @item /DEBUG=TRACEBACK
9057 @cindex @code{/DEBUG=TRACEBACK} (@command{gnatlink})
9058 This qualifier causes sufficient information to be included in the
9059 executable file to allow a traceback, but does not include the full
9060 symbol information needed by the debugger.
9062 @item /IDENTIFICATION="<string>"
9063 @code{"<string>"} specifies the string to be stored in the image file
9064 identification field in the image header.
9065 It overrides any pragma @code{Ident} specified string.
9067 @item /NOINHIBIT-EXEC
9068 Generate the executable file even if there are linker warnings.
9070 @item /NOSTART_FILES
9071 Don't link in the object file containing the ``main'' transfer address.
9072 Used when linking with a foreign language main program compiled with an
9076 Prefer linking with object libraries over sharable images, even without
9082 @node The GNAT Make Program gnatmake
9083 @chapter The GNAT Make Program @command{gnatmake}
9087 * Running gnatmake::
9088 * Switches for gnatmake::
9089 * Mode Switches for gnatmake::
9090 * Notes on the Command Line::
9091 * How gnatmake Works::
9092 * Examples of gnatmake Usage::
9095 A typical development cycle when working on an Ada program consists of
9096 the following steps:
9100 Edit some sources to fix bugs.
9106 Compile all sources affected.
9116 The third step can be tricky, because not only do the modified files
9117 @cindex Dependency rules
9118 have to be compiled, but any files depending on these files must also be
9119 recompiled. The dependency rules in Ada can be quite complex, especially
9120 in the presence of overloading, @code{use} clauses, generics and inlined
9123 @command{gnatmake} automatically takes care of the third and fourth steps
9124 of this process. It determines which sources need to be compiled,
9125 compiles them, and binds and links the resulting object files.
9127 Unlike some other Ada make programs, the dependencies are always
9128 accurately recomputed from the new sources. The source based approach of
9129 the GNAT compilation model makes this possible. This means that if
9130 changes to the source program cause corresponding changes in
9131 dependencies, they will always be tracked exactly correctly by
9134 @node Running gnatmake
9135 @section Running @command{gnatmake}
9138 The usual form of the @command{gnatmake} command is
9141 $ gnatmake @ovar{switches} @var{file_name}
9142 @ovar{file_names} @ovar{mode_switches}
9146 The only required argument is one @var{file_name}, which specifies
9147 a compilation unit that is a main program. Several @var{file_names} can be
9148 specified: this will result in several executables being built.
9149 If @code{switches} are present, they can be placed before the first
9150 @var{file_name}, between @var{file_names} or after the last @var{file_name}.
9151 If @var{mode_switches} are present, they must always be placed after
9152 the last @var{file_name} and all @code{switches}.
9154 If you are using standard file extensions (@file{.adb} and @file{.ads}), then the
9155 extension may be omitted from the @var{file_name} arguments. However, if
9156 you are using non-standard extensions, then it is required that the
9157 extension be given. A relative or absolute directory path can be
9158 specified in a @var{file_name}, in which case, the input source file will
9159 be searched for in the specified directory only. Otherwise, the input
9160 source file will first be searched in the directory where
9161 @command{gnatmake} was invoked and if it is not found, it will be search on
9162 the source path of the compiler as described in
9163 @ref{Search Paths and the Run-Time Library (RTL)}.
9165 All @command{gnatmake} output (except when you specify
9166 @option{^-M^/DEPENDENCIES_LIST^}) is to
9167 @file{stderr}. The output produced by the
9168 @option{^-M^/DEPENDENCIES_LIST^} switch is send to
9171 @node Switches for gnatmake
9172 @section Switches for @command{gnatmake}
9175 You may specify any of the following switches to @command{gnatmake}:
9181 @cindex @option{--version} @command{gnatmake}
9182 Display Copyright and version, then exit disregarding all other options.
9185 @cindex @option{--help} @command{gnatmake}
9186 If @option{--version} was not used, display usage, then exit disregarding
9190 @item --GCC=@var{compiler_name}
9191 @cindex @option{--GCC=compiler_name} (@command{gnatmake})
9192 Program used for compiling. The default is `@command{gcc}'. You need to use
9193 quotes around @var{compiler_name} if @code{compiler_name} contains
9194 spaces or other separator characters. As an example @option{--GCC="foo -x
9195 -y"} will instruct @command{gnatmake} to use @code{foo -x -y} as your
9196 compiler. A limitation of this syntax is that the name and path name of
9197 the executable itself must not include any embedded spaces. Note that
9198 switch @option{-c} is always inserted after your command name. Thus in the
9199 above example the compiler command that will be used by @command{gnatmake}
9200 will be @code{foo -c -x -y}. If several @option{--GCC=compiler_name} are
9201 used, only the last @var{compiler_name} is taken into account. However,
9202 all the additional switches are also taken into account. Thus,
9203 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9204 @option{--GCC="bar -x -y -z -t"}.
9206 @item --GNATBIND=@var{binder_name}
9207 @cindex @option{--GNATBIND=binder_name} (@command{gnatmake})
9208 Program used for binding. The default is `@code{gnatbind}'. You need to
9209 use quotes around @var{binder_name} if @var{binder_name} contains spaces
9210 or other separator characters. As an example @option{--GNATBIND="bar -x
9211 -y"} will instruct @command{gnatmake} to use @code{bar -x -y} as your
9212 binder. Binder switches that are normally appended by @command{gnatmake}
9213 to `@code{gnatbind}' are now appended to the end of @code{bar -x -y}.
9214 A limitation of this syntax is that the name and path name of the executable
9215 itself must not include any embedded spaces.
9217 @item --GNATLINK=@var{linker_name}
9218 @cindex @option{--GNATLINK=linker_name} (@command{gnatmake})
9219 Program used for linking. The default is `@command{gnatlink}'. You need to
9220 use quotes around @var{linker_name} if @var{linker_name} contains spaces
9221 or other separator characters. As an example @option{--GNATLINK="lan -x
9222 -y"} will instruct @command{gnatmake} to use @code{lan -x -y} as your
9223 linker. Linker switches that are normally appended by @command{gnatmake} to
9224 `@command{gnatlink}' are now appended to the end of @code{lan -x -y}.
9225 A limitation of this syntax is that the name and path name of the executable
9226 itself must not include any embedded spaces.
9230 @item ^-a^/ALL_FILES^
9231 @cindex @option{^-a^/ALL_FILES^} (@command{gnatmake})
9232 Consider all files in the make process, even the GNAT internal system
9233 files (for example, the predefined Ada library files), as well as any
9234 locked files. Locked files are files whose ALI file is write-protected.
9236 @command{gnatmake} does not check these files,
9237 because the assumption is that the GNAT internal files are properly up
9238 to date, and also that any write protected ALI files have been properly
9239 installed. Note that if there is an installation problem, such that one
9240 of these files is not up to date, it will be properly caught by the
9242 You may have to specify this switch if you are working on GNAT
9243 itself. The switch @option{^-a^/ALL_FILES^} is also useful
9244 in conjunction with @option{^-f^/FORCE_COMPILE^}
9245 if you need to recompile an entire application,
9246 including run-time files, using special configuration pragmas,
9247 such as a @code{Normalize_Scalars} pragma.
9250 @code{gnatmake ^-a^/ALL_FILES^} compiles all GNAT
9253 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
9256 the @code{/CHECKS=SUPPRESS_ALL /STYLE_CHECKS=GNAT} switch.
9259 @item ^-b^/ACTIONS=BIND^
9260 @cindex @option{^-b^/ACTIONS=BIND^} (@command{gnatmake})
9261 Bind only. Can be combined with @option{^-c^/ACTIONS=COMPILE^} to do
9262 compilation and binding, but no link.
9263 Can be combined with @option{^-l^/ACTIONS=LINK^}
9264 to do binding and linking. When not combined with
9265 @option{^-c^/ACTIONS=COMPILE^}
9266 all the units in the closure of the main program must have been previously
9267 compiled and must be up to date. The root unit specified by @var{file_name}
9268 may be given without extension, with the source extension or, if no GNAT
9269 Project File is specified, with the ALI file extension.
9271 @item ^-c^/ACTIONS=COMPILE^
9272 @cindex @option{^-c^/ACTIONS=COMPILE^} (@command{gnatmake})
9273 Compile only. Do not perform binding, except when @option{^-b^/ACTIONS=BIND^}
9274 is also specified. Do not perform linking, except if both
9275 @option{^-b^/ACTIONS=BIND^} and
9276 @option{^-l^/ACTIONS=LINK^} are also specified.
9277 If the root unit specified by @var{file_name} is not a main unit, this is the
9278 default. Otherwise @command{gnatmake} will attempt binding and linking
9279 unless all objects are up to date and the executable is more recent than
9283 @cindex @option{^-C^/MAPPING^} (@command{gnatmake})
9284 Use a temporary mapping file. A mapping file is a way to communicate
9285 to the compiler two mappings: from unit names to file names (without
9286 any directory information) and from file names to path names (with
9287 full directory information). A mapping file can make the compiler's
9288 file searches faster, especially if there are many source directories,
9289 or the sources are read over a slow network connection. If
9290 @option{^-P^/PROJECT_FILE^} is used, a mapping file is always used, so
9291 @option{^-C^/MAPPING^} is unnecessary; in this case the mapping file
9292 is initially populated based on the project file. If
9293 @option{^-C^/MAPPING^} is used without
9294 @option{^-P^/PROJECT_FILE^},
9295 the mapping file is initially empty. Each invocation of the compiler
9296 will add any newly accessed sources to the mapping file.
9298 @item ^-C=^/USE_MAPPING_FILE=^@var{file}
9299 @cindex @option{^-C=^/USE_MAPPING^} (@command{gnatmake})
9300 Use a specific mapping file. The file, specified as a path name (absolute or
9301 relative) by this switch, should already exist, otherwise the switch is
9302 ineffective. The specified mapping file will be communicated to the compiler.
9303 This switch is not compatible with a project file
9304 (^-P^/PROJECT_FILE=^@var{file}) or with multiple compiling processes
9305 (^-j^/PROCESSES=^nnn, when nnn is greater than 1).
9307 @item ^-d^/DISPLAY_PROGRESS^
9308 @cindex @option{^-d^/DISPLAY_PROGRESS^} (@command{gnatmake})
9309 Display progress for each source, up to date or not, as a single line
9312 completed x out of y (zz%)
9315 If the file needs to be compiled this is displayed after the invocation of
9316 the compiler. These lines are displayed even in quiet output mode.
9318 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
9319 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@command{gnatmake})
9320 Put all object files and ALI file in directory @var{dir}.
9321 If the @option{^-D^/DIRECTORY_OBJECTS^} switch is not used, all object files
9322 and ALI files go in the current working directory.
9324 This switch cannot be used when using a project file.
9328 @cindex @option{-eL} (@command{gnatmake})
9329 @cindex symbolic links
9330 Follow all symbolic links when processing project files.
9331 This should be used if your project uses symbolic links for files or
9332 directories, but is not needed in other cases.
9334 @cindex naming scheme
9335 This also assumes that no directory matches the naming scheme for files (for
9336 instance that you do not have a directory called "sources.ads" when using the
9337 default GNAT naming scheme).
9339 When you do not have to use this switch (ie by default), gnatmake is able to
9340 save a lot of system calls (several per source file and object file), which
9341 can result in a significant speed up to load and manipulate a project file,
9342 especially when using source files from a remote system.
9346 @item ^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^
9347 @cindex @option{^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^} (@command{gnatmake})
9348 Output the commands for the compiler, the binder and the linker
9349 on ^standard output^SYS$OUTPUT^,
9350 instead of ^standard error^SYS$ERROR^.
9352 @item ^-f^/FORCE_COMPILE^
9353 @cindex @option{^-f^/FORCE_COMPILE^} (@command{gnatmake})
9354 Force recompilations. Recompile all sources, even though some object
9355 files may be up to date, but don't recompile predefined or GNAT internal
9356 files or locked files (files with a write-protected ALI file),
9357 unless the @option{^-a^/ALL_FILES^} switch is also specified.
9359 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
9360 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@command{gnatmake})
9361 When using project files, if some errors or warnings are detected during
9362 parsing and verbose mode is not in effect (no use of switch
9363 ^-v^/VERBOSE^), then error lines start with the full path name of the project
9364 file, rather than its simple file name.
9367 @cindex @option{^-g^/DEBUG^} (@command{gnatmake})
9368 Enable debugging. This switch is simply passed to the compiler and to the
9371 @item ^-i^/IN_PLACE^
9372 @cindex @option{^-i^/IN_PLACE^} (@command{gnatmake})
9373 In normal mode, @command{gnatmake} compiles all object files and ALI files
9374 into the current directory. If the @option{^-i^/IN_PLACE^} switch is used,
9375 then instead object files and ALI files that already exist are overwritten
9376 in place. This means that once a large project is organized into separate
9377 directories in the desired manner, then @command{gnatmake} will automatically
9378 maintain and update this organization. If no ALI files are found on the
9379 Ada object path (@ref{Search Paths and the Run-Time Library (RTL)}),
9380 the new object and ALI files are created in the
9381 directory containing the source being compiled. If another organization
9382 is desired, where objects and sources are kept in different directories,
9383 a useful technique is to create dummy ALI files in the desired directories.
9384 When detecting such a dummy file, @command{gnatmake} will be forced to
9385 recompile the corresponding source file, and it will be put the resulting
9386 object and ALI files in the directory where it found the dummy file.
9388 @item ^-j^/PROCESSES=^@var{n}
9389 @cindex @option{^-j^/PROCESSES^} (@command{gnatmake})
9390 @cindex Parallel make
9391 Use @var{n} processes to carry out the (re)compilations. On a
9392 multiprocessor machine compilations will occur in parallel. In the
9393 event of compilation errors, messages from various compilations might
9394 get interspersed (but @command{gnatmake} will give you the full ordered
9395 list of failing compiles at the end). If this is problematic, rerun
9396 the make process with n set to 1 to get a clean list of messages.
9398 @item ^-k^/CONTINUE_ON_ERROR^
9399 @cindex @option{^-k^/CONTINUE_ON_ERROR^} (@command{gnatmake})
9400 Keep going. Continue as much as possible after a compilation error. To
9401 ease the programmer's task in case of compilation errors, the list of
9402 sources for which the compile fails is given when @command{gnatmake}
9405 If @command{gnatmake} is invoked with several @file{file_names} and with this
9406 switch, if there are compilation errors when building an executable,
9407 @command{gnatmake} will not attempt to build the following executables.
9409 @item ^-l^/ACTIONS=LINK^
9410 @cindex @option{^-l^/ACTIONS=LINK^} (@command{gnatmake})
9411 Link only. Can be combined with @option{^-b^/ACTIONS=BIND^} to binding
9412 and linking. Linking will not be performed if combined with
9413 @option{^-c^/ACTIONS=COMPILE^}
9414 but not with @option{^-b^/ACTIONS=BIND^}.
9415 When not combined with @option{^-b^/ACTIONS=BIND^}
9416 all the units in the closure of the main program must have been previously
9417 compiled and must be up to date, and the main program needs to have been bound.
9418 The root unit specified by @var{file_name}
9419 may be given without extension, with the source extension or, if no GNAT
9420 Project File is specified, with the ALI file extension.
9422 @item ^-m^/MINIMAL_RECOMPILATION^
9423 @cindex @option{^-m^/MINIMAL_RECOMPILATION^} (@command{gnatmake})
9424 Specify that the minimum necessary amount of recompilations
9425 be performed. In this mode @command{gnatmake} ignores time
9426 stamp differences when the only
9427 modifications to a source file consist in adding/removing comments,
9428 empty lines, spaces or tabs. This means that if you have changed the
9429 comments in a source file or have simply reformatted it, using this
9430 switch will tell @command{gnatmake} not to recompile files that depend on it
9431 (provided other sources on which these files depend have undergone no
9432 semantic modifications). Note that the debugging information may be
9433 out of date with respect to the sources if the @option{-m} switch causes
9434 a compilation to be switched, so the use of this switch represents a
9435 trade-off between compilation time and accurate debugging information.
9437 @item ^-M^/DEPENDENCIES_LIST^
9438 @cindex Dependencies, producing list
9439 @cindex @option{^-M^/DEPENDENCIES_LIST^} (@command{gnatmake})
9440 Check if all objects are up to date. If they are, output the object
9441 dependences to @file{stdout} in a form that can be directly exploited in
9442 a @file{Makefile}. By default, each source file is prefixed with its
9443 (relative or absolute) directory name. This name is whatever you
9444 specified in the various @option{^-aI^/SOURCE_SEARCH^}
9445 and @option{^-I^/SEARCH^} switches. If you use
9446 @code{gnatmake ^-M^/DEPENDENCIES_LIST^}
9447 @option{^-q^/QUIET^}
9448 (see below), only the source file names,
9449 without relative paths, are output. If you just specify the
9450 @option{^-M^/DEPENDENCIES_LIST^}
9451 switch, dependencies of the GNAT internal system files are omitted. This
9452 is typically what you want. If you also specify
9453 the @option{^-a^/ALL_FILES^} switch,
9454 dependencies of the GNAT internal files are also listed. Note that
9455 dependencies of the objects in external Ada libraries (see switch
9456 @option{^-aL^/SKIP_MISSING=^}@var{dir} in the following list)
9459 @item ^-n^/DO_OBJECT_CHECK^
9460 @cindex @option{^-n^/DO_OBJECT_CHECK^} (@command{gnatmake})
9461 Don't compile, bind, or link. Checks if all objects are up to date.
9462 If they are not, the full name of the first file that needs to be
9463 recompiled is printed.
9464 Repeated use of this option, followed by compiling the indicated source
9465 file, will eventually result in recompiling all required units.
9467 @item ^-o ^/EXECUTABLE=^@var{exec_name}
9468 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatmake})
9469 Output executable name. The name of the final executable program will be
9470 @var{exec_name}. If the @option{^-o^/EXECUTABLE^} switch is omitted the default
9471 name for the executable will be the name of the input file in appropriate form
9472 for an executable file on the host system.
9474 This switch cannot be used when invoking @command{gnatmake} with several
9477 @item ^-p or --create-missing-dirs^/CREATE_MISSING_DIRS^
9478 @cindex @option{^-p^/CREATE_MISSING_DIRS^} (@command{gnatmake})
9479 When using project files (^-P^/PROJECT_FILE=^@var{project}), create
9480 automatically missing object directories, library directories and exec
9483 @item ^-P^/PROJECT_FILE=^@var{project}
9484 @cindex @option{^-P^/PROJECT_FILE^} (@command{gnatmake})
9485 Use project file @var{project}. Only one such switch can be used.
9486 @xref{gnatmake and Project Files}.
9489 @cindex @option{^-q^/QUIET^} (@command{gnatmake})
9490 Quiet. When this flag is not set, the commands carried out by
9491 @command{gnatmake} are displayed.
9493 @item ^-s^/SWITCH_CHECK/^
9494 @cindex @option{^-s^/SWITCH_CHECK^} (@command{gnatmake})
9495 Recompile if compiler switches have changed since last compilation.
9496 All compiler switches but -I and -o are taken into account in the
9498 orders between different ``first letter'' switches are ignored, but
9499 orders between same switches are taken into account. For example,
9500 @option{-O -O2} is different than @option{-O2 -O}, but @option{-g -O}
9501 is equivalent to @option{-O -g}.
9503 This switch is recommended when Integrated Preprocessing is used.
9506 @cindex @option{^-u^/UNIQUE^} (@command{gnatmake})
9507 Unique. Recompile at most the main files. It implies -c. Combined with
9508 -f, it is equivalent to calling the compiler directly. Note that using
9509 ^-u^/UNIQUE^ with a project file and no main has a special meaning
9510 (@pxref{Project Files and Main Subprograms}).
9512 @item ^-U^/ALL_PROJECTS^
9513 @cindex @option{^-U^/ALL_PROJECTS^} (@command{gnatmake})
9514 When used without a project file or with one or several mains on the command
9515 line, is equivalent to ^-u^/UNIQUE^. When used with a project file and no main
9516 on the command line, all sources of all project files are checked and compiled
9517 if not up to date, and libraries are rebuilt, if necessary.
9520 @cindex @option{^-v^/REASONS^} (@command{gnatmake})
9521 Verbose. Display the reason for all recompilations @command{gnatmake}
9522 decides are necessary, with the highest verbosity level.
9524 @item ^-vl^/LOW_VERBOSITY^
9525 @cindex @option{^-vl^/LOW_VERBOSITY^} (@command{gnatmake})
9526 Verbosity level Low. Display fewer lines than in verbosity Medium.
9528 @item ^-vm^/MEDIUM_VERBOSITY^
9529 @cindex @option{^-vm^/MEDIUM_VERBOSITY^} (@command{gnatmake})
9530 Verbosity level Medium. Potentially display fewer lines than in verbosity High.
9532 @item ^-vh^/HIGH_VERBOSITY^
9533 @cindex @option{^-vm^/HIGH_VERBOSITY^} (@command{gnatmake})
9534 Verbosity level High. Equivalent to ^-v^/REASONS^.
9536 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
9537 Indicate the verbosity of the parsing of GNAT project files.
9538 @xref{Switches Related to Project Files}.
9540 @item ^-x^/NON_PROJECT_UNIT_COMPILATION^
9541 @cindex @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} (@command{gnatmake})
9542 Indicate that sources that are not part of any Project File may be compiled.
9543 Normally, when using Project Files, only sources that are part of a Project
9544 File may be compile. When this switch is used, a source outside of all Project
9545 Files may be compiled. The ALI file and the object file will be put in the
9546 object directory of the main Project. The compilation switches used will only
9547 be those specified on the command line. Even when
9548 @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} is used, mains specified on the
9549 command line need to be sources of a project file.
9551 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
9552 Indicate that external variable @var{name} has the value @var{value}.
9553 The Project Manager will use this value for occurrences of
9554 @code{external(name)} when parsing the project file.
9555 @xref{Switches Related to Project Files}.
9558 @cindex @option{^-z^/NOMAIN^} (@command{gnatmake})
9559 No main subprogram. Bind and link the program even if the unit name
9560 given on the command line is a package name. The resulting executable
9561 will execute the elaboration routines of the package and its closure,
9562 then the finalization routines.
9567 @item @command{gcc} @asis{switches}
9569 Any uppercase or multi-character switch that is not a @command{gnatmake} switch
9570 is passed to @command{gcc} (e.g.@: @option{-O}, @option{-gnato,} etc.)
9573 Any qualifier that cannot be recognized as a qualifier for @code{GNAT MAKE}
9574 but is recognizable as a valid qualifier for @code{GNAT COMPILE} is
9575 automatically treated as a compiler switch, and passed on to all
9576 compilations that are carried out.
9581 Source and library search path switches:
9585 @item ^-aI^/SOURCE_SEARCH=^@var{dir}
9586 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatmake})
9587 When looking for source files also look in directory @var{dir}.
9588 The order in which source files search is undertaken is
9589 described in @ref{Search Paths and the Run-Time Library (RTL)}.
9591 @item ^-aL^/SKIP_MISSING=^@var{dir}
9592 @cindex @option{^-aL^/SKIP_MISSING^} (@command{gnatmake})
9593 Consider @var{dir} as being an externally provided Ada library.
9594 Instructs @command{gnatmake} to skip compilation units whose @file{.ALI}
9595 files have been located in directory @var{dir}. This allows you to have
9596 missing bodies for the units in @var{dir} and to ignore out of date bodies
9597 for the same units. You still need to specify
9598 the location of the specs for these units by using the switches
9599 @option{^-aI^/SOURCE_SEARCH=^@var{dir}}
9600 or @option{^-I^/SEARCH=^@var{dir}}.
9601 Note: this switch is provided for compatibility with previous versions
9602 of @command{gnatmake}. The easier method of causing standard libraries
9603 to be excluded from consideration is to write-protect the corresponding
9606 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
9607 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatmake})
9608 When searching for library and object files, look in directory
9609 @var{dir}. The order in which library files are searched is described in
9610 @ref{Search Paths for gnatbind}.
9612 @item ^-A^/CONDITIONAL_SOURCE_SEARCH=^@var{dir}
9613 @cindex Search paths, for @command{gnatmake}
9614 @cindex @option{^-A^/CONDITIONAL_SOURCE_SEARCH^} (@command{gnatmake})
9615 Equivalent to @option{^-aL^/SKIP_MISSING=^@var{dir}
9616 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9618 @item ^-I^/SEARCH=^@var{dir}
9619 @cindex @option{^-I^/SEARCH^} (@command{gnatmake})
9620 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}
9621 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9623 @item ^-I-^/NOCURRENT_DIRECTORY^
9624 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatmake})
9625 @cindex Source files, suppressing search
9626 Do not look for source files in the directory containing the source
9627 file named in the command line.
9628 Do not look for ALI or object files in the directory
9629 where @command{gnatmake} was invoked.
9631 @item ^-L^/LIBRARY_SEARCH=^@var{dir}
9632 @cindex @option{^-L^/LIBRARY_SEARCH^} (@command{gnatmake})
9633 @cindex Linker libraries
9634 Add directory @var{dir} to the list of directories in which the linker
9635 will search for libraries. This is equivalent to
9636 @option{-largs ^-L^/LIBRARY_SEARCH=^}@var{dir}.
9638 Furthermore, under Windows, the sources pointed to by the libraries path
9639 set in the registry are not searched for.
9643 @cindex @option{-nostdinc} (@command{gnatmake})
9644 Do not look for source files in the system default directory.
9647 @cindex @option{-nostdlib} (@command{gnatmake})
9648 Do not look for library files in the system default directory.
9650 @item --RTS=@var{rts-path}
9651 @cindex @option{--RTS} (@command{gnatmake})
9652 Specifies the default location of the runtime library. GNAT looks for the
9654 in the following directories, and stops as soon as a valid runtime is found
9655 (@file{adainclude} or @file{ada_source_path}, and @file{adalib} or
9656 @file{ada_object_path} present):
9659 @item <current directory>/$rts_path
9661 @item <default-search-dir>/$rts_path
9663 @item <default-search-dir>/rts-$rts_path
9667 The selected path is handled like a normal RTS path.
9671 @node Mode Switches for gnatmake
9672 @section Mode Switches for @command{gnatmake}
9675 The mode switches (referred to as @code{mode_switches}) allow the
9676 inclusion of switches that are to be passed to the compiler itself, the
9677 binder or the linker. The effect of a mode switch is to cause all
9678 subsequent switches up to the end of the switch list, or up to the next
9679 mode switch, to be interpreted as switches to be passed on to the
9680 designated component of GNAT.
9684 @item -cargs @var{switches}
9685 @cindex @option{-cargs} (@command{gnatmake})
9686 Compiler switches. Here @var{switches} is a list of switches
9687 that are valid switches for @command{gcc}. They will be passed on to
9688 all compile steps performed by @command{gnatmake}.
9690 @item -bargs @var{switches}
9691 @cindex @option{-bargs} (@command{gnatmake})
9692 Binder switches. Here @var{switches} is a list of switches
9693 that are valid switches for @code{gnatbind}. They will be passed on to
9694 all bind steps performed by @command{gnatmake}.
9696 @item -largs @var{switches}
9697 @cindex @option{-largs} (@command{gnatmake})
9698 Linker switches. Here @var{switches} is a list of switches
9699 that are valid switches for @command{gnatlink}. They will be passed on to
9700 all link steps performed by @command{gnatmake}.
9702 @item -margs @var{switches}
9703 @cindex @option{-margs} (@command{gnatmake})
9704 Make switches. The switches are directly interpreted by @command{gnatmake},
9705 regardless of any previous occurrence of @option{-cargs}, @option{-bargs}
9709 @node Notes on the Command Line
9710 @section Notes on the Command Line
9713 This section contains some additional useful notes on the operation
9714 of the @command{gnatmake} command.
9718 @cindex Recompilation, by @command{gnatmake}
9719 If @command{gnatmake} finds no ALI files, it recompiles the main program
9720 and all other units required by the main program.
9721 This means that @command{gnatmake}
9722 can be used for the initial compile, as well as during subsequent steps of
9723 the development cycle.
9726 If you enter @code{gnatmake @var{file}.adb}, where @file{@var{file}.adb}
9727 is a subunit or body of a generic unit, @command{gnatmake} recompiles
9728 @file{@var{file}.adb} (because it finds no ALI) and stops, issuing a
9732 In @command{gnatmake} the switch @option{^-I^/SEARCH^}
9733 is used to specify both source and
9734 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
9735 instead if you just want to specify
9736 source paths only and @option{^-aO^/OBJECT_SEARCH^}
9737 if you want to specify library paths
9741 @command{gnatmake} will ignore any files whose ALI file is write-protected.
9742 This may conveniently be used to exclude standard libraries from
9743 consideration and in particular it means that the use of the
9744 @option{^-f^/FORCE_COMPILE^} switch will not recompile these files
9745 unless @option{^-a^/ALL_FILES^} is also specified.
9748 @command{gnatmake} has been designed to make the use of Ada libraries
9749 particularly convenient. Assume you have an Ada library organized
9750 as follows: @i{^obj-dir^[OBJ_DIR]^} contains the objects and ALI files for
9751 of your Ada compilation units,
9752 whereas @i{^include-dir^[INCLUDE_DIR]^} contains the
9753 specs of these units, but no bodies. Then to compile a unit
9754 stored in @code{main.adb}, which uses this Ada library you would just type
9758 $ gnatmake -aI@var{include-dir} -aL@var{obj-dir} main
9761 $ gnatmake /SOURCE_SEARCH=@i{[INCLUDE_DIR]}
9762 /SKIP_MISSING=@i{[OBJ_DIR]} main
9767 Using @command{gnatmake} along with the
9768 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}
9769 switch provides a mechanism for avoiding unnecessary recompilations. Using
9771 you can update the comments/format of your
9772 source files without having to recompile everything. Note, however, that
9773 adding or deleting lines in a source files may render its debugging
9774 info obsolete. If the file in question is a spec, the impact is rather
9775 limited, as that debugging info will only be useful during the
9776 elaboration phase of your program. For bodies the impact can be more
9777 significant. In all events, your debugger will warn you if a source file
9778 is more recent than the corresponding object, and alert you to the fact
9779 that the debugging information may be out of date.
9782 @node How gnatmake Works
9783 @section How @command{gnatmake} Works
9786 Generally @command{gnatmake} automatically performs all necessary
9787 recompilations and you don't need to worry about how it works. However,
9788 it may be useful to have some basic understanding of the @command{gnatmake}
9789 approach and in particular to understand how it uses the results of
9790 previous compilations without incorrectly depending on them.
9792 First a definition: an object file is considered @dfn{up to date} if the
9793 corresponding ALI file exists and if all the source files listed in the
9794 dependency section of this ALI file have time stamps matching those in
9795 the ALI file. This means that neither the source file itself nor any
9796 files that it depends on have been modified, and hence there is no need
9797 to recompile this file.
9799 @command{gnatmake} works by first checking if the specified main unit is up
9800 to date. If so, no compilations are required for the main unit. If not,
9801 @command{gnatmake} compiles the main program to build a new ALI file that
9802 reflects the latest sources. Then the ALI file of the main unit is
9803 examined to find all the source files on which the main program depends,
9804 and @command{gnatmake} recursively applies the above procedure on all these
9807 This process ensures that @command{gnatmake} only trusts the dependencies
9808 in an existing ALI file if they are known to be correct. Otherwise it
9809 always recompiles to determine a new, guaranteed accurate set of
9810 dependencies. As a result the program is compiled ``upside down'' from what may
9811 be more familiar as the required order of compilation in some other Ada
9812 systems. In particular, clients are compiled before the units on which
9813 they depend. The ability of GNAT to compile in any order is critical in
9814 allowing an order of compilation to be chosen that guarantees that
9815 @command{gnatmake} will recompute a correct set of new dependencies if
9818 When invoking @command{gnatmake} with several @var{file_names}, if a unit is
9819 imported by several of the executables, it will be recompiled at most once.
9821 Note: when using non-standard naming conventions
9822 (@pxref{Using Other File Names}), changing through a configuration pragmas
9823 file the version of a source and invoking @command{gnatmake} to recompile may
9824 have no effect, if the previous version of the source is still accessible
9825 by @command{gnatmake}. It may be necessary to use the switch
9826 ^-f^/FORCE_COMPILE^.
9828 @node Examples of gnatmake Usage
9829 @section Examples of @command{gnatmake} Usage
9832 @item gnatmake hello.adb
9833 Compile all files necessary to bind and link the main program
9834 @file{hello.adb} (containing unit @code{Hello}) and bind and link the
9835 resulting object files to generate an executable file @file{^hello^HELLO.EXE^}.
9837 @item gnatmake main1 main2 main3
9838 Compile all files necessary to bind and link the main programs
9839 @file{main1.adb} (containing unit @code{Main1}), @file{main2.adb}
9840 (containing unit @code{Main2}) and @file{main3.adb}
9841 (containing unit @code{Main3}) and bind and link the resulting object files
9842 to generate three executable files @file{^main1^MAIN1.EXE^},
9843 @file{^main2^MAIN2.EXE^}
9844 and @file{^main3^MAIN3.EXE^}.
9847 @item gnatmake -q Main_Unit -cargs -O2 -bargs -l
9851 @item gnatmake Main_Unit /QUIET
9852 /COMPILER_QUALIFIERS /OPTIMIZE=ALL
9853 /BINDER_QUALIFIERS /ORDER_OF_ELABORATION
9855 Compile all files necessary to bind and link the main program unit
9856 @code{Main_Unit} (from file @file{main_unit.adb}). All compilations will
9857 be done with optimization level 2 and the order of elaboration will be
9858 listed by the binder. @command{gnatmake} will operate in quiet mode, not
9859 displaying commands it is executing.
9862 @c *************************
9863 @node Improving Performance
9864 @chapter Improving Performance
9865 @cindex Improving performance
9868 This chapter presents several topics related to program performance.
9869 It first describes some of the tradeoffs that need to be considered
9870 and some of the techniques for making your program run faster.
9871 It then documents the @command{gnatelim} tool and unused subprogram/data
9872 elimination feature, which can reduce the size of program executables.
9874 Note: to invoke @command{gnatelim} with a project file, use the @code{gnat}
9875 driver (see @ref{The GNAT Driver and Project Files}).
9879 * Performance Considerations::
9880 * Text_IO Suggestions::
9881 * Reducing Size of Ada Executables with gnatelim::
9882 * Reducing Size of Executables with unused subprogram/data elimination::
9886 @c *****************************
9887 @node Performance Considerations
9888 @section Performance Considerations
9891 The GNAT system provides a number of options that allow a trade-off
9896 performance of the generated code
9899 speed of compilation
9902 minimization of dependences and recompilation
9905 the degree of run-time checking.
9909 The defaults (if no options are selected) aim at improving the speed
9910 of compilation and minimizing dependences, at the expense of performance
9911 of the generated code:
9918 no inlining of subprogram calls
9921 all run-time checks enabled except overflow and elaboration checks
9925 These options are suitable for most program development purposes. This
9926 chapter describes how you can modify these choices, and also provides
9927 some guidelines on debugging optimized code.
9930 * Controlling Run-Time Checks::
9931 * Use of Restrictions::
9932 * Optimization Levels::
9933 * Debugging Optimized Code::
9934 * Inlining of Subprograms::
9935 * Other Optimization Switches::
9936 * Optimization and Strict Aliasing::
9939 * Coverage Analysis::
9943 @node Controlling Run-Time Checks
9944 @subsection Controlling Run-Time Checks
9947 By default, GNAT generates all run-time checks, except integer overflow
9948 checks, stack overflow checks, and checks for access before elaboration on
9949 subprogram calls. The latter are not required in default mode, because all
9950 necessary checking is done at compile time.
9951 @cindex @option{-gnatp} (@command{gcc})
9952 @cindex @option{-gnato} (@command{gcc})
9953 Two gnat switches, @option{-gnatp} and @option{-gnato} allow this default to
9954 be modified. @xref{Run-Time Checks}.
9956 Our experience is that the default is suitable for most development
9959 We treat integer overflow specially because these
9960 are quite expensive and in our experience are not as important as other
9961 run-time checks in the development process. Note that division by zero
9962 is not considered an overflow check, and divide by zero checks are
9963 generated where required by default.
9965 Elaboration checks are off by default, and also not needed by default, since
9966 GNAT uses a static elaboration analysis approach that avoids the need for
9967 run-time checking. This manual contains a full chapter discussing the issue
9968 of elaboration checks, and if the default is not satisfactory for your use,
9969 you should read this chapter.
9971 For validity checks, the minimal checks required by the Ada Reference
9972 Manual (for case statements and assignments to array elements) are on
9973 by default. These can be suppressed by use of the @option{-gnatVn} switch.
9974 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
9975 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
9976 it may be reasonable to routinely use @option{-gnatVn}. Validity checks
9977 are also suppressed entirely if @option{-gnatp} is used.
9979 @cindex Overflow checks
9980 @cindex Checks, overflow
9983 @cindex pragma Suppress
9984 @cindex pragma Unsuppress
9985 Note that the setting of the switches controls the default setting of
9986 the checks. They may be modified using either @code{pragma Suppress} (to
9987 remove checks) or @code{pragma Unsuppress} (to add back suppressed
9988 checks) in the program source.
9990 @node Use of Restrictions
9991 @subsection Use of Restrictions
9994 The use of pragma Restrictions allows you to control which features are
9995 permitted in your program. Apart from the obvious point that if you avoid
9996 relatively expensive features like finalization (enforceable by the use
9997 of pragma Restrictions (No_Finalization), the use of this pragma does not
9998 affect the generated code in most cases.
10000 One notable exception to this rule is that the possibility of task abort
10001 results in some distributed overhead, particularly if finalization or
10002 exception handlers are used. The reason is that certain sections of code
10003 have to be marked as non-abortable.
10005 If you use neither the @code{abort} statement, nor asynchronous transfer
10006 of control (@code{select @dots{} then abort}), then this distributed overhead
10007 is removed, which may have a general positive effect in improving
10008 overall performance. Especially code involving frequent use of tasking
10009 constructs and controlled types will show much improved performance.
10010 The relevant restrictions pragmas are
10012 @smallexample @c ada
10013 pragma Restrictions (No_Abort_Statements);
10014 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
10018 It is recommended that these restriction pragmas be used if possible. Note
10019 that this also means that you can write code without worrying about the
10020 possibility of an immediate abort at any point.
10022 @node Optimization Levels
10023 @subsection Optimization Levels
10024 @cindex @option{^-O^/OPTIMIZE^} (@command{gcc})
10027 Without any optimization ^option,^qualifier,^
10028 the compiler's goal is to reduce the cost of
10029 compilation and to make debugging produce the expected results.
10030 Statements are independent: if you stop the program with a breakpoint between
10031 statements, you can then assign a new value to any variable or change
10032 the program counter to any other statement in the subprogram and get exactly
10033 the results you would expect from the source code.
10035 Turning on optimization makes the compiler attempt to improve the
10036 performance and/or code size at the expense of compilation time and
10037 possibly the ability to debug the program.
10039 If you use multiple
10040 ^-O options, with or without level numbers,^/OPTIMIZE qualifiers,^
10041 the last such option is the one that is effective.
10044 The default is optimization off. This results in the fastest compile
10045 times, but GNAT makes absolutely no attempt to optimize, and the
10046 generated programs are considerably larger and slower than when
10047 optimization is enabled. You can use the
10049 @option{-O} switch (the permitted forms are @option{-O0}, @option{-O1}
10050 @option{-O2}, @option{-O3}, and @option{-Os})
10053 @code{OPTIMIZE} qualifier
10055 to @command{gcc} to control the optimization level:
10058 @item ^-O0^/OPTIMIZE=NONE^
10059 No optimization (the default);
10060 generates unoptimized code but has
10061 the fastest compilation time.
10063 Note that many other compilers do fairly extensive optimization
10064 even if ``no optimization'' is specified. With gcc, it is
10065 very unusual to use ^-O0^/OPTIMIZE=NONE^ for production if
10066 execution time is of any concern, since ^-O0^/OPTIMIZE=NONE^
10067 really does mean no optimization at all. This difference between
10068 gcc and other compilers should be kept in mind when doing
10069 performance comparisons.
10071 @item ^-O1^/OPTIMIZE=SOME^
10072 Moderate optimization;
10073 optimizes reasonably well but does not
10074 degrade compilation time significantly.
10076 @item ^-O2^/OPTIMIZE=ALL^
10078 @itemx /OPTIMIZE=DEVELOPMENT
10081 generates highly optimized code and has
10082 the slowest compilation time.
10084 @item ^-O3^/OPTIMIZE=INLINING^
10085 Full optimization as in @option{-O2},
10086 and also attempts automatic inlining of small
10087 subprograms within a unit (@pxref{Inlining of Subprograms}).
10089 @item ^-Os^/OPTIMIZE=SPACE^
10090 Optimize space usage of resulting program.
10094 Higher optimization levels perform more global transformations on the
10095 program and apply more expensive analysis algorithms in order to generate
10096 faster and more compact code. The price in compilation time, and the
10097 resulting improvement in execution time,
10098 both depend on the particular application and the hardware environment.
10099 You should experiment to find the best level for your application.
10101 Since the precise set of optimizations done at each level will vary from
10102 release to release (and sometime from target to target), it is best to think
10103 of the optimization settings in general terms.
10104 @xref{Optimize Options,, Options That Control Optimization, gcc, Using
10105 the GNU Compiler Collection (GCC)}, for details about
10106 ^the @option{-O} settings and a number of @option{-f} options that^how to^
10107 individually enable or disable specific optimizations.
10109 Unlike some other compilation systems, ^@command{gcc}^GNAT^ has
10110 been tested extensively at all optimization levels. There are some bugs
10111 which appear only with optimization turned on, but there have also been
10112 bugs which show up only in @emph{unoptimized} code. Selecting a lower
10113 level of optimization does not improve the reliability of the code
10114 generator, which in practice is highly reliable at all optimization
10117 Note regarding the use of @option{-O3}: The use of this optimization level
10118 is generally discouraged with GNAT, since it often results in larger
10119 executables which run more slowly. See further discussion of this point
10120 in @ref{Inlining of Subprograms}.
10122 @node Debugging Optimized Code
10123 @subsection Debugging Optimized Code
10124 @cindex Debugging optimized code
10125 @cindex Optimization and debugging
10128 Although it is possible to do a reasonable amount of debugging at
10130 nonzero optimization levels,
10131 the higher the level the more likely that
10134 @option{/OPTIMIZE} settings other than @code{NONE},
10135 such settings will make it more likely that
10137 source-level constructs will have been eliminated by optimization.
10138 For example, if a loop is strength-reduced, the loop
10139 control variable may be completely eliminated and thus cannot be
10140 displayed in the debugger.
10141 This can only happen at @option{-O2} or @option{-O3}.
10142 Explicit temporary variables that you code might be eliminated at
10143 ^level^setting^ @option{-O1} or higher.
10145 The use of the @option{^-g^/DEBUG^} switch,
10146 @cindex @option{^-g^/DEBUG^} (@command{gcc})
10147 which is needed for source-level debugging,
10148 affects the size of the program executable on disk,
10149 and indeed the debugging information can be quite large.
10150 However, it has no effect on the generated code (and thus does not
10151 degrade performance)
10153 Since the compiler generates debugging tables for a compilation unit before
10154 it performs optimizations, the optimizing transformations may invalidate some
10155 of the debugging data. You therefore need to anticipate certain
10156 anomalous situations that may arise while debugging optimized code.
10157 These are the most common cases:
10161 @i{The ``hopping Program Counter'':} Repeated @code{step} or @code{next}
10163 the PC bouncing back and forth in the code. This may result from any of
10164 the following optimizations:
10168 @i{Common subexpression elimination:} using a single instance of code for a
10169 quantity that the source computes several times. As a result you
10170 may not be able to stop on what looks like a statement.
10173 @i{Invariant code motion:} moving an expression that does not change within a
10174 loop, to the beginning of the loop.
10177 @i{Instruction scheduling:} moving instructions so as to
10178 overlap loads and stores (typically) with other code, or in
10179 general to move computations of values closer to their uses. Often
10180 this causes you to pass an assignment statement without the assignment
10181 happening and then later bounce back to the statement when the
10182 value is actually needed. Placing a breakpoint on a line of code
10183 and then stepping over it may, therefore, not always cause all the
10184 expected side-effects.
10188 @i{The ``big leap'':} More commonly known as @emph{cross-jumping}, in which
10189 two identical pieces of code are merged and the program counter suddenly
10190 jumps to a statement that is not supposed to be executed, simply because
10191 it (and the code following) translates to the same thing as the code
10192 that @emph{was} supposed to be executed. This effect is typically seen in
10193 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
10194 a @code{break} in a C @code{^switch^switch^} statement.
10197 @i{The ``roving variable'':} The symptom is an unexpected value in a variable.
10198 There are various reasons for this effect:
10202 In a subprogram prologue, a parameter may not yet have been moved to its
10206 A variable may be dead, and its register re-used. This is
10207 probably the most common cause.
10210 As mentioned above, the assignment of a value to a variable may
10214 A variable may be eliminated entirely by value propagation or
10215 other means. In this case, GCC may incorrectly generate debugging
10216 information for the variable
10220 In general, when an unexpected value appears for a local variable or parameter
10221 you should first ascertain if that value was actually computed by
10222 your program, as opposed to being incorrectly reported by the debugger.
10224 array elements in an object designated by an access value
10225 are generally less of a problem, once you have ascertained that the access
10227 Typically, this means checking variables in the preceding code and in the
10228 calling subprogram to verify that the value observed is explainable from other
10229 values (one must apply the procedure recursively to those
10230 other values); or re-running the code and stopping a little earlier
10231 (perhaps before the call) and stepping to better see how the variable obtained
10232 the value in question; or continuing to step @emph{from} the point of the
10233 strange value to see if code motion had simply moved the variable's
10238 In light of such anomalies, a recommended technique is to use @option{-O0}
10239 early in the software development cycle, when extensive debugging capabilities
10240 are most needed, and then move to @option{-O1} and later @option{-O2} as
10241 the debugger becomes less critical.
10242 Whether to use the @option{^-g^/DEBUG^} switch in the release version is
10243 a release management issue.
10245 Note that if you use @option{-g} you can then use the @command{strip} program
10246 on the resulting executable,
10247 which removes both debugging information and global symbols.
10250 @node Inlining of Subprograms
10251 @subsection Inlining of Subprograms
10254 A call to a subprogram in the current unit is inlined if all the
10255 following conditions are met:
10259 The optimization level is at least @option{-O1}.
10262 The called subprogram is suitable for inlining: It must be small enough
10263 and not contain something that @command{gcc} cannot support in inlined
10267 @cindex pragma Inline
10269 Either @code{pragma Inline} applies to the subprogram, or it is local
10270 to the unit and called once from within it, or it is small and automatic
10271 inlining (optimization level @option{-O3}) is specified.
10275 Calls to subprograms in @code{with}'ed units are normally not inlined.
10276 To achieve actual inlining (that is, replacement of the call by the code
10277 in the body of the subprogram), the following conditions must all be true.
10281 The optimization level is at least @option{-O1}.
10284 The called subprogram is suitable for inlining: It must be small enough
10285 and not contain something that @command{gcc} cannot support in inlined
10289 The call appears in a body (not in a package spec).
10292 There is a @code{pragma Inline} for the subprogram.
10295 @cindex @option{-gnatn} (@command{gcc})
10296 The @option{^-gnatn^/INLINE^} switch
10297 is used in the @command{gcc} command line
10300 Even if all these conditions are met, it may not be possible for
10301 the compiler to inline the call, due to the length of the body,
10302 or features in the body that make it impossible for the compiler
10303 to do the inlining.
10305 Note that specifying the @option{-gnatn} switch causes additional
10306 compilation dependencies. Consider the following:
10308 @smallexample @c ada
10328 With the default behavior (no @option{-gnatn} switch specified), the
10329 compilation of the @code{Main} procedure depends only on its own source,
10330 @file{main.adb}, and the spec of the package in file @file{r.ads}. This
10331 means that editing the body of @code{R} does not require recompiling
10334 On the other hand, the call @code{R.Q} is not inlined under these
10335 circumstances. If the @option{-gnatn} switch is present when @code{Main}
10336 is compiled, the call will be inlined if the body of @code{Q} is small
10337 enough, but now @code{Main} depends on the body of @code{R} in
10338 @file{r.adb} as well as on the spec. This means that if this body is edited,
10339 the main program must be recompiled. Note that this extra dependency
10340 occurs whether or not the call is in fact inlined by @command{gcc}.
10342 The use of front end inlining with @option{-gnatN} generates similar
10343 additional dependencies.
10345 @cindex @option{^-fno-inline^/INLINE=SUPPRESS^} (@command{gcc})
10346 Note: The @option{^-fno-inline^/INLINE=SUPPRESS^} switch
10347 can be used to prevent
10348 all inlining. This switch overrides all other conditions and ensures
10349 that no inlining occurs. The extra dependences resulting from
10350 @option{-gnatn} will still be active, even if
10351 this switch is used to suppress the resulting inlining actions.
10353 @cindex @option{-fno-inline-functions} (@command{gcc})
10354 Note: The @option{-fno-inline-functions} switch can be used to prevent
10355 automatic inlining of small subprograms if @option{-O3} is used.
10357 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
10358 Note: The @option{-fno-inline-functions-called-once} switch
10359 can be used to prevent inlining of subprograms local to the unit
10360 and called once from within it if @option{-O1} is used.
10362 Note regarding the use of @option{-O3}: There is no difference in inlining
10363 behavior between @option{-O2} and @option{-O3} for subprograms with an explicit
10364 pragma @code{Inline} assuming the use of @option{-gnatn}
10365 or @option{-gnatN} (the switches that activate inlining). If you have used
10366 pragma @code{Inline} in appropriate cases, then it is usually much better
10367 to use @option{-O2} and @option{-gnatn} and avoid the use of @option{-O3} which
10368 in this case only has the effect of inlining subprograms you did not
10369 think should be inlined. We often find that the use of @option{-O3} slows
10370 down code by performing excessive inlining, leading to increased instruction
10371 cache pressure from the increased code size. So the bottom line here is
10372 that you should not automatically assume that @option{-O3} is better than
10373 @option{-O2}, and indeed you should use @option{-O3} only if tests show that
10374 it actually improves performance.
10376 @node Other Optimization Switches
10377 @subsection Other Optimization Switches
10378 @cindex Optimization Switches
10380 Since @code{GNAT} uses the @command{gcc} back end, all the specialized
10381 @command{gcc} optimization switches are potentially usable. These switches
10382 have not been extensively tested with GNAT but can generally be expected
10383 to work. Examples of switches in this category are
10384 @option{-funroll-loops} and
10385 the various target-specific @option{-m} options (in particular, it has been
10386 observed that @option{-march=pentium4} can significantly improve performance
10387 on appropriate machines). For full details of these switches, see
10388 @ref{Submodel Options,, Hardware Models and Configurations, gcc, Using
10389 the GNU Compiler Collection (GCC)}.
10391 @node Optimization and Strict Aliasing
10392 @subsection Optimization and Strict Aliasing
10394 @cindex Strict Aliasing
10395 @cindex No_Strict_Aliasing
10398 The strong typing capabilities of Ada allow an optimizer to generate
10399 efficient code in situations where other languages would be forced to
10400 make worst case assumptions preventing such optimizations. Consider
10401 the following example:
10403 @smallexample @c ada
10406 type Int1 is new Integer;
10407 type Int2 is new Integer;
10408 type Int1A is access Int1;
10409 type Int2A is access Int2;
10416 for J in Data'Range loop
10417 if Data (J) = Int1V.all then
10418 Int2V.all := Int2V.all + 1;
10427 In this example, since the variable @code{Int1V} can only access objects
10428 of type @code{Int1}, and @code{Int2V} can only access objects of type
10429 @code{Int2}, there is no possibility that the assignment to
10430 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
10431 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
10432 for all iterations of the loop and avoid the extra memory reference
10433 required to dereference it each time through the loop.
10435 This kind of optimization, called strict aliasing analysis, is
10436 triggered by specifying an optimization level of @option{-O2} or
10437 higher or @option{-Os} and allows @code{GNAT} to generate more efficient code
10438 when access values are involved.
10440 However, although this optimization is always correct in terms of
10441 the formal semantics of the Ada Reference Manual, difficulties can
10442 arise if features like @code{Unchecked_Conversion} are used to break
10443 the typing system. Consider the following complete program example:
10445 @smallexample @c ada
10448 type int1 is new integer;
10449 type int2 is new integer;
10450 type a1 is access int1;
10451 type a2 is access int2;
10456 function to_a2 (Input : a1) return a2;
10459 with Unchecked_Conversion;
10461 function to_a2 (Input : a1) return a2 is
10463 new Unchecked_Conversion (a1, a2);
10465 return to_a2u (Input);
10471 with Text_IO; use Text_IO;
10473 v1 : a1 := new int1;
10474 v2 : a2 := to_a2 (v1);
10478 put_line (int1'image (v1.all));
10484 This program prints out 0 in @option{-O0} or @option{-O1}
10485 mode, but it prints out 1 in @option{-O2} mode. That's
10486 because in strict aliasing mode, the compiler can and
10487 does assume that the assignment to @code{v2.all} could not
10488 affect the value of @code{v1.all}, since different types
10491 This behavior is not a case of non-conformance with the standard, since
10492 the Ada RM specifies that an unchecked conversion where the resulting
10493 bit pattern is not a correct value of the target type can result in an
10494 abnormal value and attempting to reference an abnormal value makes the
10495 execution of a program erroneous. That's the case here since the result
10496 does not point to an object of type @code{int2}. This means that the
10497 effect is entirely unpredictable.
10499 However, although that explanation may satisfy a language
10500 lawyer, in practice an applications programmer expects an
10501 unchecked conversion involving pointers to create true
10502 aliases and the behavior of printing 1 seems plain wrong.
10503 In this case, the strict aliasing optimization is unwelcome.
10505 Indeed the compiler recognizes this possibility, and the
10506 unchecked conversion generates a warning:
10509 p2.adb:5:07: warning: possible aliasing problem with type "a2"
10510 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
10511 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
10515 Unfortunately the problem is recognized when compiling the body of
10516 package @code{p2}, but the actual "bad" code is generated while
10517 compiling the body of @code{m} and this latter compilation does not see
10518 the suspicious @code{Unchecked_Conversion}.
10520 As implied by the warning message, there are approaches you can use to
10521 avoid the unwanted strict aliasing optimization in a case like this.
10523 One possibility is to simply avoid the use of @option{-O2}, but
10524 that is a bit drastic, since it throws away a number of useful
10525 optimizations that do not involve strict aliasing assumptions.
10527 A less drastic approach is to compile the program using the
10528 option @option{-fno-strict-aliasing}. Actually it is only the
10529 unit containing the dereferencing of the suspicious pointer
10530 that needs to be compiled. So in this case, if we compile
10531 unit @code{m} with this switch, then we get the expected
10532 value of zero printed. Analyzing which units might need
10533 the switch can be painful, so a more reasonable approach
10534 is to compile the entire program with options @option{-O2}
10535 and @option{-fno-strict-aliasing}. If the performance is
10536 satisfactory with this combination of options, then the
10537 advantage is that the entire issue of possible "wrong"
10538 optimization due to strict aliasing is avoided.
10540 To avoid the use of compiler switches, the configuration
10541 pragma @code{No_Strict_Aliasing} with no parameters may be
10542 used to specify that for all access types, the strict
10543 aliasing optimization should be suppressed.
10545 However, these approaches are still overkill, in that they causes
10546 all manipulations of all access values to be deoptimized. A more
10547 refined approach is to concentrate attention on the specific
10548 access type identified as problematic.
10550 First, if a careful analysis of uses of the pointer shows
10551 that there are no possible problematic references, then
10552 the warning can be suppressed by bracketing the
10553 instantiation of @code{Unchecked_Conversion} to turn
10556 @smallexample @c ada
10557 pragma Warnings (Off);
10559 new Unchecked_Conversion (a1, a2);
10560 pragma Warnings (On);
10564 Of course that approach is not appropriate for this particular
10565 example, since indeed there is a problematic reference. In this
10566 case we can take one of two other approaches.
10568 The first possibility is to move the instantiation of unchecked
10569 conversion to the unit in which the type is declared. In
10570 this example, we would move the instantiation of
10571 @code{Unchecked_Conversion} from the body of package
10572 @code{p2} to the spec of package @code{p1}. Now the
10573 warning disappears. That's because any use of the
10574 access type knows there is a suspicious unchecked
10575 conversion, and the strict aliasing optimization
10576 is automatically suppressed for the type.
10578 If it is not practical to move the unchecked conversion to the same unit
10579 in which the destination access type is declared (perhaps because the
10580 source type is not visible in that unit), you may use pragma
10581 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
10582 same declarative sequence as the declaration of the access type:
10584 @smallexample @c ada
10585 type a2 is access int2;
10586 pragma No_Strict_Aliasing (a2);
10590 Here again, the compiler now knows that the strict aliasing optimization
10591 should be suppressed for any reference to type @code{a2} and the
10592 expected behavior is obtained.
10594 Finally, note that although the compiler can generate warnings for
10595 simple cases of unchecked conversions, there are tricker and more
10596 indirect ways of creating type incorrect aliases which the compiler
10597 cannot detect. Examples are the use of address overlays and unchecked
10598 conversions involving composite types containing access types as
10599 components. In such cases, no warnings are generated, but there can
10600 still be aliasing problems. One safe coding practice is to forbid the
10601 use of address clauses for type overlaying, and to allow unchecked
10602 conversion only for primitive types. This is not really a significant
10603 restriction since any possible desired effect can be achieved by
10604 unchecked conversion of access values.
10606 The aliasing analysis done in strict aliasing mode can certainly
10607 have significant benefits. We have seen cases of large scale
10608 application code where the time is increased by up to 5% by turning
10609 this optimization off. If you have code that includes significant
10610 usage of unchecked conversion, you might want to just stick with
10611 @option{-O1} and avoid the entire issue. If you get adequate
10612 performance at this level of optimization level, that's probably
10613 the safest approach. If tests show that you really need higher
10614 levels of optimization, then you can experiment with @option{-O2}
10615 and @option{-O2 -fno-strict-aliasing} to see how much effect this
10616 has on size and speed of the code. If you really need to use
10617 @option{-O2} with strict aliasing in effect, then you should
10618 review any uses of unchecked conversion of access types,
10619 particularly if you are getting the warnings described above.
10622 @node Coverage Analysis
10623 @subsection Coverage Analysis
10626 GNAT supports the HP Performance Coverage Analyzer (PCA), which allows
10627 the user to determine the distribution of execution time across a program,
10628 @pxref{Profiling} for details of usage.
10632 @node Text_IO Suggestions
10633 @section @code{Text_IO} Suggestions
10634 @cindex @code{Text_IO} and performance
10637 The @code{Ada.Text_IO} package has fairly high overheads due in part to
10638 the requirement of maintaining page and line counts. If performance
10639 is critical, a recommendation is to use @code{Stream_IO} instead of
10640 @code{Text_IO} for volume output, since this package has less overhead.
10642 If @code{Text_IO} must be used, note that by default output to the standard
10643 output and standard error files is unbuffered (this provides better
10644 behavior when output statements are used for debugging, or if the
10645 progress of a program is observed by tracking the output, e.g. by
10646 using the Unix @command{tail -f} command to watch redirected output.
10648 If you are generating large volumes of output with @code{Text_IO} and
10649 performance is an important factor, use a designated file instead
10650 of the standard output file, or change the standard output file to
10651 be buffered using @code{Interfaces.C_Streams.setvbuf}.
10655 @node Reducing Size of Ada Executables with gnatelim
10656 @section Reducing Size of Ada Executables with @code{gnatelim}
10660 This section describes @command{gnatelim}, a tool which detects unused
10661 subprograms and helps the compiler to create a smaller executable for your
10666 * Running gnatelim::
10667 * Correcting the List of Eliminate Pragmas::
10668 * Making Your Executables Smaller::
10669 * Summary of the gnatelim Usage Cycle::
10672 @node About gnatelim
10673 @subsection About @code{gnatelim}
10676 When a program shares a set of Ada
10677 packages with other programs, it may happen that this program uses
10678 only a fraction of the subprograms defined in these packages. The code
10679 created for these unused subprograms increases the size of the executable.
10681 @code{gnatelim} tracks unused subprograms in an Ada program and
10682 outputs a list of GNAT-specific pragmas @code{Eliminate} marking all the
10683 subprograms that are declared but never called. By placing the list of
10684 @code{Eliminate} pragmas in the GNAT configuration file @file{gnat.adc} and
10685 recompiling your program, you may decrease the size of its executable,
10686 because the compiler will not generate the code for 'eliminated' subprograms.
10687 @xref{Pragma Eliminate,,, gnat_rm, GNAT Reference Manual}, for more
10688 information about this pragma.
10690 @code{gnatelim} needs as its input data the name of the main subprogram
10691 and a bind file for a main subprogram.
10693 To create a bind file for @code{gnatelim}, run @code{gnatbind} for
10694 the main subprogram. @code{gnatelim} can work with both Ada and C
10695 bind files; when both are present, it uses the Ada bind file.
10696 The following commands will build the program and create the bind file:
10699 $ gnatmake ^-c Main_Prog^/ACTIONS=COMPILE MAIN_PROG^
10700 $ gnatbind main_prog
10703 Note that @code{gnatelim} needs neither object nor ALI files.
10705 @node Running gnatelim
10706 @subsection Running @code{gnatelim}
10709 @code{gnatelim} has the following command-line interface:
10712 $ gnatelim @ovar{options} name
10716 @code{name} should be a name of a source file that contains the main subprogram
10717 of a program (partition).
10719 @code{gnatelim} has the following switches:
10724 @cindex @option{^-q^/QUIET^} (@command{gnatelim})
10725 Quiet mode: by default @code{gnatelim} outputs to the standard error
10726 stream the number of program units left to be processed. This option turns
10729 @item ^-v^/VERBOSE^
10730 @cindex @option{^-v^/VERBOSE^} (@command{gnatelim})
10731 Verbose mode: @code{gnatelim} version information is printed as Ada
10732 comments to the standard output stream. Also, in addition to the number of
10733 program units left @code{gnatelim} will output the name of the current unit
10737 @cindex @option{^-a^/ALL^} (@command{gnatelim})
10738 Also look for subprograms from the GNAT run time that can be eliminated. Note
10739 that when @file{gnat.adc} is produced using this switch, the entire program
10740 must be recompiled with switch @option{^-a^/ALL_FILES^} to @command{gnatmake}.
10742 @item ^-I^/INCLUDE_DIRS=^@var{dir}
10743 @cindex @option{^-I^/INCLUDE_DIRS^} (@command{gnatelim})
10744 When looking for source files also look in directory @var{dir}. Specifying
10745 @option{^-I-^/INCLUDE_DIRS=-^} instructs @code{gnatelim} not to look for
10746 sources in the current directory.
10748 @item ^-b^/BIND_FILE=^@var{bind_file}
10749 @cindex @option{^-b^/BIND_FILE^} (@command{gnatelim})
10750 Specifies @var{bind_file} as the bind file to process. If not set, the name
10751 of the bind file is computed from the full expanded Ada name
10752 of a main subprogram.
10754 @item ^-C^/CONFIG_FILE=^@var{config_file}
10755 @cindex @option{^-C^/CONFIG_FILE^} (@command{gnatelim})
10756 Specifies a file @var{config_file} that contains configuration pragmas. The
10757 file must be specified with full path.
10759 @item ^--GCC^/COMPILER^=@var{compiler_name}
10760 @cindex @option{^-GCC^/COMPILER^} (@command{gnatelim})
10761 Instructs @code{gnatelim} to use specific @command{gcc} compiler instead of one
10762 available on the path.
10764 @item ^--GNATMAKE^/GNATMAKE^=@var{gnatmake_name}
10765 @cindex @option{^--GNATMAKE^/GNATMAKE^} (@command{gnatelim})
10766 Instructs @code{gnatelim} to use specific @command{gnatmake} instead of one
10767 available on the path.
10771 @code{gnatelim} sends its output to the standard output stream, and all the
10772 tracing and debug information is sent to the standard error stream.
10773 In order to produce a proper GNAT configuration file
10774 @file{gnat.adc}, redirection must be used:
10778 $ PIPE GNAT ELIM MAIN_PROG.ADB > GNAT.ADC
10781 $ gnatelim main_prog.adb > gnat.adc
10790 $ gnatelim main_prog.adb >> gnat.adc
10794 in order to append the @code{gnatelim} output to the existing contents of
10798 @node Correcting the List of Eliminate Pragmas
10799 @subsection Correcting the List of Eliminate Pragmas
10802 In some rare cases @code{gnatelim} may try to eliminate
10803 subprograms that are actually called in the program. In this case, the
10804 compiler will generate an error message of the form:
10807 file.adb:106:07: cannot call eliminated subprogram "My_Prog"
10811 You will need to manually remove the wrong @code{Eliminate} pragmas from
10812 the @file{gnat.adc} file. You should recompile your program
10813 from scratch after that, because you need a consistent @file{gnat.adc} file
10814 during the entire compilation.
10816 @node Making Your Executables Smaller
10817 @subsection Making Your Executables Smaller
10820 In order to get a smaller executable for your program you now have to
10821 recompile the program completely with the new @file{gnat.adc} file
10822 created by @code{gnatelim} in your current directory:
10825 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10829 (Use the @option{^-f^/FORCE_COMPILE^} option for @command{gnatmake} to
10830 recompile everything
10831 with the set of pragmas @code{Eliminate} that you have obtained with
10832 @command{gnatelim}).
10834 Be aware that the set of @code{Eliminate} pragmas is specific to each
10835 program. It is not recommended to merge sets of @code{Eliminate}
10836 pragmas created for different programs in one @file{gnat.adc} file.
10838 @node Summary of the gnatelim Usage Cycle
10839 @subsection Summary of the gnatelim Usage Cycle
10842 Here is a quick summary of the steps to be taken in order to reduce
10843 the size of your executables with @code{gnatelim}. You may use
10844 other GNAT options to control the optimization level,
10845 to produce the debugging information, to set search path, etc.
10849 Produce a bind file
10852 $ gnatmake ^-c main_prog^/ACTIONS=COMPILE MAIN_PROG^
10853 $ gnatbind main_prog
10857 Generate a list of @code{Eliminate} pragmas
10860 $ PIPE GNAT ELIM MAIN_PROG > GNAT.ADC
10863 $ gnatelim main_prog >@r{[}>@r{]} gnat.adc
10868 Recompile the application
10871 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10876 @node Reducing Size of Executables with unused subprogram/data elimination
10877 @section Reducing Size of Executables with Unused Subprogram/Data Elimination
10878 @findex unused subprogram/data elimination
10881 This section describes how you can eliminate unused subprograms and data from
10882 your executable just by setting options at compilation time.
10885 * About unused subprogram/data elimination::
10886 * Compilation options::
10887 * Example of unused subprogram/data elimination::
10890 @node About unused subprogram/data elimination
10891 @subsection About unused subprogram/data elimination
10894 By default, an executable contains all code and data of its composing objects
10895 (directly linked or coming from statically linked libraries), even data or code
10896 never used by this executable.
10898 This feature will allow you to eliminate such unused code from your
10899 executable, making it smaller (in disk and in memory).
10901 This functionality is available on all Linux platforms except for the IA-64
10902 architecture and on all cross platforms using the ELF binary file format.
10903 In both cases GNU binutils version 2.16 or later are required to enable it.
10905 @node Compilation options
10906 @subsection Compilation options
10909 The operation of eliminating the unused code and data from the final executable
10910 is directly performed by the linker.
10912 In order to do this, it has to work with objects compiled with the
10914 @option{-ffunction-sections} @option{-fdata-sections}.
10915 @cindex @option{-ffunction-sections} (@command{gcc})
10916 @cindex @option{-fdata-sections} (@command{gcc})
10917 These options are usable with C and Ada files.
10918 They will place respectively each
10919 function or data in a separate section in the resulting object file.
10921 Once the objects and static libraries are created with these options, the
10922 linker can perform the dead code elimination. You can do this by setting
10923 the @option{-Wl,--gc-sections} option to gcc command or in the
10924 @option{-largs} section of @command{gnatmake}. This will perform a
10925 garbage collection of code and data never referenced.
10927 If the linker performs a partial link (@option{-r} ld linker option), then you
10928 will need to provide one or several entry point using the
10929 @option{-e} / @option{--entry} ld option.
10931 Note that objects compiled without the @option{-ffunction-sections} and
10932 @option{-fdata-sections} options can still be linked with the executable.
10933 However, no dead code elimination will be performed on those objects (they will
10936 The GNAT static library is now compiled with -ffunction-sections and
10937 -fdata-sections on some platforms. This allows you to eliminate the unused code
10938 and data of the GNAT library from your executable.
10940 @node Example of unused subprogram/data elimination
10941 @subsection Example of unused subprogram/data elimination
10944 Here is a simple example:
10946 @smallexample @c ada
10955 Used_Data : Integer;
10956 Unused_Data : Integer;
10958 procedure Used (Data : Integer);
10959 procedure Unused (Data : Integer);
10962 package body Aux is
10963 procedure Used (Data : Integer) is
10968 procedure Unused (Data : Integer) is
10970 Unused_Data := Data;
10976 @code{Unused} and @code{Unused_Data} are never referenced in this code
10977 excerpt, and hence they may be safely removed from the final executable.
10982 $ nm test | grep used
10983 020015f0 T aux__unused
10984 02005d88 B aux__unused_data
10985 020015cc T aux__used
10986 02005d84 B aux__used_data
10988 $ gnatmake test -cargs -fdata-sections -ffunction-sections \
10989 -largs -Wl,--gc-sections
10991 $ nm test | grep used
10992 02005350 T aux__used
10993 0201ffe0 B aux__used_data
10997 It can be observed that the procedure @code{Unused} and the object
10998 @code{Unused_Data} are removed by the linker when using the
10999 appropriate options.
11001 @c ********************************
11002 @node Renaming Files Using gnatchop
11003 @chapter Renaming Files Using @code{gnatchop}
11007 This chapter discusses how to handle files with multiple units by using
11008 the @code{gnatchop} utility. This utility is also useful in renaming
11009 files to meet the standard GNAT default file naming conventions.
11012 * Handling Files with Multiple Units::
11013 * Operating gnatchop in Compilation Mode::
11014 * Command Line for gnatchop::
11015 * Switches for gnatchop::
11016 * Examples of gnatchop Usage::
11019 @node Handling Files with Multiple Units
11020 @section Handling Files with Multiple Units
11023 The basic compilation model of GNAT requires that a file submitted to the
11024 compiler have only one unit and there be a strict correspondence
11025 between the file name and the unit name.
11027 The @code{gnatchop} utility allows both of these rules to be relaxed,
11028 allowing GNAT to process files which contain multiple compilation units
11029 and files with arbitrary file names. @code{gnatchop}
11030 reads the specified file and generates one or more output files,
11031 containing one unit per file. The unit and the file name correspond,
11032 as required by GNAT.
11034 If you want to permanently restructure a set of ``foreign'' files so that
11035 they match the GNAT rules, and do the remaining development using the
11036 GNAT structure, you can simply use @command{gnatchop} once, generate the
11037 new set of files and work with them from that point on.
11039 Alternatively, if you want to keep your files in the ``foreign'' format,
11040 perhaps to maintain compatibility with some other Ada compilation
11041 system, you can set up a procedure where you use @command{gnatchop} each
11042 time you compile, regarding the source files that it writes as temporary
11043 files that you throw away.
11045 Note that if your file containing multiple units starts with a byte order
11046 mark (BOM) specifying UTF-8 encoding, then the files generated by gnatchop
11047 will each start with a copy of this BOM, meaning that they can be compiled
11048 automatically in UTF-8 mode without needing to specify an explicit encoding.
11050 @node Operating gnatchop in Compilation Mode
11051 @section Operating gnatchop in Compilation Mode
11054 The basic function of @code{gnatchop} is to take a file with multiple units
11055 and split it into separate files. The boundary between files is reasonably
11056 clear, except for the issue of comments and pragmas. In default mode, the
11057 rule is that any pragmas between units belong to the previous unit, except
11058 that configuration pragmas always belong to the following unit. Any comments
11059 belong to the following unit. These rules
11060 almost always result in the right choice of
11061 the split point without needing to mark it explicitly and most users will
11062 find this default to be what they want. In this default mode it is incorrect to
11063 submit a file containing only configuration pragmas, or one that ends in
11064 configuration pragmas, to @code{gnatchop}.
11066 However, using a special option to activate ``compilation mode'',
11068 can perform another function, which is to provide exactly the semantics
11069 required by the RM for handling of configuration pragmas in a compilation.
11070 In the absence of configuration pragmas (at the main file level), this
11071 option has no effect, but it causes such configuration pragmas to be handled
11072 in a quite different manner.
11074 First, in compilation mode, if @code{gnatchop} is given a file that consists of
11075 only configuration pragmas, then this file is appended to the
11076 @file{gnat.adc} file in the current directory. This behavior provides
11077 the required behavior described in the RM for the actions to be taken
11078 on submitting such a file to the compiler, namely that these pragmas
11079 should apply to all subsequent compilations in the same compilation
11080 environment. Using GNAT, the current directory, possibly containing a
11081 @file{gnat.adc} file is the representation
11082 of a compilation environment. For more information on the
11083 @file{gnat.adc} file, see @ref{Handling of Configuration Pragmas}.
11085 Second, in compilation mode, if @code{gnatchop}
11086 is given a file that starts with
11087 configuration pragmas, and contains one or more units, then these
11088 configuration pragmas are prepended to each of the chopped files. This
11089 behavior provides the required behavior described in the RM for the
11090 actions to be taken on compiling such a file, namely that the pragmas
11091 apply to all units in the compilation, but not to subsequently compiled
11094 Finally, if configuration pragmas appear between units, they are appended
11095 to the previous unit. This results in the previous unit being illegal,
11096 since the compiler does not accept configuration pragmas that follow
11097 a unit. This provides the required RM behavior that forbids configuration
11098 pragmas other than those preceding the first compilation unit of a
11101 For most purposes, @code{gnatchop} will be used in default mode. The
11102 compilation mode described above is used only if you need exactly
11103 accurate behavior with respect to compilations, and you have files
11104 that contain multiple units and configuration pragmas. In this
11105 circumstance the use of @code{gnatchop} with the compilation mode
11106 switch provides the required behavior, and is for example the mode
11107 in which GNAT processes the ACVC tests.
11109 @node Command Line for gnatchop
11110 @section Command Line for @code{gnatchop}
11113 The @code{gnatchop} command has the form:
11116 $ gnatchop switches @var{file name} @r{[}@var{file name} @dots{}@r{]}
11121 The only required argument is the file name of the file to be chopped.
11122 There are no restrictions on the form of this file name. The file itself
11123 contains one or more Ada units, in normal GNAT format, concatenated
11124 together. As shown, more than one file may be presented to be chopped.
11126 When run in default mode, @code{gnatchop} generates one output file in
11127 the current directory for each unit in each of the files.
11129 @var{directory}, if specified, gives the name of the directory to which
11130 the output files will be written. If it is not specified, all files are
11131 written to the current directory.
11133 For example, given a
11134 file called @file{hellofiles} containing
11136 @smallexample @c ada
11141 with Text_IO; use Text_IO;
11144 Put_Line ("Hello");
11154 $ gnatchop ^hellofiles^HELLOFILES.^
11158 generates two files in the current directory, one called
11159 @file{hello.ads} containing the single line that is the procedure spec,
11160 and the other called @file{hello.adb} containing the remaining text. The
11161 original file is not affected. The generated files can be compiled in
11165 When gnatchop is invoked on a file that is empty or that contains only empty
11166 lines and/or comments, gnatchop will not fail, but will not produce any
11169 For example, given a
11170 file called @file{toto.txt} containing
11172 @smallexample @c ada
11184 $ gnatchop ^toto.txt^TOT.TXT^
11188 will not produce any new file and will result in the following warnings:
11191 toto.txt:1:01: warning: empty file, contains no compilation units
11192 no compilation units found
11193 no source files written
11196 @node Switches for gnatchop
11197 @section Switches for @code{gnatchop}
11200 @command{gnatchop} recognizes the following switches:
11206 @cindex @option{--version} @command{gnatchop}
11207 Display Copyright and version, then exit disregarding all other options.
11210 @cindex @option{--help} @command{gnatchop}
11211 If @option{--version} was not used, display usage, then exit disregarding
11214 @item ^-c^/COMPILATION^
11215 @cindex @option{^-c^/COMPILATION^} (@code{gnatchop})
11216 Causes @code{gnatchop} to operate in compilation mode, in which
11217 configuration pragmas are handled according to strict RM rules. See
11218 previous section for a full description of this mode.
11221 @item -gnat@var{xxx}
11222 This passes the given @option{-gnat@var{xxx}} switch to @code{gnat} which is
11223 used to parse the given file. Not all @var{xxx} options make sense,
11224 but for example, the use of @option{-gnati2} allows @code{gnatchop} to
11225 process a source file that uses Latin-2 coding for identifiers.
11229 Causes @code{gnatchop} to generate a brief help summary to the standard
11230 output file showing usage information.
11232 @item ^-k@var{mm}^/FILE_NAME_MAX_LENGTH=@var{mm}^
11233 @cindex @option{^-k^/FILE_NAME_MAX_LENGTH^} (@code{gnatchop})
11234 Limit generated file names to the specified number @code{mm}
11236 This is useful if the
11237 resulting set of files is required to be interoperable with systems
11238 which limit the length of file names.
11240 If no value is given, or
11241 if no @code{/FILE_NAME_MAX_LENGTH} qualifier is given,
11242 a default of 39, suitable for OpenVMS Alpha
11243 Systems, is assumed
11246 No space is allowed between the @option{-k} and the numeric value. The numeric
11247 value may be omitted in which case a default of @option{-k8},
11249 with DOS-like file systems, is used. If no @option{-k} switch
11251 there is no limit on the length of file names.
11254 @item ^-p^/PRESERVE^
11255 @cindex @option{^-p^/PRESERVE^} (@code{gnatchop})
11256 Causes the file ^modification^creation^ time stamp of the input file to be
11257 preserved and used for the time stamp of the output file(s). This may be
11258 useful for preserving coherency of time stamps in an environment where
11259 @code{gnatchop} is used as part of a standard build process.
11262 @cindex @option{^-q^/QUIET^} (@code{gnatchop})
11263 Causes output of informational messages indicating the set of generated
11264 files to be suppressed. Warnings and error messages are unaffected.
11266 @item ^-r^/REFERENCE^
11267 @cindex @option{^-r^/REFERENCE^} (@code{gnatchop})
11268 @findex Source_Reference
11269 Generate @code{Source_Reference} pragmas. Use this switch if the output
11270 files are regarded as temporary and development is to be done in terms
11271 of the original unchopped file. This switch causes
11272 @code{Source_Reference} pragmas to be inserted into each of the
11273 generated files to refers back to the original file name and line number.
11274 The result is that all error messages refer back to the original
11276 In addition, the debugging information placed into the object file (when
11277 the @option{^-g^/DEBUG^} switch of @command{gcc} or @command{gnatmake} is
11279 also refers back to this original file so that tools like profilers and
11280 debuggers will give information in terms of the original unchopped file.
11282 If the original file to be chopped itself contains
11283 a @code{Source_Reference}
11284 pragma referencing a third file, then gnatchop respects
11285 this pragma, and the generated @code{Source_Reference} pragmas
11286 in the chopped file refer to the original file, with appropriate
11287 line numbers. This is particularly useful when @code{gnatchop}
11288 is used in conjunction with @code{gnatprep} to compile files that
11289 contain preprocessing statements and multiple units.
11291 @item ^-v^/VERBOSE^
11292 @cindex @option{^-v^/VERBOSE^} (@code{gnatchop})
11293 Causes @code{gnatchop} to operate in verbose mode. The version
11294 number and copyright notice are output, as well as exact copies of
11295 the gnat1 commands spawned to obtain the chop control information.
11297 @item ^-w^/OVERWRITE^
11298 @cindex @option{^-w^/OVERWRITE^} (@code{gnatchop})
11299 Overwrite existing file names. Normally @code{gnatchop} regards it as a
11300 fatal error if there is already a file with the same name as a
11301 file it would otherwise output, in other words if the files to be
11302 chopped contain duplicated units. This switch bypasses this
11303 check, and causes all but the last instance of such duplicated
11304 units to be skipped.
11307 @item --GCC=@var{xxxx}
11308 @cindex @option{--GCC=} (@code{gnatchop})
11309 Specify the path of the GNAT parser to be used. When this switch is used,
11310 no attempt is made to add the prefix to the GNAT parser executable.
11314 @node Examples of gnatchop Usage
11315 @section Examples of @code{gnatchop} Usage
11319 @item gnatchop /OVERWRITE HELLO_S.ADA [PRERELEASE.FILES]
11322 @item gnatchop -w hello_s.ada prerelease/files
11325 Chops the source file @file{hello_s.ada}. The output files will be
11326 placed in the directory @file{^prerelease/files^[PRERELEASE.FILES]^},
11328 files with matching names in that directory (no files in the current
11329 directory are modified).
11331 @item gnatchop ^archive^ARCHIVE.^
11332 Chops the source file @file{^archive^ARCHIVE.^}
11333 into the current directory. One
11334 useful application of @code{gnatchop} is in sending sets of sources
11335 around, for example in email messages. The required sources are simply
11336 concatenated (for example, using a ^Unix @code{cat}^VMS @code{APPEND/NEW}^
11338 @command{gnatchop} is used at the other end to reconstitute the original
11341 @item gnatchop file1 file2 file3 direc
11342 Chops all units in files @file{file1}, @file{file2}, @file{file3}, placing
11343 the resulting files in the directory @file{direc}. Note that if any units
11344 occur more than once anywhere within this set of files, an error message
11345 is generated, and no files are written. To override this check, use the
11346 @option{^-w^/OVERWRITE^} switch,
11347 in which case the last occurrence in the last file will
11348 be the one that is output, and earlier duplicate occurrences for a given
11349 unit will be skipped.
11352 @node Configuration Pragmas
11353 @chapter Configuration Pragmas
11354 @cindex Configuration pragmas
11355 @cindex Pragmas, configuration
11358 Configuration pragmas include those pragmas described as
11359 such in the Ada Reference Manual, as well as
11360 implementation-dependent pragmas that are configuration pragmas.
11361 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
11362 for details on these additional GNAT-specific configuration pragmas.
11363 Most notably, the pragma @code{Source_File_Name}, which allows
11364 specifying non-default names for source files, is a configuration
11365 pragma. The following is a complete list of configuration pragmas
11366 recognized by GNAT:
11374 Assume_No_Invalid_Values
11379 Compile_Time_Warning
11381 Component_Alignment
11382 Convention_Identifier
11390 External_Name_Casing
11393 Float_Representation
11406 Priority_Specific_Dispatching
11409 Propagate_Exceptions
11412 Restricted_Run_Time
11414 Restrictions_Warnings
11417 Source_File_Name_Project
11420 Suppress_Exception_Locations
11421 Task_Dispatching_Policy
11427 Wide_Character_Encoding
11432 * Handling of Configuration Pragmas::
11433 * The Configuration Pragmas Files::
11436 @node Handling of Configuration Pragmas
11437 @section Handling of Configuration Pragmas
11439 Configuration pragmas may either appear at the start of a compilation
11440 unit, in which case they apply only to that unit, or they may apply to
11441 all compilations performed in a given compilation environment.
11443 GNAT also provides the @code{gnatchop} utility to provide an automatic
11444 way to handle configuration pragmas following the semantics for
11445 compilations (that is, files with multiple units), described in the RM.
11446 See @ref{Operating gnatchop in Compilation Mode} for details.
11447 However, for most purposes, it will be more convenient to edit the
11448 @file{gnat.adc} file that contains configuration pragmas directly,
11449 as described in the following section.
11451 @node The Configuration Pragmas Files
11452 @section The Configuration Pragmas Files
11453 @cindex @file{gnat.adc}
11456 In GNAT a compilation environment is defined by the current
11457 directory at the time that a compile command is given. This current
11458 directory is searched for a file whose name is @file{gnat.adc}. If
11459 this file is present, it is expected to contain one or more
11460 configuration pragmas that will be applied to the current compilation.
11461 However, if the switch @option{-gnatA} is used, @file{gnat.adc} is not
11464 Configuration pragmas may be entered into the @file{gnat.adc} file
11465 either by running @code{gnatchop} on a source file that consists only of
11466 configuration pragmas, or more conveniently by
11467 direct editing of the @file{gnat.adc} file, which is a standard format
11470 In addition to @file{gnat.adc}, additional files containing configuration
11471 pragmas may be applied to the current compilation using the switch
11472 @option{-gnatec}@var{path}. @var{path} must designate an existing file that
11473 contains only configuration pragmas. These configuration pragmas are
11474 in addition to those found in @file{gnat.adc} (provided @file{gnat.adc}
11475 is present and switch @option{-gnatA} is not used).
11477 It is allowed to specify several switches @option{-gnatec}, all of which
11478 will be taken into account.
11480 If you are using project file, a separate mechanism is provided using
11481 project attributes, see @ref{Specifying Configuration Pragmas} for more
11485 Of special interest to GNAT OpenVMS Alpha is the following
11486 configuration pragma:
11488 @smallexample @c ada
11490 pragma Extend_System (Aux_DEC);
11495 In the presence of this pragma, GNAT adds to the definition of the
11496 predefined package SYSTEM all the additional types and subprograms that are
11497 defined in HP Ada. See @ref{Compatibility with HP Ada} for details.
11500 @node Handling Arbitrary File Naming Conventions Using gnatname
11501 @chapter Handling Arbitrary File Naming Conventions Using @code{gnatname}
11502 @cindex Arbitrary File Naming Conventions
11505 * Arbitrary File Naming Conventions::
11506 * Running gnatname::
11507 * Switches for gnatname::
11508 * Examples of gnatname Usage::
11511 @node Arbitrary File Naming Conventions
11512 @section Arbitrary File Naming Conventions
11515 The GNAT compiler must be able to know the source file name of a compilation
11516 unit. When using the standard GNAT default file naming conventions
11517 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
11518 does not need additional information.
11521 When the source file names do not follow the standard GNAT default file naming
11522 conventions, the GNAT compiler must be given additional information through
11523 a configuration pragmas file (@pxref{Configuration Pragmas})
11525 When the non-standard file naming conventions are well-defined,
11526 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
11527 (@pxref{Alternative File Naming Schemes}) may be sufficient. However,
11528 if the file naming conventions are irregular or arbitrary, a number
11529 of pragma @code{Source_File_Name} for individual compilation units
11531 To help maintain the correspondence between compilation unit names and
11532 source file names within the compiler,
11533 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
11536 @node Running gnatname
11537 @section Running @code{gnatname}
11540 The usual form of the @code{gnatname} command is
11543 $ gnatname @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}
11544 @r{[}--and @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}@r{]}
11548 All of the arguments are optional. If invoked without any argument,
11549 @code{gnatname} will display its usage.
11552 When used with at least one naming pattern, @code{gnatname} will attempt to
11553 find all the compilation units in files that follow at least one of the
11554 naming patterns. To find these compilation units,
11555 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
11559 One or several Naming Patterns may be given as arguments to @code{gnatname}.
11560 Each Naming Pattern is enclosed between double quotes.
11561 A Naming Pattern is a regular expression similar to the wildcard patterns
11562 used in file names by the Unix shells or the DOS prompt.
11565 @code{gnatname} may be called with several sections of directories/patterns.
11566 Sections are separated by switch @code{--and}. In each section, there must be
11567 at least one pattern. If no directory is specified in a section, the current
11568 directory (or the project directory is @code{-P} is used) is implied.
11569 The options other that the directory switches and the patterns apply globally
11570 even if they are in different sections.
11573 Examples of Naming Patterns are
11582 For a more complete description of the syntax of Naming Patterns,
11583 see the second kind of regular expressions described in @file{g-regexp.ads}
11584 (the ``Glob'' regular expressions).
11587 When invoked with no switch @code{-P}, @code{gnatname} will create a
11588 configuration pragmas file @file{gnat.adc} in the current working directory,
11589 with pragmas @code{Source_File_Name} for each file that contains a valid Ada
11592 @node Switches for gnatname
11593 @section Switches for @code{gnatname}
11596 Switches for @code{gnatname} must precede any specified Naming Pattern.
11599 You may specify any of the following switches to @code{gnatname}:
11605 @cindex @option{--version} @command{gnatname}
11606 Display Copyright and version, then exit disregarding all other options.
11609 @cindex @option{--help} @command{gnatname}
11610 If @option{--version} was not used, display usage, then exit disregarding
11614 Start another section of directories/patterns.
11616 @item ^-c^/CONFIG_FILE=^@file{file}
11617 @cindex @option{^-c^/CONFIG_FILE^} (@code{gnatname})
11618 Create a configuration pragmas file @file{file} (instead of the default
11621 There may be zero, one or more space between @option{-c} and
11624 @file{file} may include directory information. @file{file} must be
11625 writable. There may be only one switch @option{^-c^/CONFIG_FILE^}.
11626 When a switch @option{^-c^/CONFIG_FILE^} is
11627 specified, no switch @option{^-P^/PROJECT_FILE^} may be specified (see below).
11629 @item ^-d^/SOURCE_DIRS=^@file{dir}
11630 @cindex @option{^-d^/SOURCE_DIRS^} (@code{gnatname})
11631 Look for source files in directory @file{dir}. There may be zero, one or more
11632 spaces between @option{^-d^/SOURCE_DIRS=^} and @file{dir}.
11633 When a switch @option{^-d^/SOURCE_DIRS^}
11634 is specified, the current working directory will not be searched for source
11635 files, unless it is explicitly specified with a @option{^-d^/SOURCE_DIRS^}
11636 or @option{^-D^/DIR_FILES^} switch.
11637 Several switches @option{^-d^/SOURCE_DIRS^} may be specified.
11638 If @file{dir} is a relative path, it is relative to the directory of
11639 the configuration pragmas file specified with switch
11640 @option{^-c^/CONFIG_FILE^},
11641 or to the directory of the project file specified with switch
11642 @option{^-P^/PROJECT_FILE^} or,
11643 if neither switch @option{^-c^/CONFIG_FILE^}
11644 nor switch @option{^-P^/PROJECT_FILE^} are specified, it is relative to the
11645 current working directory. The directory
11646 specified with switch @option{^-d^/SOURCE_DIRS^} must exist and be readable.
11648 @item ^-D^/DIRS_FILE=^@file{file}
11649 @cindex @option{^-D^/DIRS_FILE^} (@code{gnatname})
11650 Look for source files in all directories listed in text file @file{file}.
11651 There may be zero, one or more spaces between @option{^-D^/DIRS_FILE=^}
11653 @file{file} must be an existing, readable text file.
11654 Each nonempty line in @file{file} must be a directory.
11655 Specifying switch @option{^-D^/DIRS_FILE^} is equivalent to specifying as many
11656 switches @option{^-d^/SOURCE_DIRS^} as there are nonempty lines in
11659 @item ^-f^/FOREIGN_PATTERN=^@file{pattern}
11660 @cindex @option{^-f^/FOREIGN_PATTERN^} (@code{gnatname})
11661 Foreign patterns. Using this switch, it is possible to add sources of languages
11662 other than Ada to the list of sources of a project file.
11663 It is only useful if a ^-P^/PROJECT_FILE^ switch is used.
11666 gnatname ^-Pprj -f"*.c"^/PROJECT_FILE=PRJ /FOREIGN_PATTERN=*.C^ "*.ada"
11669 will look for Ada units in all files with the @file{.ada} extension,
11670 and will add to the list of file for project @file{prj.gpr} the C files
11671 with extension @file{.^c^C^}.
11674 @cindex @option{^-h^/HELP^} (@code{gnatname})
11675 Output usage (help) information. The output is written to @file{stdout}.
11677 @item ^-P^/PROJECT_FILE=^@file{proj}
11678 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatname})
11679 Create or update project file @file{proj}. There may be zero, one or more space
11680 between @option{-P} and @file{proj}. @file{proj} may include directory
11681 information. @file{proj} must be writable.
11682 There may be only one switch @option{^-P^/PROJECT_FILE^}.
11683 When a switch @option{^-P^/PROJECT_FILE^} is specified,
11684 no switch @option{^-c^/CONFIG_FILE^} may be specified.
11686 @item ^-v^/VERBOSE^
11687 @cindex @option{^-v^/VERBOSE^} (@code{gnatname})
11688 Verbose mode. Output detailed explanation of behavior to @file{stdout}.
11689 This includes name of the file written, the name of the directories to search
11690 and, for each file in those directories whose name matches at least one of
11691 the Naming Patterns, an indication of whether the file contains a unit,
11692 and if so the name of the unit.
11694 @item ^-v -v^/VERBOSE /VERBOSE^
11695 @cindex @option{^-v -v^/VERBOSE /VERBOSE^} (@code{gnatname})
11696 Very Verbose mode. In addition to the output produced in verbose mode,
11697 for each file in the searched directories whose name matches none of
11698 the Naming Patterns, an indication is given that there is no match.
11700 @item ^-x^/EXCLUDED_PATTERN=^@file{pattern}
11701 @cindex @option{^-x^/EXCLUDED_PATTERN^} (@code{gnatname})
11702 Excluded patterns. Using this switch, it is possible to exclude some files
11703 that would match the name patterns. For example,
11705 gnatname ^-x "*_nt.ada"^/EXCLUDED_PATTERN=*_nt.ada^ "*.ada"
11708 will look for Ada units in all files with the @file{.ada} extension,
11709 except those whose names end with @file{_nt.ada}.
11713 @node Examples of gnatname Usage
11714 @section Examples of @code{gnatname} Usage
11718 $ gnatname /CONFIG_FILE=[HOME.ME]NAMES.ADC /SOURCE_DIRS=SOURCES "[a-z]*.ada*"
11724 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
11729 In this example, the directory @file{^/home/me^[HOME.ME]^} must already exist
11730 and be writable. In addition, the directory
11731 @file{^/home/me/sources^[HOME.ME.SOURCES]^} (specified by
11732 @option{^-d sources^/SOURCE_DIRS=SOURCES^}) must exist and be readable.
11735 Note the optional spaces after @option{-c} and @option{-d}.
11740 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
11741 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
11744 $ gnatname /PROJECT_FILE=[HOME.ME]PROJ
11745 /EXCLUDED_PATTERN=*_nt_body.ada
11746 /SOURCE_DIRS=(SOURCES,[SOURCES.PLUS])
11747 /DIRS_FILE=COMMON_DIRS.TXT "body_*" "spec_*"
11751 Note that several switches @option{^-d^/SOURCE_DIRS^} may be used,
11752 even in conjunction with one or several switches
11753 @option{^-D^/DIRS_FILE^}. Several Naming Patterns and one excluded pattern
11754 are used in this example.
11756 @c *****************************************
11757 @c * G N A T P r o j e c t M a n a g e r *
11758 @c *****************************************
11759 @node GNAT Project Manager
11760 @chapter GNAT Project Manager
11764 * Examples of Project Files::
11765 * Project File Syntax::
11766 * Objects and Sources in Project Files::
11767 * Importing Projects::
11768 * Project Extension::
11769 * Project Hierarchy Extension::
11770 * External References in Project Files::
11771 * Packages in Project Files::
11772 * Variables from Imported Projects::
11774 * Library Projects::
11775 * Stand-alone Library Projects::
11776 * Switches Related to Project Files::
11777 * Tools Supporting Project Files::
11778 * An Extended Example::
11779 * Project File Complete Syntax::
11782 @c ****************
11783 @c * Introduction *
11784 @c ****************
11787 @section Introduction
11790 This chapter describes GNAT's @emph{Project Manager}, a facility that allows
11791 you to manage complex builds involving a number of source files, directories,
11792 and compilation options for different system configurations. In particular,
11793 project files allow you to specify:
11796 The directory or set of directories containing the source files, and/or the
11797 names of the specific source files themselves
11799 The directory in which the compiler's output
11800 (@file{ALI} files, object files, tree files) is to be placed
11802 The directory in which the executable programs is to be placed
11804 ^Switch^Switch^ settings for any of the project-enabled tools
11805 (@command{gnatmake}, compiler, binder, linker, @code{gnatls}, @code{gnatxref},
11806 @code{gnatfind}); you can apply these settings either globally or to individual
11809 The source files containing the main subprogram(s) to be built
11811 The source programming language(s) (currently Ada and/or C)
11813 Source file naming conventions; you can specify these either globally or for
11814 individual compilation units
11821 @node Project Files
11822 @subsection Project Files
11825 Project files are written in a syntax close to that of Ada, using familiar
11826 notions such as packages, context clauses, declarations, default values,
11827 assignments, and inheritance. Finally, project files can be built
11828 hierarchically from other project files, simplifying complex system
11829 integration and project reuse.
11831 A @dfn{project} is a specific set of values for various compilation properties.
11832 The settings for a given project are described by means of
11833 a @dfn{project file}, which is a text file written in an Ada-like syntax.
11834 Property values in project files are either strings or lists of strings.
11835 Properties that are not explicitly set receive default values. A project
11836 file may interrogate the values of @dfn{external variables} (user-defined
11837 command-line switches or environment variables), and it may specify property
11838 settings conditionally, based on the value of such variables.
11840 In simple cases, a project's source files depend only on other source files
11841 in the same project, or on the predefined libraries. (@emph{Dependence} is
11843 the Ada technical sense; as in one Ada unit @code{with}ing another.) However,
11844 the Project Manager also allows more sophisticated arrangements,
11845 where the source files in one project depend on source files in other
11849 One project can @emph{import} other projects containing needed source files.
11851 You can organize GNAT projects in a hierarchy: a @emph{child} project
11852 can extend a @emph{parent} project, inheriting the parent's source files and
11853 optionally overriding any of them with alternative versions
11857 More generally, the Project Manager lets you structure large development
11858 efforts into hierarchical subsystems, where build decisions are delegated
11859 to the subsystem level, and thus different compilation environments
11860 (^switch^switch^ settings) used for different subsystems.
11862 The Project Manager is invoked through the
11863 @option{^-P^/PROJECT_FILE=^@emph{projectfile}}
11864 switch to @command{gnatmake} or to the @command{^gnat^GNAT^} front driver.
11866 There may be zero, one or more spaces between @option{-P} and
11867 @option{@emph{projectfile}}.
11869 If you want to define (on the command line) an external variable that is
11870 queried by the project file, you must use the
11871 @option{^-X^/EXTERNAL_REFERENCE=^@emph{vbl}=@emph{value}} switch.
11872 The Project Manager parses and interprets the project file, and drives the
11873 invoked tool based on the project settings.
11875 The Project Manager supports a wide range of development strategies,
11876 for systems of all sizes. Here are some typical practices that are
11880 Using a common set of source files, but generating object files in different
11881 directories via different ^switch^switch^ settings
11883 Using a mostly-shared set of source files, but with different versions of
11888 The destination of an executable can be controlled inside a project file
11889 using the @option{^-o^-o^}
11891 In the absence of such a ^switch^switch^ either inside
11892 the project file or on the command line, any executable files generated by
11893 @command{gnatmake} are placed in the directory @code{Exec_Dir} specified
11894 in the project file. If no @code{Exec_Dir} is specified, they will be placed
11895 in the object directory of the project.
11897 You can use project files to achieve some of the effects of a source
11898 versioning system (for example, defining separate projects for
11899 the different sets of sources that comprise different releases) but the
11900 Project Manager is independent of any source configuration management tools
11901 that might be used by the developers.
11903 The next section introduces the main features of GNAT's project facility
11904 through a sequence of examples; subsequent sections will present the syntax
11905 and semantics in more detail. A more formal description of the project
11906 facility appears in @ref{Project File Reference,,, gnat_rm, GNAT
11909 @c *****************************
11910 @c * Examples of Project Files *
11911 @c *****************************
11913 @node Examples of Project Files
11914 @section Examples of Project Files
11916 This section illustrates some of the typical uses of project files and
11917 explains their basic structure and behavior.
11920 * Common Sources with Different ^Switches^Switches^ and Directories::
11921 * Using External Variables::
11922 * Importing Other Projects::
11923 * Extending a Project::
11926 @node Common Sources with Different ^Switches^Switches^ and Directories
11927 @subsection Common Sources with Different ^Switches^Switches^ and Directories
11931 * Specifying the Object Directory::
11932 * Specifying the Exec Directory::
11933 * Project File Packages::
11934 * Specifying ^Switch^Switch^ Settings::
11935 * Main Subprograms::
11936 * Executable File Names::
11937 * Source File Naming Conventions::
11938 * Source Language(s)::
11942 Suppose that the Ada source files @file{pack.ads}, @file{pack.adb}, and
11943 @file{proc.adb} are in the @file{/common} directory. The file
11944 @file{proc.adb} contains an Ada main subprogram @code{Proc} that @code{with}s
11945 package @code{Pack}. We want to compile these source files under two sets
11946 of ^switches^switches^:
11949 When debugging, we want to pass the @option{-g} switch to @command{gnatmake},
11950 and the @option{^-gnata^-gnata^},
11951 @option{^-gnato^-gnato^},
11952 and @option{^-gnatE^-gnatE^} switches to the
11953 compiler; the compiler's output is to appear in @file{/common/debug}
11955 When preparing a release version, we want to pass the @option{^-O2^O2^} switch
11956 to the compiler; the compiler's output is to appear in @file{/common/release}
11960 The GNAT project files shown below, respectively @file{debug.gpr} and
11961 @file{release.gpr} in the @file{/common} directory, achieve these effects.
11974 ^/common/debug^[COMMON.DEBUG]^
11979 ^/common/release^[COMMON.RELEASE]^
11984 Here are the corresponding project files:
11986 @smallexample @c projectfile
11989 for Object_Dir use "debug";
11990 for Main use ("proc");
11993 for ^Default_Switches^Default_Switches^ ("Ada")
11995 for Executable ("proc.adb") use "proc1";
12000 package Compiler is
12001 for ^Default_Switches^Default_Switches^ ("Ada")
12002 use ("-fstack-check",
12005 "^-gnatE^-gnatE^");
12011 @smallexample @c projectfile
12014 for Object_Dir use "release";
12015 for Exec_Dir use ".";
12016 for Main use ("proc");
12018 package Compiler is
12019 for ^Default_Switches^Default_Switches^ ("Ada")
12027 The name of the project defined by @file{debug.gpr} is @code{"Debug"} (case
12028 insensitive), and analogously the project defined by @file{release.gpr} is
12029 @code{"Release"}. For consistency the file should have the same name as the
12030 project, and the project file's extension should be @code{"gpr"}. These
12031 conventions are not required, but a warning is issued if they are not followed.
12033 If the current directory is @file{^/temp^[TEMP]^}, then the command
12035 gnatmake ^-P/common/debug.gpr^/PROJECT_FILE=[COMMON]DEBUG^
12039 generates object and ALI files in @file{^/common/debug^[COMMON.DEBUG]^},
12040 as well as the @code{^proc1^PROC1.EXE^} executable,
12041 using the ^switch^switch^ settings defined in the project file.
12043 Likewise, the command
12045 gnatmake ^-P/common/release.gpr^/PROJECT_FILE=[COMMON]RELEASE^
12049 generates object and ALI files in @file{^/common/release^[COMMON.RELEASE]^},
12050 and the @code{^proc^PROC.EXE^}
12051 executable in @file{^/common^[COMMON]^},
12052 using the ^switch^switch^ settings from the project file.
12055 @unnumberedsubsubsec Source Files
12058 If a project file does not explicitly specify a set of source directories or
12059 a set of source files, then by default the project's source files are the
12060 Ada source files in the project file directory. Thus @file{pack.ads},
12061 @file{pack.adb}, and @file{proc.adb} are the source files for both projects.
12063 @node Specifying the Object Directory
12064 @unnumberedsubsubsec Specifying the Object Directory
12067 Several project properties are modeled by Ada-style @emph{attributes};
12068 a property is defined by supplying the equivalent of an Ada attribute
12069 definition clause in the project file.
12070 A project's object directory is another such a property; the corresponding
12071 attribute is @code{Object_Dir}, and its value is also a string expression,
12072 specified either as absolute or relative. In the later case,
12073 it is relative to the project file directory. Thus the compiler's
12074 output is directed to @file{^/common/debug^[COMMON.DEBUG]^}
12075 (for the @code{Debug} project)
12076 and to @file{^/common/release^[COMMON.RELEASE]^}
12077 (for the @code{Release} project).
12078 If @code{Object_Dir} is not specified, then the default is the project file
12081 @node Specifying the Exec Directory
12082 @unnumberedsubsubsec Specifying the Exec Directory
12085 A project's exec directory is another property; the corresponding
12086 attribute is @code{Exec_Dir}, and its value is also a string expression,
12087 either specified as relative or absolute. If @code{Exec_Dir} is not specified,
12088 then the default is the object directory (which may also be the project file
12089 directory if attribute @code{Object_Dir} is not specified). Thus the executable
12090 is placed in @file{^/common/debug^[COMMON.DEBUG]^}
12091 for the @code{Debug} project (attribute @code{Exec_Dir} not specified)
12092 and in @file{^/common^[COMMON]^} for the @code{Release} project.
12094 @node Project File Packages
12095 @unnumberedsubsubsec Project File Packages
12098 A GNAT tool that is integrated with the Project Manager is modeled by a
12099 corresponding package in the project file. In the example above,
12100 The @code{Debug} project defines the packages @code{Builder}
12101 (for @command{gnatmake}) and @code{Compiler};
12102 the @code{Release} project defines only the @code{Compiler} package.
12104 The Ada-like package syntax is not to be taken literally. Although packages in
12105 project files bear a surface resemblance to packages in Ada source code, the
12106 notation is simply a way to convey a grouping of properties for a named
12107 entity. Indeed, the package names permitted in project files are restricted
12108 to a predefined set, corresponding to the project-aware tools, and the contents
12109 of packages are limited to a small set of constructs.
12110 The packages in the example above contain attribute definitions.
12112 @node Specifying ^Switch^Switch^ Settings
12113 @unnumberedsubsubsec Specifying ^Switch^Switch^ Settings
12116 ^Switch^Switch^ settings for a project-aware tool can be specified through
12117 attributes in the package that corresponds to the tool.
12118 The example above illustrates one of the relevant attributes,
12119 @code{^Default_Switches^Default_Switches^}, which is defined in packages
12120 in both project files.
12121 Unlike simple attributes like @code{Source_Dirs},
12122 @code{^Default_Switches^Default_Switches^} is
12123 known as an @emph{associative array}. When you define this attribute, you must
12124 supply an ``index'' (a literal string), and the effect of the attribute
12125 definition is to set the value of the array at the specified index.
12126 For the @code{^Default_Switches^Default_Switches^} attribute,
12127 the index is a programming language (in our case, Ada),
12128 and the value specified (after @code{use}) must be a list
12129 of string expressions.
12131 The attributes permitted in project files are restricted to a predefined set.
12132 Some may appear at project level, others in packages.
12133 For any attribute that is an associative array, the index must always be a
12134 literal string, but the restrictions on this string (e.g., a file name or a
12135 language name) depend on the individual attribute.
12136 Also depending on the attribute, its specified value will need to be either a
12137 string or a string list.
12139 In the @code{Debug} project, we set the switches for two tools,
12140 @command{gnatmake} and the compiler, and thus we include the two corresponding
12141 packages; each package defines the @code{^Default_Switches^Default_Switches^}
12142 attribute with index @code{"Ada"}.
12143 Note that the package corresponding to
12144 @command{gnatmake} is named @code{Builder}. The @code{Release} project is
12145 similar, but only includes the @code{Compiler} package.
12147 In project @code{Debug} above, the ^switches^switches^ starting with
12148 @option{-gnat} that are specified in package @code{Compiler}
12149 could have been placed in package @code{Builder}, since @command{gnatmake}
12150 transmits all such ^switches^switches^ to the compiler.
12152 @node Main Subprograms
12153 @unnumberedsubsubsec Main Subprograms
12156 One of the specifiable properties of a project is a list of files that contain
12157 main subprograms. This property is captured in the @code{Main} attribute,
12158 whose value is a list of strings. If a project defines the @code{Main}
12159 attribute, it is not necessary to identify the main subprogram(s) when
12160 invoking @command{gnatmake} (@pxref{gnatmake and Project Files}).
12162 @node Executable File Names
12163 @unnumberedsubsubsec Executable File Names
12166 By default, the executable file name corresponding to a main source is
12167 deduced from the main source file name. Through the attributes
12168 @code{Executable} and @code{Executable_Suffix} of package @code{Builder},
12169 it is possible to change this default.
12170 In project @code{Debug} above, the executable file name
12171 for main source @file{^proc.adb^PROC.ADB^} is
12172 @file{^proc1^PROC1.EXE^}.
12173 Attribute @code{Executable_Suffix}, when specified, may change the suffix
12174 of the executable files, when no attribute @code{Executable} applies:
12175 its value replace the platform-specific executable suffix.
12176 Attributes @code{Executable} and @code{Executable_Suffix} are the only ways to
12177 specify a non-default executable file name when several mains are built at once
12178 in a single @command{gnatmake} command.
12180 @node Source File Naming Conventions
12181 @unnumberedsubsubsec Source File Naming Conventions
12184 Since the project files above do not specify any source file naming
12185 conventions, the GNAT defaults are used. The mechanism for defining source
12186 file naming conventions -- a package named @code{Naming} --
12187 is described below (@pxref{Naming Schemes}).
12189 @node Source Language(s)
12190 @unnumberedsubsubsec Source Language(s)
12193 Since the project files do not specify a @code{Languages} attribute, by
12194 default the GNAT tools assume that the language of the project file is Ada.
12195 More generally, a project can comprise source files
12196 in Ada, C, and/or other languages.
12198 @node Using External Variables
12199 @subsection Using External Variables
12202 Instead of supplying different project files for debug and release, we can
12203 define a single project file that queries an external variable (set either
12204 on the command line or via an ^environment variable^logical name^) in order to
12205 conditionally define the appropriate settings. Again, assume that the
12206 source files @file{pack.ads}, @file{pack.adb}, and @file{proc.adb} are
12207 located in directory @file{^/common^[COMMON]^}. The following project file,
12208 @file{build.gpr}, queries the external variable named @code{STYLE} and
12209 defines an object directory and ^switch^switch^ settings based on whether
12210 the value is @code{"deb"} (debug) or @code{"rel"} (release), and where
12211 the default is @code{"deb"}.
12213 @smallexample @c projectfile
12216 for Main use ("proc");
12218 type Style_Type is ("deb", "rel");
12219 Style : Style_Type := external ("STYLE", "deb");
12223 for Object_Dir use "debug";
12226 for Object_Dir use "release";
12227 for Exec_Dir use ".";
12236 for ^Default_Switches^Default_Switches^ ("Ada")
12238 for Executable ("proc") use "proc1";
12247 package Compiler is
12251 for ^Default_Switches^Default_Switches^ ("Ada")
12252 use ("^-gnata^-gnata^",
12254 "^-gnatE^-gnatE^");
12257 for ^Default_Switches^Default_Switches^ ("Ada")
12268 @code{Style_Type} is an example of a @emph{string type}, which is the project
12269 file analog of an Ada enumeration type but whose components are string literals
12270 rather than identifiers. @code{Style} is declared as a variable of this type.
12272 The form @code{external("STYLE", "deb")} is known as an
12273 @emph{external reference}; its first argument is the name of an
12274 @emph{external variable}, and the second argument is a default value to be
12275 used if the external variable doesn't exist. You can define an external
12276 variable on the command line via the @option{^-X^/EXTERNAL_REFERENCE^} switch,
12277 or you can use ^an environment variable^a logical name^
12278 as an external variable.
12280 Each @code{case} construct is expanded by the Project Manager based on the
12281 value of @code{Style}. Thus the command
12284 gnatmake -P/common/build.gpr -XSTYLE=deb
12290 gnatmake /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=deb
12295 is equivalent to the @command{gnatmake} invocation using the project file
12296 @file{debug.gpr} in the earlier example. So is the command
12298 gnatmake ^-P/common/build.gpr^/PROJECT_FILE=[COMMON]BUILD.GPR^
12302 since @code{"deb"} is the default for @code{STYLE}.
12308 gnatmake -P/common/build.gpr -XSTYLE=rel
12314 GNAT MAKE /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=rel
12319 is equivalent to the @command{gnatmake} invocation using the project file
12320 @file{release.gpr} in the earlier example.
12322 @node Importing Other Projects
12323 @subsection Importing Other Projects
12324 @cindex @code{ADA_PROJECT_PATH}
12325 @cindex @code{GPR_PROJECT_PATH}
12328 A compilation unit in a source file in one project may depend on compilation
12329 units in source files in other projects. To compile this unit under
12330 control of a project file, the
12331 dependent project must @emph{import} the projects containing the needed source
12333 This effect is obtained using syntax similar to an Ada @code{with} clause,
12334 but where @code{with}ed entities are strings that denote project files.
12336 As an example, suppose that the two projects @code{GUI_Proj} and
12337 @code{Comm_Proj} are defined in the project files @file{gui_proj.gpr} and
12338 @file{comm_proj.gpr} in directories @file{^/gui^[GUI]^}
12339 and @file{^/comm^[COMM]^}, respectively.
12340 Suppose that the source files for @code{GUI_Proj} are
12341 @file{gui.ads} and @file{gui.adb}, and that the source files for
12342 @code{Comm_Proj} are @file{comm.ads} and @file{comm.adb}, where each set of
12343 files is located in its respective project file directory. Schematically:
12362 We want to develop an application in directory @file{^/app^[APP]^} that
12363 @code{with} the packages @code{GUI} and @code{Comm}, using the properties of
12364 the corresponding project files (e.g.@: the ^switch^switch^ settings
12365 and object directory).
12366 Skeletal code for a main procedure might be something like the following:
12368 @smallexample @c ada
12371 procedure App_Main is
12380 Here is a project file, @file{app_proj.gpr}, that achieves the desired
12383 @smallexample @c projectfile
12385 with "/gui/gui_proj", "/comm/comm_proj";
12386 project App_Proj is
12387 for Main use ("app_main");
12393 Building an executable is achieved through the command:
12395 gnatmake ^-P/app/app_proj^/PROJECT_FILE=[APP]APP_PROJ^
12398 which will generate the @code{^app_main^APP_MAIN.EXE^} executable
12399 in the directory where @file{app_proj.gpr} resides.
12401 If an imported project file uses the standard extension (@code{^gpr^GPR^}) then
12402 (as illustrated above) the @code{with} clause can omit the extension.
12404 Our example specified an absolute path for each imported project file.
12405 Alternatively, the directory name of an imported object can be omitted
12409 The imported project file is in the same directory as the importing project
12412 You have defined one or two ^environment variables^logical names^
12413 that includes the directory containing
12414 the needed project file. The syntax of @code{GPR_PROJECT_PATH} and
12415 @code{ADA_PROJECT_PATH} is the same as
12416 the syntax of @code{ADA_INCLUDE_PATH} and @code{ADA_OBJECTS_PATH}: a list of
12417 directory names separated by colons (semicolons on Windows).
12421 Thus, if we define @code{ADA_PROJECT_PATH} or @code{GPR_PROJECT_PATH}
12422 to include @file{^/gui^[GUI]^} and
12423 @file{^/comm^[COMM]^}, then our project file @file{app_proj.gpr} can be written
12426 @smallexample @c projectfile
12428 with "gui_proj", "comm_proj";
12429 project App_Proj is
12430 for Main use ("app_main");
12436 Importing other projects can create ambiguities.
12437 For example, the same unit might be present in different imported projects, or
12438 it might be present in both the importing project and in an imported project.
12439 Both of these conditions are errors. Note that in the current version of
12440 the Project Manager, it is illegal to have an ambiguous unit even if the
12441 unit is never referenced by the importing project. This restriction may be
12442 relaxed in a future release.
12444 @node Extending a Project
12445 @subsection Extending a Project
12448 In large software systems it is common to have multiple
12449 implementations of a common interface; in Ada terms, multiple versions of a
12450 package body for the same spec. For example, one implementation
12451 might be safe for use in tasking programs, while another might only be used
12452 in sequential applications. This can be modeled in GNAT using the concept
12453 of @emph{project extension}. If one project (the ``child'') @emph{extends}
12454 another project (the ``parent'') then by default all source files of the
12455 parent project are inherited by the child, but the child project can
12456 override any of the parent's source files with new versions, and can also
12457 add new files. This facility is the project analog of a type extension in
12458 Object-Oriented Programming. Project hierarchies are permitted (a child
12459 project may be the parent of yet another project), and a project that
12460 inherits one project can also import other projects.
12462 As an example, suppose that directory @file{^/seq^[SEQ]^} contains the project
12463 file @file{seq_proj.gpr} as well as the source files @file{pack.ads},
12464 @file{pack.adb}, and @file{proc.adb}:
12477 Note that the project file can simply be empty (that is, no attribute or
12478 package is defined):
12480 @smallexample @c projectfile
12482 project Seq_Proj is
12488 implying that its source files are all the Ada source files in the project
12491 Suppose we want to supply an alternate version of @file{pack.adb}, in
12492 directory @file{^/tasking^[TASKING]^}, but use the existing versions of
12493 @file{pack.ads} and @file{proc.adb}. We can define a project
12494 @code{Tasking_Proj} that inherits @code{Seq_Proj}:
12498 ^/tasking^[TASKING]^
12504 project Tasking_Proj extends "/seq/seq_proj" is
12510 The version of @file{pack.adb} used in a build depends on which project file
12513 Note that we could have obtained the desired behavior using project import
12514 rather than project inheritance; a @code{base} project would contain the
12515 sources for @file{pack.ads} and @file{proc.adb}, a sequential project would
12516 import @code{base} and add @file{pack.adb}, and likewise a tasking project
12517 would import @code{base} and add a different version of @file{pack.adb}. The
12518 choice depends on whether other sources in the original project need to be
12519 overridden. If they do, then project extension is necessary, otherwise,
12520 importing is sufficient.
12523 In a project file that extends another project file, it is possible to
12524 indicate that an inherited source is not part of the sources of the extending
12525 project. This is necessary sometimes when a package spec has been overloaded
12526 and no longer requires a body: in this case, it is necessary to indicate that
12527 the inherited body is not part of the sources of the project, otherwise there
12528 will be a compilation error when compiling the spec.
12530 For that purpose, the attribute @code{Excluded_Source_Files} is used.
12531 Its value is a string list: a list of file names. It is also possible to use
12532 attribute @code{Excluded_Source_List_File}. Its value is a single string:
12533 the file name of a text file containing a list of file names, one per line.
12535 @smallexample @c @projectfile
12536 project B extends "a" is
12537 for Source_Files use ("pkg.ads");
12538 -- New spec of Pkg does not need a completion
12539 for Excluded_Source_Files use ("pkg.adb");
12543 Attribute @code{Excluded_Source_Files} may also be used to check if a source
12544 is still needed: if it is possible to build using @command{gnatmake} when such
12545 a source is put in attribute @code{Excluded_Source_Files} of a project P, then
12546 it is possible to remove the source completely from a system that includes
12549 @c ***********************
12550 @c * Project File Syntax *
12551 @c ***********************
12553 @node Project File Syntax
12554 @section Project File Syntax
12558 * Qualified Projects::
12564 * Associative Array Attributes::
12565 * case Constructions::
12569 This section describes the structure of project files.
12571 A project may be an @emph{independent project}, entirely defined by a single
12572 project file. Any Ada source file in an independent project depends only
12573 on the predefined library and other Ada source files in the same project.
12576 A project may also @dfn{depend on} other projects, in either or both of
12577 the following ways:
12579 @item It may import any number of projects
12580 @item It may extend at most one other project
12584 The dependence relation is a directed acyclic graph (the subgraph reflecting
12585 the ``extends'' relation is a tree).
12587 A project's @dfn{immediate sources} are the source files directly defined by
12588 that project, either implicitly by residing in the project file's directory,
12589 or explicitly through any of the source-related attributes described below.
12590 More generally, a project @var{proj}'s @dfn{sources} are the immediate sources
12591 of @var{proj} together with the immediate sources (unless overridden) of any
12592 project on which @var{proj} depends (either directly or indirectly).
12595 @subsection Basic Syntax
12598 As seen in the earlier examples, project files have an Ada-like syntax.
12599 The minimal project file is:
12600 @smallexample @c projectfile
12609 The identifier @code{Empty} is the name of the project.
12610 This project name must be present after the reserved
12611 word @code{end} at the end of the project file, followed by a semi-colon.
12613 Any name in a project file, such as the project name or a variable name,
12614 has the same syntax as an Ada identifier.
12616 The reserved words of project files are the Ada 95 reserved words plus
12617 @code{extends}, @code{external}, and @code{project}. Note that the only Ada
12618 reserved words currently used in project file syntax are:
12654 Comments in project files have the same syntax as in Ada, two consecutive
12655 hyphens through the end of the line.
12657 @node Qualified Projects
12658 @subsection Qualified Projects
12661 Before the reserved @code{project}, there may be one or two "qualifiers", that
12662 is identifiers or other reserved words, to qualify the project.
12664 The current list of qualifiers is:
12668 @code{abstract}: qualify a project with no sources. A qualified abstract
12669 project must either have no declaration of attributes @code{Source_Dirs},
12670 @code{Source_Files}, @code{Languages} or @code{Source_List_File}, or one of
12671 @code{Source_Dirs}, @code{Source_Files}, or @code{Languages} must be declared
12672 as empty. If it extends another project, the project it extends must also be a
12673 qualified abstract project.
12676 @code{standard}: a standard project is a non library project with sources.
12679 @code{aggregate}: for future extension
12682 @code{aggregate library}: for future extension
12685 @code{library}: a library project must declare both attributes
12686 @code{Library_Name} and @code{Library_Dir}.
12689 @code{configuration}: a configuration project cannot be in a project tree.
12693 @subsection Packages
12696 A project file may contain @emph{packages}. The name of a package must be one
12697 of the identifiers from the following list. A package
12698 with a given name may only appear once in a project file. Package names are
12699 case insensitive. The following package names are legal:
12715 @code{Cross_Reference}
12719 @code{Pretty_Printer}
12729 @code{Language_Processing}
12733 In its simplest form, a package may be empty:
12735 @smallexample @c projectfile
12745 A package may contain @emph{attribute declarations},
12746 @emph{variable declarations} and @emph{case constructions}, as will be
12749 When there is ambiguity between a project name and a package name,
12750 the name always designates the project. To avoid possible confusion, it is
12751 always a good idea to avoid naming a project with one of the
12752 names allowed for packages or any name that starts with @code{gnat}.
12755 @subsection Expressions
12758 An @emph{expression} is either a @emph{string expression} or a
12759 @emph{string list expression}.
12761 A @emph{string expression} is either a @emph{simple string expression} or a
12762 @emph{compound string expression}.
12764 A @emph{simple string expression} is one of the following:
12766 @item A literal string; e.g.@: @code{"comm/my_proj.gpr"}
12767 @item A string-valued variable reference (@pxref{Variables})
12768 @item A string-valued attribute reference (@pxref{Attributes})
12769 @item An external reference (@pxref{External References in Project Files})
12773 A @emph{compound string expression} is a concatenation of string expressions,
12774 using the operator @code{"&"}
12776 Path & "/" & File_Name & ".ads"
12780 A @emph{string list expression} is either a
12781 @emph{simple string list expression} or a
12782 @emph{compound string list expression}.
12784 A @emph{simple string list expression} is one of the following:
12786 @item A parenthesized list of zero or more string expressions,
12787 separated by commas
12789 File_Names := (File_Name, "gnat.adc", File_Name & ".orig");
12792 @item A string list-valued variable reference
12793 @item A string list-valued attribute reference
12797 A @emph{compound string list expression} is the concatenation (using
12798 @code{"&"}) of a simple string list expression and an expression. Note that
12799 each term in a compound string list expression, except the first, may be
12800 either a string expression or a string list expression.
12802 @smallexample @c projectfile
12804 File_Name_List := () & File_Name; -- One string in this list
12805 Extended_File_Name_List := File_Name_List & (File_Name & ".orig");
12807 Big_List := File_Name_List & Extended_File_Name_List;
12808 -- Concatenation of two string lists: three strings
12809 Illegal_List := "gnat.adc" & Extended_File_Name_List;
12810 -- Illegal: must start with a string list
12815 @subsection String Types
12818 A @emph{string type declaration} introduces a discrete set of string literals.
12819 If a string variable is declared to have this type, its value
12820 is restricted to the given set of literals.
12822 Here is an example of a string type declaration:
12824 @smallexample @c projectfile
12825 type OS is ("NT", "nt", "Unix", "GNU/Linux", "other OS");
12829 Variables of a string type are called @emph{typed variables}; all other
12830 variables are called @emph{untyped variables}. Typed variables are
12831 particularly useful in @code{case} constructions, to support conditional
12832 attribute declarations.
12833 (@pxref{case Constructions}).
12835 The string literals in the list are case sensitive and must all be different.
12836 They may include any graphic characters allowed in Ada, including spaces.
12838 A string type may only be declared at the project level, not inside a package.
12840 A string type may be referenced by its name if it has been declared in the same
12841 project file, or by an expanded name whose prefix is the name of the project
12842 in which it is declared.
12845 @subsection Variables
12848 A variable may be declared at the project file level, or within a package.
12849 Here are some examples of variable declarations:
12851 @smallexample @c projectfile
12853 This_OS : OS := external ("OS"); -- a typed variable declaration
12854 That_OS := "GNU/Linux"; -- an untyped variable declaration
12859 The syntax of a @emph{typed variable declaration} is identical to the Ada
12860 syntax for an object declaration. By contrast, the syntax of an untyped
12861 variable declaration is identical to an Ada assignment statement. In fact,
12862 variable declarations in project files have some of the characteristics of
12863 an assignment, in that successive declarations for the same variable are
12864 allowed. Untyped variable declarations do establish the expected kind of the
12865 variable (string or string list), and successive declarations for it must
12866 respect the initial kind.
12869 A string variable declaration (typed or untyped) declares a variable
12870 whose value is a string. This variable may be used as a string expression.
12871 @smallexample @c projectfile
12872 File_Name := "readme.txt";
12873 Saved_File_Name := File_Name & ".saved";
12877 A string list variable declaration declares a variable whose value is a list
12878 of strings. The list may contain any number (zero or more) of strings.
12880 @smallexample @c projectfile
12882 List_With_One_Element := ("^-gnaty^-gnaty^");
12883 List_With_Two_Elements := List_With_One_Element & "^-gnatg^-gnatg^";
12884 Long_List := ("main.ada", "pack1_.ada", "pack1.ada", "pack2_.ada"
12885 "pack2.ada", "util_.ada", "util.ada");
12889 The same typed variable may not be declared more than once at project level,
12890 and it may not be declared more than once in any package; it is in effect
12893 The same untyped variable may be declared several times. Declarations are
12894 elaborated in the order in which they appear, so the new value replaces
12895 the old one, and any subsequent reference to the variable uses the new value.
12896 However, as noted above, if a variable has been declared as a string, all
12898 declarations must give it a string value. Similarly, if a variable has
12899 been declared as a string list, all subsequent declarations
12900 must give it a string list value.
12902 A @emph{variable reference} may take several forms:
12905 @item The simple variable name, for a variable in the current package (if any)
12906 or in the current project
12907 @item An expanded name, whose prefix is a context name.
12911 A @emph{context} may be one of the following:
12914 @item The name of an existing package in the current project
12915 @item The name of an imported project of the current project
12916 @item The name of an ancestor project (i.e., a project extended by the current
12917 project, either directly or indirectly)
12918 @item An expanded name whose prefix is an imported/parent project name, and
12919 whose selector is a package name in that project.
12923 A variable reference may be used in an expression.
12926 @subsection Attributes
12929 A project (and its packages) may have @emph{attributes} that define
12930 the project's properties. Some attributes have values that are strings;
12931 others have values that are string lists.
12933 There are two categories of attributes: @emph{simple attributes}
12934 and @emph{associative arrays} (@pxref{Associative Array Attributes}).
12936 Legal project attribute names, and attribute names for each legal package are
12937 listed below. Attributes names are case-insensitive.
12939 The following attributes are defined on projects (all are simple attributes):
12941 @multitable @columnfractions .4 .3
12942 @item @emph{Attribute Name}
12944 @item @code{Source_Files}
12946 @item @code{Source_Dirs}
12948 @item @code{Source_List_File}
12950 @item @code{Object_Dir}
12952 @item @code{Exec_Dir}
12954 @item @code{Excluded_Source_Dirs}
12956 @item @code{Excluded_Source_Files}
12958 @item @code{Excluded_Source_List_File}
12960 @item @code{Languages}
12964 @item @code{Library_Dir}
12966 @item @code{Library_Name}
12968 @item @code{Library_Kind}
12970 @item @code{Library_Version}
12972 @item @code{Library_Interface}
12974 @item @code{Library_Auto_Init}
12976 @item @code{Library_Options}
12978 @item @code{Library_Src_Dir}
12980 @item @code{Library_ALI_Dir}
12982 @item @code{Library_GCC}
12984 @item @code{Library_Symbol_File}
12986 @item @code{Library_Symbol_Policy}
12988 @item @code{Library_Reference_Symbol_File}
12990 @item @code{Externally_Built}
12995 The following attributes are defined for package @code{Naming}
12996 (@pxref{Naming Schemes}):
12998 @multitable @columnfractions .4 .2 .2 .2
12999 @item Attribute Name @tab Category @tab Index @tab Value
13000 @item @code{Spec_Suffix}
13001 @tab associative array
13004 @item @code{Body_Suffix}
13005 @tab associative array
13008 @item @code{Separate_Suffix}
13009 @tab simple attribute
13012 @item @code{Casing}
13013 @tab simple attribute
13016 @item @code{Dot_Replacement}
13017 @tab simple attribute
13021 @tab associative array
13025 @tab associative array
13028 @item @code{Specification_Exceptions}
13029 @tab associative array
13032 @item @code{Implementation_Exceptions}
13033 @tab associative array
13039 The following attributes are defined for packages @code{Builder},
13040 @code{Compiler}, @code{Binder},
13041 @code{Linker}, @code{Cross_Reference}, and @code{Finder}
13042 (@pxref{^Switches^Switches^ and Project Files}).
13044 @multitable @columnfractions .4 .2 .2 .2
13045 @item Attribute Name @tab Category @tab Index @tab Value
13046 @item @code{^Default_Switches^Default_Switches^}
13047 @tab associative array
13050 @item @code{^Switches^Switches^}
13051 @tab associative array
13057 In addition, package @code{Compiler} has a single string attribute
13058 @code{Local_Configuration_Pragmas} and package @code{Builder} has a single
13059 string attribute @code{Global_Configuration_Pragmas}.
13062 Each simple attribute has a default value: the empty string (for string-valued
13063 attributes) and the empty list (for string list-valued attributes).
13065 An attribute declaration defines a new value for an attribute.
13067 Examples of simple attribute declarations:
13069 @smallexample @c projectfile
13070 for Object_Dir use "objects";
13071 for Source_Dirs use ("units", "test/drivers");
13075 The syntax of a @dfn{simple attribute declaration} is similar to that of an
13076 attribute definition clause in Ada.
13078 Attributes references may be appear in expressions.
13079 The general form for such a reference is @code{<entity>'<attribute>}:
13080 Associative array attributes are functions. Associative
13081 array attribute references must have an argument that is a string literal.
13085 @smallexample @c projectfile
13087 Naming'Dot_Replacement
13088 Imported_Project'Source_Dirs
13089 Imported_Project.Naming'Casing
13090 Builder'^Default_Switches^Default_Switches^("Ada")
13094 The prefix of an attribute may be:
13096 @item @code{project} for an attribute of the current project
13097 @item The name of an existing package of the current project
13098 @item The name of an imported project
13099 @item The name of a parent project that is extended by the current project
13100 @item An expanded name whose prefix is imported/parent project name,
13101 and whose selector is a package name
13106 @smallexample @c projectfile
13109 for Source_Dirs use project'Source_Dirs & "units";
13110 for Source_Dirs use project'Source_Dirs & "test/drivers"
13116 In the first attribute declaration, initially the attribute @code{Source_Dirs}
13117 has the default value: an empty string list. After this declaration,
13118 @code{Source_Dirs} is a string list of one element: @code{"units"}.
13119 After the second attribute declaration @code{Source_Dirs} is a string list of
13120 two elements: @code{"units"} and @code{"test/drivers"}.
13122 Note: this example is for illustration only. In practice,
13123 the project file would contain only one attribute declaration:
13125 @smallexample @c projectfile
13126 for Source_Dirs use ("units", "test/drivers");
13129 @node Associative Array Attributes
13130 @subsection Associative Array Attributes
13133 Some attributes are defined as @emph{associative arrays}. An associative
13134 array may be regarded as a function that takes a string as a parameter
13135 and delivers a string or string list value as its result.
13137 Here are some examples of single associative array attribute associations:
13139 @smallexample @c projectfile
13140 for Body ("main") use "Main.ada";
13141 for ^Switches^Switches^ ("main.ada")
13143 "^-gnatv^-gnatv^");
13144 for ^Switches^Switches^ ("main.ada")
13145 use Builder'^Switches^Switches^ ("main.ada")
13150 Like untyped variables and simple attributes, associative array attributes
13151 may be declared several times. Each declaration supplies a new value for the
13152 attribute, and replaces the previous setting.
13155 An associative array attribute may be declared as a full associative array
13156 declaration, with the value of the same attribute in an imported or extended
13159 @smallexample @c projectfile
13161 for Default_Switches use Default.Builder'Default_Switches;
13166 In this example, @code{Default} must be either a project imported by the
13167 current project, or the project that the current project extends. If the
13168 attribute is in a package (in this case, in package @code{Builder}), the same
13169 package needs to be specified.
13172 A full associative array declaration replaces any other declaration for the
13173 attribute, including other full associative array declaration. Single
13174 associative array associations may be declare after a full associative
13175 declaration, modifying the value for a single association of the attribute.
13177 @node case Constructions
13178 @subsection @code{case} Constructions
13181 A @code{case} construction is used in a project file to effect conditional
13183 Here is a typical example:
13185 @smallexample @c projectfile
13188 type OS_Type is ("GNU/Linux", "Unix", "NT", "VMS");
13190 OS : OS_Type := external ("OS", "GNU/Linux");
13194 package Compiler is
13196 when "GNU/Linux" | "Unix" =>
13197 for ^Default_Switches^Default_Switches^ ("Ada")
13198 use ("^-gnath^-gnath^");
13200 for ^Default_Switches^Default_Switches^ ("Ada")
13201 use ("^-gnatP^-gnatP^");
13210 The syntax of a @code{case} construction is based on the Ada case statement
13211 (although there is no @code{null} construction for empty alternatives).
13213 The case expression must be a typed string variable.
13214 Each alternative comprises the reserved word @code{when}, either a list of
13215 literal strings separated by the @code{"|"} character or the reserved word
13216 @code{others}, and the @code{"=>"} token.
13217 Each literal string must belong to the string type that is the type of the
13219 An @code{others} alternative, if present, must occur last.
13221 After each @code{=>}, there are zero or more constructions. The only
13222 constructions allowed in a case construction are other case constructions,
13223 attribute declarations and variable declarations. String type declarations and
13224 package declarations are not allowed. Variable declarations are restricted to
13225 variables that have already been declared before the case construction.
13227 The value of the case variable is often given by an external reference
13228 (@pxref{External References in Project Files}).
13230 @c ****************************************
13231 @c * Objects and Sources in Project Files *
13232 @c ****************************************
13234 @node Objects and Sources in Project Files
13235 @section Objects and Sources in Project Files
13238 * Object Directory::
13240 * Source Directories::
13241 * Source File Names::
13245 Each project has exactly one object directory and one or more source
13246 directories. The source directories must contain at least one source file,
13247 unless the project file explicitly specifies that no source files are present
13248 (@pxref{Source File Names}).
13250 @node Object Directory
13251 @subsection Object Directory
13254 The object directory for a project is the directory containing the compiler's
13255 output (such as @file{ALI} files and object files) for the project's immediate
13258 The object directory is given by the value of the attribute @code{Object_Dir}
13259 in the project file.
13261 @smallexample @c projectfile
13262 for Object_Dir use "objects";
13266 The attribute @code{Object_Dir} has a string value, the path name of the object
13267 directory. The path name may be absolute or relative to the directory of the
13268 project file. This directory must already exist, and be readable and writable.
13270 By default, when the attribute @code{Object_Dir} is not given an explicit value
13271 or when its value is the empty string, the object directory is the same as the
13272 directory containing the project file.
13274 @node Exec Directory
13275 @subsection Exec Directory
13278 The exec directory for a project is the directory containing the executables
13279 for the project's main subprograms.
13281 The exec directory is given by the value of the attribute @code{Exec_Dir}
13282 in the project file.
13284 @smallexample @c projectfile
13285 for Exec_Dir use "executables";
13289 The attribute @code{Exec_Dir} has a string value, the path name of the exec
13290 directory. The path name may be absolute or relative to the directory of the
13291 project file. This directory must already exist, and be writable.
13293 By default, when the attribute @code{Exec_Dir} is not given an explicit value
13294 or when its value is the empty string, the exec directory is the same as the
13295 object directory of the project file.
13297 @node Source Directories
13298 @subsection Source Directories
13301 The source directories of a project are specified by the project file
13302 attribute @code{Source_Dirs}.
13304 This attribute's value is a string list. If the attribute is not given an
13305 explicit value, then there is only one source directory, the one where the
13306 project file resides.
13308 A @code{Source_Dirs} attribute that is explicitly defined to be the empty list,
13311 @smallexample @c projectfile
13312 for Source_Dirs use ();
13316 indicates that the project contains no source files.
13318 Otherwise, each string in the string list designates one or more
13319 source directories.
13321 @smallexample @c projectfile
13322 for Source_Dirs use ("sources", "test/drivers");
13326 If a string in the list ends with @code{"/**"}, then the directory whose path
13327 name precedes the two asterisks, as well as all its subdirectories
13328 (recursively), are source directories.
13330 @smallexample @c projectfile
13331 for Source_Dirs use ("/system/sources/**");
13335 Here the directory @code{/system/sources} and all of its subdirectories
13336 (recursively) are source directories.
13338 To specify that the source directories are the directory of the project file
13339 and all of its subdirectories, you can declare @code{Source_Dirs} as follows:
13340 @smallexample @c projectfile
13341 for Source_Dirs use ("./**");
13345 Each of the source directories must exist and be readable.
13347 @node Source File Names
13348 @subsection Source File Names
13351 In a project that contains source files, their names may be specified by the
13352 attributes @code{Source_Files} (a string list) or @code{Source_List_File}
13353 (a string). Source file names never include any directory information.
13355 If the attribute @code{Source_Files} is given an explicit value, then each
13356 element of the list is a source file name.
13358 @smallexample @c projectfile
13359 for Source_Files use ("main.adb");
13360 for Source_Files use ("main.adb", "pack1.ads", "pack2.adb");
13364 If the attribute @code{Source_Files} is not given an explicit value,
13365 but the attribute @code{Source_List_File} is given a string value,
13366 then the source file names are contained in the text file whose path name
13367 (absolute or relative to the directory of the project file) is the
13368 value of the attribute @code{Source_List_File}.
13370 Each line in the file that is not empty or is not a comment
13371 contains a source file name.
13373 @smallexample @c projectfile
13374 for Source_List_File use "source_list.txt";
13378 By default, if neither the attribute @code{Source_Files} nor the attribute
13379 @code{Source_List_File} is given an explicit value, then each file in the
13380 source directories that conforms to the project's naming scheme
13381 (@pxref{Naming Schemes}) is an immediate source of the project.
13383 A warning is issued if both attributes @code{Source_Files} and
13384 @code{Source_List_File} are given explicit values. In this case, the attribute
13385 @code{Source_Files} prevails.
13387 Each source file name must be the name of one existing source file
13388 in one of the source directories.
13390 A @code{Source_Files} attribute whose value is an empty list
13391 indicates that there are no source files in the project.
13393 If the order of the source directories is known statically, that is if
13394 @code{"/**"} is not used in the string list @code{Source_Dirs}, then there may
13395 be several files with the same source file name. In this case, only the file
13396 in the first directory is considered as an immediate source of the project
13397 file. If the order of the source directories is not known statically, it is
13398 an error to have several files with the same source file name.
13400 Projects can be specified to have no Ada source
13401 files: the value of @code{Source_Dirs} or @code{Source_Files} may be an empty
13402 list, or the @code{"Ada"} may be absent from @code{Languages}:
13404 @smallexample @c projectfile
13405 for Source_Dirs use ();
13406 for Source_Files use ();
13407 for Languages use ("C", "C++");
13411 Otherwise, a project must contain at least one immediate source.
13413 Projects with no source files are useful as template packages
13414 (@pxref{Packages in Project Files}) for other projects; in particular to
13415 define a package @code{Naming} (@pxref{Naming Schemes}).
13417 @c ****************************
13418 @c * Importing Projects *
13419 @c ****************************
13421 @node Importing Projects
13422 @section Importing Projects
13423 @cindex @code{ADA_PROJECT_PATH}
13424 @cindex @code{GPR_PROJECT_PATH}
13427 An immediate source of a project P may depend on source files that
13428 are neither immediate sources of P nor in the predefined library.
13429 To get this effect, P must @emph{import} the projects that contain the needed
13432 @smallexample @c projectfile
13434 with "project1", "utilities.gpr";
13435 with "/namings/apex.gpr";
13442 As can be seen in this example, the syntax for importing projects is similar
13443 to the syntax for importing compilation units in Ada. However, project files
13444 use literal strings instead of names, and the @code{with} clause identifies
13445 project files rather than packages.
13447 Each literal string is the file name or path name (absolute or relative) of a
13448 project file. If a string corresponds to a file name, with no path or a
13449 relative path, then its location is determined by the @emph{project path}. The
13450 latter can be queried using @code{gnatls -v}. It contains:
13454 In first position, the directory containing the current project file.
13456 In last position, the default project directory. This default project directory
13457 is part of the GNAT installation and is the standard place to install project
13458 files giving access to standard support libraries.
13460 @ref{Installing a library}
13464 In between, all the directories referenced in the
13465 ^environment variables^logical names^ @env{GPR_PROJECT_PATH}
13466 and @env{ADA_PROJECT_PATH} if they exist, and in that order.
13470 If a relative pathname is used, as in
13472 @smallexample @c projectfile
13477 then the full path for the project is constructed by concatenating this
13478 relative path to those in the project path, in order, until a matching file is
13479 found. Any symbolic link will be fully resolved in the directory of the
13480 importing project file before the imported project file is examined.
13482 If the @code{with}'ed project file name does not have an extension,
13483 the default is @file{^.gpr^.GPR^}. If a file with this extension is not found,
13484 then the file name as specified in the @code{with} clause (no extension) will
13485 be used. In the above example, if a file @code{project1.gpr} is found, then it
13486 will be used; otherwise, if a file @code{^project1^PROJECT1^} exists
13487 then it will be used; if neither file exists, this is an error.
13489 A warning is issued if the name of the project file does not match the
13490 name of the project; this check is case insensitive.
13492 Any source file that is an immediate source of the imported project can be
13493 used by the immediate sources of the importing project, transitively. Thus
13494 if @code{A} imports @code{B}, and @code{B} imports @code{C}, the immediate
13495 sources of @code{A} may depend on the immediate sources of @code{C}, even if
13496 @code{A} does not import @code{C} explicitly. However, this is not recommended,
13497 because if and when @code{B} ceases to import @code{C}, some sources in
13498 @code{A} will no longer compile.
13500 A side effect of this capability is that normally cyclic dependencies are not
13501 permitted: if @code{A} imports @code{B} (directly or indirectly) then @code{B}
13502 is not allowed to import @code{A}. However, there are cases when cyclic
13503 dependencies would be beneficial. For these cases, another form of import
13504 between projects exists, the @code{limited with}: a project @code{A} that
13505 imports a project @code{B} with a straight @code{with} may also be imported,
13506 directly or indirectly, by @code{B} on the condition that imports from @code{B}
13507 to @code{A} include at least one @code{limited with}.
13509 @smallexample @c 0projectfile
13515 limited with "../a/a.gpr";
13523 limited with "../a/a.gpr";
13529 In the above legal example, there are two project cycles:
13532 @item A -> C -> D -> A
13536 In each of these cycle there is one @code{limited with}: import of @code{A}
13537 from @code{B} and import of @code{A} from @code{D}.
13539 The difference between straight @code{with} and @code{limited with} is that
13540 the name of a project imported with a @code{limited with} cannot be used in the
13541 project that imports it. In particular, its packages cannot be renamed and
13542 its variables cannot be referred to.
13544 An exception to the above rules for @code{limited with} is that for the main
13545 project specified to @command{gnatmake} or to the @command{GNAT} driver a
13546 @code{limited with} is equivalent to a straight @code{with}. For example,
13547 in the example above, projects @code{B} and @code{D} could not be main
13548 projects for @command{gnatmake} or to the @command{GNAT} driver, because they
13549 each have a @code{limited with} that is the only one in a cycle of importing
13552 @c *********************
13553 @c * Project Extension *
13554 @c *********************
13556 @node Project Extension
13557 @section Project Extension
13560 During development of a large system, it is sometimes necessary to use
13561 modified versions of some of the source files, without changing the original
13562 sources. This can be achieved through the @emph{project extension} facility.
13564 @smallexample @c projectfile
13565 project Modified_Utilities extends "/baseline/utilities.gpr" is @dots{}
13569 A project extension declaration introduces an extending project
13570 (the @emph{child}) and a project being extended (the @emph{parent}).
13572 By default, a child project inherits all the sources of its parent.
13573 However, inherited sources can be overridden: a unit in a parent is hidden
13574 by a unit of the same name in the child.
13576 Inherited sources are considered to be sources (but not immediate sources)
13577 of the child project; see @ref{Project File Syntax}.
13579 An inherited source file retains any switches specified in the parent project.
13581 For example if the project @code{Utilities} contains the spec and the
13582 body of an Ada package @code{Util_IO}, then the project
13583 @code{Modified_Utilities} can contain a new body for package @code{Util_IO}.
13584 The original body of @code{Util_IO} will not be considered in program builds.
13585 However, the package spec will still be found in the project
13588 A child project can have only one parent, except when it is qualified as
13589 abstract. But it may import any number of other projects.
13591 A project is not allowed to import directly or indirectly at the same time a
13592 child project and any of its ancestors.
13594 @c *******************************
13595 @c * Project Hierarchy Extension *
13596 @c *******************************
13598 @node Project Hierarchy Extension
13599 @section Project Hierarchy Extension
13602 When extending a large system spanning multiple projects, it is often
13603 inconvenient to extend every project in the hierarchy that is impacted by a
13604 small change introduced. In such cases, it is possible to create a virtual
13605 extension of entire hierarchy using @code{extends all} relationship.
13607 When the project is extended using @code{extends all} inheritance, all projects
13608 that are imported by it, both directly and indirectly, are considered virtually
13609 extended. That is, the Project Manager creates "virtual projects"
13610 that extend every project in the hierarchy; all these virtual projects have
13611 no sources of their own and have as object directory the object directory of
13612 the root of "extending all" project.
13614 It is possible to explicitly extend one or more projects in the hierarchy
13615 in order to modify the sources. These extending projects must be imported by
13616 the "extending all" project, which will replace the corresponding virtual
13617 projects with the explicit ones.
13619 When building such a project hierarchy extension, the Project Manager will
13620 ensure that both modified sources and sources in virtual extending projects
13621 that depend on them, are recompiled.
13623 By means of example, consider the following hierarchy of projects.
13627 project A, containing package P1
13629 project B importing A and containing package P2 which depends on P1
13631 project C importing B and containing package P3 which depends on P2
13635 We want to modify packages P1 and P3.
13637 This project hierarchy will need to be extended as follows:
13641 Create project A1 that extends A, placing modified P1 there:
13643 @smallexample @c 0projectfile
13644 project A1 extends "(@dots{})/A" is
13649 Create project C1 that "extends all" C and imports A1, placing modified
13652 @smallexample @c 0projectfile
13653 with "(@dots{})/A1";
13654 project C1 extends all "(@dots{})/C" is
13659 When you build project C1, your entire modified project space will be
13660 recompiled, including the virtual project B1 that has been impacted by the
13661 "extending all" inheritance of project C.
13663 Note that if a Library Project in the hierarchy is virtually extended,
13664 the virtual project that extends the Library Project is not a Library Project.
13666 @c ****************************************
13667 @c * External References in Project Files *
13668 @c ****************************************
13670 @node External References in Project Files
13671 @section External References in Project Files
13674 A project file may contain references to external variables; such references
13675 are called @emph{external references}.
13677 An external variable is either defined as part of the environment (an
13678 environment variable in Unix, for example) or else specified on the command
13679 line via the @option{^-X^/EXTERNAL_REFERENCE=^@emph{vbl}=@emph{value}} switch.
13680 If both, then the command line value is used.
13682 The value of an external reference is obtained by means of the built-in
13683 function @code{external}, which returns a string value.
13684 This function has two forms:
13686 @item @code{external (external_variable_name)}
13687 @item @code{external (external_variable_name, default_value)}
13691 Each parameter must be a string literal. For example:
13693 @smallexample @c projectfile
13695 external ("OS", "GNU/Linux")
13699 In the form with one parameter, the function returns the value of
13700 the external variable given as parameter. If this name is not present in the
13701 environment, the function returns an empty string.
13703 In the form with two string parameters, the second argument is
13704 the value returned when the variable given as the first argument is not
13705 present in the environment. In the example above, if @code{"OS"} is not
13706 the name of ^an environment variable^a logical name^ and is not passed on
13707 the command line, then the returned value is @code{"GNU/Linux"}.
13709 An external reference may be part of a string expression or of a string
13710 list expression, and can therefore appear in a variable declaration or
13711 an attribute declaration.
13713 @smallexample @c projectfile
13715 type Mode_Type is ("Debug", "Release");
13716 Mode : Mode_Type := external ("MODE");
13723 @c *****************************
13724 @c * Packages in Project Files *
13725 @c *****************************
13727 @node Packages in Project Files
13728 @section Packages in Project Files
13731 A @emph{package} defines the settings for project-aware tools within a
13733 For each such tool one can declare a package; the names for these
13734 packages are preset (@pxref{Packages}).
13735 A package may contain variable declarations, attribute declarations, and case
13738 @smallexample @c projectfile
13741 package Builder is -- used by gnatmake
13742 for ^Default_Switches^Default_Switches^ ("Ada")
13751 The syntax of package declarations mimics that of package in Ada.
13753 Most of the packages have an attribute
13754 @code{^Default_Switches^Default_Switches^}.
13755 This attribute is an associative array, and its value is a string list.
13756 The index of the associative array is the name of a programming language (case
13757 insensitive). This attribute indicates the ^switch^switch^
13758 or ^switches^switches^ to be used
13759 with the corresponding tool.
13761 Some packages also have another attribute, @code{^Switches^Switches^},
13762 an associative array whose value is a string list.
13763 The index is the name of a source file.
13764 This attribute indicates the ^switch^switch^
13765 or ^switches^switches^ to be used by the corresponding
13766 tool when dealing with this specific file.
13768 Further information on these ^switch^switch^-related attributes is found in
13769 @ref{^Switches^Switches^ and Project Files}.
13771 A package may be declared as a @emph{renaming} of another package; e.g., from
13772 the project file for an imported project.
13774 @smallexample @c projectfile
13776 with "/global/apex.gpr";
13778 package Naming renames Apex.Naming;
13785 Packages that are renamed in other project files often come from project files
13786 that have no sources: they are just used as templates. Any modification in the
13787 template will be reflected automatically in all the project files that rename
13788 a package from the template.
13790 In addition to the tool-oriented packages, you can also declare a package
13791 named @code{Naming} to establish specialized source file naming conventions
13792 (@pxref{Naming Schemes}).
13794 @c ************************************
13795 @c * Variables from Imported Projects *
13796 @c ************************************
13798 @node Variables from Imported Projects
13799 @section Variables from Imported Projects
13802 An attribute or variable defined in an imported or parent project can
13803 be used in expressions in the importing / extending project.
13804 Such an attribute or variable is denoted by an expanded name whose prefix
13805 is either the name of the project or the expanded name of a package within
13808 @smallexample @c projectfile
13811 project Main extends "base" is
13812 Var1 := Imported.Var;
13813 Var2 := Base.Var & ".new";
13818 for ^Default_Switches^Default_Switches^ ("Ada")
13819 use Imported.Builder'Ada_^Switches^Switches^ &
13820 "^-gnatg^-gnatg^" &
13826 package Compiler is
13827 for ^Default_Switches^Default_Switches^ ("Ada")
13828 use Base.Compiler'Ada_^Switches^Switches^;
13839 The value of @code{Var1} is a copy of the variable @code{Var} defined
13840 in the project file @file{"imported.gpr"}
13842 the value of @code{Var2} is a copy of the value of variable @code{Var}
13843 defined in the project file @file{base.gpr}, concatenated with @code{".new"}
13845 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
13846 @code{Builder} is a string list that includes in its value a copy of the value
13847 of @code{Ada_^Switches^Switches^} defined in the @code{Builder} package
13848 in project file @file{imported.gpr} plus two new elements:
13849 @option{"^-gnatg^-gnatg^"}
13850 and @option{"^-v^-v^"};
13852 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
13853 @code{Compiler} is a copy of the variable @code{Ada_^Switches^Switches^}
13854 defined in the @code{Compiler} package in project file @file{base.gpr},
13855 the project being extended.
13858 @c ******************
13859 @c * Naming Schemes *
13860 @c ******************
13862 @node Naming Schemes
13863 @section Naming Schemes
13866 Sometimes an Ada software system is ported from a foreign compilation
13867 environment to GNAT, and the file names do not use the default GNAT
13868 conventions. Instead of changing all the file names (which for a variety
13869 of reasons might not be possible), you can define the relevant file
13870 naming scheme in the @code{Naming} package in your project file.
13873 Note that the use of pragmas described in
13874 @ref{Alternative File Naming Schemes} by mean of a configuration
13875 pragmas file is not supported when using project files. You must use
13876 the features described in this paragraph. You can however use specify
13877 other configuration pragmas (@pxref{Specifying Configuration Pragmas}).
13880 For example, the following
13881 package models the Apex file naming rules:
13883 @smallexample @c projectfile
13886 for Casing use "lowercase";
13887 for Dot_Replacement use ".";
13888 for Spec_Suffix ("Ada") use ".1.ada";
13889 for Body_Suffix ("Ada") use ".2.ada";
13896 For example, the following package models the HP Ada file naming rules:
13898 @smallexample @c projectfile
13901 for Casing use "lowercase";
13902 for Dot_Replacement use "__";
13903 for Spec_Suffix ("Ada") use "_.^ada^ada^";
13904 for Body_Suffix ("Ada") use ".^ada^ada^";
13910 (Note that @code{Casing} is @code{"lowercase"} because GNAT gets the file
13911 names in lower case)
13915 You can define the following attributes in package @code{Naming}:
13919 @item @code{Casing}
13920 This must be a string with one of the three values @code{"lowercase"},
13921 @code{"uppercase"} or @code{"mixedcase"}; these strings are case insensitive.
13924 If @code{Casing} is not specified, then the default is @code{"lowercase"}.
13926 @item @code{Dot_Replacement}
13927 This must be a string whose value satisfies the following conditions:
13930 @item It must not be empty
13931 @item It cannot start or end with an alphanumeric character
13932 @item It cannot be a single underscore
13933 @item It cannot start with an underscore followed by an alphanumeric
13934 @item It cannot contain a dot @code{'.'} except if the entire string
13939 If @code{Dot_Replacement} is not specified, then the default is @code{"-"}.
13941 @item @code{Spec_Suffix}
13942 This is an associative array (indexed by the programming language name, case
13943 insensitive) whose value is a string that must satisfy the following
13947 @item It must not be empty
13948 @item It must include at least one dot
13951 If @code{Spec_Suffix ("Ada")} is not specified, then the default is
13952 @code{"^.ads^.ADS^"}.
13954 @item @code{Body_Suffix}
13955 This is an associative array (indexed by the programming language name, case
13956 insensitive) whose value is a string that must satisfy the following
13960 @item It must not be empty
13961 @item It must include at least one dot
13962 @item It cannot be the same as @code{Spec_Suffix ("Ada")}
13965 If @code{Body_Suffix ("Ada")} and @code{Spec_Suffix ("Ada")} end with the
13966 same string, then a file name that ends with the longest of these two suffixes
13967 will be a body if the longest suffix is @code{Body_Suffix ("Ada")} or a spec
13968 if the longest suffix is @code{Spec_Suffix ("Ada")}.
13970 If the suffix does not start with a '.', a file with a name exactly equal
13971 to the suffix will also be part of the project (for instance if you define
13972 the suffix as @code{Makefile}, a file called @file{Makefile} will be part
13973 of the project. This is not interesting in general when using projects to
13974 compile. However, it might become useful when a project is also used to
13975 find the list of source files in an editor, like the GNAT Programming System
13978 If @code{Body_Suffix ("Ada")} is not specified, then the default is
13979 @code{"^.adb^.ADB^"}.
13981 @item @code{Separate_Suffix}
13982 This must be a string whose value satisfies the same conditions as
13983 @code{Body_Suffix}. The same "longest suffix" rules apply.
13986 If @code{Separate_Suffix ("Ada")} is not specified, then it defaults to same
13987 value as @code{Body_Suffix ("Ada")}.
13991 You can use the associative array attribute @code{Spec} to define
13992 the source file name for an individual Ada compilation unit's spec. The array
13993 index must be a string literal that identifies the Ada unit (case insensitive).
13994 The value of this attribute must be a string that identifies the file that
13995 contains this unit's spec (case sensitive or insensitive depending on the
13998 @smallexample @c projectfile
13999 for Spec ("MyPack.MyChild") use "mypack.mychild.spec";
14002 When the source file contains several units, you can indicate at what
14003 position the unit occurs in the file, with the following. The first unit
14004 in the file has index 1
14006 @smallexample @c projectfile
14007 for Body ("top") use "foo.a" at 1;
14008 for Body ("foo") use "foo.a" at 2;
14013 You can use the associative array attribute @code{Body} to
14014 define the source file name for an individual Ada compilation unit's body
14015 (possibly a subunit). The array index must be a string literal that identifies
14016 the Ada unit (case insensitive). The value of this attribute must be a string
14017 that identifies the file that contains this unit's body or subunit (case
14018 sensitive or insensitive depending on the operating system).
14020 @smallexample @c projectfile
14021 for Body ("MyPack.MyChild") use "mypack.mychild.body";
14025 @c ********************
14026 @c * Library Projects *
14027 @c ********************
14029 @node Library Projects
14030 @section Library Projects
14033 @emph{Library projects} are projects whose object code is placed in a library.
14034 (Note that this facility is not yet supported on all platforms).
14036 @code{gnatmake} or @code{gprbuild} will collect all object files into a
14037 single archive, which might either be a shared or a static library. This
14038 library can later on be linked with multiple executables, potentially
14039 reducing their sizes.
14041 If your project file specifies languages other than Ada, but you are still
14042 using @code{gnatmake} to compile and link, the latter will not try to
14043 compile your sources other than Ada (you should use @code{gprbuild} if that
14044 is your intent). However, @code{gnatmake} will automatically link all object
14045 files found in the object directory, whether or not they were compiled from
14046 an Ada source file. This specific behavior only applies when multiple
14047 languages are specified.
14049 To create a library project, you need to define in its project file
14050 two project-level attributes: @code{Library_Name} and @code{Library_Dir}.
14051 Additionally, you may define other library-related attributes such as
14052 @code{Library_Kind}, @code{Library_Version}, @code{Library_Interface},
14053 @code{Library_Auto_Init}, @code{Library_Options} and @code{Library_GCC}.
14055 The @code{Library_Name} attribute has a string value. There is no restriction
14056 on the name of a library. It is the responsibility of the developer to
14057 choose a name that will be accepted by the platform. It is recommended to
14058 choose names that could be Ada identifiers; such names are almost guaranteed
14059 to be acceptable on all platforms.
14061 The @code{Library_Dir} attribute has a string value that designates the path
14062 (absolute or relative) of the directory where the library will reside.
14063 It must designate an existing directory, and this directory must be writable,
14064 different from the project's object directory and from any source directory
14065 in the project tree.
14067 If both @code{Library_Name} and @code{Library_Dir} are specified and
14068 are legal, then the project file defines a library project. The optional
14069 library-related attributes are checked only for such project files.
14071 The @code{Library_Kind} attribute has a string value that must be one of the
14072 following (case insensitive): @code{"static"}, @code{"dynamic"} or
14073 @code{"relocatable"} (which is a synonym for @code{"dynamic"}). If this
14074 attribute is not specified, the library is a static library, that is
14075 an archive of object files that can be potentially linked into a
14076 static executable. Otherwise, the library may be dynamic or
14077 relocatable, that is a library that is loaded only at the start of execution.
14079 If you need to build both a static and a dynamic library, you should use two
14080 different object directories, since in some cases some extra code needs to
14081 be generated for the latter. For such cases, it is recommended to either use
14082 two different project files, or a single one which uses external variables
14083 to indicate what kind of library should be build.
14085 The @code{Library_ALI_Dir} attribute may be specified to indicate the
14086 directory where the ALI files of the library will be copied. When it is
14087 not specified, the ALI files are copied to the directory specified in
14088 attribute @code{Library_Dir}. The directory specified by @code{Library_ALI_Dir}
14089 must be writable and different from the project's object directory and from
14090 any source directory in the project tree.
14092 The @code{Library_Version} attribute has a string value whose interpretation
14093 is platform dependent. It has no effect on VMS and Windows. On Unix, it is
14094 used only for dynamic/relocatable libraries as the internal name of the
14095 library (the @code{"soname"}). If the library file name (built from the
14096 @code{Library_Name}) is different from the @code{Library_Version}, then the
14097 library file will be a symbolic link to the actual file whose name will be
14098 @code{Library_Version}.
14102 @smallexample @c projectfile
14108 for Library_Dir use "lib_dir";
14109 for Library_Name use "dummy";
14110 for Library_Kind use "relocatable";
14111 for Library_Version use "libdummy.so." & Version;
14118 Directory @file{lib_dir} will contain the internal library file whose name
14119 will be @file{libdummy.so.1}, and @file{libdummy.so} will be a symbolic link to
14120 @file{libdummy.so.1}.
14122 When @command{gnatmake} detects that a project file
14123 is a library project file, it will check all immediate sources of the project
14124 and rebuild the library if any of the sources have been recompiled.
14126 Standard project files can import library project files. In such cases,
14127 the libraries will only be rebuilt if some of its sources are recompiled
14128 because they are in the closure of some other source in an importing project.
14129 Sources of the library project files that are not in such a closure will
14130 not be checked, unless the full library is checked, because one of its sources
14131 needs to be recompiled.
14133 For instance, assume the project file @code{A} imports the library project file
14134 @code{L}. The immediate sources of A are @file{a1.adb}, @file{a2.ads} and
14135 @file{a2.adb}. The immediate sources of L are @file{l1.ads}, @file{l1.adb},
14136 @file{l2.ads}, @file{l2.adb}.
14138 If @file{l1.adb} has been modified, then the library associated with @code{L}
14139 will be rebuilt when compiling all the immediate sources of @code{A} only
14140 if @file{a1.ads}, @file{a2.ads} or @file{a2.adb} includes a statement
14143 To be sure that all the sources in the library associated with @code{L} are
14144 up to date, and that all the sources of project @code{A} are also up to date,
14145 the following two commands needs to be used:
14152 When a library is built or rebuilt, an attempt is made first to delete all
14153 files in the library directory.
14154 All @file{ALI} files will also be copied from the object directory to the
14155 library directory. To build executables, @command{gnatmake} will use the
14156 library rather than the individual object files.
14159 It is also possible to create library project files for third-party libraries
14160 that are precompiled and cannot be compiled locally thanks to the
14161 @code{externally_built} attribute. (See @ref{Installing a library}).
14164 @c *******************************
14165 @c * Stand-alone Library Projects *
14166 @c *******************************
14168 @node Stand-alone Library Projects
14169 @section Stand-alone Library Projects
14172 A Stand-alone Library is a library that contains the necessary code to
14173 elaborate the Ada units that are included in the library. A Stand-alone
14174 Library is suitable to be used in an executable when the main is not
14175 in Ada. However, Stand-alone Libraries may also be used with an Ada main
14178 A Stand-alone Library Project is a Library Project where the library is
14179 a Stand-alone Library.
14181 To be a Stand-alone Library Project, in addition to the two attributes
14182 that make a project a Library Project (@code{Library_Name} and
14183 @code{Library_Dir}, see @ref{Library Projects}), the attribute
14184 @code{Library_Interface} must be defined.
14186 @smallexample @c projectfile
14188 for Library_Dir use "lib_dir";
14189 for Library_Name use "dummy";
14190 for Library_Interface use ("int1", "int1.child");
14194 Attribute @code{Library_Interface} has a nonempty string list value,
14195 each string in the list designating a unit contained in an immediate source
14196 of the project file.
14198 When a Stand-alone Library is built, first the binder is invoked to build
14199 a package whose name depends on the library name
14200 (^b~dummy.ads/b^B$DUMMY.ADS/B^ in the example above).
14201 This binder-generated package includes initialization and
14202 finalization procedures whose
14203 names depend on the library name (dummyinit and dummyfinal in the example
14204 above). The object corresponding to this package is included in the library.
14206 A dynamic or relocatable Stand-alone Library is automatically initialized
14207 if automatic initialization of Stand-alone Libraries is supported on the
14208 platform and if attribute @code{Library_Auto_Init} is not specified or
14209 is specified with the value "true". A static Stand-alone Library is never
14210 automatically initialized.
14212 Single string attribute @code{Library_Auto_Init} may be specified with only
14213 two possible values: "false" or "true" (case-insensitive). Specifying
14214 "false" for attribute @code{Library_Auto_Init} will prevent automatic
14215 initialization of dynamic or relocatable libraries.
14217 When a non-automatically initialized Stand-alone Library is used
14218 in an executable, its initialization procedure must be called before
14219 any service of the library is used.
14220 When the main subprogram is in Ada, it may mean that the initialization
14221 procedure has to be called during elaboration of another package.
14223 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
14224 (those that are listed in attribute @code{Library_Interface}) are copied to
14225 the Library Directory. As a consequence, only the Interface Units may be
14226 imported from Ada units outside of the library. If other units are imported,
14227 the binding phase will fail.
14229 When a Stand-Alone Library is bound, the switches that are specified in
14230 the attribute @code{Default_Switches ("Ada")} in package @code{Binder} are
14231 used in the call to @command{gnatbind}.
14233 The string list attribute @code{Library_Options} may be used to specified
14234 additional switches to the call to @command{gcc} to link the library.
14236 The attribute @code{Library_Src_Dir}, may be specified for a
14237 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
14238 single string value. Its value must be the path (absolute or relative to the
14239 project directory) of an existing directory. This directory cannot be the
14240 object directory or one of the source directories, but it can be the same as
14241 the library directory. The sources of the Interface
14242 Units of the library, necessary to an Ada client of the library, will be
14243 copied to the designated directory, called Interface Copy directory.
14244 These sources includes the specs of the Interface Units, but they may also
14245 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
14246 are used, or when there is a generic units in the spec. Before the sources
14247 are copied to the Interface Copy directory, an attempt is made to delete all
14248 files in the Interface Copy directory.
14250 @c *************************************
14251 @c * Switches Related to Project Files *
14252 @c *************************************
14253 @node Switches Related to Project Files
14254 @section Switches Related to Project Files
14257 The following switches are used by GNAT tools that support project files:
14261 @item ^-P^/PROJECT_FILE=^@var{project}
14262 @cindex @option{^-P^/PROJECT_FILE^} (any project-aware tool)
14263 Indicates the name of a project file. This project file will be parsed with
14264 the verbosity indicated by @option{^-vP^MESSAGE_PROJECT_FILES=^@emph{x}},
14265 if any, and using the external references indicated
14266 by @option{^-X^/EXTERNAL_REFERENCE^} switches, if any.
14268 There may zero, one or more spaces between @option{-P} and @var{project}.
14272 There must be only one @option{^-P^/PROJECT_FILE^} switch on the command line.
14275 Since the Project Manager parses the project file only after all the switches
14276 on the command line are checked, the order of the switches
14277 @option{^-P^/PROJECT_FILE^},
14278 @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}}
14279 or @option{^-X^/EXTERNAL_REFERENCE^} is not significant.
14281 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
14282 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (any project-aware tool)
14283 Indicates that external variable @var{name} has the value @var{value}.
14284 The Project Manager will use this value for occurrences of
14285 @code{external(name)} when parsing the project file.
14289 If @var{name} or @var{value} includes a space, then @var{name=value} should be
14290 put between quotes.
14298 Several @option{^-X^/EXTERNAL_REFERENCE^} switches can be used simultaneously.
14299 If several @option{^-X^/EXTERNAL_REFERENCE^} switches specify the same
14300 @var{name}, only the last one is used.
14303 An external variable specified with a @option{^-X^/EXTERNAL_REFERENCE^} switch
14304 takes precedence over the value of the same name in the environment.
14306 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
14307 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (any project-aware tool)
14308 Indicates the verbosity of the parsing of GNAT project files.
14311 @option{-vP0} means Default;
14312 @option{-vP1} means Medium;
14313 @option{-vP2} means High.
14317 There are three possible options for this qualifier: DEFAULT, MEDIUM and
14322 The default is ^Default^DEFAULT^: no output for syntactically correct
14325 If several @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}} switches are present,
14326 only the last one is used.
14328 @item ^-aP^/ADD_PROJECT_SEARCH_DIR=^<dir>
14329 @cindex @option{^-aP^/ADD_PROJECT_SEARCH_DIR=^} (any project-aware tool)
14330 Add directory <dir> at the beginning of the project search path, in order,
14331 after the current working directory.
14335 @cindex @option{-eL} (any project-aware tool)
14336 Follow all symbolic links when processing project files.
14339 @item ^--subdirs^/SUBDIRS^=<subdir>
14340 @cindex @option{^--subdirs^/SUBDIRS^=} (gnatmake and gnatclean)
14341 This switch is recognized by gnatmake and gnatclean. It indicate that the real
14342 directories (except the source directories) are the subdirectories <subdir>
14343 of the directories specified in the project files. This applies in particular
14344 to object directories, library directories and exec directories. If the
14345 subdirectories do not exist, they are created automatically.
14349 @c **********************************
14350 @c * Tools Supporting Project Files *
14351 @c **********************************
14353 @node Tools Supporting Project Files
14354 @section Tools Supporting Project Files
14357 * gnatmake and Project Files::
14358 * The GNAT Driver and Project Files::
14361 @node gnatmake and Project Files
14362 @subsection gnatmake and Project Files
14365 This section covers several topics related to @command{gnatmake} and
14366 project files: defining ^switches^switches^ for @command{gnatmake}
14367 and for the tools that it invokes; specifying configuration pragmas;
14368 the use of the @code{Main} attribute; building and rebuilding library project
14372 * ^Switches^Switches^ and Project Files::
14373 * Specifying Configuration Pragmas::
14374 * Project Files and Main Subprograms::
14375 * Library Project Files::
14378 @node ^Switches^Switches^ and Project Files
14379 @subsubsection ^Switches^Switches^ and Project Files
14382 It is not currently possible to specify VMS style qualifiers in the project
14383 files; only Unix style ^switches^switches^ may be specified.
14387 For each of the packages @code{Builder}, @code{Compiler}, @code{Binder}, and
14388 @code{Linker}, you can specify a @code{^Default_Switches^Default_Switches^}
14389 attribute, a @code{^Switches^Switches^} attribute, or both;
14390 as their names imply, these ^switch^switch^-related
14391 attributes affect the ^switches^switches^ that are used for each of these GNAT
14393 @command{gnatmake} is invoked. As will be explained below, these
14394 component-specific ^switches^switches^ precede
14395 the ^switches^switches^ provided on the @command{gnatmake} command line.
14397 The @code{^Default_Switches^Default_Switches^} attribute is an associative
14398 array indexed by language name (case insensitive) whose value is a string list.
14401 @smallexample @c projectfile
14403 package Compiler is
14404 for ^Default_Switches^Default_Switches^ ("Ada")
14405 use ("^-gnaty^-gnaty^",
14412 The @code{^Switches^Switches^} attribute is also an associative array,
14413 indexed by a file name (which may or may not be case sensitive, depending
14414 on the operating system) whose value is a string list. For example:
14416 @smallexample @c projectfile
14419 for ^Switches^Switches^ ("main1.adb")
14421 for ^Switches^Switches^ ("main2.adb")
14428 For the @code{Builder} package, the file names must designate source files
14429 for main subprograms. For the @code{Binder} and @code{Linker} packages, the
14430 file names must designate @file{ALI} or source files for main subprograms.
14431 In each case just the file name without an explicit extension is acceptable.
14433 For each tool used in a program build (@command{gnatmake}, the compiler, the
14434 binder, and the linker), the corresponding package @dfn{contributes} a set of
14435 ^switches^switches^ for each file on which the tool is invoked, based on the
14436 ^switch^switch^-related attributes defined in the package.
14437 In particular, the ^switches^switches^
14438 that each of these packages contributes for a given file @var{f} comprise:
14442 the value of attribute @code{^Switches^Switches^ (@var{f})},
14443 if it is specified in the package for the given file,
14445 otherwise, the value of @code{^Default_Switches^Default_Switches^ ("Ada")},
14446 if it is specified in the package.
14450 If neither of these attributes is defined in the package, then the package does
14451 not contribute any ^switches^switches^ for the given file.
14453 When @command{gnatmake} is invoked on a file, the ^switches^switches^ comprise
14454 two sets, in the following order: those contributed for the file
14455 by the @code{Builder} package;
14456 and the switches passed on the command line.
14458 When @command{gnatmake} invokes a tool (compiler, binder, linker) on a file,
14459 the ^switches^switches^ passed to the tool comprise three sets,
14460 in the following order:
14464 the applicable ^switches^switches^ contributed for the file
14465 by the @code{Builder} package in the project file supplied on the command line;
14468 those contributed for the file by the package (in the relevant project file --
14469 see below) corresponding to the tool; and
14472 the applicable switches passed on the command line.
14476 The term @emph{applicable ^switches^switches^} reflects the fact that
14477 @command{gnatmake} ^switches^switches^ may or may not be passed to individual
14478 tools, depending on the individual ^switch^switch^.
14480 @command{gnatmake} may invoke the compiler on source files from different
14481 projects. The Project Manager will use the appropriate project file to
14482 determine the @code{Compiler} package for each source file being compiled.
14483 Likewise for the @code{Binder} and @code{Linker} packages.
14485 As an example, consider the following package in a project file:
14487 @smallexample @c projectfile
14490 package Compiler is
14491 for ^Default_Switches^Default_Switches^ ("Ada")
14493 for ^Switches^Switches^ ("a.adb")
14495 for ^Switches^Switches^ ("b.adb")
14497 "^-gnaty^-gnaty^");
14504 If @command{gnatmake} is invoked with this project file, and it needs to
14505 compile, say, the files @file{a.adb}, @file{b.adb}, and @file{c.adb}, then
14506 @file{a.adb} will be compiled with the ^switch^switch^
14507 @option{^-O1^-O1^},
14508 @file{b.adb} with ^switches^switches^
14510 and @option{^-gnaty^-gnaty^},
14511 and @file{c.adb} with @option{^-g^-g^}.
14513 The following example illustrates the ordering of the ^switches^switches^
14514 contributed by different packages:
14516 @smallexample @c projectfile
14520 for ^Switches^Switches^ ("main.adb")
14528 package Compiler is
14529 for ^Switches^Switches^ ("main.adb")
14537 If you issue the command:
14540 gnatmake ^-Pproj2^/PROJECT_FILE=PROJ2^ -O0 main
14544 then the compiler will be invoked on @file{main.adb} with the following
14545 sequence of ^switches^switches^
14548 ^-g -O1 -O2 -O0^-g -O1 -O2 -O0^
14551 with the last @option{^-O^-O^}
14552 ^switch^switch^ having precedence over the earlier ones;
14553 several other ^switches^switches^
14554 (such as @option{^-c^-c^}) are added implicitly.
14556 The ^switches^switches^
14558 and @option{^-O1^-O1^} are contributed by package
14559 @code{Builder}, @option{^-O2^-O2^} is contributed
14560 by the package @code{Compiler}
14561 and @option{^-O0^-O0^} comes from the command line.
14563 The @option{^-g^-g^}
14564 ^switch^switch^ will also be passed in the invocation of
14565 @command{Gnatlink.}
14567 A final example illustrates switch contributions from packages in different
14570 @smallexample @c projectfile
14573 for Source_Files use ("pack.ads", "pack.adb");
14574 package Compiler is
14575 for ^Default_Switches^Default_Switches^ ("Ada")
14576 use ("^-gnata^-gnata^");
14584 for Source_Files use ("foo_main.adb", "bar_main.adb");
14586 for ^Switches^Switches^ ("foo_main.adb")
14594 -- Ada source file:
14596 procedure Foo_Main is
14604 gnatmake ^-PProj4^/PROJECT_FILE=PROJ4^ foo_main.adb -cargs -gnato
14608 then the ^switches^switches^ passed to the compiler for @file{foo_main.adb} are
14609 @option{^-g^-g^} (contributed by the package @code{Proj4.Builder}) and
14610 @option{^-gnato^-gnato^} (passed on the command line).
14611 When the imported package @code{Pack} is compiled, the ^switches^switches^ used
14612 are @option{^-g^-g^} from @code{Proj4.Builder},
14613 @option{^-gnata^-gnata^} (contributed from package @code{Proj3.Compiler},
14614 and @option{^-gnato^-gnato^} from the command line.
14617 When using @command{gnatmake} with project files, some ^switches^switches^ or
14618 arguments may be expressed as relative paths. As the working directory where
14619 compilation occurs may change, these relative paths are converted to absolute
14620 paths. For the ^switches^switches^ found in a project file, the relative paths
14621 are relative to the project file directory, for the switches on the command
14622 line, they are relative to the directory where @command{gnatmake} is invoked.
14623 The ^switches^switches^ for which this occurs are:
14629 ^-aI^-aI^, as well as all arguments that are not switches (arguments to
14631 ^-o^-o^, object files specified in package @code{Linker} or after
14632 -largs on the command line). The exception to this rule is the ^switch^switch^
14633 ^--RTS=^--RTS=^ for which a relative path argument is never converted.
14635 @node Specifying Configuration Pragmas
14636 @subsubsection Specifying Configuration Pragmas
14638 When using @command{gnatmake} with project files, if there exists a file
14639 @file{gnat.adc} that contains configuration pragmas, this file will be
14642 Configuration pragmas can be defined by means of the following attributes in
14643 project files: @code{Global_Configuration_Pragmas} in package @code{Builder}
14644 and @code{Local_Configuration_Pragmas} in package @code{Compiler}.
14646 Both these attributes are single string attributes. Their values is the path
14647 name of a file containing configuration pragmas. If a path name is relative,
14648 then it is relative to the project directory of the project file where the
14649 attribute is defined.
14651 When compiling a source, the configuration pragmas used are, in order,
14652 those listed in the file designated by attribute
14653 @code{Global_Configuration_Pragmas} in package @code{Builder} of the main
14654 project file, if it is specified, and those listed in the file designated by
14655 attribute @code{Local_Configuration_Pragmas} in package @code{Compiler} of
14656 the project file of the source, if it exists.
14658 @node Project Files and Main Subprograms
14659 @subsubsection Project Files and Main Subprograms
14662 When using a project file, you can invoke @command{gnatmake}
14663 with one or several main subprograms, by specifying their source files on the
14667 gnatmake ^-P^/PROJECT_FILE=^prj main1 main2 main3
14671 Each of these needs to be a source file of the same project, except
14672 when the switch ^-u^/UNIQUE^ is used.
14675 When ^-u^/UNIQUE^ is not used, all the mains need to be sources of the
14676 same project, one of the project in the tree rooted at the project specified
14677 on the command line. The package @code{Builder} of this common project, the
14678 "main project" is the one that is considered by @command{gnatmake}.
14681 When ^-u^/UNIQUE^ is used, the specified source files may be in projects
14682 imported directly or indirectly by the project specified on the command line.
14683 Note that if such a source file is not part of the project specified on the
14684 command line, the ^switches^switches^ found in package @code{Builder} of the
14685 project specified on the command line, if any, that are transmitted
14686 to the compiler will still be used, not those found in the project file of
14690 When using a project file, you can also invoke @command{gnatmake} without
14691 explicitly specifying any main, and the effect depends on whether you have
14692 defined the @code{Main} attribute. This attribute has a string list value,
14693 where each element in the list is the name of a source file (the file
14694 extension is optional) that contains a unit that can be a main subprogram.
14696 If the @code{Main} attribute is defined in a project file as a non-empty
14697 string list and the switch @option{^-u^/UNIQUE^} is not used on the command
14698 line, then invoking @command{gnatmake} with this project file but without any
14699 main on the command line is equivalent to invoking @command{gnatmake} with all
14700 the file names in the @code{Main} attribute on the command line.
14703 @smallexample @c projectfile
14706 for Main use ("main1", "main2", "main3");
14712 With this project file, @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^"}
14714 @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^ main1 main2 main3"}.
14716 When the project attribute @code{Main} is not specified, or is specified
14717 as an empty string list, or when the switch @option{-u} is used on the command
14718 line, then invoking @command{gnatmake} with no main on the command line will
14719 result in all immediate sources of the project file being checked, and
14720 potentially recompiled. Depending on the presence of the switch @option{-u},
14721 sources from other project files on which the immediate sources of the main
14722 project file depend are also checked and potentially recompiled. In other
14723 words, the @option{-u} switch is applied to all of the immediate sources of the
14726 When no main is specified on the command line and attribute @code{Main} exists
14727 and includes several mains, or when several mains are specified on the
14728 command line, the default ^switches^switches^ in package @code{Builder} will
14729 be used for all mains, even if there are specific ^switches^switches^
14730 specified for one or several mains.
14732 But the ^switches^switches^ from package @code{Binder} or @code{Linker} will be
14733 the specific ^switches^switches^ for each main, if they are specified.
14735 @node Library Project Files
14736 @subsubsection Library Project Files
14739 When @command{gnatmake} is invoked with a main project file that is a library
14740 project file, it is not allowed to specify one or more mains on the command
14744 When a library project file is specified, switches ^-b^/ACTION=BIND^ and
14745 ^-l^/ACTION=LINK^ have special meanings.
14748 @item ^-b^/ACTION=BIND^ is only allowed for stand-alone libraries. It indicates
14749 to @command{gnatmake} that @command{gnatbind} should be invoked for the
14752 @item ^-l^/ACTION=LINK^ may be used for all library projects. It indicates
14753 to @command{gnatmake} that the binder generated file should be compiled
14754 (in the case of a stand-alone library) and that the library should be built.
14758 @node The GNAT Driver and Project Files
14759 @subsection The GNAT Driver and Project Files
14762 A number of GNAT tools, other than @command{^gnatmake^gnatmake^}
14763 can benefit from project files:
14764 @command{^gnatbind^gnatbind^},
14765 @command{^gnatcheck^gnatcheck^}),
14766 @command{^gnatclean^gnatclean^}),
14767 @command{^gnatelim^gnatelim^},
14768 @command{^gnatfind^gnatfind^},
14769 @command{^gnatlink^gnatlink^},
14770 @command{^gnatls^gnatls^},
14771 @command{^gnatmetric^gnatmetric^},
14772 @command{^gnatpp^gnatpp^},
14773 @command{^gnatstub^gnatstub^},
14774 and @command{^gnatxref^gnatxref^}. However, none of these tools can be invoked
14775 directly with a project file switch (@option{^-P^/PROJECT_FILE=^}).
14776 They must be invoked through the @command{gnat} driver.
14778 The @command{gnat} driver is a wrapper that accepts a number of commands and
14779 calls the corresponding tool. It was designed initially for VMS platforms (to
14780 convert VMS qualifiers to Unix-style switches), but it is now available on all
14783 On non-VMS platforms, the @command{gnat} driver accepts the following commands
14784 (case insensitive):
14788 BIND to invoke @command{^gnatbind^gnatbind^}
14790 CHOP to invoke @command{^gnatchop^gnatchop^}
14792 CLEAN to invoke @command{^gnatclean^gnatclean^}
14794 COMP or COMPILE to invoke the compiler
14796 ELIM to invoke @command{^gnatelim^gnatelim^}
14798 FIND to invoke @command{^gnatfind^gnatfind^}
14800 KR or KRUNCH to invoke @command{^gnatkr^gnatkr^}
14802 LINK to invoke @command{^gnatlink^gnatlink^}
14804 LS or LIST to invoke @command{^gnatls^gnatls^}
14806 MAKE to invoke @command{^gnatmake^gnatmake^}
14808 NAME to invoke @command{^gnatname^gnatname^}
14810 PREP or PREPROCESS to invoke @command{^gnatprep^gnatprep^}
14812 PP or PRETTY to invoke @command{^gnatpp^gnatpp^}
14814 METRIC to invoke @command{^gnatmetric^gnatmetric^}
14816 STUB to invoke @command{^gnatstub^gnatstub^}
14818 XREF to invoke @command{^gnatxref^gnatxref^}
14822 (note that the compiler is invoked using the command
14823 @command{^gnatmake -f -u -c^gnatmake -f -u -c^}).
14826 On non-VMS platforms, between @command{gnat} and the command, two
14827 special switches may be used:
14831 @command{-v} to display the invocation of the tool.
14833 @command{-dn} to prevent the @command{gnat} driver from removing
14834 the temporary files it has created. These temporary files are
14835 configuration files and temporary file list files.
14839 The command may be followed by switches and arguments for the invoked
14843 gnat bind -C main.ali
14849 Switches may also be put in text files, one switch per line, and the text
14850 files may be specified with their path name preceded by '@@'.
14853 gnat bind @@args.txt main.ali
14857 In addition, for commands BIND, COMP or COMPILE, FIND, ELIM, LS or LIST, LINK,
14858 METRIC, PP or PRETTY, STUB and XREF, the project file related switches
14859 (@option{^-P^/PROJECT_FILE^},
14860 @option{^-X^/EXTERNAL_REFERENCE^} and
14861 @option{^-vP^/MESSAGES_PROJECT_FILE=^x}) may be used in addition to
14862 the switches of the invoking tool.
14865 When GNAT PP or GNAT PRETTY is used with a project file, but with no source
14866 specified on the command line, it invokes @command{^gnatpp^gnatpp^} with all
14867 the immediate sources of the specified project file.
14870 When GNAT METRIC is used with a project file, but with no source
14871 specified on the command line, it invokes @command{^gnatmetric^gnatmetric^}
14872 with all the immediate sources of the specified project file and with
14873 @option{^-d^/DIRECTORY^} with the parameter pointing to the object directory
14877 In addition, when GNAT PP, GNAT PRETTY or GNAT METRIC is used with
14878 a project file, no source is specified on the command line and
14879 switch ^-U^/ALL_PROJECTS^ is specified on the command line, then
14880 the underlying tool (^gnatpp^gnatpp^ or
14881 ^gnatmetric^gnatmetric^) is invoked for all sources of all projects,
14882 not only for the immediate sources of the main project.
14884 (-U stands for Universal or Union of the project files of the project tree)
14888 For each of the following commands, there is optionally a corresponding
14889 package in the main project.
14893 package @code{Binder} for command BIND (invoking @code{^gnatbind^gnatbind^})
14896 package @code{Check} for command CHECK (invoking
14897 @code{^gnatcheck^gnatcheck^})
14900 package @code{Compiler} for command COMP or COMPILE (invoking the compiler)
14903 package @code{Cross_Reference} for command XREF (invoking
14904 @code{^gnatxref^gnatxref^})
14907 package @code{Eliminate} for command ELIM (invoking
14908 @code{^gnatelim^gnatelim^})
14911 package @code{Finder} for command FIND (invoking @code{^gnatfind^gnatfind^})
14914 package @code{Gnatls} for command LS or LIST (invoking @code{^gnatls^gnatls^})
14917 package @code{Gnatstub} for command STUB
14918 (invoking @code{^gnatstub^gnatstub^})
14921 package @code{Linker} for command LINK (invoking @code{^gnatlink^gnatlink^})
14924 package @code{Metrics} for command METRIC
14925 (invoking @code{^gnatmetric^gnatmetric^})
14928 package @code{Pretty_Printer} for command PP or PRETTY
14929 (invoking @code{^gnatpp^gnatpp^})
14934 Package @code{Gnatls} has a unique attribute @code{^Switches^Switches^},
14935 a simple variable with a string list value. It contains ^switches^switches^
14936 for the invocation of @code{^gnatls^gnatls^}.
14938 @smallexample @c projectfile
14942 for ^Switches^Switches^
14951 All other packages have two attribute @code{^Switches^Switches^} and
14952 @code{^Default_Switches^Default_Switches^}.
14955 @code{^Switches^Switches^} is an associative array attribute, indexed by the
14956 source file name, that has a string list value: the ^switches^switches^ to be
14957 used when the tool corresponding to the package is invoked for the specific
14961 @code{^Default_Switches^Default_Switches^} is an associative array attribute,
14962 indexed by the programming language that has a string list value.
14963 @code{^Default_Switches^Default_Switches^ ("Ada")} contains the
14964 ^switches^switches^ for the invocation of the tool corresponding
14965 to the package, except if a specific @code{^Switches^Switches^} attribute
14966 is specified for the source file.
14968 @smallexample @c projectfile
14972 for Source_Dirs use ("./**");
14975 for ^Switches^Switches^ use
14982 package Compiler is
14983 for ^Default_Switches^Default_Switches^ ("Ada")
14984 use ("^-gnatv^-gnatv^",
14985 "^-gnatwa^-gnatwa^");
14991 for ^Default_Switches^Default_Switches^ ("Ada")
14999 for ^Default_Switches^Default_Switches^ ("Ada")
15001 for ^Switches^Switches^ ("main.adb")
15010 for ^Default_Switches^Default_Switches^ ("Ada")
15017 package Cross_Reference is
15018 for ^Default_Switches^Default_Switches^ ("Ada")
15023 end Cross_Reference;
15029 With the above project file, commands such as
15032 ^gnat comp -Pproj main^GNAT COMP /PROJECT_FILE=PROJ MAIN^
15033 ^gnat ls -Pproj main^GNAT LIST /PROJECT_FILE=PROJ MAIN^
15034 ^gnat xref -Pproj main^GNAT XREF /PROJECT_FILE=PROJ MAIN^
15035 ^gnat bind -Pproj main.ali^GNAT BIND /PROJECT_FILE=PROJ MAIN.ALI^
15036 ^gnat link -Pproj main.ali^GNAT LINK /PROJECT_FILE=PROJ MAIN.ALI^
15040 will set up the environment properly and invoke the tool with the switches
15041 found in the package corresponding to the tool:
15042 @code{^Default_Switches^Default_Switches^ ("Ada")} for all tools,
15043 except @code{^Switches^Switches^ ("main.adb")}
15044 for @code{^gnatlink^gnatlink^}.
15045 It is also possible to invoke some of the tools,
15046 @code{^gnatcheck^gnatcheck^}),
15047 @code{^gnatmetric^gnatmetric^}),
15048 and @code{^gnatpp^gnatpp^})
15049 on a set of project units thanks to the combination of the switches
15050 @option{-P}, @option{-U} and possibly the main unit when one is interested
15051 in its closure. For instance,
15055 will compute the metrics for all the immediate units of project
15058 gnat metric -Pproj -U
15060 will compute the metrics for all the units of the closure of projects
15061 rooted at @code{proj}.
15063 gnat metric -Pproj -U main_unit
15065 will compute the metrics for the closure of units rooted at
15066 @code{main_unit}. This last possibility relies implicitly
15067 on @command{gnatbind}'s option @option{-R}.
15069 @c **********************
15070 @node An Extended Example
15071 @section An Extended Example
15074 Suppose that we have two programs, @var{prog1} and @var{prog2},
15075 whose sources are in corresponding directories. We would like
15076 to build them with a single @command{gnatmake} command, and we want to place
15077 their object files into @file{build} subdirectories of the source directories.
15078 Furthermore, we want to have to have two separate subdirectories
15079 in @file{build} -- @file{release} and @file{debug} -- which will contain
15080 the object files compiled with different set of compilation flags.
15082 In other words, we have the following structure:
15099 Here are the project files that we must place in a directory @file{main}
15100 to maintain this structure:
15104 @item We create a @code{Common} project with a package @code{Compiler} that
15105 specifies the compilation ^switches^switches^:
15110 @b{project} Common @b{is}
15112 @b{for} Source_Dirs @b{use} (); -- No source files
15116 @b{type} Build_Type @b{is} ("release", "debug");
15117 Build : Build_Type := External ("BUILD", "debug");
15120 @b{package} Compiler @b{is}
15121 @b{case} Build @b{is}
15122 @b{when} "release" =>
15123 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
15124 @b{use} ("^-O2^-O2^");
15125 @b{when} "debug" =>
15126 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
15127 @b{use} ("^-g^-g^");
15135 @item We create separate projects for the two programs:
15142 @b{project} Prog1 @b{is}
15144 @b{for} Source_Dirs @b{use} ("prog1");
15145 @b{for} Object_Dir @b{use} "prog1/build/" & Common.Build;
15147 @b{package} Compiler @b{renames} Common.Compiler;
15158 @b{project} Prog2 @b{is}
15160 @b{for} Source_Dirs @b{use} ("prog2");
15161 @b{for} Object_Dir @b{use} "prog2/build/" & Common.Build;
15163 @b{package} Compiler @b{renames} Common.Compiler;
15169 @item We create a wrapping project @code{Main}:
15178 @b{project} Main @b{is}
15180 @b{package} Compiler @b{renames} Common.Compiler;
15186 @item Finally we need to create a dummy procedure that @code{with}s (either
15187 explicitly or implicitly) all the sources of our two programs.
15192 Now we can build the programs using the command
15195 gnatmake ^-P^/PROJECT_FILE=^main dummy
15199 for the Debug mode, or
15203 gnatmake -Pmain -XBUILD=release
15209 GNAT MAKE /PROJECT_FILE=main /EXTERNAL_REFERENCE=BUILD=release
15214 for the Release mode.
15216 @c ********************************
15217 @c * Project File Complete Syntax *
15218 @c ********************************
15220 @node Project File Complete Syntax
15221 @section Project File Complete Syntax
15225 context_clause project_declaration
15231 @b{with} path_name @{ , path_name @} ;
15236 project_declaration ::=
15237 simple_project_declaration | project_extension
15239 simple_project_declaration ::=
15240 @b{project} <project_>simple_name @b{is}
15241 @{declarative_item@}
15242 @b{end} <project_>simple_name;
15244 project_extension ::=
15245 @b{project} <project_>simple_name @b{extends} path_name @b{is}
15246 @{declarative_item@}
15247 @b{end} <project_>simple_name;
15249 declarative_item ::=
15250 package_declaration |
15251 typed_string_declaration |
15252 other_declarative_item
15254 package_declaration ::=
15255 package_spec | package_renaming
15258 @b{package} package_identifier @b{is}
15259 @{simple_declarative_item@}
15260 @b{end} package_identifier ;
15262 package_identifier ::=
15263 @code{Naming} | @code{Builder} | @code{Compiler} | @code{Binder} |
15264 @code{Linker} | @code{Finder} | @code{Cross_Reference} |
15265 @code{^gnatls^gnatls^} | @code{IDE} | @code{Pretty_Printer}
15267 package_renaming ::==
15268 @b{package} package_identifier @b{renames}
15269 <project_>simple_name.package_identifier ;
15271 typed_string_declaration ::=
15272 @b{type} <typed_string_>_simple_name @b{is}
15273 ( string_literal @{, string_literal@} );
15275 other_declarative_item ::=
15276 attribute_declaration |
15277 typed_variable_declaration |
15278 variable_declaration |
15281 attribute_declaration ::=
15282 full_associative_array_declaration |
15283 @b{for} attribute_designator @b{use} expression ;
15285 full_associative_array_declaration ::=
15286 @b{for} <associative_array_attribute_>simple_name @b{use}
15287 <project_>simple_name [ . <package_>simple_Name ] ' <attribute_>simple_name ;
15289 attribute_designator ::=
15290 <simple_attribute_>simple_name |
15291 <associative_array_attribute_>simple_name ( string_literal )
15293 typed_variable_declaration ::=
15294 <typed_variable_>simple_name : <typed_string_>name := string_expression ;
15296 variable_declaration ::=
15297 <variable_>simple_name := expression;
15307 attribute_reference
15313 ( <string_>expression @{ , <string_>expression @} )
15316 @b{external} ( string_literal [, string_literal] )
15318 attribute_reference ::=
15319 attribute_prefix ' <simple_attribute_>simple_name [ ( literal_string ) ]
15321 attribute_prefix ::=
15323 <project_>simple_name | package_identifier |
15324 <project_>simple_name . package_identifier
15326 case_construction ::=
15327 @b{case} <typed_variable_>name @b{is}
15332 @b{when} discrete_choice_list =>
15333 @{case_construction | attribute_declaration@}
15335 discrete_choice_list ::=
15336 string_literal @{| string_literal@} |
15340 simple_name @{. simple_name@}
15343 identifier (same as Ada)
15347 @node The Cross-Referencing Tools gnatxref and gnatfind
15348 @chapter The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
15353 The compiler generates cross-referencing information (unless
15354 you set the @samp{-gnatx} switch), which are saved in the @file{.ali} files.
15355 This information indicates where in the source each entity is declared and
15356 referenced. Note that entities in package Standard are not included, but
15357 entities in all other predefined units are included in the output.
15359 Before using any of these two tools, you need to compile successfully your
15360 application, so that GNAT gets a chance to generate the cross-referencing
15363 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
15364 information to provide the user with the capability to easily locate the
15365 declaration and references to an entity. These tools are quite similar,
15366 the difference being that @code{gnatfind} is intended for locating
15367 definitions and/or references to a specified entity or entities, whereas
15368 @code{gnatxref} is oriented to generating a full report of all
15371 To use these tools, you must not compile your application using the
15372 @option{-gnatx} switch on the @command{gnatmake} command line
15373 (@pxref{The GNAT Make Program gnatmake}). Otherwise, cross-referencing
15374 information will not be generated.
15376 Note: to invoke @code{gnatxref} or @code{gnatfind} with a project file,
15377 use the @code{gnat} driver (see @ref{The GNAT Driver and Project Files}).
15380 * gnatxref Switches::
15381 * gnatfind Switches::
15382 * Project Files for gnatxref and gnatfind::
15383 * Regular Expressions in gnatfind and gnatxref::
15384 * Examples of gnatxref Usage::
15385 * Examples of gnatfind Usage::
15388 @node gnatxref Switches
15389 @section @code{gnatxref} Switches
15392 The command invocation for @code{gnatxref} is:
15394 $ gnatxref @ovar{switches} @var{sourcefile1} @r{[}@var{sourcefile2} @dots{}@r{]}
15403 identifies the source files for which a report is to be generated. The
15404 ``with''ed units will be processed too. You must provide at least one file.
15406 These file names are considered to be regular expressions, so for instance
15407 specifying @file{source*.adb} is the same as giving every file in the current
15408 directory whose name starts with @file{source} and whose extension is
15411 You shouldn't specify any directory name, just base names. @command{gnatxref}
15412 and @command{gnatfind} will be able to locate these files by themselves using
15413 the source path. If you specify directories, no result is produced.
15418 The switches can be:
15422 @cindex @option{--version} @command{gnatxref}
15423 Display Copyright and version, then exit disregarding all other options.
15426 @cindex @option{--help} @command{gnatxref}
15427 If @option{--version} was not used, display usage, then exit disregarding
15430 @item ^-a^/ALL_FILES^
15431 @cindex @option{^-a^/ALL_FILES^} (@command{gnatxref})
15432 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
15433 the read-only files found in the library search path. Otherwise, these files
15434 will be ignored. This option can be used to protect Gnat sources or your own
15435 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
15436 much faster, and their output much smaller. Read-only here refers to access
15437 or permissions status in the file system for the current user.
15440 @cindex @option{-aIDIR} (@command{gnatxref})
15441 When looking for source files also look in directory DIR. The order in which
15442 source file search is undertaken is the same as for @command{gnatmake}.
15445 @cindex @option{-aODIR} (@command{gnatxref})
15446 When searching for library and object files, look in directory
15447 DIR. The order in which library files are searched is the same as for
15448 @command{gnatmake}.
15451 @cindex @option{-nostdinc} (@command{gnatxref})
15452 Do not look for sources in the system default directory.
15455 @cindex @option{-nostdlib} (@command{gnatxref})
15456 Do not look for library files in the system default directory.
15458 @item --RTS=@var{rts-path}
15459 @cindex @option{--RTS} (@command{gnatxref})
15460 Specifies the default location of the runtime library. Same meaning as the
15461 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
15463 @item ^-d^/DERIVED_TYPES^
15464 @cindex @option{^-d^/DERIVED_TYPES^} (@command{gnatxref})
15465 If this switch is set @code{gnatxref} will output the parent type
15466 reference for each matching derived types.
15468 @item ^-f^/FULL_PATHNAME^
15469 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatxref})
15470 If this switch is set, the output file names will be preceded by their
15471 directory (if the file was found in the search path). If this switch is
15472 not set, the directory will not be printed.
15474 @item ^-g^/IGNORE_LOCALS^
15475 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatxref})
15476 If this switch is set, information is output only for library-level
15477 entities, ignoring local entities. The use of this switch may accelerate
15478 @code{gnatfind} and @code{gnatxref}.
15481 @cindex @option{-IDIR} (@command{gnatxref})
15482 Equivalent to @samp{-aODIR -aIDIR}.
15485 @cindex @option{-pFILE} (@command{gnatxref})
15486 Specify a project file to use @xref{Project Files}.
15487 If you need to use the @file{.gpr}
15488 project files, you should use gnatxref through the GNAT driver
15489 (@command{gnat xref -Pproject}).
15491 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
15492 project file in the current directory.
15494 If a project file is either specified or found by the tools, then the content
15495 of the source directory and object directory lines are added as if they
15496 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^}
15497 and @samp{^-aO^OBJECT_SEARCH^}.
15499 Output only unused symbols. This may be really useful if you give your
15500 main compilation unit on the command line, as @code{gnatxref} will then
15501 display every unused entity and 'with'ed package.
15505 Instead of producing the default output, @code{gnatxref} will generate a
15506 @file{tags} file that can be used by vi. For examples how to use this
15507 feature, see @ref{Examples of gnatxref Usage}. The tags file is output
15508 to the standard output, thus you will have to redirect it to a file.
15514 All these switches may be in any order on the command line, and may even
15515 appear after the file names. They need not be separated by spaces, thus
15516 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
15517 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
15519 @node gnatfind Switches
15520 @section @code{gnatfind} Switches
15523 The command line for @code{gnatfind} is:
15526 $ gnatfind @ovar{switches} @var{pattern}@r{[}:@var{sourcefile}@r{[}:@var{line}@r{[}:@var{column}@r{]]]}
15527 @r{[}@var{file1} @var{file2} @dots{}]
15535 An entity will be output only if it matches the regular expression found
15536 in @var{pattern}, see @ref{Regular Expressions in gnatfind and gnatxref}.
15538 Omitting the pattern is equivalent to specifying @samp{*}, which
15539 will match any entity. Note that if you do not provide a pattern, you
15540 have to provide both a sourcefile and a line.
15542 Entity names are given in Latin-1, with uppercase/lowercase equivalence
15543 for matching purposes. At the current time there is no support for
15544 8-bit codes other than Latin-1, or for wide characters in identifiers.
15547 @code{gnatfind} will look for references, bodies or declarations
15548 of symbols referenced in @file{@var{sourcefile}}, at line @var{line}
15549 and column @var{column}. See @ref{Examples of gnatfind Usage}
15550 for syntax examples.
15553 is a decimal integer identifying the line number containing
15554 the reference to the entity (or entities) to be located.
15557 is a decimal integer identifying the exact location on the
15558 line of the first character of the identifier for the
15559 entity reference. Columns are numbered from 1.
15561 @item file1 file2 @dots{}
15562 The search will be restricted to these source files. If none are given, then
15563 the search will be done for every library file in the search path.
15564 These file must appear only after the pattern or sourcefile.
15566 These file names are considered to be regular expressions, so for instance
15567 specifying @file{source*.adb} is the same as giving every file in the current
15568 directory whose name starts with @file{source} and whose extension is
15571 The location of the spec of the entity will always be displayed, even if it
15572 isn't in one of @file{@var{file1}}, @file{@var{file2}},@enddots{} The
15573 occurrences of the entity in the separate units of the ones given on the
15574 command line will also be displayed.
15576 Note that if you specify at least one file in this part, @code{gnatfind} may
15577 sometimes not be able to find the body of the subprograms.
15582 At least one of 'sourcefile' or 'pattern' has to be present on
15585 The following switches are available:
15589 @cindex @option{--version} @command{gnatfind}
15590 Display Copyright and version, then exit disregarding all other options.
15593 @cindex @option{--help} @command{gnatfind}
15594 If @option{--version} was not used, display usage, then exit disregarding
15597 @item ^-a^/ALL_FILES^
15598 @cindex @option{^-a^/ALL_FILES^} (@command{gnatfind})
15599 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
15600 the read-only files found in the library search path. Otherwise, these files
15601 will be ignored. This option can be used to protect Gnat sources or your own
15602 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
15603 much faster, and their output much smaller. Read-only here refers to access
15604 or permission status in the file system for the current user.
15607 @cindex @option{-aIDIR} (@command{gnatfind})
15608 When looking for source files also look in directory DIR. The order in which
15609 source file search is undertaken is the same as for @command{gnatmake}.
15612 @cindex @option{-aODIR} (@command{gnatfind})
15613 When searching for library and object files, look in directory
15614 DIR. The order in which library files are searched is the same as for
15615 @command{gnatmake}.
15618 @cindex @option{-nostdinc} (@command{gnatfind})
15619 Do not look for sources in the system default directory.
15622 @cindex @option{-nostdlib} (@command{gnatfind})
15623 Do not look for library files in the system default directory.
15625 @item --ext=@var{extension}
15626 @cindex @option{--ext} (@command{gnatfind})
15627 Specify an alternate ali file extension. The default is @code{ali} and other
15628 extensions (e.g. @code{sli} for SPARK library files) may be specified via this
15629 switch. Note that if this switch overrides the default, which means that only
15630 the new extension will be considered.
15632 @item --RTS=@var{rts-path}
15633 @cindex @option{--RTS} (@command{gnatfind})
15634 Specifies the default location of the runtime library. Same meaning as the
15635 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
15637 @item ^-d^/DERIVED_TYPE_INFORMATION^
15638 @cindex @option{^-d^/DERIVED_TYPE_INFORMATION^} (@code{gnatfind})
15639 If this switch is set, then @code{gnatfind} will output the parent type
15640 reference for each matching derived types.
15642 @item ^-e^/EXPRESSIONS^
15643 @cindex @option{^-e^/EXPRESSIONS^} (@command{gnatfind})
15644 By default, @code{gnatfind} accept the simple regular expression set for
15645 @samp{pattern}. If this switch is set, then the pattern will be
15646 considered as full Unix-style regular expression.
15648 @item ^-f^/FULL_PATHNAME^
15649 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatfind})
15650 If this switch is set, the output file names will be preceded by their
15651 directory (if the file was found in the search path). If this switch is
15652 not set, the directory will not be printed.
15654 @item ^-g^/IGNORE_LOCALS^
15655 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatfind})
15656 If this switch is set, information is output only for library-level
15657 entities, ignoring local entities. The use of this switch may accelerate
15658 @code{gnatfind} and @code{gnatxref}.
15661 @cindex @option{-IDIR} (@command{gnatfind})
15662 Equivalent to @samp{-aODIR -aIDIR}.
15665 @cindex @option{-pFILE} (@command{gnatfind})
15666 Specify a project file (@pxref{Project Files}) to use.
15667 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
15668 project file in the current directory.
15670 If a project file is either specified or found by the tools, then the content
15671 of the source directory and object directory lines are added as if they
15672 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^} and
15673 @samp{^-aO^/OBJECT_SEARCH^}.
15675 @item ^-r^/REFERENCES^
15676 @cindex @option{^-r^/REFERENCES^} (@command{gnatfind})
15677 By default, @code{gnatfind} will output only the information about the
15678 declaration, body or type completion of the entities. If this switch is
15679 set, the @code{gnatfind} will locate every reference to the entities in
15680 the files specified on the command line (or in every file in the search
15681 path if no file is given on the command line).
15683 @item ^-s^/PRINT_LINES^
15684 @cindex @option{^-s^/PRINT_LINES^} (@command{gnatfind})
15685 If this switch is set, then @code{gnatfind} will output the content
15686 of the Ada source file lines were the entity was found.
15688 @item ^-t^/TYPE_HIERARCHY^
15689 @cindex @option{^-t^/TYPE_HIERARCHY^} (@command{gnatfind})
15690 If this switch is set, then @code{gnatfind} will output the type hierarchy for
15691 the specified type. It act like -d option but recursively from parent
15692 type to parent type. When this switch is set it is not possible to
15693 specify more than one file.
15698 All these switches may be in any order on the command line, and may even
15699 appear after the file names. They need not be separated by spaces, thus
15700 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
15701 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
15703 As stated previously, gnatfind will search in every directory in the
15704 search path. You can force it to look only in the current directory if
15705 you specify @code{*} at the end of the command line.
15707 @node Project Files for gnatxref and gnatfind
15708 @section Project Files for @command{gnatxref} and @command{gnatfind}
15711 Project files allow a programmer to specify how to compile its
15712 application, where to find sources, etc. These files are used
15714 primarily by GPS, but they can also be used
15717 @code{gnatxref} and @code{gnatfind}.
15719 A project file name must end with @file{.gpr}. If a single one is
15720 present in the current directory, then @code{gnatxref} and @code{gnatfind} will
15721 extract the information from it. If multiple project files are found, none of
15722 them is read, and you have to use the @samp{-p} switch to specify the one
15725 The following lines can be included, even though most of them have default
15726 values which can be used in most cases.
15727 The lines can be entered in any order in the file.
15728 Except for @file{src_dir} and @file{obj_dir}, you can only have one instance of
15729 each line. If you have multiple instances, only the last one is taken into
15734 [default: @code{"^./^[]^"}]
15735 specifies a directory where to look for source files. Multiple @code{src_dir}
15736 lines can be specified and they will be searched in the order they
15740 [default: @code{"^./^[]^"}]
15741 specifies a directory where to look for object and library files. Multiple
15742 @code{obj_dir} lines can be specified, and they will be searched in the order
15745 @item comp_opt=SWITCHES
15746 [default: @code{""}]
15747 creates a variable which can be referred to subsequently by using
15748 the @code{$@{comp_opt@}} notation. This is intended to store the default
15749 switches given to @command{gnatmake} and @command{gcc}.
15751 @item bind_opt=SWITCHES
15752 [default: @code{""}]
15753 creates a variable which can be referred to subsequently by using
15754 the @samp{$@{bind_opt@}} notation. This is intended to store the default
15755 switches given to @command{gnatbind}.
15757 @item link_opt=SWITCHES
15758 [default: @code{""}]
15759 creates a variable which can be referred to subsequently by using
15760 the @samp{$@{link_opt@}} notation. This is intended to store the default
15761 switches given to @command{gnatlink}.
15763 @item main=EXECUTABLE
15764 [default: @code{""}]
15765 specifies the name of the executable for the application. This variable can
15766 be referred to in the following lines by using the @samp{$@{main@}} notation.
15769 @item comp_cmd=COMMAND
15770 [default: @code{"GNAT COMPILE /SEARCH=$@{src_dir@} /DEBUG /TRY_SEMANTICS"}]
15773 @item comp_cmd=COMMAND
15774 [default: @code{"gcc -c -I$@{src_dir@} -g -gnatq"}]
15776 specifies the command used to compile a single file in the application.
15779 @item make_cmd=COMMAND
15780 [default: @code{"GNAT MAKE $@{main@}
15781 /SOURCE_SEARCH=$@{src_dir@} /OBJECT_SEARCH=$@{obj_dir@}
15782 /DEBUG /TRY_SEMANTICS /COMPILER_QUALIFIERS $@{comp_opt@}
15783 /BINDER_QUALIFIERS $@{bind_opt@} /LINKER_QUALIFIERS $@{link_opt@}"}]
15786 @item make_cmd=COMMAND
15787 [default: @code{"gnatmake $@{main@} -aI$@{src_dir@}
15788 -aO$@{obj_dir@} -g -gnatq -cargs $@{comp_opt@}
15789 -bargs $@{bind_opt@} -largs $@{link_opt@}"}]
15791 specifies the command used to recompile the whole application.
15793 @item run_cmd=COMMAND
15794 [default: @code{"$@{main@}"}]
15795 specifies the command used to run the application.
15797 @item debug_cmd=COMMAND
15798 [default: @code{"gdb $@{main@}"}]
15799 specifies the command used to debug the application
15804 @command{gnatxref} and @command{gnatfind} only take into account the
15805 @code{src_dir} and @code{obj_dir} lines, and ignore the others.
15807 @node Regular Expressions in gnatfind and gnatxref
15808 @section Regular Expressions in @code{gnatfind} and @code{gnatxref}
15811 As specified in the section about @command{gnatfind}, the pattern can be a
15812 regular expression. Actually, there are to set of regular expressions
15813 which are recognized by the program:
15816 @item globbing patterns
15817 These are the most usual regular expression. They are the same that you
15818 generally used in a Unix shell command line, or in a DOS session.
15820 Here is a more formal grammar:
15827 term ::= elmt -- matches elmt
15828 term ::= elmt elmt -- concatenation (elmt then elmt)
15829 term ::= * -- any string of 0 or more characters
15830 term ::= ? -- matches any character
15831 term ::= [char @{char@}] -- matches any character listed
15832 term ::= [char - char] -- matches any character in range
15836 @item full regular expression
15837 The second set of regular expressions is much more powerful. This is the
15838 type of regular expressions recognized by utilities such a @file{grep}.
15840 The following is the form of a regular expression, expressed in Ada
15841 reference manual style BNF is as follows
15848 regexp ::= term @{| term@} -- alternation (term or term @dots{})
15850 term ::= item @{item@} -- concatenation (item then item)
15852 item ::= elmt -- match elmt
15853 item ::= elmt * -- zero or more elmt's
15854 item ::= elmt + -- one or more elmt's
15855 item ::= elmt ? -- matches elmt or nothing
15858 elmt ::= nschar -- matches given character
15859 elmt ::= [nschar @{nschar@}] -- matches any character listed
15860 elmt ::= [^^^ nschar @{nschar@}] -- matches any character not listed
15861 elmt ::= [char - char] -- matches chars in given range
15862 elmt ::= \ char -- matches given character
15863 elmt ::= . -- matches any single character
15864 elmt ::= ( regexp ) -- parens used for grouping
15866 char ::= any character, including special characters
15867 nschar ::= any character except ()[].*+?^^^
15871 Following are a few examples:
15875 will match any of the two strings @samp{abcde} and @samp{fghi},
15878 will match any string like @samp{abd}, @samp{abcd}, @samp{abccd},
15879 @samp{abcccd}, and so on,
15882 will match any string which has only lowercase characters in it (and at
15883 least one character.
15888 @node Examples of gnatxref Usage
15889 @section Examples of @code{gnatxref} Usage
15891 @subsection General Usage
15894 For the following examples, we will consider the following units:
15896 @smallexample @c ada
15902 3: procedure Foo (B : in Integer);
15909 1: package body Main is
15910 2: procedure Foo (B : in Integer) is
15921 2: procedure Print (B : Integer);
15930 The first thing to do is to recompile your application (for instance, in
15931 that case just by doing a @samp{gnatmake main}, so that GNAT generates
15932 the cross-referencing information.
15933 You can then issue any of the following commands:
15935 @item gnatxref main.adb
15936 @code{gnatxref} generates cross-reference information for main.adb
15937 and every unit 'with'ed by main.adb.
15939 The output would be:
15947 Decl: main.ads 3:20
15948 Body: main.adb 2:20
15949 Ref: main.adb 4:13 5:13 6:19
15952 Ref: main.adb 6:8 7:8
15962 Decl: main.ads 3:15
15963 Body: main.adb 2:15
15966 Body: main.adb 1:14
15969 Ref: main.adb 6:12 7:12
15973 that is the entity @code{Main} is declared in main.ads, line 2, column 9,
15974 its body is in main.adb, line 1, column 14 and is not referenced any where.
15976 The entity @code{Print} is declared in bar.ads, line 2, column 15 and it
15977 it referenced in main.adb, line 6 column 12 and line 7 column 12.
15979 @item gnatxref package1.adb package2.ads
15980 @code{gnatxref} will generates cross-reference information for
15981 package1.adb, package2.ads and any other package 'with'ed by any
15987 @subsection Using gnatxref with vi
15989 @code{gnatxref} can generate a tags file output, which can be used
15990 directly from @command{vi}. Note that the standard version of @command{vi}
15991 will not work properly with overloaded symbols. Consider using another
15992 free implementation of @command{vi}, such as @command{vim}.
15995 $ gnatxref -v gnatfind.adb > tags
15999 will generate the tags file for @code{gnatfind} itself (if the sources
16000 are in the search path!).
16002 From @command{vi}, you can then use the command @samp{:tag @var{entity}}
16003 (replacing @var{entity} by whatever you are looking for), and vi will
16004 display a new file with the corresponding declaration of entity.
16007 @node Examples of gnatfind Usage
16008 @section Examples of @code{gnatfind} Usage
16012 @item gnatfind ^-f^/FULL_PATHNAME^ xyz:main.adb
16013 Find declarations for all entities xyz referenced at least once in
16014 main.adb. The references are search in every library file in the search
16017 The directories will be printed as well (as the @samp{^-f^/FULL_PATHNAME^}
16020 The output will look like:
16022 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
16023 ^directory/^[directory]^main.adb:24:10: xyz <= body
16024 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
16028 that is to say, one of the entities xyz found in main.adb is declared at
16029 line 12 of main.ads (and its body is in main.adb), and another one is
16030 declared at line 45 of foo.ads
16032 @item gnatfind ^-fs^/FULL_PATHNAME/SOURCE_LINE^ xyz:main.adb
16033 This is the same command as the previous one, instead @code{gnatfind} will
16034 display the content of the Ada source file lines.
16036 The output will look like:
16039 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
16041 ^directory/^[directory]^main.adb:24:10: xyz <= body
16043 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
16048 This can make it easier to find exactly the location your are looking
16051 @item gnatfind ^-r^/REFERENCES^ "*x*":main.ads:123 foo.adb
16052 Find references to all entities containing an x that are
16053 referenced on line 123 of main.ads.
16054 The references will be searched only in main.ads and foo.adb.
16056 @item gnatfind main.ads:123
16057 Find declarations and bodies for all entities that are referenced on
16058 line 123 of main.ads.
16060 This is the same as @code{gnatfind "*":main.adb:123}.
16062 @item gnatfind ^mydir/^[mydir]^main.adb:123:45
16063 Find the declaration for the entity referenced at column 45 in
16064 line 123 of file main.adb in directory mydir. Note that it
16065 is usual to omit the identifier name when the column is given,
16066 since the column position identifies a unique reference.
16068 The column has to be the beginning of the identifier, and should not
16069 point to any character in the middle of the identifier.
16073 @c *********************************
16074 @node The GNAT Pretty-Printer gnatpp
16075 @chapter The GNAT Pretty-Printer @command{gnatpp}
16077 @cindex Pretty-Printer
16080 ^The @command{gnatpp} tool^GNAT PRETTY^ is an ASIS-based utility
16081 for source reformatting / pretty-printing.
16082 It takes an Ada source file as input and generates a reformatted
16084 You can specify various style directives via switches; e.g.,
16085 identifier case conventions, rules of indentation, and comment layout.
16087 To produce a reformatted file, @command{gnatpp} generates and uses the ASIS
16088 tree for the input source and thus requires the input to be syntactically and
16089 semantically legal.
16090 If this condition is not met, @command{gnatpp} will terminate with an
16091 error message; no output file will be generated.
16093 If the source files presented to @command{gnatpp} contain
16094 preprocessing directives, then the output file will
16095 correspond to the generated source after all
16096 preprocessing is carried out. There is no way
16097 using @command{gnatpp} to obtain pretty printed files that
16098 include the preprocessing directives.
16100 If the compilation unit
16101 contained in the input source depends semantically upon units located
16102 outside the current directory, you have to provide the source search path
16103 when invoking @command{gnatpp}, if these units are contained in files with
16104 names that do not follow the GNAT file naming rules, you have to provide
16105 the configuration file describing the corresponding naming scheme;
16106 see the description of the @command{gnatpp}
16107 switches below. Another possibility is to use a project file and to
16108 call @command{gnatpp} through the @command{gnat} driver
16110 The @command{gnatpp} command has the form
16113 $ gnatpp @ovar{switches} @var{filename}
16120 @var{switches} is an optional sequence of switches defining such properties as
16121 the formatting rules, the source search path, and the destination for the
16125 @var{filename} is the name (including the extension) of the source file to
16126 reformat; ``wildcards'' or several file names on the same gnatpp command are
16127 allowed. The file name may contain path information; it does not have to
16128 follow the GNAT file naming rules
16132 * Switches for gnatpp::
16133 * Formatting Rules::
16136 @node Switches for gnatpp
16137 @section Switches for @command{gnatpp}
16140 The following subsections describe the various switches accepted by
16141 @command{gnatpp}, organized by category.
16144 You specify a switch by supplying a name and generally also a value.
16145 In many cases the values for a switch with a given name are incompatible with
16147 (for example the switch that controls the casing of a reserved word may have
16148 exactly one value: upper case, lower case, or
16149 mixed case) and thus exactly one such switch can be in effect for an
16150 invocation of @command{gnatpp}.
16151 If more than one is supplied, the last one is used.
16152 However, some values for the same switch are mutually compatible.
16153 You may supply several such switches to @command{gnatpp}, but then
16154 each must be specified in full, with both the name and the value.
16155 Abbreviated forms (the name appearing once, followed by each value) are
16157 For example, to set
16158 the alignment of the assignment delimiter both in declarations and in
16159 assignment statements, you must write @option{-A2A3}
16160 (or @option{-A2 -A3}), but not @option{-A23}.
16164 In many cases the set of options for a given qualifier are incompatible with
16165 each other (for example the qualifier that controls the casing of a reserved
16166 word may have exactly one option, which specifies either upper case, lower
16167 case, or mixed case), and thus exactly one such option can be in effect for
16168 an invocation of @command{gnatpp}.
16169 If more than one is supplied, the last one is used.
16170 However, some qualifiers have options that are mutually compatible,
16171 and then you may then supply several such options when invoking
16175 In most cases, it is obvious whether or not the
16176 ^values for a switch with a given name^options for a given qualifier^
16177 are compatible with each other.
16178 When the semantics might not be evident, the summaries below explicitly
16179 indicate the effect.
16182 * Alignment Control::
16184 * Construct Layout Control::
16185 * General Text Layout Control::
16186 * Other Formatting Options::
16187 * Setting the Source Search Path::
16188 * Output File Control::
16189 * Other gnatpp Switches::
16192 @node Alignment Control
16193 @subsection Alignment Control
16194 @cindex Alignment control in @command{gnatpp}
16197 Programs can be easier to read if certain constructs are vertically aligned.
16198 By default all alignments are set ON.
16199 Through the @option{^-A0^/ALIGN=OFF^} switch you may reset the default to
16200 OFF, and then use one or more of the other
16201 ^@option{-A@var{n}} switches^@option{/ALIGN} options^
16202 to activate alignment for specific constructs.
16205 @cindex @option{^-A@var{n}^/ALIGN^} (@command{gnatpp})
16209 Set all alignments to ON
16212 @item ^-A0^/ALIGN=OFF^
16213 Set all alignments to OFF
16215 @item ^-A1^/ALIGN=COLONS^
16216 Align @code{:} in declarations
16218 @item ^-A2^/ALIGN=DECLARATIONS^
16219 Align @code{:=} in initializations in declarations
16221 @item ^-A3^/ALIGN=STATEMENTS^
16222 Align @code{:=} in assignment statements
16224 @item ^-A4^/ALIGN=ARROWS^
16225 Align @code{=>} in associations
16227 @item ^-A5^/ALIGN=COMPONENT_CLAUSES^
16228 Align @code{at} keywords in the component clauses in record
16229 representation clauses
16233 The @option{^-A^/ALIGN^} switches are mutually compatible; any combination
16236 @node Casing Control
16237 @subsection Casing Control
16238 @cindex Casing control in @command{gnatpp}
16241 @command{gnatpp} allows you to specify the casing for reserved words,
16242 pragma names, attribute designators and identifiers.
16243 For identifiers you may define a
16244 general rule for name casing but also override this rule
16245 via a set of dictionary files.
16247 Three types of casing are supported: lower case, upper case, and mixed case.
16248 Lower and upper case are self-explanatory (but since some letters in
16249 Latin1 and other GNAT-supported character sets
16250 exist only in lower-case form, an upper case conversion will have no
16252 ``Mixed case'' means that the first letter, and also each letter immediately
16253 following an underscore, are converted to their uppercase forms;
16254 all the other letters are converted to their lowercase forms.
16257 @cindex @option{^-a@var{x}^/ATTRIBUTE^} (@command{gnatpp})
16258 @item ^-aL^/ATTRIBUTE_CASING=LOWER_CASE^
16259 Attribute designators are lower case
16261 @item ^-aU^/ATTRIBUTE_CASING=UPPER_CASE^
16262 Attribute designators are upper case
16264 @item ^-aM^/ATTRIBUTE_CASING=MIXED_CASE^
16265 Attribute designators are mixed case (this is the default)
16267 @cindex @option{^-k@var{x}^/KEYWORD_CASING^} (@command{gnatpp})
16268 @item ^-kL^/KEYWORD_CASING=LOWER_CASE^
16269 Keywords (technically, these are known in Ada as @emph{reserved words}) are
16270 lower case (this is the default)
16272 @item ^-kU^/KEYWORD_CASING=UPPER_CASE^
16273 Keywords are upper case
16275 @cindex @option{^-n@var{x}^/NAME_CASING^} (@command{gnatpp})
16276 @item ^-nD^/NAME_CASING=AS_DECLARED^
16277 Name casing for defining occurrences are as they appear in the source file
16278 (this is the default)
16280 @item ^-nU^/NAME_CASING=UPPER_CASE^
16281 Names are in upper case
16283 @item ^-nL^/NAME_CASING=LOWER_CASE^
16284 Names are in lower case
16286 @item ^-nM^/NAME_CASING=MIXED_CASE^
16287 Names are in mixed case
16289 @cindex @option{^-p@var{x}^/PRAGMA_CASING^} (@command{gnatpp})
16290 @item ^-pL^/PRAGMA_CASING=LOWER_CASE^
16291 Pragma names are lower case
16293 @item ^-pU^/PRAGMA_CASING=UPPER_CASE^
16294 Pragma names are upper case
16296 @item ^-pM^/PRAGMA_CASING=MIXED_CASE^
16297 Pragma names are mixed case (this is the default)
16299 @item ^-D@var{file}^/DICTIONARY=@var{file}^
16300 @cindex @option{^-D^/DICTIONARY^} (@command{gnatpp})
16301 Use @var{file} as a @emph{dictionary file} that defines
16302 the casing for a set of specified names,
16303 thereby overriding the effect on these names by
16304 any explicit or implicit
16305 ^-n^/NAME_CASING^ switch.
16306 To supply more than one dictionary file,
16307 use ^several @option{-D} switches^a list of files as options^.
16310 @option{gnatpp} implicitly uses a @emph{default dictionary file}
16311 to define the casing for the Ada predefined names and
16312 the names declared in the GNAT libraries.
16314 @item ^-D-^/SPECIFIC_CASING^
16315 @cindex @option{^-D-^/SPECIFIC_CASING^} (@command{gnatpp})
16316 Do not use the default dictionary file;
16317 instead, use the casing
16318 defined by a @option{^-n^/NAME_CASING^} switch and any explicit
16323 The structure of a dictionary file, and details on the conventions
16324 used in the default dictionary file, are defined in @ref{Name Casing}.
16326 The @option{^-D-^/SPECIFIC_CASING^} and
16327 @option{^-D@var{file}^/DICTIONARY=@var{file}^} switches are mutually
16330 @node Construct Layout Control
16331 @subsection Construct Layout Control
16332 @cindex Layout control in @command{gnatpp}
16335 This group of @command{gnatpp} switches controls the layout of comments and
16336 complex syntactic constructs. See @ref{Formatting Comments} for details
16340 @cindex @option{^-c@var{n}^/COMMENTS_LAYOUT^} (@command{gnatpp})
16341 @item ^-c0^/COMMENTS_LAYOUT=UNTOUCHED^
16342 All the comments remain unchanged
16344 @item ^-c1^/COMMENTS_LAYOUT=DEFAULT^
16345 GNAT-style comment line indentation (this is the default).
16347 @item ^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^
16348 Reference-manual comment line indentation.
16350 @item ^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^
16351 GNAT-style comment beginning
16353 @item ^-c4^/COMMENTS_LAYOUT=REFORMAT^
16354 Reformat comment blocks
16356 @item ^-c5^/COMMENTS_LAYOUT=KEEP_SPECIAL^
16357 Keep unchanged special form comments
16359 Reformat comment blocks
16361 @cindex @option{^-l@var{n}^/CONSTRUCT_LAYOUT^} (@command{gnatpp})
16362 @item ^-l1^/CONSTRUCT_LAYOUT=GNAT^
16363 GNAT-style layout (this is the default)
16365 @item ^-l2^/CONSTRUCT_LAYOUT=COMPACT^
16368 @item ^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^
16371 @cindex @option{^-N^/NOTABS^} (@command{gnatpp})
16373 All the VT characters are removed from the comment text. All the HT characters
16374 are expanded with the sequences of space characters to get to the next tab
16377 @cindex @option{^--no-separate-is^/NO_SEPARATE_IS^} (@command{gnatpp})
16378 @item ^--no-separate-is^/NO_SEPARATE_IS^
16379 Do not place the keyword @code{is} on a separate line in a subprogram body in
16380 case if the spec occupies more then one line.
16382 @cindex @option{^--separate-label^/SEPARATE_LABEL^} (@command{gnatpp})
16383 @item ^--separate-label^/SEPARATE_LABEL^
16384 Place statement label(s) on a separate line, with the following statement
16387 @cindex @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} (@command{gnatpp})
16388 @item ^--separate-loop-then^/SEPARATE_LOOP_THEN^
16389 Place the keyword @code{loop} in FOR and WHILE loop statements and the
16390 keyword @code{then} in IF statements on a separate line.
16392 @cindex @option{^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^} (@command{gnatpp})
16393 @item ^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^
16394 Do not place the keyword @code{loop} in FOR and WHILE loop statements and the
16395 keyword @code{then} in IF statements on a separate line. This option is
16396 incompatible with @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} option.
16398 @cindex @option{^--use-on-new-line^/USE_ON_NEW_LINE^} (@command{gnatpp})
16399 @item ^--use-on-new-line^/USE_ON_NEW_LINE^
16400 Start each USE clause in a context clause from a separate line.
16402 @cindex @option{^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^} (@command{gnatpp})
16403 @item ^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^
16404 Use a separate line for a loop or block statement name, but do not use an extra
16405 indentation level for the statement itself.
16411 The @option{-c1} and @option{-c2} switches are incompatible.
16412 The @option{-c3} and @option{-c4} switches are compatible with each other and
16413 also with @option{-c1} and @option{-c2}. The @option{-c0} switch disables all
16414 the other comment formatting switches.
16416 The @option{-l1}, @option{-l2}, and @option{-l3} switches are incompatible.
16421 For the @option{/COMMENTS_LAYOUT} qualifier:
16424 The @option{DEFAULT} and @option{STANDARD_INDENT} options are incompatible.
16426 The @option{GNAT_BEGINNING} and @option{REFORMAT} options are compatible with
16427 each other and also with @option{DEFAULT} and @option{STANDARD_INDENT}.
16431 The @option{GNAT}, @option{COMPACT}, and @option{UNCOMPACT} options for the
16432 @option{/CONSTRUCT_LAYOUT} qualifier are incompatible.
16435 @node General Text Layout Control
16436 @subsection General Text Layout Control
16439 These switches allow control over line length and indentation.
16442 @item ^-M@var{nnn}^/LINE_LENGTH_MAX=@var{nnn}^
16443 @cindex @option{^-M^/LINE_LENGTH^} (@command{gnatpp})
16444 Maximum line length, @var{nnn} from 32@dots{}256, the default value is 79
16446 @item ^-i@var{nnn}^/INDENTATION_LEVEL=@var{nnn}^
16447 @cindex @option{^-i^/INDENTATION_LEVEL^} (@command{gnatpp})
16448 Indentation level, @var{nnn} from 1@dots{}9, the default value is 3
16450 @item ^-cl@var{nnn}^/CONTINUATION_INDENT=@var{nnn}^
16451 @cindex @option{^-cl^/CONTINUATION_INDENT^} (@command{gnatpp})
16452 Indentation level for continuation lines (relative to the line being
16453 continued), @var{nnn} from 1@dots{}9.
16455 value is one less then the (normal) indentation level, unless the
16456 indentation is set to 1 (in which case the default value for continuation
16457 line indentation is also 1)
16460 @node Other Formatting Options
16461 @subsection Other Formatting Options
16464 These switches control the inclusion of missing end/exit labels, and
16465 the indentation level in @b{case} statements.
16468 @item ^-e^/NO_MISSED_LABELS^
16469 @cindex @option{^-e^/NO_MISSED_LABELS^} (@command{gnatpp})
16470 Do not insert missing end/exit labels. An end label is the name of
16471 a construct that may optionally be repeated at the end of the
16472 construct's declaration;
16473 e.g., the names of packages, subprograms, and tasks.
16474 An exit label is the name of a loop that may appear as target
16475 of an exit statement within the loop.
16476 By default, @command{gnatpp} inserts these end/exit labels when
16477 they are absent from the original source. This option suppresses such
16478 insertion, so that the formatted source reflects the original.
16480 @item ^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^
16481 @cindex @option{^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^} (@command{gnatpp})
16482 Insert a Form Feed character after a pragma Page.
16484 @item ^-T@var{nnn}^/MAX_INDENT=@var{nnn}^
16485 @cindex @option{^-T^/MAX_INDENT^} (@command{gnatpp})
16486 Do not use an additional indentation level for @b{case} alternatives
16487 and variants if there are @var{nnn} or more (the default
16489 If @var{nnn} is 0, an additional indentation level is
16490 used for @b{case} alternatives and variants regardless of their number.
16493 @node Setting the Source Search Path
16494 @subsection Setting the Source Search Path
16497 To define the search path for the input source file, @command{gnatpp}
16498 uses the same switches as the GNAT compiler, with the same effects.
16501 @item ^-I^/SEARCH=^@var{dir}
16502 @cindex @option{^-I^/SEARCH^} (@code{gnatpp})
16503 The same as the corresponding gcc switch
16505 @item ^-I-^/NOCURRENT_DIRECTORY^
16506 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatpp})
16507 The same as the corresponding gcc switch
16509 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE^=@var{path}
16510 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@code{gnatpp})
16511 The same as the corresponding gcc switch
16513 @item ^--RTS^/RUNTIME_SYSTEM^=@var{path}
16514 @cindex @option{^--RTS^/RUNTIME_SYSTEM^} (@code{gnatpp})
16515 The same as the corresponding gcc switch
16519 @node Output File Control
16520 @subsection Output File Control
16523 By default the output is sent to the file whose name is obtained by appending
16524 the ^@file{.pp}^@file{$PP}^ suffix to the name of the input file
16525 (if the file with this name already exists, it is unconditionally overwritten).
16526 Thus if the input file is @file{^my_ada_proc.adb^MY_ADA_PROC.ADB^} then
16527 @command{gnatpp} will produce @file{^my_ada_proc.adb.pp^MY_ADA_PROC.ADB$PP^}
16529 The output may be redirected by the following switches:
16532 @item ^-pipe^/STANDARD_OUTPUT^
16533 @cindex @option{^-pipe^/STANDARD_OUTPUT^} (@code{gnatpp})
16534 Send the output to @code{Standard_Output}
16536 @item ^-o @var{output_file}^/OUTPUT=@var{output_file}^
16537 @cindex @option{^-o^/OUTPUT^} (@code{gnatpp})
16538 Write the output into @var{output_file}.
16539 If @var{output_file} already exists, @command{gnatpp} terminates without
16540 reading or processing the input file.
16542 @item ^-of ^/FORCED_OUTPUT=^@var{output_file}
16543 @cindex @option{^-of^/FORCED_OUTPUT^} (@code{gnatpp})
16544 Write the output into @var{output_file}, overwriting the existing file
16545 (if one is present).
16547 @item ^-r^/REPLACE^
16548 @cindex @option{^-r^/REPLACE^} (@code{gnatpp})
16549 Replace the input source file with the reformatted output, and copy the
16550 original input source into the file whose name is obtained by appending the
16551 ^@file{.npp}^@file{$NPP}^ suffix to the name of the input file.
16552 If a file with this name already exists, @command{gnatpp} terminates without
16553 reading or processing the input file.
16555 @item ^-rf^/OVERRIDING_REPLACE^
16556 @cindex @option{^-rf^/OVERRIDING_REPLACE^} (@code{gnatpp})
16557 Like @option{^-r^/REPLACE^} except that if the file with the specified name
16558 already exists, it is overwritten.
16560 @item ^-rnb^/REPLACE_NO_BACKUP^
16561 @cindex @option{^-rnb^/REPLACE_NO_BACKUP^} (@code{gnatpp})
16562 Replace the input source file with the reformatted output without
16563 creating any backup copy of the input source.
16565 @item ^--eol=@var{xxx}^/END_OF_LINE=@var{xxx}^
16566 @cindex @option{^--eol^/END_OF_LINE^} (@code{gnatpp})
16567 Specifies the format of the reformatted output file. The @var{xxx}
16568 ^string specified with the switch^option^ may be either
16570 @item ``@option{^dos^DOS^}'' MS DOS style, lines end with CR LF characters
16571 @item ``@option{^crlf^CRLF^}''
16572 the same as @option{^crlf^CRLF^}
16573 @item ``@option{^unix^UNIX^}'' UNIX style, lines end with LF character
16574 @item ``@option{^lf^LF^}''
16575 the same as @option{^unix^UNIX^}
16578 @item ^-W^/RESULT_ENCODING=^@var{e}
16579 @cindex @option{^-W^/RESULT_ENCODING=^} (@command{gnatpp})
16580 Specify the wide character encoding method used to write the code in the
16582 @var{e} is one of the following:
16590 Upper half encoding
16592 @item ^s^SHIFT_JIS^
16602 Brackets encoding (default value)
16608 Options @option{^-pipe^/STANDARD_OUTPUT^},
16609 @option{^-o^/OUTPUT^} and
16610 @option{^-of^/FORCED_OUTPUT^} are allowed only if the call to gnatpp
16611 contains only one file to reformat.
16613 @option{^--eol^/END_OF_LINE^}
16615 @option{^-W^/RESULT_ENCODING^}
16616 cannot be used together
16617 with @option{^-pipe^/STANDARD_OUTPUT^} option.
16619 @node Other gnatpp Switches
16620 @subsection Other @code{gnatpp} Switches
16623 The additional @command{gnatpp} switches are defined in this subsection.
16626 @item ^-files @var{filename}^/FILES=@var{output_file}^
16627 @cindex @option{^-files^/FILES^} (@code{gnatpp})
16628 Take the argument source files from the specified file. This file should be an
16629 ordinary textual file containing file names separated by spaces or
16630 line breaks. You can use this switch more then once in the same call to
16631 @command{gnatpp}. You also can combine this switch with explicit list of
16634 @item ^-v^/VERBOSE^
16635 @cindex @option{^-v^/VERBOSE^} (@code{gnatpp})
16637 @command{gnatpp} generates version information and then
16638 a trace of the actions it takes to produce or obtain the ASIS tree.
16640 @item ^-w^/WARNINGS^
16641 @cindex @option{^-w^/WARNINGS^} (@code{gnatpp})
16643 @command{gnatpp} generates a warning whenever it cannot provide
16644 a required layout in the result source.
16647 @node Formatting Rules
16648 @section Formatting Rules
16651 The following subsections show how @command{gnatpp} treats ``white space'',
16652 comments, program layout, and name casing.
16653 They provide the detailed descriptions of the switches shown above.
16656 * White Space and Empty Lines::
16657 * Formatting Comments::
16658 * Construct Layout::
16662 @node White Space and Empty Lines
16663 @subsection White Space and Empty Lines
16666 @command{gnatpp} does not have an option to control space characters.
16667 It will add or remove spaces according to the style illustrated by the
16668 examples in the @cite{Ada Reference Manual}.
16670 The only format effectors
16671 (see @cite{Ada Reference Manual}, paragraph 2.1(13))
16672 that will appear in the output file are platform-specific line breaks,
16673 and also format effectors within (but not at the end of) comments.
16674 In particular, each horizontal tab character that is not inside
16675 a comment will be treated as a space and thus will appear in the
16676 output file as zero or more spaces depending on
16677 the reformatting of the line in which it appears.
16678 The only exception is a Form Feed character, which is inserted after a
16679 pragma @code{Page} when @option{-ff} is set.
16681 The output file will contain no lines with trailing ``white space'' (spaces,
16684 Empty lines in the original source are preserved
16685 only if they separate declarations or statements.
16686 In such contexts, a
16687 sequence of two or more empty lines is replaced by exactly one empty line.
16688 Note that a blank line will be removed if it separates two ``comment blocks''
16689 (a comment block is a sequence of whole-line comments).
16690 In order to preserve a visual separation between comment blocks, use an
16691 ``empty comment'' (a line comprising only hyphens) rather than an empty line.
16692 Likewise, if for some reason you wish to have a sequence of empty lines,
16693 use a sequence of empty comments instead.
16695 @node Formatting Comments
16696 @subsection Formatting Comments
16699 Comments in Ada code are of two kinds:
16702 a @emph{whole-line comment}, which appears by itself (possibly preceded by
16703 ``white space'') on a line
16706 an @emph{end-of-line comment}, which follows some other Ada lexical element
16711 The indentation of a whole-line comment is that of either
16712 the preceding or following line in
16713 the formatted source, depending on switch settings as will be described below.
16715 For an end-of-line comment, @command{gnatpp} leaves the same number of spaces
16716 between the end of the preceding Ada lexical element and the beginning
16717 of the comment as appear in the original source,
16718 unless either the comment has to be split to
16719 satisfy the line length limitation, or else the next line contains a
16720 whole line comment that is considered a continuation of this end-of-line
16721 comment (because it starts at the same position).
16723 cases, the start of the end-of-line comment is moved right to the nearest
16724 multiple of the indentation level.
16725 This may result in a ``line overflow'' (the right-shifted comment extending
16726 beyond the maximum line length), in which case the comment is split as
16729 There is a difference between @option{^-c1^/COMMENTS_LAYOUT=DEFAULT^}
16730 (GNAT-style comment line indentation)
16731 and @option{^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^}
16732 (reference-manual comment line indentation).
16733 With reference-manual style, a whole-line comment is indented as if it
16734 were a declaration or statement at the same place
16735 (i.e., according to the indentation of the preceding line(s)).
16736 With GNAT style, a whole-line comment that is immediately followed by an
16737 @b{if} or @b{case} statement alternative, a record variant, or the reserved
16738 word @b{begin}, is indented based on the construct that follows it.
16741 @smallexample @c ada
16753 Reference-manual indentation produces:
16755 @smallexample @c ada
16767 while GNAT-style indentation produces:
16769 @smallexample @c ada
16781 The @option{^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^} switch
16782 (GNAT style comment beginning) has the following
16787 For each whole-line comment that does not end with two hyphens,
16788 @command{gnatpp} inserts spaces if necessary after the starting two hyphens
16789 to ensure that there are at least two spaces between these hyphens and the
16790 first non-blank character of the comment.
16794 For an end-of-line comment, if in the original source the next line is a
16795 whole-line comment that starts at the same position
16796 as the end-of-line comment,
16797 then the whole-line comment (and all whole-line comments
16798 that follow it and that start at the same position)
16799 will start at this position in the output file.
16802 That is, if in the original source we have:
16804 @smallexample @c ada
16807 A := B + C; -- B must be in the range Low1..High1
16808 -- C must be in the range Low2..High2
16809 --B+C will be in the range Low1+Low2..High1+High2
16815 Then in the formatted source we get
16817 @smallexample @c ada
16820 A := B + C; -- B must be in the range Low1..High1
16821 -- C must be in the range Low2..High2
16822 -- B+C will be in the range Low1+Low2..High1+High2
16828 A comment that exceeds the line length limit will be split.
16830 @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} (reformat comment blocks) is set and
16831 the line belongs to a reformattable block, splitting the line generates a
16832 @command{gnatpp} warning.
16833 The @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} switch specifies that whole-line
16834 comments may be reformatted in typical
16835 word processor style (that is, moving words between lines and putting as
16836 many words in a line as possible).
16839 The @option{^-c5^/COMMENTS_LAYOUT=KEEP_SPECIAL^} switch specifies, that comments
16840 that has a special format (that is, a character that is neither a letter nor digit
16841 not white space nor line break immediately following the leading @code{--} of
16842 the comment) should be without any change moved from the argument source
16843 into reformatted source. This switch allows to preserve comments that are used
16844 as a special marks in the code (e.g.@: SPARK annotation).
16846 @node Construct Layout
16847 @subsection Construct Layout
16850 In several cases the suggested layout in the Ada Reference Manual includes
16851 an extra level of indentation that many programmers prefer to avoid. The
16852 affected cases include:
16856 @item Record type declaration (RM 3.8)
16858 @item Record representation clause (RM 13.5.1)
16860 @item Loop statement in case if a loop has a statement identifier (RM 5.6)
16862 @item Block statement in case if a block has a statement identifier (RM 5.6)
16866 In compact mode (when GNAT style layout or compact layout is set),
16867 the pretty printer uses one level of indentation instead
16868 of two. This is achieved in the record definition and record representation
16869 clause cases by putting the @code{record} keyword on the same line as the
16870 start of the declaration or representation clause, and in the block and loop
16871 case by putting the block or loop header on the same line as the statement
16875 The difference between GNAT style @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^}
16876 and compact @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^}
16877 layout on the one hand, and uncompact layout
16878 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} on the other hand,
16879 can be illustrated by the following examples:
16883 @multitable @columnfractions .5 .5
16884 @item @i{GNAT style, compact layout} @tab @i{Uncompact layout}
16887 @smallexample @c ada
16894 @smallexample @c ada
16903 @smallexample @c ada
16905 a at 0 range 0 .. 31;
16906 b at 4 range 0 .. 31;
16910 @smallexample @c ada
16913 a at 0 range 0 .. 31;
16914 b at 4 range 0 .. 31;
16919 @smallexample @c ada
16927 @smallexample @c ada
16937 @smallexample @c ada
16938 Clear : for J in 1 .. 10 loop
16943 @smallexample @c ada
16945 for J in 1 .. 10 loop
16956 GNAT style, compact layout Uncompact layout
16958 type q is record type q is
16959 a : integer; record
16960 b : integer; a : integer;
16961 end record; b : integer;
16964 for q use record for q use
16965 a at 0 range 0 .. 31; record
16966 b at 4 range 0 .. 31; a at 0 range 0 .. 31;
16967 end record; b at 4 range 0 .. 31;
16970 Block : declare Block :
16971 A : Integer := 3; declare
16972 begin A : Integer := 3;
16974 end Block; Proc (A, A);
16977 Clear : for J in 1 .. 10 loop Clear :
16978 A (J) := 0; for J in 1 .. 10 loop
16979 end loop Clear; A (J) := 0;
16986 A further difference between GNAT style layout and compact layout is that
16987 GNAT style layout inserts empty lines as separation for
16988 compound statements, return statements and bodies.
16990 Note that the layout specified by
16991 @option{^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^}
16992 for named block and loop statements overrides the layout defined by these
16993 constructs by @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^},
16994 @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^} or
16995 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} option.
16998 @subsection Name Casing
17001 @command{gnatpp} always converts the usage occurrence of a (simple) name to
17002 the same casing as the corresponding defining identifier.
17004 You control the casing for defining occurrences via the
17005 @option{^-n^/NAME_CASING^} switch.
17007 With @option{-nD} (``as declared'', which is the default),
17010 With @option{/NAME_CASING=AS_DECLARED}, which is the default,
17012 defining occurrences appear exactly as in the source file
17013 where they are declared.
17014 The other ^values for this switch^options for this qualifier^ ---
17015 @option{^-nU^UPPER_CASE^},
17016 @option{^-nL^LOWER_CASE^},
17017 @option{^-nM^MIXED_CASE^} ---
17019 ^upper, lower, or mixed case, respectively^the corresponding casing^.
17020 If @command{gnatpp} changes the casing of a defining
17021 occurrence, it analogously changes the casing of all the
17022 usage occurrences of this name.
17024 If the defining occurrence of a name is not in the source compilation unit
17025 currently being processed by @command{gnatpp}, the casing of each reference to
17026 this name is changed according to the value of the @option{^-n^/NAME_CASING^}
17027 switch (subject to the dictionary file mechanism described below).
17028 Thus @command{gnatpp} acts as though the @option{^-n^/NAME_CASING^} switch
17030 casing for the defining occurrence of the name.
17032 Some names may need to be spelled with casing conventions that are not
17033 covered by the upper-, lower-, and mixed-case transformations.
17034 You can arrange correct casing by placing such names in a
17035 @emph{dictionary file},
17036 and then supplying a @option{^-D^/DICTIONARY^} switch.
17037 The casing of names from dictionary files overrides
17038 any @option{^-n^/NAME_CASING^} switch.
17040 To handle the casing of Ada predefined names and the names from GNAT libraries,
17041 @command{gnatpp} assumes a default dictionary file.
17042 The name of each predefined entity is spelled with the same casing as is used
17043 for the entity in the @cite{Ada Reference Manual}.
17044 The name of each entity in the GNAT libraries is spelled with the same casing
17045 as is used in the declaration of that entity.
17047 The @w{@option{^-D-^/SPECIFIC_CASING^}} switch suppresses the use of the
17048 default dictionary file.
17049 Instead, the casing for predefined and GNAT-defined names will be established
17050 by the @option{^-n^/NAME_CASING^} switch or explicit dictionary files.
17051 For example, by default the names @code{Ada.Text_IO} and @code{GNAT.OS_Lib}
17052 will appear as just shown,
17053 even in the presence of a @option{^-nU^/NAME_CASING=UPPER_CASE^} switch.
17054 To ensure that even such names are rendered in uppercase,
17055 additionally supply the @w{@option{^-D-^/SPECIFIC_CASING^}} switch
17056 (or else, less conveniently, place these names in upper case in a dictionary
17059 A dictionary file is
17060 a plain text file; each line in this file can be either a blank line
17061 (containing only space characters and ASCII.HT characters), an Ada comment
17062 line, or the specification of exactly one @emph{casing schema}.
17064 A casing schema is a string that has the following syntax:
17068 @var{casing_schema} ::= @var{identifier} | *@var{simple_identifier}*
17070 @var{simple_identifier} ::= @var{letter}@{@var{letter_or_digit}@}
17075 (See @cite{Ada Reference Manual}, Section 2.3) for the definition of the
17076 @var{identifier} lexical element and the @var{letter_or_digit} category.)
17078 The casing schema string can be followed by white space and/or an Ada-style
17079 comment; any amount of white space is allowed before the string.
17081 If a dictionary file is passed as
17083 the value of a @option{-D@var{file}} switch
17086 an option to the @option{/DICTIONARY} qualifier
17089 simple name and every identifier, @command{gnatpp} checks if the dictionary
17090 defines the casing for the name or for some of its parts (the term ``subword''
17091 is used below to denote the part of a name which is delimited by ``_'' or by
17092 the beginning or end of the word and which does not contain any ``_'' inside):
17096 if the whole name is in the dictionary, @command{gnatpp} uses for this name
17097 the casing defined by the dictionary; no subwords are checked for this word
17100 for every subword @command{gnatpp} checks if the dictionary contains the
17101 corresponding string of the form @code{*@var{simple_identifier}*},
17102 and if it does, the casing of this @var{simple_identifier} is used
17106 if the whole name does not contain any ``_'' inside, and if for this name
17107 the dictionary contains two entries - one of the form @var{identifier},
17108 and another - of the form *@var{simple_identifier}*, then the first one
17109 is applied to define the casing of this name
17112 if more than one dictionary file is passed as @command{gnatpp} switches, each
17113 dictionary adds new casing exceptions and overrides all the existing casing
17114 exceptions set by the previous dictionaries
17117 when @command{gnatpp} checks if the word or subword is in the dictionary,
17118 this check is not case sensitive
17122 For example, suppose we have the following source to reformat:
17124 @smallexample @c ada
17127 name1 : integer := 1;
17128 name4_name3_name2 : integer := 2;
17129 name2_name3_name4 : Boolean;
17132 name2_name3_name4 := name4_name3_name2 > name1;
17138 And suppose we have two dictionaries:
17155 If @command{gnatpp} is called with the following switches:
17159 @command{gnatpp -nM -D dict1 -D dict2 test.adb}
17162 @command{gnatpp test.adb /NAME_CASING=MIXED_CASE /DICTIONARY=(dict1, dict2)}
17167 then we will get the following name casing in the @command{gnatpp} output:
17169 @smallexample @c ada
17172 NAME1 : Integer := 1;
17173 Name4_NAME3_Name2 : Integer := 2;
17174 Name2_NAME3_Name4 : Boolean;
17177 Name2_NAME3_Name4 := Name4_NAME3_Name2 > NAME1;
17182 @c *********************************
17183 @node The GNAT Metric Tool gnatmetric
17184 @chapter The GNAT Metric Tool @command{gnatmetric}
17186 @cindex Metric tool
17189 ^The @command{gnatmetric} tool^@command{GNAT METRIC}^ is an ASIS-based utility
17190 for computing various program metrics.
17191 It takes an Ada source file as input and generates a file containing the
17192 metrics data as output. Various switches control which
17193 metrics are computed and output.
17195 @command{gnatmetric} generates and uses the ASIS
17196 tree for the input source and thus requires the input to be syntactically and
17197 semantically legal.
17198 If this condition is not met, @command{gnatmetric} will generate
17199 an error message; no metric information for this file will be
17200 computed and reported.
17202 If the compilation unit contained in the input source depends semantically
17203 upon units in files located outside the current directory, you have to provide
17204 the source search path when invoking @command{gnatmetric}.
17205 If it depends semantically upon units that are contained
17206 in files with names that do not follow the GNAT file naming rules, you have to
17207 provide the configuration file describing the corresponding naming scheme (see
17208 the description of the @command{gnatmetric} switches below.)
17209 Alternatively, you may use a project file and invoke @command{gnatmetric}
17210 through the @command{gnat} driver.
17212 The @command{gnatmetric} command has the form
17215 $ gnatmetric @ovar{switches} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
17222 @var{switches} specify the metrics to compute and define the destination for
17226 Each @var{filename} is the name (including the extension) of a source
17227 file to process. ``Wildcards'' are allowed, and
17228 the file name may contain path information.
17229 If no @var{filename} is supplied, then the @var{switches} list must contain
17231 @option{-files} switch (@pxref{Other gnatmetric Switches}).
17232 Including both a @option{-files} switch and one or more
17233 @var{filename} arguments is permitted.
17236 @samp{-cargs @var{gcc_switches}} is a list of switches for
17237 @command{gcc}. They will be passed on to all compiler invocations made by
17238 @command{gnatmetric} to generate the ASIS trees. Here you can provide
17239 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
17240 and use the @option{-gnatec} switch to set the configuration file.
17244 * Switches for gnatmetric::
17247 @node Switches for gnatmetric
17248 @section Switches for @command{gnatmetric}
17251 The following subsections describe the various switches accepted by
17252 @command{gnatmetric}, organized by category.
17255 * Output Files Control::
17256 * Disable Metrics For Local Units::
17257 * Specifying a set of metrics to compute::
17258 * Other gnatmetric Switches::
17259 * Generate project-wide metrics::
17262 @node Output Files Control
17263 @subsection Output File Control
17264 @cindex Output file control in @command{gnatmetric}
17267 @command{gnatmetric} has two output formats. It can generate a
17268 textual (human-readable) form, and also XML. By default only textual
17269 output is generated.
17271 When generating the output in textual form, @command{gnatmetric} creates
17272 for each Ada source file a corresponding text file
17273 containing the computed metrics, except for the case when the set of metrics
17274 specified by gnatmetric parameters consists only of metrics that are computed
17275 for the whole set of analyzed sources, but not for each Ada source.
17276 By default, this file is placed in the same directory as where the source
17277 file is located, and its name is obtained
17278 by appending the ^@file{.metrix}^@file{$METRIX}^ suffix to the name of the
17281 All the output information generated in XML format is placed in a single
17282 file. By default this file is placed in the current directory and has the
17283 name ^@file{metrix.xml}^@file{METRIX$XML}^.
17285 Some of the computed metrics are summed over the units passed to
17286 @command{gnatmetric}; for example, the total number of lines of code.
17287 By default this information is sent to @file{stdout}, but a file
17288 can be specified with the @option{-og} switch.
17290 The following switches control the @command{gnatmetric} output:
17293 @cindex @option{^-x^/XML^} (@command{gnatmetric})
17295 Generate the XML output
17297 @cindex @option{^-xs^/XSD^} (@command{gnatmetric})
17299 Generate the XML output and the XML schema file that describes the structure
17300 of the XML metric report, this schema is assigned to the XML file. The schema
17301 file has the same name as the XML output file with @file{.xml} suffix replaced
17304 @cindex @option{^-nt^/NO_TEXT^} (@command{gnatmetric})
17305 @item ^-nt^/NO_TEXT^
17306 Do not generate the output in text form (implies @option{^-x^/XML^})
17308 @cindex @option{^-d^/DIRECTORY^} (@command{gnatmetric})
17309 @item ^-d @var{output_dir}^/DIRECTORY=@var{output_dir}^
17310 Put textual files with detailed metrics into @var{output_dir}
17312 @cindex @option{^-o^/SUFFIX_DETAILS^} (@command{gnatmetric})
17313 @item ^-o @var{file_suffix}^/SUFFIX_DETAILS=@var{file_suffix}^
17314 Use @var{file_suffix}, instead of ^@file{.metrix}^@file{$METRIX}^
17315 in the name of the output file.
17317 @cindex @option{^-og^/GLOBAL_OUTPUT^} (@command{gnatmetric})
17318 @item ^-og @var{file_name}^/GLOBAL_OUTPUT=@var{file_name}^
17319 Put global metrics into @var{file_name}
17321 @cindex @option{^-ox^/XML_OUTPUT^} (@command{gnatmetric})
17322 @item ^-ox @var{file_name}^/XML_OUTPUT=@var{file_name}^
17323 Put the XML output into @var{file_name} (also implies @option{^-x^/XML^})
17325 @cindex @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} (@command{gnatmetric})
17326 @item ^-sfn^/SHORT_SOURCE_FILE_NAME^
17327 Use ``short'' source file names in the output. (The @command{gnatmetric}
17328 output includes the name(s) of the Ada source file(s) from which the metrics
17329 are computed. By default each name includes the absolute path. The
17330 @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} switch causes @command{gnatmetric}
17331 to exclude all directory information from the file names that are output.)
17335 @node Disable Metrics For Local Units
17336 @subsection Disable Metrics For Local Units
17337 @cindex Disable Metrics For Local Units in @command{gnatmetric}
17340 @command{gnatmetric} relies on the GNAT compilation model @minus{}
17342 unit per one source file. It computes line metrics for the whole source
17343 file, and it also computes syntax
17344 and complexity metrics for the file's outermost unit.
17346 By default, @command{gnatmetric} will also compute all metrics for certain
17347 kinds of locally declared program units:
17351 subprogram (and generic subprogram) bodies;
17354 package (and generic package) specs and bodies;
17357 task object and type specifications and bodies;
17360 protected object and type specifications and bodies.
17364 These kinds of entities will be referred to as
17365 @emph{eligible local program units}, or simply @emph{eligible local units},
17366 @cindex Eligible local unit (for @command{gnatmetric})
17367 in the discussion below.
17369 Note that a subprogram declaration, generic instantiation,
17370 or renaming declaration only receives metrics
17371 computation when it appear as the outermost entity
17374 Suppression of metrics computation for eligible local units can be
17375 obtained via the following switch:
17378 @cindex @option{^-n@var{x}^/SUPPRESS^} (@command{gnatmetric})
17379 @item ^-nolocal^/SUPPRESS=LOCAL_DETAILS^
17380 Do not compute detailed metrics for eligible local program units
17384 @node Specifying a set of metrics to compute
17385 @subsection Specifying a set of metrics to compute
17388 By default all the metrics are computed and reported. The switches
17389 described in this subsection allow you to control, on an individual
17390 basis, whether metrics are computed and
17391 reported. If at least one positive metric
17392 switch is specified (that is, a switch that defines that a given
17393 metric or set of metrics is to be computed), then only
17394 explicitly specified metrics are reported.
17397 * Line Metrics Control::
17398 * Syntax Metrics Control::
17399 * Complexity Metrics Control::
17400 * Object-Oriented Metrics Control::
17403 @node Line Metrics Control
17404 @subsubsection Line Metrics Control
17405 @cindex Line metrics control in @command{gnatmetric}
17408 For any (legal) source file, and for each of its
17409 eligible local program units, @command{gnatmetric} computes the following
17414 the total number of lines;
17417 the total number of code lines (i.e., non-blank lines that are not comments)
17420 the number of comment lines
17423 the number of code lines containing end-of-line comments;
17426 the comment percentage: the ratio between the number of lines that contain
17427 comments and the number of all non-blank lines, expressed as a percentage;
17430 the number of empty lines and lines containing only space characters and/or
17431 format effectors (blank lines)
17434 the average number of code lines in subprogram bodies, task bodies, entry
17435 bodies and statement sequences in package bodies (this metric is only computed
17436 across the whole set of the analyzed units)
17441 @command{gnatmetric} sums the values of the line metrics for all the
17442 files being processed and then generates the cumulative results. The tool
17443 also computes for all the files being processed the average number of code
17446 You can use the following switches to select the specific line metrics
17447 to be computed and reported.
17450 @cindex @option{^--lines@var{x}^/LINE_COUNT_METRICS^} (@command{gnatmetric})
17453 @cindex @option{--no-lines@var{x}}
17456 @item ^--lines-all^/LINE_COUNT_METRICS=ALL^
17457 Report all the line metrics
17459 @item ^--no-lines-all^/LINE_COUNT_METRICS=NONE^
17460 Do not report any of line metrics
17462 @item ^--lines^/LINE_COUNT_METRICS=ALL_LINES^
17463 Report the number of all lines
17465 @item ^--no-lines^/LINE_COUNT_METRICS=NOALL_LINES^
17466 Do not report the number of all lines
17468 @item ^--lines-code^/LINE_COUNT_METRICS=CODE_LINES^
17469 Report the number of code lines
17471 @item ^--no-lines-code^/LINE_COUNT_METRICS=NOCODE_LINES^
17472 Do not report the number of code lines
17474 @item ^--lines-comment^/LINE_COUNT_METRICS=COMMENT_LINES^
17475 Report the number of comment lines
17477 @item ^--no-lines-comment^/LINE_COUNT_METRICS=NOCOMMENT_LINES^
17478 Do not report the number of comment lines
17480 @item ^--lines-eol-comment^/LINE_COUNT_METRICS=CODE_COMMENT_LINES^
17481 Report the number of code lines containing
17482 end-of-line comments
17484 @item ^--no-lines-eol-comment^/LINE_COUNT_METRICS=NOCODE_COMMENT_LINES^
17485 Do not report the number of code lines containing
17486 end-of-line comments
17488 @item ^--lines-ratio^/LINE_COUNT_METRICS=COMMENT_PERCENTAGE^
17489 Report the comment percentage in the program text
17491 @item ^--no-lines-ratio^/LINE_COUNT_METRICS=NOCOMMENT_PERCENTAGE^
17492 Do not report the comment percentage in the program text
17494 @item ^--lines-blank^/LINE_COUNT_METRICS=BLANK_LINES^
17495 Report the number of blank lines
17497 @item ^--no-lines-blank^/LINE_COUNT_METRICS=NOBLANK_LINES^
17498 Do not report the number of blank lines
17500 @item ^--lines-average^/LINE_COUNT_METRICS=AVERAGE_BODY_LINES^
17501 Report the average number of code lines in subprogram bodies, task bodies,
17502 entry bodies and statement sequences in package bodies. The metric is computed
17503 and reported for the whole set of processed Ada sources only.
17505 @item ^--no-lines-average^/LINE_COUNT_METRICS=NOAVERAGE_BODY_LINES^
17506 Do not report the average number of code lines in subprogram bodies,
17507 task bodies, entry bodies and statement sequences in package bodies.
17511 @node Syntax Metrics Control
17512 @subsubsection Syntax Metrics Control
17513 @cindex Syntax metrics control in @command{gnatmetric}
17516 @command{gnatmetric} computes various syntactic metrics for the
17517 outermost unit and for each eligible local unit:
17520 @item LSLOC (``Logical Source Lines Of Code'')
17521 The total number of declarations and the total number of statements
17523 @item Maximal static nesting level of inner program units
17525 @cite{Ada Reference Manual}, 10.1(1), ``A program unit is either a
17526 package, a task unit, a protected unit, a
17527 protected entry, a generic unit, or an explicitly declared subprogram other
17528 than an enumeration literal.''
17530 @item Maximal nesting level of composite syntactic constructs
17531 This corresponds to the notion of the
17532 maximum nesting level in the GNAT built-in style checks
17533 (@pxref{Style Checking})
17537 For the outermost unit in the file, @command{gnatmetric} additionally computes
17538 the following metrics:
17541 @item Public subprograms
17542 This metric is computed for package specs. It is the
17543 number of subprograms and generic subprograms declared in the visible
17544 part (including the visible part of nested packages, protected objects, and
17547 @item All subprograms
17548 This metric is computed for bodies and subunits. The
17549 metric is equal to a total number of subprogram bodies in the compilation
17551 Neither generic instantiations nor renamings-as-a-body nor body stubs
17552 are counted. Any subprogram body is counted, independently of its nesting
17553 level and enclosing constructs. Generic bodies and bodies of protected
17554 subprograms are counted in the same way as ``usual'' subprogram bodies.
17557 This metric is computed for package specs and
17558 generic package declarations. It is the total number of types
17559 that can be referenced from outside this compilation unit, plus the
17560 number of types from all the visible parts of all the visible generic
17561 packages. Generic formal types are not counted. Only types, not subtypes,
17565 Along with the total number of public types, the following
17566 types are counted and reported separately:
17573 Root tagged types (abstract, non-abstract, private, non-private). Type
17574 extensions are @emph{not} counted
17577 Private types (including private extensions)
17588 This metric is computed for any compilation unit. It is equal to the total
17589 number of the declarations of different types given in the compilation unit.
17590 The private and the corresponding full type declaration are counted as one
17591 type declaration. Incomplete type declarations and generic formal types
17593 No distinction is made among different kinds of types (abstract,
17594 private etc.); the total number of types is computed and reported.
17599 By default, all the syntax metrics are computed and reported. You can use the
17600 following switches to select specific syntax metrics.
17604 @cindex @option{^--syntax@var{x}^/SYNTAX_METRICS^} (@command{gnatmetric})
17607 @cindex @option{--no-syntax@var{x}} (@command{gnatmetric})
17610 @item ^--syntax-all^/SYNTAX_METRICS=ALL^
17611 Report all the syntax metrics
17613 @item ^--no-syntax-all^/SYNTAX_METRICS=NONE^
17614 Do not report any of syntax metrics
17616 @item ^--declarations^/SYNTAX_METRICS=DECLARATIONS^
17617 Report the total number of declarations
17619 @item ^--no-declarations^/SYNTAX_METRICS=NODECLARATIONS^
17620 Do not report the total number of declarations
17622 @item ^--statements^/SYNTAX_METRICS=STATEMENTS^
17623 Report the total number of statements
17625 @item ^--no-statements^/SYNTAX_METRICS=NOSTATEMENTS^
17626 Do not report the total number of statements
17628 @item ^--public-subprograms^/SYNTAX_METRICS=PUBLIC_SUBPROGRAMS^
17629 Report the number of public subprograms in a compilation unit
17631 @item ^--no-public-subprograms^/SYNTAX_METRICS=NOPUBLIC_SUBPROGRAMS^
17632 Do not report the number of public subprograms in a compilation unit
17634 @item ^--all-subprograms^/SYNTAX_METRICS=ALL_SUBPROGRAMS^
17635 Report the number of all the subprograms in a compilation unit
17637 @item ^--no-all-subprograms^/SYNTAX_METRICS=NOALL_SUBPROGRAMS^
17638 Do not report the number of all the subprograms in a compilation unit
17640 @item ^--public-types^/SYNTAX_METRICS=PUBLIC_TYPES^
17641 Report the number of public types in a compilation unit
17643 @item ^--no-public-types^/SYNTAX_METRICS=NOPUBLIC_TYPES^
17644 Do not report the number of public types in a compilation unit
17646 @item ^--all-types^/SYNTAX_METRICS=ALL_TYPES^
17647 Report the number of all the types in a compilation unit
17649 @item ^--no-all-types^/SYNTAX_METRICS=NOALL_TYPES^
17650 Do not report the number of all the types in a compilation unit
17652 @item ^--unit-nesting^/SYNTAX_METRICS=UNIT_NESTING^
17653 Report the maximal program unit nesting level
17655 @item ^--no-unit-nesting^/SYNTAX_METRICS=UNIT_NESTING_OFF^
17656 Do not report the maximal program unit nesting level
17658 @item ^--construct-nesting^/SYNTAX_METRICS=CONSTRUCT_NESTING^
17659 Report the maximal construct nesting level
17661 @item ^--no-construct-nesting^/SYNTAX_METRICS=NOCONSTRUCT_NESTING^
17662 Do not report the maximal construct nesting level
17666 @node Complexity Metrics Control
17667 @subsubsection Complexity Metrics Control
17668 @cindex Complexity metrics control in @command{gnatmetric}
17671 For a program unit that is an executable body (a subprogram body (including
17672 generic bodies), task body, entry body or a package body containing
17673 its own statement sequence) @command{gnatmetric} computes the following
17674 complexity metrics:
17678 McCabe cyclomatic complexity;
17681 McCabe essential complexity;
17684 maximal loop nesting level
17689 The McCabe complexity metrics are defined
17690 in @url{http://www.mccabe.com/pdf/nist235r.pdf}
17692 According to McCabe, both control statements and short-circuit control forms
17693 should be taken into account when computing cyclomatic complexity. For each
17694 body, we compute three metric values:
17698 the complexity introduced by control
17699 statements only, without taking into account short-circuit forms,
17702 the complexity introduced by short-circuit control forms only, and
17706 cyclomatic complexity, which is the sum of these two values.
17710 When computing cyclomatic and essential complexity, @command{gnatmetric} skips
17711 the code in the exception handlers and in all the nested program units.
17713 By default, all the complexity metrics are computed and reported.
17714 For more fine-grained control you can use
17715 the following switches:
17718 @cindex @option{^-complexity@var{x}^/COMPLEXITY_METRICS^} (@command{gnatmetric})
17721 @cindex @option{--no-complexity@var{x}}
17724 @item ^--complexity-all^/COMPLEXITY_METRICS=ALL^
17725 Report all the complexity metrics
17727 @item ^--no-complexity-all^/COMPLEXITY_METRICS=NONE^
17728 Do not report any of complexity metrics
17730 @item ^--complexity-cyclomatic^/COMPLEXITY_METRICS=CYCLOMATIC^
17731 Report the McCabe Cyclomatic Complexity
17733 @item ^--no-complexity-cyclomatic^/COMPLEXITY_METRICS=NOCYCLOMATIC^
17734 Do not report the McCabe Cyclomatic Complexity
17736 @item ^--complexity-essential^/COMPLEXITY_METRICS=ESSENTIAL^
17737 Report the Essential Complexity
17739 @item ^--no-complexity-essential^/COMPLEXITY_METRICS=NOESSENTIAL^
17740 Do not report the Essential Complexity
17742 @item ^--loop-nesting^/COMPLEXITY_METRICS=LOOP_NESTING_ON^
17743 Report maximal loop nesting level
17745 @item ^--no-loop-nesting^/COMPLEXITY_METRICS=NOLOOP_NESTING^
17746 Do not report maximal loop nesting level
17748 @item ^--complexity-average^/COMPLEXITY_METRICS=AVERAGE_COMPLEXITY^
17749 Report the average McCabe Cyclomatic Complexity for all the subprogram bodies,
17750 task bodies, entry bodies and statement sequences in package bodies.
17751 The metric is computed and reported for whole set of processed Ada sources
17754 @item ^--no-complexity-average^/COMPLEXITY_METRICS=NOAVERAGE_COMPLEXITY^
17755 Do not report the average McCabe Cyclomatic Complexity for all the subprogram
17756 bodies, task bodies, entry bodies and statement sequences in package bodies
17758 @cindex @option{^-ne^/NO_EXITS_AS_GOTOS^} (@command{gnatmetric})
17759 @item ^-ne^/NO_EXITS_AS_GOTOS^
17760 Do not consider @code{exit} statements as @code{goto}s when
17761 computing Essential Complexity
17763 @item ^--extra-exit-points^/EXTRA_EXIT_POINTS^
17764 Report the extra exit points for subprogram bodies. As an exit point, this
17765 metric counts @code{return} statements and raise statements in case when the
17766 raised exception is not handled in the same body. In case of a function this
17767 metric subtracts 1 from the number of exit points, because a function body
17768 must contain at least one @code{return} statement.
17770 @item ^--no-extra-exit-points^/NOEXTRA_EXIT_POINTS^
17771 Do not report the extra exit points for subprogram bodies
17775 @node Object-Oriented Metrics Control
17776 @subsubsection Object-Oriented Metrics Control
17777 @cindex Object-Oriented metrics control in @command{gnatmetric}
17780 @cindex Coupling metrics (in in @command{gnatmetric})
17781 Coupling metrics are object-oriented metrics that measure the
17782 dependencies between a given class (or a group of classes) and the
17783 ``external world'' (that is, the other classes in the program). In this
17784 subsection the term ``class'' is used in its
17785 traditional object-oriented programming sense
17786 (an instantiable module that contains data and/or method members).
17787 A @emph{category} (of classes)
17788 is a group of closely related classes that are reused and/or
17791 A class @code{K}'s @emph{efferent coupling} is the number of classes
17792 that @code{K} depends upon.
17793 A category's efferent coupling is the number of classes outside the
17794 category that the classes inside the category depend upon.
17796 A class @code{K}'s @emph{afferent coupling} is the number of classes
17797 that depend upon @code{K}.
17798 A category's afferent coupling is the number of classes outside the
17799 category that depend on classes belonging to the category.
17801 Ada's implementation of the object-oriented paradigm does not use the
17802 traditional class notion, so the definition of the coupling
17803 metrics for Ada maps the class and class category notions
17804 onto Ada constructs.
17806 For the coupling metrics, several kinds of modules -- a library package,
17807 a library generic package, and a library generic package instantiation --
17808 that define a tagged type or an interface type are
17809 considered to be a class. A category consists of a library package (or
17810 a library generic package) that defines a tagged or an interface type,
17811 together with all its descendant (generic) packages that define tagged
17812 or interface types. For any package counted as a class,
17813 its body and subunits (if any) are considered
17814 together with its spec when counting the dependencies, and coupling
17815 metrics are reported for spec units only. For dependencies
17816 between classes, the Ada semantic dependencies are considered.
17817 For coupling metrics, only dependencies on units that are considered as
17818 classes, are considered.
17820 When computing coupling metrics, @command{gnatmetric} counts only
17821 dependencies between units that are arguments of the gnatmetric call.
17822 Coupling metrics are program-wide (or project-wide) metrics, so to
17823 get a valid result, you should call @command{gnatmetric} for
17824 the whole set of sources that make up your program. It can be done
17825 by calling @command{gnatmetric} from the GNAT driver with @option{-U}
17826 option (see See @ref{The GNAT Driver and Project Files} for details.
17828 By default, all the coupling metrics are disabled. You can use the following
17829 switches to specify the coupling metrics to be computed and reported:
17834 @cindex @option{--package@var{x}} (@command{gnatmetric})
17835 @cindex @option{--no-package@var{x}} (@command{gnatmetric})
17836 @cindex @option{--category@var{x}} (@command{gnatmetric})
17837 @cindex @option{--no-category@var{x}} (@command{gnatmetric})
17841 @cindex @option{/COUPLING_METRICS} (@command{gnatmetric})
17844 @item ^--coupling-all^/COUPLING_METRICS=ALL^
17845 Report all the coupling metrics
17847 @item ^--no-coupling-all^/COUPLING_METRICS=NONE^
17848 Do not report any of metrics
17850 @item ^--package-efferent-coupling^/COUPLING_METRICS=PACKAGE_EFFERENT^
17851 Report package efferent coupling
17853 @item ^--no-package-efferent-coupling^/COUPLING_METRICS=NOPACKAGE_EFFERENT^
17854 Do not report package efferent coupling
17856 @item ^--package-afferent-coupling^/COUPLING_METRICS=PACKAGE_AFFERENT^
17857 Report package afferent coupling
17859 @item ^--no-package-afferent-coupling^/COUPLING_METRICS=NOPACKAGE_AFFERENT^
17860 Do not report package afferent coupling
17862 @item ^--category-efferent-coupling^/COUPLING_METRICS=CATEGORY_EFFERENT^
17863 Report category efferent coupling
17865 @item ^--no-category-efferent-coupling^/COUPLING_METRICS=NOCATEGORY_EFFERENT^
17866 Do not report category efferent coupling
17868 @item ^--category-afferent-coupling^/COUPLING_METRICS=CATEGORY_AFFERENT^
17869 Report category afferent coupling
17871 @item ^--no-category-afferent-coupling^/COUPLING_METRICS=NOCATEGORY_AFFERENT^
17872 Do not report category afferent coupling
17876 @node Other gnatmetric Switches
17877 @subsection Other @code{gnatmetric} Switches
17880 Additional @command{gnatmetric} switches are as follows:
17883 @item ^-files @var{filename}^/FILES=@var{filename}^
17884 @cindex @option{^-files^/FILES^} (@code{gnatmetric})
17885 Take the argument source files from the specified file. This file should be an
17886 ordinary text file containing file names separated by spaces or
17887 line breaks. You can use this switch more then once in the same call to
17888 @command{gnatmetric}. You also can combine this switch with
17889 an explicit list of files.
17891 @item ^-v^/VERBOSE^
17892 @cindex @option{^-v^/VERBOSE^} (@code{gnatmetric})
17894 @command{gnatmetric} generates version information and then
17895 a trace of sources being processed.
17897 @item ^-dv^/DEBUG_OUTPUT^
17898 @cindex @option{^-dv^/DEBUG_OUTPUT^} (@code{gnatmetric})
17900 @command{gnatmetric} generates various messages useful to understand what
17901 happens during the metrics computation
17904 @cindex @option{^-q^/QUIET^} (@code{gnatmetric})
17908 @node Generate project-wide metrics
17909 @subsection Generate project-wide metrics
17911 In order to compute metrics on all units of a given project, you can use
17912 the @command{gnat} driver along with the @option{-P} option:
17918 If the project @code{proj} depends upon other projects, you can compute
17919 the metrics on the project closure using the @option{-U} option:
17921 gnat metric -Pproj -U
17925 Finally, if not all the units are relevant to a particular main
17926 program in the project closure, you can generate metrics for the set
17927 of units needed to create a given main program (unit closure) using
17928 the @option{-U} option followed by the name of the main unit:
17930 gnat metric -Pproj -U main
17934 @c ***********************************
17935 @node File Name Krunching Using gnatkr
17936 @chapter File Name Krunching Using @code{gnatkr}
17940 This chapter discusses the method used by the compiler to shorten
17941 the default file names chosen for Ada units so that they do not
17942 exceed the maximum length permitted. It also describes the
17943 @code{gnatkr} utility that can be used to determine the result of
17944 applying this shortening.
17948 * Krunching Method::
17949 * Examples of gnatkr Usage::
17953 @section About @code{gnatkr}
17956 The default file naming rule in GNAT
17957 is that the file name must be derived from
17958 the unit name. The exact default rule is as follows:
17961 Take the unit name and replace all dots by hyphens.
17963 If such a replacement occurs in the
17964 second character position of a name, and the first character is
17965 ^@samp{a}, @samp{g}, @samp{s}, or @samp{i}, ^@samp{A}, @samp{G}, @samp{S}, or @samp{I},^
17966 then replace the dot by the character
17967 ^@samp{~} (tilde)^@samp{$} (dollar sign)^
17968 instead of a minus.
17970 The reason for this exception is to avoid clashes
17971 with the standard names for children of System, Ada, Interfaces,
17972 and GNAT, which use the prefixes
17973 ^@samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},^@samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},^
17976 The @option{^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{nn}}
17977 switch of the compiler activates a ``krunching''
17978 circuit that limits file names to nn characters (where nn is a decimal
17979 integer). For example, using OpenVMS,
17980 where the maximum file name length is
17981 39, the value of nn is usually set to 39, but if you want to generate
17982 a set of files that would be usable if ported to a system with some
17983 different maximum file length, then a different value can be specified.
17984 The default value of 39 for OpenVMS need not be specified.
17986 The @code{gnatkr} utility can be used to determine the krunched name for
17987 a given file, when krunched to a specified maximum length.
17990 @section Using @code{gnatkr}
17993 The @code{gnatkr} command has the form
17997 $ gnatkr @var{name} @ovar{length}
18003 $ gnatkr @var{name} /COUNT=nn
18008 @var{name} is the uncrunched file name, derived from the name of the unit
18009 in the standard manner described in the previous section (i.e., in particular
18010 all dots are replaced by hyphens). The file name may or may not have an
18011 extension (defined as a suffix of the form period followed by arbitrary
18012 characters other than period). If an extension is present then it will
18013 be preserved in the output. For example, when krunching @file{hellofile.ads}
18014 to eight characters, the result will be hellofil.ads.
18016 Note: for compatibility with previous versions of @code{gnatkr} dots may
18017 appear in the name instead of hyphens, but the last dot will always be
18018 taken as the start of an extension. So if @code{gnatkr} is given an argument
18019 such as @file{Hello.World.adb} it will be treated exactly as if the first
18020 period had been a hyphen, and for example krunching to eight characters
18021 gives the result @file{hellworl.adb}.
18023 Note that the result is always all lower case (except on OpenVMS where it is
18024 all upper case). Characters of the other case are folded as required.
18026 @var{length} represents the length of the krunched name. The default
18027 when no argument is given is ^8^39^ characters. A length of zero stands for
18028 unlimited, in other words do not chop except for system files where the
18029 implied crunching length is always eight characters.
18032 The output is the krunched name. The output has an extension only if the
18033 original argument was a file name with an extension.
18035 @node Krunching Method
18036 @section Krunching Method
18039 The initial file name is determined by the name of the unit that the file
18040 contains. The name is formed by taking the full expanded name of the
18041 unit and replacing the separating dots with hyphens and
18042 using ^lowercase^uppercase^
18043 for all letters, except that a hyphen in the second character position is
18044 replaced by a ^tilde^dollar sign^ if the first character is
18045 ^@samp{a}, @samp{i}, @samp{g}, or @samp{s}^@samp{A}, @samp{I}, @samp{G}, or @samp{S}^.
18046 The extension is @code{.ads} for a
18047 spec and @code{.adb} for a body.
18048 Krunching does not affect the extension, but the file name is shortened to
18049 the specified length by following these rules:
18053 The name is divided into segments separated by hyphens, tildes or
18054 underscores and all hyphens, tildes, and underscores are
18055 eliminated. If this leaves the name short enough, we are done.
18058 If the name is too long, the longest segment is located (left-most
18059 if there are two of equal length), and shortened by dropping
18060 its last character. This is repeated until the name is short enough.
18062 As an example, consider the krunching of @*@file{our-strings-wide_fixed.adb}
18063 to fit the name into 8 characters as required by some operating systems.
18066 our-strings-wide_fixed 22
18067 our strings wide fixed 19
18068 our string wide fixed 18
18069 our strin wide fixed 17
18070 our stri wide fixed 16
18071 our stri wide fixe 15
18072 our str wide fixe 14
18073 our str wid fixe 13
18079 Final file name: oustwifi.adb
18083 The file names for all predefined units are always krunched to eight
18084 characters. The krunching of these predefined units uses the following
18085 special prefix replacements:
18089 replaced by @file{^a^A^-}
18092 replaced by @file{^g^G^-}
18095 replaced by @file{^i^I^-}
18098 replaced by @file{^s^S^-}
18101 These system files have a hyphen in the second character position. That
18102 is why normal user files replace such a character with a
18103 ^tilde^dollar sign^, to
18104 avoid confusion with system file names.
18106 As an example of this special rule, consider
18107 @*@file{ada-strings-wide_fixed.adb}, which gets krunched as follows:
18110 ada-strings-wide_fixed 22
18111 a- strings wide fixed 18
18112 a- string wide fixed 17
18113 a- strin wide fixed 16
18114 a- stri wide fixed 15
18115 a- stri wide fixe 14
18116 a- str wide fixe 13
18122 Final file name: a-stwifi.adb
18126 Of course no file shortening algorithm can guarantee uniqueness over all
18127 possible unit names, and if file name krunching is used then it is your
18128 responsibility to ensure that no name clashes occur. The utility
18129 program @code{gnatkr} is supplied for conveniently determining the
18130 krunched name of a file.
18132 @node Examples of gnatkr Usage
18133 @section Examples of @code{gnatkr} Usage
18140 $ gnatkr very_long_unit_name.ads --> velounna.ads
18141 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
18142 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
18143 $ gnatkr grandparent-parent-child --> grparchi
18145 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
18146 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
18149 @node Preprocessing Using gnatprep
18150 @chapter Preprocessing Using @code{gnatprep}
18154 This chapter discusses how to use GNAT's @code{gnatprep} utility for simple
18156 Although designed for use with GNAT, @code{gnatprep} does not depend on any
18157 special GNAT features.
18158 For further discussion of conditional compilation in general, see
18159 @ref{Conditional Compilation}.
18162 * Preprocessing Symbols::
18164 * Switches for gnatprep::
18165 * Form of Definitions File::
18166 * Form of Input Text for gnatprep::
18169 @node Preprocessing Symbols
18170 @section Preprocessing Symbols
18173 Preprocessing symbols are defined in definition files and referred to in
18174 sources to be preprocessed. A Preprocessing symbol is an identifier, following
18175 normal Ada (case-insensitive) rules for its syntax, with the restriction that
18176 all characters need to be in the ASCII set (no accented letters).
18178 @node Using gnatprep
18179 @section Using @code{gnatprep}
18182 To call @code{gnatprep} use
18185 $ gnatprep @ovar{switches} @var{infile} @var{outfile} @ovar{deffile}
18192 is an optional sequence of switches as described in the next section.
18195 is the full name of the input file, which is an Ada source
18196 file containing preprocessor directives.
18199 is the full name of the output file, which is an Ada source
18200 in standard Ada form. When used with GNAT, this file name will
18201 normally have an ads or adb suffix.
18204 is the full name of a text file containing definitions of
18205 preprocessing symbols to be referenced by the preprocessor. This argument is
18206 optional, and can be replaced by the use of the @option{-D} switch.
18210 @node Switches for gnatprep
18211 @section Switches for @code{gnatprep}
18216 @item ^-b^/BLANK_LINES^
18217 @cindex @option{^-b^/BLANK_LINES^} (@command{gnatprep})
18218 Causes both preprocessor lines and the lines deleted by
18219 preprocessing to be replaced by blank lines in the output source file,
18220 preserving line numbers in the output file.
18222 @item ^-c^/COMMENTS^
18223 @cindex @option{^-c^/COMMENTS^} (@command{gnatprep})
18224 Causes both preprocessor lines and the lines deleted
18225 by preprocessing to be retained in the output source as comments marked
18226 with the special string @code{"--! "}. This option will result in line numbers
18227 being preserved in the output file.
18229 @item ^-C^/REPLACE_IN_COMMENTS^
18230 @cindex @option{^-C^/REPLACE_IN_COMMENTS^} (@command{gnatprep})
18231 Causes comments to be scanned. Normally comments are ignored by gnatprep.
18232 If this option is specified, then comments are scanned and any $symbol
18233 substitutions performed as in program text. This is particularly useful
18234 when structured comments are used (e.g., when writing programs in the
18235 SPARK dialect of Ada). Note that this switch is not available when
18236 doing integrated preprocessing (it would be useless in this context
18237 since comments are ignored by the compiler in any case).
18239 @item ^-Dsymbol=value^/ASSOCIATE="symbol=value"^
18240 @cindex @option{^-D^/ASSOCIATE^} (@command{gnatprep})
18241 Defines a new preprocessing symbol, associated with value. If no value is given
18242 on the command line, then symbol is considered to be @code{True}. This switch
18243 can be used in place of a definition file.
18247 @cindex @option{/REMOVE} (@command{gnatprep})
18248 This is the default setting which causes lines deleted by preprocessing
18249 to be entirely removed from the output file.
18252 @item ^-r^/REFERENCE^
18253 @cindex @option{^-r^/REFERENCE^} (@command{gnatprep})
18254 Causes a @code{Source_Reference} pragma to be generated that
18255 references the original input file, so that error messages will use
18256 the file name of this original file. The use of this switch implies
18257 that preprocessor lines are not to be removed from the file, so its
18258 use will force @option{^-b^/BLANK_LINES^} mode if
18259 @option{^-c^/COMMENTS^}
18260 has not been specified explicitly.
18262 Note that if the file to be preprocessed contains multiple units, then
18263 it will be necessary to @code{gnatchop} the output file from
18264 @code{gnatprep}. If a @code{Source_Reference} pragma is present
18265 in the preprocessed file, it will be respected by
18266 @code{gnatchop ^-r^/REFERENCE^}
18267 so that the final chopped files will correctly refer to the original
18268 input source file for @code{gnatprep}.
18270 @item ^-s^/SYMBOLS^
18271 @cindex @option{^-s^/SYMBOLS^} (@command{gnatprep})
18272 Causes a sorted list of symbol names and values to be
18273 listed on the standard output file.
18275 @item ^-u^/UNDEFINED^
18276 @cindex @option{^-u^/UNDEFINED^} (@command{gnatprep})
18277 Causes undefined symbols to be treated as having the value FALSE in the context
18278 of a preprocessor test. In the absence of this option, an undefined symbol in
18279 a @code{#if} or @code{#elsif} test will be treated as an error.
18285 Note: if neither @option{-b} nor @option{-c} is present,
18286 then preprocessor lines and
18287 deleted lines are completely removed from the output, unless -r is
18288 specified, in which case -b is assumed.
18291 @node Form of Definitions File
18292 @section Form of Definitions File
18295 The definitions file contains lines of the form
18302 where symbol is a preprocessing symbol, and value is one of the following:
18306 Empty, corresponding to a null substitution
18308 A string literal using normal Ada syntax
18310 Any sequence of characters from the set
18311 (letters, digits, period, underline).
18315 Comment lines may also appear in the definitions file, starting with
18316 the usual @code{--},
18317 and comments may be added to the definitions lines.
18319 @node Form of Input Text for gnatprep
18320 @section Form of Input Text for @code{gnatprep}
18323 The input text may contain preprocessor conditional inclusion lines,
18324 as well as general symbol substitution sequences.
18326 The preprocessor conditional inclusion commands have the form
18331 #if @i{expression} @r{[}then@r{]}
18333 #elsif @i{expression} @r{[}then@r{]}
18335 #elsif @i{expression} @r{[}then@r{]}
18346 In this example, @i{expression} is defined by the following grammar:
18348 @i{expression} ::= <symbol>
18349 @i{expression} ::= <symbol> = "<value>"
18350 @i{expression} ::= <symbol> = <symbol>
18351 @i{expression} ::= <symbol> 'Defined
18352 @i{expression} ::= not @i{expression}
18353 @i{expression} ::= @i{expression} and @i{expression}
18354 @i{expression} ::= @i{expression} or @i{expression}
18355 @i{expression} ::= @i{expression} and then @i{expression}
18356 @i{expression} ::= @i{expression} or else @i{expression}
18357 @i{expression} ::= ( @i{expression} )
18360 The following restriction exists: it is not allowed to have "and" or "or"
18361 following "not" in the same expression without parentheses. For example, this
18368 This should be one of the following:
18376 For the first test (@i{expression} ::= <symbol>) the symbol must have
18377 either the value true or false, that is to say the right-hand of the
18378 symbol definition must be one of the (case-insensitive) literals
18379 @code{True} or @code{False}. If the value is true, then the
18380 corresponding lines are included, and if the value is false, they are
18383 The test (@i{expression} ::= <symbol> @code{'Defined}) is true only if
18384 the symbol has been defined in the definition file or by a @option{-D}
18385 switch on the command line. Otherwise, the test is false.
18387 The equality tests are case insensitive, as are all the preprocessor lines.
18389 If the symbol referenced is not defined in the symbol definitions file,
18390 then the effect depends on whether or not switch @option{-u}
18391 is specified. If so, then the symbol is treated as if it had the value
18392 false and the test fails. If this switch is not specified, then
18393 it is an error to reference an undefined symbol. It is also an error to
18394 reference a symbol that is defined with a value other than @code{True}
18397 The use of the @code{not} operator inverts the sense of this logical test.
18398 The @code{not} operator cannot be combined with the @code{or} or @code{and}
18399 operators, without parentheses. For example, "if not X or Y then" is not
18400 allowed, but "if (not X) or Y then" and "if not (X or Y) then" are.
18402 The @code{then} keyword is optional as shown
18404 The @code{#} must be the first non-blank character on a line, but
18405 otherwise the format is free form. Spaces or tabs may appear between
18406 the @code{#} and the keyword. The keywords and the symbols are case
18407 insensitive as in normal Ada code. Comments may be used on a
18408 preprocessor line, but other than that, no other tokens may appear on a
18409 preprocessor line. Any number of @code{elsif} clauses can be present,
18410 including none at all. The @code{else} is optional, as in Ada.
18412 The @code{#} marking the start of a preprocessor line must be the first
18413 non-blank character on the line, i.e., it must be preceded only by
18414 spaces or horizontal tabs.
18416 Symbol substitution outside of preprocessor lines is obtained by using
18424 anywhere within a source line, except in a comment or within a
18425 string literal. The identifier
18426 following the @code{$} must match one of the symbols defined in the symbol
18427 definition file, and the result is to substitute the value of the
18428 symbol in place of @code{$symbol} in the output file.
18430 Note that although the substitution of strings within a string literal
18431 is not possible, it is possible to have a symbol whose defined value is
18432 a string literal. So instead of setting XYZ to @code{hello} and writing:
18435 Header : String := "$XYZ";
18439 you should set XYZ to @code{"hello"} and write:
18442 Header : String := $XYZ;
18446 and then the substitution will occur as desired.
18449 @node The GNAT Run-Time Library Builder gnatlbr
18450 @chapter The GNAT Run-Time Library Builder @code{gnatlbr}
18452 @cindex Library builder
18455 @code{gnatlbr} is a tool for rebuilding the GNAT run time with user
18456 supplied configuration pragmas.
18459 * Running gnatlbr::
18460 * Switches for gnatlbr::
18461 * Examples of gnatlbr Usage::
18464 @node Running gnatlbr
18465 @section Running @code{gnatlbr}
18468 The @code{gnatlbr} command has the form
18471 $ GNAT LIBRARY /@r{[}CREATE@r{|}SET@r{|}DELETE@r{]}=directory @r{[}/CONFIG=file@r{]}
18474 @node Switches for gnatlbr
18475 @section Switches for @code{gnatlbr}
18478 @code{gnatlbr} recognizes the following switches:
18482 @item /CREATE=directory
18483 @cindex @code{/CREATE} (@code{gnatlbr})
18484 Create the new run-time library in the specified directory.
18486 @item /SET=directory
18487 @cindex @code{/SET} (@code{gnatlbr})
18488 Make the library in the specified directory the current run-time library.
18490 @item /DELETE=directory
18491 @cindex @code{/DELETE} (@code{gnatlbr})
18492 Delete the run-time library in the specified directory.
18495 @cindex @code{/CONFIG} (@code{gnatlbr})
18496 With /CREATE: Use the configuration pragmas in the specified file when
18497 building the library.
18499 With /SET: Use the configuration pragmas in the specified file when
18504 @node Examples of gnatlbr Usage
18505 @section Example of @code{gnatlbr} Usage
18508 Contents of VAXFLOAT.ADC:
18509 pragma Float_Representation (VAX_Float);
18511 $ GNAT LIBRARY /CREATE=[.VAXFLOAT] /CONFIG=VAXFLOAT.ADC
18513 GNAT LIBRARY rebuilds the run-time library in directory [.VAXFLOAT]
18518 @node The GNAT Library Browser gnatls
18519 @chapter The GNAT Library Browser @code{gnatls}
18521 @cindex Library browser
18524 @code{gnatls} is a tool that outputs information about compiled
18525 units. It gives the relationship between objects, unit names and source
18526 files. It can also be used to check the source dependencies of a unit
18527 as well as various characteristics.
18529 Note: to invoke @code{gnatls} with a project file, use the @code{gnat}
18530 driver (see @ref{The GNAT Driver and Project Files}).
18534 * Switches for gnatls::
18535 * Examples of gnatls Usage::
18538 @node Running gnatls
18539 @section Running @code{gnatls}
18542 The @code{gnatls} command has the form
18545 $ gnatls switches @var{object_or_ali_file}
18549 The main argument is the list of object or @file{ali} files
18550 (@pxref{The Ada Library Information Files})
18551 for which information is requested.
18553 In normal mode, without additional option, @code{gnatls} produces a
18554 four-column listing. Each line represents information for a specific
18555 object. The first column gives the full path of the object, the second
18556 column gives the name of the principal unit in this object, the third
18557 column gives the status of the source and the fourth column gives the
18558 full path of the source representing this unit.
18559 Here is a simple example of use:
18563 ^./^[]^demo1.o demo1 DIF demo1.adb
18564 ^./^[]^demo2.o demo2 OK demo2.adb
18565 ^./^[]^hello.o h1 OK hello.adb
18566 ^./^[]^instr-child.o instr.child MOK instr-child.adb
18567 ^./^[]^instr.o instr OK instr.adb
18568 ^./^[]^tef.o tef DIF tef.adb
18569 ^./^[]^text_io_example.o text_io_example OK text_io_example.adb
18570 ^./^[]^tgef.o tgef DIF tgef.adb
18574 The first line can be interpreted as follows: the main unit which is
18576 object file @file{demo1.o} is demo1, whose main source is in
18577 @file{demo1.adb}. Furthermore, the version of the source used for the
18578 compilation of demo1 has been modified (DIF). Each source file has a status
18579 qualifier which can be:
18582 @item OK (unchanged)
18583 The version of the source file used for the compilation of the
18584 specified unit corresponds exactly to the actual source file.
18586 @item MOK (slightly modified)
18587 The version of the source file used for the compilation of the
18588 specified unit differs from the actual source file but not enough to
18589 require recompilation. If you use gnatmake with the qualifier
18590 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}, a file marked
18591 MOK will not be recompiled.
18593 @item DIF (modified)
18594 No version of the source found on the path corresponds to the source
18595 used to build this object.
18597 @item ??? (file not found)
18598 No source file was found for this unit.
18600 @item HID (hidden, unchanged version not first on PATH)
18601 The version of the source that corresponds exactly to the source used
18602 for compilation has been found on the path but it is hidden by another
18603 version of the same source that has been modified.
18607 @node Switches for gnatls
18608 @section Switches for @code{gnatls}
18611 @code{gnatls} recognizes the following switches:
18615 @cindex @option{--version} @command{gnatls}
18616 Display Copyright and version, then exit disregarding all other options.
18619 @cindex @option{--help} @command{gnatls}
18620 If @option{--version} was not used, display usage, then exit disregarding
18623 @item ^-a^/ALL_UNITS^
18624 @cindex @option{^-a^/ALL_UNITS^} (@code{gnatls})
18625 Consider all units, including those of the predefined Ada library.
18626 Especially useful with @option{^-d^/DEPENDENCIES^}.
18628 @item ^-d^/DEPENDENCIES^
18629 @cindex @option{^-d^/DEPENDENCIES^} (@code{gnatls})
18630 List sources from which specified units depend on.
18632 @item ^-h^/OUTPUT=OPTIONS^
18633 @cindex @option{^-h^/OUTPUT=OPTIONS^} (@code{gnatls})
18634 Output the list of options.
18636 @item ^-o^/OUTPUT=OBJECTS^
18637 @cindex @option{^-o^/OUTPUT=OBJECTS^} (@code{gnatls})
18638 Only output information about object files.
18640 @item ^-s^/OUTPUT=SOURCES^
18641 @cindex @option{^-s^/OUTPUT=SOURCES^} (@code{gnatls})
18642 Only output information about source files.
18644 @item ^-u^/OUTPUT=UNITS^
18645 @cindex @option{^-u^/OUTPUT=UNITS^} (@code{gnatls})
18646 Only output information about compilation units.
18648 @item ^-files^/FILES^=@var{file}
18649 @cindex @option{^-files^/FILES^} (@code{gnatls})
18650 Take as arguments the files listed in text file @var{file}.
18651 Text file @var{file} may contain empty lines that are ignored.
18652 Each nonempty line should contain the name of an existing file.
18653 Several such switches may be specified simultaneously.
18655 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
18656 @itemx ^-aI^/SOURCE_SEARCH=^@var{dir}
18657 @itemx ^-I^/SEARCH=^@var{dir}
18658 @itemx ^-I-^/NOCURRENT_DIRECTORY^
18660 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatls})
18661 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatls})
18662 @cindex @option{^-I^/SEARCH^} (@code{gnatls})
18663 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatls})
18664 Source path manipulation. Same meaning as the equivalent @command{gnatmake}
18665 flags (@pxref{Switches for gnatmake}).
18667 @item --RTS=@var{rts-path}
18668 @cindex @option{--RTS} (@code{gnatls})
18669 Specifies the default location of the runtime library. Same meaning as the
18670 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
18672 @item ^-v^/OUTPUT=VERBOSE^
18673 @cindex @option{^-v^/OUTPUT=VERBOSE^} (@code{gnatls})
18674 Verbose mode. Output the complete source, object and project paths. Do not use
18675 the default column layout but instead use long format giving as much as
18676 information possible on each requested units, including special
18677 characteristics such as:
18680 @item Preelaborable
18681 The unit is preelaborable in the Ada sense.
18684 No elaboration code has been produced by the compiler for this unit.
18687 The unit is pure in the Ada sense.
18689 @item Elaborate_Body
18690 The unit contains a pragma Elaborate_Body.
18693 The unit contains a pragma Remote_Types.
18695 @item Shared_Passive
18696 The unit contains a pragma Shared_Passive.
18699 This unit is part of the predefined environment and cannot be modified
18702 @item Remote_Call_Interface
18703 The unit contains a pragma Remote_Call_Interface.
18709 @node Examples of gnatls Usage
18710 @section Example of @code{gnatls} Usage
18714 Example of using the verbose switch. Note how the source and
18715 object paths are affected by the -I switch.
18718 $ gnatls -v -I.. demo1.o
18720 GNATLS 5.03w (20041123-34)
18721 Copyright 1997-2004 Free Software Foundation, Inc.
18723 Source Search Path:
18724 <Current_Directory>
18726 /home/comar/local/adainclude/
18728 Object Search Path:
18729 <Current_Directory>
18731 /home/comar/local/lib/gcc-lib/x86-linux/3.4.3/adalib/
18733 Project Search Path:
18734 <Current_Directory>
18735 /home/comar/local/lib/gnat/
18740 Kind => subprogram body
18741 Flags => No_Elab_Code
18742 Source => demo1.adb modified
18746 The following is an example of use of the dependency list.
18747 Note the use of the -s switch
18748 which gives a straight list of source files. This can be useful for
18749 building specialized scripts.
18752 $ gnatls -d demo2.o
18753 ./demo2.o demo2 OK demo2.adb
18759 $ gnatls -d -s -a demo1.o
18761 /home/comar/local/adainclude/ada.ads
18762 /home/comar/local/adainclude/a-finali.ads
18763 /home/comar/local/adainclude/a-filico.ads
18764 /home/comar/local/adainclude/a-stream.ads
18765 /home/comar/local/adainclude/a-tags.ads
18768 /home/comar/local/adainclude/gnat.ads
18769 /home/comar/local/adainclude/g-io.ads
18771 /home/comar/local/adainclude/system.ads
18772 /home/comar/local/adainclude/s-exctab.ads
18773 /home/comar/local/adainclude/s-finimp.ads
18774 /home/comar/local/adainclude/s-finroo.ads
18775 /home/comar/local/adainclude/s-secsta.ads
18776 /home/comar/local/adainclude/s-stalib.ads
18777 /home/comar/local/adainclude/s-stoele.ads
18778 /home/comar/local/adainclude/s-stratt.ads
18779 /home/comar/local/adainclude/s-tasoli.ads
18780 /home/comar/local/adainclude/s-unstyp.ads
18781 /home/comar/local/adainclude/unchconv.ads
18787 GNAT LIST /DEPENDENCIES /OUTPUT=SOURCES /ALL_UNITS DEMO1.ADB
18789 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]ada.ads
18790 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-finali.ads
18791 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-filico.ads
18792 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-stream.ads
18793 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-tags.ads
18797 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]gnat.ads
18798 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]g-io.ads
18800 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]system.ads
18801 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-exctab.ads
18802 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finimp.ads
18803 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finroo.ads
18804 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-secsta.ads
18805 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stalib.ads
18806 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stoele.ads
18807 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stratt.ads
18808 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-tasoli.ads
18809 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-unstyp.ads
18810 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]unchconv.ads
18814 @node Cleaning Up Using gnatclean
18815 @chapter Cleaning Up Using @code{gnatclean}
18817 @cindex Cleaning tool
18820 @code{gnatclean} is a tool that allows the deletion of files produced by the
18821 compiler, binder and linker, including ALI files, object files, tree files,
18822 expanded source files, library files, interface copy source files, binder
18823 generated files and executable files.
18826 * Running gnatclean::
18827 * Switches for gnatclean::
18828 @c * Examples of gnatclean Usage::
18831 @node Running gnatclean
18832 @section Running @code{gnatclean}
18835 The @code{gnatclean} command has the form:
18838 $ gnatclean switches @var{names}
18842 @var{names} is a list of source file names. Suffixes @code{.^ads^ADS^} and
18843 @code{^adb^ADB^} may be omitted. If a project file is specified using switch
18844 @code{^-P^/PROJECT_FILE=^}, then @var{names} may be completely omitted.
18847 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
18848 if switch @code{^-c^/COMPILER_FILES_ONLY^} is not specified, by the binder and
18849 the linker. In informative-only mode, specified by switch
18850 @code{^-n^/NODELETE^}, the list of files that would have been deleted in
18851 normal mode is listed, but no file is actually deleted.
18853 @node Switches for gnatclean
18854 @section Switches for @code{gnatclean}
18857 @code{gnatclean} recognizes the following switches:
18861 @cindex @option{--version} @command{gnatclean}
18862 Display Copyright and version, then exit disregarding all other options.
18865 @cindex @option{--help} @command{gnatclean}
18866 If @option{--version} was not used, display usage, then exit disregarding
18869 @item ^-c^/COMPILER_FILES_ONLY^
18870 @cindex @option{^-c^/COMPILER_FILES_ONLY^} (@code{gnatclean})
18871 Only attempt to delete the files produced by the compiler, not those produced
18872 by the binder or the linker. The files that are not to be deleted are library
18873 files, interface copy files, binder generated files and executable files.
18875 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
18876 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatclean})
18877 Indicate that ALI and object files should normally be found in directory
18880 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
18881 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatclean})
18882 When using project files, if some errors or warnings are detected during
18883 parsing and verbose mode is not in effect (no use of switch
18884 ^-v^/VERBOSE^), then error lines start with the full path name of the project
18885 file, rather than its simple file name.
18888 @cindex @option{^-h^/HELP^} (@code{gnatclean})
18889 Output a message explaining the usage of @code{^gnatclean^gnatclean^}.
18891 @item ^-n^/NODELETE^
18892 @cindex @option{^-n^/NODELETE^} (@code{gnatclean})
18893 Informative-only mode. Do not delete any files. Output the list of the files
18894 that would have been deleted if this switch was not specified.
18896 @item ^-P^/PROJECT_FILE=^@var{project}
18897 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatclean})
18898 Use project file @var{project}. Only one such switch can be used.
18899 When cleaning a project file, the files produced by the compilation of the
18900 immediate sources or inherited sources of the project files are to be
18901 deleted. This is not depending on the presence or not of executable names
18902 on the command line.
18905 @cindex @option{^-q^/QUIET^} (@code{gnatclean})
18906 Quiet output. If there are no errors, do not output anything, except in
18907 verbose mode (switch ^-v^/VERBOSE^) or in informative-only mode
18908 (switch ^-n^/NODELETE^).
18910 @item ^-r^/RECURSIVE^
18911 @cindex @option{^-r^/RECURSIVE^} (@code{gnatclean})
18912 When a project file is specified (using switch ^-P^/PROJECT_FILE=^),
18913 clean all imported and extended project files, recursively. If this switch
18914 is not specified, only the files related to the main project file are to be
18915 deleted. This switch has no effect if no project file is specified.
18917 @item ^-v^/VERBOSE^
18918 @cindex @option{^-v^/VERBOSE^} (@code{gnatclean})
18921 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
18922 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (@code{gnatclean})
18923 Indicates the verbosity of the parsing of GNAT project files.
18924 @xref{Switches Related to Project Files}.
18926 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
18927 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (@code{gnatclean})
18928 Indicates that external variable @var{name} has the value @var{value}.
18929 The Project Manager will use this value for occurrences of
18930 @code{external(name)} when parsing the project file.
18931 @xref{Switches Related to Project Files}.
18933 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
18934 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatclean})
18935 When searching for ALI and object files, look in directory
18938 @item ^-I^/SEARCH=^@var{dir}
18939 @cindex @option{^-I^/SEARCH^} (@code{gnatclean})
18940 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}}.
18942 @item ^-I-^/NOCURRENT_DIRECTORY^
18943 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatclean})
18944 @cindex Source files, suppressing search
18945 Do not look for ALI or object files in the directory
18946 where @code{gnatclean} was invoked.
18950 @c @node Examples of gnatclean Usage
18951 @c @section Examples of @code{gnatclean} Usage
18954 @node GNAT and Libraries
18955 @chapter GNAT and Libraries
18956 @cindex Library, building, installing, using
18959 This chapter describes how to build and use libraries with GNAT, and also shows
18960 how to recompile the GNAT run-time library. You should be familiar with the
18961 Project Manager facility (@pxref{GNAT Project Manager}) before reading this
18965 * Introduction to Libraries in GNAT::
18966 * General Ada Libraries::
18967 * Stand-alone Ada Libraries::
18968 * Rebuilding the GNAT Run-Time Library::
18971 @node Introduction to Libraries in GNAT
18972 @section Introduction to Libraries in GNAT
18975 A library is, conceptually, a collection of objects which does not have its
18976 own main thread of execution, but rather provides certain services to the
18977 applications that use it. A library can be either statically linked with the
18978 application, in which case its code is directly included in the application,
18979 or, on platforms that support it, be dynamically linked, in which case
18980 its code is shared by all applications making use of this library.
18982 GNAT supports both types of libraries.
18983 In the static case, the compiled code can be provided in different ways. The
18984 simplest approach is to provide directly the set of objects resulting from
18985 compilation of the library source files. Alternatively, you can group the
18986 objects into an archive using whatever commands are provided by the operating
18987 system. For the latter case, the objects are grouped into a shared library.
18989 In the GNAT environment, a library has three types of components:
18995 @xref{The Ada Library Information Files}.
18997 Object files, an archive or a shared library.
19001 A GNAT library may expose all its source files, which is useful for
19002 documentation purposes. Alternatively, it may expose only the units needed by
19003 an external user to make use of the library. That is to say, the specs
19004 reflecting the library services along with all the units needed to compile
19005 those specs, which can include generic bodies or any body implementing an
19006 inlined routine. In the case of @emph{stand-alone libraries} those exposed
19007 units are called @emph{interface units} (@pxref{Stand-alone Ada Libraries}).
19009 All compilation units comprising an application, including those in a library,
19010 need to be elaborated in an order partially defined by Ada's semantics. GNAT
19011 computes the elaboration order from the @file{ALI} files and this is why they
19012 constitute a mandatory part of GNAT libraries.
19013 @emph{Stand-alone libraries} are the exception to this rule because a specific
19014 library elaboration routine is produced independently of the application(s)
19017 @node General Ada Libraries
19018 @section General Ada Libraries
19021 * Building a library::
19022 * Installing a library::
19023 * Using a library::
19026 @node Building a library
19027 @subsection Building a library
19030 The easiest way to build a library is to use the Project Manager,
19031 which supports a special type of project called a @emph{Library Project}
19032 (@pxref{Library Projects}).
19034 A project is considered a library project, when two project-level attributes
19035 are defined in it: @code{Library_Name} and @code{Library_Dir}. In order to
19036 control different aspects of library configuration, additional optional
19037 project-level attributes can be specified:
19040 This attribute controls whether the library is to be static or dynamic
19042 @item Library_Version
19043 This attribute specifies the library version; this value is used
19044 during dynamic linking of shared libraries to determine if the currently
19045 installed versions of the binaries are compatible.
19047 @item Library_Options
19049 These attributes specify additional low-level options to be used during
19050 library generation, and redefine the actual application used to generate
19055 The GNAT Project Manager takes full care of the library maintenance task,
19056 including recompilation of the source files for which objects do not exist
19057 or are not up to date, assembly of the library archive, and installation of
19058 the library (i.e., copying associated source, object and @file{ALI} files
19059 to the specified location).
19061 Here is a simple library project file:
19062 @smallexample @c ada
19064 for Source_Dirs use ("src1", "src2");
19065 for Object_Dir use "obj";
19066 for Library_Name use "mylib";
19067 for Library_Dir use "lib";
19068 for Library_Kind use "dynamic";
19073 and the compilation command to build and install the library:
19075 @smallexample @c ada
19076 $ gnatmake -Pmy_lib
19080 It is not entirely trivial to perform manually all the steps required to
19081 produce a library. We recommend that you use the GNAT Project Manager
19082 for this task. In special cases where this is not desired, the necessary
19083 steps are discussed below.
19085 There are various possibilities for compiling the units that make up the
19086 library: for example with a Makefile (@pxref{Using the GNU make Utility}) or
19087 with a conventional script. For simple libraries, it is also possible to create
19088 a dummy main program which depends upon all the packages that comprise the
19089 interface of the library. This dummy main program can then be given to
19090 @command{gnatmake}, which will ensure that all necessary objects are built.
19092 After this task is accomplished, you should follow the standard procedure
19093 of the underlying operating system to produce the static or shared library.
19095 Here is an example of such a dummy program:
19096 @smallexample @c ada
19098 with My_Lib.Service1;
19099 with My_Lib.Service2;
19100 with My_Lib.Service3;
19101 procedure My_Lib_Dummy is
19109 Here are the generic commands that will build an archive or a shared library.
19112 # compiling the library
19113 $ gnatmake -c my_lib_dummy.adb
19115 # we don't need the dummy object itself
19116 $ rm my_lib_dummy.o my_lib_dummy.ali
19118 # create an archive with the remaining objects
19119 $ ar rc libmy_lib.a *.o
19120 # some systems may require "ranlib" to be run as well
19122 # or create a shared library
19123 $ gcc -shared -o libmy_lib.so *.o
19124 # some systems may require the code to have been compiled with -fPIC
19126 # remove the object files that are now in the library
19129 # Make the ALI files read-only so that gnatmake will not try to
19130 # regenerate the objects that are in the library
19135 Please note that the library must have a name of the form @file{lib@var{xxx}.a}
19136 or @file{lib@var{xxx}.so} (or @file{lib@var{xxx}.dll} on Windows) in order to
19137 be accessed by the directive @option{-l@var{xxx}} at link time.
19139 @node Installing a library
19140 @subsection Installing a library
19141 @cindex @code{ADA_PROJECT_PATH}
19142 @cindex @code{GPR_PROJECT_PATH}
19145 If you use project files, library installation is part of the library build
19146 process. Thus no further action is needed in order to make use of the
19147 libraries that are built as part of the general application build. A usable
19148 version of the library is installed in the directory specified by the
19149 @code{Library_Dir} attribute of the library project file.
19151 You may want to install a library in a context different from where the library
19152 is built. This situation arises with third party suppliers, who may want
19153 to distribute a library in binary form where the user is not expected to be
19154 able to recompile the library. The simplest option in this case is to provide
19155 a project file slightly different from the one used to build the library, by
19156 using the @code{externally_built} attribute. For instance, the project
19157 file used to build the library in the previous section can be changed into the
19158 following one when the library is installed:
19160 @smallexample @c projectfile
19162 for Source_Dirs use ("src1", "src2");
19163 for Library_Name use "mylib";
19164 for Library_Dir use "lib";
19165 for Library_Kind use "dynamic";
19166 for Externally_Built use "true";
19171 This project file assumes that the directories @file{src1},
19172 @file{src2}, and @file{lib} exist in
19173 the directory containing the project file. The @code{externally_built}
19174 attribute makes it clear to the GNAT builder that it should not attempt to
19175 recompile any of the units from this library. It allows the library provider to
19176 restrict the source set to the minimum necessary for clients to make use of the
19177 library as described in the first section of this chapter. It is the
19178 responsibility of the library provider to install the necessary sources, ALI
19179 files and libraries in the directories mentioned in the project file. For
19180 convenience, the user's library project file should be installed in a location
19181 that will be searched automatically by the GNAT
19182 builder. These are the directories referenced in the @env{GPR_PROJECT_PATH}
19183 environment variable (@pxref{Importing Projects}), and also the default GNAT
19184 library location that can be queried with @command{gnatls -v} and is usually of
19185 the form $gnat_install_root/lib/gnat.
19187 When project files are not an option, it is also possible, but not recommended,
19188 to install the library so that the sources needed to use the library are on the
19189 Ada source path and the ALI files & libraries be on the Ada Object path (see
19190 @ref{Search Paths and the Run-Time Library (RTL)}. Alternatively, the system
19191 administrator can place general-purpose libraries in the default compiler
19192 paths, by specifying the libraries' location in the configuration files
19193 @file{ada_source_path} and @file{ada_object_path}. These configuration files
19194 must be located in the GNAT installation tree at the same place as the gcc spec
19195 file. The location of the gcc spec file can be determined as follows:
19201 The configuration files mentioned above have a simple format: each line
19202 must contain one unique directory name.
19203 Those names are added to the corresponding path
19204 in their order of appearance in the file. The names can be either absolute
19205 or relative; in the latter case, they are relative to where theses files
19208 The files @file{ada_source_path} and @file{ada_object_path} might not be
19210 GNAT installation, in which case, GNAT will look for its run-time library in
19211 the directories @file{adainclude} (for the sources) and @file{adalib} (for the
19212 objects and @file{ALI} files). When the files exist, the compiler does not
19213 look in @file{adainclude} and @file{adalib}, and thus the
19214 @file{ada_source_path} file
19215 must contain the location for the GNAT run-time sources (which can simply
19216 be @file{adainclude}). In the same way, the @file{ada_object_path} file must
19217 contain the location for the GNAT run-time objects (which can simply
19220 You can also specify a new default path to the run-time library at compilation
19221 time with the switch @option{--RTS=rts-path}. You can thus choose / change
19222 the run-time library you want your program to be compiled with. This switch is
19223 recognized by @command{gcc}, @command{gnatmake}, @command{gnatbind},
19224 @command{gnatls}, @command{gnatfind} and @command{gnatxref}.
19226 It is possible to install a library before or after the standard GNAT
19227 library, by reordering the lines in the configuration files. In general, a
19228 library must be installed before the GNAT library if it redefines
19231 @node Using a library
19232 @subsection Using a library
19234 @noindent Once again, the project facility greatly simplifies the use of
19235 libraries. In this context, using a library is just a matter of adding a
19236 @code{with} clause in the user project. For instance, to make use of the
19237 library @code{My_Lib} shown in examples in earlier sections, you can
19240 @smallexample @c projectfile
19247 Even if you have a third-party, non-Ada library, you can still use GNAT's
19248 Project Manager facility to provide a wrapper for it. For example, the
19249 following project, when @code{with}ed by your main project, will link with the
19250 third-party library @file{liba.a}:
19252 @smallexample @c projectfile
19255 for Externally_Built use "true";
19256 for Source_Files use ();
19257 for Library_Dir use "lib";
19258 for Library_Name use "a";
19259 for Library_Kind use "static";
19263 This is an alternative to the use of @code{pragma Linker_Options}. It is
19264 especially interesting in the context of systems with several interdependent
19265 static libraries where finding a proper linker order is not easy and best be
19266 left to the tools having visibility over project dependence information.
19269 In order to use an Ada library manually, you need to make sure that this
19270 library is on both your source and object path
19271 (see @ref{Search Paths and the Run-Time Library (RTL)}
19272 and @ref{Search Paths for gnatbind}). Furthermore, when the objects are grouped
19273 in an archive or a shared library, you need to specify the desired
19274 library at link time.
19276 For example, you can use the library @file{mylib} installed in
19277 @file{/dir/my_lib_src} and @file{/dir/my_lib_obj} with the following commands:
19280 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \
19285 This can be expressed more simply:
19290 when the following conditions are met:
19293 @file{/dir/my_lib_src} has been added by the user to the environment
19294 variable @env{ADA_INCLUDE_PATH}, or by the administrator to the file
19295 @file{ada_source_path}
19297 @file{/dir/my_lib_obj} has been added by the user to the environment
19298 variable @env{ADA_OBJECTS_PATH}, or by the administrator to the file
19299 @file{ada_object_path}
19301 a pragma @code{Linker_Options} has been added to one of the sources.
19304 @smallexample @c ada
19305 pragma Linker_Options ("-lmy_lib");
19309 @node Stand-alone Ada Libraries
19310 @section Stand-alone Ada Libraries
19311 @cindex Stand-alone library, building, using
19314 * Introduction to Stand-alone Libraries::
19315 * Building a Stand-alone Library::
19316 * Creating a Stand-alone Library to be used in a non-Ada context::
19317 * Restrictions in Stand-alone Libraries::
19320 @node Introduction to Stand-alone Libraries
19321 @subsection Introduction to Stand-alone Libraries
19324 A Stand-alone Library (abbreviated ``SAL'') is a library that contains the
19326 elaborate the Ada units that are included in the library. In contrast with
19327 an ordinary library, which consists of all sources, objects and @file{ALI}
19329 library, a SAL may specify a restricted subset of compilation units
19330 to serve as a library interface. In this case, the fully
19331 self-sufficient set of files will normally consist of an objects
19332 archive, the sources of interface units' specs, and the @file{ALI}
19333 files of interface units.
19334 If an interface spec contains a generic unit or an inlined subprogram,
19336 source must also be provided; if the units that must be provided in the source
19337 form depend on other units, the source and @file{ALI} files of those must
19340 The main purpose of a SAL is to minimize the recompilation overhead of client
19341 applications when a new version of the library is installed. Specifically,
19342 if the interface sources have not changed, client applications do not need to
19343 be recompiled. If, furthermore, a SAL is provided in the shared form and its
19344 version, controlled by @code{Library_Version} attribute, is not changed,
19345 then the clients do not need to be relinked.
19347 SALs also allow the library providers to minimize the amount of library source
19348 text exposed to the clients. Such ``information hiding'' might be useful or
19349 necessary for various reasons.
19351 Stand-alone libraries are also well suited to be used in an executable whose
19352 main routine is not written in Ada.
19354 @node Building a Stand-alone Library
19355 @subsection Building a Stand-alone Library
19358 GNAT's Project facility provides a simple way of building and installing
19359 stand-alone libraries; see @ref{Stand-alone Library Projects}.
19360 To be a Stand-alone Library Project, in addition to the two attributes
19361 that make a project a Library Project (@code{Library_Name} and
19362 @code{Library_Dir}; see @ref{Library Projects}), the attribute
19363 @code{Library_Interface} must be defined. For example:
19365 @smallexample @c projectfile
19367 for Library_Dir use "lib_dir";
19368 for Library_Name use "dummy";
19369 for Library_Interface use ("int1", "int1.child");
19374 Attribute @code{Library_Interface} has a non-empty string list value,
19375 each string in the list designating a unit contained in an immediate source
19376 of the project file.
19378 When a Stand-alone Library is built, first the binder is invoked to build
19379 a package whose name depends on the library name
19380 (@file{^b~dummy.ads/b^B$DUMMY.ADS/B^} in the example above).
19381 This binder-generated package includes initialization and
19382 finalization procedures whose
19383 names depend on the library name (@code{dummyinit} and @code{dummyfinal}
19385 above). The object corresponding to this package is included in the library.
19387 You must ensure timely (e.g., prior to any use of interfaces in the SAL)
19388 calling of these procedures if a static SAL is built, or if a shared SAL
19390 with the project-level attribute @code{Library_Auto_Init} set to
19393 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
19394 (those that are listed in attribute @code{Library_Interface}) are copied to
19395 the Library Directory. As a consequence, only the Interface Units may be
19396 imported from Ada units outside of the library. If other units are imported,
19397 the binding phase will fail.
19399 The attribute @code{Library_Src_Dir} may be specified for a
19400 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
19401 single string value. Its value must be the path (absolute or relative to the
19402 project directory) of an existing directory. This directory cannot be the
19403 object directory or one of the source directories, but it can be the same as
19404 the library directory. The sources of the Interface
19405 Units of the library that are needed by an Ada client of the library will be
19406 copied to the designated directory, called the Interface Copy directory.
19407 These sources include the specs of the Interface Units, but they may also
19408 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
19409 are used, or when there is a generic unit in the spec. Before the sources
19410 are copied to the Interface Copy directory, an attempt is made to delete all
19411 files in the Interface Copy directory.
19413 Building stand-alone libraries by hand is somewhat tedious, but for those
19414 occasions when it is necessary here are the steps that you need to perform:
19417 Compile all library sources.
19420 Invoke the binder with the switch @option{-n} (No Ada main program),
19421 with all the @file{ALI} files of the interfaces, and
19422 with the switch @option{-L} to give specific names to the @code{init}
19423 and @code{final} procedures. For example:
19425 gnatbind -n int1.ali int2.ali -Lsal1
19429 Compile the binder generated file:
19435 Link the dynamic library with all the necessary object files,
19436 indicating to the linker the names of the @code{init} (and possibly
19437 @code{final}) procedures for automatic initialization (and finalization).
19438 The built library should be placed in a directory different from
19439 the object directory.
19442 Copy the @code{ALI} files of the interface to the library directory,
19443 add in this copy an indication that it is an interface to a SAL
19444 (i.e., add a word @option{SL} on the line in the @file{ALI} file that starts
19445 with letter ``P'') and make the modified copy of the @file{ALI} file
19450 Using SALs is not different from using other libraries
19451 (see @ref{Using a library}).
19453 @node Creating a Stand-alone Library to be used in a non-Ada context
19454 @subsection Creating a Stand-alone Library to be used in a non-Ada context
19457 It is easy to adapt the SAL build procedure discussed above for use of a SAL in
19460 The only extra step required is to ensure that library interface subprograms
19461 are compatible with the main program, by means of @code{pragma Export}
19462 or @code{pragma Convention}.
19464 Here is an example of simple library interface for use with C main program:
19466 @smallexample @c ada
19467 package My_Package is
19469 procedure Do_Something;
19470 pragma Export (C, Do_Something, "do_something");
19472 procedure Do_Something_Else;
19473 pragma Export (C, Do_Something_Else, "do_something_else");
19479 On the foreign language side, you must provide a ``foreign'' view of the
19480 library interface; remember that it should contain elaboration routines in
19481 addition to interface subprograms.
19483 The example below shows the content of @code{mylib_interface.h} (note
19484 that there is no rule for the naming of this file, any name can be used)
19486 /* the library elaboration procedure */
19487 extern void mylibinit (void);
19489 /* the library finalization procedure */
19490 extern void mylibfinal (void);
19492 /* the interface exported by the library */
19493 extern void do_something (void);
19494 extern void do_something_else (void);
19498 Libraries built as explained above can be used from any program, provided
19499 that the elaboration procedures (named @code{mylibinit} in the previous
19500 example) are called before the library services are used. Any number of
19501 libraries can be used simultaneously, as long as the elaboration
19502 procedure of each library is called.
19504 Below is an example of a C program that uses the @code{mylib} library.
19507 #include "mylib_interface.h"
19512 /* First, elaborate the library before using it */
19515 /* Main program, using the library exported entities */
19517 do_something_else ();
19519 /* Library finalization at the end of the program */
19526 Note that invoking any library finalization procedure generated by
19527 @code{gnatbind} shuts down the Ada run-time environment.
19529 finalization of all Ada libraries must be performed at the end of the program.
19530 No call to these libraries or to the Ada run-time library should be made
19531 after the finalization phase.
19533 @node Restrictions in Stand-alone Libraries
19534 @subsection Restrictions in Stand-alone Libraries
19537 The pragmas listed below should be used with caution inside libraries,
19538 as they can create incompatibilities with other Ada libraries:
19540 @item pragma @code{Locking_Policy}
19541 @item pragma @code{Queuing_Policy}
19542 @item pragma @code{Task_Dispatching_Policy}
19543 @item pragma @code{Unreserve_All_Interrupts}
19547 When using a library that contains such pragmas, the user must make sure
19548 that all libraries use the same pragmas with the same values. Otherwise,
19549 @code{Program_Error} will
19550 be raised during the elaboration of the conflicting
19551 libraries. The usage of these pragmas and its consequences for the user
19552 should therefore be well documented.
19554 Similarly, the traceback in the exception occurrence mechanism should be
19555 enabled or disabled in a consistent manner across all libraries.
19556 Otherwise, Program_Error will be raised during the elaboration of the
19557 conflicting libraries.
19559 If the @code{Version} or @code{Body_Version}
19560 attributes are used inside a library, then you need to
19561 perform a @code{gnatbind} step that specifies all @file{ALI} files in all
19562 libraries, so that version identifiers can be properly computed.
19563 In practice these attributes are rarely used, so this is unlikely
19564 to be a consideration.
19566 @node Rebuilding the GNAT Run-Time Library
19567 @section Rebuilding the GNAT Run-Time Library
19568 @cindex GNAT Run-Time Library, rebuilding
19569 @cindex Building the GNAT Run-Time Library
19570 @cindex Rebuilding the GNAT Run-Time Library
19571 @cindex Run-Time Library, rebuilding
19574 It may be useful to recompile the GNAT library in various contexts, the
19575 most important one being the use of partition-wide configuration pragmas
19576 such as @code{Normalize_Scalars}. A special Makefile called
19577 @code{Makefile.adalib} is provided to that effect and can be found in
19578 the directory containing the GNAT library. The location of this
19579 directory depends on the way the GNAT environment has been installed and can
19580 be determined by means of the command:
19587 The last entry in the object search path usually contains the
19588 gnat library. This Makefile contains its own documentation and in
19589 particular the set of instructions needed to rebuild a new library and
19592 @node Using the GNU make Utility
19593 @chapter Using the GNU @code{make} Utility
19597 This chapter offers some examples of makefiles that solve specific
19598 problems. It does not explain how to write a makefile (@pxref{Top,, GNU
19599 make, make, GNU @code{make}}), nor does it try to replace the
19600 @command{gnatmake} utility (@pxref{The GNAT Make Program gnatmake}).
19602 All the examples in this section are specific to the GNU version of
19603 make. Although @command{make} is a standard utility, and the basic language
19604 is the same, these examples use some advanced features found only in
19608 * Using gnatmake in a Makefile::
19609 * Automatically Creating a List of Directories::
19610 * Generating the Command Line Switches::
19611 * Overcoming Command Line Length Limits::
19614 @node Using gnatmake in a Makefile
19615 @section Using gnatmake in a Makefile
19620 Complex project organizations can be handled in a very powerful way by
19621 using GNU make combined with gnatmake. For instance, here is a Makefile
19622 which allows you to build each subsystem of a big project into a separate
19623 shared library. Such a makefile allows you to significantly reduce the link
19624 time of very big applications while maintaining full coherence at
19625 each step of the build process.
19627 The list of dependencies are handled automatically by
19628 @command{gnatmake}. The Makefile is simply used to call gnatmake in each of
19629 the appropriate directories.
19631 Note that you should also read the example on how to automatically
19632 create the list of directories
19633 (@pxref{Automatically Creating a List of Directories})
19634 which might help you in case your project has a lot of subdirectories.
19639 @font@heightrm=cmr8
19642 ## This Makefile is intended to be used with the following directory
19644 ## - The sources are split into a series of csc (computer software components)
19645 ## Each of these csc is put in its own directory.
19646 ## Their name are referenced by the directory names.
19647 ## They will be compiled into shared library (although this would also work
19648 ## with static libraries
19649 ## - The main program (and possibly other packages that do not belong to any
19650 ## csc is put in the top level directory (where the Makefile is).
19651 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
19652 ## \_ second_csc (sources) __ lib (will contain the library)
19654 ## Although this Makefile is build for shared library, it is easy to modify
19655 ## to build partial link objects instead (modify the lines with -shared and
19658 ## With this makefile, you can change any file in the system or add any new
19659 ## file, and everything will be recompiled correctly (only the relevant shared
19660 ## objects will be recompiled, and the main program will be re-linked).
19662 # The list of computer software component for your project. This might be
19663 # generated automatically.
19666 # Name of the main program (no extension)
19669 # If we need to build objects with -fPIC, uncomment the following line
19672 # The following variable should give the directory containing libgnat.so
19673 # You can get this directory through 'gnatls -v'. This is usually the last
19674 # directory in the Object_Path.
19677 # The directories for the libraries
19678 # (This macro expands the list of CSC to the list of shared libraries, you
19679 # could simply use the expanded form:
19680 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
19681 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
19683 $@{MAIN@}: objects $@{LIB_DIR@}
19684 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
19685 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
19688 # recompile the sources
19689 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
19691 # Note: In a future version of GNAT, the following commands will be simplified
19692 # by a new tool, gnatmlib
19694 mkdir -p $@{dir $@@ @}
19695 cd $@{dir $@@ @} && gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
19696 cd $@{dir $@@ @} && cp -f ../*.ali .
19698 # The dependencies for the modules
19699 # Note that we have to force the expansion of *.o, since in some cases
19700 # make won't be able to do it itself.
19701 aa/lib/libaa.so: $@{wildcard aa/*.o@}
19702 bb/lib/libbb.so: $@{wildcard bb/*.o@}
19703 cc/lib/libcc.so: $@{wildcard cc/*.o@}
19705 # Make sure all of the shared libraries are in the path before starting the
19708 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
19711 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
19712 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
19713 $@{RM@} $@{CSC_LIST:%=%/*.o@}
19714 $@{RM@} *.o *.ali $@{MAIN@}
19717 @node Automatically Creating a List of Directories
19718 @section Automatically Creating a List of Directories
19721 In most makefiles, you will have to specify a list of directories, and
19722 store it in a variable. For small projects, it is often easier to
19723 specify each of them by hand, since you then have full control over what
19724 is the proper order for these directories, which ones should be
19727 However, in larger projects, which might involve hundreds of
19728 subdirectories, it might be more convenient to generate this list
19731 The example below presents two methods. The first one, although less
19732 general, gives you more control over the list. It involves wildcard
19733 characters, that are automatically expanded by @command{make}. Its
19734 shortcoming is that you need to explicitly specify some of the
19735 organization of your project, such as for instance the directory tree
19736 depth, whether some directories are found in a separate tree, @enddots{}
19738 The second method is the most general one. It requires an external
19739 program, called @command{find}, which is standard on all Unix systems. All
19740 the directories found under a given root directory will be added to the
19746 @font@heightrm=cmr8
19749 # The examples below are based on the following directory hierarchy:
19750 # All the directories can contain any number of files
19751 # ROOT_DIRECTORY -> a -> aa -> aaa
19754 # -> b -> ba -> baa
19757 # This Makefile creates a variable called DIRS, that can be reused any time
19758 # you need this list (see the other examples in this section)
19760 # The root of your project's directory hierarchy
19764 # First method: specify explicitly the list of directories
19765 # This allows you to specify any subset of all the directories you need.
19768 DIRS := a/aa/ a/ab/ b/ba/
19771 # Second method: use wildcards
19772 # Note that the argument(s) to wildcard below should end with a '/'.
19773 # Since wildcards also return file names, we have to filter them out
19774 # to avoid duplicate directory names.
19775 # We thus use make's @code{dir} and @code{sort} functions.
19776 # It sets DIRs to the following value (note that the directories aaa and baa
19777 # are not given, unless you change the arguments to wildcard).
19778 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
19781 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
19782 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
19785 # Third method: use an external program
19786 # This command is much faster if run on local disks, avoiding NFS slowdowns.
19787 # This is the most complete command: it sets DIRs to the following value:
19788 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
19791 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
19795 @node Generating the Command Line Switches
19796 @section Generating the Command Line Switches
19799 Once you have created the list of directories as explained in the
19800 previous section (@pxref{Automatically Creating a List of Directories}),
19801 you can easily generate the command line arguments to pass to gnatmake.
19803 For the sake of completeness, this example assumes that the source path
19804 is not the same as the object path, and that you have two separate lists
19808 # see "Automatically creating a list of directories" to create
19813 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
19814 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
19817 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
19820 @node Overcoming Command Line Length Limits
19821 @section Overcoming Command Line Length Limits
19824 One problem that might be encountered on big projects is that many
19825 operating systems limit the length of the command line. It is thus hard to give
19826 gnatmake the list of source and object directories.
19828 This example shows how you can set up environment variables, which will
19829 make @command{gnatmake} behave exactly as if the directories had been
19830 specified on the command line, but have a much higher length limit (or
19831 even none on most systems).
19833 It assumes that you have created a list of directories in your Makefile,
19834 using one of the methods presented in
19835 @ref{Automatically Creating a List of Directories}.
19836 For the sake of completeness, we assume that the object
19837 path (where the ALI files are found) is different from the sources patch.
19839 Note a small trick in the Makefile below: for efficiency reasons, we
19840 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
19841 expanded immediately by @code{make}. This way we overcome the standard
19842 make behavior which is to expand the variables only when they are
19845 On Windows, if you are using the standard Windows command shell, you must
19846 replace colons with semicolons in the assignments to these variables.
19851 @font@heightrm=cmr8
19854 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECT_PATH.
19855 # This is the same thing as putting the -I arguments on the command line.
19856 # (the equivalent of using -aI on the command line would be to define
19857 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECT_PATH).
19858 # You can of course have different values for these variables.
19860 # Note also that we need to keep the previous values of these variables, since
19861 # they might have been set before running 'make' to specify where the GNAT
19862 # library is installed.
19864 # see "Automatically creating a list of directories" to create these
19870 space:=$@{empty@} $@{empty@}
19871 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
19872 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
19873 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
19874 ADA_OBJECT_PATH += $@{OBJECT_LIST@}
19875 export ADA_INCLUDE_PATH
19876 export ADA_OBJECT_PATH
19883 @node Memory Management Issues
19884 @chapter Memory Management Issues
19887 This chapter describes some useful memory pools provided in the GNAT library
19888 and in particular the GNAT Debug Pool facility, which can be used to detect
19889 incorrect uses of access values (including ``dangling references'').
19891 It also describes the @command{gnatmem} tool, which can be used to track down
19896 * Some Useful Memory Pools::
19897 * The GNAT Debug Pool Facility::
19899 * The gnatmem Tool::
19903 @node Some Useful Memory Pools
19904 @section Some Useful Memory Pools
19905 @findex Memory Pool
19906 @cindex storage, pool
19909 The @code{System.Pool_Global} package offers the Unbounded_No_Reclaim_Pool
19910 storage pool. Allocations use the standard system call @code{malloc} while
19911 deallocations use the standard system call @code{free}. No reclamation is
19912 performed when the pool goes out of scope. For performance reasons, the
19913 standard default Ada allocators/deallocators do not use any explicit storage
19914 pools but if they did, they could use this storage pool without any change in
19915 behavior. That is why this storage pool is used when the user
19916 manages to make the default implicit allocator explicit as in this example:
19917 @smallexample @c ada
19918 type T1 is access Something;
19919 -- no Storage pool is defined for T2
19920 type T2 is access Something_Else;
19921 for T2'Storage_Pool use T1'Storage_Pool;
19922 -- the above is equivalent to
19923 for T2'Storage_Pool use System.Pool_Global.Global_Pool_Object;
19927 The @code{System.Pool_Local} package offers the Unbounded_Reclaim_Pool storage
19928 pool. The allocation strategy is similar to @code{Pool_Local}'s
19929 except that the all
19930 storage allocated with this pool is reclaimed when the pool object goes out of
19931 scope. This pool provides a explicit mechanism similar to the implicit one
19932 provided by several Ada 83 compilers for allocations performed through a local
19933 access type and whose purpose was to reclaim memory when exiting the
19934 scope of a given local access. As an example, the following program does not
19935 leak memory even though it does not perform explicit deallocation:
19937 @smallexample @c ada
19938 with System.Pool_Local;
19939 procedure Pooloc1 is
19940 procedure Internal is
19941 type A is access Integer;
19942 X : System.Pool_Local.Unbounded_Reclaim_Pool;
19943 for A'Storage_Pool use X;
19946 for I in 1 .. 50 loop
19951 for I in 1 .. 100 loop
19958 The @code{System.Pool_Size} package implements the Stack_Bounded_Pool used when
19959 @code{Storage_Size} is specified for an access type.
19960 The whole storage for the pool is
19961 allocated at once, usually on the stack at the point where the access type is
19962 elaborated. It is automatically reclaimed when exiting the scope where the
19963 access type is defined. This package is not intended to be used directly by the
19964 user and it is implicitly used for each such declaration:
19966 @smallexample @c ada
19967 type T1 is access Something;
19968 for T1'Storage_Size use 10_000;
19971 @node The GNAT Debug Pool Facility
19972 @section The GNAT Debug Pool Facility
19974 @cindex storage, pool, memory corruption
19977 The use of unchecked deallocation and unchecked conversion can easily
19978 lead to incorrect memory references. The problems generated by such
19979 references are usually difficult to tackle because the symptoms can be
19980 very remote from the origin of the problem. In such cases, it is
19981 very helpful to detect the problem as early as possible. This is the
19982 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
19984 In order to use the GNAT specific debugging pool, the user must
19985 associate a debug pool object with each of the access types that may be
19986 related to suspected memory problems. See Ada Reference Manual 13.11.
19987 @smallexample @c ada
19988 type Ptr is access Some_Type;
19989 Pool : GNAT.Debug_Pools.Debug_Pool;
19990 for Ptr'Storage_Pool use Pool;
19994 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
19995 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
19996 allow the user to redefine allocation and deallocation strategies. They
19997 also provide a checkpoint for each dereference, through the use of
19998 the primitive operation @code{Dereference} which is implicitly called at
19999 each dereference of an access value.
20001 Once an access type has been associated with a debug pool, operations on
20002 values of the type may raise four distinct exceptions,
20003 which correspond to four potential kinds of memory corruption:
20006 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
20008 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
20010 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
20012 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage }
20016 For types associated with a Debug_Pool, dynamic allocation is performed using
20017 the standard GNAT allocation routine. References to all allocated chunks of
20018 memory are kept in an internal dictionary. Several deallocation strategies are
20019 provided, whereupon the user can choose to release the memory to the system,
20020 keep it allocated for further invalid access checks, or fill it with an easily
20021 recognizable pattern for debug sessions. The memory pattern is the old IBM
20022 hexadecimal convention: @code{16#DEADBEEF#}.
20024 See the documentation in the file g-debpoo.ads for more information on the
20025 various strategies.
20027 Upon each dereference, a check is made that the access value denotes a
20028 properly allocated memory location. Here is a complete example of use of
20029 @code{Debug_Pools}, that includes typical instances of memory corruption:
20030 @smallexample @c ada
20034 with Gnat.Io; use Gnat.Io;
20035 with Unchecked_Deallocation;
20036 with Unchecked_Conversion;
20037 with GNAT.Debug_Pools;
20038 with System.Storage_Elements;
20039 with Ada.Exceptions; use Ada.Exceptions;
20040 procedure Debug_Pool_Test is
20042 type T is access Integer;
20043 type U is access all T;
20045 P : GNAT.Debug_Pools.Debug_Pool;
20046 for T'Storage_Pool use P;
20048 procedure Free is new Unchecked_Deallocation (Integer, T);
20049 function UC is new Unchecked_Conversion (U, T);
20052 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
20062 Put_Line (Integer'Image(B.all));
20064 when E : others => Put_Line ("raised: " & Exception_Name (E));
20069 when E : others => Put_Line ("raised: " & Exception_Name (E));
20073 Put_Line (Integer'Image(B.all));
20075 when E : others => Put_Line ("raised: " & Exception_Name (E));
20080 when E : others => Put_Line ("raised: " & Exception_Name (E));
20083 end Debug_Pool_Test;
20087 The debug pool mechanism provides the following precise diagnostics on the
20088 execution of this erroneous program:
20091 Total allocated bytes : 0
20092 Total deallocated bytes : 0
20093 Current Water Mark: 0
20097 Total allocated bytes : 8
20098 Total deallocated bytes : 0
20099 Current Water Mark: 8
20102 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
20103 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
20104 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
20105 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
20107 Total allocated bytes : 8
20108 Total deallocated bytes : 4
20109 Current Water Mark: 4
20114 @node The gnatmem Tool
20115 @section The @command{gnatmem} Tool
20119 The @code{gnatmem} utility monitors dynamic allocation and
20120 deallocation activity in a program, and displays information about
20121 incorrect deallocations and possible sources of memory leaks.
20122 It is designed to work in association with a static runtime library
20123 only and in this context provides three types of information:
20126 General information concerning memory management, such as the total
20127 number of allocations and deallocations, the amount of allocated
20128 memory and the high water mark, i.e.@: the largest amount of allocated
20129 memory in the course of program execution.
20132 Backtraces for all incorrect deallocations, that is to say deallocations
20133 which do not correspond to a valid allocation.
20136 Information on each allocation that is potentially the origin of a memory
20141 * Running gnatmem::
20142 * Switches for gnatmem::
20143 * Example of gnatmem Usage::
20146 @node Running gnatmem
20147 @subsection Running @code{gnatmem}
20150 @code{gnatmem} makes use of the output created by the special version of
20151 allocation and deallocation routines that record call information. This
20152 allows to obtain accurate dynamic memory usage history at a minimal cost to
20153 the execution speed. Note however, that @code{gnatmem} is not supported on
20154 all platforms (currently, it is supported on AIX, HP-UX, GNU/Linux,
20155 Solaris and Windows NT/2000/XP (x86).
20158 The @code{gnatmem} command has the form
20161 $ gnatmem @ovar{switches} user_program
20165 The program must have been linked with the instrumented version of the
20166 allocation and deallocation routines. This is done by linking with the
20167 @file{libgmem.a} library. For correct symbolic backtrace information,
20168 the user program should be compiled with debugging options
20169 (see @ref{Switches for gcc}). For example to build @file{my_program}:
20172 $ gnatmake -g my_program -largs -lgmem
20176 As library @file{libgmem.a} contains an alternate body for package
20177 @code{System.Memory}, @file{s-memory.adb} should not be compiled and linked
20178 when an executable is linked with library @file{libgmem.a}. It is then not
20179 recommended to use @command{gnatmake} with switch @option{^-a^/ALL_FILES^}.
20182 When @file{my_program} is executed, the file @file{gmem.out} is produced.
20183 This file contains information about all allocations and deallocations
20184 performed by the program. It is produced by the instrumented allocations and
20185 deallocations routines and will be used by @code{gnatmem}.
20187 In order to produce symbolic backtrace information for allocations and
20188 deallocations performed by the GNAT run-time library, you need to use a
20189 version of that library that has been compiled with the @option{-g} switch
20190 (see @ref{Rebuilding the GNAT Run-Time Library}).
20192 Gnatmem must be supplied with the @file{gmem.out} file and the executable to
20193 examine. If the location of @file{gmem.out} file was not explicitly supplied by
20194 @option{-i} switch, gnatmem will assume that this file can be found in the
20195 current directory. For example, after you have executed @file{my_program},
20196 @file{gmem.out} can be analyzed by @code{gnatmem} using the command:
20199 $ gnatmem my_program
20203 This will produce the output with the following format:
20205 *************** debut cc
20207 $ gnatmem my_program
20211 Total number of allocations : 45
20212 Total number of deallocations : 6
20213 Final Water Mark (non freed mem) : 11.29 Kilobytes
20214 High Water Mark : 11.40 Kilobytes
20219 Allocation Root # 2
20220 -------------------
20221 Number of non freed allocations : 11
20222 Final Water Mark (non freed mem) : 1.16 Kilobytes
20223 High Water Mark : 1.27 Kilobytes
20225 my_program.adb:23 my_program.alloc
20231 The first block of output gives general information. In this case, the
20232 Ada construct ``@code{@b{new}}'' was executed 45 times, and only 6 calls to an
20233 Unchecked_Deallocation routine occurred.
20236 Subsequent paragraphs display information on all allocation roots.
20237 An allocation root is a specific point in the execution of the program
20238 that generates some dynamic allocation, such as a ``@code{@b{new}}''
20239 construct. This root is represented by an execution backtrace (or subprogram
20240 call stack). By default the backtrace depth for allocations roots is 1, so
20241 that a root corresponds exactly to a source location. The backtrace can
20242 be made deeper, to make the root more specific.
20244 @node Switches for gnatmem
20245 @subsection Switches for @code{gnatmem}
20248 @code{gnatmem} recognizes the following switches:
20253 @cindex @option{-q} (@code{gnatmem})
20254 Quiet. Gives the minimum output needed to identify the origin of the
20255 memory leaks. Omits statistical information.
20258 @cindex @var{N} (@code{gnatmem})
20259 N is an integer literal (usually between 1 and 10) which controls the
20260 depth of the backtraces defining allocation root. The default value for
20261 N is 1. The deeper the backtrace, the more precise the localization of
20262 the root. Note that the total number of roots can depend on this
20263 parameter. This parameter must be specified @emph{before} the name of the
20264 executable to be analyzed, to avoid ambiguity.
20267 @cindex @option{-b} (@code{gnatmem})
20268 This switch has the same effect as just depth parameter.
20270 @item -i @var{file}
20271 @cindex @option{-i} (@code{gnatmem})
20272 Do the @code{gnatmem} processing starting from @file{file}, rather than
20273 @file{gmem.out} in the current directory.
20276 @cindex @option{-m} (@code{gnatmem})
20277 This switch causes @code{gnatmem} to mask the allocation roots that have less
20278 than n leaks. The default value is 1. Specifying the value of 0 will allow to
20279 examine even the roots that didn't result in leaks.
20282 @cindex @option{-s} (@code{gnatmem})
20283 This switch causes @code{gnatmem} to sort the allocation roots according to the
20284 specified order of sort criteria, each identified by a single letter. The
20285 currently supported criteria are @code{n, h, w} standing respectively for
20286 number of unfreed allocations, high watermark, and final watermark
20287 corresponding to a specific root. The default order is @code{nwh}.
20291 @node Example of gnatmem Usage
20292 @subsection Example of @code{gnatmem} Usage
20295 The following example shows the use of @code{gnatmem}
20296 on a simple memory-leaking program.
20297 Suppose that we have the following Ada program:
20299 @smallexample @c ada
20302 with Unchecked_Deallocation;
20303 procedure Test_Gm is
20305 type T is array (1..1000) of Integer;
20306 type Ptr is access T;
20307 procedure Free is new Unchecked_Deallocation (T, Ptr);
20310 procedure My_Alloc is
20315 procedure My_DeAlloc is
20323 for I in 1 .. 5 loop
20324 for J in I .. 5 loop
20335 The program needs to be compiled with debugging option and linked with
20336 @code{gmem} library:
20339 $ gnatmake -g test_gm -largs -lgmem
20343 Then we execute the program as usual:
20350 Then @code{gnatmem} is invoked simply with
20356 which produces the following output (result may vary on different platforms):
20361 Total number of allocations : 18
20362 Total number of deallocations : 5
20363 Final Water Mark (non freed mem) : 53.00 Kilobytes
20364 High Water Mark : 56.90 Kilobytes
20366 Allocation Root # 1
20367 -------------------
20368 Number of non freed allocations : 11
20369 Final Water Mark (non freed mem) : 42.97 Kilobytes
20370 High Water Mark : 46.88 Kilobytes
20372 test_gm.adb:11 test_gm.my_alloc
20374 Allocation Root # 2
20375 -------------------
20376 Number of non freed allocations : 1
20377 Final Water Mark (non freed mem) : 10.02 Kilobytes
20378 High Water Mark : 10.02 Kilobytes
20380 s-secsta.adb:81 system.secondary_stack.ss_init
20382 Allocation Root # 3
20383 -------------------
20384 Number of non freed allocations : 1
20385 Final Water Mark (non freed mem) : 12 Bytes
20386 High Water Mark : 12 Bytes
20388 s-secsta.adb:181 system.secondary_stack.ss_init
20392 Note that the GNAT run time contains itself a certain number of
20393 allocations that have no corresponding deallocation,
20394 as shown here for root #2 and root
20395 #3. This is a normal behavior when the number of non-freed allocations
20396 is one, it allocates dynamic data structures that the run time needs for
20397 the complete lifetime of the program. Note also that there is only one
20398 allocation root in the user program with a single line back trace:
20399 test_gm.adb:11 test_gm.my_alloc, whereas a careful analysis of the
20400 program shows that 'My_Alloc' is called at 2 different points in the
20401 source (line 21 and line 24). If those two allocation roots need to be
20402 distinguished, the backtrace depth parameter can be used:
20405 $ gnatmem 3 test_gm
20409 which will give the following output:
20414 Total number of allocations : 18
20415 Total number of deallocations : 5
20416 Final Water Mark (non freed mem) : 53.00 Kilobytes
20417 High Water Mark : 56.90 Kilobytes
20419 Allocation Root # 1
20420 -------------------
20421 Number of non freed allocations : 10
20422 Final Water Mark (non freed mem) : 39.06 Kilobytes
20423 High Water Mark : 42.97 Kilobytes
20425 test_gm.adb:11 test_gm.my_alloc
20426 test_gm.adb:24 test_gm
20427 b_test_gm.c:52 main
20429 Allocation Root # 2
20430 -------------------
20431 Number of non freed allocations : 1
20432 Final Water Mark (non freed mem) : 10.02 Kilobytes
20433 High Water Mark : 10.02 Kilobytes
20435 s-secsta.adb:81 system.secondary_stack.ss_init
20436 s-secsta.adb:283 <system__secondary_stack___elabb>
20437 b_test_gm.c:33 adainit
20439 Allocation Root # 3
20440 -------------------
20441 Number of non freed allocations : 1
20442 Final Water Mark (non freed mem) : 3.91 Kilobytes
20443 High Water Mark : 3.91 Kilobytes
20445 test_gm.adb:11 test_gm.my_alloc
20446 test_gm.adb:21 test_gm
20447 b_test_gm.c:52 main
20449 Allocation Root # 4
20450 -------------------
20451 Number of non freed allocations : 1
20452 Final Water Mark (non freed mem) : 12 Bytes
20453 High Water Mark : 12 Bytes
20455 s-secsta.adb:181 system.secondary_stack.ss_init
20456 s-secsta.adb:283 <system__secondary_stack___elabb>
20457 b_test_gm.c:33 adainit
20461 The allocation root #1 of the first example has been split in 2 roots #1
20462 and #3 thanks to the more precise associated backtrace.
20466 @node Stack Related Facilities
20467 @chapter Stack Related Facilities
20470 This chapter describes some useful tools associated with stack
20471 checking and analysis. In
20472 particular, it deals with dynamic and static stack usage measurements.
20475 * Stack Overflow Checking::
20476 * Static Stack Usage Analysis::
20477 * Dynamic Stack Usage Analysis::
20480 @node Stack Overflow Checking
20481 @section Stack Overflow Checking
20482 @cindex Stack Overflow Checking
20483 @cindex -fstack-check
20486 For most operating systems, @command{gcc} does not perform stack overflow
20487 checking by default. This means that if the main environment task or
20488 some other task exceeds the available stack space, then unpredictable
20489 behavior will occur. Most native systems offer some level of protection by
20490 adding a guard page at the end of each task stack. This mechanism is usually
20491 not enough for dealing properly with stack overflow situations because
20492 a large local variable could ``jump'' above the guard page.
20493 Furthermore, when the
20494 guard page is hit, there may not be any space left on the stack for executing
20495 the exception propagation code. Enabling stack checking avoids
20498 To activate stack checking, compile all units with the gcc option
20499 @option{-fstack-check}. For example:
20502 gcc -c -fstack-check package1.adb
20506 Units compiled with this option will generate extra instructions to check
20507 that any use of the stack (for procedure calls or for declaring local
20508 variables in declare blocks) does not exceed the available stack space.
20509 If the space is exceeded, then a @code{Storage_Error} exception is raised.
20511 For declared tasks, the stack size is controlled by the size
20512 given in an applicable @code{Storage_Size} pragma or by the value specified
20513 at bind time with @option{-d} (@pxref{Switches for gnatbind}) or is set to
20514 the default size as defined in the GNAT runtime otherwise.
20516 For the environment task, the stack size depends on
20517 system defaults and is unknown to the compiler. Stack checking
20518 may still work correctly if a fixed
20519 size stack is allocated, but this cannot be guaranteed.
20521 To ensure that a clean exception is signalled for stack
20522 overflow, set the environment variable
20523 @env{GNAT_STACK_LIMIT} to indicate the maximum
20524 stack area that can be used, as in:
20525 @cindex GNAT_STACK_LIMIT
20528 SET GNAT_STACK_LIMIT 1600
20532 The limit is given in kilobytes, so the above declaration would
20533 set the stack limit of the environment task to 1.6 megabytes.
20534 Note that the only purpose of this usage is to limit the amount
20535 of stack used by the environment task. If it is necessary to
20536 increase the amount of stack for the environment task, then this
20537 is an operating systems issue, and must be addressed with the
20538 appropriate operating systems commands.
20541 To have a fixed size stack in the environment task, the stack must be put
20542 in the P0 address space and its size specified. Use these switches to
20546 gnatmake my_progs -largs "-Wl,--opt=STACK=4000,/p0image"
20550 The quotes are required to keep case. The number after @samp{STACK=} is the
20551 size of the environmental task stack in pagelets (512 bytes). In this example
20552 the stack size is about 2 megabytes.
20555 A consequence of the @option{/p0image} qualifier is also to makes RMS buffers
20556 be placed in P0 space. Refer to @cite{HP OpenVMS Linker Utility Manual} for
20557 more details about the @option{/p0image} qualifier and the @option{stack}
20561 @node Static Stack Usage Analysis
20562 @section Static Stack Usage Analysis
20563 @cindex Static Stack Usage Analysis
20564 @cindex -fstack-usage
20567 A unit compiled with @option{-fstack-usage} will generate an extra file
20569 the maximum amount of stack used, on a per-function basis.
20570 The file has the same
20571 basename as the target object file with a @file{.su} extension.
20572 Each line of this file is made up of three fields:
20576 The name of the function.
20580 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
20583 The second field corresponds to the size of the known part of the function
20586 The qualifier @code{static} means that the function frame size
20588 It usually means that all local variables have a static size.
20589 In this case, the second field is a reliable measure of the function stack
20592 The qualifier @code{dynamic} means that the function frame size is not static.
20593 It happens mainly when some local variables have a dynamic size. When this
20594 qualifier appears alone, the second field is not a reliable measure
20595 of the function stack analysis. When it is qualified with @code{bounded}, it
20596 means that the second field is a reliable maximum of the function stack
20599 @node Dynamic Stack Usage Analysis
20600 @section Dynamic Stack Usage Analysis
20603 It is possible to measure the maximum amount of stack used by a task, by
20604 adding a switch to @command{gnatbind}, as:
20607 $ gnatbind -u0 file
20611 With this option, at each task termination, its stack usage is output on
20613 It is not always convenient to output the stack usage when the program
20614 is still running. Hence, it is possible to delay this output until program
20615 termination. for a given number of tasks specified as the argument of the
20616 @option{-u} option. For instance:
20619 $ gnatbind -u100 file
20623 will buffer the stack usage information of the first 100 tasks to terminate and
20624 output this info at program termination. Results are displayed in four
20628 Index | Task Name | Stack Size | Stack Usage [Value +/- Variation]
20635 is a number associated with each task.
20638 is the name of the task analyzed.
20641 is the maximum size for the stack.
20644 is the measure done by the stack analyzer. In order to prevent overflow, the stack
20645 is not entirely analyzed, and it's not possible to know exactly how
20646 much has actually been used. The report thus contains the theoretical stack usage
20647 (Value) and the possible variation (Variation) around this value.
20652 The environment task stack, e.g., the stack that contains the main unit, is
20653 only processed when the environment variable GNAT_STACK_LIMIT is set.
20656 @c *********************************
20658 @c *********************************
20659 @node Verifying Properties Using gnatcheck
20660 @chapter Verifying Properties Using @command{gnatcheck}
20662 @cindex @command{gnatcheck}
20665 The @command{gnatcheck} tool is an ASIS-based utility that checks properties
20666 of Ada source files according to a given set of semantic rules.
20669 In order to check compliance with a given rule, @command{gnatcheck} has to
20670 semantically analyze the Ada sources.
20671 Therefore, checks can only be performed on
20672 legal Ada units. Moreover, when a unit depends semantically upon units located
20673 outside the current directory, the source search path has to be provided when
20674 calling @command{gnatcheck}, either through a specified project file or
20675 through @command{gnatcheck} switches as described below.
20677 A number of rules are predefined in @command{gnatcheck} and are described
20678 later in this chapter.
20679 You can also add new rules, by modifying the @command{gnatcheck} code and
20680 rebuilding the tool. In order to add a simple rule making some local checks,
20681 a small amount of straightforward ASIS-based programming is usually needed.
20683 Project support for @command{gnatcheck} is provided by the GNAT
20684 driver (see @ref{The GNAT Driver and Project Files}).
20686 Invoking @command{gnatcheck} on the command line has the form:
20689 $ gnatcheck @ovar{switches} @{@var{filename}@}
20690 @r{[}^-files^/FILES^=@{@var{arg_list_filename}@}@r{]}
20691 @r{[}-cargs @var{gcc_switches}@r{]} -rules @var{rule_options}
20698 @var{switches} specify the general tool options
20701 Each @var{filename} is the name (including the extension) of a source
20702 file to process. ``Wildcards'' are allowed, and
20703 the file name may contain path information.
20706 Each @var{arg_list_filename} is the name (including the extension) of a text
20707 file containing the names of the source files to process, separated by spaces
20711 @var{gcc_switches} is a list of switches for
20712 @command{gcc}. They will be passed on to all compiler invocations made by
20713 @command{gnatcheck} to generate the ASIS trees. Here you can provide
20714 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
20715 and use the @option{-gnatec} switch to set the configuration file.
20718 @var{rule_options} is a list of options for controlling a set of
20719 rules to be checked by @command{gnatcheck} (@pxref{gnatcheck Rule Options}).
20723 Either a @file{@var{filename}} or an @file{@var{arg_list_filename}} must be
20727 * Format of the Report File::
20728 * General gnatcheck Switches::
20729 * gnatcheck Rule Options::
20730 * Adding the Results of Compiler Checks to gnatcheck Output::
20731 * Project-Wide Checks::
20733 * Predefined Rules::
20736 @node Format of the Report File
20737 @section Format of the Report File
20738 @cindex Report file (for @code{gnatcheck})
20741 The @command{gnatcheck} tool outputs on @file{stdout} all messages concerning
20743 It also creates a text file that
20744 contains the complete report of the last gnatcheck run. By default this file
20745 is named named @file{^gnatcheck.out^GNATCHECK.OUT^} and it is located in the
20746 current directory; the @option{^-o^/OUTPUT^} option can be used to change the
20747 name and/or location of the report file. This report contains:
20749 @item date and time of @command{gnatcheck} run, the version of
20750 the tool that has generated this report and the full parameters
20751 of the @command{gnatcheck} invocation;
20752 @item list of enabled rules;
20753 @item total number of detected violations;
20754 @item list of source files where rule violations have been detected;
20755 @item list of source files where no violations have been detected.
20758 @node General gnatcheck Switches
20759 @section General @command{gnatcheck} Switches
20762 The following switches control the general @command{gnatcheck} behavior
20766 @cindex @option{^-a^/ALL^} (@command{gnatcheck})
20768 Process all units including those with read-only ALI files such as
20769 those from the GNAT Run-Time library.
20773 @cindex @option{-d} (@command{gnatcheck})
20778 @cindex @option{-dd} (@command{gnatcheck})
20780 Progress indicator mode (for use in GPS).
20783 @cindex @option{^-h^/HELP^} (@command{gnatcheck})
20785 List the predefined and user-defined rules. For more details see
20786 @ref{Predefined Rules}.
20788 @cindex @option{^-l^/LOCS^} (@command{gnatcheck})
20790 Use full source locations references in the report file. For a construct from
20791 a generic instantiation a full source location is a chain from the location
20792 of this construct in the generic unit to the place where this unit is
20795 @cindex @option{^-log^/LOG^} (@command{gnatcheck})
20797 Duplicate all the output sent to @file{stderr} into a log file. The log file
20798 is named @file{gnatcheck.log} and is located in the current directory.
20800 @cindex @option{^-m^/DIAGNOSTIC_LIMIT^} (@command{gnatcheck})
20801 @item ^-m@i{nnnn}^/DIAGNOSTIC_LIMIT=@i{nnnn}^
20802 Maximum number of diagnostics to be sent to @file{stdout}, where @i{nnnn} is in
20803 the range 0@dots{}1000;
20804 the default value is 500. Zero means that there is no limitation on
20805 the number of diagnostic messages to be output.
20807 @cindex @option{^-q^/QUIET^} (@command{gnatcheck})
20809 Quiet mode. All the diagnostics about rule violations are placed in the
20810 @command{gnatcheck} report file only, without duplication on @file{stdout}.
20812 @cindex @option{^-s^/SHORT^} (@command{gnatcheck})
20814 Short format of the report file (no version information, no list of applied
20815 rules, no list of checked sources is included)
20817 @cindex @option{^--include-file=@var{file}^/INCLUDE_FILE=@var{file}^} (@command{gnatcheck})
20818 @item ^--include-file^/INCLUDE_FILE^
20819 Append the content of the specified text file to the report file
20821 @cindex @option{^-t^/TIME^} (@command{gnatcheck})
20823 Print out execution time.
20825 @cindex @option{^-v^/VERBOSE^} (@command{gnatcheck})
20826 @item ^-v^/VERBOSE^
20827 Verbose mode; @command{gnatcheck} generates version information and then
20828 a trace of sources being processed.
20830 @cindex @option{^-o ^/OUTPUT^} (@command{gnatcheck})
20831 @item ^-o ^/OUTPUT=^@var{report_file}
20832 Set name of report file file to @var{report_file} .
20837 Note that if any of the options @option{^-s1^/COMPILER_STYLE^},
20838 @option{^-s2^/BY_RULES^} or
20839 @option{^-s3^/BY_FILES_BY_RULES^} is specified,
20840 then the @command{gnatcheck} report file will only contain sections
20841 explicitly denoted by these options.
20843 @node gnatcheck Rule Options
20844 @section @command{gnatcheck} Rule Options
20847 The following options control the processing performed by
20848 @command{gnatcheck}.
20851 @cindex @option{+ALL} (@command{gnatcheck})
20853 Turn all the rule checks ON.
20855 @cindex @option{-ALL} (@command{gnatcheck})
20857 Turn all the rule checks OFF.
20859 @cindex @option{+R} (@command{gnatcheck})
20860 @item +R@var{rule_id}@r{[}:@var{param}@r{]}
20861 Turn on the check for a specified rule with the specified parameter, if any.
20862 @var{rule_id} must be the identifier of one of the currently implemented rules
20863 (use @option{^-h^/HELP^} for the list of implemented rules). Rule identifiers
20864 are not case-sensitive. The @var{param} item must
20865 be a string representing a valid parameter(s) for the specified rule.
20866 If it contains any space characters then this string must be enclosed in
20869 @cindex @option{-R} (@command{gnatcheck})
20870 @item -R@var{rule_id}@r{[}:@var{param}@r{]}
20871 Turn off the check for a specified rule with the specified parameter, if any.
20873 @cindex @option{-from} (@command{gnatcheck})
20874 @item -from=@var{rule_option_filename}
20875 Read the rule options from the text file @var{rule_option_filename}, referred
20876 to as a ``coding standard file'' below.
20881 The default behavior is that all the rule checks are disabled.
20883 A coding standard file is a text file that contains a set of rule options
20885 @cindex Coding standard file (for @code{gnatcheck})
20886 The file may contain empty lines and Ada-style comments (comment
20887 lines and end-of-line comments). There can be several rule options on a
20888 single line (separated by a space).
20890 A coding standard file may reference other coding standard files by including
20891 more @option{-from=@var{rule_option_filename}}
20892 options, each such option being replaced with the content of the
20893 corresponding coding standard file during processing. In case a
20894 cycle is detected (that is, @file{@var{rule_file_1}} reads rule options
20895 from @file{@var{rule_file_2}}, and @file{@var{rule_file_2}} reads
20896 (directly or indirectly) rule options from @file{@var{rule_file_1}}),
20897 processing fails with an error message.
20900 @node Adding the Results of Compiler Checks to gnatcheck Output
20901 @section Adding the Results of Compiler Checks to @command{gnatcheck} Output
20904 The @command{gnatcheck} tool can include in the generated diagnostic messages
20906 the report file the results of the checks performed by the compiler. Though
20907 disabled by default, this effect may be obtained by using @option{+R} with
20908 the following rule identifiers and parameters:
20912 To record restrictions violations (which are performed by the compiler if the
20913 pragma @code{Restrictions} or @code{Restriction_Warnings} are given),
20914 use the @code{Restrictions} rule
20915 with the same parameters as pragma
20916 @code{Restrictions} or @code{Restriction_Warnings}.
20919 To record compiler style checks (@pxref{Style Checking}), use the
20920 @code{Style_Checks} rule.
20921 This rule takes a parameter in one of the following forms:
20925 which enables the standard style checks corresponding to the @option{-gnatyy}
20926 GNAT style check option, or
20929 a string with the same
20930 structure and semantics as the @code{string_LITERAL} parameter of the
20931 GNAT pragma @code{Style_Checks}
20932 (for further information about this pragma,
20933 @pxref{Pragma Style_Checks,,, gnat_rm, GNAT Reference Manual}).
20938 @code{+RStyle_Checks:O} rule option activates
20939 the compiler style check that corresponds to
20940 @code{-gnatyO} style check option.
20943 To record compiler warnings (@pxref{Warning Message Control}), use the
20944 @code{Warnings} rule with a parameter that is a valid
20945 @i{static_string_expression} argument of the GNAT pragma @code{Warnings}
20946 (for further information about this pragma,
20947 @pxref{Pragma Warnings,,,gnat_rm, GNAT Reference Manual}).
20948 Note that in case of gnatcheck
20949 's' parameter, that corresponds to the GNAT @option{-gnatws} option, disables
20950 all the specific warnings, but not suppresses the warning mode,
20951 and 'e' parameter, corresponding to @option{-gnatwe} that means
20952 "treat warnings as errors", does not have any effect.
20956 To disable a specific restriction check, use @code{-RStyle_Checks} gnatcheck
20957 option with the corresponding restriction name as a parameter. @code{-R} is
20958 not available for @code{Style_Checks} and @code{Warnings} options, to disable
20959 warnings and style checks, use the corresponding warning and style options.
20961 @node Project-Wide Checks
20962 @section Project-Wide Checks
20963 @cindex Project-wide checks (for @command{gnatcheck})
20966 In order to perform checks on all units of a given project, you can use
20967 the GNAT driver along with the @option{-P} option:
20969 gnat check -Pproj -rules -from=my_rules
20973 If the project @code{proj} depends upon other projects, you can perform
20974 checks on the project closure using the @option{-U} option:
20976 gnat check -Pproj -U -rules -from=my_rules
20980 Finally, if not all the units are relevant to a particular main
20981 program in the project closure, you can perform checks for the set
20982 of units needed to create a given main program (unit closure) using
20983 the @option{-U} option followed by the name of the main unit:
20985 gnat check -Pproj -U main -rules -from=my_rules
20989 @node Rule exemption
20990 @section Rule exemption
20991 @cindex Rule exemption (for @command{gnatcheck})
20994 One of the most useful applications of @command{gnatcheck} is to
20995 automate the enforcement of project-specific coding standards,
20996 for example in safety-critical systems where particular features
20997 must be restricted in order to simplify the certification effort.
20998 However, it may sometimes be appropriate to violate a coding standard rule,
20999 and in such cases the rationale for the violation should be provided
21000 in the source program itself so that the individuals
21001 reviewing or maintaining the program can immediately understand the intent.
21003 The @command{gnatcheck} tool supports this practice with the notion of
21004 a ``rule exemption'' covering a specific source code section. Normally
21005 rule violation messages are issued both on @file{stderr}
21006 and in a report file. In contrast, exempted violations are not listed on
21007 @file{stderr}; thus users invoking @command{gnatcheck} interactively
21008 (e.g. in its GPS interface) do not need to pay attention to known and
21009 justified violations. However, exempted violations along with their
21010 justification are documented in a special section of the report file that
21011 @command{gnatcheck} generates.
21014 * Using pragma Annotate to Control Rule Exemption::
21015 * gnatcheck Annotations Rules::
21018 @node Using pragma Annotate to Control Rule Exemption
21019 @subsection Using pragma @code{Annotate} to Control Rule Exemption
21020 @cindex Using pragma Annotate to control rule exemption
21023 Rule exemption is controlled by pragma @code{Annotate} when its first
21024 argument is ``gnatcheck''. The syntax of @command{gnatcheck}'s
21025 exemption control annotations is as follows:
21027 @smallexample @c ada
21029 pragma Annotate (gnatcheck, @i{exemption_control}, @i{Rule_Name}, [@i{justification}]);
21031 @i{exemption_control} ::= Exempt_On | Exempt_Off
21033 @i{Rule_Name} ::= string_literal
21035 @i{justification} ::= string_literal
21040 When a @command{gnatcheck} annotation has more then four arguments,
21041 @command{gnatcheck} issues a warning and ignores the additional arguments.
21042 If the additional arguments do not follow the syntax above,
21043 @command{gnatcheck} emits a warning and ignores the annotation.
21045 The @i{@code{Rule_Name}} argument should be the name of some existing
21046 @command{gnatcheck} rule.
21047 Otherwise a warning message is generated and the pragma is
21048 ignored. If @code{Rule_Name} denotes a rule that is not activated by the given
21049 @command{gnatcheck} call, the pragma is ignored and no warning is issued.
21051 A source code section where an exemption is active for a given rule is
21052 delimited by an @code{exempt_on} and @code{exempt_off} annotation pair:
21054 @smallexample @c ada
21055 pragma Annotate (gnatcheck, Exempt_On, Rule_Name, "justification");
21056 -- source code section
21057 pragma Annotate (gnatcheck, Exempt_Off, Rule_Name);
21061 @node gnatcheck Annotations Rules
21062 @subsection @command{gnatcheck} Annotations Rules
21063 @cindex @command{gnatcheck} annotations rules
21068 An ``Exempt_Off'' annotation can only appear after a corresponding
21069 ``Exempt_On'' annotation.
21072 Exempted source code sections are only based on the source location of the
21073 annotations. Any source construct between the two
21074 annotations is part of the exempted source code section.
21077 Exempted source code sections for different rules are independent. They can
21078 be nested or intersect with one another without limitation.
21079 Creating nested or intersecting source code sections for the same rule is
21083 Malformed exempted source code sections are reported by a warning, and
21084 the corresponding rule exemptions are ignored.
21087 When an exempted source code section does not contain at least one violation
21088 of the exempted rule, a warning is emitted on @file{stderr}.
21091 If an ``Exempt_On'' annotation pragma does not have a matching
21092 ``Exempt_Off'' annotation pragma in the same compilation unit, then the
21093 exemption for the given rule is ignored and a warning is issued.
21097 @node Predefined Rules
21098 @section Predefined Rules
21099 @cindex Predefined rules (for @command{gnatcheck})
21102 @c (Jan 2007) Since the global rules are still under development and are not
21103 @c documented, there is no point in explaining the difference between
21104 @c global and local rules
21106 A rule in @command{gnatcheck} is either local or global.
21107 A @emph{local rule} is a rule that applies to a well-defined section
21108 of a program and that can be checked by analyzing only this section.
21109 A @emph{global rule} requires analysis of some global properties of the
21110 whole program (mostly related to the program call graph).
21111 As of @value{NOW}, the implementation of global rules should be
21112 considered to be at a preliminary stage. You can use the
21113 @option{+GLOBAL} option to enable all the global rules, and the
21114 @option{-GLOBAL} rule option to disable all the global rules.
21116 All the global rules in the list below are
21117 so indicated by marking them ``GLOBAL''.
21118 This +GLOBAL and -GLOBAL options are not
21119 included in the list of gnatcheck options above, because at the moment they
21120 are considered as a temporary debug options.
21122 @command{gnatcheck} performs rule checks for generic
21123 instances only for global rules. This limitation may be relaxed in a later
21128 The following subsections document the rules implemented in
21129 @command{gnatcheck}.
21130 The subsection title is the same as the rule identifier, which may be
21131 used as a parameter of the @option{+R} or @option{-R} options.
21135 * Abstract_Type_Declarations::
21136 * Anonymous_Arrays::
21137 * Anonymous_Subtypes::
21139 * Boolean_Relational_Operators::
21141 * Ceiling_Violations::
21143 * Complex_Inlined_Subprograms::
21144 * Controlled_Type_Declarations::
21145 * Declarations_In_Blocks::
21146 * Deep_Inheritance_Hierarchies::
21147 * Deeply_Nested_Generics::
21148 * Deeply_Nested_Inlining::
21150 * Deeply_Nested_Local_Inlining::
21152 * Default_Parameters::
21153 * Direct_Calls_To_Primitives::
21154 * Discriminated_Records::
21155 * Enumeration_Ranges_In_CASE_Statements::
21156 * Exceptions_As_Control_Flow::
21157 * Exits_From_Conditional_Loops::
21158 * EXIT_Statements_With_No_Loop_Name::
21159 * Expanded_Loop_Exit_Names::
21160 * Explicit_Full_Discrete_Ranges::
21161 * Float_Equality_Checks::
21162 * Forbidden_Attributes::
21163 * Forbidden_Pragmas::
21164 * Function_Style_Procedures::
21165 * Generics_In_Subprograms::
21166 * GOTO_Statements::
21167 * Implicit_IN_Mode_Parameters::
21168 * Implicit_SMALL_For_Fixed_Point_Types::
21169 * Improperly_Located_Instantiations::
21170 * Improper_Returns::
21171 * Library_Level_Subprograms::
21174 * Improperly_Called_Protected_Entries::
21177 * Misnamed_Controlling_Parameters::
21178 * Misnamed_Identifiers::
21179 * Multiple_Entries_In_Protected_Definitions::
21181 * Non_Qualified_Aggregates::
21182 * Non_Short_Circuit_Operators::
21183 * Non_SPARK_Attributes::
21184 * Non_Tagged_Derived_Types::
21185 * Non_Visible_Exceptions::
21186 * Numeric_Literals::
21187 * OTHERS_In_Aggregates::
21188 * OTHERS_In_CASE_Statements::
21189 * OTHERS_In_Exception_Handlers::
21190 * Outer_Loop_Exits::
21191 * Overloaded_Operators::
21192 * Overly_Nested_Control_Structures::
21193 * Parameters_Out_Of_Order::
21194 * Positional_Actuals_For_Defaulted_Generic_Parameters::
21195 * Positional_Actuals_For_Defaulted_Parameters::
21196 * Positional_Components::
21197 * Positional_Generic_Parameters::
21198 * Positional_Parameters::
21199 * Predefined_Numeric_Types::
21200 * Raising_External_Exceptions::
21201 * Raising_Predefined_Exceptions::
21202 * Separate_Numeric_Error_Handlers::
21205 * Side_Effect_Functions::
21208 * Too_Many_Parents::
21209 * Unassigned_OUT_Parameters::
21210 * Uncommented_BEGIN_In_Package_Bodies::
21211 * Unconditional_Exits::
21212 * Unconstrained_Array_Returns::
21213 * Universal_Ranges::
21214 * Unnamed_Blocks_And_Loops::
21216 * Unused_Subprograms::
21218 * USE_PACKAGE_Clauses::
21219 * Visible_Components::
21220 * Volatile_Objects_Without_Address_Clauses::
21224 @node Abstract_Type_Declarations
21225 @subsection @code{Abstract_Type_Declarations}
21226 @cindex @code{Abstract_Type_Declarations} rule (for @command{gnatcheck})
21229 Flag all declarations of abstract types. For an abstract private
21230 type, both the private and full type declarations are flagged.
21232 This rule has no parameters.
21235 @node Anonymous_Arrays
21236 @subsection @code{Anonymous_Arrays}
21237 @cindex @code{Anonymous_Arrays} rule (for @command{gnatcheck})
21240 Flag all anonymous array type definitions (by Ada semantics these can only
21241 occur in object declarations).
21243 This rule has no parameters.
21245 @node Anonymous_Subtypes
21246 @subsection @code{Anonymous_Subtypes}
21247 @cindex @code{Anonymous_Subtypes} rule (for @command{gnatcheck})
21250 Flag all uses of anonymous subtypes (except cases when subtype indication
21251 is a part of a record component definition, and this subtype indication
21252 depends on a discriminant). A use of an anonymous subtype is
21253 any instance of a subtype indication with a constraint, other than one
21254 that occurs immediately within a subtype declaration. Any use of a range
21255 other than as a constraint used immediately within a subtype declaration
21256 is considered as an anonymous subtype.
21258 An effect of this rule is that @code{for} loops such as the following are
21259 flagged (since @code{1..N} is formally a ``range''):
21261 @smallexample @c ada
21262 for I in 1 .. N loop
21268 Declaring an explicit subtype solves the problem:
21270 @smallexample @c ada
21271 subtype S is Integer range 1..N;
21279 This rule has no parameters.
21282 @subsection @code{Blocks}
21283 @cindex @code{Blocks} rule (for @command{gnatcheck})
21286 Flag each block statement.
21288 This rule has no parameters.
21290 @node Boolean_Relational_Operators
21291 @subsection @code{Boolean_Relational_Operators}
21292 @cindex @code{Boolean_Relational_Operators} rule (for @command{gnatcheck})
21295 Flag each call to a predefined relational operator (``<'', ``>'', ``<='',
21296 ``>='', ``='' and ``/='') for the predefined Boolean type.
21297 (This rule is useful in enforcing the SPARK language restrictions.)
21299 Calls to predefined relational operators of any type derived from
21300 @code{Standard.Boolean} are not detected. Calls to user-defined functions
21301 with these designators, and uses of operators that are renamings
21302 of the predefined relational operators for @code{Standard.Boolean},
21303 are likewise not detected.
21305 This rule has no parameters.
21308 @node Ceiling_Violations
21309 @subsection @code{Ceiling5_Violations} (under construction, GLOBAL)
21310 @cindex @code{Ceiling_Violations} rule (for @command{gnatcheck})
21313 Flag invocations of a protected operation by a task whose priority exceeds
21314 the protected object's ceiling.
21316 As of @value{NOW}, this rule has the following limitations:
21321 We consider only pragmas Priority and Interrupt_Priority as means to define
21322 a task/protected operation priority. We do not consider the effect of using
21323 Ada.Dynamic_Priorities.Set_Priority procedure;
21326 We consider only base task priorities, and no priority inheritance. That is,
21327 we do not make a difference between calls issued during task activation and
21328 execution of the sequence of statements from task body;
21331 Any situation when the priority of protected operation caller is set by a
21332 dynamic expression (that is, the corresponding Priority or
21333 Interrupt_Priority pragma has a non-static expression as an argument) we
21334 treat as a priority inconsistency (and, therefore, detect this situation).
21338 At the moment the notion of the main subprogram is not implemented in
21339 gnatcheck, so any pragma Priority in a library level subprogram body (in case
21340 if this subprogram can be a main subprogram of a partition) changes the
21341 priority of an environment task. So if we have more then one such pragma in
21342 the set of processed sources, the pragma that is processed last, defines the
21343 priority of an environment task.
21345 This rule has no parameters.
21348 @node Controlled_Type_Declarations
21349 @subsection @code{Controlled_Type_Declarations}
21350 @cindex @code{Controlled_Type_Declarations} rule (for @command{gnatcheck})
21353 Flag all declarations of controlled types. A declaration of a private type
21354 is flagged if its full declaration declares a controlled type. A declaration
21355 of a derived type is flagged if its ancestor type is controlled. Subtype
21356 declarations are not checked. A declaration of a type that itself is not a
21357 descendant of a type declared in @code{Ada.Finalization} but has a controlled
21358 component is not checked.
21360 This rule has no parameters.
21363 @node Complex_Inlined_Subprograms
21364 @subsection @code{Complex_Inlined_Subprograms}
21365 @cindex @code{Complex_Inlined_Subprograms} rule (for @command{gnatcheck})
21368 Flags a subprogram (or generic subprogram) if
21369 pragma Inline is applied to the subprogram and at least one of the following
21374 it contains at least one complex declaration such as a subprogram body,
21375 package, task, protected declaration, or a generic instantiation
21376 (except instantiation of @code{Ada.Unchecked_Conversion});
21379 it contains at least one complex statement such as a loop, a case
21380 or a if statement, or a short circuit control form;
21383 the number of statements exceeds
21384 a value specified by the @option{N} rule parameter;
21388 This rule has the following (mandatory) parameter for the @option{+R} option:
21392 Positive integer specifying the maximum allowed total number of statements
21393 in the subprogram body.
21397 @node Declarations_In_Blocks
21398 @subsection @code{Declarations_In_Blocks}
21399 @cindex @code{Declarations_In_Blocks} rule (for @command{gnatcheck})
21402 Flag all block statements containing local declarations. A @code{declare}
21403 block with an empty @i{declarative_part} or with a @i{declarative part}
21404 containing only pragmas and/or @code{use} clauses is not flagged.
21406 This rule has no parameters.
21409 @node Deep_Inheritance_Hierarchies
21410 @subsection @code{Deep_Inheritance_Hierarchies}
21411 @cindex @code{Deep_Inheritance_Hierarchies} rule (for @command{gnatcheck})
21414 Flags a tagged derived type declaration or an interface type declaration if
21415 its depth (in its inheritance
21416 hierarchy) exceeds the value specified by the @option{N} rule parameter.
21418 The inheritance depth of a tagged type or interface type is defined as 0 for
21419 a type with no parent and no progenitor, and otherwise as 1 + max of the
21420 depths of the immediate parent and immediate progenitors.
21422 This rule does not flag private extension
21423 declarations. In the case of a private extension, the corresponding full
21424 declaration is checked.
21426 This rule has the following (mandatory) parameter for the @option{+R} option:
21430 Integer not less than -1 specifying the maximal allowed depth of any inheritance
21431 hierarchy. If the rule parameter is set to -1, the rule flags all the declarations
21432 of tagged and interface types.
21436 @node Deeply_Nested_Generics
21437 @subsection @code{Deeply_Nested_Generics}
21438 @cindex @code{Deeply_Nested_Generics} rule (for @command{gnatcheck})
21441 Flags a generic declaration nested in another generic declaration if
21442 the nesting level of the inner generic exceeds
21443 a value specified by the @option{N} rule parameter.
21444 The nesting level is the number of generic declaratons that enclose the given
21445 (generic) declaration. Formal packages are not flagged by this rule.
21447 This rule has the following (mandatory) parameters for the @option{+R} option:
21451 Positive integer specifying the maximal allowed nesting level
21452 for a generic declaration.
21455 @node Deeply_Nested_Inlining
21456 @subsection @code{Deeply_Nested_Inlining}
21457 @cindex @code{Deeply_Nested_Inlining} rule (for @command{gnatcheck})
21460 Flags a subprogram (or generic subprogram) if
21461 pragma Inline has been applied to the subprogram but the subprogram
21462 calls to another inlined subprogram that results in nested inlining
21463 with nesting depth exceeding the value specified by the
21464 @option{N} rule parameter.
21466 This rule requires the global analysis of all the compilation units that
21467 are @command{gnatcheck} arguments; such analysis may affect the tool's
21470 This rule has the following (mandatory) parameter for the @option{+R} option:
21474 Positive integer specifying the maximal allowed level of nested inlining.
21479 @node Deeply_Nested_Local_Inlining
21480 @subsection @code{Deeply_Nested_Local_Inlining}
21481 @cindex @code{Deeply_Nested_Local_Inlining} rule (for @command{gnatcheck})
21484 Flags a subprogram body if a pragma @code{Inline} is applied to the
21485 corresponding subprogram (or generic subprogram) and the body contains a call
21486 to another inlined subprogram that results in nested inlining with nesting
21487 depth more then a value specified by the @option{N} rule parameter.
21488 This rule is similar to @code{Deeply_Nested_Inlining} rule, but it
21489 assumes that calls to subprograms in
21490 with'ed units are not inlided, so all the analysis of the depth of inlining is
21491 limited by the compilation unit where the subprogram body is located and the
21492 units it depends semantically upon. Such analysis may be usefull for the case
21493 when neiter @option{-gnatn} nor @option{-gnatN} option is used when building
21496 This rule has the following (mandatory) parameters for the @option{+R} option:
21500 Positive integer specifying the maximal allowed level of nested inlining.
21505 @node Default_Parameters
21506 @subsection @code{Default_Parameters}
21507 @cindex @code{Default_Parameters} rule (for @command{gnatcheck})
21510 Flag all default expressions for subprogram parameters. Parameter
21511 declarations of formal and generic subprograms are also checked.
21513 This rule has no parameters.
21516 @node Direct_Calls_To_Primitives
21517 @subsection @code{Direct_Calls_To_Primitives}
21518 @cindex @code{Direct_Calls_To_Primitives} rule (for @command{gnatcheck})
21521 Flags any non-dispatching call to a dispatching primitive operation, except
21522 for the common idiom where a primitive subprogram for a tagged type
21523 directly calls the same primitive subprogram of the type's immediate ancestor.
21525 This rule has no parameters.
21528 @node Discriminated_Records
21529 @subsection @code{Discriminated_Records}
21530 @cindex @code{Discriminated_Records} rule (for @command{gnatcheck})
21533 Flag all declarations of record types with discriminants. Only the
21534 declarations of record and record extension types are checked. Incomplete,
21535 formal, private, derived and private extension type declarations are not
21536 checked. Task and protected type declarations also are not checked.
21538 This rule has no parameters.
21541 @node Enumeration_Ranges_In_CASE_Statements
21542 @subsection @code{Enumeration_Ranges_In_CASE_Statements}
21543 @cindex @code{Enumeration_Ranges_In_CASE_Statements} (for @command{gnatcheck})
21546 Flag each use of a range of enumeration literals as a choice in a
21547 @code{case} statement.
21548 All forms for specifying a range (explicit ranges
21549 such as @code{A .. B}, subtype marks and @code{'Range} attributes) are flagged.
21550 An enumeration range is
21551 flagged even if contains exactly one enumeration value or no values at all. A
21552 type derived from an enumeration type is considered as an enumeration type.
21554 This rule helps prevent maintenance problems arising from adding an
21555 enumeration value to a type and having it implicitly handled by an existing
21556 @code{case} statement with an enumeration range that includes the new literal.
21558 This rule has no parameters.
21561 @node Exceptions_As_Control_Flow
21562 @subsection @code{Exceptions_As_Control_Flow}
21563 @cindex @code{Exceptions_As_Control_Flow} (for @command{gnatcheck})
21566 Flag each place where an exception is explicitly raised and handled in the
21567 same subprogram body. A @code{raise} statement in an exception handler,
21568 package body, task body or entry body is not flagged.
21570 The rule has no parameters.
21572 @node Exits_From_Conditional_Loops
21573 @subsection @code{Exits_From_Conditional_Loops}
21574 @cindex @code{Exits_From_Conditional_Loops} (for @command{gnatcheck})
21577 Flag any exit statement if it transfers the control out of a @code{for} loop
21578 or a @code{while} loop. This includes cases when the @code{exit} statement
21579 applies to a @code{FOR} or @code{while} loop, and cases when it is enclosed
21580 in some @code{for} or @code{while} loop, but transfers the control from some
21581 outer (inconditional) @code{loop} statement.
21583 The rule has no parameters.
21586 @node EXIT_Statements_With_No_Loop_Name
21587 @subsection @code{EXIT_Statements_With_No_Loop_Name}
21588 @cindex @code{EXIT_Statements_With_No_Loop_Name} (for @command{gnatcheck})
21591 Flag each @code{exit} statement that does not specify the name of the loop
21594 The rule has no parameters.
21597 @node Expanded_Loop_Exit_Names
21598 @subsection @code{Expanded_Loop_Exit_Names}
21599 @cindex @code{Expanded_Loop_Exit_Names} rule (for @command{gnatcheck})
21602 Flag all expanded loop names in @code{exit} statements.
21604 This rule has no parameters.
21606 @node Explicit_Full_Discrete_Ranges
21607 @subsection @code{Explicit_Full_Discrete_Ranges}
21608 @cindex @code{Explicit_Full_Discrete_Ranges} rule (for @command{gnatcheck})
21611 Flag each discrete range that has the form @code{A'First .. A'Last}.
21613 This rule has no parameters.
21615 @node Float_Equality_Checks
21616 @subsection @code{Float_Equality_Checks}
21617 @cindex @code{Float_Equality_Checks} rule (for @command{gnatcheck})
21620 Flag all calls to the predefined equality operations for floating-point types.
21621 Both ``@code{=}'' and ``@code{/=}'' operations are checked.
21622 User-defined equality operations are not flagged, nor are ``@code{=}''
21623 and ``@code{/=}'' operations for fixed-point types.
21625 This rule has no parameters.
21628 @node Forbidden_Attributes
21629 @subsection @code{Forbidden_Attributes}
21630 @cindex @code{Forbidden_Attributes} rule (for @command{gnatcheck})
21633 Flag each use of the specified attributes. The attributes to be detected are
21634 named in the rule's parameters.
21636 This rule has the following parameters:
21639 @item For the @option{+R} option
21642 @item @emph{Attribute_Designator}
21643 Adds the specified attribute to the set of attributes to be detected and sets
21644 the detection checks for all the specified attributes ON.
21645 If @emph{Attribute_Designator}
21646 does not denote any attribute defined in the Ada standard
21648 @ref{Implementation Defined Attributes,,, gnat_rm, GNAT Reference
21649 Manual}, it is treated as the name of unknown attribute.
21652 All the GNAT-specific attributes are detected; this sets
21653 the detection checks for all the specified attributes ON.
21656 All attributes are detected; this sets the rule ON.
21659 @item For the @option{-R} option
21661 @item @emph{Attribute_Designator}
21662 Removes the specified attribute from the set of attributes to be
21663 detected without affecting detection checks for
21664 other attributes. If @emph{Attribute_Designator} does not correspond to any
21665 attribute defined in the Ada standard or in
21666 @ref{Implementation Defined Attributes,,, gnat_rm, GNAT Reference Manual},
21667 this option is treated as turning OFF detection of all unknown attributes.
21670 Turn OFF detection of all GNAT-specific attributes
21673 Clear the list of the attributes to be detected and
21679 Parameters are not case sensitive. If @emph{Attribute_Designator} does not
21680 have the syntax of an Ada identifier and therefore can not be considered as a
21681 (part of an) attribute designator, a diagnostic message is generated and the
21682 corresponding parameter is ignored. (If an attribute allows a static
21683 expression to be a part of the attribute designator, this expression is
21684 ignored by this rule.)
21686 When more then one parameter is given in the same rule option, the parameters
21687 must be separated by commas.
21689 If more then one option for this rule is specified for the gnatcheck call, a
21690 new option overrides the previous one(s).
21692 The @option{+R} option with no parameters turns the rule ON, with the set of
21693 attributes to be detected defined by the previous rule options.
21694 (By default this set is empty, so if the only option specified for the rule is
21695 @option{+RForbidden_Attributes} (with
21696 no parameter), then the rule is enabled, but it does not detect anything).
21697 The @option{-R} option with no parameter turns the rule OFF, but it does not
21698 affect the set of attributes to be detected.
21701 @node Forbidden_Pragmas
21702 @subsection @code{Forbidden_Pragmas}
21703 @cindex @code{Forbidden_Pragmas} rule (for @command{gnatcheck})
21706 Flag each use of the specified pragmas. The pragmas to be detected
21707 are named in the rule's parameters.
21709 This rule has the following parameters:
21712 @item For the @option{+R} option
21715 @item @emph{Pragma_Name}
21716 Adds the specified pragma to the set of pragmas to be
21717 checked and sets the checks for all the specified pragmas
21718 ON. @emph{Pragma_Name} is treated as a name of a pragma. If it
21719 does not correspond to any pragma name defined in the Ada
21720 standard or to the name of a GNAT-specific pragma defined
21721 in @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference
21722 Manual}, it is treated as the name of unknown pragma.
21725 All the GNAT-specific pragmas are detected; this sets
21726 the checks for all the specified pragmas ON.
21729 All pragmas are detected; this sets the rule ON.
21732 @item For the @option{-R} option
21734 @item @emph{Pragma_Name}
21735 Removes the specified pragma from the set of pragmas to be
21736 checked without affecting checks for
21737 other pragmas. @emph{Pragma_Name} is treated as a name
21738 of a pragma. If it does not correspond to any pragma
21739 defined in the Ada standard or to any name defined in
21740 @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
21741 this option is treated as turning OFF detection of all unknown pragmas.
21744 Turn OFF detection of all GNAT-specific pragmas
21747 Clear the list of the pragmas to be detected and
21753 Parameters are not case sensitive. If @emph{Pragma_Name} does not have
21754 the syntax of an Ada identifier and therefore can not be considered
21755 as a pragma name, a diagnostic message is generated and the corresponding
21756 parameter is ignored.
21758 When more then one parameter is given in the same rule option, the parameters
21759 must be separated by a comma.
21761 If more then one option for this rule is specified for the @command{gnatcheck}
21762 call, a new option overrides the previous one(s).
21764 The @option{+R} option with no parameters turns the rule ON with the set of
21765 pragmas to be detected defined by the previous rule options.
21766 (By default this set is empty, so if the only option specified for the rule is
21767 @option{+RForbidden_Pragmas} (with
21768 no parameter), then the rule is enabled, but it does not detect anything).
21769 The @option{-R} option with no parameter turns the rule OFF, but it does not
21770 affect the set of pragmas to be detected.
21775 @node Function_Style_Procedures
21776 @subsection @code{Function_Style_Procedures}
21777 @cindex @code{Function_Style_Procedures} rule (for @command{gnatcheck})
21780 Flag each procedure that can be rewritten as a function. A procedure can be
21781 converted into a function if it has exactly one parameter of mode @code{out}
21782 and no parameters of mode @code{in out}. Procedure declarations,
21783 formal procedure declarations, and generic procedure declarations are always
21785 bodies and body stubs are flagged only if they do not have corresponding
21786 separate declarations. Procedure renamings and procedure instantiations are
21789 If a procedure can be rewritten as a function, but its @code{out} parameter is
21790 of a limited type, it is not flagged.
21792 Protected procedures are not flagged. Null procedures also are not flagged.
21794 This rule has no parameters.
21797 @node Generics_In_Subprograms
21798 @subsection @code{Generics_In_Subprograms}
21799 @cindex @code{Generics_In_Subprograms} rule (for @command{gnatcheck})
21802 Flag each declaration of a generic unit in a subprogram. Generic
21803 declarations in the bodies of generic subprograms are also flagged.
21804 A generic unit nested in another generic unit is not flagged.
21805 If a generic unit is
21806 declared in a local package that is declared in a subprogram body, the
21807 generic unit is flagged.
21809 This rule has no parameters.
21812 @node GOTO_Statements
21813 @subsection @code{GOTO_Statements}
21814 @cindex @code{GOTO_Statements} rule (for @command{gnatcheck})
21817 Flag each occurrence of a @code{goto} statement.
21819 This rule has no parameters.
21822 @node Implicit_IN_Mode_Parameters
21823 @subsection @code{Implicit_IN_Mode_Parameters}
21824 @cindex @code{Implicit_IN_Mode_Parameters} rule (for @command{gnatcheck})
21827 Flag each occurrence of a formal parameter with an implicit @code{in} mode.
21828 Note that @code{access} parameters, although they technically behave
21829 like @code{in} parameters, are not flagged.
21831 This rule has no parameters.
21834 @node Implicit_SMALL_For_Fixed_Point_Types
21835 @subsection @code{Implicit_SMALL_For_Fixed_Point_Types}
21836 @cindex @code{Implicit_SMALL_For_Fixed_Point_Types} rule (for @command{gnatcheck})
21839 Flag each fixed point type declaration that lacks an explicit
21840 representation clause to define its @code{'Small} value.
21841 Since @code{'Small} can be defined only for ordinary fixed point types,
21842 decimal fixed point type declarations are not checked.
21844 This rule has no parameters.
21847 @node Improperly_Located_Instantiations
21848 @subsection @code{Improperly_Located_Instantiations}
21849 @cindex @code{Improperly_Located_Instantiations} rule (for @command{gnatcheck})
21852 Flag all generic instantiations in library-level package specs
21853 (including library generic packages) and in all subprogram bodies.
21855 Instantiations in task and entry bodies are not flagged. Instantiations in the
21856 bodies of protected subprograms are flagged.
21858 This rule has no parameters.
21862 @node Improper_Returns
21863 @subsection @code{Improper_Returns}
21864 @cindex @code{Improper_Returns} rule (for @command{gnatcheck})
21867 Flag each explicit @code{return} statement in procedures, and
21868 multiple @code{return} statements in functions.
21869 Diagnostic messages are generated for all @code{return} statements
21870 in a procedure (thus each procedure must be written so that it
21871 returns implicitly at the end of its statement part),
21872 and for all @code{return} statements in a function after the first one.
21873 This rule supports the stylistic convention that each subprogram
21874 should have no more than one point of normal return.
21876 This rule has no parameters.
21879 @node Library_Level_Subprograms
21880 @subsection @code{Library_Level_Subprograms}
21881 @cindex @code{Library_Level_Subprograms} rule (for @command{gnatcheck})
21884 Flag all library-level subprograms (including generic subprogram instantiations).
21886 This rule has no parameters.
21889 @node Local_Packages
21890 @subsection @code{Local_Packages}
21891 @cindex @code{Local_Packages} rule (for @command{gnatcheck})
21894 Flag all local packages declared in package and generic package
21896 Local packages in bodies are not flagged.
21898 This rule has no parameters.
21901 @node Improperly_Called_Protected_Entries
21902 @subsection @code{Improperly_Called_Protected_Entries} (under construction, GLOBAL)
21903 @cindex @code{Improperly_Called_Protected_Entries} rule (for @command{gnatcheck})
21906 Flag each protected entry that can be called from more than one task.
21908 This rule has no parameters.
21912 @subsection @code{Metrics}
21913 @cindex @code{Metrics} rule (for @command{gnatcheck})
21916 There is a set of checks based on computing a metric value and comparing the
21917 result with the specified upper (or lower, depending on a specific metric)
21918 value specified for a given metric. A construct is flagged if a given metric
21919 is applicable (can be computed) for it and the computed value is greater
21920 then (lover then) the specified upper (lower) bound.
21922 The name of any metric-based rule consists of the prefix @code{Metrics_}
21923 followed by the name of the corresponding metric (see the table below).
21924 For @option{+R} option, each metric-based rule has a numeric parameter
21925 specifying the bound (integer or real, depending on a metric), @option{-R}
21926 option for metric rules does not have a parameter.
21928 The following table shows the metric names for that the corresponding
21929 metrics-based checks are supported by gnatcheck, including the
21930 constraint that must be satisfied by the bound that is specified for the check
21931 and what bound - upper (U) or lower (L) - should be specified.
21933 @multitable {@code{Cyclomatic_Complexity}}{Cyclomatic complexity}{Positive integer}
21935 @headitem Check Name @tab Description @tab Bounds Value
21938 @item @b{Check Name} @tab @b{Description} @tab @b{Bounds Value}
21940 @c Above conditional code is workaround to bug in texi2html (Feb 2008)
21941 @item @code{Essential_Complexity} @tab Essential complexity @tab Positive integer (U)
21942 @item @code{Cyclomatic_Complexity} @tab Cyclomatic complexity @tab Positive integer (U)
21943 @item @code{LSLOC} @tab Logical Source Lines of Code @tab Positive integer (U)
21947 The meaning and the computed values for all these metrics are exactly
21948 the same as for the corresponding metrics in @command{gnatmetric}.
21950 @emph{Example:} the rule
21952 +RMetrics_Cyclomatic_Complexity : 7
21955 means that all bodies with cyclomatic complexity exceeding 7 will be flagged.
21957 To turn OFF the check for cyclomatic complexity metric, use the following option:
21959 -RMetrics_Cyclomatic_Complexity
21963 @node Misnamed_Controlling_Parameters
21964 @subsection @code{Misnamed_Controlling_Parameters}
21965 @cindex @code{Misnamed_Controlling_Parameters} rule (for @command{gnatcheck})
21968 Flags a declaration of a dispatching operation, if the first parameter is
21969 not a controlling one and its name is not @code{This} (the check for
21970 parameter name is not case-sensitive). Declarations of dispatching functions
21971 with controlling result and no controlling parameter are never flagged.
21973 A subprogram body declaration, subprogram renaming declaration or subprogram
21974 body stub is flagged only if it is not a completion of a prior subprogram
21977 This rule has no parameters.
21981 @node Misnamed_Identifiers
21982 @subsection @code{Misnamed_Identifiers}
21983 @cindex @code{Misnamed_Identifiers} rule (for @command{gnatcheck})
21986 Flag the declaration of each identifier that does not have a suffix
21987 corresponding to the kind of entity being declared.
21988 The following declarations are checked:
21995 subtype declarations
21998 constant declarations (but not number declarations)
22001 package renaming declarations (but not generic package renaming
22006 This rule may have parameters. When used without parameters, the rule enforces
22007 the following checks:
22011 type-defining names end with @code{_T}, unless the type is an access type,
22012 in which case the suffix must be @code{_A}
22014 constant names end with @code{_C}
22016 names defining package renamings end with @code{_R}
22020 Defining identifiers from incomplete type declarations are never flagged.
22022 For a private type declaration (including private extensions), the defining
22023 identifier from the private type declaration is checked against the type
22024 suffix (even if the corresponding full declaration is an access type
22025 declaration), and the defining identifier from the corresponding full type
22026 declaration is not checked.
22029 For a deferred constant, the defining name in the corresponding full constant
22030 declaration is not checked.
22032 Defining names of formal types are not checked.
22034 The rule may have the following parameters:
22038 For the @option{+R} option:
22041 Sets the default listed above for all the names to be checked.
22043 @item Type_Suffix=@emph{string}
22044 Specifies the suffix for a type name.
22046 @item Access_Suffix=@emph{string}
22047 Specifies the suffix for an access type name. If
22048 this parameter is set, it overrides for access
22049 types the suffix set by the @code{Type_Suffix} parameter.
22050 For access types, @emph{string} may have the following format:
22051 @emph{suffix1(suffix2)}. That means that an access type name
22052 should have the @emph{suffix1} suffix except for the case when
22053 the designated type is also an access type, in this case the
22054 type name should have the @emph{suffix1 & suffix2} suffix.
22056 @item Class_Access_Suffix=@emph{string}
22057 Specifies the suffix for the name of an access type that points to some class-wide
22058 type. If this parameter is set, it overrides for such access
22059 types the suffix set by the @code{Type_Suffix} or @code{Access_Suffix}
22062 @item Class_Subtype_Suffix=@emph{string}
22063 Specifies the suffix for the name of a subtype that denotes a class-wide type.
22065 @item Constant_Suffix=@emph{string}
22066 Specifies the suffix for a constant name.
22068 @item Renaming_Suffix=@emph{string}
22069 Specifies the suffix for a package renaming name.
22073 For the @option{-R} option:
22076 Remove all the suffixes specified for the
22077 identifier suffix checks, whether by default or
22078 as specified by other rule parameters. All the
22079 checks for this rule are disabled as a result.
22082 Removes the suffix specified for types. This
22083 disables checks for types but does not disable
22084 any other checks for this rule (including the
22085 check for access type names if @code{Access_Suffix} is
22088 @item Access_Suffix
22089 Removes the suffix specified for access types.
22090 This disables checks for access type names but
22091 does not disable any other checks for this rule.
22092 If @code{Type_Suffix} is set, access type names are
22093 checked as ordinary type names.
22095 @item Class_Access_Suffix
22096 Removes the suffix specified for access types pointing to class-wide
22097 type. This disables specific checks for names of access types pointing to
22098 class-wide types but does not disable any other checks for this rule.
22099 If @code{Type_Suffix} is set, access type names are
22100 checked as ordinary type names. If @code{Access_Suffix} is set, these
22101 access types are checked as any other access type name.
22103 @item Class_Subtype_Suffix=@emph{string}
22104 Removes the suffix specified for subtype names.
22105 This disables checks for subtype names but
22106 does not disable any other checks for this rule.
22108 @item Constant_Suffix
22109 Removes the suffix specified for constants. This
22110 disables checks for constant names but does not
22111 disable any other checks for this rule.
22113 @item Renaming_Suffix
22114 Removes the suffix specified for package
22115 renamings. This disables checks for package
22116 renamings but does not disable any other checks
22122 If more than one parameter is used, parameters must be separated by commas.
22124 If more than one option is specified for the @command{gnatcheck} invocation,
22125 a new option overrides the previous one(s).
22127 The @option{+RMisnamed_Identifiers} option (with no parameter) enables
22129 name suffixes specified by previous options used for this rule.
22131 The @option{-RMisnamed_Identifiers} option (with no parameter) disables
22132 all the checks but keeps
22133 all the suffixes specified by previous options used for this rule.
22135 The @emph{string} value must be a valid suffix for an Ada identifier (after
22136 trimming all the leading and trailing space characters, if any).
22137 Parameters are not case sensitive, except the @emph{string} part.
22139 If any error is detected in a rule parameter, the parameter is ignored.
22140 In such a case the options that are set for the rule are not
22145 @node Multiple_Entries_In_Protected_Definitions
22146 @subsection @code{Multiple_Entries_In_Protected_Definitions}
22147 @cindex @code{Multiple_Entries_In_Protected_Definitions} rule (for @command{gnatcheck})
22150 Flag each protected definition (i.e., each protected object/type declaration)
22151 that defines more than one entry.
22152 Diagnostic messages are generated for all the entry declarations
22153 except the first one. An entry family is counted as one entry. Entries from
22154 the private part of the protected definition are also checked.
22156 This rule has no parameters.
22159 @subsection @code{Name_Clashes}
22160 @cindex @code{Name_Clashes} rule (for @command{gnatcheck})
22163 Check that certain names are not used as defining identifiers. To activate
22164 this rule, you need to supply a reference to the dictionary file(s) as a rule
22165 parameter(s) (more then one dictionary file can be specified). If no
22166 dictionary file is set, this rule will not cause anything to be flagged.
22167 Only defining occurrences, not references, are checked.
22168 The check is not case-sensitive.
22170 This rule is enabled by default, but without setting any corresponding
22171 dictionary file(s); thus the default effect is to do no checks.
22173 A dictionary file is a plain text file. The maximum line length for this file
22174 is 1024 characters. If the line is longer then this limit, extra characters
22177 Each line can be either an empty line, a comment line, or a line containing
22178 a list of identifiers separated by space or HT characters.
22179 A comment is an Ada-style comment (from @code{--} to end-of-line).
22180 Identifiers must follow the Ada syntax for identifiers.
22181 A line containing one or more identifiers may end with a comment.
22183 @node Non_Qualified_Aggregates
22184 @subsection @code{Non_Qualified_Aggregates}
22185 @cindex @code{Non_Qualified_Aggregates} rule (for @command{gnatcheck})
22188 Flag each non-qualified aggregate.
22189 A non-qualified aggregate is an
22190 aggregate that is not the expression of a qualified expression. A
22191 string literal is not considered an aggregate, but an array
22192 aggregate of a string type is considered as a normal aggregate.
22193 Aggregates of anonymous array types are not flagged.
22195 This rule has no parameters.
22198 @node Non_Short_Circuit_Operators
22199 @subsection @code{Non_Short_Circuit_Operators}
22200 @cindex @code{Non_Short_Circuit_Operators} rule (for @command{gnatcheck})
22203 Flag all calls to predefined @code{and} and @code{or} operators for
22204 any boolean type. Calls to
22205 user-defined @code{and} and @code{or} and to operators defined by renaming
22206 declarations are not flagged. Calls to predefined @code{and} and @code{or}
22207 operators for modular types or boolean array types are not flagged.
22209 This rule has no parameters.
22213 @node Non_SPARK_Attributes
22214 @subsection @code{Non_SPARK_Attributes}
22215 @cindex @code{Non_SPARK_Attributes} rule (for @command{gnatcheck})
22218 The SPARK language defines the following subset of Ada 95 attribute
22219 designators as those that can be used in SPARK programs. The use of
22220 any other attribute is flagged.
22223 @item @code{'Adjacent}
22226 @item @code{'Ceiling}
22227 @item @code{'Component_Size}
22228 @item @code{'Compose}
22229 @item @code{'Copy_Sign}
22230 @item @code{'Delta}
22231 @item @code{'Denorm}
22232 @item @code{'Digits}
22233 @item @code{'Exponent}
22234 @item @code{'First}
22235 @item @code{'Floor}
22237 @item @code{'Fraction}
22239 @item @code{'Leading_Part}
22240 @item @code{'Length}
22241 @item @code{'Machine}
22242 @item @code{'Machine_Emax}
22243 @item @code{'Machine_Emin}
22244 @item @code{'Machine_Mantissa}
22245 @item @code{'Machine_Overflows}
22246 @item @code{'Machine_Radix}
22247 @item @code{'Machine_Rounds}
22250 @item @code{'Model}
22251 @item @code{'Model_Emin}
22252 @item @code{'Model_Epsilon}
22253 @item @code{'Model_Mantissa}
22254 @item @code{'Model_Small}
22255 @item @code{'Modulus}
22258 @item @code{'Range}
22259 @item @code{'Remainder}
22260 @item @code{'Rounding}
22261 @item @code{'Safe_First}
22262 @item @code{'Safe_Last}
22263 @item @code{'Scaling}
22264 @item @code{'Signed_Zeros}
22266 @item @code{'Small}
22268 @item @code{'Truncation}
22269 @item @code{'Unbiased_Rounding}
22271 @item @code{'Valid}
22275 This rule has no parameters.
22278 @node Non_Tagged_Derived_Types
22279 @subsection @code{Non_Tagged_Derived_Types}
22280 @cindex @code{Non_Tagged_Derived_Types} rule (for @command{gnatcheck})
22283 Flag all derived type declarations that do not have a record extension part.
22285 This rule has no parameters.
22289 @node Non_Visible_Exceptions
22290 @subsection @code{Non_Visible_Exceptions}
22291 @cindex @code{Non_Visible_Exceptions} rule (for @command{gnatcheck})
22294 Flag constructs leading to the possibility of propagating an exception
22295 out of the scope in which the exception is declared.
22296 Two cases are detected:
22300 An exception declaration in a subprogram body, task body or block
22301 statement is flagged if the body or statement does not contain a handler for
22302 that exception or a handler with an @code{others} choice.
22305 A @code{raise} statement in an exception handler of a subprogram body,
22306 task body or block statement is flagged if it (re)raises a locally
22307 declared exception. This may occur under the following circumstances:
22310 it explicitly raises a locally declared exception, or
22312 it does not specify an exception name (i.e., it is simply @code{raise;})
22313 and the enclosing handler contains a locally declared exception in its
22319 Renamings of local exceptions are not flagged.
22321 This rule has no parameters.
22324 @node Numeric_Literals
22325 @subsection @code{Numeric_Literals}
22326 @cindex @code{Numeric_Literals} rule (for @command{gnatcheck})
22329 Flag each use of a numeric literal in an index expression, and in any
22330 circumstance except for the following:
22334 a literal occurring in the initialization expression for a constant
22335 declaration or a named number declaration, or
22338 an integer literal that is less than or equal to a value
22339 specified by the @option{N} rule parameter.
22343 This rule may have the following parameters for the @option{+R} option:
22347 @emph{N} is an integer literal used as the maximal value that is not flagged
22348 (i.e., integer literals not exceeding this value are allowed)
22351 All integer literals are flagged
22355 If no parameters are set, the maximum unflagged value is 1.
22357 The last specified check limit (or the fact that there is no limit at
22358 all) is used when multiple @option{+R} options appear.
22360 The @option{-R} option for this rule has no parameters.
22361 It disables the rule but retains the last specified maximum unflagged value.
22362 If the @option{+R} option subsequently appears, this value is used as the
22363 threshold for the check.
22366 @node OTHERS_In_Aggregates
22367 @subsection @code{OTHERS_In_Aggregates}
22368 @cindex @code{OTHERS_In_Aggregates} rule (for @command{gnatcheck})
22371 Flag each use of an @code{others} choice in extension aggregates.
22372 In record and array aggregates, an @code{others} choice is flagged unless
22373 it is used to refer to all components, or to all but one component.
22375 If, in case of a named array aggregate, there are two associations, one
22376 with an @code{others} choice and another with a discrete range, the
22377 @code{others} choice is flagged even if the discrete range specifies
22378 exactly one component; for example, @code{(1..1 => 0, others => 1)}.
22380 This rule has no parameters.
22382 @node OTHERS_In_CASE_Statements
22383 @subsection @code{OTHERS_In_CASE_Statements}
22384 @cindex @code{OTHERS_In_CASE_Statements} rule (for @command{gnatcheck})
22387 Flag any use of an @code{others} choice in a @code{case} statement.
22389 This rule has no parameters.
22391 @node OTHERS_In_Exception_Handlers
22392 @subsection @code{OTHERS_In_Exception_Handlers}
22393 @cindex @code{OTHERS_In_Exception_Handlers} rule (for @command{gnatcheck})
22396 Flag any use of an @code{others} choice in an exception handler.
22398 This rule has no parameters.
22401 @node Outer_Loop_Exits
22402 @subsection @code{Outer_Loop_Exits}
22403 @cindex @code{Outer_Loop_Exits} rule (for @command{gnatcheck})
22406 Flag each @code{exit} statement containing a loop name that is not the name
22407 of the immediately enclosing @code{loop} statement.
22409 This rule has no parameters.
22412 @node Overloaded_Operators
22413 @subsection @code{Overloaded_Operators}
22414 @cindex @code{Overloaded_Operators} rule (for @command{gnatcheck})
22417 Flag each function declaration that overloads an operator symbol.
22418 A function body is checked only if the body does not have a
22419 separate spec. Formal functions are also checked. For a
22420 renaming declaration, only renaming-as-declaration is checked
22422 This rule has no parameters.
22425 @node Overly_Nested_Control_Structures
22426 @subsection @code{Overly_Nested_Control_Structures}
22427 @cindex @code{Overly_Nested_Control_Structures} rule (for @command{gnatcheck})
22430 Flag each control structure whose nesting level exceeds the value provided
22431 in the rule parameter.
22433 The control structures checked are the following:
22436 @item @code{if} statement
22437 @item @code{case} statement
22438 @item @code{loop} statement
22439 @item Selective accept statement
22440 @item Timed entry call statement
22441 @item Conditional entry call
22442 @item Asynchronous select statement
22446 The rule has the following parameter for the @option{+R} option:
22450 Positive integer specifying the maximal control structure nesting
22451 level that is not flagged
22455 If the parameter for the @option{+R} option is not specified or
22456 if it is not a positive integer, @option{+R} option is ignored.
22458 If more then one option is specified for the gnatcheck call, the later option and
22459 new parameter override the previous one(s).
22462 @node Parameters_Out_Of_Order
22463 @subsection @code{Parameters_Out_Of_Order}
22464 @cindex @code{Parameters_Out_Of_Order} rule (for @command{gnatcheck})
22467 Flag each subprogram and entry declaration whose formal parameters are not
22468 ordered according to the following scheme:
22472 @item @code{in} and @code{access} parameters first,
22473 then @code{in out} parameters,
22474 and then @code{out} parameters;
22476 @item for @code{in} mode, parameters with default initialization expressions
22481 Only the first violation of the described order is flagged.
22483 The following constructs are checked:
22486 @item subprogram declarations (including null procedures);
22487 @item generic subprogram declarations;
22488 @item formal subprogram declarations;
22489 @item entry declarations;
22490 @item subprogram bodies and subprogram body stubs that do not
22491 have separate specifications
22495 Subprogram renamings are not checked.
22497 This rule has no parameters.
22500 @node Positional_Actuals_For_Defaulted_Generic_Parameters
22501 @subsection @code{Positional_Actuals_For_Defaulted_Generic_Parameters}
22502 @cindex @code{Positional_Actuals_For_Defaulted_Generic_Parameters} rule (for @command{gnatcheck})
22505 Flag each generic actual parameter corresponding to a generic formal
22506 parameter with a default initialization, if positional notation is used.
22508 This rule has no parameters.
22510 @node Positional_Actuals_For_Defaulted_Parameters
22511 @subsection @code{Positional_Actuals_For_Defaulted_Parameters}
22512 @cindex @code{Positional_Actuals_For_Defaulted_Parameters} rule (for @command{gnatcheck})
22515 Flag each actual parameter to a subprogram or entry call where the
22516 corresponding formal parameter has a default expression, if positional
22519 This rule has no parameters.
22521 @node Positional_Components
22522 @subsection @code{Positional_Components}
22523 @cindex @code{Positional_Components} rule (for @command{gnatcheck})
22526 Flag each array, record and extension aggregate that includes positional
22529 This rule has no parameters.
22532 @node Positional_Generic_Parameters
22533 @subsection @code{Positional_Generic_Parameters}
22534 @cindex @code{Positional_Generic_Parameters} rule (for @command{gnatcheck})
22537 Flag each positional actual generic parameter except for the case when
22538 the generic unit being iinstantiated has exactly one generic formal
22541 This rule has no parameters.
22544 @node Positional_Parameters
22545 @subsection @code{Positional_Parameters}
22546 @cindex @code{Positional_Parameters} rule (for @command{gnatcheck})
22549 Flag each positional parameter notation in a subprogram or entry call,
22550 except for the following:
22554 Parameters of calls to of prefix or infix operators are not flagged
22556 If the called subprogram or entry has only one formal parameter,
22557 the parameter of the call is not flagged;
22559 If a subprogram call uses the @emph{Object.Operation} notation, then
22562 the first parameter (that is, @emph{Object}) is not flagged;
22564 if the called subprogram has only two parameters, the second parameter
22565 of the call is not flagged;
22570 This rule has no parameters.
22575 @node Predefined_Numeric_Types
22576 @subsection @code{Predefined_Numeric_Types}
22577 @cindex @code{Predefined_Numeric_Types} rule (for @command{gnatcheck})
22580 Flag each explicit use of the name of any numeric type or subtype defined
22581 in package @code{Standard}.
22583 The rationale for this rule is to detect when the
22584 program may depend on platform-specific characteristics of the implementation
22585 of the predefined numeric types. Note that this rule is over-pessimistic;
22586 for example, a program that uses @code{String} indexing
22587 likely needs a variable of type @code{Integer}.
22588 Another example is the flagging of predefined numeric types with explicit
22591 @smallexample @c ada
22592 subtype My_Integer is Integer range Left .. Right;
22593 Vy_Var : My_Integer;
22597 This rule detects only numeric types and subtypes defined in
22598 @code{Standard}. The use of numeric types and subtypes defined in other
22599 predefined packages (such as @code{System.Any_Priority} or
22600 @code{Ada.Text_IO.Count}) is not flagged
22602 This rule has no parameters.
22606 @node Raising_External_Exceptions
22607 @subsection @code{Raising_External_Exceptions}
22608 @cindex @code{Raising_External_Exceptions} rule (for @command{gnatcheck})
22611 Flag any @code{raise} statement, in a program unit declared in a library
22612 package or in a generic library package, for an exception that is
22613 neither a predefined exception nor an exception that is also declared (or
22614 renamed) in the visible part of the package.
22616 This rule has no parameters.
22620 @node Raising_Predefined_Exceptions
22621 @subsection @code{Raising_Predefined_Exceptions}
22622 @cindex @code{Raising_Predefined_Exceptions} rule (for @command{gnatcheck})
22625 Flag each @code{raise} statement that raises a predefined exception
22626 (i.e., one of the exceptions @code{Constraint_Error}, @code{Numeric_Error},
22627 @code{Program_Error}, @code{Storage_Error}, or @code{Tasking_Error}).
22629 This rule has no parameters.
22631 @node Separate_Numeric_Error_Handlers
22632 @subsection @code{Separate_Numeric_Error_Handlers}
22633 @cindex @code{Separate_Numeric_Error_Handlers} rule (for @command{gnatcheck})
22636 Flags each exception handler that contains a choice for
22637 the predefined @code{Constraint_Error} exception, but does not contain
22638 the choice for the predefined @code{Numeric_Error} exception, or
22639 that contains the choice for @code{Numeric_Error}, but does not contain the
22640 choice for @code{Constraint_Error}.
22642 This rule has no parameters.
22646 @subsection @code{Recursion} (under construction, GLOBAL)
22647 @cindex @code{Recursion} rule (for @command{gnatcheck})
22650 Flag recursive subprograms (cycles in the call graph). Declarations, and not
22651 calls, of recursive subprograms are detected.
22653 This rule has no parameters.
22657 @node Side_Effect_Functions
22658 @subsection @code{Side_Effect_Functions} (under construction, GLOBAL)
22659 @cindex @code{Side_Effect_Functions} rule (for @command{gnatcheck})
22662 Flag functions with side effects.
22664 We define a side effect as changing any data object that is not local for the
22665 body of this function.
22667 At the moment, we do NOT consider a side effect any input-output operations
22668 (changing a state or a content of any file).
22670 We do not consider protected functions for this rule (???)
22672 There are the following sources of side effect:
22675 @item Explicit (or direct) side-effect:
22679 direct assignment to a non-local variable;
22682 direct call to an entity that is known to change some data object that is
22683 not local for the body of this function (Note, that if F1 calls F2 and F2
22684 does have a side effect, this does not automatically mean that F1 also
22685 have a side effect, because it may be the case that F2 is declared in
22686 F1's body and it changes some data object that is global for F2, but
22690 @item Indirect side-effect:
22693 Subprogram calls implicitly issued by:
22696 computing initialization expressions from type declarations as a part
22697 of object elaboration or allocator evaluation;
22699 computing implicit parameters of subprogram or entry calls or generic
22704 activation of a task that change some non-local data object (directly or
22708 elaboration code of a package that is a result of a package instantiation;
22711 controlled objects;
22714 @item Situations when we can suspect a side-effect, but the full static check
22715 is either impossible or too hard:
22718 assignment to access variables or to the objects pointed by access
22722 call to a subprogram pointed by access-to-subprogram value
22730 This rule has no parameters.
22734 @subsection @code{Slices}
22735 @cindex @code{Slices} rule (for @command{gnatcheck})
22738 Flag all uses of array slicing
22740 This rule has no parameters.
22743 @node Too_Many_Parents
22744 @subsection @code{Too_Many_Parents}
22745 @cindex @code{Too_Many_Parents} rule (for @command{gnatcheck})
22748 Flags any type declaration, single task declaration or single protected
22749 declaration that has more then @option{N} parents, @option{N} is a parameter
22751 A parent here is either a (sub)type denoted by the subtype mark from the
22752 parent_subtype_indication (in case of a derived type declaration), or
22753 any of the progenitors from the interface list, if any.
22755 This rule has the following (mandatory) parameters for the @option{+R} option:
22759 Positive integer specifying the maximal allowed number of parents.
22763 @node Unassigned_OUT_Parameters
22764 @subsection @code{Unassigned_OUT_Parameters}
22765 @cindex @code{Unassigned_OUT_Parameters} rule (for @command{gnatcheck})
22768 Flags procedures' @code{out} parameters that are not assigned, and
22769 identifies the contexts in which the assignments are missing.
22771 An @code{out} parameter is flagged in the statements in the procedure
22772 body's handled sequence of statements (before the procedure body's
22773 @code{exception} part, if any) if this sequence of statements contains
22774 no assignments to the parameter.
22776 An @code{out} parameter is flagged in an exception handler in the exception
22777 part of the procedure body's handled sequence of statements if the handler
22778 contains no assignment to the parameter.
22780 Bodies of generic procedures are also considered.
22782 The following are treated as assignments to an @code{out} parameter:
22786 an assignment statement, with the parameter or some component as the target;
22789 passing the parameter (or one of its components) as an @code{out} or
22790 @code{in out} parameter.
22794 This rule does not have any parameters.
22798 @node Uncommented_BEGIN_In_Package_Bodies
22799 @subsection @code{Uncommented_BEGIN_In_Package_Bodies}
22800 @cindex @code{Uncommented_BEGIN_In_Package_Bodies} rule (for @command{gnatcheck})
22803 Flags each package body with declarations and a statement part that does not
22804 include a trailing comment on the line containing the @code{begin} keyword;
22805 this trailing comment needs to specify the package name and nothing else.
22806 The @code{begin} is not flagged if the package body does not
22807 contain any declarations.
22809 If the @code{begin} keyword is placed on the
22810 same line as the last declaration or the first statement, it is flagged
22811 independently of whether the line contains a trailing comment. The
22812 diagnostic message is attached to the line containing the first statement.
22814 This rule has no parameters.
22816 @node Unconditional_Exits
22817 @subsection @code{Unconditional_Exits}
22818 @cindex @code{Unconditional_Exits} rule (for @command{gnatcheck})
22821 Flag unconditional @code{exit} statements.
22823 This rule has no parameters.
22825 @node Unconstrained_Array_Returns
22826 @subsection @code{Unconstrained_Array_Returns}
22827 @cindex @code{Unconstrained_Array_Returns} rule (for @command{gnatcheck})
22830 Flag each function returning an unconstrained array. Function declarations,
22831 function bodies (and body stubs) having no separate specifications,
22832 and generic function instantiations are checked.
22833 Function calls and function renamings are
22836 Generic function declarations, and function declarations in generic
22837 packages are not checked, instead this rule checks the results of
22838 generic instantiations (that is, expanded specification and expanded
22839 body corresponding to an instantiation).
22841 This rule has no parameters.
22843 @node Universal_Ranges
22844 @subsection @code{Universal_Ranges}
22845 @cindex @code{Universal_Ranges} rule (for @command{gnatcheck})
22848 Flag discrete ranges that are a part of an index constraint, constrained
22849 array definition, or @code{for}-loop parameter specification, and whose bounds
22850 are both of type @i{universal_integer}. Ranges that have at least one
22851 bound of a specific type (such as @code{1 .. N}, where @code{N} is a variable
22852 or an expression of non-universal type) are not flagged.
22854 This rule has no parameters.
22857 @node Unnamed_Blocks_And_Loops
22858 @subsection @code{Unnamed_Blocks_And_Loops}
22859 @cindex @code{Unnamed_Blocks_And_Loops} rule (for @command{gnatcheck})
22862 Flag each unnamed block statement and loop statement.
22864 The rule has no parameters.
22869 @node Unused_Subprograms
22870 @subsection @code{Unused_Subprograms} (under construction, GLOBAL)
22871 @cindex @code{Unused_Subprograms} rule (for @command{gnatcheck})
22874 Flag all unused subprograms.
22876 This rule has no parameters.
22882 @node USE_PACKAGE_Clauses
22883 @subsection @code{USE_PACKAGE_Clauses}
22884 @cindex @code{USE_PACKAGE_Clauses} rule (for @command{gnatcheck})
22887 Flag all @code{use} clauses for packages; @code{use type} clauses are
22890 This rule has no parameters.
22893 @node Visible_Components
22894 @subsection @code{Visible_Components}
22895 @cindex @code{Visible_Components} rule (for @command{gnatcheck})
22898 Flags all the type declarations located in the visible part of a library
22899 package or a library generic package that can declare a visible component. A
22900 type is considered as declaring a visible component if it contains a record
22901 definition by its own or as a part of a record extension. Type declaration is
22902 flagged even if it contains a record definition that defines no components.
22904 Declarations located in private parts of local (generic) packages are not
22905 flagged. Declarations in private packages are not flagged.
22907 This rule has no parameters.
22910 @node Volatile_Objects_Without_Address_Clauses
22911 @subsection @code{Volatile_Objects_Without_Address_Clauses}
22912 @cindex @code{Volatile_Objects_Without_Address_Clauses} rule (for @command{gnatcheck})
22915 Flag each volatile object that does not have an address clause.
22917 The following check is made: if the pragma @code{Volatile} is applied to a
22918 data object or to its type, then an address clause must
22919 be supplied for this object.
22921 This rule does not check the components of data objects,
22922 array components that are volatile as a result of the pragma
22923 @code{Volatile_Components}, or objects that are volatile because
22924 they are atomic as a result of pragmas @code{Atomic} or
22925 @code{Atomic_Components}.
22927 Only variable declarations, and not constant declarations, are checked.
22929 This rule has no parameters.
22932 @c *********************************
22933 @node Creating Sample Bodies Using gnatstub
22934 @chapter Creating Sample Bodies Using @command{gnatstub}
22938 @command{gnatstub} creates body stubs, that is, empty but compilable bodies
22939 for library unit declarations.
22941 Note: to invoke @code{gnatstub} with a project file, use the @code{gnat}
22942 driver (see @ref{The GNAT Driver and Project Files}).
22944 To create a body stub, @command{gnatstub} has to compile the library
22945 unit declaration. Therefore, bodies can be created only for legal
22946 library units. Moreover, if a library unit depends semantically upon
22947 units located outside the current directory, you have to provide
22948 the source search path when calling @command{gnatstub}, see the description
22949 of @command{gnatstub} switches below.
22951 By default, all the program unit body stubs generated by @code{gnatstub}
22952 raise the predefined @code{Program_Error} exception, which will catch
22953 accidental calls of generated stubs. This behavior can be changed with
22954 option @option{^--no-exception^/NO_EXCEPTION^} (see below).
22957 * Running gnatstub::
22958 * Switches for gnatstub::
22961 @node Running gnatstub
22962 @section Running @command{gnatstub}
22965 @command{gnatstub} has the command-line interface of the form
22968 $ gnatstub @ovar{switches} @var{filename} @ovar{directory}
22975 is the name of the source file that contains a library unit declaration
22976 for which a body must be created. The file name may contain the path
22978 The file name does not have to follow the GNAT file name conventions. If the
22980 does not follow GNAT file naming conventions, the name of the body file must
22982 explicitly as the value of the @option{^-o^/BODY=^@var{body-name}} option.
22983 If the file name follows the GNAT file naming
22984 conventions and the name of the body file is not provided,
22987 of the body file from the argument file name by replacing the @file{.ads}
22989 with the @file{.adb} suffix.
22992 indicates the directory in which the body stub is to be placed (the default
22997 is an optional sequence of switches as described in the next section
23000 @node Switches for gnatstub
23001 @section Switches for @command{gnatstub}
23007 @cindex @option{^-f^/FULL^} (@command{gnatstub})
23008 If the destination directory already contains a file with the name of the
23010 for the argument spec file, replace it with the generated body stub.
23012 @item ^-hs^/HEADER=SPEC^
23013 @cindex @option{^-hs^/HEADER=SPEC^} (@command{gnatstub})
23014 Put the comment header (i.e., all the comments preceding the
23015 compilation unit) from the source of the library unit declaration
23016 into the body stub.
23018 @item ^-hg^/HEADER=GENERAL^
23019 @cindex @option{^-hg^/HEADER=GENERAL^} (@command{gnatstub})
23020 Put a sample comment header into the body stub.
23022 @item ^--header-file=@var{filename}^/FROM_HEADER_FILE=@var{filename}^
23023 @cindex @option{^--header-file^/FROM_HEADER_FILE=^} (@command{gnatstub})
23024 Use the content of the file as the comment header for a generated body stub.
23028 @cindex @option{-IDIR} (@command{gnatstub})
23030 @cindex @option{-I-} (@command{gnatstub})
23033 @item /NOCURRENT_DIRECTORY
23034 @cindex @option{/NOCURRENT_DIRECTORY} (@command{gnatstub})
23036 ^These switches have ^This switch has^ the same meaning as in calls to
23038 ^They define ^It defines ^ the source search path in the call to
23039 @command{gcc} issued
23040 by @command{gnatstub} to compile an argument source file.
23042 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE=^@var{PATH}
23043 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatstub})
23044 This switch has the same meaning as in calls to @command{gcc}.
23045 It defines the additional configuration file to be passed to the call to
23046 @command{gcc} issued
23047 by @command{gnatstub} to compile an argument source file.
23049 @item ^-gnatyM^/MAX_LINE_LENGTH=^@var{n}
23050 @cindex @option{^-gnatyM^/MAX_LINE_LENGTH^} (@command{gnatstub})
23051 (@var{n} is a non-negative integer). Set the maximum line length in the
23052 body stub to @var{n}; the default is 79. The maximum value that can be
23053 specified is 32767. Note that in the special case of configuration
23054 pragma files, the maximum is always 32767 regardless of whether or
23055 not this switch appears.
23057 @item ^-gnaty^/STYLE_CHECKS=^@var{n}
23058 @cindex @option{^-gnaty^/STYLE_CHECKS=^} (@command{gnatstub})
23059 (@var{n} is a non-negative integer from 1 to 9). Set the indentation level in
23060 the generated body sample to @var{n}.
23061 The default indentation is 3.
23063 @item ^-gnatyo^/ORDERED_SUBPROGRAMS^
23064 @cindex @option{^-gnato^/ORDERED_SUBPROGRAMS^} (@command{gnatstub})
23065 Order local bodies alphabetically. (By default local bodies are ordered
23066 in the same way as the corresponding local specs in the argument spec file.)
23068 @item ^-i^/INDENTATION=^@var{n}
23069 @cindex @option{^-i^/INDENTATION^} (@command{gnatstub})
23070 Same as @option{^-gnaty^/STYLE_CHECKS=^@var{n}}
23072 @item ^-k^/TREE_FILE=SAVE^
23073 @cindex @option{^-k^/TREE_FILE=SAVE^} (@command{gnatstub})
23074 Do not remove the tree file (i.e., the snapshot of the compiler internal
23075 structures used by @command{gnatstub}) after creating the body stub.
23077 @item ^-l^/LINE_LENGTH=^@var{n}
23078 @cindex @option{^-l^/LINE_LENGTH^} (@command{gnatstub})
23079 Same as @option{^-gnatyM^/MAX_LINE_LENGTH=^@var{n}}
23081 @item ^--no-exception^/NO_EXCEPTION^
23082 @cindex @option{^--no-exception^/NO_EXCEPTION^} (@command{gnatstub})
23083 Avoind raising PROGRAM_ERROR in the generated bodies of program unit stubs.
23084 This is not always possible for function stubs.
23086 @item ^--no-local-header^/NO_LOCAL_HEADER^
23087 @cindex @option{^--no-local-header^/NO_LOCAL_HEADER^} (@command{gnatstub})
23088 Do not place local comment header with unit name before body stub for a
23091 @item ^-o ^/BODY=^@var{body-name}
23092 @cindex @option{^-o^/BODY^} (@command{gnatstub})
23093 Body file name. This should be set if the argument file name does not
23095 the GNAT file naming
23096 conventions. If this switch is omitted the default name for the body will be
23098 from the argument file name according to the GNAT file naming conventions.
23101 @cindex @option{^-q^/QUIET^} (@command{gnatstub})
23102 Quiet mode: do not generate a confirmation when a body is
23103 successfully created, and do not generate a message when a body is not
23107 @item ^-r^/TREE_FILE=REUSE^
23108 @cindex @option{^-r^/TREE_FILE=REUSE^} (@command{gnatstub})
23109 Reuse the tree file (if it exists) instead of creating it. Instead of
23110 creating the tree file for the library unit declaration, @command{gnatstub}
23111 tries to find it in the current directory and use it for creating
23112 a body. If the tree file is not found, no body is created. This option
23113 also implies @option{^-k^/SAVE^}, whether or not
23114 the latter is set explicitly.
23116 @item ^-t^/TREE_FILE=OVERWRITE^
23117 @cindex @option{^-t^/TREE_FILE=OVERWRITE^} (@command{gnatstub})
23118 Overwrite the existing tree file. If the current directory already
23119 contains the file which, according to the GNAT file naming rules should
23120 be considered as a tree file for the argument source file,
23122 will refuse to create the tree file needed to create a sample body
23123 unless this option is set.
23125 @item ^-v^/VERBOSE^
23126 @cindex @option{^-v^/VERBOSE^} (@command{gnatstub})
23127 Verbose mode: generate version information.
23131 @c *********************************
23132 @node Generating Ada Bindings for C and C++ headers
23133 @chapter Generating Ada Bindings for C and C++ headers
23137 GNAT now comes with a new experimental binding generator for C and C++
23138 headers which is intended to do 95% of the tedious work of generating
23139 Ada specs from C or C++ header files. Note that this still is a work in
23140 progress, not designed to generate 100% correct Ada specs.
23142 The code generated is using the Ada 2005 syntax, which makes it
23143 easier to interface with other languages than previous versions of Ada.
23146 * Running the binding generator::
23147 * Generating bindings for C++ headers::
23151 @node Running the binding generator
23152 @section Running the binding generator
23155 The binding generator is part of the @command{gcc} compiler and can be
23156 invoked via the @option{-fdump-ada-spec} switch, which will generate Ada
23157 spec files for the header files specified on the command line, and all
23158 header files needed by these files transitivitely. For example:
23161 $ g++ -c -fdump-ada-spec -C /usr/include/time.h
23162 $ gcc -c -gnat05 *.ads
23165 will generate, under GNU/Linux, the following files: @file{time_h.ads},
23166 @file{bits_time_h.ads}, @file{stddef_h.ads}, @file{bits_types_h.ads} which
23167 correspond to the files @file{/usr/include/time.h},
23168 @file{/usr/include/bits/time.h}, etc@dots{}, and will then compile in Ada 2005
23169 mode these Ada specs.
23171 The @code{-C} switch tells @command{gcc} to extract comments from headers,
23172 and will attempt to generate corresponding Ada comments.
23174 If you want to generate a single Ada file and not the transitive closure, you
23175 can use instead the @option{-fdump-ada-spec-slim} switch.
23177 Note that we recommend when possible to use the @command{g++} driver to
23178 generate bindings, even for most C headers, since this will in general
23179 generate better Ada specs. For generating bindings for C++ headers, it is
23180 mandatory to use the @command{g++} command, or @command{gcc -x c++} which
23181 is equivalent in this case. If @command{g++} cannot work on your C headers
23182 because of incompatibilities between C and C++, then you can fallback to
23183 @command{gcc} instead.
23185 For an example of better bindings generated from the C++ front-end,
23186 the name of the parameters (when available) are actually ignored by the C
23187 front-end. Consider the following C header:
23190 extern void foo (int variable);
23193 with the C front-end, @code{variable} is ignored, and the above is handled as:
23196 extern void foo (int);
23199 generating a generic:
23202 procedure foo (param1 : int);
23205 with the C++ front-end, the name is available, and we generate:
23208 procedure foo (variable : int);
23211 In some cases, the generated bindings will be more complete or more meaningful
23212 when defining some macros, which you can do via the @option{-D} switch. This
23213 is for example the case with @file{Xlib.h} under GNU/Linux:
23216 g++ -c -fdump-ada-spec -DXLIB_ILLEGAL_ACCESS -C /usr/include/X11/Xlib.h
23219 The above will generate more complete bindings than a straight call without
23220 the @option{-DXLIB_ILLEGAL_ACCESS} switch.
23222 In other cases, it is not possible to parse a header file in a stand alone
23223 manner, because other include files need to be included first. In this
23224 case, the solution is to create a small header file including the needed
23225 @code{#include} and possible @code{#define} directives. For example, to
23226 generate Ada bindings for @file{readline/readline.h}, you need to first
23227 include @file{stdio.h}, so you can create a file with the following two
23228 lines in e.g. @file{readline1.h}:
23232 #include <readline/readline.h>
23235 and then generate Ada bindings from this file:
23238 $ g++ -c -fdump-ada-spec readline1.h
23241 @node Generating bindings for C++ headers
23242 @section Generating bindings for C++ headers
23245 Generating bindings for C++ headers is done using the same options, always
23246 with the @command{g++} compiler.
23248 In this mode, C++ classes will be mapped to Ada tagged types, constructors
23249 will be mapped using the @code{CPP_Constructor} pragma, and when possible,
23250 multiple inheritance of abstract classes will be mapped to Ada interfaces
23251 (@xref{Interfacing to C++,,,gnat_rm, GNAT Reference Manual}, for additional
23252 information on interfacing to C++).
23254 For example, given the following C++ header file:
23261 virtual int Number_Of_Teeth () = 0;
23266 virtual void Set_Owner (char* Name) = 0;
23272 virtual void Set_Age (int New_Age);
23275 class Dog : Animal, Carnivore, Domestic @{
23280 virtual int Number_Of_Teeth ();
23281 virtual void Set_Owner (char* Name);
23289 The corresponding Ada code is generated:
23291 @smallexample @c ada
23294 package Class_Carnivore is
23295 type Carnivore is limited interface;
23296 pragma Import (CPP, Carnivore);
23298 function Number_Of_Teeth (this : access Carnivore) return int is abstract;
23300 use Class_Carnivore;
23302 package Class_Domestic is
23303 type Domestic is limited interface;
23304 pragma Import (CPP, Domestic);
23306 procedure Set_Owner
23307 (this : access Domestic;
23308 Name : Interfaces.C.Strings.chars_ptr) is abstract;
23310 use Class_Domestic;
23312 package Class_Animal is
23313 type Animal is tagged limited record
23314 Age_Count : aliased int;
23316 pragma Import (CPP, Animal);
23318 procedure Set_Age (this : access Animal; New_Age : int);
23319 pragma Import (CPP, Set_Age, "_ZN6Animal7Set_AgeEi");
23323 package Class_Dog is
23324 type Dog is new Animal and Carnivore and Domestic with record
23325 Tooth_Count : aliased int;
23326 Owner : Interfaces.C.Strings.chars_ptr;
23328 pragma Import (CPP, Dog);
23330 function Number_Of_Teeth (this : access Dog) return int;
23331 pragma Import (CPP, Number_Of_Teeth, "_ZN3Dog15Number_Of_TeethEv");
23333 procedure Set_Owner
23334 (this : access Dog; Name : Interfaces.C.Strings.chars_ptr);
23335 pragma Import (CPP, Set_Owner, "_ZN3Dog9Set_OwnerEPc");
23337 function New_Dog return Dog;
23338 pragma CPP_Constructor (New_Dog);
23339 pragma Import (CPP, New_Dog, "_ZN3DogC1Ev");
23350 @item -fdump-ada-spec
23351 @cindex @option{-fdump-ada-spec} (@command{gcc})
23352 Generate Ada spec files for the given header files transitively (including
23353 all header files that these headers depend upon).
23355 @item -fdump-ada-spec-slim
23356 @cindex @option{-fdump-ada-spec-slim} (@command{gcc})
23357 Generate Ada spec files for the header files specified on the command line
23361 @cindex @option{-C} (@command{gcc})
23362 Extract comments from headers and generate Ada comments in the Ada spec files.
23365 @node Other Utility Programs
23366 @chapter Other Utility Programs
23369 This chapter discusses some other utility programs available in the Ada
23373 * Using Other Utility Programs with GNAT::
23374 * The External Symbol Naming Scheme of GNAT::
23375 * Converting Ada Files to html with gnathtml::
23376 * Installing gnathtml::
23383 @node Using Other Utility Programs with GNAT
23384 @section Using Other Utility Programs with GNAT
23387 The object files generated by GNAT are in standard system format and in
23388 particular the debugging information uses this format. This means
23389 programs generated by GNAT can be used with existing utilities that
23390 depend on these formats.
23393 In general, any utility program that works with C will also often work with
23394 Ada programs generated by GNAT. This includes software utilities such as
23395 gprof (a profiling program), @code{gdb} (the FSF debugger), and utilities such
23399 @node The External Symbol Naming Scheme of GNAT
23400 @section The External Symbol Naming Scheme of GNAT
23403 In order to interpret the output from GNAT, when using tools that are
23404 originally intended for use with other languages, it is useful to
23405 understand the conventions used to generate link names from the Ada
23408 All link names are in all lowercase letters. With the exception of library
23409 procedure names, the mechanism used is simply to use the full expanded
23410 Ada name with dots replaced by double underscores. For example, suppose
23411 we have the following package spec:
23413 @smallexample @c ada
23424 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
23425 the corresponding link name is @code{qrs__mn}.
23427 Of course if a @code{pragma Export} is used this may be overridden:
23429 @smallexample @c ada
23434 pragma Export (Var1, C, External_Name => "var1_name");
23436 pragma Export (Var2, C, Link_Name => "var2_link_name");
23443 In this case, the link name for @var{Var1} is whatever link name the
23444 C compiler would assign for the C function @var{var1_name}. This typically
23445 would be either @var{var1_name} or @var{_var1_name}, depending on operating
23446 system conventions, but other possibilities exist. The link name for
23447 @var{Var2} is @var{var2_link_name}, and this is not operating system
23451 One exception occurs for library level procedures. A potential ambiguity
23452 arises between the required name @code{_main} for the C main program,
23453 and the name we would otherwise assign to an Ada library level procedure
23454 called @code{Main} (which might well not be the main program).
23456 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
23457 names. So if we have a library level procedure such as
23459 @smallexample @c ada
23462 procedure Hello (S : String);
23468 the external name of this procedure will be @var{_ada_hello}.
23471 @node Converting Ada Files to html with gnathtml
23472 @section Converting Ada Files to HTML with @code{gnathtml}
23475 This @code{Perl} script allows Ada source files to be browsed using
23476 standard Web browsers. For installation procedure, see the section
23477 @xref{Installing gnathtml}.
23479 Ada reserved keywords are highlighted in a bold font and Ada comments in
23480 a blue font. Unless your program was compiled with the gcc @option{-gnatx}
23481 switch to suppress the generation of cross-referencing information, user
23482 defined variables and types will appear in a different color; you will
23483 be able to click on any identifier and go to its declaration.
23485 The command line is as follow:
23487 $ perl gnathtml.pl @ovar{^switches^options^} @var{ada-files}
23491 You can pass it as many Ada files as you want. @code{gnathtml} will generate
23492 an html file for every ada file, and a global file called @file{index.htm}.
23493 This file is an index of every identifier defined in the files.
23495 The available ^switches^options^ are the following ones:
23499 @cindex @option{-83} (@code{gnathtml})
23500 Only the Ada 83 subset of keywords will be highlighted.
23502 @item -cc @var{color}
23503 @cindex @option{-cc} (@code{gnathtml})
23504 This option allows you to change the color used for comments. The default
23505 value is green. The color argument can be any name accepted by html.
23508 @cindex @option{-d} (@code{gnathtml})
23509 If the Ada files depend on some other files (for instance through
23510 @code{with} clauses, the latter files will also be converted to html.
23511 Only the files in the user project will be converted to html, not the files
23512 in the run-time library itself.
23515 @cindex @option{-D} (@code{gnathtml})
23516 This command is the same as @option{-d} above, but @command{gnathtml} will
23517 also look for files in the run-time library, and generate html files for them.
23519 @item -ext @var{extension}
23520 @cindex @option{-ext} (@code{gnathtml})
23521 This option allows you to change the extension of the generated HTML files.
23522 If you do not specify an extension, it will default to @file{htm}.
23525 @cindex @option{-f} (@code{gnathtml})
23526 By default, gnathtml will generate html links only for global entities
23527 ('with'ed units, global variables and types,@dots{}). If you specify
23528 @option{-f} on the command line, then links will be generated for local
23531 @item -l @var{number}
23532 @cindex @option{-l} (@code{gnathtml})
23533 If this ^switch^option^ is provided and @var{number} is not 0, then
23534 @code{gnathtml} will number the html files every @var{number} line.
23537 @cindex @option{-I} (@code{gnathtml})
23538 Specify a directory to search for library files (@file{.ALI} files) and
23539 source files. You can provide several -I switches on the command line,
23540 and the directories will be parsed in the order of the command line.
23543 @cindex @option{-o} (@code{gnathtml})
23544 Specify the output directory for html files. By default, gnathtml will
23545 saved the generated html files in a subdirectory named @file{html/}.
23547 @item -p @var{file}
23548 @cindex @option{-p} (@code{gnathtml})
23549 If you are using Emacs and the most recent Emacs Ada mode, which provides
23550 a full Integrated Development Environment for compiling, checking,
23551 running and debugging applications, you may use @file{.gpr} files
23552 to give the directories where Emacs can find sources and object files.
23554 Using this ^switch^option^, you can tell gnathtml to use these files.
23555 This allows you to get an html version of your application, even if it
23556 is spread over multiple directories.
23558 @item -sc @var{color}
23559 @cindex @option{-sc} (@code{gnathtml})
23560 This ^switch^option^ allows you to change the color used for symbol
23562 The default value is red. The color argument can be any name accepted by html.
23564 @item -t @var{file}
23565 @cindex @option{-t} (@code{gnathtml})
23566 This ^switch^option^ provides the name of a file. This file contains a list of
23567 file names to be converted, and the effect is exactly as though they had
23568 appeared explicitly on the command line. This
23569 is the recommended way to work around the command line length limit on some
23574 @node Installing gnathtml
23575 @section Installing @code{gnathtml}
23578 @code{Perl} needs to be installed on your machine to run this script.
23579 @code{Perl} is freely available for almost every architecture and
23580 Operating System via the Internet.
23582 On Unix systems, you may want to modify the first line of the script
23583 @code{gnathtml}, to explicitly tell the Operating system where Perl
23584 is. The syntax of this line is:
23586 #!full_path_name_to_perl
23590 Alternatively, you may run the script using the following command line:
23593 $ perl gnathtml.pl @ovar{switches} @var{files}
23602 The GNAT distribution provides an Ada 95 template for the HP Language
23603 Sensitive Editor (LSE), a component of DECset. In order to
23604 access it, invoke LSE with the qualifier /ENVIRONMENT=GNU:[LIB]ADA95.ENV.
23611 GNAT supports The HP Performance Coverage Analyzer (PCA), a component
23612 of DECset. To use it proceed as outlined under ``HELP PCA'', except for running
23613 the collection phase with the /DEBUG qualifier.
23616 $ GNAT MAKE /DEBUG <PROGRAM_NAME>
23617 $ DEFINE LIB$DEBUG PCA$COLLECTOR
23618 $ RUN/DEBUG <PROGRAM_NAME>
23624 @c ******************************
23625 @node Code Coverage and Profiling
23626 @chapter Code Coverage and Profiling
23627 @cindex Code Coverage
23631 This chapter describes how to use @code{gcov} - coverage testing tool - and
23632 @code{gprof} - profiler tool - on your Ada programs.
23635 * Code Coverage of Ada Programs using gcov::
23636 * Profiling an Ada Program using gprof::
23639 @node Code Coverage of Ada Programs using gcov
23640 @section Code Coverage of Ada Programs using gcov
23642 @cindex -fprofile-arcs
23643 @cindex -ftest-coverage
23645 @cindex Code Coverage
23648 @code{gcov} is a test coverage program: it analyzes the execution of a given
23649 program on selected tests, to help you determine the portions of the program
23650 that are still untested.
23652 @code{gcov} is part of the GCC suite, and is described in detail in the GCC
23653 User's Guide. You can refer to this documentation for a more complete
23656 This chapter provides a quick startup guide, and
23657 details some Gnat-specific features.
23660 * Quick startup guide::
23664 @node Quick startup guide
23665 @subsection Quick startup guide
23667 In order to perform coverage analysis of a program using @code{gcov}, 3
23672 Code instrumentation during the compilation process
23674 Execution of the instrumented program
23676 Execution of the @code{gcov} tool to generate the result.
23679 The code instrumentation needed by gcov is created at the object level:
23680 The source code is not modified in any way, because the instrumentation code is
23681 inserted by gcc during the compilation process. To compile your code with code
23682 coverage activated, you need to recompile your whole project using the
23684 @code{-fprofile-arcs} and @code{-ftest-coverage}, and link it using
23685 @code{-fprofile-arcs}.
23688 $ gnatmake -P my_project.gpr -f -cargs -fprofile-arcs -ftest-coverage \
23689 -largs -fprofile-arcs
23692 This compilation process will create @file{.gcno} files together with
23693 the usual object files.
23695 Once the program is compiled with coverage instrumentation, you can
23696 run it as many times as needed - on portions of a test suite for
23697 example. The first execution will produce @file{.gcda} files at the
23698 same location as the @file{.gcno} files. The following executions
23699 will update those files, so that a cumulative result of the covered
23700 portions of the program is generated.
23702 Finally, you need to call the @code{gcov} tool. The different options of
23703 @code{gcov} are available in the GCC User's Guide, section 'Invoking gcov'.
23705 This will create annotated source files with a @file{.gcov} extension:
23706 @file{my_main.adb} file will be analysed in @file{my_main.adb.gcov}.
23708 @node Gnat specifics
23709 @subsection Gnat specifics
23711 Because Ada semantics, portions of the source code may be shared among
23712 several object files. This is the case for example when generics are
23713 involved, when inlining is active or when declarations generate initialisation
23714 calls. In order to take
23715 into account this shared code, you need to call @code{gcov} on all
23716 source files of the tested program at once.
23718 The list of source files might exceed the system's maximum command line
23719 length. In order to bypass this limitation, a new mechanism has been
23720 implemented in @code{gcov}: you can now list all your project's files into a
23721 text file, and provide this file to gcov as a parameter, preceded by a @@
23722 (e.g. @samp{gcov @@mysrclist.txt}).
23724 Note that on AIX compiling a static library with @code{-fprofile-arcs} is
23725 not supported as there can be unresolved symbols during the final link.
23727 @node Profiling an Ada Program using gprof
23728 @section Profiling an Ada Program using gprof
23734 This section is not meant to be an exhaustive documentation of @code{gprof}.
23735 Full documentation for it can be found in the GNU Profiler User's Guide
23736 documentation that is part of this GNAT distribution.
23738 Profiling a program helps determine the parts of a program that are executed
23739 most often, and are therefore the most time-consuming.
23741 @code{gprof} is the standard GNU profiling tool; it has been enhanced to
23742 better handle Ada programs and multitasking.
23743 It is currently supported on the following platforms
23748 solaris sparc/sparc64/x86
23754 In order to profile a program using @code{gprof}, 3 steps are needed:
23758 Code instrumentation, requiring a full recompilation of the project with the
23761 Execution of the program under the analysis conditions, i.e. with the desired
23764 Analysis of the results using the @code{gprof} tool.
23768 The following sections detail the different steps, and indicate how
23769 to interpret the results:
23771 * Compilation for profiling::
23772 * Program execution::
23774 * Interpretation of profiling results::
23777 @node Compilation for profiling
23778 @subsection Compilation for profiling
23782 In order to profile a program the first step is to tell the compiler
23783 to generate the necessary profiling information. The compiler switch to be used
23784 is @code{-pg}, which must be added to other compilation switches. This
23785 switch needs to be specified both during compilation and link stages, and can
23786 be specified once when using gnatmake:
23789 gnatmake -f -pg -P my_project
23793 Note that only the objects that were compiled with the @samp{-pg} switch will be
23794 profiled; if you need to profile your whole project, use the
23795 @samp{-f} gnatmake switch to force full recompilation.
23797 @node Program execution
23798 @subsection Program execution
23801 Once the program has been compiled for profiling, you can run it as usual.
23803 The only constraint imposed by profiling is that the program must terminate
23804 normally. An interrupted program (via a Ctrl-C, kill, etc.) will not be
23807 Once the program completes execution, a data file called @file{gmon.out} is
23808 generated in the directory where the program was launched from. If this file
23809 already exists, it will be overwritten.
23811 @node Running gprof
23812 @subsection Running gprof
23815 The @code{gprof} tool is called as follow:
23818 gprof my_prog gmon.out
23829 The complete form of the gprof command line is the following:
23832 gprof [^switches^options^] [executable [data-file]]
23836 @code{gprof} supports numerous ^switch^options^. The order of these
23837 ^switch^options^ does not matter. The full list of options can be found in
23838 the GNU Profiler User's Guide documentation that comes with this documentation.
23840 The following is the subset of those switches that is most relevant:
23844 @item --demangle[=@var{style}]
23845 @itemx --no-demangle
23846 @cindex @option{--demangle} (@code{gprof})
23847 These options control whether symbol names should be demangled when
23848 printing output. The default is to demangle C++ symbols. The
23849 @code{--no-demangle} option may be used to turn off demangling. Different
23850 compilers have different mangling styles. The optional demangling style
23851 argument can be used to choose an appropriate demangling style for your
23852 compiler, in particular Ada symbols generated by GNAT can be demangled using
23853 @code{--demangle=gnat}.
23855 @item -e @var{function_name}
23856 @cindex @option{-e} (@code{gprof})
23857 The @samp{-e @var{function}} option tells @code{gprof} not to print
23858 information about the function @var{function_name} (and its
23859 children@dots{}) in the call graph. The function will still be listed
23860 as a child of any functions that call it, but its index number will be
23861 shown as @samp{[not printed]}. More than one @samp{-e} option may be
23862 given; only one @var{function_name} may be indicated with each @samp{-e}
23865 @item -E @var{function_name}
23866 @cindex @option{-E} (@code{gprof})
23867 The @code{-E @var{function}} option works like the @code{-e} option, but
23868 execution time spent in the function (and children who were not called from
23869 anywhere else), will not be used to compute the percentages-of-time for
23870 the call graph. More than one @samp{-E} option may be given; only one
23871 @var{function_name} may be indicated with each @samp{-E} option.
23873 @item -f @var{function_name}
23874 @cindex @option{-f} (@code{gprof})
23875 The @samp{-f @var{function}} option causes @code{gprof} to limit the
23876 call graph to the function @var{function_name} and its children (and
23877 their children@dots{}). More than one @samp{-f} option may be given;
23878 only one @var{function_name} may be indicated with each @samp{-f}
23881 @item -F @var{function_name}
23882 @cindex @option{-F} (@code{gprof})
23883 The @samp{-F @var{function}} option works like the @code{-f} option, but
23884 only time spent in the function and its children (and their
23885 children@dots{}) will be used to determine total-time and
23886 percentages-of-time for the call graph. More than one @samp{-F} option
23887 may be given; only one @var{function_name} may be indicated with each
23888 @samp{-F} option. The @samp{-F} option overrides the @samp{-E} option.
23892 @node Interpretation of profiling results
23893 @subsection Interpretation of profiling results
23897 The results of the profiling analysis are represented by two arrays: the
23898 'flat profile' and the 'call graph'. Full documentation of those outputs
23899 can be found in the GNU Profiler User's Guide.
23901 The flat profile shows the time spent in each function of the program, and how
23902 many time it has been called. This allows you to locate easily the most
23903 time-consuming functions.
23905 The call graph shows, for each subprogram, the subprograms that call it,
23906 and the subprograms that it calls. It also provides an estimate of the time
23907 spent in each of those callers/called subprograms.
23910 @c ******************************
23911 @node Running and Debugging Ada Programs
23912 @chapter Running and Debugging Ada Programs
23916 This chapter discusses how to debug Ada programs.
23918 It applies to GNAT on the Alpha OpenVMS platform;
23919 for I64 OpenVMS please refer to the @cite{OpenVMS Debugger Manual},
23920 since HP has implemented Ada support in the OpenVMS debugger on I64.
23923 An incorrect Ada program may be handled in three ways by the GNAT compiler:
23927 The illegality may be a violation of the static semantics of Ada. In
23928 that case GNAT diagnoses the constructs in the program that are illegal.
23929 It is then a straightforward matter for the user to modify those parts of
23933 The illegality may be a violation of the dynamic semantics of Ada. In
23934 that case the program compiles and executes, but may generate incorrect
23935 results, or may terminate abnormally with some exception.
23938 When presented with a program that contains convoluted errors, GNAT
23939 itself may terminate abnormally without providing full diagnostics on
23940 the incorrect user program.
23944 * The GNAT Debugger GDB::
23946 * Introduction to GDB Commands::
23947 * Using Ada Expressions::
23948 * Calling User-Defined Subprograms::
23949 * Using the Next Command in a Function::
23952 * Debugging Generic Units::
23953 * GNAT Abnormal Termination or Failure to Terminate::
23954 * Naming Conventions for GNAT Source Files::
23955 * Getting Internal Debugging Information::
23956 * Stack Traceback::
23962 @node The GNAT Debugger GDB
23963 @section The GNAT Debugger GDB
23966 @code{GDB} is a general purpose, platform-independent debugger that
23967 can be used to debug mixed-language programs compiled with @command{gcc},
23968 and in particular is capable of debugging Ada programs compiled with
23969 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
23970 complex Ada data structures.
23972 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
23974 located in the GNU:[DOCS] directory,
23976 for full details on the usage of @code{GDB}, including a section on
23977 its usage on programs. This manual should be consulted for full
23978 details. The section that follows is a brief introduction to the
23979 philosophy and use of @code{GDB}.
23981 When GNAT programs are compiled, the compiler optionally writes debugging
23982 information into the generated object file, including information on
23983 line numbers, and on declared types and variables. This information is
23984 separate from the generated code. It makes the object files considerably
23985 larger, but it does not add to the size of the actual executable that
23986 will be loaded into memory, and has no impact on run-time performance. The
23987 generation of debug information is triggered by the use of the
23988 ^-g^/DEBUG^ switch in the @command{gcc} or @command{gnatmake} command
23989 used to carry out the compilations. It is important to emphasize that
23990 the use of these options does not change the generated code.
23992 The debugging information is written in standard system formats that
23993 are used by many tools, including debuggers and profilers. The format
23994 of the information is typically designed to describe C types and
23995 semantics, but GNAT implements a translation scheme which allows full
23996 details about Ada types and variables to be encoded into these
23997 standard C formats. Details of this encoding scheme may be found in
23998 the file exp_dbug.ads in the GNAT source distribution. However, the
23999 details of this encoding are, in general, of no interest to a user,
24000 since @code{GDB} automatically performs the necessary decoding.
24002 When a program is bound and linked, the debugging information is
24003 collected from the object files, and stored in the executable image of
24004 the program. Again, this process significantly increases the size of
24005 the generated executable file, but it does not increase the size of
24006 the executable program itself. Furthermore, if this program is run in
24007 the normal manner, it runs exactly as if the debug information were
24008 not present, and takes no more actual memory.
24010 However, if the program is run under control of @code{GDB}, the
24011 debugger is activated. The image of the program is loaded, at which
24012 point it is ready to run. If a run command is given, then the program
24013 will run exactly as it would have if @code{GDB} were not present. This
24014 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
24015 entirely non-intrusive until a breakpoint is encountered. If no
24016 breakpoint is ever hit, the program will run exactly as it would if no
24017 debugger were present. When a breakpoint is hit, @code{GDB} accesses
24018 the debugging information and can respond to user commands to inspect
24019 variables, and more generally to report on the state of execution.
24023 @section Running GDB
24026 This section describes how to initiate the debugger.
24027 @c The above sentence is really just filler, but it was otherwise
24028 @c clumsy to get the first paragraph nonindented given the conditional
24029 @c nature of the description
24032 The debugger can be launched from a @code{GPS} menu or
24033 directly from the command line. The description below covers the latter use.
24034 All the commands shown can be used in the @code{GPS} debug console window,
24035 but there are usually more GUI-based ways to achieve the same effect.
24038 The command to run @code{GDB} is
24041 $ ^gdb program^GDB PROGRAM^
24045 where @code{^program^PROGRAM^} is the name of the executable file. This
24046 activates the debugger and results in a prompt for debugger commands.
24047 The simplest command is simply @code{run}, which causes the program to run
24048 exactly as if the debugger were not present. The following section
24049 describes some of the additional commands that can be given to @code{GDB}.
24051 @c *******************************
24052 @node Introduction to GDB Commands
24053 @section Introduction to GDB Commands
24056 @code{GDB} contains a large repertoire of commands. @xref{Top,,
24057 Debugging with GDB, gdb, Debugging with GDB},
24059 located in the GNU:[DOCS] directory,
24061 for extensive documentation on the use
24062 of these commands, together with examples of their use. Furthermore,
24063 the command @command{help} invoked from within GDB activates a simple help
24064 facility which summarizes the available commands and their options.
24065 In this section we summarize a few of the most commonly
24066 used commands to give an idea of what @code{GDB} is about. You should create
24067 a simple program with debugging information and experiment with the use of
24068 these @code{GDB} commands on the program as you read through the
24072 @item set args @var{arguments}
24073 The @var{arguments} list above is a list of arguments to be passed to
24074 the program on a subsequent run command, just as though the arguments
24075 had been entered on a normal invocation of the program. The @code{set args}
24076 command is not needed if the program does not require arguments.
24079 The @code{run} command causes execution of the program to start from
24080 the beginning. If the program is already running, that is to say if
24081 you are currently positioned at a breakpoint, then a prompt will ask
24082 for confirmation that you want to abandon the current execution and
24085 @item breakpoint @var{location}
24086 The breakpoint command sets a breakpoint, that is to say a point at which
24087 execution will halt and @code{GDB} will await further
24088 commands. @var{location} is
24089 either a line number within a file, given in the format @code{file:linenumber},
24090 or it is the name of a subprogram. If you request that a breakpoint be set on
24091 a subprogram that is overloaded, a prompt will ask you to specify on which of
24092 those subprograms you want to breakpoint. You can also
24093 specify that all of them should be breakpointed. If the program is run
24094 and execution encounters the breakpoint, then the program
24095 stops and @code{GDB} signals that the breakpoint was encountered by
24096 printing the line of code before which the program is halted.
24098 @item breakpoint exception @var{name}
24099 A special form of the breakpoint command which breakpoints whenever
24100 exception @var{name} is raised.
24101 If @var{name} is omitted,
24102 then a breakpoint will occur when any exception is raised.
24104 @item print @var{expression}
24105 This will print the value of the given expression. Most simple
24106 Ada expression formats are properly handled by @code{GDB}, so the expression
24107 can contain function calls, variables, operators, and attribute references.
24110 Continues execution following a breakpoint, until the next breakpoint or the
24111 termination of the program.
24114 Executes a single line after a breakpoint. If the next statement
24115 is a subprogram call, execution continues into (the first statement of)
24116 the called subprogram.
24119 Executes a single line. If this line is a subprogram call, executes and
24120 returns from the call.
24123 Lists a few lines around the current source location. In practice, it
24124 is usually more convenient to have a separate edit window open with the
24125 relevant source file displayed. Successive applications of this command
24126 print subsequent lines. The command can be given an argument which is a
24127 line number, in which case it displays a few lines around the specified one.
24130 Displays a backtrace of the call chain. This command is typically
24131 used after a breakpoint has occurred, to examine the sequence of calls that
24132 leads to the current breakpoint. The display includes one line for each
24133 activation record (frame) corresponding to an active subprogram.
24136 At a breakpoint, @code{GDB} can display the values of variables local
24137 to the current frame. The command @code{up} can be used to
24138 examine the contents of other active frames, by moving the focus up
24139 the stack, that is to say from callee to caller, one frame at a time.
24142 Moves the focus of @code{GDB} down from the frame currently being
24143 examined to the frame of its callee (the reverse of the previous command),
24145 @item frame @var{n}
24146 Inspect the frame with the given number. The value 0 denotes the frame
24147 of the current breakpoint, that is to say the top of the call stack.
24152 The above list is a very short introduction to the commands that
24153 @code{GDB} provides. Important additional capabilities, including conditional
24154 breakpoints, the ability to execute command sequences on a breakpoint,
24155 the ability to debug at the machine instruction level and many other
24156 features are described in detail in @ref{Top,, Debugging with GDB, gdb,
24157 Debugging with GDB}. Note that most commands can be abbreviated
24158 (for example, c for continue, bt for backtrace).
24160 @node Using Ada Expressions
24161 @section Using Ada Expressions
24162 @cindex Ada expressions
24165 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
24166 extensions. The philosophy behind the design of this subset is
24170 That @code{GDB} should provide basic literals and access to operations for
24171 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
24172 leaving more sophisticated computations to subprograms written into the
24173 program (which therefore may be called from @code{GDB}).
24176 That type safety and strict adherence to Ada language restrictions
24177 are not particularly important to the @code{GDB} user.
24180 That brevity is important to the @code{GDB} user.
24184 Thus, for brevity, the debugger acts as if there were
24185 implicit @code{with} and @code{use} clauses in effect for all user-written
24186 packages, thus making it unnecessary to fully qualify most names with
24187 their packages, regardless of context. Where this causes ambiguity,
24188 @code{GDB} asks the user's intent.
24190 For details on the supported Ada syntax, see @ref{Top,, Debugging with
24191 GDB, gdb, Debugging with GDB}.
24193 @node Calling User-Defined Subprograms
24194 @section Calling User-Defined Subprograms
24197 An important capability of @code{GDB} is the ability to call user-defined
24198 subprograms while debugging. This is achieved simply by entering
24199 a subprogram call statement in the form:
24202 call subprogram-name (parameters)
24206 The keyword @code{call} can be omitted in the normal case where the
24207 @code{subprogram-name} does not coincide with any of the predefined
24208 @code{GDB} commands.
24210 The effect is to invoke the given subprogram, passing it the
24211 list of parameters that is supplied. The parameters can be expressions and
24212 can include variables from the program being debugged. The
24213 subprogram must be defined
24214 at the library level within your program, and @code{GDB} will call the
24215 subprogram within the environment of your program execution (which
24216 means that the subprogram is free to access or even modify variables
24217 within your program).
24219 The most important use of this facility is in allowing the inclusion of
24220 debugging routines that are tailored to particular data structures
24221 in your program. Such debugging routines can be written to provide a suitably
24222 high-level description of an abstract type, rather than a low-level dump
24223 of its physical layout. After all, the standard
24224 @code{GDB print} command only knows the physical layout of your
24225 types, not their abstract meaning. Debugging routines can provide information
24226 at the desired semantic level and are thus enormously useful.
24228 For example, when debugging GNAT itself, it is crucial to have access to
24229 the contents of the tree nodes used to represent the program internally.
24230 But tree nodes are represented simply by an integer value (which in turn
24231 is an index into a table of nodes).
24232 Using the @code{print} command on a tree node would simply print this integer
24233 value, which is not very useful. But the PN routine (defined in file
24234 treepr.adb in the GNAT sources) takes a tree node as input, and displays
24235 a useful high level representation of the tree node, which includes the
24236 syntactic category of the node, its position in the source, the integers
24237 that denote descendant nodes and parent node, as well as varied
24238 semantic information. To study this example in more detail, you might want to
24239 look at the body of the PN procedure in the stated file.
24241 @node Using the Next Command in a Function
24242 @section Using the Next Command in a Function
24245 When you use the @code{next} command in a function, the current source
24246 location will advance to the next statement as usual. A special case
24247 arises in the case of a @code{return} statement.
24249 Part of the code for a return statement is the ``epilog'' of the function.
24250 This is the code that returns to the caller. There is only one copy of
24251 this epilog code, and it is typically associated with the last return
24252 statement in the function if there is more than one return. In some
24253 implementations, this epilog is associated with the first statement
24256 The result is that if you use the @code{next} command from a return
24257 statement that is not the last return statement of the function you
24258 may see a strange apparent jump to the last return statement or to
24259 the start of the function. You should simply ignore this odd jump.
24260 The value returned is always that from the first return statement
24261 that was stepped through.
24263 @node Ada Exceptions
24264 @section Breaking on Ada Exceptions
24268 You can set breakpoints that trip when your program raises
24269 selected exceptions.
24272 @item break exception
24273 Set a breakpoint that trips whenever (any task in the) program raises
24276 @item break exception @var{name}
24277 Set a breakpoint that trips whenever (any task in the) program raises
24278 the exception @var{name}.
24280 @item break exception unhandled
24281 Set a breakpoint that trips whenever (any task in the) program raises an
24282 exception for which there is no handler.
24284 @item info exceptions
24285 @itemx info exceptions @var{regexp}
24286 The @code{info exceptions} command permits the user to examine all defined
24287 exceptions within Ada programs. With a regular expression, @var{regexp}, as
24288 argument, prints out only those exceptions whose name matches @var{regexp}.
24296 @code{GDB} allows the following task-related commands:
24300 This command shows a list of current Ada tasks, as in the following example:
24307 ID TID P-ID Thread Pri State Name
24308 1 8088000 0 807e000 15 Child Activation Wait main_task
24309 2 80a4000 1 80ae000 15 Accept/Select Wait b
24310 3 809a800 1 80a4800 15 Child Activation Wait a
24311 * 4 80ae800 3 80b8000 15 Running c
24315 In this listing, the asterisk before the first task indicates it to be the
24316 currently running task. The first column lists the task ID that is used
24317 to refer to tasks in the following commands.
24319 @item break @var{linespec} task @var{taskid}
24320 @itemx break @var{linespec} task @var{taskid} if @dots{}
24321 @cindex Breakpoints and tasks
24322 These commands are like the @code{break @dots{} thread @dots{}}.
24323 @var{linespec} specifies source lines.
24325 Use the qualifier @samp{task @var{taskid}} with a breakpoint command
24326 to specify that you only want @code{GDB} to stop the program when a
24327 particular Ada task reaches this breakpoint. @var{taskid} is one of the
24328 numeric task identifiers assigned by @code{GDB}, shown in the first
24329 column of the @samp{info tasks} display.
24331 If you do not specify @samp{task @var{taskid}} when you set a
24332 breakpoint, the breakpoint applies to @emph{all} tasks of your
24335 You can use the @code{task} qualifier on conditional breakpoints as
24336 well; in this case, place @samp{task @var{taskid}} before the
24337 breakpoint condition (before the @code{if}).
24339 @item task @var{taskno}
24340 @cindex Task switching
24342 This command allows to switch to the task referred by @var{taskno}. In
24343 particular, This allows to browse the backtrace of the specified
24344 task. It is advised to switch back to the original task before
24345 continuing execution otherwise the scheduling of the program may be
24350 For more detailed information on the tasking support,
24351 see @ref{Top,, Debugging with GDB, gdb, Debugging with GDB}.
24353 @node Debugging Generic Units
24354 @section Debugging Generic Units
24355 @cindex Debugging Generic Units
24359 GNAT always uses code expansion for generic instantiation. This means that
24360 each time an instantiation occurs, a complete copy of the original code is
24361 made, with appropriate substitutions of formals by actuals.
24363 It is not possible to refer to the original generic entities in
24364 @code{GDB}, but it is always possible to debug a particular instance of
24365 a generic, by using the appropriate expanded names. For example, if we have
24367 @smallexample @c ada
24372 generic package k is
24373 procedure kp (v1 : in out integer);
24377 procedure kp (v1 : in out integer) is
24383 package k1 is new k;
24384 package k2 is new k;
24386 var : integer := 1;
24399 Then to break on a call to procedure kp in the k2 instance, simply
24403 (gdb) break g.k2.kp
24407 When the breakpoint occurs, you can step through the code of the
24408 instance in the normal manner and examine the values of local variables, as for
24411 @node GNAT Abnormal Termination or Failure to Terminate
24412 @section GNAT Abnormal Termination or Failure to Terminate
24413 @cindex GNAT Abnormal Termination or Failure to Terminate
24416 When presented with programs that contain serious errors in syntax
24418 GNAT may on rare occasions experience problems in operation, such
24420 segmentation fault or illegal memory access, raising an internal
24421 exception, terminating abnormally, or failing to terminate at all.
24422 In such cases, you can activate
24423 various features of GNAT that can help you pinpoint the construct in your
24424 program that is the likely source of the problem.
24426 The following strategies are presented in increasing order of
24427 difficulty, corresponding to your experience in using GNAT and your
24428 familiarity with compiler internals.
24432 Run @command{gcc} with the @option{-gnatf}. This first
24433 switch causes all errors on a given line to be reported. In its absence,
24434 only the first error on a line is displayed.
24436 The @option{-gnatdO} switch causes errors to be displayed as soon as they
24437 are encountered, rather than after compilation is terminated. If GNAT
24438 terminates prematurely or goes into an infinite loop, the last error
24439 message displayed may help to pinpoint the culprit.
24442 Run @command{gcc} with the @option{^-v (verbose)^/VERBOSE^} switch. In this
24443 mode, @command{gcc} produces ongoing information about the progress of the
24444 compilation and provides the name of each procedure as code is
24445 generated. This switch allows you to find which Ada procedure was being
24446 compiled when it encountered a code generation problem.
24449 @cindex @option{-gnatdc} switch
24450 Run @command{gcc} with the @option{-gnatdc} switch. This is a GNAT specific
24451 switch that does for the front-end what @option{^-v^VERBOSE^} does
24452 for the back end. The system prints the name of each unit,
24453 either a compilation unit or nested unit, as it is being analyzed.
24455 Finally, you can start
24456 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
24457 front-end of GNAT, and can be run independently (normally it is just
24458 called from @command{gcc}). You can use @code{gdb} on @code{gnat1} as you
24459 would on a C program (but @pxref{The GNAT Debugger GDB} for caveats). The
24460 @code{where} command is the first line of attack; the variable
24461 @code{lineno} (seen by @code{print lineno}), used by the second phase of
24462 @code{gnat1} and by the @command{gcc} backend, indicates the source line at
24463 which the execution stopped, and @code{input_file name} indicates the name of
24467 @node Naming Conventions for GNAT Source Files
24468 @section Naming Conventions for GNAT Source Files
24471 In order to examine the workings of the GNAT system, the following
24472 brief description of its organization may be helpful:
24476 Files with prefix @file{^sc^SC^} contain the lexical scanner.
24479 All files prefixed with @file{^par^PAR^} are components of the parser. The
24480 numbers correspond to chapters of the Ada Reference Manual. For example,
24481 parsing of select statements can be found in @file{par-ch9.adb}.
24484 All files prefixed with @file{^sem^SEM^} perform semantic analysis. The
24485 numbers correspond to chapters of the Ada standard. For example, all
24486 issues involving context clauses can be found in @file{sem_ch10.adb}. In
24487 addition, some features of the language require sufficient special processing
24488 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
24489 dynamic dispatching, etc.
24492 All files prefixed with @file{^exp^EXP^} perform normalization and
24493 expansion of the intermediate representation (abstract syntax tree, or AST).
24494 these files use the same numbering scheme as the parser and semantics files.
24495 For example, the construction of record initialization procedures is done in
24496 @file{exp_ch3.adb}.
24499 The files prefixed with @file{^bind^BIND^} implement the binder, which
24500 verifies the consistency of the compilation, determines an order of
24501 elaboration, and generates the bind file.
24504 The files @file{atree.ads} and @file{atree.adb} detail the low-level
24505 data structures used by the front-end.
24508 The files @file{sinfo.ads} and @file{sinfo.adb} detail the structure of
24509 the abstract syntax tree as produced by the parser.
24512 The files @file{einfo.ads} and @file{einfo.adb} detail the attributes of
24513 all entities, computed during semantic analysis.
24516 Library management issues are dealt with in files with prefix
24522 Ada files with the prefix @file{^a-^A-^} are children of @code{Ada}, as
24523 defined in Annex A.
24528 Files with prefix @file{^i-^I-^} are children of @code{Interfaces}, as
24529 defined in Annex B.
24533 Files with prefix @file{^s-^S-^} are children of @code{System}. This includes
24534 both language-defined children and GNAT run-time routines.
24538 Files with prefix @file{^g-^G-^} are children of @code{GNAT}. These are useful
24539 general-purpose packages, fully documented in their specs. All
24540 the other @file{.c} files are modifications of common @command{gcc} files.
24543 @node Getting Internal Debugging Information
24544 @section Getting Internal Debugging Information
24547 Most compilers have internal debugging switches and modes. GNAT
24548 does also, except GNAT internal debugging switches and modes are not
24549 secret. A summary and full description of all the compiler and binder
24550 debug flags are in the file @file{debug.adb}. You must obtain the
24551 sources of the compiler to see the full detailed effects of these flags.
24553 The switches that print the source of the program (reconstructed from
24554 the internal tree) are of general interest for user programs, as are the
24556 the full internal tree, and the entity table (the symbol table
24557 information). The reconstructed source provides a readable version of the
24558 program after the front-end has completed analysis and expansion,
24559 and is useful when studying the performance of specific constructs.
24560 For example, constraint checks are indicated, complex aggregates
24561 are replaced with loops and assignments, and tasking primitives
24562 are replaced with run-time calls.
24564 @node Stack Traceback
24565 @section Stack Traceback
24567 @cindex stack traceback
24568 @cindex stack unwinding
24571 Traceback is a mechanism to display the sequence of subprogram calls that
24572 leads to a specified execution point in a program. Often (but not always)
24573 the execution point is an instruction at which an exception has been raised.
24574 This mechanism is also known as @i{stack unwinding} because it obtains
24575 its information by scanning the run-time stack and recovering the activation
24576 records of all active subprograms. Stack unwinding is one of the most
24577 important tools for program debugging.
24579 The first entry stored in traceback corresponds to the deepest calling level,
24580 that is to say the subprogram currently executing the instruction
24581 from which we want to obtain the traceback.
24583 Note that there is no runtime performance penalty when stack traceback
24584 is enabled, and no exception is raised during program execution.
24587 * Non-Symbolic Traceback::
24588 * Symbolic Traceback::
24591 @node Non-Symbolic Traceback
24592 @subsection Non-Symbolic Traceback
24593 @cindex traceback, non-symbolic
24596 Note: this feature is not supported on all platforms. See
24597 @file{GNAT.Traceback spec in g-traceb.ads} for a complete list of supported
24601 * Tracebacks From an Unhandled Exception::
24602 * Tracebacks From Exception Occurrences (non-symbolic)::
24603 * Tracebacks From Anywhere in a Program (non-symbolic)::
24606 @node Tracebacks From an Unhandled Exception
24607 @subsubsection Tracebacks From an Unhandled Exception
24610 A runtime non-symbolic traceback is a list of addresses of call instructions.
24611 To enable this feature you must use the @option{-E}
24612 @code{gnatbind}'s option. With this option a stack traceback is stored as part
24613 of exception information. You can retrieve this information using the
24614 @code{addr2line} tool.
24616 Here is a simple example:
24618 @smallexample @c ada
24624 raise Constraint_Error;
24639 $ gnatmake stb -bargs -E
24642 Execution terminated by unhandled exception
24643 Exception name: CONSTRAINT_ERROR
24645 Call stack traceback locations:
24646 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
24650 As we see the traceback lists a sequence of addresses for the unhandled
24651 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
24652 guess that this exception come from procedure P1. To translate these
24653 addresses into the source lines where the calls appear, the
24654 @code{addr2line} tool, described below, is invaluable. The use of this tool
24655 requires the program to be compiled with debug information.
24658 $ gnatmake -g stb -bargs -E
24661 Execution terminated by unhandled exception
24662 Exception name: CONSTRAINT_ERROR
24664 Call stack traceback locations:
24665 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
24667 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
24668 0x4011f1 0x77e892a4
24670 00401373 at d:/stb/stb.adb:5
24671 0040138B at d:/stb/stb.adb:10
24672 0040139C at d:/stb/stb.adb:14
24673 00401335 at d:/stb/b~stb.adb:104
24674 004011C4 at /build/@dots{}/crt1.c:200
24675 004011F1 at /build/@dots{}/crt1.c:222
24676 77E892A4 in ?? at ??:0
24680 The @code{addr2line} tool has several other useful options:
24684 to get the function name corresponding to any location
24686 @item --demangle=gnat
24687 to use the gnat decoding mode for the function names. Note that
24688 for binutils version 2.9.x the option is simply @option{--demangle}.
24692 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
24693 0x40139c 0x401335 0x4011c4 0x4011f1
24695 00401373 in stb.p1 at d:/stb/stb.adb:5
24696 0040138B in stb.p2 at d:/stb/stb.adb:10
24697 0040139C in stb at d:/stb/stb.adb:14
24698 00401335 in main at d:/stb/b~stb.adb:104
24699 004011C4 in <__mingw_CRTStartup> at /build/@dots{}/crt1.c:200
24700 004011F1 in <mainCRTStartup> at /build/@dots{}/crt1.c:222
24704 From this traceback we can see that the exception was raised in
24705 @file{stb.adb} at line 5, which was reached from a procedure call in
24706 @file{stb.adb} at line 10, and so on. The @file{b~std.adb} is the binder file,
24707 which contains the call to the main program.
24708 @xref{Running gnatbind}. The remaining entries are assorted runtime routines,
24709 and the output will vary from platform to platform.
24711 It is also possible to use @code{GDB} with these traceback addresses to debug
24712 the program. For example, we can break at a given code location, as reported
24713 in the stack traceback:
24719 Furthermore, this feature is not implemented inside Windows DLL. Only
24720 the non-symbolic traceback is reported in this case.
24723 (gdb) break *0x401373
24724 Breakpoint 1 at 0x401373: file stb.adb, line 5.
24728 It is important to note that the stack traceback addresses
24729 do not change when debug information is included. This is particularly useful
24730 because it makes it possible to release software without debug information (to
24731 minimize object size), get a field report that includes a stack traceback
24732 whenever an internal bug occurs, and then be able to retrieve the sequence
24733 of calls with the same program compiled with debug information.
24735 @node Tracebacks From Exception Occurrences (non-symbolic)
24736 @subsubsection Tracebacks From Exception Occurrences
24739 Non-symbolic tracebacks are obtained by using the @option{-E} binder argument.
24740 The stack traceback is attached to the exception information string, and can
24741 be retrieved in an exception handler within the Ada program, by means of the
24742 Ada facilities defined in @code{Ada.Exceptions}. Here is a simple example:
24744 @smallexample @c ada
24746 with Ada.Exceptions;
24751 use Ada.Exceptions;
24759 Text_IO.Put_Line (Exception_Information (E));
24773 This program will output:
24778 Exception name: CONSTRAINT_ERROR
24779 Message: stb.adb:12
24780 Call stack traceback locations:
24781 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
24784 @node Tracebacks From Anywhere in a Program (non-symbolic)
24785 @subsubsection Tracebacks From Anywhere in a Program
24788 It is also possible to retrieve a stack traceback from anywhere in a
24789 program. For this you need to
24790 use the @code{GNAT.Traceback} API. This package includes a procedure called
24791 @code{Call_Chain} that computes a complete stack traceback, as well as useful
24792 display procedures described below. It is not necessary to use the
24793 @option{-E gnatbind} option in this case, because the stack traceback mechanism
24794 is invoked explicitly.
24797 In the following example we compute a traceback at a specific location in
24798 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
24799 convert addresses to strings:
24801 @smallexample @c ada
24803 with GNAT.Traceback;
24804 with GNAT.Debug_Utilities;
24810 use GNAT.Traceback;
24813 TB : Tracebacks_Array (1 .. 10);
24814 -- We are asking for a maximum of 10 stack frames.
24816 -- Len will receive the actual number of stack frames returned.
24818 Call_Chain (TB, Len);
24820 Text_IO.Put ("In STB.P1 : ");
24822 for K in 1 .. Len loop
24823 Text_IO.Put (Debug_Utilities.Image (TB (K)));
24844 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
24845 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
24849 You can then get further information by invoking the @code{addr2line}
24850 tool as described earlier (note that the hexadecimal addresses
24851 need to be specified in C format, with a leading ``0x'').
24853 @node Symbolic Traceback
24854 @subsection Symbolic Traceback
24855 @cindex traceback, symbolic
24858 A symbolic traceback is a stack traceback in which procedure names are
24859 associated with each code location.
24862 Note that this feature is not supported on all platforms. See
24863 @file{GNAT.Traceback.Symbolic spec in g-trasym.ads} for a complete
24864 list of currently supported platforms.
24867 Note that the symbolic traceback requires that the program be compiled
24868 with debug information. If it is not compiled with debug information
24869 only the non-symbolic information will be valid.
24872 * Tracebacks From Exception Occurrences (symbolic)::
24873 * Tracebacks From Anywhere in a Program (symbolic)::
24876 @node Tracebacks From Exception Occurrences (symbolic)
24877 @subsubsection Tracebacks From Exception Occurrences
24879 @smallexample @c ada
24881 with GNAT.Traceback.Symbolic;
24887 raise Constraint_Error;
24904 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
24909 $ gnatmake -g .\stb -bargs -E -largs -lgnat -laddr2line -lintl
24912 0040149F in stb.p1 at stb.adb:8
24913 004014B7 in stb.p2 at stb.adb:13
24914 004014CF in stb.p3 at stb.adb:18
24915 004015DD in ada.stb at stb.adb:22
24916 00401461 in main at b~stb.adb:168
24917 004011C4 in __mingw_CRTStartup at crt1.c:200
24918 004011F1 in mainCRTStartup at crt1.c:222
24919 77E892A4 in ?? at ??:0
24923 In the above example the ``.\'' syntax in the @command{gnatmake} command
24924 is currently required by @command{addr2line} for files that are in
24925 the current working directory.
24926 Moreover, the exact sequence of linker options may vary from platform
24928 The above @option{-largs} section is for Windows platforms. By contrast,
24929 under Unix there is no need for the @option{-largs} section.
24930 Differences across platforms are due to details of linker implementation.
24932 @node Tracebacks From Anywhere in a Program (symbolic)
24933 @subsubsection Tracebacks From Anywhere in a Program
24936 It is possible to get a symbolic stack traceback
24937 from anywhere in a program, just as for non-symbolic tracebacks.
24938 The first step is to obtain a non-symbolic
24939 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
24940 information. Here is an example:
24942 @smallexample @c ada
24944 with GNAT.Traceback;
24945 with GNAT.Traceback.Symbolic;
24950 use GNAT.Traceback;
24951 use GNAT.Traceback.Symbolic;
24954 TB : Tracebacks_Array (1 .. 10);
24955 -- We are asking for a maximum of 10 stack frames.
24957 -- Len will receive the actual number of stack frames returned.
24959 Call_Chain (TB, Len);
24960 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
24973 @c ******************************
24975 @node Compatibility with HP Ada
24976 @chapter Compatibility with HP Ada
24977 @cindex Compatibility
24982 @cindex Compatibility between GNAT and HP Ada
24983 This chapter compares HP Ada (formerly known as ``DEC Ada'')
24984 for OpenVMS Alpha and GNAT for OpenVMS for Alpha and for I64.
24985 GNAT is highly compatible
24986 with HP Ada, and it should generally be straightforward to port code
24987 from the HP Ada environment to GNAT. However, there are a few language
24988 and implementation differences of which the user must be aware. These
24989 differences are discussed in this chapter. In
24990 addition, the operating environment and command structure for the
24991 compiler are different, and these differences are also discussed.
24993 For further details on these and other compatibility issues,
24994 see Appendix E of the HP publication
24995 @cite{HP Ada, Technical Overview and Comparison on HP Platforms}.
24997 Except where otherwise indicated, the description of GNAT for OpenVMS
24998 applies to both the Alpha and I64 platforms.
25000 For information on porting Ada code from GNAT on Alpha OpenVMS to GNAT on
25001 I64 OpenVMS, see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
25003 The discussion in this chapter addresses specifically the implementation
25004 of Ada 83 for HP OpenVMS Alpha Systems. In cases where the implementation
25005 of HP Ada differs between OpenVMS Alpha Systems and OpenVMS VAX Systems,
25006 GNAT always follows the Alpha implementation.
25008 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
25009 attributes are recognized, although only a subset of them can sensibly
25010 be implemented. The description of pragmas in
25011 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
25012 indicates whether or not they are applicable to non-VMS systems.
25015 * Ada Language Compatibility::
25016 * Differences in the Definition of Package System::
25017 * Language-Related Features::
25018 * The Package STANDARD::
25019 * The Package SYSTEM::
25020 * Tasking and Task-Related Features::
25021 * Pragmas and Pragma-Related Features::
25022 * Library of Predefined Units::
25024 * Main Program Definition::
25025 * Implementation-Defined Attributes::
25026 * Compiler and Run-Time Interfacing::
25027 * Program Compilation and Library Management::
25029 * Implementation Limits::
25030 * Tools and Utilities::
25033 @node Ada Language Compatibility
25034 @section Ada Language Compatibility
25037 GNAT handles Ada 95 and Ada 2005 as well as Ada 83, whereas HP Ada is only
25038 for Ada 83. Ada 95 and Ada 2005 are almost completely upwards compatible
25039 with Ada 83, and therefore Ada 83 programs will compile
25040 and run under GNAT with
25041 no changes or only minor changes. The @cite{Annotated Ada Reference Manual}
25042 provides details on specific incompatibilities.
25044 GNAT provides the switch @option{/83} on the @command{GNAT COMPILE} command,
25045 as well as the pragma @code{ADA_83}, to force the compiler to
25046 operate in Ada 83 mode. This mode does not guarantee complete
25047 conformance to Ada 83, but in practice is sufficient to
25048 eliminate most sources of incompatibilities.
25049 In particular, it eliminates the recognition of the
25050 additional Ada 95 and Ada 2005 keywords, so that their use as identifiers
25051 in Ada 83 programs is legal, and handles the cases of packages
25052 with optional bodies, and generics that instantiate unconstrained
25053 types without the use of @code{(<>)}.
25055 @node Differences in the Definition of Package System
25056 @section Differences in the Definition of Package @code{System}
25059 An Ada compiler is allowed to add
25060 implementation-dependent declarations to package @code{System}.
25062 GNAT does not take advantage of this permission, and the version of
25063 @code{System} provided by GNAT exactly matches that defined in the Ada
25066 However, HP Ada adds an extensive set of declarations to package
25068 as fully documented in the HP Ada manuals. To minimize changes required
25069 for programs that make use of these extensions, GNAT provides the pragma
25070 @code{Extend_System} for extending the definition of package System. By using:
25071 @cindex pragma @code{Extend_System}
25072 @cindex @code{Extend_System} pragma
25074 @smallexample @c ada
25077 pragma Extend_System (Aux_DEC);
25083 the set of definitions in @code{System} is extended to include those in
25084 package @code{System.Aux_DEC}.
25085 @cindex @code{System.Aux_DEC} package
25086 @cindex @code{Aux_DEC} package (child of @code{System})
25087 These definitions are incorporated directly into package @code{System},
25088 as though they had been declared there. For a
25089 list of the declarations added, see the spec of this package,
25090 which can be found in the file @file{s-auxdec.ads} in the GNAT library.
25091 @cindex @file{s-auxdec.ads} file
25092 The pragma @code{Extend_System} is a configuration pragma, which means that
25093 it can be placed in the file @file{gnat.adc}, so that it will automatically
25094 apply to all subsequent compilations. See @ref{Configuration Pragmas},
25095 for further details.
25097 An alternative approach that avoids the use of the non-standard
25098 @code{Extend_System} pragma is to add a context clause to the unit that
25099 references these facilities:
25101 @smallexample @c ada
25103 with System.Aux_DEC;
25104 use System.Aux_DEC;
25109 The effect is not quite semantically identical to incorporating
25110 the declarations directly into package @code{System},
25111 but most programs will not notice a difference
25112 unless they use prefix notation (e.g.@: @code{System.Integer_8})
25113 to reference the entities directly in package @code{System}.
25114 For units containing such references,
25115 the prefixes must either be removed, or the pragma @code{Extend_System}
25118 @node Language-Related Features
25119 @section Language-Related Features
25122 The following sections highlight differences in types,
25123 representations of types, operations, alignment, and
25127 * Integer Types and Representations::
25128 * Floating-Point Types and Representations::
25129 * Pragmas Float_Representation and Long_Float::
25130 * Fixed-Point Types and Representations::
25131 * Record and Array Component Alignment::
25132 * Address Clauses::
25133 * Other Representation Clauses::
25136 @node Integer Types and Representations
25137 @subsection Integer Types and Representations
25140 The set of predefined integer types is identical in HP Ada and GNAT.
25141 Furthermore the representation of these integer types is also identical,
25142 including the capability of size clauses forcing biased representation.
25145 HP Ada for OpenVMS Alpha systems has defined the
25146 following additional integer types in package @code{System}:
25163 @code{LARGEST_INTEGER}
25167 In GNAT, the first four of these types may be obtained from the
25168 standard Ada package @code{Interfaces}.
25169 Alternatively, by use of the pragma @code{Extend_System}, identical
25170 declarations can be referenced directly in package @code{System}.
25171 On both GNAT and HP Ada, the maximum integer size is 64 bits.
25173 @node Floating-Point Types and Representations
25174 @subsection Floating-Point Types and Representations
25175 @cindex Floating-Point types
25178 The set of predefined floating-point types is identical in HP Ada and GNAT.
25179 Furthermore the representation of these floating-point
25180 types is also identical. One important difference is that the default
25181 representation for HP Ada is @code{VAX_Float}, but the default representation
25184 Specific types may be declared to be @code{VAX_Float} or IEEE, using the
25185 pragma @code{Float_Representation} as described in the HP Ada
25187 For example, the declarations:
25189 @smallexample @c ada
25191 type F_Float is digits 6;
25192 pragma Float_Representation (VAX_Float, F_Float);
25197 declares a type @code{F_Float} that will be represented in @code{VAX_Float}
25199 This set of declarations actually appears in @code{System.Aux_DEC},
25201 the full set of additional floating-point declarations provided in
25202 the HP Ada version of package @code{System}.
25203 This and similar declarations may be accessed in a user program
25204 by using pragma @code{Extend_System}. The use of this
25205 pragma, and the related pragma @code{Long_Float} is described in further
25206 detail in the following section.
25208 @node Pragmas Float_Representation and Long_Float
25209 @subsection Pragmas @code{Float_Representation} and @code{Long_Float}
25212 HP Ada provides the pragma @code{Float_Representation}, which
25213 acts as a program library switch to allow control over
25214 the internal representation chosen for the predefined
25215 floating-point types declared in the package @code{Standard}.
25216 The format of this pragma is as follows:
25218 @smallexample @c ada
25220 pragma Float_Representation(VAX_Float | IEEE_Float);
25225 This pragma controls the representation of floating-point
25230 @code{VAX_Float} specifies that floating-point
25231 types are represented by default with the VAX system hardware types
25232 @code{F-floating}, @code{D-floating}, @code{G-floating}.
25233 Note that the @code{H-floating}
25234 type was available only on VAX systems, and is not available
25235 in either HP Ada or GNAT.
25238 @code{IEEE_Float} specifies that floating-point
25239 types are represented by default with the IEEE single and
25240 double floating-point types.
25244 GNAT provides an identical implementation of the pragma
25245 @code{Float_Representation}, except that it functions as a
25246 configuration pragma. Note that the
25247 notion of configuration pragma corresponds closely to the
25248 HP Ada notion of a program library switch.
25250 When no pragma is used in GNAT, the default is @code{IEEE_Float},
25252 from HP Ada 83, where the default is @code{VAX_Float}. In addition, the
25253 predefined libraries in GNAT are built using @code{IEEE_Float}, so it is not
25254 advisable to change the format of numbers passed to standard library
25255 routines, and if necessary explicit type conversions may be needed.
25257 The use of @code{IEEE_Float} is recommended in GNAT since it is more
25258 efficient, and (given that it conforms to an international standard)
25259 potentially more portable.
25260 The situation in which @code{VAX_Float} may be useful is in interfacing
25261 to existing code and data that expect the use of @code{VAX_Float}.
25262 In such a situation use the predefined @code{VAX_Float}
25263 types in package @code{System}, as extended by
25264 @code{Extend_System}. For example, use @code{System.F_Float}
25265 to specify the 32-bit @code{F-Float} format.
25268 On OpenVMS systems, HP Ada provides the pragma @code{Long_Float}
25269 to allow control over the internal representation chosen
25270 for the predefined type @code{Long_Float} and for floating-point
25271 type declarations with digits specified in the range 7 .. 15.
25272 The format of this pragma is as follows:
25274 @smallexample @c ada
25276 pragma Long_Float (D_FLOAT | G_FLOAT);
25280 @node Fixed-Point Types and Representations
25281 @subsection Fixed-Point Types and Representations
25284 On HP Ada for OpenVMS Alpha systems, rounding is
25285 away from zero for both positive and negative numbers.
25286 Therefore, @code{+0.5} rounds to @code{1},
25287 and @code{-0.5} rounds to @code{-1}.
25289 On GNAT the results of operations
25290 on fixed-point types are in accordance with the Ada
25291 rules. In particular, results of operations on decimal
25292 fixed-point types are truncated.
25294 @node Record and Array Component Alignment
25295 @subsection Record and Array Component Alignment
25298 On HP Ada for OpenVMS Alpha, all non-composite components
25299 are aligned on natural boundaries. For example, 1-byte
25300 components are aligned on byte boundaries, 2-byte
25301 components on 2-byte boundaries, 4-byte components on 4-byte
25302 byte boundaries, and so on. The OpenVMS Alpha hardware
25303 runs more efficiently with naturally aligned data.
25305 On GNAT, alignment rules are compatible
25306 with HP Ada for OpenVMS Alpha.
25308 @node Address Clauses
25309 @subsection Address Clauses
25312 In HP Ada and GNAT, address clauses are supported for
25313 objects and imported subprograms.
25314 The predefined type @code{System.Address} is a private type
25315 in both compilers on Alpha OpenVMS, with the same representation
25316 (it is simply a machine pointer). Addition, subtraction, and comparison
25317 operations are available in the standard Ada package
25318 @code{System.Storage_Elements}, or in package @code{System}
25319 if it is extended to include @code{System.Aux_DEC} using a
25320 pragma @code{Extend_System} as previously described.
25322 Note that code that @code{with}'s both this extended package @code{System}
25323 and the package @code{System.Storage_Elements} should not @code{use}
25324 both packages, or ambiguities will result. In general it is better
25325 not to mix these two sets of facilities. The Ada package was
25326 designed specifically to provide the kind of features that HP Ada
25327 adds directly to package @code{System}.
25329 The type @code{System.Address} is a 64-bit integer type in GNAT for
25330 I64 OpenVMS. For more information,
25331 see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
25333 GNAT is compatible with HP Ada in its handling of address
25334 clauses, except for some limitations in
25335 the form of address clauses for composite objects with
25336 initialization. Such address clauses are easily replaced
25337 by the use of an explicitly-defined constant as described
25338 in the Ada Reference Manual (13.1(22)). For example, the sequence
25341 @smallexample @c ada
25343 X, Y : Integer := Init_Func;
25344 Q : String (X .. Y) := "abc";
25346 for Q'Address use Compute_Address;
25351 will be rejected by GNAT, since the address cannot be computed at the time
25352 that @code{Q} is declared. To achieve the intended effect, write instead:
25354 @smallexample @c ada
25357 X, Y : Integer := Init_Func;
25358 Q_Address : constant Address := Compute_Address;
25359 Q : String (X .. Y) := "abc";
25361 for Q'Address use Q_Address;
25367 which will be accepted by GNAT (and other Ada compilers), and is also
25368 compatible with Ada 83. A fuller description of the restrictions
25369 on address specifications is found in @ref{Top, GNAT Reference Manual,
25370 About This Guide, gnat_rm, GNAT Reference Manual}.
25372 @node Other Representation Clauses
25373 @subsection Other Representation Clauses
25376 GNAT implements in a compatible manner all the representation
25377 clauses supported by HP Ada. In addition, GNAT
25378 implements the representation clause forms that were introduced in Ada 95,
25379 including @code{COMPONENT_SIZE} and @code{SIZE} clauses for objects.
25381 @node The Package STANDARD
25382 @section The Package @code{STANDARD}
25385 The package @code{STANDARD}, as implemented by HP Ada, is fully
25386 described in the @cite{Ada Reference Manual} and in the
25387 @cite{HP Ada Language Reference Manual}. As implemented by GNAT, the
25388 package @code{STANDARD} is described in the @cite{Ada Reference Manual}.
25390 In addition, HP Ada supports the Latin-1 character set in
25391 the type @code{CHARACTER}. GNAT supports the Latin-1 character set
25392 in the type @code{CHARACTER} and also Unicode (ISO 10646 BMP) in
25393 the type @code{WIDE_CHARACTER}.
25395 The floating-point types supported by GNAT are those
25396 supported by HP Ada, but the defaults are different, and are controlled by
25397 pragmas. See @ref{Floating-Point Types and Representations}, for details.
25399 @node The Package SYSTEM
25400 @section The Package @code{SYSTEM}
25403 HP Ada provides a specific version of the package
25404 @code{SYSTEM} for each platform on which the language is implemented.
25405 For the complete spec of the package @code{SYSTEM}, see
25406 Appendix F of the @cite{HP Ada Language Reference Manual}.
25408 On HP Ada, the package @code{SYSTEM} includes the following conversion
25411 @item @code{TO_ADDRESS(INTEGER)}
25413 @item @code{TO_ADDRESS(UNSIGNED_LONGWORD)}
25415 @item @code{TO_ADDRESS(}@i{universal_integer}@code{)}
25417 @item @code{TO_INTEGER(ADDRESS)}
25419 @item @code{TO_UNSIGNED_LONGWORD(ADDRESS)}
25421 @item Function @code{IMPORT_VALUE return UNSIGNED_LONGWORD} and the
25422 functions @code{IMPORT_ADDRESS} and @code{IMPORT_LARGEST_VALUE}
25426 By default, GNAT supplies a version of @code{SYSTEM} that matches
25427 the definition given in the @cite{Ada Reference Manual}.
25429 is a subset of the HP system definitions, which is as
25430 close as possible to the original definitions. The only difference
25431 is that the definition of @code{SYSTEM_NAME} is different:
25433 @smallexample @c ada
25435 type Name is (SYSTEM_NAME_GNAT);
25436 System_Name : constant Name := SYSTEM_NAME_GNAT;
25441 Also, GNAT adds the Ada declarations for
25442 @code{BIT_ORDER} and @code{DEFAULT_BIT_ORDER}.
25444 However, the use of the following pragma causes GNAT
25445 to extend the definition of package @code{SYSTEM} so that it
25446 encompasses the full set of HP-specific extensions,
25447 including the functions listed above:
25449 @smallexample @c ada
25451 pragma Extend_System (Aux_DEC);
25456 The pragma @code{Extend_System} is a configuration pragma that
25457 is most conveniently placed in the @file{gnat.adc} file. @xref{Pragma
25458 Extend_System,,, gnat_rm, GNAT Reference Manual}, for further details.
25460 HP Ada does not allow the recompilation of the package
25461 @code{SYSTEM}. Instead HP Ada provides several pragmas
25462 (@code{SYSTEM_NAME}, @code{STORAGE_UNIT}, and @code{MEMORY_SIZE})
25463 to modify values in the package @code{SYSTEM}.
25464 On OpenVMS Alpha systems, the pragma
25465 @code{SYSTEM_NAME} takes the enumeration literal @code{OPENVMS_AXP} as
25466 its single argument.
25468 GNAT does permit the recompilation of package @code{SYSTEM} using
25469 the special switch @option{-gnatg}, and this switch can be used if
25470 it is necessary to modify the definitions in @code{SYSTEM}. GNAT does
25471 not permit the specification of @code{SYSTEM_NAME}, @code{STORAGE_UNIT}
25472 or @code{MEMORY_SIZE} by any other means.
25474 On GNAT systems, the pragma @code{SYSTEM_NAME} takes the
25475 enumeration literal @code{SYSTEM_NAME_GNAT}.
25477 The definitions provided by the use of
25479 @smallexample @c ada
25480 pragma Extend_System (AUX_Dec);
25484 are virtually identical to those provided by the HP Ada 83 package
25485 @code{SYSTEM}. One important difference is that the name of the
25487 function for type @code{UNSIGNED_LONGWORD} is changed to
25488 @code{TO_ADDRESS_LONG}.
25489 @xref{Address Clauses,,, gnat_rm, GNAT Reference Manual}, for a
25490 discussion of why this change was necessary.
25493 The version of @code{TO_ADDRESS} taking a @i{universal_integer} argument
25495 an extension to Ada 83 not strictly compatible with the reference manual.
25496 GNAT, in order to be exactly compatible with the standard,
25497 does not provide this capability. In HP Ada 83, the
25498 point of this definition is to deal with a call like:
25500 @smallexample @c ada
25501 TO_ADDRESS (16#12777#);
25505 Normally, according to Ada 83 semantics, one would expect this to be
25506 ambiguous, since it matches both the @code{INTEGER} and
25507 @code{UNSIGNED_LONGWORD} forms of @code{TO_ADDRESS}.
25508 However, in HP Ada 83, there is no ambiguity, since the
25509 definition using @i{universal_integer} takes precedence.
25511 In GNAT, since the version with @i{universal_integer} cannot be supplied,
25513 not possible to be 100% compatible. Since there are many programs using
25514 numeric constants for the argument to @code{TO_ADDRESS}, the decision in
25516 to change the name of the function in the @code{UNSIGNED_LONGWORD} case,
25517 so the declarations provided in the GNAT version of @code{AUX_Dec} are:
25519 @smallexample @c ada
25520 function To_Address (X : Integer) return Address;
25521 pragma Pure_Function (To_Address);
25523 function To_Address_Long (X : Unsigned_Longword) return Address;
25524 pragma Pure_Function (To_Address_Long);
25528 This means that programs using @code{TO_ADDRESS} for
25529 @code{UNSIGNED_LONGWORD} must change the name to @code{TO_ADDRESS_LONG}.
25531 @node Tasking and Task-Related Features
25532 @section Tasking and Task-Related Features
25535 This section compares the treatment of tasking in GNAT
25536 and in HP Ada for OpenVMS Alpha.
25537 The GNAT description applies to both Alpha and I64 OpenVMS.
25538 For detailed information on tasking in
25539 HP Ada, see the @cite{HP Ada Language Reference Manual} and the
25540 relevant run-time reference manual.
25543 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
25544 * Assigning Task IDs::
25545 * Task IDs and Delays::
25546 * Task-Related Pragmas::
25547 * Scheduling and Task Priority::
25549 * External Interrupts::
25552 @node Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
25553 @subsection Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
25556 On OpenVMS Alpha systems, each Ada task (except a passive
25557 task) is implemented as a single stream of execution
25558 that is created and managed by the kernel. On these
25559 systems, HP Ada tasking support is based on DECthreads,
25560 an implementation of the POSIX standard for threads.
25562 Also, on OpenVMS Alpha systems, HP Ada tasks and foreign
25563 code that calls DECthreads routines can be used together.
25564 The interaction between Ada tasks and DECthreads routines
25565 can have some benefits. For example when on OpenVMS Alpha,
25566 HP Ada can call C code that is already threaded.
25568 GNAT uses the facilities of DECthreads,
25569 and Ada tasks are mapped to threads.
25571 @node Assigning Task IDs
25572 @subsection Assigning Task IDs
25575 The HP Ada Run-Time Library always assigns @code{%TASK 1} to
25576 the environment task that executes the main program. On
25577 OpenVMS Alpha systems, @code{%TASK 0} is often used for tasks
25578 that have been created but are not yet activated.
25580 On OpenVMS Alpha systems, task IDs are assigned at
25581 activation. On GNAT systems, task IDs are also assigned at
25582 task creation but do not have the same form or values as
25583 task ID values in HP Ada. There is no null task, and the
25584 environment task does not have a specific task ID value.
25586 @node Task IDs and Delays
25587 @subsection Task IDs and Delays
25590 On OpenVMS Alpha systems, tasking delays are implemented
25591 using Timer System Services. The Task ID is used for the
25592 identification of the timer request (the @code{REQIDT} parameter).
25593 If Timers are used in the application take care not to use
25594 @code{0} for the identification, because cancelling such a timer
25595 will cancel all timers and may lead to unpredictable results.
25597 @node Task-Related Pragmas
25598 @subsection Task-Related Pragmas
25601 Ada supplies the pragma @code{TASK_STORAGE}, which allows
25602 specification of the size of the guard area for a task
25603 stack. (The guard area forms an area of memory that has no
25604 read or write access and thus helps in the detection of
25605 stack overflow.) On OpenVMS Alpha systems, if the pragma
25606 @code{TASK_STORAGE} specifies a value of zero, a minimal guard
25607 area is created. In the absence of a pragma @code{TASK_STORAGE},
25608 a default guard area is created.
25610 GNAT supplies the following task-related pragmas:
25613 @item @code{TASK_INFO}
25615 This pragma appears within a task definition and
25616 applies to the task in which it appears. The argument
25617 must be of type @code{SYSTEM.TASK_INFO.TASK_INFO_TYPE}.
25619 @item @code{TASK_STORAGE}
25621 GNAT implements pragma @code{TASK_STORAGE} in the same way as HP Ada.
25622 Both HP Ada and GNAT supply the pragmas @code{PASSIVE},
25623 @code{SUPPRESS}, and @code{VOLATILE}.
25625 @node Scheduling and Task Priority
25626 @subsection Scheduling and Task Priority
25629 HP Ada implements the Ada language requirement that
25630 when two tasks are eligible for execution and they have
25631 different priorities, the lower priority task does not
25632 execute while the higher priority task is waiting. The HP
25633 Ada Run-Time Library keeps a task running until either the
25634 task is suspended or a higher priority task becomes ready.
25636 On OpenVMS Alpha systems, the default strategy is round-
25637 robin with preemption. Tasks of equal priority take turns
25638 at the processor. A task is run for a certain period of
25639 time and then placed at the tail of the ready queue for
25640 its priority level.
25642 HP Ada provides the implementation-defined pragma @code{TIME_SLICE},
25643 which can be used to enable or disable round-robin
25644 scheduling of tasks with the same priority.
25645 See the relevant HP Ada run-time reference manual for
25646 information on using the pragmas to control HP Ada task
25649 GNAT follows the scheduling rules of Annex D (Real-Time
25650 Annex) of the @cite{Ada Reference Manual}. In general, this
25651 scheduling strategy is fully compatible with HP Ada
25652 although it provides some additional constraints (as
25653 fully documented in Annex D).
25654 GNAT implements time slicing control in a manner compatible with
25655 HP Ada 83, by means of the pragma @code{Time_Slice}, whose semantics
25656 are identical to the HP Ada 83 pragma of the same name.
25657 Note that it is not possible to mix GNAT tasking and
25658 HP Ada 83 tasking in the same program, since the two run-time
25659 libraries are not compatible.
25661 @node The Task Stack
25662 @subsection The Task Stack
25665 In HP Ada, a task stack is allocated each time a
25666 non-passive task is activated. As soon as the task is
25667 terminated, the storage for the task stack is deallocated.
25668 If you specify a size of zero (bytes) with @code{T'STORAGE_SIZE},
25669 a default stack size is used. Also, regardless of the size
25670 specified, some additional space is allocated for task
25671 management purposes. On OpenVMS Alpha systems, at least
25672 one page is allocated.
25674 GNAT handles task stacks in a similar manner. In accordance with
25675 the Ada rules, it provides the pragma @code{STORAGE_SIZE} as
25676 an alternative method for controlling the task stack size.
25677 The specification of the attribute @code{T'STORAGE_SIZE} is also
25678 supported in a manner compatible with HP Ada.
25680 @node External Interrupts
25681 @subsection External Interrupts
25684 On HP Ada, external interrupts can be associated with task entries.
25685 GNAT is compatible with HP Ada in its handling of external interrupts.
25687 @node Pragmas and Pragma-Related Features
25688 @section Pragmas and Pragma-Related Features
25691 Both HP Ada and GNAT supply all language-defined pragmas
25692 as specified by the Ada 83 standard. GNAT also supplies all
25693 language-defined pragmas introduced by Ada 95 and Ada 2005.
25694 In addition, GNAT implements the implementation-defined pragmas
25698 @item @code{AST_ENTRY}
25700 @item @code{COMMON_OBJECT}
25702 @item @code{COMPONENT_ALIGNMENT}
25704 @item @code{EXPORT_EXCEPTION}
25706 @item @code{EXPORT_FUNCTION}
25708 @item @code{EXPORT_OBJECT}
25710 @item @code{EXPORT_PROCEDURE}
25712 @item @code{EXPORT_VALUED_PROCEDURE}
25714 @item @code{FLOAT_REPRESENTATION}
25718 @item @code{IMPORT_EXCEPTION}
25720 @item @code{IMPORT_FUNCTION}
25722 @item @code{IMPORT_OBJECT}
25724 @item @code{IMPORT_PROCEDURE}
25726 @item @code{IMPORT_VALUED_PROCEDURE}
25728 @item @code{INLINE_GENERIC}
25730 @item @code{INTERFACE_NAME}
25732 @item @code{LONG_FLOAT}
25734 @item @code{MAIN_STORAGE}
25736 @item @code{PASSIVE}
25738 @item @code{PSECT_OBJECT}
25740 @item @code{SHARE_GENERIC}
25742 @item @code{SUPPRESS_ALL}
25744 @item @code{TASK_STORAGE}
25746 @item @code{TIME_SLICE}
25752 These pragmas are all fully implemented, with the exception of @code{TITLE},
25753 @code{PASSIVE}, and @code{SHARE_GENERIC}, which are
25754 recognized, but which have no
25755 effect in GNAT. The effect of @code{PASSIVE} may be obtained by the
25756 use of Ada protected objects. In GNAT, all generics are inlined.
25758 Unlike HP Ada, the GNAT ``@code{EXPORT_}@i{subprogram}'' pragmas require
25759 a separate subprogram specification which must appear before the
25762 GNAT also supplies a number of implementation-defined pragmas as follows:
25764 @item @code{ABORT_DEFER}
25766 @item @code{ADA_83}
25768 @item @code{ADA_95}
25770 @item @code{ADA_05}
25772 @item @code{ANNOTATE}
25774 @item @code{ASSERT}
25776 @item @code{C_PASS_BY_COPY}
25778 @item @code{CPP_CLASS}
25780 @item @code{CPP_CONSTRUCTOR}
25782 @item @code{CPP_DESTRUCTOR}
25786 @item @code{EXTEND_SYSTEM}
25788 @item @code{LINKER_ALIAS}
25790 @item @code{LINKER_SECTION}
25792 @item @code{MACHINE_ATTRIBUTE}
25794 @item @code{NO_RETURN}
25796 @item @code{PURE_FUNCTION}
25798 @item @code{SOURCE_FILE_NAME}
25800 @item @code{SOURCE_REFERENCE}
25802 @item @code{TASK_INFO}
25804 @item @code{UNCHECKED_UNION}
25806 @item @code{UNIMPLEMENTED_UNIT}
25808 @item @code{UNIVERSAL_DATA}
25810 @item @code{UNSUPPRESS}
25812 @item @code{WARNINGS}
25814 @item @code{WEAK_EXTERNAL}
25818 For full details on these GNAT implementation-defined pragmas,
25819 see @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference
25823 * Restrictions on the Pragma INLINE::
25824 * Restrictions on the Pragma INTERFACE::
25825 * Restrictions on the Pragma SYSTEM_NAME::
25828 @node Restrictions on the Pragma INLINE
25829 @subsection Restrictions on Pragma @code{INLINE}
25832 HP Ada enforces the following restrictions on the pragma @code{INLINE}:
25834 @item Parameters cannot have a task type.
25836 @item Function results cannot be task types, unconstrained
25837 array types, or unconstrained types with discriminants.
25839 @item Bodies cannot declare the following:
25841 @item Subprogram body or stub (imported subprogram is allowed)
25845 @item Generic declarations
25847 @item Instantiations
25851 @item Access types (types derived from access types allowed)
25853 @item Array or record types
25855 @item Dependent tasks
25857 @item Direct recursive calls of subprogram or containing
25858 subprogram, directly or via a renaming
25864 In GNAT, the only restriction on pragma @code{INLINE} is that the
25865 body must occur before the call if both are in the same
25866 unit, and the size must be appropriately small. There are
25867 no other specific restrictions which cause subprograms to
25868 be incapable of being inlined.
25870 @node Restrictions on the Pragma INTERFACE
25871 @subsection Restrictions on Pragma @code{INTERFACE}
25874 The following restrictions on pragma @code{INTERFACE}
25875 are enforced by both HP Ada and GNAT:
25877 @item Languages accepted: Ada, Bliss, C, Fortran, Default.
25878 Default is the default on OpenVMS Alpha systems.
25880 @item Parameter passing: Language specifies default
25881 mechanisms but can be overridden with an @code{EXPORT} pragma.
25884 @item Ada: Use internal Ada rules.
25886 @item Bliss, C: Parameters must be mode @code{in}; cannot be
25887 record or task type. Result cannot be a string, an
25888 array, or a record.
25890 @item Fortran: Parameters cannot have a task type. Result cannot
25891 be a string, an array, or a record.
25896 GNAT is entirely upwards compatible with HP Ada, and in addition allows
25897 record parameters for all languages.
25899 @node Restrictions on the Pragma SYSTEM_NAME
25900 @subsection Restrictions on Pragma @code{SYSTEM_NAME}
25903 For HP Ada for OpenVMS Alpha, the enumeration literal
25904 for the type @code{NAME} is @code{OPENVMS_AXP}.
25905 In GNAT, the enumeration
25906 literal for the type @code{NAME} is @code{SYSTEM_NAME_GNAT}.
25908 @node Library of Predefined Units
25909 @section Library of Predefined Units
25912 A library of predefined units is provided as part of the
25913 HP Ada and GNAT implementations. HP Ada does not provide
25914 the package @code{MACHINE_CODE} but instead recommends importing
25917 The GNAT versions of the HP Ada Run-Time Library (@code{ADA$PREDEFINED:})
25918 units are taken from the OpenVMS Alpha version, not the OpenVMS VAX
25920 The HP Ada Predefined Library units are modified to remove post-Ada 83
25921 incompatibilities and to make them interoperable with GNAT
25922 (@pxref{Changes to DECLIB}, for details).
25923 The units are located in the @file{DECLIB} directory.
25925 The GNAT RTL is contained in
25926 the @file{ADALIB} directory, and
25927 the default search path is set up to find @code{DECLIB} units in preference
25928 to @code{ADALIB} units with the same name (@code{TEXT_IO},
25929 @code{SEQUENTIAL_IO}, and @code{DIRECT_IO}, for example).
25932 * Changes to DECLIB::
25935 @node Changes to DECLIB
25936 @subsection Changes to @code{DECLIB}
25939 The changes made to the HP Ada predefined library for GNAT and post-Ada 83
25940 compatibility are minor and include the following:
25943 @item Adjusting the location of pragmas and record representation
25944 clauses to obey Ada 95 (and thus Ada 2005) rules
25946 @item Adding the proper notation to generic formal parameters
25947 that take unconstrained types in instantiation
25949 @item Adding pragma @code{ELABORATE_BODY} to package specs
25950 that have package bodies not otherwise allowed
25952 @item Replacing occurrences of the identifier ``@code{PROTECTED}'' by
25953 ``@code{PROTECTD}''.
25954 Currently these are found only in the @code{STARLET} package spec.
25956 @item Changing @code{SYSTEM.ADDRESS} to @code{SYSTEM.SHORT_ADDRESS}
25957 where the address size is constrained to 32 bits.
25961 None of the above changes is visible to users.
25967 On OpenVMS Alpha, HP Ada provides the following strongly-typed bindings:
25970 @item Command Language Interpreter (CLI interface)
25972 @item DECtalk Run-Time Library (DTK interface)
25974 @item Librarian utility routines (LBR interface)
25976 @item General Purpose Run-Time Library (LIB interface)
25978 @item Math Run-Time Library (MTH interface)
25980 @item National Character Set Run-Time Library (NCS interface)
25982 @item Compiled Code Support Run-Time Library (OTS interface)
25984 @item Parallel Processing Run-Time Library (PPL interface)
25986 @item Screen Management Run-Time Library (SMG interface)
25988 @item Sort Run-Time Library (SOR interface)
25990 @item String Run-Time Library (STR interface)
25992 @item STARLET System Library
25995 @item X Window System Version 11R4 and 11R5 (X, XLIB interface)
25997 @item X Windows Toolkit (XT interface)
25999 @item X/Motif Version 1.1.3 and 1.2 (XM interface)
26003 GNAT provides implementations of these HP bindings in the @code{DECLIB}
26004 directory, on both the Alpha and I64 OpenVMS platforms.
26006 The X/Motif bindings used to build @code{DECLIB} are whatever versions are
26008 HP Ada @file{ADA$PREDEFINED} directory with extension @file{.ADC}.
26009 A pragma @code{Linker_Options} has been added to packages @code{Xm},
26010 @code{Xt}, and @code{X_Lib}
26011 causing the default X/Motif sharable image libraries to be linked in. This
26012 is done via options files named @file{xm.opt}, @file{xt.opt}, and
26013 @file{x_lib.opt} (also located in the @file{DECLIB} directory).
26015 It may be necessary to edit these options files to update or correct the
26016 library names if, for example, the newer X/Motif bindings from
26017 @file{ADA$EXAMPLES}
26018 had been (previous to installing GNAT) copied and renamed to supersede the
26019 default @file{ADA$PREDEFINED} versions.
26022 * Shared Libraries and Options Files::
26023 * Interfaces to C::
26026 @node Shared Libraries and Options Files
26027 @subsection Shared Libraries and Options Files
26030 When using the HP Ada
26031 predefined X and Motif bindings, the linking with their sharable images is
26032 done automatically by @command{GNAT LINK}.
26033 When using other X and Motif bindings, you need
26034 to add the corresponding sharable images to the command line for
26035 @code{GNAT LINK}. When linking with shared libraries, or with
26036 @file{.OPT} files, you must
26037 also add them to the command line for @command{GNAT LINK}.
26039 A shared library to be used with GNAT is built in the same way as other
26040 libraries under VMS. The VMS Link command can be used in standard fashion.
26042 @node Interfaces to C
26043 @subsection Interfaces to C
26047 provides the following Ada types and operations:
26050 @item C types package (@code{C_TYPES})
26052 @item C strings (@code{C_TYPES.NULL_TERMINATED})
26054 @item Other_types (@code{SHORT_INT})
26058 Interfacing to C with GNAT, you can use the above approach
26059 described for HP Ada or the facilities of Annex B of
26060 the @cite{Ada Reference Manual} (packages @code{INTERFACES.C},
26061 @code{INTERFACES.C.STRINGS} and @code{INTERFACES.C.POINTERS}). For more
26062 information, see @ref{Interfacing to C,,, gnat_rm, GNAT Reference Manual}.
26064 The @option{-gnatF} qualifier forces default and explicit
26065 @code{External_Name} parameters in pragmas @code{Import} and @code{Export}
26066 to be uppercased for compatibility with the default behavior
26067 of HP C. The qualifier has no effect on @code{Link_Name} parameters.
26069 @node Main Program Definition
26070 @section Main Program Definition
26073 The following section discusses differences in the
26074 definition of main programs on HP Ada and GNAT.
26075 On HP Ada, main programs are defined to meet the
26076 following conditions:
26078 @item Procedure with no formal parameters (returns @code{0} upon
26081 @item Procedure with no formal parameters (returns @code{42} when
26082 an unhandled exception is raised)
26084 @item Function with no formal parameters whose returned value
26085 is of a discrete type
26087 @item Procedure with one @code{out} formal of a discrete type for
26088 which a specification of pragma @code{EXPORT_VALUED_PROCEDURE} is given.
26093 When declared with the pragma @code{EXPORT_VALUED_PROCEDURE},
26094 a main function or main procedure returns a discrete
26095 value whose size is less than 64 bits (32 on VAX systems),
26096 the value is zero- or sign-extended as appropriate.
26097 On GNAT, main programs are defined as follows:
26099 @item Must be a non-generic, parameterless subprogram that
26100 is either a procedure or function returning an Ada
26101 @code{STANDARD.INTEGER} (the predefined type)
26103 @item Cannot be a generic subprogram or an instantiation of a
26107 @node Implementation-Defined Attributes
26108 @section Implementation-Defined Attributes
26111 GNAT provides all HP Ada implementation-defined
26114 @node Compiler and Run-Time Interfacing
26115 @section Compiler and Run-Time Interfacing
26118 HP Ada provides the following qualifiers to pass options to the linker
26121 @item @option{/WAIT} and @option{/SUBMIT}
26123 @item @option{/COMMAND}
26125 @item @option{/@r{[}NO@r{]}MAP}
26127 @item @option{/OUTPUT=@var{file-spec}}
26129 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
26133 To pass options to the linker, GNAT provides the following
26137 @item @option{/EXECUTABLE=@var{exec-name}}
26139 @item @option{/VERBOSE}
26141 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
26145 For more information on these switches, see
26146 @ref{Switches for gnatlink}.
26147 In HP Ada, the command-line switch @option{/OPTIMIZE} is available
26148 to control optimization. HP Ada also supplies the
26151 @item @code{OPTIMIZE}
26153 @item @code{INLINE}
26155 @item @code{INLINE_GENERIC}
26157 @item @code{SUPPRESS_ALL}
26159 @item @code{PASSIVE}
26163 In GNAT, optimization is controlled strictly by command
26164 line parameters, as described in the corresponding section of this guide.
26165 The HP pragmas for control of optimization are
26166 recognized but ignored.
26168 Note that in GNAT, the default is optimization off, whereas in HP Ada
26169 the default is that optimization is turned on.
26171 @node Program Compilation and Library Management
26172 @section Program Compilation and Library Management
26175 HP Ada and GNAT provide a comparable set of commands to
26176 build programs. HP Ada also provides a program library,
26177 which is a concept that does not exist on GNAT. Instead,
26178 GNAT provides directories of sources that are compiled as
26181 The following table summarizes
26182 the HP Ada commands and provides
26183 equivalent GNAT commands. In this table, some GNAT
26184 equivalents reflect the fact that GNAT does not use the
26185 concept of a program library. Instead, it uses a model
26186 in which collections of source and object files are used
26187 in a manner consistent with other languages like C and
26188 Fortran. Therefore, standard system file commands are used
26189 to manipulate these elements. Those GNAT commands are marked with
26191 Note that, unlike HP Ada, none of the GNAT commands accepts wild cards.
26194 @multitable @columnfractions .35 .65
26196 @item @emph{HP Ada Command}
26197 @tab @emph{GNAT Equivalent / Description}
26199 @item @command{ADA}
26200 @tab @command{GNAT COMPILE}@*
26201 Invokes the compiler to compile one or more Ada source files.
26203 @item @command{ACS ATTACH}@*
26204 @tab [No equivalent]@*
26205 Switches control of terminal from current process running the program
26208 @item @command{ACS CHECK}
26209 @tab @command{GNAT MAKE /DEPENDENCY_LIST}@*
26210 Forms the execution closure of one
26211 or more compiled units and checks completeness and currency.
26213 @item @command{ACS COMPILE}
26214 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
26215 Forms the execution closure of one or
26216 more specified units, checks completeness and currency,
26217 identifies units that have revised source files, compiles same,
26218 and recompiles units that are or will become obsolete.
26219 Also completes incomplete generic instantiations.
26221 @item @command{ACS COPY FOREIGN}
26223 Copies a foreign object file into the program library as a
26226 @item @command{ACS COPY UNIT}
26228 Copies a compiled unit from one program library to another.
26230 @item @command{ACS CREATE LIBRARY}
26231 @tab Create /directory (*)@*
26232 Creates a program library.
26234 @item @command{ACS CREATE SUBLIBRARY}
26235 @tab Create /directory (*)@*
26236 Creates a program sublibrary.
26238 @item @command{ACS DELETE LIBRARY}
26240 Deletes a program library and its contents.
26242 @item @command{ACS DELETE SUBLIBRARY}
26244 Deletes a program sublibrary and its contents.
26246 @item @command{ACS DELETE UNIT}
26247 @tab Delete file (*)@*
26248 On OpenVMS systems, deletes one or more compiled units from
26249 the current program library.
26251 @item @command{ACS DIRECTORY}
26252 @tab Directory (*)@*
26253 On OpenVMS systems, lists units contained in the current
26256 @item @command{ACS ENTER FOREIGN}
26258 Allows the import of a foreign body as an Ada library
26259 spec and enters a reference to a pointer.
26261 @item @command{ACS ENTER UNIT}
26263 Enters a reference (pointer) from the current program library to
26264 a unit compiled into another program library.
26266 @item @command{ACS EXIT}
26267 @tab [No equivalent]@*
26268 Exits from the program library manager.
26270 @item @command{ACS EXPORT}
26272 Creates an object file that contains system-specific object code
26273 for one or more units. With GNAT, object files can simply be copied
26274 into the desired directory.
26276 @item @command{ACS EXTRACT SOURCE}
26278 Allows access to the copied source file for each Ada compilation unit
26280 @item @command{ACS HELP}
26281 @tab @command{HELP GNAT}@*
26282 Provides online help.
26284 @item @command{ACS LINK}
26285 @tab @command{GNAT LINK}@*
26286 Links an object file containing Ada units into an executable file.
26288 @item @command{ACS LOAD}
26290 Loads (partially compiles) Ada units into the program library.
26291 Allows loading a program from a collection of files into a library
26292 without knowing the relationship among units.
26294 @item @command{ACS MERGE}
26296 Merges into the current program library, one or more units from
26297 another library where they were modified.
26299 @item @command{ACS RECOMPILE}
26300 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
26301 Recompiles from external or copied source files any obsolete
26302 unit in the closure. Also, completes any incomplete generic
26305 @item @command{ACS REENTER}
26306 @tab @command{GNAT MAKE}@*
26307 Reenters current references to units compiled after last entered
26308 with the @command{ACS ENTER UNIT} command.
26310 @item @command{ACS SET LIBRARY}
26311 @tab Set default (*)@*
26312 Defines a program library to be the compilation context as well
26313 as the target library for compiler output and commands in general.
26315 @item @command{ACS SET PRAGMA}
26316 @tab Edit @file{gnat.adc} (*)@*
26317 Redefines specified values of the library characteristics
26318 @code{LONG_ FLOAT}, @code{MEMORY_SIZE}, @code{SYSTEM_NAME},
26319 and @code{Float_Representation}.
26321 @item @command{ACS SET SOURCE}
26322 @tab Define @code{ADA_INCLUDE_PATH} path (*)@*
26323 Defines the source file search list for the @command{ACS COMPILE} command.
26325 @item @command{ACS SHOW LIBRARY}
26326 @tab Directory (*)@*
26327 Lists information about one or more program libraries.
26329 @item @command{ACS SHOW PROGRAM}
26330 @tab [No equivalent]@*
26331 Lists information about the execution closure of one or
26332 more units in the program library.
26334 @item @command{ACS SHOW SOURCE}
26335 @tab Show logical @code{ADA_INCLUDE_PATH}@*
26336 Shows the source file search used when compiling units.
26338 @item @command{ACS SHOW VERSION}
26339 @tab Compile with @option{VERBOSE} option
26340 Displays the version number of the compiler and program library
26343 @item @command{ACS SPAWN}
26344 @tab [No equivalent]@*
26345 Creates a subprocess of the current process (same as @command{DCL SPAWN}
26348 @item @command{ACS VERIFY}
26349 @tab [No equivalent]@*
26350 Performs a series of consistency checks on a program library to
26351 determine whether the library structure and library files are in
26358 @section Input-Output
26361 On OpenVMS Alpha systems, HP Ada uses OpenVMS Record
26362 Management Services (RMS) to perform operations on
26366 HP Ada and GNAT predefine an identical set of input-
26367 output packages. To make the use of the
26368 generic @code{TEXT_IO} operations more convenient, HP Ada
26369 provides predefined library packages that instantiate the
26370 integer and floating-point operations for the predefined
26371 integer and floating-point types as shown in the following table.
26373 @multitable @columnfractions .45 .55
26374 @item @emph{Package Name} @tab Instantiation
26376 @item @code{INTEGER_TEXT_IO}
26377 @tab @code{INTEGER_IO(INTEGER)}
26379 @item @code{SHORT_INTEGER_TEXT_IO}
26380 @tab @code{INTEGER_IO(SHORT_INTEGER)}
26382 @item @code{SHORT_SHORT_INTEGER_TEXT_IO}
26383 @tab @code{INTEGER_IO(SHORT_SHORT_INTEGER)}
26385 @item @code{FLOAT_TEXT_IO}
26386 @tab @code{FLOAT_IO(FLOAT)}
26388 @item @code{LONG_FLOAT_TEXT_IO}
26389 @tab @code{FLOAT_IO(LONG_FLOAT)}
26393 The HP Ada predefined packages and their operations
26394 are implemented using OpenVMS Alpha files and input-output
26395 facilities. HP Ada supports asynchronous input-output on OpenVMS Alpha.
26396 Familiarity with the following is recommended:
26398 @item RMS file organizations and access methods
26400 @item OpenVMS file specifications and directories
26402 @item OpenVMS File Definition Language (FDL)
26406 GNAT provides I/O facilities that are completely
26407 compatible with HP Ada. The distribution includes the
26408 standard HP Ada versions of all I/O packages, operating
26409 in a manner compatible with HP Ada. In particular, the
26410 following packages are by default the HP Ada (Ada 83)
26411 versions of these packages rather than the renamings
26412 suggested in Annex J of the Ada Reference Manual:
26414 @item @code{TEXT_IO}
26416 @item @code{SEQUENTIAL_IO}
26418 @item @code{DIRECT_IO}
26422 The use of the standard child package syntax (for
26423 example, @code{ADA.TEXT_IO}) retrieves the post-Ada 83 versions of these
26425 GNAT provides HP-compatible predefined instantiations
26426 of the @code{TEXT_IO} packages, and also
26427 provides the standard predefined instantiations required
26428 by the @cite{Ada Reference Manual}.
26430 For further information on how GNAT interfaces to the file
26431 system or how I/O is implemented in programs written in
26432 mixed languages, see @ref{Implementation of the Standard I/O,,,
26433 gnat_rm, GNAT Reference Manual}.
26434 This chapter covers the following:
26436 @item Standard I/O packages
26438 @item @code{FORM} strings
26440 @item @code{ADA.DIRECT_IO}
26442 @item @code{ADA.SEQUENTIAL_IO}
26444 @item @code{ADA.TEXT_IO}
26446 @item Stream pointer positioning
26448 @item Reading and writing non-regular files
26450 @item @code{GET_IMMEDIATE}
26452 @item Treating @code{TEXT_IO} files as streams
26459 @node Implementation Limits
26460 @section Implementation Limits
26463 The following table lists implementation limits for HP Ada
26465 @multitable @columnfractions .60 .20 .20
26467 @item @emph{Compilation Parameter}
26472 @item In a subprogram or entry declaration, maximum number of
26473 formal parameters that are of an unconstrained record type
26478 @item Maximum identifier length (number of characters)
26483 @item Maximum number of characters in a source line
26488 @item Maximum collection size (number of bytes)
26493 @item Maximum number of discriminants for a record type
26498 @item Maximum number of formal parameters in an entry or
26499 subprogram declaration
26504 @item Maximum number of dimensions in an array type
26509 @item Maximum number of library units and subunits in a compilation.
26514 @item Maximum number of library units and subunits in an execution.
26519 @item Maximum number of objects declared with the pragma @code{COMMON_OBJECT}
26520 or @code{PSECT_OBJECT}
26525 @item Maximum number of enumeration literals in an enumeration type
26531 @item Maximum number of lines in a source file
26536 @item Maximum number of bits in any object
26541 @item Maximum size of the static portion of a stack frame (approximate)
26546 @node Tools and Utilities
26547 @section Tools and Utilities
26550 The following table lists some of the OpenVMS development tools
26551 available for HP Ada, and the corresponding tools for
26552 use with @value{EDITION} on Alpha and I64 platforms.
26553 Aside from the debugger, all the OpenVMS tools identified are part
26554 of the DECset package.
26557 @c Specify table in TeX since Texinfo does a poor job
26561 \settabs\+Language-Sensitive Editor\quad
26562 &Product with HP Ada\quad
26565 &\it Product with HP Ada
26566 & \it Product with GNAT Pro\cr
26568 \+Code Management System
26572 \+Language-Sensitive Editor
26574 & emacs or HP LSE (Alpha)\cr
26584 & OpenVMS Debug (I64)\cr
26586 \+Source Code Analyzer /
26603 \+Coverage Analyzer
26607 \+Module Management
26609 & Not applicable\cr
26619 @c This is the Texinfo version of the table. It renders poorly in pdf, hence
26620 @c the TeX version above for the printed version
26622 @c @multitable @columnfractions .3 .4 .4
26623 @multitable {Source Code Analyzer /}{Tool with HP Ada}{Tool with GNAT Pro}
26625 @tab @i{Tool with HP Ada}
26626 @tab @i{Tool with @value{EDITION}}
26627 @item Code Management@*System
26630 @item Language-Sensitive@*Editor
26632 @tab emacs or HP LSE (Alpha)
26641 @tab OpenVMS Debug (I64)
26642 @item Source Code Analyzer /@*Cross Referencer
26646 @tab HP Digital Test@*Manager (DTM)
26648 @item Performance and@*Coverage Analyzer
26651 @item Module Management@*System
26653 @tab Not applicable
26660 @c **************************************
26661 @node Platform-Specific Information for the Run-Time Libraries
26662 @appendix Platform-Specific Information for the Run-Time Libraries
26663 @cindex Tasking and threads libraries
26664 @cindex Threads libraries and tasking
26665 @cindex Run-time libraries (platform-specific information)
26668 The GNAT run-time implementation may vary with respect to both the
26669 underlying threads library and the exception handling scheme.
26670 For threads support, one or more of the following are supplied:
26672 @item @b{native threads library}, a binding to the thread package from
26673 the underlying operating system
26675 @item @b{pthreads library} (Sparc Solaris only), a binding to the Solaris
26676 POSIX thread package
26680 For exception handling, either or both of two models are supplied:
26682 @item @b{Zero-Cost Exceptions} (``ZCX''),@footnote{
26683 Most programs should experience a substantial speed improvement by
26684 being compiled with a ZCX run-time.
26685 This is especially true for
26686 tasking applications or applications with many exception handlers.}
26687 @cindex Zero-Cost Exceptions
26688 @cindex ZCX (Zero-Cost Exceptions)
26689 which uses binder-generated tables that
26690 are interrogated at run time to locate a handler
26692 @item @b{setjmp / longjmp} (``SJLJ''),
26693 @cindex setjmp/longjmp Exception Model
26694 @cindex SJLJ (setjmp/longjmp Exception Model)
26695 which uses dynamically-set data to establish
26696 the set of handlers
26700 This appendix summarizes which combinations of threads and exception support
26701 are supplied on various GNAT platforms.
26702 It then shows how to select a particular library either
26703 permanently or temporarily,
26704 explains the properties of (and tradeoffs among) the various threads
26705 libraries, and provides some additional
26706 information about several specific platforms.
26709 * Summary of Run-Time Configurations::
26710 * Specifying a Run-Time Library::
26711 * Choosing the Scheduling Policy::
26712 * Solaris-Specific Considerations::
26713 * Linux-Specific Considerations::
26714 * AIX-Specific Considerations::
26715 * Irix-Specific Considerations::
26716 * RTX-Specific Considerations::
26719 @node Summary of Run-Time Configurations
26720 @section Summary of Run-Time Configurations
26722 @multitable @columnfractions .30 .70
26723 @item @b{alpha-openvms}
26724 @item @code{@ @ }@i{rts-native (default)}
26725 @item @code{@ @ @ @ }Tasking @tab native VMS threads
26726 @item @code{@ @ @ @ }Exceptions @tab ZCX
26728 @item @b{alpha-tru64}
26729 @item @code{@ @ }@i{rts-native (default)}
26730 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
26731 @item @code{@ @ @ @ }Exceptions @tab ZCX
26733 @item @code{@ @ }@i{rts-sjlj}
26734 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
26735 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26737 @item @b{ia64-hp_linux}
26738 @item @code{@ @ }@i{rts-native (default)}
26739 @item @code{@ @ @ @ }Tasking @tab pthread library
26740 @item @code{@ @ @ @ }Exceptions @tab ZCX
26742 @item @b{ia64-hpux}
26743 @item @code{@ @ }@i{rts-native (default)}
26744 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
26745 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26747 @item @b{ia64-openvms}
26748 @item @code{@ @ }@i{rts-native (default)}
26749 @item @code{@ @ @ @ }Tasking @tab native VMS threads
26750 @item @code{@ @ @ @ }Exceptions @tab ZCX
26752 @item @b{ia64-sgi_linux}
26753 @item @code{@ @ }@i{rts-native (default)}
26754 @item @code{@ @ @ @ }Tasking @tab pthread library
26755 @item @code{@ @ @ @ }Exceptions @tab ZCX
26757 @item @b{mips-irix}
26758 @item @code{@ @ }@i{rts-native (default)}
26759 @item @code{@ @ @ @ }Tasking @tab native IRIX threads
26760 @item @code{@ @ @ @ }Exceptions @tab ZCX
26763 @item @code{@ @ }@i{rts-native (default)}
26764 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
26765 @item @code{@ @ @ @ }Exceptions @tab ZCX
26767 @item @code{@ @ }@i{rts-sjlj}
26768 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
26769 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26772 @item @code{@ @ }@i{rts-native (default)}
26773 @item @code{@ @ @ @ }Tasking @tab native AIX threads
26774 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26776 @item @b{ppc-darwin}
26777 @item @code{@ @ }@i{rts-native (default)}
26778 @item @code{@ @ @ @ }Tasking @tab native MacOS threads
26779 @item @code{@ @ @ @ }Exceptions @tab ZCX
26781 @item @b{sparc-solaris} @tab
26782 @item @code{@ @ }@i{rts-native (default)}
26783 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
26784 @item @code{@ @ @ @ }Exceptions @tab ZCX
26786 @item @code{@ @ }@i{rts-pthread}
26787 @item @code{@ @ @ @ }Tasking @tab pthread library
26788 @item @code{@ @ @ @ }Exceptions @tab ZCX
26790 @item @code{@ @ }@i{rts-sjlj}
26791 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
26792 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26794 @item @b{sparc64-solaris} @tab
26795 @item @code{@ @ }@i{rts-native (default)}
26796 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
26797 @item @code{@ @ @ @ }Exceptions @tab ZCX
26799 @item @b{x86-linux}
26800 @item @code{@ @ }@i{rts-native (default)}
26801 @item @code{@ @ @ @ }Tasking @tab pthread library
26802 @item @code{@ @ @ @ }Exceptions @tab ZCX
26804 @item @code{@ @ }@i{rts-sjlj}
26805 @item @code{@ @ @ @ }Tasking @tab pthread library
26806 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26809 @item @code{@ @ }@i{rts-native (default)}
26810 @item @code{@ @ @ @ }Tasking @tab native LynxOS threads
26811 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26813 @item @b{x86-solaris}
26814 @item @code{@ @ }@i{rts-native (default)}
26815 @item @code{@ @ @ @ }Tasking @tab native Solaris threads
26816 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26818 @item @b{x86-windows}
26819 @item @code{@ @ }@i{rts-native (default)}
26820 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
26821 @item @code{@ @ @ @ }Exceptions @tab ZCX
26823 @item @code{@ @ }@i{rts-sjlj (default)}
26824 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
26825 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26827 @item @b{x86-windows-rtx}
26828 @item @code{@ @ }@i{rts-rtx-rtss (default)}
26829 @item @code{@ @ @ @ }Tasking @tab RTX real-time subsystem RTSS threads (kernel mode)
26830 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26832 @item @code{@ @ }@i{rts-rtx-w32}
26833 @item @code{@ @ @ @ }Tasking @tab RTX Win32 threads (user mode)
26834 @item @code{@ @ @ @ }Exceptions @tab ZCX
26836 @item @b{x86_64-linux}
26837 @item @code{@ @ }@i{rts-native (default)}
26838 @item @code{@ @ @ @ }Tasking @tab pthread library
26839 @item @code{@ @ @ @ }Exceptions @tab ZCX
26841 @item @code{@ @ }@i{rts-sjlj}
26842 @item @code{@ @ @ @ }Tasking @tab pthread library
26843 @item @code{@ @ @ @ }Exceptions @tab SJLJ
26847 @node Specifying a Run-Time Library
26848 @section Specifying a Run-Time Library
26851 The @file{adainclude} subdirectory containing the sources of the GNAT
26852 run-time library, and the @file{adalib} subdirectory containing the
26853 @file{ALI} files and the static and/or shared GNAT library, are located
26854 in the gcc target-dependent area:
26857 target=$prefix/lib/gcc/gcc-@i{dumpmachine}/gcc-@i{dumpversion}/
26861 As indicated above, on some platforms several run-time libraries are supplied.
26862 These libraries are installed in the target dependent area and
26863 contain a complete source and binary subdirectory. The detailed description
26864 below explains the differences between the different libraries in terms of
26865 their thread support.
26867 The default run-time library (when GNAT is installed) is @emph{rts-native}.
26868 This default run time is selected by the means of soft links.
26869 For example on x86-linux:
26875 +--- adainclude----------+
26877 +--- adalib-----------+ |
26879 +--- rts-native | |
26881 | +--- adainclude <---+
26883 | +--- adalib <----+
26894 If the @i{rts-sjlj} library is to be selected on a permanent basis,
26895 these soft links can be modified with the following commands:
26899 $ rm -f adainclude adalib
26900 $ ln -s rts-sjlj/adainclude adainclude
26901 $ ln -s rts-sjlj/adalib adalib
26905 Alternatively, you can specify @file{rts-sjlj/adainclude} in the file
26906 @file{$target/ada_source_path} and @file{rts-sjlj/adalib} in
26907 @file{$target/ada_object_path}.
26909 Selecting another run-time library temporarily can be
26910 achieved by using the @option{--RTS} switch, e.g., @option{--RTS=sjlj}
26911 @cindex @option{--RTS} option
26913 @node Choosing the Scheduling Policy
26914 @section Choosing the Scheduling Policy
26917 When using a POSIX threads implementation, you have a choice of several
26918 scheduling policies: @code{SCHED_FIFO},
26919 @cindex @code{SCHED_FIFO} scheduling policy
26921 @cindex @code{SCHED_RR} scheduling policy
26922 and @code{SCHED_OTHER}.
26923 @cindex @code{SCHED_OTHER} scheduling policy
26924 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
26925 or @code{SCHED_RR} requires special (e.g., root) privileges.
26927 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
26929 @cindex @code{SCHED_FIFO} scheduling policy
26930 you can use one of the following:
26934 @code{pragma Time_Slice (0.0)}
26935 @cindex pragma Time_Slice
26937 the corresponding binder option @option{-T0}
26938 @cindex @option{-T0} option
26940 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
26941 @cindex pragma Task_Dispatching_Policy
26945 To specify @code{SCHED_RR},
26946 @cindex @code{SCHED_RR} scheduling policy
26947 you should use @code{pragma Time_Slice} with a
26948 value greater than @code{0.0}, or else use the corresponding @option{-T}
26951 @node Solaris-Specific Considerations
26952 @section Solaris-Specific Considerations
26953 @cindex Solaris Sparc threads libraries
26956 This section addresses some topics related to the various threads libraries
26960 * Solaris Threads Issues::
26963 @node Solaris Threads Issues
26964 @subsection Solaris Threads Issues
26967 GNAT under Solaris/Sparc 32 bits comes with an alternate tasking run-time
26968 library based on POSIX threads --- @emph{rts-pthread}.
26969 @cindex rts-pthread threads library
26970 This run-time library has the advantage of being mostly shared across all
26971 POSIX-compliant thread implementations, and it also provides under
26972 @w{Solaris 8} the @code{PTHREAD_PRIO_INHERIT}
26973 @cindex @code{PTHREAD_PRIO_INHERIT} policy (under rts-pthread)
26974 and @code{PTHREAD_PRIO_PROTECT}
26975 @cindex @code{PTHREAD_PRIO_PROTECT} policy (under rts-pthread)
26976 semantics that can be selected using the predefined pragma
26977 @code{Locking_Policy}
26978 @cindex pragma Locking_Policy (under rts-pthread)
26980 @code{Inheritance_Locking} and @code{Ceiling_Locking} as the policy.
26981 @cindex @code{Inheritance_Locking} (under rts-pthread)
26982 @cindex @code{Ceiling_Locking} (under rts-pthread)
26984 As explained above, the native run-time library is based on the Solaris thread
26985 library (@code{libthread}) and is the default library.
26987 When the Solaris threads library is used (this is the default), programs
26988 compiled with GNAT can automatically take advantage of
26989 and can thus execute on multiple processors.
26990 The user can alternatively specify a processor on which the program should run
26991 to emulate a single-processor system. The multiprocessor / uniprocessor choice
26993 setting the environment variable @env{GNAT_PROCESSOR}
26994 @cindex @env{GNAT_PROCESSOR} environment variable (on Sparc Solaris)
26995 to one of the following:
26999 Use the default configuration (run the program on all
27000 available processors) - this is the same as having @code{GNAT_PROCESSOR}
27004 Let the run-time implementation choose one processor and run the program on
27007 @item 0 .. Last_Proc
27008 Run the program on the specified processor.
27009 @code{Last_Proc} is equal to @code{_SC_NPROCESSORS_CONF - 1}
27010 (where @code{_SC_NPROCESSORS_CONF} is a system variable).
27013 @node Linux-Specific Considerations
27014 @section Linux-Specific Considerations
27015 @cindex Linux threads libraries
27018 On GNU/Linux without NPTL support (usually system with GNU C Library
27019 older than 2.3), the signal model is not POSIX compliant, which means
27020 that to send a signal to the process, you need to send the signal to all
27021 threads, e.g.@: by using @code{killpg()}.
27023 @node AIX-Specific Considerations
27024 @section AIX-Specific Considerations
27025 @cindex AIX resolver library
27028 On AIX, the resolver library initializes some internal structure on
27029 the first call to @code{get*by*} functions, which are used to implement
27030 @code{GNAT.Sockets.Get_Host_By_Name} and
27031 @code{GNAT.Sockets.Get_Host_By_Address}.
27032 If such initialization occurs within an Ada task, and the stack size for
27033 the task is the default size, a stack overflow may occur.
27035 To avoid this overflow, the user should either ensure that the first call
27036 to @code{GNAT.Sockets.Get_Host_By_Name} or
27037 @code{GNAT.Sockets.Get_Host_By_Addrss}
27038 occurs in the environment task, or use @code{pragma Storage_Size} to
27039 specify a sufficiently large size for the stack of the task that contains
27042 @node Irix-Specific Considerations
27043 @section Irix-Specific Considerations
27044 @cindex Irix libraries
27047 The GCC support libraries coming with the Irix compiler have moved to
27048 their canonical place with respect to the general Irix ABI related
27049 conventions. Running applications built with the default shared GNAT
27050 run-time now requires the LD_LIBRARY_PATH environment variable to
27051 include this location. A possible way to achieve this is to issue the
27052 following command line on a bash prompt:
27056 $ LD_LIBRARY_PATH=$LD_LIBRARY_PATH:`dirname \`gcc --print-file-name=libgcc_s.so\``
27060 @node RTX-Specific Considerations
27061 @section RTX-Specific Considerations
27062 @cindex RTX libraries
27065 The Real-time Extension (RTX) to Windows is based on the Windows Win32
27066 API. Applications can be built to work in two different modes:
27070 Windows executables that run in Ring 3 to utilize memory protection
27071 (@emph{rts-rtx-w32}).
27074 Real-time subsystem (RTSS) executables that run in Ring 0, where
27075 performance can be optimized with RTSS applications taking precedent
27076 over all Windows applications (@emph{rts-rtx-rtss}).
27080 @c *******************************
27081 @node Example of Binder Output File
27082 @appendix Example of Binder Output File
27085 This Appendix displays the source code for @command{gnatbind}'s output
27086 file generated for a simple ``Hello World'' program.
27087 Comments have been added for clarification purposes.
27089 @smallexample @c adanocomment
27093 -- The package is called Ada_Main unless this name is actually used
27094 -- as a unit name in the partition, in which case some other unique
27098 package ada_main is
27100 Elab_Final_Code : Integer;
27101 pragma Import (C, Elab_Final_Code, "__gnat_inside_elab_final_code");
27103 -- The main program saves the parameters (argument count,
27104 -- argument values, environment pointer) in global variables
27105 -- for later access by other units including
27106 -- Ada.Command_Line.
27108 gnat_argc : Integer;
27109 gnat_argv : System.Address;
27110 gnat_envp : System.Address;
27112 -- The actual variables are stored in a library routine. This
27113 -- is useful for some shared library situations, where there
27114 -- are problems if variables are not in the library.
27116 pragma Import (C, gnat_argc);
27117 pragma Import (C, gnat_argv);
27118 pragma Import (C, gnat_envp);
27120 -- The exit status is similarly an external location
27122 gnat_exit_status : Integer;
27123 pragma Import (C, gnat_exit_status);
27125 GNAT_Version : constant String :=
27126 "GNAT Version: 6.0.0w (20061115)";
27127 pragma Export (C, GNAT_Version, "__gnat_version");
27129 -- This is the generated adafinal routine that performs
27130 -- finalization at the end of execution. In the case where
27131 -- Ada is the main program, this main program makes a call
27132 -- to adafinal at program termination.
27134 procedure adafinal;
27135 pragma Export (C, adafinal, "adafinal");
27137 -- This is the generated adainit routine that performs
27138 -- initialization at the start of execution. In the case
27139 -- where Ada is the main program, this main program makes
27140 -- a call to adainit at program startup.
27143 pragma Export (C, adainit, "adainit");
27145 -- This routine is called at the start of execution. It is
27146 -- a dummy routine that is used by the debugger to breakpoint
27147 -- at the start of execution.
27149 procedure Break_Start;
27150 pragma Import (C, Break_Start, "__gnat_break_start");
27152 -- This is the actual generated main program (it would be
27153 -- suppressed if the no main program switch were used). As
27154 -- required by standard system conventions, this program has
27155 -- the external name main.
27159 argv : System.Address;
27160 envp : System.Address)
27162 pragma Export (C, main, "main");
27164 -- The following set of constants give the version
27165 -- identification values for every unit in the bound
27166 -- partition. This identification is computed from all
27167 -- dependent semantic units, and corresponds to the
27168 -- string that would be returned by use of the
27169 -- Body_Version or Version attributes.
27171 type Version_32 is mod 2 ** 32;
27172 u00001 : constant Version_32 := 16#7880BEB3#;
27173 u00002 : constant Version_32 := 16#0D24CBD0#;
27174 u00003 : constant Version_32 := 16#3283DBEB#;
27175 u00004 : constant Version_32 := 16#2359F9ED#;
27176 u00005 : constant Version_32 := 16#664FB847#;
27177 u00006 : constant Version_32 := 16#68E803DF#;
27178 u00007 : constant Version_32 := 16#5572E604#;
27179 u00008 : constant Version_32 := 16#46B173D8#;
27180 u00009 : constant Version_32 := 16#156A40CF#;
27181 u00010 : constant Version_32 := 16#033DABE0#;
27182 u00011 : constant Version_32 := 16#6AB38FEA#;
27183 u00012 : constant Version_32 := 16#22B6217D#;
27184 u00013 : constant Version_32 := 16#68A22947#;
27185 u00014 : constant Version_32 := 16#18CC4A56#;
27186 u00015 : constant Version_32 := 16#08258E1B#;
27187 u00016 : constant Version_32 := 16#367D5222#;
27188 u00017 : constant Version_32 := 16#20C9ECA4#;
27189 u00018 : constant Version_32 := 16#50D32CB6#;
27190 u00019 : constant Version_32 := 16#39A8BB77#;
27191 u00020 : constant Version_32 := 16#5CF8FA2B#;
27192 u00021 : constant Version_32 := 16#2F1EB794#;
27193 u00022 : constant Version_32 := 16#31AB6444#;
27194 u00023 : constant Version_32 := 16#1574B6E9#;
27195 u00024 : constant Version_32 := 16#5109C189#;
27196 u00025 : constant Version_32 := 16#56D770CD#;
27197 u00026 : constant Version_32 := 16#02F9DE3D#;
27198 u00027 : constant Version_32 := 16#08AB6B2C#;
27199 u00028 : constant Version_32 := 16#3FA37670#;
27200 u00029 : constant Version_32 := 16#476457A0#;
27201 u00030 : constant Version_32 := 16#731E1B6E#;
27202 u00031 : constant Version_32 := 16#23C2E789#;
27203 u00032 : constant Version_32 := 16#0F1BD6A1#;
27204 u00033 : constant Version_32 := 16#7C25DE96#;
27205 u00034 : constant Version_32 := 16#39ADFFA2#;
27206 u00035 : constant Version_32 := 16#571DE3E7#;
27207 u00036 : constant Version_32 := 16#5EB646AB#;
27208 u00037 : constant Version_32 := 16#4249379B#;
27209 u00038 : constant Version_32 := 16#0357E00A#;
27210 u00039 : constant Version_32 := 16#3784FB72#;
27211 u00040 : constant Version_32 := 16#2E723019#;
27212 u00041 : constant Version_32 := 16#623358EA#;
27213 u00042 : constant Version_32 := 16#107F9465#;
27214 u00043 : constant Version_32 := 16#6843F68A#;
27215 u00044 : constant Version_32 := 16#63305874#;
27216 u00045 : constant Version_32 := 16#31E56CE1#;
27217 u00046 : constant Version_32 := 16#02917970#;
27218 u00047 : constant Version_32 := 16#6CCBA70E#;
27219 u00048 : constant Version_32 := 16#41CD4204#;
27220 u00049 : constant Version_32 := 16#572E3F58#;
27221 u00050 : constant Version_32 := 16#20729FF5#;
27222 u00051 : constant Version_32 := 16#1D4F93E8#;
27223 u00052 : constant Version_32 := 16#30B2EC3D#;
27224 u00053 : constant Version_32 := 16#34054F96#;
27225 u00054 : constant Version_32 := 16#5A199860#;
27226 u00055 : constant Version_32 := 16#0E7F912B#;
27227 u00056 : constant Version_32 := 16#5760634A#;
27228 u00057 : constant Version_32 := 16#5D851835#;
27230 -- The following Export pragmas export the version numbers
27231 -- with symbolic names ending in B (for body) or S
27232 -- (for spec) so that they can be located in a link. The
27233 -- information provided here is sufficient to track down
27234 -- the exact versions of units used in a given build.
27236 pragma Export (C, u00001, "helloB");
27237 pragma Export (C, u00002, "system__standard_libraryB");
27238 pragma Export (C, u00003, "system__standard_libraryS");
27239 pragma Export (C, u00004, "adaS");
27240 pragma Export (C, u00005, "ada__text_ioB");
27241 pragma Export (C, u00006, "ada__text_ioS");
27242 pragma Export (C, u00007, "ada__exceptionsB");
27243 pragma Export (C, u00008, "ada__exceptionsS");
27244 pragma Export (C, u00009, "gnatS");
27245 pragma Export (C, u00010, "gnat__heap_sort_aB");
27246 pragma Export (C, u00011, "gnat__heap_sort_aS");
27247 pragma Export (C, u00012, "systemS");
27248 pragma Export (C, u00013, "system__exception_tableB");
27249 pragma Export (C, u00014, "system__exception_tableS");
27250 pragma Export (C, u00015, "gnat__htableB");
27251 pragma Export (C, u00016, "gnat__htableS");
27252 pragma Export (C, u00017, "system__exceptionsS");
27253 pragma Export (C, u00018, "system__machine_state_operationsB");
27254 pragma Export (C, u00019, "system__machine_state_operationsS");
27255 pragma Export (C, u00020, "system__machine_codeS");
27256 pragma Export (C, u00021, "system__storage_elementsB");
27257 pragma Export (C, u00022, "system__storage_elementsS");
27258 pragma Export (C, u00023, "system__secondary_stackB");
27259 pragma Export (C, u00024, "system__secondary_stackS");
27260 pragma Export (C, u00025, "system__parametersB");
27261 pragma Export (C, u00026, "system__parametersS");
27262 pragma Export (C, u00027, "system__soft_linksB");
27263 pragma Export (C, u00028, "system__soft_linksS");
27264 pragma Export (C, u00029, "system__stack_checkingB");
27265 pragma Export (C, u00030, "system__stack_checkingS");
27266 pragma Export (C, u00031, "system__tracebackB");
27267 pragma Export (C, u00032, "system__tracebackS");
27268 pragma Export (C, u00033, "ada__streamsS");
27269 pragma Export (C, u00034, "ada__tagsB");
27270 pragma Export (C, u00035, "ada__tagsS");
27271 pragma Export (C, u00036, "system__string_opsB");
27272 pragma Export (C, u00037, "system__string_opsS");
27273 pragma Export (C, u00038, "interfacesS");
27274 pragma Export (C, u00039, "interfaces__c_streamsB");
27275 pragma Export (C, u00040, "interfaces__c_streamsS");
27276 pragma Export (C, u00041, "system__file_ioB");
27277 pragma Export (C, u00042, "system__file_ioS");
27278 pragma Export (C, u00043, "ada__finalizationB");
27279 pragma Export (C, u00044, "ada__finalizationS");
27280 pragma Export (C, u00045, "system__finalization_rootB");
27281 pragma Export (C, u00046, "system__finalization_rootS");
27282 pragma Export (C, u00047, "system__finalization_implementationB");
27283 pragma Export (C, u00048, "system__finalization_implementationS");
27284 pragma Export (C, u00049, "system__string_ops_concat_3B");
27285 pragma Export (C, u00050, "system__string_ops_concat_3S");
27286 pragma Export (C, u00051, "system__stream_attributesB");
27287 pragma Export (C, u00052, "system__stream_attributesS");
27288 pragma Export (C, u00053, "ada__io_exceptionsS");
27289 pragma Export (C, u00054, "system__unsigned_typesS");
27290 pragma Export (C, u00055, "system__file_control_blockS");
27291 pragma Export (C, u00056, "ada__finalization__list_controllerB");
27292 pragma Export (C, u00057, "ada__finalization__list_controllerS");
27294 -- BEGIN ELABORATION ORDER
27297 -- gnat.heap_sort_a (spec)
27298 -- gnat.heap_sort_a (body)
27299 -- gnat.htable (spec)
27300 -- gnat.htable (body)
27301 -- interfaces (spec)
27303 -- system.machine_code (spec)
27304 -- system.parameters (spec)
27305 -- system.parameters (body)
27306 -- interfaces.c_streams (spec)
27307 -- interfaces.c_streams (body)
27308 -- system.standard_library (spec)
27309 -- ada.exceptions (spec)
27310 -- system.exception_table (spec)
27311 -- system.exception_table (body)
27312 -- ada.io_exceptions (spec)
27313 -- system.exceptions (spec)
27314 -- system.storage_elements (spec)
27315 -- system.storage_elements (body)
27316 -- system.machine_state_operations (spec)
27317 -- system.machine_state_operations (body)
27318 -- system.secondary_stack (spec)
27319 -- system.stack_checking (spec)
27320 -- system.soft_links (spec)
27321 -- system.soft_links (body)
27322 -- system.stack_checking (body)
27323 -- system.secondary_stack (body)
27324 -- system.standard_library (body)
27325 -- system.string_ops (spec)
27326 -- system.string_ops (body)
27329 -- ada.streams (spec)
27330 -- system.finalization_root (spec)
27331 -- system.finalization_root (body)
27332 -- system.string_ops_concat_3 (spec)
27333 -- system.string_ops_concat_3 (body)
27334 -- system.traceback (spec)
27335 -- system.traceback (body)
27336 -- ada.exceptions (body)
27337 -- system.unsigned_types (spec)
27338 -- system.stream_attributes (spec)
27339 -- system.stream_attributes (body)
27340 -- system.finalization_implementation (spec)
27341 -- system.finalization_implementation (body)
27342 -- ada.finalization (spec)
27343 -- ada.finalization (body)
27344 -- ada.finalization.list_controller (spec)
27345 -- ada.finalization.list_controller (body)
27346 -- system.file_control_block (spec)
27347 -- system.file_io (spec)
27348 -- system.file_io (body)
27349 -- ada.text_io (spec)
27350 -- ada.text_io (body)
27352 -- END ELABORATION ORDER
27356 -- The following source file name pragmas allow the generated file
27357 -- names to be unique for different main programs. They are needed
27358 -- since the package name will always be Ada_Main.
27360 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
27361 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
27363 -- Generated package body for Ada_Main starts here
27365 package body ada_main is
27367 -- The actual finalization is performed by calling the
27368 -- library routine in System.Standard_Library.Adafinal
27370 procedure Do_Finalize;
27371 pragma Import (C, Do_Finalize, "system__standard_library__adafinal");
27378 procedure adainit is
27380 -- These booleans are set to True once the associated unit has
27381 -- been elaborated. It is also used to avoid elaborating the
27382 -- same unit twice.
27385 pragma Import (Ada, E040, "interfaces__c_streams_E");
27388 pragma Import (Ada, E008, "ada__exceptions_E");
27391 pragma Import (Ada, E014, "system__exception_table_E");
27394 pragma Import (Ada, E053, "ada__io_exceptions_E");
27397 pragma Import (Ada, E017, "system__exceptions_E");
27400 pragma Import (Ada, E024, "system__secondary_stack_E");
27403 pragma Import (Ada, E030, "system__stack_checking_E");
27406 pragma Import (Ada, E028, "system__soft_links_E");
27409 pragma Import (Ada, E035, "ada__tags_E");
27412 pragma Import (Ada, E033, "ada__streams_E");
27415 pragma Import (Ada, E046, "system__finalization_root_E");
27418 pragma Import (Ada, E048, "system__finalization_implementation_E");
27421 pragma Import (Ada, E044, "ada__finalization_E");
27424 pragma Import (Ada, E057, "ada__finalization__list_controller_E");
27427 pragma Import (Ada, E055, "system__file_control_block_E");
27430 pragma Import (Ada, E042, "system__file_io_E");
27433 pragma Import (Ada, E006, "ada__text_io_E");
27435 -- Set_Globals is a library routine that stores away the
27436 -- value of the indicated set of global values in global
27437 -- variables within the library.
27439 procedure Set_Globals
27440 (Main_Priority : Integer;
27441 Time_Slice_Value : Integer;
27442 WC_Encoding : Character;
27443 Locking_Policy : Character;
27444 Queuing_Policy : Character;
27445 Task_Dispatching_Policy : Character;
27446 Adafinal : System.Address;
27447 Unreserve_All_Interrupts : Integer;
27448 Exception_Tracebacks : Integer);
27449 @findex __gnat_set_globals
27450 pragma Import (C, Set_Globals, "__gnat_set_globals");
27452 -- SDP_Table_Build is a library routine used to build the
27453 -- exception tables. See unit Ada.Exceptions in files
27454 -- a-except.ads/adb for full details of how zero cost
27455 -- exception handling works. This procedure, the call to
27456 -- it, and the two following tables are all omitted if the
27457 -- build is in longjmp/setjmp exception mode.
27459 @findex SDP_Table_Build
27460 @findex Zero Cost Exceptions
27461 procedure SDP_Table_Build
27462 (SDP_Addresses : System.Address;
27463 SDP_Count : Natural;
27464 Elab_Addresses : System.Address;
27465 Elab_Addr_Count : Natural);
27466 pragma Import (C, SDP_Table_Build, "__gnat_SDP_Table_Build");
27468 -- Table of Unit_Exception_Table addresses. Used for zero
27469 -- cost exception handling to build the top level table.
27471 ST : aliased constant array (1 .. 23) of System.Address := (
27473 Ada.Text_Io'UET_Address,
27474 Ada.Exceptions'UET_Address,
27475 Gnat.Heap_Sort_A'UET_Address,
27476 System.Exception_Table'UET_Address,
27477 System.Machine_State_Operations'UET_Address,
27478 System.Secondary_Stack'UET_Address,
27479 System.Parameters'UET_Address,
27480 System.Soft_Links'UET_Address,
27481 System.Stack_Checking'UET_Address,
27482 System.Traceback'UET_Address,
27483 Ada.Streams'UET_Address,
27484 Ada.Tags'UET_Address,
27485 System.String_Ops'UET_Address,
27486 Interfaces.C_Streams'UET_Address,
27487 System.File_Io'UET_Address,
27488 Ada.Finalization'UET_Address,
27489 System.Finalization_Root'UET_Address,
27490 System.Finalization_Implementation'UET_Address,
27491 System.String_Ops_Concat_3'UET_Address,
27492 System.Stream_Attributes'UET_Address,
27493 System.File_Control_Block'UET_Address,
27494 Ada.Finalization.List_Controller'UET_Address);
27496 -- Table of addresses of elaboration routines. Used for
27497 -- zero cost exception handling to make sure these
27498 -- addresses are included in the top level procedure
27501 EA : aliased constant array (1 .. 23) of System.Address := (
27502 adainit'Code_Address,
27503 Do_Finalize'Code_Address,
27504 Ada.Exceptions'Elab_Spec'Address,
27505 System.Exceptions'Elab_Spec'Address,
27506 Interfaces.C_Streams'Elab_Spec'Address,
27507 System.Exception_Table'Elab_Body'Address,
27508 Ada.Io_Exceptions'Elab_Spec'Address,
27509 System.Stack_Checking'Elab_Spec'Address,
27510 System.Soft_Links'Elab_Body'Address,
27511 System.Secondary_Stack'Elab_Body'Address,
27512 Ada.Tags'Elab_Spec'Address,
27513 Ada.Tags'Elab_Body'Address,
27514 Ada.Streams'Elab_Spec'Address,
27515 System.Finalization_Root'Elab_Spec'Address,
27516 Ada.Exceptions'Elab_Body'Address,
27517 System.Finalization_Implementation'Elab_Spec'Address,
27518 System.Finalization_Implementation'Elab_Body'Address,
27519 Ada.Finalization'Elab_Spec'Address,
27520 Ada.Finalization.List_Controller'Elab_Spec'Address,
27521 System.File_Control_Block'Elab_Spec'Address,
27522 System.File_Io'Elab_Body'Address,
27523 Ada.Text_Io'Elab_Spec'Address,
27524 Ada.Text_Io'Elab_Body'Address);
27526 -- Start of processing for adainit
27530 -- Call SDP_Table_Build to build the top level procedure
27531 -- table for zero cost exception handling (omitted in
27532 -- longjmp/setjmp mode).
27534 SDP_Table_Build (ST'Address, 23, EA'Address, 23);
27536 -- Call Set_Globals to record various information for
27537 -- this partition. The values are derived by the binder
27538 -- from information stored in the ali files by the compiler.
27540 @findex __gnat_set_globals
27542 (Main_Priority => -1,
27543 -- Priority of main program, -1 if no pragma Priority used
27545 Time_Slice_Value => -1,
27546 -- Time slice from Time_Slice pragma, -1 if none used
27548 WC_Encoding => 'b',
27549 -- Wide_Character encoding used, default is brackets
27551 Locking_Policy => ' ',
27552 -- Locking_Policy used, default of space means not
27553 -- specified, otherwise it is the first character of
27554 -- the policy name.
27556 Queuing_Policy => ' ',
27557 -- Queuing_Policy used, default of space means not
27558 -- specified, otherwise it is the first character of
27559 -- the policy name.
27561 Task_Dispatching_Policy => ' ',
27562 -- Task_Dispatching_Policy used, default of space means
27563 -- not specified, otherwise first character of the
27566 Adafinal => System.Null_Address,
27567 -- Address of Adafinal routine, not used anymore
27569 Unreserve_All_Interrupts => 0,
27570 -- Set true if pragma Unreserve_All_Interrupts was used
27572 Exception_Tracebacks => 0);
27573 -- Indicates if exception tracebacks are enabled
27575 Elab_Final_Code := 1;
27577 -- Now we have the elaboration calls for all units in the partition.
27578 -- The Elab_Spec and Elab_Body attributes generate references to the
27579 -- implicit elaboration procedures generated by the compiler for
27580 -- each unit that requires elaboration.
27583 Interfaces.C_Streams'Elab_Spec;
27587 Ada.Exceptions'Elab_Spec;
27590 System.Exception_Table'Elab_Body;
27594 Ada.Io_Exceptions'Elab_Spec;
27598 System.Exceptions'Elab_Spec;
27602 System.Stack_Checking'Elab_Spec;
27605 System.Soft_Links'Elab_Body;
27610 System.Secondary_Stack'Elab_Body;
27614 Ada.Tags'Elab_Spec;
27617 Ada.Tags'Elab_Body;
27621 Ada.Streams'Elab_Spec;
27625 System.Finalization_Root'Elab_Spec;
27629 Ada.Exceptions'Elab_Body;
27633 System.Finalization_Implementation'Elab_Spec;
27636 System.Finalization_Implementation'Elab_Body;
27640 Ada.Finalization'Elab_Spec;
27644 Ada.Finalization.List_Controller'Elab_Spec;
27648 System.File_Control_Block'Elab_Spec;
27652 System.File_Io'Elab_Body;
27656 Ada.Text_Io'Elab_Spec;
27659 Ada.Text_Io'Elab_Body;
27663 Elab_Final_Code := 0;
27671 procedure adafinal is
27680 -- main is actually a function, as in the ANSI C standard,
27681 -- defined to return the exit status. The three parameters
27682 -- are the argument count, argument values and environment
27685 @findex Main Program
27688 argv : System.Address;
27689 envp : System.Address)
27692 -- The initialize routine performs low level system
27693 -- initialization using a standard library routine which
27694 -- sets up signal handling and performs any other
27695 -- required setup. The routine can be found in file
27698 @findex __gnat_initialize
27699 procedure initialize;
27700 pragma Import (C, initialize, "__gnat_initialize");
27702 -- The finalize routine performs low level system
27703 -- finalization using a standard library routine. The
27704 -- routine is found in file a-final.c and in the standard
27705 -- distribution is a dummy routine that does nothing, so
27706 -- really this is a hook for special user finalization.
27708 @findex __gnat_finalize
27709 procedure finalize;
27710 pragma Import (C, finalize, "__gnat_finalize");
27712 -- We get to the main program of the partition by using
27713 -- pragma Import because if we try to with the unit and
27714 -- call it Ada style, then not only do we waste time
27715 -- recompiling it, but also, we don't really know the right
27716 -- switches (e.g.@: identifier character set) to be used
27719 procedure Ada_Main_Program;
27720 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
27722 -- Start of processing for main
27725 -- Save global variables
27731 -- Call low level system initialization
27735 -- Call our generated Ada initialization routine
27739 -- This is the point at which we want the debugger to get
27744 -- Now we call the main program of the partition
27748 -- Perform Ada finalization
27752 -- Perform low level system finalization
27756 -- Return the proper exit status
27757 return (gnat_exit_status);
27760 -- This section is entirely comments, so it has no effect on the
27761 -- compilation of the Ada_Main package. It provides the list of
27762 -- object files and linker options, as well as some standard
27763 -- libraries needed for the link. The gnatlink utility parses
27764 -- this b~hello.adb file to read these comment lines to generate
27765 -- the appropriate command line arguments for the call to the
27766 -- system linker. The BEGIN/END lines are used for sentinels for
27767 -- this parsing operation.
27769 -- The exact file names will of course depend on the environment,
27770 -- host/target and location of files on the host system.
27772 @findex Object file list
27773 -- BEGIN Object file/option list
27776 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
27777 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
27778 -- END Object file/option list
27784 The Ada code in the above example is exactly what is generated by the
27785 binder. We have added comments to more clearly indicate the function
27786 of each part of the generated @code{Ada_Main} package.
27788 The code is standard Ada in all respects, and can be processed by any
27789 tools that handle Ada. In particular, it is possible to use the debugger
27790 in Ada mode to debug the generated @code{Ada_Main} package. For example,
27791 suppose that for reasons that you do not understand, your program is crashing
27792 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
27793 you can place a breakpoint on the call:
27795 @smallexample @c ada
27796 Ada.Text_Io'Elab_Body;
27800 and trace the elaboration routine for this package to find out where
27801 the problem might be (more usually of course you would be debugging
27802 elaboration code in your own application).
27804 @node Elaboration Order Handling in GNAT
27805 @appendix Elaboration Order Handling in GNAT
27806 @cindex Order of elaboration
27807 @cindex Elaboration control
27810 * Elaboration Code::
27811 * Checking the Elaboration Order::
27812 * Controlling the Elaboration Order::
27813 * Controlling Elaboration in GNAT - Internal Calls::
27814 * Controlling Elaboration in GNAT - External Calls::
27815 * Default Behavior in GNAT - Ensuring Safety::
27816 * Treatment of Pragma Elaborate::
27817 * Elaboration Issues for Library Tasks::
27818 * Mixing Elaboration Models::
27819 * What to Do If the Default Elaboration Behavior Fails::
27820 * Elaboration for Access-to-Subprogram Values::
27821 * Summary of Procedures for Elaboration Control::
27822 * Other Elaboration Order Considerations::
27826 This chapter describes the handling of elaboration code in Ada and
27827 in GNAT, and discusses how the order of elaboration of program units can
27828 be controlled in GNAT, either automatically or with explicit programming
27831 @node Elaboration Code
27832 @section Elaboration Code
27835 Ada provides rather general mechanisms for executing code at elaboration
27836 time, that is to say before the main program starts executing. Such code arises
27840 @item Initializers for variables.
27841 Variables declared at the library level, in package specs or bodies, can
27842 require initialization that is performed at elaboration time, as in:
27843 @smallexample @c ada
27845 Sqrt_Half : Float := Sqrt (0.5);
27849 @item Package initialization code
27850 Code in a @code{BEGIN-END} section at the outer level of a package body is
27851 executed as part of the package body elaboration code.
27853 @item Library level task allocators
27854 Tasks that are declared using task allocators at the library level
27855 start executing immediately and hence can execute at elaboration time.
27859 Subprogram calls are possible in any of these contexts, which means that
27860 any arbitrary part of the program may be executed as part of the elaboration
27861 code. It is even possible to write a program which does all its work at
27862 elaboration time, with a null main program, although stylistically this
27863 would usually be considered an inappropriate way to structure
27866 An important concern arises in the context of elaboration code:
27867 we have to be sure that it is executed in an appropriate order. What we
27868 have is a series of elaboration code sections, potentially one section
27869 for each unit in the program. It is important that these execute
27870 in the correct order. Correctness here means that, taking the above
27871 example of the declaration of @code{Sqrt_Half},
27872 if some other piece of
27873 elaboration code references @code{Sqrt_Half},
27874 then it must run after the
27875 section of elaboration code that contains the declaration of
27878 There would never be any order of elaboration problem if we made a rule
27879 that whenever you @code{with} a unit, you must elaborate both the spec and body
27880 of that unit before elaborating the unit doing the @code{with}'ing:
27882 @smallexample @c ada
27886 package Unit_2 is @dots{}
27892 would require that both the body and spec of @code{Unit_1} be elaborated
27893 before the spec of @code{Unit_2}. However, a rule like that would be far too
27894 restrictive. In particular, it would make it impossible to have routines
27895 in separate packages that were mutually recursive.
27897 You might think that a clever enough compiler could look at the actual
27898 elaboration code and determine an appropriate correct order of elaboration,
27899 but in the general case, this is not possible. Consider the following
27902 In the body of @code{Unit_1}, we have a procedure @code{Func_1}
27904 the variable @code{Sqrt_1}, which is declared in the elaboration code
27905 of the body of @code{Unit_1}:
27907 @smallexample @c ada
27909 Sqrt_1 : Float := Sqrt (0.1);
27914 The elaboration code of the body of @code{Unit_1} also contains:
27916 @smallexample @c ada
27919 if expression_1 = 1 then
27920 Q := Unit_2.Func_2;
27927 @code{Unit_2} is exactly parallel,
27928 it has a procedure @code{Func_2} that references
27929 the variable @code{Sqrt_2}, which is declared in the elaboration code of
27930 the body @code{Unit_2}:
27932 @smallexample @c ada
27934 Sqrt_2 : Float := Sqrt (0.1);
27939 The elaboration code of the body of @code{Unit_2} also contains:
27941 @smallexample @c ada
27944 if expression_2 = 2 then
27945 Q := Unit_1.Func_1;
27952 Now the question is, which of the following orders of elaboration is
27977 If you carefully analyze the flow here, you will see that you cannot tell
27978 at compile time the answer to this question.
27979 If @code{expression_1} is not equal to 1,
27980 and @code{expression_2} is not equal to 2,
27981 then either order is acceptable, because neither of the function calls is
27982 executed. If both tests evaluate to true, then neither order is acceptable
27983 and in fact there is no correct order.
27985 If one of the two expressions is true, and the other is false, then one
27986 of the above orders is correct, and the other is incorrect. For example,
27987 if @code{expression_1} /= 1 and @code{expression_2} = 2,
27988 then the call to @code{Func_1}
27989 will occur, but not the call to @code{Func_2.}
27990 This means that it is essential
27991 to elaborate the body of @code{Unit_1} before
27992 the body of @code{Unit_2}, so the first
27993 order of elaboration is correct and the second is wrong.
27995 By making @code{expression_1} and @code{expression_2}
27996 depend on input data, or perhaps
27997 the time of day, we can make it impossible for the compiler or binder
27998 to figure out which of these expressions will be true, and hence it
27999 is impossible to guarantee a safe order of elaboration at run time.
28001 @node Checking the Elaboration Order
28002 @section Checking the Elaboration Order
28005 In some languages that involve the same kind of elaboration problems,
28006 e.g.@: Java and C++, the programmer is expected to worry about these
28007 ordering problems himself, and it is common to
28008 write a program in which an incorrect elaboration order gives
28009 surprising results, because it references variables before they
28011 Ada is designed to be a safe language, and a programmer-beware approach is
28012 clearly not sufficient. Consequently, the language provides three lines
28016 @item Standard rules
28017 Some standard rules restrict the possible choice of elaboration
28018 order. In particular, if you @code{with} a unit, then its spec is always
28019 elaborated before the unit doing the @code{with}. Similarly, a parent
28020 spec is always elaborated before the child spec, and finally
28021 a spec is always elaborated before its corresponding body.
28023 @item Dynamic elaboration checks
28024 @cindex Elaboration checks
28025 @cindex Checks, elaboration
28026 Dynamic checks are made at run time, so that if some entity is accessed
28027 before it is elaborated (typically by means of a subprogram call)
28028 then the exception (@code{Program_Error}) is raised.
28030 @item Elaboration control
28031 Facilities are provided for the programmer to specify the desired order
28035 Let's look at these facilities in more detail. First, the rules for
28036 dynamic checking. One possible rule would be simply to say that the
28037 exception is raised if you access a variable which has not yet been
28038 elaborated. The trouble with this approach is that it could require
28039 expensive checks on every variable reference. Instead Ada has two
28040 rules which are a little more restrictive, but easier to check, and
28044 @item Restrictions on calls
28045 A subprogram can only be called at elaboration time if its body
28046 has been elaborated. The rules for elaboration given above guarantee
28047 that the spec of the subprogram has been elaborated before the
28048 call, but not the body. If this rule is violated, then the
28049 exception @code{Program_Error} is raised.
28051 @item Restrictions on instantiations
28052 A generic unit can only be instantiated if the body of the generic
28053 unit has been elaborated. Again, the rules for elaboration given above
28054 guarantee that the spec of the generic unit has been elaborated
28055 before the instantiation, but not the body. If this rule is
28056 violated, then the exception @code{Program_Error} is raised.
28060 The idea is that if the body has been elaborated, then any variables
28061 it references must have been elaborated; by checking for the body being
28062 elaborated we guarantee that none of its references causes any
28063 trouble. As we noted above, this is a little too restrictive, because a
28064 subprogram that has no non-local references in its body may in fact be safe
28065 to call. However, it really would be unsafe to rely on this, because
28066 it would mean that the caller was aware of details of the implementation
28067 in the body. This goes against the basic tenets of Ada.
28069 A plausible implementation can be described as follows.
28070 A Boolean variable is associated with each subprogram
28071 and each generic unit. This variable is initialized to False, and is set to
28072 True at the point body is elaborated. Every call or instantiation checks the
28073 variable, and raises @code{Program_Error} if the variable is False.
28075 Note that one might think that it would be good enough to have one Boolean
28076 variable for each package, but that would not deal with cases of trying
28077 to call a body in the same package as the call
28078 that has not been elaborated yet.
28079 Of course a compiler may be able to do enough analysis to optimize away
28080 some of the Boolean variables as unnecessary, and @code{GNAT} indeed
28081 does such optimizations, but still the easiest conceptual model is to
28082 think of there being one variable per subprogram.
28084 @node Controlling the Elaboration Order
28085 @section Controlling the Elaboration Order
28088 In the previous section we discussed the rules in Ada which ensure
28089 that @code{Program_Error} is raised if an incorrect elaboration order is
28090 chosen. This prevents erroneous executions, but we need mechanisms to
28091 specify a correct execution and avoid the exception altogether.
28092 To achieve this, Ada provides a number of features for controlling
28093 the order of elaboration. We discuss these features in this section.
28095 First, there are several ways of indicating to the compiler that a given
28096 unit has no elaboration problems:
28099 @item packages that do not require a body
28100 A library package that does not require a body does not permit
28101 a body (this rule was introduced in Ada 95).
28102 Thus if we have a such a package, as in:
28104 @smallexample @c ada
28107 package Definitions is
28109 type m is new integer;
28111 type a is array (1 .. 10) of m;
28112 type b is array (1 .. 20) of m;
28120 A package that @code{with}'s @code{Definitions} may safely instantiate
28121 @code{Definitions.Subp} because the compiler can determine that there
28122 definitely is no package body to worry about in this case
28125 @cindex pragma Pure
28127 Places sufficient restrictions on a unit to guarantee that
28128 no call to any subprogram in the unit can result in an
28129 elaboration problem. This means that the compiler does not need
28130 to worry about the point of elaboration of such units, and in
28131 particular, does not need to check any calls to any subprograms
28134 @item pragma Preelaborate
28135 @findex Preelaborate
28136 @cindex pragma Preelaborate
28137 This pragma places slightly less stringent restrictions on a unit than
28139 but these restrictions are still sufficient to ensure that there
28140 are no elaboration problems with any calls to the unit.
28142 @item pragma Elaborate_Body
28143 @findex Elaborate_Body
28144 @cindex pragma Elaborate_Body
28145 This pragma requires that the body of a unit be elaborated immediately
28146 after its spec. Suppose a unit @code{A} has such a pragma,
28147 and unit @code{B} does
28148 a @code{with} of unit @code{A}. Recall that the standard rules require
28149 the spec of unit @code{A}
28150 to be elaborated before the @code{with}'ing unit; given the pragma in
28151 @code{A}, we also know that the body of @code{A}
28152 will be elaborated before @code{B}, so
28153 that calls to @code{A} are safe and do not need a check.
28158 unlike pragma @code{Pure} and pragma @code{Preelaborate},
28160 @code{Elaborate_Body} does not guarantee that the program is
28161 free of elaboration problems, because it may not be possible
28162 to satisfy the requested elaboration order.
28163 Let's go back to the example with @code{Unit_1} and @code{Unit_2}.
28165 marks @code{Unit_1} as @code{Elaborate_Body},
28166 and not @code{Unit_2,} then the order of
28167 elaboration will be:
28179 Now that means that the call to @code{Func_1} in @code{Unit_2}
28180 need not be checked,
28181 it must be safe. But the call to @code{Func_2} in
28182 @code{Unit_1} may still fail if
28183 @code{Expression_1} is equal to 1,
28184 and the programmer must still take
28185 responsibility for this not being the case.
28187 If all units carry a pragma @code{Elaborate_Body}, then all problems are
28188 eliminated, except for calls entirely within a body, which are
28189 in any case fully under programmer control. However, using the pragma
28190 everywhere is not always possible.
28191 In particular, for our @code{Unit_1}/@code{Unit_2} example, if
28192 we marked both of them as having pragma @code{Elaborate_Body}, then
28193 clearly there would be no possible elaboration order.
28195 The above pragmas allow a server to guarantee safe use by clients, and
28196 clearly this is the preferable approach. Consequently a good rule
28197 is to mark units as @code{Pure} or @code{Preelaborate} if possible,
28198 and if this is not possible,
28199 mark them as @code{Elaborate_Body} if possible.
28200 As we have seen, there are situations where neither of these
28201 three pragmas can be used.
28202 So we also provide methods for clients to control the
28203 order of elaboration of the servers on which they depend:
28206 @item pragma Elaborate (unit)
28208 @cindex pragma Elaborate
28209 This pragma is placed in the context clause, after a @code{with} clause,
28210 and it requires that the body of the named unit be elaborated before
28211 the unit in which the pragma occurs. The idea is to use this pragma
28212 if the current unit calls at elaboration time, directly or indirectly,
28213 some subprogram in the named unit.
28215 @item pragma Elaborate_All (unit)
28216 @findex Elaborate_All
28217 @cindex pragma Elaborate_All
28218 This is a stronger version of the Elaborate pragma. Consider the
28222 Unit A @code{with}'s unit B and calls B.Func in elab code
28223 Unit B @code{with}'s unit C, and B.Func calls C.Func
28227 Now if we put a pragma @code{Elaborate (B)}
28228 in unit @code{A}, this ensures that the
28229 body of @code{B} is elaborated before the call, but not the
28230 body of @code{C}, so
28231 the call to @code{C.Func} could still cause @code{Program_Error} to
28234 The effect of a pragma @code{Elaborate_All} is stronger, it requires
28235 not only that the body of the named unit be elaborated before the
28236 unit doing the @code{with}, but also the bodies of all units that the
28237 named unit uses, following @code{with} links transitively. For example,
28238 if we put a pragma @code{Elaborate_All (B)} in unit @code{A},
28240 not only that the body of @code{B} be elaborated before @code{A},
28242 body of @code{C}, because @code{B} @code{with}'s @code{C}.
28246 We are now in a position to give a usage rule in Ada for avoiding
28247 elaboration problems, at least if dynamic dispatching and access to
28248 subprogram values are not used. We will handle these cases separately
28251 The rule is simple. If a unit has elaboration code that can directly or
28252 indirectly make a call to a subprogram in a @code{with}'ed unit, or instantiate
28253 a generic package in a @code{with}'ed unit,
28254 then if the @code{with}'ed unit does not have
28255 pragma @code{Pure} or @code{Preelaborate}, then the client should have
28256 a pragma @code{Elaborate_All}
28257 for the @code{with}'ed unit. By following this rule a client is
28258 assured that calls can be made without risk of an exception.
28260 For generic subprogram instantiations, the rule can be relaxed to
28261 require only a pragma @code{Elaborate} since elaborating the body
28262 of a subprogram cannot cause any transitive elaboration (we are
28263 not calling the subprogram in this case, just elaborating its
28266 If this rule is not followed, then a program may be in one of four
28270 @item No order exists
28271 No order of elaboration exists which follows the rules, taking into
28272 account any @code{Elaborate}, @code{Elaborate_All},
28273 or @code{Elaborate_Body} pragmas. In
28274 this case, an Ada compiler must diagnose the situation at bind
28275 time, and refuse to build an executable program.
28277 @item One or more orders exist, all incorrect
28278 One or more acceptable elaboration orders exist, and all of them
28279 generate an elaboration order problem. In this case, the binder
28280 can build an executable program, but @code{Program_Error} will be raised
28281 when the program is run.
28283 @item Several orders exist, some right, some incorrect
28284 One or more acceptable elaboration orders exists, and some of them
28285 work, and some do not. The programmer has not controlled
28286 the order of elaboration, so the binder may or may not pick one of
28287 the correct orders, and the program may or may not raise an
28288 exception when it is run. This is the worst case, because it means
28289 that the program may fail when moved to another compiler, or even
28290 another version of the same compiler.
28292 @item One or more orders exists, all correct
28293 One ore more acceptable elaboration orders exist, and all of them
28294 work. In this case the program runs successfully. This state of
28295 affairs can be guaranteed by following the rule we gave above, but
28296 may be true even if the rule is not followed.
28300 Note that one additional advantage of following our rules on the use
28301 of @code{Elaborate} and @code{Elaborate_All}
28302 is that the program continues to stay in the ideal (all orders OK) state
28303 even if maintenance
28304 changes some bodies of some units. Conversely, if a program that does
28305 not follow this rule happens to be safe at some point, this state of affairs
28306 may deteriorate silently as a result of maintenance changes.
28308 You may have noticed that the above discussion did not mention
28309 the use of @code{Elaborate_Body}. This was a deliberate omission. If you
28310 @code{with} an @code{Elaborate_Body} unit, it still may be the case that
28311 code in the body makes calls to some other unit, so it is still necessary
28312 to use @code{Elaborate_All} on such units.
28314 @node Controlling Elaboration in GNAT - Internal Calls
28315 @section Controlling Elaboration in GNAT - Internal Calls
28318 In the case of internal calls, i.e., calls within a single package, the
28319 programmer has full control over the order of elaboration, and it is up
28320 to the programmer to elaborate declarations in an appropriate order. For
28323 @smallexample @c ada
28326 function One return Float;
28330 function One return Float is
28339 will obviously raise @code{Program_Error} at run time, because function
28340 One will be called before its body is elaborated. In this case GNAT will
28341 generate a warning that the call will raise @code{Program_Error}:
28347 2. function One return Float;
28349 4. Q : Float := One;
28351 >>> warning: cannot call "One" before body is elaborated
28352 >>> warning: Program_Error will be raised at run time
28355 6. function One return Float is
28368 Note that in this particular case, it is likely that the call is safe, because
28369 the function @code{One} does not access any global variables.
28370 Nevertheless in Ada, we do not want the validity of the check to depend on
28371 the contents of the body (think about the separate compilation case), so this
28372 is still wrong, as we discussed in the previous sections.
28374 The error is easily corrected by rearranging the declarations so that the
28375 body of @code{One} appears before the declaration containing the call
28376 (note that in Ada 95 and Ada 2005,
28377 declarations can appear in any order, so there is no restriction that
28378 would prevent this reordering, and if we write:
28380 @smallexample @c ada
28383 function One return Float;
28385 function One return Float is
28396 then all is well, no warning is generated, and no
28397 @code{Program_Error} exception
28399 Things are more complicated when a chain of subprograms is executed:
28401 @smallexample @c ada
28404 function A return Integer;
28405 function B return Integer;
28406 function C return Integer;
28408 function B return Integer is begin return A; end;
28409 function C return Integer is begin return B; end;
28413 function A return Integer is begin return 1; end;
28419 Now the call to @code{C}
28420 at elaboration time in the declaration of @code{X} is correct, because
28421 the body of @code{C} is already elaborated,
28422 and the call to @code{B} within the body of
28423 @code{C} is correct, but the call
28424 to @code{A} within the body of @code{B} is incorrect, because the body
28425 of @code{A} has not been elaborated, so @code{Program_Error}
28426 will be raised on the call to @code{A}.
28427 In this case GNAT will generate a
28428 warning that @code{Program_Error} may be
28429 raised at the point of the call. Let's look at the warning:
28435 2. function A return Integer;
28436 3. function B return Integer;
28437 4. function C return Integer;
28439 6. function B return Integer is begin return A; end;
28441 >>> warning: call to "A" before body is elaborated may
28442 raise Program_Error
28443 >>> warning: "B" called at line 7
28444 >>> warning: "C" called at line 9
28446 7. function C return Integer is begin return B; end;
28448 9. X : Integer := C;
28450 11. function A return Integer is begin return 1; end;
28460 Note that the message here says ``may raise'', instead of the direct case,
28461 where the message says ``will be raised''. That's because whether
28463 actually called depends in general on run-time flow of control.
28464 For example, if the body of @code{B} said
28466 @smallexample @c ada
28469 function B return Integer is
28471 if some-condition-depending-on-input-data then
28482 then we could not know until run time whether the incorrect call to A would
28483 actually occur, so @code{Program_Error} might
28484 or might not be raised. It is possible for a compiler to
28485 do a better job of analyzing bodies, to
28486 determine whether or not @code{Program_Error}
28487 might be raised, but it certainly
28488 couldn't do a perfect job (that would require solving the halting problem
28489 and is provably impossible), and because this is a warning anyway, it does
28490 not seem worth the effort to do the analysis. Cases in which it
28491 would be relevant are rare.
28493 In practice, warnings of either of the forms given
28494 above will usually correspond to
28495 real errors, and should be examined carefully and eliminated.
28496 In the rare case where a warning is bogus, it can be suppressed by any of
28497 the following methods:
28501 Compile with the @option{-gnatws} switch set
28504 Suppress @code{Elaboration_Check} for the called subprogram
28507 Use pragma @code{Warnings_Off} to turn warnings off for the call
28511 For the internal elaboration check case,
28512 GNAT by default generates the
28513 necessary run-time checks to ensure
28514 that @code{Program_Error} is raised if any
28515 call fails an elaboration check. Of course this can only happen if a
28516 warning has been issued as described above. The use of pragma
28517 @code{Suppress (Elaboration_Check)} may (but is not guaranteed to) suppress
28518 some of these checks, meaning that it may be possible (but is not
28519 guaranteed) for a program to be able to call a subprogram whose body
28520 is not yet elaborated, without raising a @code{Program_Error} exception.
28522 @node Controlling Elaboration in GNAT - External Calls
28523 @section Controlling Elaboration in GNAT - External Calls
28526 The previous section discussed the case in which the execution of a
28527 particular thread of elaboration code occurred entirely within a
28528 single unit. This is the easy case to handle, because a programmer
28529 has direct and total control over the order of elaboration, and
28530 furthermore, checks need only be generated in cases which are rare
28531 and which the compiler can easily detect.
28532 The situation is more complex when separate compilation is taken into account.
28533 Consider the following:
28535 @smallexample @c ada
28539 function Sqrt (Arg : Float) return Float;
28542 package body Math is
28543 function Sqrt (Arg : Float) return Float is
28552 X : Float := Math.Sqrt (0.5);
28565 where @code{Main} is the main program. When this program is executed, the
28566 elaboration code must first be executed, and one of the jobs of the
28567 binder is to determine the order in which the units of a program are
28568 to be elaborated. In this case we have four units: the spec and body
28570 the spec of @code{Stuff} and the body of @code{Main}).
28571 In what order should the four separate sections of elaboration code
28574 There are some restrictions in the order of elaboration that the binder
28575 can choose. In particular, if unit U has a @code{with}
28576 for a package @code{X}, then you
28577 are assured that the spec of @code{X}
28578 is elaborated before U , but you are
28579 not assured that the body of @code{X}
28580 is elaborated before U.
28581 This means that in the above case, the binder is allowed to choose the
28592 but that's not good, because now the call to @code{Math.Sqrt}
28593 that happens during
28594 the elaboration of the @code{Stuff}
28595 spec happens before the body of @code{Math.Sqrt} is
28596 elaborated, and hence causes @code{Program_Error} exception to be raised.
28597 At first glance, one might say that the binder is misbehaving, because
28598 obviously you want to elaborate the body of something you @code{with}
28600 that is not a general rule that can be followed in all cases. Consider
28602 @smallexample @c ada
28605 package X is @dots{}
28607 package Y is @dots{}
28610 package body Y is @dots{}
28613 package body X is @dots{}
28619 This is a common arrangement, and, apart from the order of elaboration
28620 problems that might arise in connection with elaboration code, this works fine.
28621 A rule that says that you must first elaborate the body of anything you
28622 @code{with} cannot work in this case:
28623 the body of @code{X} @code{with}'s @code{Y},
28624 which means you would have to
28625 elaborate the body of @code{Y} first, but that @code{with}'s @code{X},
28627 you have to elaborate the body of @code{X} first, but @dots{} and we have a
28628 loop that cannot be broken.
28630 It is true that the binder can in many cases guess an order of elaboration
28631 that is unlikely to cause a @code{Program_Error}
28632 exception to be raised, and it tries to do so (in the
28633 above example of @code{Math/Stuff/Spec}, the GNAT binder will
28635 elaborate the body of @code{Math} right after its spec, so all will be well).
28637 However, a program that blindly relies on the binder to be helpful can
28638 get into trouble, as we discussed in the previous sections, so
28640 provides a number of facilities for assisting the programmer in
28641 developing programs that are robust with respect to elaboration order.
28643 @node Default Behavior in GNAT - Ensuring Safety
28644 @section Default Behavior in GNAT - Ensuring Safety
28647 The default behavior in GNAT ensures elaboration safety. In its
28648 default mode GNAT implements the
28649 rule we previously described as the right approach. Let's restate it:
28653 @emph{If a unit has elaboration code that can directly or indirectly make a
28654 call to a subprogram in a @code{with}'ed unit, or instantiate a generic
28655 package in a @code{with}'ed unit, then if the @code{with}'ed unit
28656 does not have pragma @code{Pure} or
28657 @code{Preelaborate}, then the client should have an
28658 @code{Elaborate_All} pragma for the @code{with}'ed unit.}
28660 @emph{In the case of instantiating a generic subprogram, it is always
28661 sufficient to have only an @code{Elaborate} pragma for the
28662 @code{with}'ed unit.}
28666 By following this rule a client is assured that calls and instantiations
28667 can be made without risk of an exception.
28669 In this mode GNAT traces all calls that are potentially made from
28670 elaboration code, and puts in any missing implicit @code{Elaborate}
28671 and @code{Elaborate_All} pragmas.
28672 The advantage of this approach is that no elaboration problems
28673 are possible if the binder can find an elaboration order that is
28674 consistent with these implicit @code{Elaborate} and
28675 @code{Elaborate_All} pragmas. The
28676 disadvantage of this approach is that no such order may exist.
28678 If the binder does not generate any diagnostics, then it means that it has
28679 found an elaboration order that is guaranteed to be safe. However, the binder
28680 may still be relying on implicitly generated @code{Elaborate} and
28681 @code{Elaborate_All} pragmas so portability to other compilers than GNAT is not
28684 If it is important to guarantee portability, then the compilations should
28687 (warn on elaboration problems) switch. This will cause warning messages
28688 to be generated indicating the missing @code{Elaborate} and
28689 @code{Elaborate_All} pragmas.
28690 Consider the following source program:
28692 @smallexample @c ada
28697 m : integer := k.r;
28704 where it is clear that there
28705 should be a pragma @code{Elaborate_All}
28706 for unit @code{k}. An implicit pragma will be generated, and it is
28707 likely that the binder will be able to honor it. However, if you want
28708 to port this program to some other Ada compiler than GNAT.
28709 it is safer to include the pragma explicitly in the source. If this
28710 unit is compiled with the
28712 switch, then the compiler outputs a warning:
28719 3. m : integer := k.r;
28721 >>> warning: call to "r" may raise Program_Error
28722 >>> warning: missing pragma Elaborate_All for "k"
28730 and these warnings can be used as a guide for supplying manually
28731 the missing pragmas. It is usually a bad idea to use this warning
28732 option during development. That's because it will warn you when
28733 you need to put in a pragma, but cannot warn you when it is time
28734 to take it out. So the use of pragma @code{Elaborate_All} may lead to
28735 unnecessary dependencies and even false circularities.
28737 This default mode is more restrictive than the Ada Reference
28738 Manual, and it is possible to construct programs which will compile
28739 using the dynamic model described there, but will run into a
28740 circularity using the safer static model we have described.
28742 Of course any Ada compiler must be able to operate in a mode
28743 consistent with the requirements of the Ada Reference Manual,
28744 and in particular must have the capability of implementing the
28745 standard dynamic model of elaboration with run-time checks.
28747 In GNAT, this standard mode can be achieved either by the use of
28748 the @option{-gnatE} switch on the compiler (@command{gcc} or
28749 @command{gnatmake}) command, or by the use of the configuration pragma:
28751 @smallexample @c ada
28752 pragma Elaboration_Checks (DYNAMIC);
28756 Either approach will cause the unit affected to be compiled using the
28757 standard dynamic run-time elaboration checks described in the Ada
28758 Reference Manual. The static model is generally preferable, since it
28759 is clearly safer to rely on compile and link time checks rather than
28760 run-time checks. However, in the case of legacy code, it may be
28761 difficult to meet the requirements of the static model. This
28762 issue is further discussed in
28763 @ref{What to Do If the Default Elaboration Behavior Fails}.
28765 Note that the static model provides a strict subset of the allowed
28766 behavior and programs of the Ada Reference Manual, so if you do
28767 adhere to the static model and no circularities exist,
28768 then you are assured that your program will
28769 work using the dynamic model, providing that you remove any
28770 pragma Elaborate statements from the source.
28772 @node Treatment of Pragma Elaborate
28773 @section Treatment of Pragma Elaborate
28774 @cindex Pragma Elaborate
28777 The use of @code{pragma Elaborate}
28778 should generally be avoided in Ada 95 and Ada 2005 programs,
28779 since there is no guarantee that transitive calls
28780 will be properly handled. Indeed at one point, this pragma was placed
28781 in Annex J (Obsolescent Features), on the grounds that it is never useful.
28783 Now that's a bit restrictive. In practice, the case in which
28784 @code{pragma Elaborate} is useful is when the caller knows that there
28785 are no transitive calls, or that the called unit contains all necessary
28786 transitive @code{pragma Elaborate} statements, and legacy code often
28787 contains such uses.
28789 Strictly speaking the static mode in GNAT should ignore such pragmas,
28790 since there is no assurance at compile time that the necessary safety
28791 conditions are met. In practice, this would cause GNAT to be incompatible
28792 with correctly written Ada 83 code that had all necessary
28793 @code{pragma Elaborate} statements in place. Consequently, we made the
28794 decision that GNAT in its default mode will believe that if it encounters
28795 a @code{pragma Elaborate} then the programmer knows what they are doing,
28796 and it will trust that no elaboration errors can occur.
28798 The result of this decision is two-fold. First to be safe using the
28799 static mode, you should remove all @code{pragma Elaborate} statements.
28800 Second, when fixing circularities in existing code, you can selectively
28801 use @code{pragma Elaborate} statements to convince the static mode of
28802 GNAT that it need not generate an implicit @code{pragma Elaborate_All}
28805 When using the static mode with @option{-gnatwl}, any use of
28806 @code{pragma Elaborate} will generate a warning about possible
28809 @node Elaboration Issues for Library Tasks
28810 @section Elaboration Issues for Library Tasks
28811 @cindex Library tasks, elaboration issues
28812 @cindex Elaboration of library tasks
28815 In this section we examine special elaboration issues that arise for
28816 programs that declare library level tasks.
28818 Generally the model of execution of an Ada program is that all units are
28819 elaborated, and then execution of the program starts. However, the
28820 declaration of library tasks definitely does not fit this model. The
28821 reason for this is that library tasks start as soon as they are declared
28822 (more precisely, as soon as the statement part of the enclosing package
28823 body is reached), that is to say before elaboration
28824 of the program is complete. This means that if such a task calls a
28825 subprogram, or an entry in another task, the callee may or may not be
28826 elaborated yet, and in the standard
28827 Reference Manual model of dynamic elaboration checks, you can even
28828 get timing dependent Program_Error exceptions, since there can be
28829 a race between the elaboration code and the task code.
28831 The static model of elaboration in GNAT seeks to avoid all such
28832 dynamic behavior, by being conservative, and the conservative
28833 approach in this particular case is to assume that all the code
28834 in a task body is potentially executed at elaboration time if
28835 a task is declared at the library level.
28837 This can definitely result in unexpected circularities. Consider
28838 the following example
28840 @smallexample @c ada
28846 type My_Int is new Integer;
28848 function Ident (M : My_Int) return My_Int;
28852 package body Decls is
28853 task body Lib_Task is
28859 function Ident (M : My_Int) return My_Int is
28867 procedure Put_Val (Arg : Decls.My_Int);
28871 package body Utils is
28872 procedure Put_Val (Arg : Decls.My_Int) is
28874 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
28881 Decls.Lib_Task.Start;
28886 If the above example is compiled in the default static elaboration
28887 mode, then a circularity occurs. The circularity comes from the call
28888 @code{Utils.Put_Val} in the task body of @code{Decls.Lib_Task}. Since
28889 this call occurs in elaboration code, we need an implicit pragma
28890 @code{Elaborate_All} for @code{Utils}. This means that not only must
28891 the spec and body of @code{Utils} be elaborated before the body
28892 of @code{Decls}, but also the spec and body of any unit that is
28893 @code{with'ed} by the body of @code{Utils} must also be elaborated before
28894 the body of @code{Decls}. This is the transitive implication of
28895 pragma @code{Elaborate_All} and it makes sense, because in general
28896 the body of @code{Put_Val} might have a call to something in a
28897 @code{with'ed} unit.
28899 In this case, the body of Utils (actually its spec) @code{with's}
28900 @code{Decls}. Unfortunately this means that the body of @code{Decls}
28901 must be elaborated before itself, in case there is a call from the
28902 body of @code{Utils}.
28904 Here is the exact chain of events we are worrying about:
28908 In the body of @code{Decls} a call is made from within the body of a library
28909 task to a subprogram in the package @code{Utils}. Since this call may
28910 occur at elaboration time (given that the task is activated at elaboration
28911 time), we have to assume the worst, i.e., that the
28912 call does happen at elaboration time.
28915 This means that the body and spec of @code{Util} must be elaborated before
28916 the body of @code{Decls} so that this call does not cause an access before
28920 Within the body of @code{Util}, specifically within the body of
28921 @code{Util.Put_Val} there may be calls to any unit @code{with}'ed
28925 One such @code{with}'ed package is package @code{Decls}, so there
28926 might be a call to a subprogram in @code{Decls} in @code{Put_Val}.
28927 In fact there is such a call in this example, but we would have to
28928 assume that there was such a call even if it were not there, since
28929 we are not supposed to write the body of @code{Decls} knowing what
28930 is in the body of @code{Utils}; certainly in the case of the
28931 static elaboration model, the compiler does not know what is in
28932 other bodies and must assume the worst.
28935 This means that the spec and body of @code{Decls} must also be
28936 elaborated before we elaborate the unit containing the call, but
28937 that unit is @code{Decls}! This means that the body of @code{Decls}
28938 must be elaborated before itself, and that's a circularity.
28942 Indeed, if you add an explicit pragma @code{Elaborate_All} for @code{Utils} in
28943 the body of @code{Decls} you will get a true Ada Reference Manual
28944 circularity that makes the program illegal.
28946 In practice, we have found that problems with the static model of
28947 elaboration in existing code often arise from library tasks, so
28948 we must address this particular situation.
28950 Note that if we compile and run the program above, using the dynamic model of
28951 elaboration (that is to say use the @option{-gnatE} switch),
28952 then it compiles, binds,
28953 links, and runs, printing the expected result of 2. Therefore in some sense
28954 the circularity here is only apparent, and we need to capture
28955 the properties of this program that distinguish it from other library-level
28956 tasks that have real elaboration problems.
28958 We have four possible answers to this question:
28963 Use the dynamic model of elaboration.
28965 If we use the @option{-gnatE} switch, then as noted above, the program works.
28966 Why is this? If we examine the task body, it is apparent that the task cannot
28968 @code{accept} statement until after elaboration has been completed, because
28969 the corresponding entry call comes from the main program, not earlier.
28970 This is why the dynamic model works here. But that's really giving
28971 up on a precise analysis, and we prefer to take this approach only if we cannot
28973 problem in any other manner. So let us examine two ways to reorganize
28974 the program to avoid the potential elaboration problem.
28977 Split library tasks into separate packages.
28979 Write separate packages, so that library tasks are isolated from
28980 other declarations as much as possible. Let us look at a variation on
28983 @smallexample @c ada
28991 package body Decls1 is
28992 task body Lib_Task is
29000 type My_Int is new Integer;
29001 function Ident (M : My_Int) return My_Int;
29005 package body Decls2 is
29006 function Ident (M : My_Int) return My_Int is
29014 procedure Put_Val (Arg : Decls2.My_Int);
29018 package body Utils is
29019 procedure Put_Val (Arg : Decls2.My_Int) is
29021 Text_IO.Put_Line (Decls2.My_Int'Image (Decls2.Ident (Arg)));
29028 Decls1.Lib_Task.Start;
29033 All we have done is to split @code{Decls} into two packages, one
29034 containing the library task, and one containing everything else. Now
29035 there is no cycle, and the program compiles, binds, links and executes
29036 using the default static model of elaboration.
29039 Declare separate task types.
29041 A significant part of the problem arises because of the use of the
29042 single task declaration form. This means that the elaboration of
29043 the task type, and the elaboration of the task itself (i.e.@: the
29044 creation of the task) happen at the same time. A good rule
29045 of style in Ada is to always create explicit task types. By
29046 following the additional step of placing task objects in separate
29047 packages from the task type declaration, many elaboration problems
29048 are avoided. Here is another modified example of the example program:
29050 @smallexample @c ada
29052 task type Lib_Task_Type is
29056 type My_Int is new Integer;
29058 function Ident (M : My_Int) return My_Int;
29062 package body Decls is
29063 task body Lib_Task_Type is
29069 function Ident (M : My_Int) return My_Int is
29077 procedure Put_Val (Arg : Decls.My_Int);
29081 package body Utils is
29082 procedure Put_Val (Arg : Decls.My_Int) is
29084 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
29090 Lib_Task : Decls.Lib_Task_Type;
29096 Declst.Lib_Task.Start;
29101 What we have done here is to replace the @code{task} declaration in
29102 package @code{Decls} with a @code{task type} declaration. Then we
29103 introduce a separate package @code{Declst} to contain the actual
29104 task object. This separates the elaboration issues for
29105 the @code{task type}
29106 declaration, which causes no trouble, from the elaboration issues
29107 of the task object, which is also unproblematic, since it is now independent
29108 of the elaboration of @code{Utils}.
29109 This separation of concerns also corresponds to
29110 a generally sound engineering principle of separating declarations
29111 from instances. This version of the program also compiles, binds, links,
29112 and executes, generating the expected output.
29115 Use No_Entry_Calls_In_Elaboration_Code restriction.
29116 @cindex No_Entry_Calls_In_Elaboration_Code
29118 The previous two approaches described how a program can be restructured
29119 to avoid the special problems caused by library task bodies. in practice,
29120 however, such restructuring may be difficult to apply to existing legacy code,
29121 so we must consider solutions that do not require massive rewriting.
29123 Let us consider more carefully why our original sample program works
29124 under the dynamic model of elaboration. The reason is that the code
29125 in the task body blocks immediately on the @code{accept}
29126 statement. Now of course there is nothing to prohibit elaboration
29127 code from making entry calls (for example from another library level task),
29128 so we cannot tell in isolation that
29129 the task will not execute the accept statement during elaboration.
29131 However, in practice it is very unusual to see elaboration code
29132 make any entry calls, and the pattern of tasks starting
29133 at elaboration time and then immediately blocking on @code{accept} or
29134 @code{select} statements is very common. What this means is that
29135 the compiler is being too pessimistic when it analyzes the
29136 whole package body as though it might be executed at elaboration
29139 If we know that the elaboration code contains no entry calls, (a very safe
29140 assumption most of the time, that could almost be made the default
29141 behavior), then we can compile all units of the program under control
29142 of the following configuration pragma:
29145 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
29149 This pragma can be placed in the @file{gnat.adc} file in the usual
29150 manner. If we take our original unmodified program and compile it
29151 in the presence of a @file{gnat.adc} containing the above pragma,
29152 then once again, we can compile, bind, link, and execute, obtaining
29153 the expected result. In the presence of this pragma, the compiler does
29154 not trace calls in a task body, that appear after the first @code{accept}
29155 or @code{select} statement, and therefore does not report a potential
29156 circularity in the original program.
29158 The compiler will check to the extent it can that the above
29159 restriction is not violated, but it is not always possible to do a
29160 complete check at compile time, so it is important to use this
29161 pragma only if the stated restriction is in fact met, that is to say
29162 no task receives an entry call before elaboration of all units is completed.
29166 @node Mixing Elaboration Models
29167 @section Mixing Elaboration Models
29169 So far, we have assumed that the entire program is either compiled
29170 using the dynamic model or static model, ensuring consistency. It
29171 is possible to mix the two models, but rules have to be followed
29172 if this mixing is done to ensure that elaboration checks are not
29175 The basic rule is that @emph{a unit compiled with the static model cannot
29176 be @code{with'ed} by a unit compiled with the dynamic model}. The
29177 reason for this is that in the static model, a unit assumes that
29178 its clients guarantee to use (the equivalent of) pragma
29179 @code{Elaborate_All} so that no elaboration checks are required
29180 in inner subprograms, and this assumption is violated if the
29181 client is compiled with dynamic checks.
29183 The precise rule is as follows. A unit that is compiled with dynamic
29184 checks can only @code{with} a unit that meets at least one of the
29185 following criteria:
29190 The @code{with'ed} unit is itself compiled with dynamic elaboration
29191 checks (that is with the @option{-gnatE} switch.
29194 The @code{with'ed} unit is an internal GNAT implementation unit from
29195 the System, Interfaces, Ada, or GNAT hierarchies.
29198 The @code{with'ed} unit has pragma Preelaborate or pragma Pure.
29201 The @code{with'ing} unit (that is the client) has an explicit pragma
29202 @code{Elaborate_All} for the @code{with'ed} unit.
29207 If this rule is violated, that is if a unit with dynamic elaboration
29208 checks @code{with's} a unit that does not meet one of the above four
29209 criteria, then the binder (@code{gnatbind}) will issue a warning
29210 similar to that in the following example:
29213 warning: "x.ads" has dynamic elaboration checks and with's
29214 warning: "y.ads" which has static elaboration checks
29218 These warnings indicate that the rule has been violated, and that as a result
29219 elaboration checks may be missed in the resulting executable file.
29220 This warning may be suppressed using the @option{-ws} binder switch
29221 in the usual manner.
29223 One useful application of this mixing rule is in the case of a subsystem
29224 which does not itself @code{with} units from the remainder of the
29225 application. In this case, the entire subsystem can be compiled with
29226 dynamic checks to resolve a circularity in the subsystem, while
29227 allowing the main application that uses this subsystem to be compiled
29228 using the more reliable default static model.
29230 @node What to Do If the Default Elaboration Behavior Fails
29231 @section What to Do If the Default Elaboration Behavior Fails
29234 If the binder cannot find an acceptable order, it outputs detailed
29235 diagnostics. For example:
29241 error: elaboration circularity detected
29242 info: "proc (body)" must be elaborated before "pack (body)"
29243 info: reason: Elaborate_All probably needed in unit "pack (body)"
29244 info: recompile "pack (body)" with -gnatwl
29245 info: for full details
29246 info: "proc (body)"
29247 info: is needed by its spec:
29248 info: "proc (spec)"
29249 info: which is withed by:
29250 info: "pack (body)"
29251 info: "pack (body)" must be elaborated before "proc (body)"
29252 info: reason: pragma Elaborate in unit "proc (body)"
29258 In this case we have a cycle that the binder cannot break. On the one
29259 hand, there is an explicit pragma Elaborate in @code{proc} for
29260 @code{pack}. This means that the body of @code{pack} must be elaborated
29261 before the body of @code{proc}. On the other hand, there is elaboration
29262 code in @code{pack} that calls a subprogram in @code{proc}. This means
29263 that for maximum safety, there should really be a pragma
29264 Elaborate_All in @code{pack} for @code{proc} which would require that
29265 the body of @code{proc} be elaborated before the body of
29266 @code{pack}. Clearly both requirements cannot be satisfied.
29267 Faced with a circularity of this kind, you have three different options.
29270 @item Fix the program
29271 The most desirable option from the point of view of long-term maintenance
29272 is to rearrange the program so that the elaboration problems are avoided.
29273 One useful technique is to place the elaboration code into separate
29274 child packages. Another is to move some of the initialization code to
29275 explicitly called subprograms, where the program controls the order
29276 of initialization explicitly. Although this is the most desirable option,
29277 it may be impractical and involve too much modification, especially in
29278 the case of complex legacy code.
29280 @item Perform dynamic checks
29281 If the compilations are done using the
29283 (dynamic elaboration check) switch, then GNAT behaves in a quite different
29284 manner. Dynamic checks are generated for all calls that could possibly result
29285 in raising an exception. With this switch, the compiler does not generate
29286 implicit @code{Elaborate} or @code{Elaborate_All} pragmas. The behavior then is
29287 exactly as specified in the @cite{Ada Reference Manual}.
29288 The binder will generate
29289 an executable program that may or may not raise @code{Program_Error}, and then
29290 it is the programmer's job to ensure that it does not raise an exception. Note
29291 that it is important to compile all units with the switch, it cannot be used
29294 @item Suppress checks
29295 The drawback of dynamic checks is that they generate a
29296 significant overhead at run time, both in space and time. If you
29297 are absolutely sure that your program cannot raise any elaboration
29298 exceptions, and you still want to use the dynamic elaboration model,
29299 then you can use the configuration pragma
29300 @code{Suppress (Elaboration_Check)} to suppress all such checks. For
29301 example this pragma could be placed in the @file{gnat.adc} file.
29303 @item Suppress checks selectively
29304 When you know that certain calls or instantiations in elaboration code cannot
29305 possibly lead to an elaboration error, and the binder nevertheless complains
29306 about implicit @code{Elaborate} and @code{Elaborate_All} pragmas that lead to
29307 elaboration circularities, it is possible to remove those warnings locally and
29308 obtain a program that will bind. Clearly this can be unsafe, and it is the
29309 responsibility of the programmer to make sure that the resulting program has no
29310 elaboration anomalies. The pragma @code{Suppress (Elaboration_Check)} can be
29311 used with different granularity to suppress warnings and break elaboration
29316 Place the pragma that names the called subprogram in the declarative part
29317 that contains the call.
29320 Place the pragma in the declarative part, without naming an entity. This
29321 disables warnings on all calls in the corresponding declarative region.
29324 Place the pragma in the package spec that declares the called subprogram,
29325 and name the subprogram. This disables warnings on all elaboration calls to
29329 Place the pragma in the package spec that declares the called subprogram,
29330 without naming any entity. This disables warnings on all elaboration calls to
29331 all subprograms declared in this spec.
29333 @item Use Pragma Elaborate
29334 As previously described in section @xref{Treatment of Pragma Elaborate},
29335 GNAT in static mode assumes that a @code{pragma} Elaborate indicates correctly
29336 that no elaboration checks are required on calls to the designated unit.
29337 There may be cases in which the caller knows that no transitive calls
29338 can occur, so that a @code{pragma Elaborate} will be sufficient in a
29339 case where @code{pragma Elaborate_All} would cause a circularity.
29343 These five cases are listed in order of decreasing safety, and therefore
29344 require increasing programmer care in their application. Consider the
29347 @smallexample @c adanocomment
29349 function F1 return Integer;
29354 function F2 return Integer;
29355 function Pure (x : integer) return integer;
29356 -- pragma Suppress (Elaboration_Check, On => Pure); -- (3)
29357 -- pragma Suppress (Elaboration_Check); -- (4)
29361 package body Pack1 is
29362 function F1 return Integer is
29366 Val : integer := Pack2.Pure (11); -- Elab. call (1)
29369 -- pragma Suppress(Elaboration_Check, Pack2.F2); -- (1)
29370 -- pragma Suppress(Elaboration_Check); -- (2)
29372 X1 := Pack2.F2 + 1; -- Elab. call (2)
29377 package body Pack2 is
29378 function F2 return Integer is
29382 function Pure (x : integer) return integer is
29384 return x ** 3 - 3 * x;
29388 with Pack1, Ada.Text_IO;
29391 Ada.Text_IO.Put_Line(Pack1.X1'Img); -- 101
29394 In the absence of any pragmas, an attempt to bind this program produces
29395 the following diagnostics:
29401 error: elaboration circularity detected
29402 info: "pack1 (body)" must be elaborated before "pack1 (body)"
29403 info: reason: Elaborate_All probably needed in unit "pack1 (body)"
29404 info: recompile "pack1 (body)" with -gnatwl for full details
29405 info: "pack1 (body)"
29406 info: must be elaborated along with its spec:
29407 info: "pack1 (spec)"
29408 info: which is withed by:
29409 info: "pack2 (body)"
29410 info: which must be elaborated along with its spec:
29411 info: "pack2 (spec)"
29412 info: which is withed by:
29413 info: "pack1 (body)"
29416 The sources of the circularity are the two calls to @code{Pack2.Pure} and
29417 @code{Pack2.F2} in the body of @code{Pack1}. We can see that the call to
29418 F2 is safe, even though F2 calls F1, because the call appears after the
29419 elaboration of the body of F1. Therefore the pragma (1) is safe, and will
29420 remove the warning on the call. It is also possible to use pragma (2)
29421 because there are no other potentially unsafe calls in the block.
29424 The call to @code{Pure} is safe because this function does not depend on the
29425 state of @code{Pack2}. Therefore any call to this function is safe, and it
29426 is correct to place pragma (3) in the corresponding package spec.
29429 Finally, we could place pragma (4) in the spec of @code{Pack2} to disable
29430 warnings on all calls to functions declared therein. Note that this is not
29431 necessarily safe, and requires more detailed examination of the subprogram
29432 bodies involved. In particular, a call to @code{F2} requires that @code{F1}
29433 be already elaborated.
29437 It is hard to generalize on which of these four approaches should be
29438 taken. Obviously if it is possible to fix the program so that the default
29439 treatment works, this is preferable, but this may not always be practical.
29440 It is certainly simple enough to use
29442 but the danger in this case is that, even if the GNAT binder
29443 finds a correct elaboration order, it may not always do so,
29444 and certainly a binder from another Ada compiler might not. A
29445 combination of testing and analysis (for which the warnings generated
29448 switch can be useful) must be used to ensure that the program is free
29449 of errors. One switch that is useful in this testing is the
29450 @option{^-p (pessimistic elaboration order)^/PESSIMISTIC_ELABORATION_ORDER^}
29453 Normally the binder tries to find an order that has the best chance
29454 of avoiding elaboration problems. However, if this switch is used, the binder
29455 plays a devil's advocate role, and tries to choose the order that
29456 has the best chance of failing. If your program works even with this
29457 switch, then it has a better chance of being error free, but this is still
29460 For an example of this approach in action, consider the C-tests (executable
29461 tests) from the ACVC suite. If these are compiled and run with the default
29462 treatment, then all but one of them succeed without generating any error
29463 diagnostics from the binder. However, there is one test that fails, and
29464 this is not surprising, because the whole point of this test is to ensure
29465 that the compiler can handle cases where it is impossible to determine
29466 a correct order statically, and it checks that an exception is indeed
29467 raised at run time.
29469 This one test must be compiled and run using the
29471 switch, and then it passes. Alternatively, the entire suite can
29472 be run using this switch. It is never wrong to run with the dynamic
29473 elaboration switch if your code is correct, and we assume that the
29474 C-tests are indeed correct (it is less efficient, but efficiency is
29475 not a factor in running the ACVC tests.)
29477 @node Elaboration for Access-to-Subprogram Values
29478 @section Elaboration for Access-to-Subprogram Values
29479 @cindex Access-to-subprogram
29482 Access-to-subprogram types (introduced in Ada 95) complicate
29483 the handling of elaboration. The trouble is that it becomes
29484 impossible to tell at compile time which procedure
29485 is being called. This means that it is not possible for the binder
29486 to analyze the elaboration requirements in this case.
29488 If at the point at which the access value is created
29489 (i.e., the evaluation of @code{P'Access} for a subprogram @code{P}),
29490 the body of the subprogram is
29491 known to have been elaborated, then the access value is safe, and its use
29492 does not require a check. This may be achieved by appropriate arrangement
29493 of the order of declarations if the subprogram is in the current unit,
29494 or, if the subprogram is in another unit, by using pragma
29495 @code{Pure}, @code{Preelaborate}, or @code{Elaborate_Body}
29496 on the referenced unit.
29498 If the referenced body is not known to have been elaborated at the point
29499 the access value is created, then any use of the access value must do a
29500 dynamic check, and this dynamic check will fail and raise a
29501 @code{Program_Error} exception if the body has not been elaborated yet.
29502 GNAT will generate the necessary checks, and in addition, if the
29504 switch is set, will generate warnings that such checks are required.
29506 The use of dynamic dispatching for tagged types similarly generates
29507 a requirement for dynamic checks, and premature calls to any primitive
29508 operation of a tagged type before the body of the operation has been
29509 elaborated, will result in the raising of @code{Program_Error}.
29511 @node Summary of Procedures for Elaboration Control
29512 @section Summary of Procedures for Elaboration Control
29513 @cindex Elaboration control
29516 First, compile your program with the default options, using none of
29517 the special elaboration control switches. If the binder successfully
29518 binds your program, then you can be confident that, apart from issues
29519 raised by the use of access-to-subprogram types and dynamic dispatching,
29520 the program is free of elaboration errors. If it is important that the
29521 program be portable, then use the
29523 switch to generate warnings about missing @code{Elaborate} or
29524 @code{Elaborate_All} pragmas, and supply the missing pragmas.
29526 If the program fails to bind using the default static elaboration
29527 handling, then you can fix the program to eliminate the binder
29528 message, or recompile the entire program with the
29529 @option{-gnatE} switch to generate dynamic elaboration checks,
29530 and, if you are sure there really are no elaboration problems,
29531 use a global pragma @code{Suppress (Elaboration_Check)}.
29533 @node Other Elaboration Order Considerations
29534 @section Other Elaboration Order Considerations
29536 This section has been entirely concerned with the issue of finding a valid
29537 elaboration order, as defined by the Ada Reference Manual. In a case
29538 where several elaboration orders are valid, the task is to find one
29539 of the possible valid elaboration orders (and the static model in GNAT
29540 will ensure that this is achieved).
29542 The purpose of the elaboration rules in the Ada Reference Manual is to
29543 make sure that no entity is accessed before it has been elaborated. For
29544 a subprogram, this means that the spec and body must have been elaborated
29545 before the subprogram is called. For an object, this means that the object
29546 must have been elaborated before its value is read or written. A violation
29547 of either of these two requirements is an access before elaboration order,
29548 and this section has been all about avoiding such errors.
29550 In the case where more than one order of elaboration is possible, in the
29551 sense that access before elaboration errors are avoided, then any one of
29552 the orders is ``correct'' in the sense that it meets the requirements of
29553 the Ada Reference Manual, and no such error occurs.
29555 However, it may be the case for a given program, that there are
29556 constraints on the order of elaboration that come not from consideration
29557 of avoiding elaboration errors, but rather from extra-lingual logic
29558 requirements. Consider this example:
29560 @smallexample @c ada
29561 with Init_Constants;
29562 package Constants is
29567 package Init_Constants is
29568 procedure P; -- require a body
29569 end Init_Constants;
29572 package body Init_Constants is
29573 procedure P is begin null; end;
29577 end Init_Constants;
29581 Z : Integer := Constants.X + Constants.Y;
29585 with Text_IO; use Text_IO;
29588 Put_Line (Calc.Z'Img);
29593 In this example, there is more than one valid order of elaboration. For
29594 example both the following are correct orders:
29597 Init_Constants spec
29600 Init_Constants body
29605 Init_Constants spec
29606 Init_Constants body
29613 There is no language rule to prefer one or the other, both are correct
29614 from an order of elaboration point of view. But the programmatic effects
29615 of the two orders are very different. In the first, the elaboration routine
29616 of @code{Calc} initializes @code{Z} to zero, and then the main program
29617 runs with this value of zero. But in the second order, the elaboration
29618 routine of @code{Calc} runs after the body of Init_Constants has set
29619 @code{X} and @code{Y} and thus @code{Z} is set to 7 before @code{Main}
29622 One could perhaps by applying pretty clever non-artificial intelligence
29623 to the situation guess that it is more likely that the second order of
29624 elaboration is the one desired, but there is no formal linguistic reason
29625 to prefer one over the other. In fact in this particular case, GNAT will
29626 prefer the second order, because of the rule that bodies are elaborated
29627 as soon as possible, but it's just luck that this is what was wanted
29628 (if indeed the second order was preferred).
29630 If the program cares about the order of elaboration routines in a case like
29631 this, it is important to specify the order required. In this particular
29632 case, that could have been achieved by adding to the spec of Calc:
29634 @smallexample @c ada
29635 pragma Elaborate_All (Constants);
29639 which requires that the body (if any) and spec of @code{Constants},
29640 as well as the body and spec of any unit @code{with}'ed by
29641 @code{Constants} be elaborated before @code{Calc} is elaborated.
29643 Clearly no automatic method can always guess which alternative you require,
29644 and if you are working with legacy code that had constraints of this kind
29645 which were not properly specified by adding @code{Elaborate} or
29646 @code{Elaborate_All} pragmas, then indeed it is possible that two different
29647 compilers can choose different orders.
29649 However, GNAT does attempt to diagnose the common situation where there
29650 are uninitialized variables in the visible part of a package spec, and the
29651 corresponding package body has an elaboration block that directly or
29652 indirectly initialized one or more of these variables. This is the situation
29653 in which a pragma Elaborate_Body is usually desirable, and GNAT will generate
29654 a warning that suggests this addition if it detects this situation.
29656 The @code{gnatbind}
29657 @option{^-p^/PESSIMISTIC_ELABORATION^} switch may be useful in smoking
29658 out problems. This switch causes bodies to be elaborated as late as possible
29659 instead of as early as possible. In the example above, it would have forced
29660 the choice of the first elaboration order. If you get different results
29661 when using this switch, and particularly if one set of results is right,
29662 and one is wrong as far as you are concerned, it shows that you have some
29663 missing @code{Elaborate} pragmas. For the example above, we have the
29667 gnatmake -f -q main
29670 gnatmake -f -q main -bargs -p
29676 It is of course quite unlikely that both these results are correct, so
29677 it is up to you in a case like this to investigate the source of the
29678 difference, by looking at the two elaboration orders that are chosen,
29679 and figuring out which is correct, and then adding the necessary
29680 @code{Elaborate} or @code{Elaborate_All} pragmas to ensure the desired order.
29684 @c *******************************
29685 @node Conditional Compilation
29686 @appendix Conditional Compilation
29687 @c *******************************
29688 @cindex Conditional compilation
29691 It is often necessary to arrange for a single source program
29692 to serve multiple purposes, where it is compiled in different
29693 ways to achieve these different goals. Some examples of the
29694 need for this feature are
29697 @item Adapting a program to a different hardware environment
29698 @item Adapting a program to a different target architecture
29699 @item Turning debugging features on and off
29700 @item Arranging for a program to compile with different compilers
29704 In C, or C++, the typical approach would be to use the preprocessor
29705 that is defined as part of the language. The Ada language does not
29706 contain such a feature. This is not an oversight, but rather a very
29707 deliberate design decision, based on the experience that overuse of
29708 the preprocessing features in C and C++ can result in programs that
29709 are extremely difficult to maintain. For example, if we have ten
29710 switches that can be on or off, this means that there are a thousand
29711 separate programs, any one of which might not even be syntactically
29712 correct, and even if syntactically correct, the resulting program
29713 might not work correctly. Testing all combinations can quickly become
29716 Nevertheless, the need to tailor programs certainly exists, and in
29717 this Appendix we will discuss how this can
29718 be achieved using Ada in general, and GNAT in particular.
29721 * Use of Boolean Constants::
29722 * Debugging - A Special Case::
29723 * Conditionalizing Declarations::
29724 * Use of Alternative Implementations::
29728 @node Use of Boolean Constants
29729 @section Use of Boolean Constants
29732 In the case where the difference is simply which code
29733 sequence is executed, the cleanest solution is to use Boolean
29734 constants to control which code is executed.
29736 @smallexample @c ada
29738 FP_Initialize_Required : constant Boolean := True;
29740 if FP_Initialize_Required then
29747 Not only will the code inside the @code{if} statement not be executed if
29748 the constant Boolean is @code{False}, but it will also be completely
29749 deleted from the program.
29750 However, the code is only deleted after the @code{if} statement
29751 has been checked for syntactic and semantic correctness.
29752 (In contrast, with preprocessors the code is deleted before the
29753 compiler ever gets to see it, so it is not checked until the switch
29755 @cindex Preprocessors (contrasted with conditional compilation)
29757 Typically the Boolean constants will be in a separate package,
29760 @smallexample @c ada
29763 FP_Initialize_Required : constant Boolean := True;
29764 Reset_Available : constant Boolean := False;
29771 The @code{Config} package exists in multiple forms for the various targets,
29772 with an appropriate script selecting the version of @code{Config} needed.
29773 Then any other unit requiring conditional compilation can do a @code{with}
29774 of @code{Config} to make the constants visible.
29777 @node Debugging - A Special Case
29778 @section Debugging - A Special Case
29781 A common use of conditional code is to execute statements (for example
29782 dynamic checks, or output of intermediate results) under control of a
29783 debug switch, so that the debugging behavior can be turned on and off.
29784 This can be done using a Boolean constant to control whether the code
29787 @smallexample @c ada
29790 Put_Line ("got to the first stage!");
29798 @smallexample @c ada
29800 if Debugging and then Temperature > 999.0 then
29801 raise Temperature_Crazy;
29807 Since this is a common case, there are special features to deal with
29808 this in a convenient manner. For the case of tests, Ada 2005 has added
29809 a pragma @code{Assert} that can be used for such tests. This pragma is modeled
29810 @cindex pragma @code{Assert}
29811 on the @code{Assert} pragma that has always been available in GNAT, so this
29812 feature may be used with GNAT even if you are not using Ada 2005 features.
29813 The use of pragma @code{Assert} is described in
29814 @ref{Pragma Assert,,, gnat_rm, GNAT Reference Manual}, but as an
29815 example, the last test could be written:
29817 @smallexample @c ada
29818 pragma Assert (Temperature <= 999.0, "Temperature Crazy");
29824 @smallexample @c ada
29825 pragma Assert (Temperature <= 999.0);
29829 In both cases, if assertions are active and the temperature is excessive,
29830 the exception @code{Assert_Failure} will be raised, with the given string in
29831 the first case or a string indicating the location of the pragma in the second
29832 case used as the exception message.
29834 You can turn assertions on and off by using the @code{Assertion_Policy}
29836 @cindex pragma @code{Assertion_Policy}
29837 This is an Ada 2005 pragma which is implemented in all modes by
29838 GNAT, but only in the latest versions of GNAT which include Ada 2005
29839 capability. Alternatively, you can use the @option{-gnata} switch
29840 @cindex @option{-gnata} switch
29841 to enable assertions from the command line (this is recognized by all versions
29844 For the example above with the @code{Put_Line}, the GNAT-specific pragma
29845 @code{Debug} can be used:
29846 @cindex pragma @code{Debug}
29848 @smallexample @c ada
29849 pragma Debug (Put_Line ("got to the first stage!"));
29853 If debug pragmas are enabled, the argument, which must be of the form of
29854 a procedure call, is executed (in this case, @code{Put_Line} will be called).
29855 Only one call can be present, but of course a special debugging procedure
29856 containing any code you like can be included in the program and then
29857 called in a pragma @code{Debug} argument as needed.
29859 One advantage of pragma @code{Debug} over the @code{if Debugging then}
29860 construct is that pragma @code{Debug} can appear in declarative contexts,
29861 such as at the very beginning of a procedure, before local declarations have
29864 Debug pragmas are enabled using either the @option{-gnata} switch that also
29865 controls assertions, or with a separate Debug_Policy pragma.
29866 @cindex pragma @code{Debug_Policy}
29867 The latter pragma is new in the Ada 2005 versions of GNAT (but it can be used
29868 in Ada 95 and Ada 83 programs as well), and is analogous to
29869 pragma @code{Assertion_Policy} to control assertions.
29871 @code{Assertion_Policy} and @code{Debug_Policy} are configuration pragmas,
29872 and thus they can appear in @file{gnat.adc} if you are not using a
29873 project file, or in the file designated to contain configuration pragmas
29875 They then apply to all subsequent compilations. In practice the use of
29876 the @option{-gnata} switch is often the most convenient method of controlling
29877 the status of these pragmas.
29879 Note that a pragma is not a statement, so in contexts where a statement
29880 sequence is required, you can't just write a pragma on its own. You have
29881 to add a @code{null} statement.
29883 @smallexample @c ada
29886 @dots{} -- some statements
29888 pragma Assert (Num_Cases < 10);
29895 @node Conditionalizing Declarations
29896 @section Conditionalizing Declarations
29899 In some cases, it may be necessary to conditionalize declarations to meet
29900 different requirements. For example we might want a bit string whose length
29901 is set to meet some hardware message requirement.
29903 In some cases, it may be possible to do this using declare blocks controlled
29904 by conditional constants:
29906 @smallexample @c ada
29908 if Small_Machine then
29910 X : Bit_String (1 .. 10);
29916 X : Large_Bit_String (1 .. 1000);
29925 Note that in this approach, both declarations are analyzed by the
29926 compiler so this can only be used where both declarations are legal,
29927 even though one of them will not be used.
29929 Another approach is to define integer constants, e.g.@: @code{Bits_Per_Word}, or
29930 Boolean constants, e.g.@: @code{Little_Endian}, and then write declarations
29931 that are parameterized by these constants. For example
29933 @smallexample @c ada
29936 Field1 at 0 range Boolean'Pos (Little_Endian) * 10 .. Bits_Per_Word;
29942 If @code{Bits_Per_Word} is set to 32, this generates either
29944 @smallexample @c ada
29947 Field1 at 0 range 0 .. 32;
29953 for the big endian case, or
29955 @smallexample @c ada
29958 Field1 at 0 range 10 .. 32;
29964 for the little endian case. Since a powerful subset of Ada expression
29965 notation is usable for creating static constants, clever use of this
29966 feature can often solve quite difficult problems in conditionalizing
29967 compilation (note incidentally that in Ada 95, the little endian
29968 constant was introduced as @code{System.Default_Bit_Order}, so you do not
29969 need to define this one yourself).
29972 @node Use of Alternative Implementations
29973 @section Use of Alternative Implementations
29976 In some cases, none of the approaches described above are adequate. This
29977 can occur for example if the set of declarations required is radically
29978 different for two different configurations.
29980 In this situation, the official Ada way of dealing with conditionalizing
29981 such code is to write separate units for the different cases. As long as
29982 this does not result in excessive duplication of code, this can be done
29983 without creating maintenance problems. The approach is to share common
29984 code as far as possible, and then isolate the code and declarations
29985 that are different. Subunits are often a convenient method for breaking
29986 out a piece of a unit that is to be conditionalized, with separate files
29987 for different versions of the subunit for different targets, where the
29988 build script selects the right one to give to the compiler.
29989 @cindex Subunits (and conditional compilation)
29991 As an example, consider a situation where a new feature in Ada 2005
29992 allows something to be done in a really nice way. But your code must be able
29993 to compile with an Ada 95 compiler. Conceptually you want to say:
29995 @smallexample @c ada
29998 @dots{} neat Ada 2005 code
30000 @dots{} not quite as neat Ada 95 code
30006 where @code{Ada_2005} is a Boolean constant.
30008 But this won't work when @code{Ada_2005} is set to @code{False},
30009 since the @code{then} clause will be illegal for an Ada 95 compiler.
30010 (Recall that although such unreachable code would eventually be deleted
30011 by the compiler, it still needs to be legal. If it uses features
30012 introduced in Ada 2005, it will be illegal in Ada 95.)
30014 So instead we write
30016 @smallexample @c ada
30017 procedure Insert is separate;
30021 Then we have two files for the subunit @code{Insert}, with the two sets of
30023 If the package containing this is called @code{File_Queries}, then we might
30027 @item @file{file_queries-insert-2005.adb}
30028 @item @file{file_queries-insert-95.adb}
30032 and the build script renames the appropriate file to
30035 file_queries-insert.adb
30039 and then carries out the compilation.
30041 This can also be done with project files' naming schemes. For example:
30043 @smallexample @c project
30044 For Body ("File_Queries.Insert") use "file_queries-insert-2005.ada";
30048 Note also that with project files it is desirable to use a different extension
30049 than @file{ads} / @file{adb} for alternative versions. Otherwise a naming
30050 conflict may arise through another commonly used feature: to declare as part
30051 of the project a set of directories containing all the sources obeying the
30052 default naming scheme.
30054 The use of alternative units is certainly feasible in all situations,
30055 and for example the Ada part of the GNAT run-time is conditionalized
30056 based on the target architecture using this approach. As a specific example,
30057 consider the implementation of the AST feature in VMS. There is one
30065 which is the same for all architectures, and three bodies:
30069 used for all non-VMS operating systems
30070 @item s-asthan-vms-alpha.adb
30071 used for VMS on the Alpha
30072 @item s-asthan-vms-ia64.adb
30073 used for VMS on the ia64
30077 The dummy version @file{s-asthan.adb} simply raises exceptions noting that
30078 this operating system feature is not available, and the two remaining
30079 versions interface with the corresponding versions of VMS to provide
30080 VMS-compatible AST handling. The GNAT build script knows the architecture
30081 and operating system, and automatically selects the right version,
30082 renaming it if necessary to @file{s-asthan.adb} before the run-time build.
30084 Another style for arranging alternative implementations is through Ada's
30085 access-to-subprogram facility.
30086 In case some functionality is to be conditionally included,
30087 you can declare an access-to-procedure variable @code{Ref} that is initialized
30088 to designate a ``do nothing'' procedure, and then invoke @code{Ref.all}
30090 In some library package, set @code{Ref} to @code{Proc'Access} for some
30091 procedure @code{Proc} that performs the relevant processing.
30092 The initialization only occurs if the library package is included in the
30094 The same idea can also be implemented using tagged types and dispatching
30098 @node Preprocessing
30099 @section Preprocessing
30100 @cindex Preprocessing
30103 Although it is quite possible to conditionalize code without the use of
30104 C-style preprocessing, as described earlier in this section, it is
30105 nevertheless convenient in some cases to use the C approach. Moreover,
30106 older Ada compilers have often provided some preprocessing capability,
30107 so legacy code may depend on this approach, even though it is not
30110 To accommodate such use, GNAT provides a preprocessor (modeled to a large
30111 extent on the various preprocessors that have been used
30112 with legacy code on other compilers, to enable easier transition).
30114 The preprocessor may be used in two separate modes. It can be used quite
30115 separately from the compiler, to generate a separate output source file
30116 that is then fed to the compiler as a separate step. This is the
30117 @code{gnatprep} utility, whose use is fully described in
30118 @ref{Preprocessing Using gnatprep}.
30119 @cindex @code{gnatprep}
30121 The preprocessing language allows such constructs as
30125 #if DEBUG or PRIORITY > 4 then
30126 bunch of declarations
30128 completely different bunch of declarations
30134 The values of the symbols @code{DEBUG} and @code{PRIORITY} can be
30135 defined either on the command line or in a separate file.
30137 The other way of running the preprocessor is even closer to the C style and
30138 often more convenient. In this approach the preprocessing is integrated into
30139 the compilation process. The compiler is fed the preprocessor input which
30140 includes @code{#if} lines etc, and then the compiler carries out the
30141 preprocessing internally and processes the resulting output.
30142 For more details on this approach, see @ref{Integrated Preprocessing}.
30145 @c *******************************
30146 @node Inline Assembler
30147 @appendix Inline Assembler
30148 @c *******************************
30151 If you need to write low-level software that interacts directly
30152 with the hardware, Ada provides two ways to incorporate assembly
30153 language code into your program. First, you can import and invoke
30154 external routines written in assembly language, an Ada feature fully
30155 supported by GNAT@. However, for small sections of code it may be simpler
30156 or more efficient to include assembly language statements directly
30157 in your Ada source program, using the facilities of the implementation-defined
30158 package @code{System.Machine_Code}, which incorporates the gcc
30159 Inline Assembler. The Inline Assembler approach offers a number of advantages,
30160 including the following:
30163 @item No need to use non-Ada tools
30164 @item Consistent interface over different targets
30165 @item Automatic usage of the proper calling conventions
30166 @item Access to Ada constants and variables
30167 @item Definition of intrinsic routines
30168 @item Possibility of inlining a subprogram comprising assembler code
30169 @item Code optimizer can take Inline Assembler code into account
30172 This chapter presents a series of examples to show you how to use
30173 the Inline Assembler. Although it focuses on the Intel x86,
30174 the general approach applies also to other processors.
30175 It is assumed that you are familiar with Ada
30176 and with assembly language programming.
30179 * Basic Assembler Syntax::
30180 * A Simple Example of Inline Assembler::
30181 * Output Variables in Inline Assembler::
30182 * Input Variables in Inline Assembler::
30183 * Inlining Inline Assembler Code::
30184 * Other Asm Functionality::
30187 @c ---------------------------------------------------------------------------
30188 @node Basic Assembler Syntax
30189 @section Basic Assembler Syntax
30192 The assembler used by GNAT and gcc is based not on the Intel assembly
30193 language, but rather on a language that descends from the AT&T Unix
30194 assembler @emph{as} (and which is often referred to as ``AT&T syntax'').
30195 The following table summarizes the main features of @emph{as} syntax
30196 and points out the differences from the Intel conventions.
30197 See the gcc @emph{as} and @emph{gas} (an @emph{as} macro
30198 pre-processor) documentation for further information.
30201 @item Register names
30202 gcc / @emph{as}: Prefix with ``%''; for example @code{%eax}
30204 Intel: No extra punctuation; for example @code{eax}
30206 @item Immediate operand
30207 gcc / @emph{as}: Prefix with ``$''; for example @code{$4}
30209 Intel: No extra punctuation; for example @code{4}
30212 gcc / @emph{as}: Prefix with ``$''; for example @code{$loc}
30214 Intel: No extra punctuation; for example @code{loc}
30216 @item Memory contents
30217 gcc / @emph{as}: No extra punctuation; for example @code{loc}
30219 Intel: Square brackets; for example @code{[loc]}
30221 @item Register contents
30222 gcc / @emph{as}: Parentheses; for example @code{(%eax)}
30224 Intel: Square brackets; for example @code{[eax]}
30226 @item Hexadecimal numbers
30227 gcc / @emph{as}: Leading ``0x'' (C language syntax); for example @code{0xA0}
30229 Intel: Trailing ``h''; for example @code{A0h}
30232 gcc / @emph{as}: Explicit in op code; for example @code{movw} to move
30235 Intel: Implicit, deduced by assembler; for example @code{mov}
30237 @item Instruction repetition
30238 gcc / @emph{as}: Split into two lines; for example
30244 Intel: Keep on one line; for example @code{rep stosl}
30246 @item Order of operands
30247 gcc / @emph{as}: Source first; for example @code{movw $4, %eax}
30249 Intel: Destination first; for example @code{mov eax, 4}
30252 @c ---------------------------------------------------------------------------
30253 @node A Simple Example of Inline Assembler
30254 @section A Simple Example of Inline Assembler
30257 The following example will generate a single assembly language statement,
30258 @code{nop}, which does nothing. Despite its lack of run-time effect,
30259 the example will be useful in illustrating the basics of
30260 the Inline Assembler facility.
30262 @smallexample @c ada
30264 with System.Machine_Code; use System.Machine_Code;
30265 procedure Nothing is
30272 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
30273 here it takes one parameter, a @emph{template string} that must be a static
30274 expression and that will form the generated instruction.
30275 @code{Asm} may be regarded as a compile-time procedure that parses
30276 the template string and additional parameters (none here),
30277 from which it generates a sequence of assembly language instructions.
30279 The examples in this chapter will illustrate several of the forms
30280 for invoking @code{Asm}; a complete specification of the syntax
30281 is found in @ref{Machine Code Insertions,,, gnat_rm, GNAT Reference
30284 Under the standard GNAT conventions, the @code{Nothing} procedure
30285 should be in a file named @file{nothing.adb}.
30286 You can build the executable in the usual way:
30290 However, the interesting aspect of this example is not its run-time behavior
30291 but rather the generated assembly code.
30292 To see this output, invoke the compiler as follows:
30294 gcc -c -S -fomit-frame-pointer -gnatp @file{nothing.adb}
30296 where the options are:
30300 compile only (no bind or link)
30302 generate assembler listing
30303 @item -fomit-frame-pointer
30304 do not set up separate stack frames
30306 do not add runtime checks
30309 This gives a human-readable assembler version of the code. The resulting
30310 file will have the same name as the Ada source file, but with a @code{.s}
30311 extension. In our example, the file @file{nothing.s} has the following
30316 .file "nothing.adb"
30318 ___gnu_compiled_ada:
30321 .globl __ada_nothing
30333 The assembly code you included is clearly indicated by
30334 the compiler, between the @code{#APP} and @code{#NO_APP}
30335 delimiters. The character before the 'APP' and 'NOAPP'
30336 can differ on different targets. For example, GNU/Linux uses '#APP' while
30337 on NT you will see '/APP'.
30339 If you make a mistake in your assembler code (such as using the
30340 wrong size modifier, or using a wrong operand for the instruction) GNAT
30341 will report this error in a temporary file, which will be deleted when
30342 the compilation is finished. Generating an assembler file will help
30343 in such cases, since you can assemble this file separately using the
30344 @emph{as} assembler that comes with gcc.
30346 Assembling the file using the command
30349 as @file{nothing.s}
30352 will give you error messages whose lines correspond to the assembler
30353 input file, so you can easily find and correct any mistakes you made.
30354 If there are no errors, @emph{as} will generate an object file
30355 @file{nothing.out}.
30357 @c ---------------------------------------------------------------------------
30358 @node Output Variables in Inline Assembler
30359 @section Output Variables in Inline Assembler
30362 The examples in this section, showing how to access the processor flags,
30363 illustrate how to specify the destination operands for assembly language
30366 @smallexample @c ada
30368 with Interfaces; use Interfaces;
30369 with Ada.Text_IO; use Ada.Text_IO;
30370 with System.Machine_Code; use System.Machine_Code;
30371 procedure Get_Flags is
30372 Flags : Unsigned_32;
30375 Asm ("pushfl" & LF & HT & -- push flags on stack
30376 "popl %%eax" & LF & HT & -- load eax with flags
30377 "movl %%eax, %0", -- store flags in variable
30378 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
30379 Put_Line ("Flags register:" & Flags'Img);
30384 In order to have a nicely aligned assembly listing, we have separated
30385 multiple assembler statements in the Asm template string with linefeed
30386 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
30387 The resulting section of the assembly output file is:
30394 movl %eax, -40(%ebp)
30399 It would have been legal to write the Asm invocation as:
30402 Asm ("pushfl popl %%eax movl %%eax, %0")
30405 but in the generated assembler file, this would come out as:
30409 pushfl popl %eax movl %eax, -40(%ebp)
30413 which is not so convenient for the human reader.
30415 We use Ada comments
30416 at the end of each line to explain what the assembler instructions
30417 actually do. This is a useful convention.
30419 When writing Inline Assembler instructions, you need to precede each register
30420 and variable name with a percent sign. Since the assembler already requires
30421 a percent sign at the beginning of a register name, you need two consecutive
30422 percent signs for such names in the Asm template string, thus @code{%%eax}.
30423 In the generated assembly code, one of the percent signs will be stripped off.
30425 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
30426 variables: operands you later define using @code{Input} or @code{Output}
30427 parameters to @code{Asm}.
30428 An output variable is illustrated in
30429 the third statement in the Asm template string:
30433 The intent is to store the contents of the eax register in a variable that can
30434 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
30435 necessarily work, since the compiler might optimize by using a register
30436 to hold Flags, and the expansion of the @code{movl} instruction would not be
30437 aware of this optimization. The solution is not to store the result directly
30438 but rather to advise the compiler to choose the correct operand form;
30439 that is the purpose of the @code{%0} output variable.
30441 Information about the output variable is supplied in the @code{Outputs}
30442 parameter to @code{Asm}:
30444 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
30447 The output is defined by the @code{Asm_Output} attribute of the target type;
30448 the general format is
30450 Type'Asm_Output (constraint_string, variable_name)
30453 The constraint string directs the compiler how
30454 to store/access the associated variable. In the example
30456 Unsigned_32'Asm_Output ("=m", Flags);
30458 the @code{"m"} (memory) constraint tells the compiler that the variable
30459 @code{Flags} should be stored in a memory variable, thus preventing
30460 the optimizer from keeping it in a register. In contrast,
30462 Unsigned_32'Asm_Output ("=r", Flags);
30464 uses the @code{"r"} (register) constraint, telling the compiler to
30465 store the variable in a register.
30467 If the constraint is preceded by the equal character (@strong{=}), it tells
30468 the compiler that the variable will be used to store data into it.
30470 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
30471 allowing the optimizer to choose whatever it deems best.
30473 There are a fairly large number of constraints, but the ones that are
30474 most useful (for the Intel x86 processor) are the following:
30480 global (i.e.@: can be stored anywhere)
30498 use one of eax, ebx, ecx or edx
30500 use one of eax, ebx, ecx, edx, esi or edi
30503 The full set of constraints is described in the gcc and @emph{as}
30504 documentation; note that it is possible to combine certain constraints
30505 in one constraint string.
30507 You specify the association of an output variable with an assembler operand
30508 through the @code{%}@emph{n} notation, where @emph{n} is a non-negative
30510 @smallexample @c ada
30512 Asm ("pushfl" & LF & HT & -- push flags on stack
30513 "popl %%eax" & LF & HT & -- load eax with flags
30514 "movl %%eax, %0", -- store flags in variable
30515 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
30519 @code{%0} will be replaced in the expanded code by the appropriate operand,
30521 the compiler decided for the @code{Flags} variable.
30523 In general, you may have any number of output variables:
30526 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
30528 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
30529 of @code{Asm_Output} attributes
30533 @smallexample @c ada
30535 Asm ("movl %%eax, %0" & LF & HT &
30536 "movl %%ebx, %1" & LF & HT &
30538 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
30539 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
30540 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
30544 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
30545 in the Ada program.
30547 As a variation on the @code{Get_Flags} example, we can use the constraints
30548 string to direct the compiler to store the eax register into the @code{Flags}
30549 variable, instead of including the store instruction explicitly in the
30550 @code{Asm} template string:
30552 @smallexample @c ada
30554 with Interfaces; use Interfaces;
30555 with Ada.Text_IO; use Ada.Text_IO;
30556 with System.Machine_Code; use System.Machine_Code;
30557 procedure Get_Flags_2 is
30558 Flags : Unsigned_32;
30561 Asm ("pushfl" & LF & HT & -- push flags on stack
30562 "popl %%eax", -- save flags in eax
30563 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
30564 Put_Line ("Flags register:" & Flags'Img);
30570 The @code{"a"} constraint tells the compiler that the @code{Flags}
30571 variable will come from the eax register. Here is the resulting code:
30579 movl %eax,-40(%ebp)
30584 The compiler generated the store of eax into Flags after
30585 expanding the assembler code.
30587 Actually, there was no need to pop the flags into the eax register;
30588 more simply, we could just pop the flags directly into the program variable:
30590 @smallexample @c ada
30592 with Interfaces; use Interfaces;
30593 with Ada.Text_IO; use Ada.Text_IO;
30594 with System.Machine_Code; use System.Machine_Code;
30595 procedure Get_Flags_3 is
30596 Flags : Unsigned_32;
30599 Asm ("pushfl" & LF & HT & -- push flags on stack
30600 "pop %0", -- save flags in Flags
30601 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
30602 Put_Line ("Flags register:" & Flags'Img);
30607 @c ---------------------------------------------------------------------------
30608 @node Input Variables in Inline Assembler
30609 @section Input Variables in Inline Assembler
30612 The example in this section illustrates how to specify the source operands
30613 for assembly language statements.
30614 The program simply increments its input value by 1:
30616 @smallexample @c ada
30618 with Interfaces; use Interfaces;
30619 with Ada.Text_IO; use Ada.Text_IO;
30620 with System.Machine_Code; use System.Machine_Code;
30621 procedure Increment is
30623 function Incr (Value : Unsigned_32) return Unsigned_32 is
30624 Result : Unsigned_32;
30627 Inputs => Unsigned_32'Asm_Input ("a", Value),
30628 Outputs => Unsigned_32'Asm_Output ("=a", Result));
30632 Value : Unsigned_32;
30636 Put_Line ("Value before is" & Value'Img);
30637 Value := Incr (Value);
30638 Put_Line ("Value after is" & Value'Img);
30643 The @code{Outputs} parameter to @code{Asm} specifies
30644 that the result will be in the eax register and that it is to be stored
30645 in the @code{Result} variable.
30647 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
30648 but with an @code{Asm_Input} attribute.
30649 The @code{"="} constraint, indicating an output value, is not present.
30651 You can have multiple input variables, in the same way that you can have more
30652 than one output variable.
30654 The parameter count (%0, %1) etc, now starts at the first input
30655 statement, and continues with the output statements.
30656 When both parameters use the same variable, the
30657 compiler will treat them as the same %n operand, which is the case here.
30659 Just as the @code{Outputs} parameter causes the register to be stored into the
30660 target variable after execution of the assembler statements, so does the
30661 @code{Inputs} parameter cause its variable to be loaded into the register
30662 before execution of the assembler statements.
30664 Thus the effect of the @code{Asm} invocation is:
30666 @item load the 32-bit value of @code{Value} into eax
30667 @item execute the @code{incl %eax} instruction
30668 @item store the contents of eax into the @code{Result} variable
30671 The resulting assembler file (with @option{-O2} optimization) contains:
30674 _increment__incr.1:
30687 @c ---------------------------------------------------------------------------
30688 @node Inlining Inline Assembler Code
30689 @section Inlining Inline Assembler Code
30692 For a short subprogram such as the @code{Incr} function in the previous
30693 section, the overhead of the call and return (creating / deleting the stack
30694 frame) can be significant, compared to the amount of code in the subprogram
30695 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
30696 which directs the compiler to expand invocations of the subprogram at the
30697 point(s) of call, instead of setting up a stack frame for out-of-line calls.
30698 Here is the resulting program:
30700 @smallexample @c ada
30702 with Interfaces; use Interfaces;
30703 with Ada.Text_IO; use Ada.Text_IO;
30704 with System.Machine_Code; use System.Machine_Code;
30705 procedure Increment_2 is
30707 function Incr (Value : Unsigned_32) return Unsigned_32 is
30708 Result : Unsigned_32;
30711 Inputs => Unsigned_32'Asm_Input ("a", Value),
30712 Outputs => Unsigned_32'Asm_Output ("=a", Result));
30715 pragma Inline (Increment);
30717 Value : Unsigned_32;
30721 Put_Line ("Value before is" & Value'Img);
30722 Value := Increment (Value);
30723 Put_Line ("Value after is" & Value'Img);
30728 Compile the program with both optimization (@option{-O2}) and inlining
30729 (@option{-gnatn}) enabled.
30731 The @code{Incr} function is still compiled as usual, but at the
30732 point in @code{Increment} where our function used to be called:
30737 call _increment__incr.1
30742 the code for the function body directly appears:
30755 thus saving the overhead of stack frame setup and an out-of-line call.
30757 @c ---------------------------------------------------------------------------
30758 @node Other Asm Functionality
30759 @section Other @code{Asm} Functionality
30762 This section describes two important parameters to the @code{Asm}
30763 procedure: @code{Clobber}, which identifies register usage;
30764 and @code{Volatile}, which inhibits unwanted optimizations.
30767 * The Clobber Parameter::
30768 * The Volatile Parameter::
30771 @c ---------------------------------------------------------------------------
30772 @node The Clobber Parameter
30773 @subsection The @code{Clobber} Parameter
30776 One of the dangers of intermixing assembly language and a compiled language
30777 such as Ada is that the compiler needs to be aware of which registers are
30778 being used by the assembly code. In some cases, such as the earlier examples,
30779 the constraint string is sufficient to indicate register usage (e.g.,
30781 the eax register). But more generally, the compiler needs an explicit
30782 identification of the registers that are used by the Inline Assembly
30785 Using a register that the compiler doesn't know about
30786 could be a side effect of an instruction (like @code{mull}
30787 storing its result in both eax and edx).
30788 It can also arise from explicit register usage in your
30789 assembly code; for example:
30792 Asm ("movl %0, %%ebx" & LF & HT &
30794 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
30795 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out));
30799 where the compiler (since it does not analyze the @code{Asm} template string)
30800 does not know you are using the ebx register.
30802 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
30803 to identify the registers that will be used by your assembly code:
30807 Asm ("movl %0, %%ebx" & LF & HT &
30809 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
30810 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
30815 The Clobber parameter is a static string expression specifying the
30816 register(s) you are using. Note that register names are @emph{not} prefixed
30817 by a percent sign. Also, if more than one register is used then their names
30818 are separated by commas; e.g., @code{"eax, ebx"}
30820 The @code{Clobber} parameter has several additional uses:
30822 @item Use ``register'' name @code{cc} to indicate that flags might have changed
30823 @item Use ``register'' name @code{memory} if you changed a memory location
30826 @c ---------------------------------------------------------------------------
30827 @node The Volatile Parameter
30828 @subsection The @code{Volatile} Parameter
30829 @cindex Volatile parameter
30832 Compiler optimizations in the presence of Inline Assembler may sometimes have
30833 unwanted effects. For example, when an @code{Asm} invocation with an input
30834 variable is inside a loop, the compiler might move the loading of the input
30835 variable outside the loop, regarding it as a one-time initialization.
30837 If this effect is not desired, you can disable such optimizations by setting
30838 the @code{Volatile} parameter to @code{True}; for example:
30840 @smallexample @c ada
30842 Asm ("movl %0, %%ebx" & LF & HT &
30844 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
30845 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
30851 By default, @code{Volatile} is set to @code{False} unless there is no
30852 @code{Outputs} parameter.
30854 Although setting @code{Volatile} to @code{True} prevents unwanted
30855 optimizations, it will also disable other optimizations that might be
30856 important for efficiency. In general, you should set @code{Volatile}
30857 to @code{True} only if the compiler's optimizations have created
30859 @c END OF INLINE ASSEMBLER CHAPTER
30860 @c ===============================
30862 @c ***********************************
30863 @c * Compatibility and Porting Guide *
30864 @c ***********************************
30865 @node Compatibility and Porting Guide
30866 @appendix Compatibility and Porting Guide
30869 This chapter describes the compatibility issues that may arise between
30870 GNAT and other Ada compilation systems (including those for Ada 83),
30871 and shows how GNAT can expedite porting
30872 applications developed in other Ada environments.
30875 * Compatibility with Ada 83::
30876 * Compatibility between Ada 95 and Ada 2005::
30877 * Implementation-dependent characteristics::
30878 * Compatibility with Other Ada Systems::
30879 * Representation Clauses::
30881 @c Brief section is only in non-VMS version
30882 @c Full chapter is in VMS version
30883 * Compatibility with HP Ada 83::
30886 * Transitioning to 64-Bit GNAT for OpenVMS::
30890 @node Compatibility with Ada 83
30891 @section Compatibility with Ada 83
30892 @cindex Compatibility (between Ada 83 and Ada 95 / Ada 2005)
30895 Ada 95 and Ada 2005 are highly upwards compatible with Ada 83. In
30896 particular, the design intention was that the difficulties associated
30897 with moving from Ada 83 to Ada 95 or Ada 2005 should be no greater than those
30898 that occur when moving from one Ada 83 system to another.
30900 However, there are a number of points at which there are minor
30901 incompatibilities. The @cite{Ada 95 Annotated Reference Manual} contains
30902 full details of these issues,
30903 and should be consulted for a complete treatment.
30905 following subsections treat the most likely issues to be encountered.
30908 * Legal Ada 83 programs that are illegal in Ada 95::
30909 * More deterministic semantics::
30910 * Changed semantics::
30911 * Other language compatibility issues::
30914 @node Legal Ada 83 programs that are illegal in Ada 95
30915 @subsection Legal Ada 83 programs that are illegal in Ada 95
30917 Some legal Ada 83 programs are illegal (i.e., they will fail to compile) in
30918 Ada 95 and thus also in Ada 2005:
30921 @item Character literals
30922 Some uses of character literals are ambiguous. Since Ada 95 has introduced
30923 @code{Wide_Character} as a new predefined character type, some uses of
30924 character literals that were legal in Ada 83 are illegal in Ada 95.
30926 @smallexample @c ada
30927 for Char in 'A' .. 'Z' loop @dots{} end loop;
30931 The problem is that @code{'A'} and @code{'Z'} could be from either
30932 @code{Character} or @code{Wide_Character}. The simplest correction
30933 is to make the type explicit; e.g.:
30934 @smallexample @c ada
30935 for Char in Character range 'A' .. 'Z' loop @dots{} end loop;
30938 @item New reserved words
30939 The identifiers @code{abstract}, @code{aliased}, @code{protected},
30940 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
30941 Existing Ada 83 code using any of these identifiers must be edited to
30942 use some alternative name.
30944 @item Freezing rules
30945 The rules in Ada 95 are slightly different with regard to the point at
30946 which entities are frozen, and representation pragmas and clauses are
30947 not permitted past the freeze point. This shows up most typically in
30948 the form of an error message complaining that a representation item
30949 appears too late, and the appropriate corrective action is to move
30950 the item nearer to the declaration of the entity to which it refers.
30952 A particular case is that representation pragmas
30955 extended HP Ada 83 compatibility pragmas such as @code{Export_Procedure})
30957 cannot be applied to a subprogram body. If necessary, a separate subprogram
30958 declaration must be introduced to which the pragma can be applied.
30960 @item Optional bodies for library packages
30961 In Ada 83, a package that did not require a package body was nevertheless
30962 allowed to have one. This lead to certain surprises in compiling large
30963 systems (situations in which the body could be unexpectedly ignored by the
30964 binder). In Ada 95, if a package does not require a body then it is not
30965 permitted to have a body. To fix this problem, simply remove a redundant
30966 body if it is empty, or, if it is non-empty, introduce a dummy declaration
30967 into the spec that makes the body required. One approach is to add a private
30968 part to the package declaration (if necessary), and define a parameterless
30969 procedure called @code{Requires_Body}, which must then be given a dummy
30970 procedure body in the package body, which then becomes required.
30971 Another approach (assuming that this does not introduce elaboration
30972 circularities) is to add an @code{Elaborate_Body} pragma to the package spec,
30973 since one effect of this pragma is to require the presence of a package body.
30975 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
30976 In Ada 95, the exception @code{Numeric_Error} is a renaming of
30977 @code{Constraint_Error}.
30978 This means that it is illegal to have separate exception handlers for
30979 the two exceptions. The fix is simply to remove the handler for the
30980 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
30981 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
30983 @item Indefinite subtypes in generics
30984 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String})
30985 as the actual for a generic formal private type, but then the instantiation
30986 would be illegal if there were any instances of declarations of variables
30987 of this type in the generic body. In Ada 95, to avoid this clear violation
30988 of the methodological principle known as the ``contract model'',
30989 the generic declaration explicitly indicates whether
30990 or not such instantiations are permitted. If a generic formal parameter
30991 has explicit unknown discriminants, indicated by using @code{(<>)} after the
30992 type name, then it can be instantiated with indefinite types, but no
30993 stand-alone variables can be declared of this type. Any attempt to declare
30994 such a variable will result in an illegality at the time the generic is
30995 declared. If the @code{(<>)} notation is not used, then it is illegal
30996 to instantiate the generic with an indefinite type.
30997 This is the potential incompatibility issue when porting Ada 83 code to Ada 95.
30998 It will show up as a compile time error, and
30999 the fix is usually simply to add the @code{(<>)} to the generic declaration.
31002 @node More deterministic semantics
31003 @subsection More deterministic semantics
31007 Conversions from real types to integer types round away from 0. In Ada 83
31008 the conversion Integer(2.5) could deliver either 2 or 3 as its value. This
31009 implementation freedom was intended to support unbiased rounding in
31010 statistical applications, but in practice it interfered with portability.
31011 In Ada 95 the conversion semantics are unambiguous, and rounding away from 0
31012 is required. Numeric code may be affected by this change in semantics.
31013 Note, though, that this issue is no worse than already existed in Ada 83
31014 when porting code from one vendor to another.
31017 The Real-Time Annex introduces a set of policies that define the behavior of
31018 features that were implementation dependent in Ada 83, such as the order in
31019 which open select branches are executed.
31022 @node Changed semantics
31023 @subsection Changed semantics
31026 The worst kind of incompatibility is one where a program that is legal in
31027 Ada 83 is also legal in Ada 95 but can have an effect in Ada 95 that was not
31028 possible in Ada 83. Fortunately this is extremely rare, but the one
31029 situation that you should be alert to is the change in the predefined type
31030 @code{Character} from 7-bit ASCII to 8-bit Latin-1.
31033 @item Range of type @code{Character}
31034 The range of @code{Standard.Character} is now the full 256 characters
31035 of Latin-1, whereas in most Ada 83 implementations it was restricted
31036 to 128 characters. Although some of the effects of
31037 this change will be manifest in compile-time rejection of legal
31038 Ada 83 programs it is possible for a working Ada 83 program to have
31039 a different effect in Ada 95, one that was not permitted in Ada 83.
31040 As an example, the expression
31041 @code{Character'Pos(Character'Last)} returned @code{127} in Ada 83 and now
31042 delivers @code{255} as its value.
31043 In general, you should look at the logic of any
31044 character-processing Ada 83 program and see whether it needs to be adapted
31045 to work correctly with Latin-1. Note that the predefined Ada 95 API has a
31046 character handling package that may be relevant if code needs to be adapted
31047 to account for the additional Latin-1 elements.
31048 The desirable fix is to
31049 modify the program to accommodate the full character set, but in some cases
31050 it may be convenient to define a subtype or derived type of Character that
31051 covers only the restricted range.
31055 @node Other language compatibility issues
31056 @subsection Other language compatibility issues
31059 @item @option{-gnat83} switch
31060 All implementations of GNAT provide a switch that causes GNAT to operate
31061 in Ada 83 mode. In this mode, some but not all compatibility problems
31062 of the type described above are handled automatically. For example, the
31063 new reserved words introduced in Ada 95 and Ada 2005 are treated simply
31064 as identifiers as in Ada 83.
31066 in practice, it is usually advisable to make the necessary modifications
31067 to the program to remove the need for using this switch.
31068 See @ref{Compiling Different Versions of Ada}.
31070 @item Support for removed Ada 83 pragmas and attributes
31071 A number of pragmas and attributes from Ada 83 were removed from Ada 95,
31072 generally because they were replaced by other mechanisms. Ada 95 and Ada 2005
31073 compilers are allowed, but not required, to implement these missing
31074 elements. In contrast with some other compilers, GNAT implements all
31075 such pragmas and attributes, eliminating this compatibility concern. These
31076 include @code{pragma Interface} and the floating point type attributes
31077 (@code{Emax}, @code{Mantissa}, etc.), among other items.
31081 @node Compatibility between Ada 95 and Ada 2005
31082 @section Compatibility between Ada 95 and Ada 2005
31083 @cindex Compatibility between Ada 95 and Ada 2005
31086 Although Ada 2005 was designed to be upwards compatible with Ada 95, there are
31087 a number of incompatibilities. Several are enumerated below;
31088 for a complete description please see the
31089 Annotated Ada 2005 Reference Manual, or section 9.1.1 in
31090 @cite{Rationale for Ada 2005}.
31093 @item New reserved words.
31094 The words @code{interface}, @code{overriding} and @code{synchronized} are
31095 reserved in Ada 2005.
31096 A pre-Ada 2005 program that uses any of these as an identifier will be
31099 @item New declarations in predefined packages.
31100 A number of packages in the predefined environment contain new declarations:
31101 @code{Ada.Exceptions}, @code{Ada.Real_Time}, @code{Ada.Strings},
31102 @code{Ada.Strings.Fixed}, @code{Ada.Strings.Bounded},
31103 @code{Ada.Strings.Unbounded}, @code{Ada.Strings.Wide_Fixed},
31104 @code{Ada.Strings.Wide_Bounded}, @code{Ada.Strings.Wide_Unbounded},
31105 @code{Ada.Tags}, @code{Ada.Text_IO}, and @code{Interfaces.C}.
31106 If an Ada 95 program does a @code{with} and @code{use} of any of these
31107 packages, the new declarations may cause name clashes.
31109 @item Access parameters.
31110 A nondispatching subprogram with an access parameter cannot be renamed
31111 as a dispatching operation. This was permitted in Ada 95.
31113 @item Access types, discriminants, and constraints.
31114 Rule changes in this area have led to some incompatibilities; for example,
31115 constrained subtypes of some access types are not permitted in Ada 2005.
31117 @item Aggregates for limited types.
31118 The allowance of aggregates for limited types in Ada 2005 raises the
31119 possibility of ambiguities in legal Ada 95 programs, since additional types
31120 now need to be considered in expression resolution.
31122 @item Fixed-point multiplication and division.
31123 Certain expressions involving ``*'' or ``/'' for a fixed-point type, which
31124 were legal in Ada 95 and invoked the predefined versions of these operations,
31126 The ambiguity may be resolved either by applying a type conversion to the
31127 expression, or by explicitly invoking the operation from package
31130 @item Return-by-reference types.
31131 The Ada 95 return-by-reference mechanism has been removed. Instead, the user
31132 can declare a function returning a value from an anonymous access type.
31136 @node Implementation-dependent characteristics
31137 @section Implementation-dependent characteristics
31139 Although the Ada language defines the semantics of each construct as
31140 precisely as practical, in some situations (for example for reasons of
31141 efficiency, or where the effect is heavily dependent on the host or target
31142 platform) the implementation is allowed some freedom. In porting Ada 83
31143 code to GNAT, you need to be aware of whether / how the existing code
31144 exercised such implementation dependencies. Such characteristics fall into
31145 several categories, and GNAT offers specific support in assisting the
31146 transition from certain Ada 83 compilers.
31149 * Implementation-defined pragmas::
31150 * Implementation-defined attributes::
31152 * Elaboration order::
31153 * Target-specific aspects::
31156 @node Implementation-defined pragmas
31157 @subsection Implementation-defined pragmas
31160 Ada compilers are allowed to supplement the language-defined pragmas, and
31161 these are a potential source of non-portability. All GNAT-defined pragmas
31162 are described in @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT
31163 Reference Manual}, and these include several that are specifically
31164 intended to correspond to other vendors' Ada 83 pragmas.
31165 For migrating from VADS, the pragma @code{Use_VADS_Size} may be useful.
31166 For compatibility with HP Ada 83, GNAT supplies the pragmas
31167 @code{Extend_System}, @code{Ident}, @code{Inline_Generic},
31168 @code{Interface_Name}, @code{Passive}, @code{Suppress_All},
31169 and @code{Volatile}.
31170 Other relevant pragmas include @code{External} and @code{Link_With}.
31171 Some vendor-specific
31172 Ada 83 pragmas (@code{Share_Generic}, @code{Subtitle}, and @code{Title}) are
31174 avoiding compiler rejection of units that contain such pragmas; they are not
31175 relevant in a GNAT context and hence are not otherwise implemented.
31177 @node Implementation-defined attributes
31178 @subsection Implementation-defined attributes
31180 Analogous to pragmas, the set of attributes may be extended by an
31181 implementation. All GNAT-defined attributes are described in
31182 @ref{Implementation Defined Attributes,,, gnat_rm, GNAT Reference
31183 Manual}, and these include several that are specifically intended
31184 to correspond to other vendors' Ada 83 attributes. For migrating from VADS,
31185 the attribute @code{VADS_Size} may be useful. For compatibility with HP
31186 Ada 83, GNAT supplies the attributes @code{Bit}, @code{Machine_Size} and
31190 @subsection Libraries
31192 Vendors may supply libraries to supplement the standard Ada API. If Ada 83
31193 code uses vendor-specific libraries then there are several ways to manage
31194 this in Ada 95 or Ada 2005:
31197 If the source code for the libraries (specs and bodies) are
31198 available, then the libraries can be migrated in the same way as the
31201 If the source code for the specs but not the bodies are
31202 available, then you can reimplement the bodies.
31204 Some features introduced by Ada 95 obviate the need for library support. For
31205 example most Ada 83 vendors supplied a package for unsigned integers. The
31206 Ada 95 modular type feature is the preferred way to handle this need, so
31207 instead of migrating or reimplementing the unsigned integer package it may
31208 be preferable to retrofit the application using modular types.
31211 @node Elaboration order
31212 @subsection Elaboration order
31214 The implementation can choose any elaboration order consistent with the unit
31215 dependency relationship. This freedom means that some orders can result in
31216 Program_Error being raised due to an ``Access Before Elaboration'': an attempt
31217 to invoke a subprogram its body has been elaborated, or to instantiate a
31218 generic before the generic body has been elaborated. By default GNAT
31219 attempts to choose a safe order (one that will not encounter access before
31220 elaboration problems) by implicitly inserting @code{Elaborate} or
31221 @code{Elaborate_All} pragmas where
31222 needed. However, this can lead to the creation of elaboration circularities
31223 and a resulting rejection of the program by gnatbind. This issue is
31224 thoroughly described in @ref{Elaboration Order Handling in GNAT}.
31225 In brief, there are several
31226 ways to deal with this situation:
31230 Modify the program to eliminate the circularities, e.g.@: by moving
31231 elaboration-time code into explicitly-invoked procedures
31233 Constrain the elaboration order by including explicit @code{Elaborate_Body} or
31234 @code{Elaborate} pragmas, and then inhibit the generation of implicit
31235 @code{Elaborate_All}
31236 pragmas either globally (as an effect of the @option{-gnatE} switch) or locally
31237 (by selectively suppressing elaboration checks via pragma
31238 @code{Suppress(Elaboration_Check)} when it is safe to do so).
31241 @node Target-specific aspects
31242 @subsection Target-specific aspects
31244 Low-level applications need to deal with machine addresses, data
31245 representations, interfacing with assembler code, and similar issues. If
31246 such an Ada 83 application is being ported to different target hardware (for
31247 example where the byte endianness has changed) then you will need to
31248 carefully examine the program logic; the porting effort will heavily depend
31249 on the robustness of the original design. Moreover, Ada 95 (and thus
31250 Ada 2005) are sometimes
31251 incompatible with typical Ada 83 compiler practices regarding implicit
31252 packing, the meaning of the Size attribute, and the size of access values.
31253 GNAT's approach to these issues is described in @ref{Representation Clauses}.
31255 @node Compatibility with Other Ada Systems
31256 @section Compatibility with Other Ada Systems
31259 If programs avoid the use of implementation dependent and
31260 implementation defined features, as documented in the @cite{Ada
31261 Reference Manual}, there should be a high degree of portability between
31262 GNAT and other Ada systems. The following are specific items which
31263 have proved troublesome in moving Ada 95 programs from GNAT to other Ada 95
31264 compilers, but do not affect porting code to GNAT@.
31265 (As of @value{NOW}, GNAT is the only compiler available for Ada 2005;
31266 the following issues may or may not arise for Ada 2005 programs
31267 when other compilers appear.)
31270 @item Ada 83 Pragmas and Attributes
31271 Ada 95 compilers are allowed, but not required, to implement the missing
31272 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
31273 GNAT implements all such pragmas and attributes, eliminating this as
31274 a compatibility concern, but some other Ada 95 compilers reject these
31275 pragmas and attributes.
31277 @item Specialized Needs Annexes
31278 GNAT implements the full set of special needs annexes. At the
31279 current time, it is the only Ada 95 compiler to do so. This means that
31280 programs making use of these features may not be portable to other Ada
31281 95 compilation systems.
31283 @item Representation Clauses
31284 Some other Ada 95 compilers implement only the minimal set of
31285 representation clauses required by the Ada 95 reference manual. GNAT goes
31286 far beyond this minimal set, as described in the next section.
31289 @node Representation Clauses
31290 @section Representation Clauses
31293 The Ada 83 reference manual was quite vague in describing both the minimal
31294 required implementation of representation clauses, and also their precise
31295 effects. Ada 95 (and thus also Ada 2005) are much more explicit, but the
31296 minimal set of capabilities required is still quite limited.
31298 GNAT implements the full required set of capabilities in
31299 Ada 95 and Ada 2005, but also goes much further, and in particular
31300 an effort has been made to be compatible with existing Ada 83 usage to the
31301 greatest extent possible.
31303 A few cases exist in which Ada 83 compiler behavior is incompatible with
31304 the requirements in Ada 95 (and thus also Ada 2005). These are instances of
31305 intentional or accidental dependence on specific implementation dependent
31306 characteristics of these Ada 83 compilers. The following is a list of
31307 the cases most likely to arise in existing Ada 83 code.
31310 @item Implicit Packing
31311 Some Ada 83 compilers allowed a Size specification to cause implicit
31312 packing of an array or record. This could cause expensive implicit
31313 conversions for change of representation in the presence of derived
31314 types, and the Ada design intends to avoid this possibility.
31315 Subsequent AI's were issued to make it clear that such implicit
31316 change of representation in response to a Size clause is inadvisable,
31317 and this recommendation is represented explicitly in the Ada 95 (and Ada 2005)
31318 Reference Manuals as implementation advice that is followed by GNAT@.
31319 The problem will show up as an error
31320 message rejecting the size clause. The fix is simply to provide
31321 the explicit pragma @code{Pack}, or for more fine tuned control, provide
31322 a Component_Size clause.
31324 @item Meaning of Size Attribute
31325 The Size attribute in Ada 95 (and Ada 2005) for discrete types is defined as
31326 the minimal number of bits required to hold values of the type. For example,
31327 on a 32-bit machine, the size of @code{Natural} will typically be 31 and not
31328 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
31329 some 32 in this situation. This problem will usually show up as a compile
31330 time error, but not always. It is a good idea to check all uses of the
31331 'Size attribute when porting Ada 83 code. The GNAT specific attribute
31332 Object_Size can provide a useful way of duplicating the behavior of
31333 some Ada 83 compiler systems.
31335 @item Size of Access Types
31336 A common assumption in Ada 83 code is that an access type is in fact a pointer,
31337 and that therefore it will be the same size as a System.Address value. This
31338 assumption is true for GNAT in most cases with one exception. For the case of
31339 a pointer to an unconstrained array type (where the bounds may vary from one
31340 value of the access type to another), the default is to use a ``fat pointer'',
31341 which is represented as two separate pointers, one to the bounds, and one to
31342 the array. This representation has a number of advantages, including improved
31343 efficiency. However, it may cause some difficulties in porting existing Ada 83
31344 code which makes the assumption that, for example, pointers fit in 32 bits on
31345 a machine with 32-bit addressing.
31347 To get around this problem, GNAT also permits the use of ``thin pointers'' for
31348 access types in this case (where the designated type is an unconstrained array
31349 type). These thin pointers are indeed the same size as a System.Address value.
31350 To specify a thin pointer, use a size clause for the type, for example:
31352 @smallexample @c ada
31353 type X is access all String;
31354 for X'Size use Standard'Address_Size;
31358 which will cause the type X to be represented using a single pointer.
31359 When using this representation, the bounds are right behind the array.
31360 This representation is slightly less efficient, and does not allow quite
31361 such flexibility in the use of foreign pointers or in using the
31362 Unrestricted_Access attribute to create pointers to non-aliased objects.
31363 But for any standard portable use of the access type it will work in
31364 a functionally correct manner and allow porting of existing code.
31365 Note that another way of forcing a thin pointer representation
31366 is to use a component size clause for the element size in an array,
31367 or a record representation clause for an access field in a record.
31371 @c This brief section is only in the non-VMS version
31372 @c The complete chapter on HP Ada is in the VMS version
31373 @node Compatibility with HP Ada 83
31374 @section Compatibility with HP Ada 83
31377 The VMS version of GNAT fully implements all the pragmas and attributes
31378 provided by HP Ada 83, as well as providing the standard HP Ada 83
31379 libraries, including Starlet. In addition, data layouts and parameter
31380 passing conventions are highly compatible. This means that porting
31381 existing HP Ada 83 code to GNAT in VMS systems should be easier than
31382 most other porting efforts. The following are some of the most
31383 significant differences between GNAT and HP Ada 83.
31386 @item Default floating-point representation
31387 In GNAT, the default floating-point format is IEEE, whereas in HP Ada 83,
31388 it is VMS format. GNAT does implement the necessary pragmas
31389 (Long_Float, Float_Representation) for changing this default.
31392 The package System in GNAT exactly corresponds to the definition in the
31393 Ada 95 reference manual, which means that it excludes many of the
31394 HP Ada 83 extensions. However, a separate package Aux_DEC is provided
31395 that contains the additional definitions, and a special pragma,
31396 Extend_System allows this package to be treated transparently as an
31397 extension of package System.
31400 The definitions provided by Aux_DEC are exactly compatible with those
31401 in the HP Ada 83 version of System, with one exception.
31402 HP Ada provides the following declarations:
31404 @smallexample @c ada
31405 TO_ADDRESS (INTEGER)
31406 TO_ADDRESS (UNSIGNED_LONGWORD)
31407 TO_ADDRESS (@i{universal_integer})
31411 The version of TO_ADDRESS taking a @i{universal integer} argument is in fact
31412 an extension to Ada 83 not strictly compatible with the reference manual.
31413 In GNAT, we are constrained to be exactly compatible with the standard,
31414 and this means we cannot provide this capability. In HP Ada 83, the
31415 point of this definition is to deal with a call like:
31417 @smallexample @c ada
31418 TO_ADDRESS (16#12777#);
31422 Normally, according to the Ada 83 standard, one would expect this to be
31423 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
31424 of TO_ADDRESS@. However, in HP Ada 83, there is no ambiguity, since the
31425 definition using @i{universal_integer} takes precedence.
31427 In GNAT, since the version with @i{universal_integer} cannot be supplied, it
31428 is not possible to be 100% compatible. Since there are many programs using
31429 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
31430 to change the name of the function in the UNSIGNED_LONGWORD case, so the
31431 declarations provided in the GNAT version of AUX_Dec are:
31433 @smallexample @c ada
31434 function To_Address (X : Integer) return Address;
31435 pragma Pure_Function (To_Address);
31437 function To_Address_Long (X : Unsigned_Longword)
31439 pragma Pure_Function (To_Address_Long);
31443 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
31444 change the name to TO_ADDRESS_LONG@.
31446 @item Task_Id values
31447 The Task_Id values assigned will be different in the two systems, and GNAT
31448 does not provide a specified value for the Task_Id of the environment task,
31449 which in GNAT is treated like any other declared task.
31453 For full details on these and other less significant compatibility issues,
31454 see appendix E of the HP publication entitled @cite{HP Ada, Technical
31455 Overview and Comparison on HP Platforms}.
31457 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
31458 attributes are recognized, although only a subset of them can sensibly
31459 be implemented. The description of pragmas in @ref{Implementation
31460 Defined Pragmas,,, gnat_rm, GNAT Reference Manual}
31461 indicates whether or not they are applicable to non-VMS systems.
31465 @node Transitioning to 64-Bit GNAT for OpenVMS
31466 @section Transitioning to 64-Bit @value{EDITION} for OpenVMS
31469 This section is meant to assist users of pre-2006 @value{EDITION}
31470 for Alpha OpenVMS who are transitioning to 64-bit @value{EDITION},
31471 the version of the GNAT technology supplied in 2006 and later for
31472 OpenVMS on both Alpha and I64.
31475 * Introduction to transitioning::
31476 * Migration of 32 bit code::
31477 * Taking advantage of 64 bit addressing::
31478 * Technical details::
31481 @node Introduction to transitioning
31482 @subsection Introduction
31485 64-bit @value{EDITION} for Open VMS has been designed to meet
31490 Providing a full conforming implementation of Ada 95 and Ada 2005
31493 Allowing maximum backward compatibility, thus easing migration of existing
31497 Supplying a path for exploiting the full 64-bit address range
31501 Ada's strong typing semantics has made it
31502 impractical to have different 32-bit and 64-bit modes. As soon as
31503 one object could possibly be outside the 32-bit address space, this
31504 would make it necessary for the @code{System.Address} type to be 64 bits.
31505 In particular, this would cause inconsistencies if 32-bit code is
31506 called from 64-bit code that raises an exception.
31508 This issue has been resolved by always using 64-bit addressing
31509 at the system level, but allowing for automatic conversions between
31510 32-bit and 64-bit addresses where required. Thus users who
31511 do not currently require 64-bit addressing capabilities, can
31512 recompile their code with only minimal changes (and indeed
31513 if the code is written in portable Ada, with no assumptions about
31514 the size of the @code{Address} type, then no changes at all are necessary).
31516 this approach provides a simple, gradual upgrade path to future
31517 use of larger memories than available for 32-bit systems.
31518 Also, newly written applications or libraries will by default
31519 be fully compatible with future systems exploiting 64-bit
31520 addressing capabilities.
31522 @ref{Migration of 32 bit code}, will focus on porting applications
31523 that do not require more than 2 GB of
31524 addressable memory. This code will be referred to as
31525 @emph{32-bit code}.
31526 For applications intending to exploit the full 64-bit address space,
31527 @ref{Taking advantage of 64 bit addressing},
31528 will consider further changes that may be required.
31529 Such code will be referred to below as @emph{64-bit code}.
31531 @node Migration of 32 bit code
31532 @subsection Migration of 32-bit code
31537 * Unchecked conversions::
31538 * Predefined constants::
31539 * Interfacing with C::
31540 * Experience with source compatibility::
31543 @node Address types
31544 @subsubsection Address types
31547 To solve the problem of mixing 64-bit and 32-bit addressing,
31548 while maintaining maximum backward compatibility, the following
31549 approach has been taken:
31553 @code{System.Address} always has a size of 64 bits
31556 @code{System.Short_Address} is a 32-bit subtype of @code{System.Address}
31560 Since @code{System.Short_Address} is a subtype of @code{System.Address},
31561 a @code{Short_Address}
31562 may be used where an @code{Address} is required, and vice versa, without
31563 needing explicit type conversions.
31564 By virtue of the Open VMS parameter passing conventions,
31566 and exported subprograms that have 32-bit address parameters are
31567 compatible with those that have 64-bit address parameters.
31568 (See @ref{Making code 64 bit clean} for details.)
31570 The areas that may need attention are those where record types have
31571 been defined that contain components of the type @code{System.Address}, and
31572 where objects of this type are passed to code expecting a record layout with
31575 Different compilers on different platforms cannot be
31576 expected to represent the same type in the same way,
31577 since alignment constraints
31578 and other system-dependent properties affect the compiler's decision.
31579 For that reason, Ada code
31580 generally uses representation clauses to specify the expected
31581 layout where required.
31583 If such a representation clause uses 32 bits for a component having
31584 the type @code{System.Address}, 64-bit @value{EDITION} for OpenVMS
31585 will detect that error and produce a specific diagnostic message.
31586 The developer should then determine whether the representation
31587 should be 64 bits or not and make either of two changes:
31588 change the size to 64 bits and leave the type as @code{System.Address}, or
31589 leave the size as 32 bits and change the type to @code{System.Short_Address}.
31590 Since @code{Short_Address} is a subtype of @code{Address}, no changes are
31591 required in any code setting or accessing the field; the compiler will
31592 automatically perform any needed conversions between address
31596 @subsubsection Access types
31599 By default, objects designated by access values are always
31600 allocated in the 32-bit
31601 address space. Thus legacy code will never contain
31602 any objects that are not addressable with 32-bit addresses, and
31603 the compiler will never raise exceptions as result of mixing
31604 32-bit and 64-bit addresses.
31606 However, the access values themselves are represented in 64 bits, for optimum
31607 performance and future compatibility with 64-bit code. As was
31608 the case with @code{System.Address}, the compiler will give an error message
31609 if an object or record component has a representation clause that
31610 requires the access value to fit in 32 bits. In such a situation,
31611 an explicit size clause for the access type, specifying 32 bits,
31612 will have the desired effect.
31614 General access types (declared with @code{access all}) can never be
31615 32 bits, as values of such types must be able to refer to any object
31616 of the designated type,
31617 including objects residing outside the 32-bit address range.
31618 Existing Ada 83 code will not contain such type definitions,
31619 however, since general access types were introduced in Ada 95.
31621 @node Unchecked conversions
31622 @subsubsection Unchecked conversions
31625 In the case of an @code{Unchecked_Conversion} where the source type is a
31626 64-bit access type or the type @code{System.Address}, and the target
31627 type is a 32-bit type, the compiler will generate a warning.
31628 Even though the generated code will still perform the required
31629 conversions, it is highly recommended in these cases to use
31630 respectively a 32-bit access type or @code{System.Short_Address}
31631 as the source type.
31633 @node Predefined constants
31634 @subsubsection Predefined constants
31637 The following table shows the correspondence between pre-2006 versions of
31638 @value{EDITION} on Alpha OpenVMS (``Old'') and 64-bit @value{EDITION}
31641 @multitable {@code{System.Short_Memory_Size}} {2**32} {2**64}
31642 @item @b{Constant} @tab @b{Old} @tab @b{New}
31643 @item @code{System.Word_Size} @tab 32 @tab 64
31644 @item @code{System.Memory_Size} @tab 2**32 @tab 2**64
31645 @item @code{System.Short_Memory_Size} @tab 2**32 @tab 2**32
31646 @item @code{System.Address_Size} @tab 32 @tab 64
31650 If you need to refer to the specific
31651 memory size of a 32-bit implementation, instead of the
31652 actual memory size, use @code{System.Short_Memory_Size}
31653 rather than @code{System.Memory_Size}.
31654 Similarly, references to @code{System.Address_Size} may need
31655 to be replaced by @code{System.Short_Address'Size}.
31656 The program @command{gnatfind} may be useful for locating
31657 references to the above constants, so that you can verify that they
31660 @node Interfacing with C
31661 @subsubsection Interfacing with C
31664 In order to minimize the impact of the transition to 64-bit addresses on
31665 legacy programs, some fundamental types in the @code{Interfaces.C}
31666 package hierarchy continue to be represented in 32 bits.
31667 These types are: @code{ptrdiff_t}, @code{size_t}, and @code{chars_ptr}.
31668 This eases integration with the default HP C layout choices, for example
31669 as found in the system routines in @code{DECC$SHR.EXE}.
31670 Because of this implementation choice, the type fully compatible with
31671 @code{chars_ptr} is now @code{Short_Address} and not @code{Address}.
31672 Depending on the context the compiler will issue a
31673 warning or an error when type @code{Address} is used, alerting the user to a
31674 potential problem. Otherwise 32-bit programs that use
31675 @code{Interfaces.C} should normally not require code modifications
31677 The other issue arising with C interfacing concerns pragma @code{Convention}.
31678 For VMS 64-bit systems, there is an issue of the appropriate default size
31679 of C convention pointers in the absence of an explicit size clause. The HP
31680 C compiler can choose either 32 or 64 bits depending on compiler options.
31681 GNAT chooses 32-bits rather than 64-bits in the default case where no size
31682 clause is given. This proves a better choice for porting 32-bit legacy
31683 applications. In order to have a 64-bit representation, it is necessary to
31684 specify a size representation clause. For example:
31686 @smallexample @c ada
31687 type int_star is access Interfaces.C.int;
31688 pragma Convention(C, int_star);
31689 for int_star'Size use 64; -- Necessary to get 64 and not 32 bits
31692 @node Experience with source compatibility
31693 @subsubsection Experience with source compatibility
31696 The Security Server and STARLET on I64 provide an interesting ``test case''
31697 for source compatibility issues, since it is in such system code
31698 where assumptions about @code{Address} size might be expected to occur.
31699 Indeed, there were a small number of occasions in the Security Server
31700 file @file{jibdef.ads}
31701 where a representation clause for a record type specified
31702 32 bits for a component of type @code{Address}.
31703 All of these errors were detected by the compiler.
31704 The repair was obvious and immediate; to simply replace @code{Address} by
31705 @code{Short_Address}.
31707 In the case of STARLET, there were several record types that should
31708 have had representation clauses but did not. In these record types
31709 there was an implicit assumption that an @code{Address} value occupied
31711 These compiled without error, but their usage resulted in run-time error
31712 returns from STARLET system calls.
31713 Future GNAT technology enhancements may include a tool that detects and flags
31714 these sorts of potential source code porting problems.
31716 @c ****************************************
31717 @node Taking advantage of 64 bit addressing
31718 @subsection Taking advantage of 64-bit addressing
31721 * Making code 64 bit clean::
31722 * Allocating memory from the 64 bit storage pool::
31723 * Restrictions on use of 64 bit objects::
31724 * Using 64 bit storage pools by default::
31725 * General access types::
31726 * STARLET and other predefined libraries::
31729 @node Making code 64 bit clean
31730 @subsubsection Making code 64-bit clean
31733 In order to prevent problems that may occur when (parts of) a
31734 system start using memory outside the 32-bit address range,
31735 we recommend some additional guidelines:
31739 For imported subprograms that take parameters of the
31740 type @code{System.Address}, ensure that these subprograms can
31741 indeed handle 64-bit addresses. If not, or when in doubt,
31742 change the subprogram declaration to specify
31743 @code{System.Short_Address} instead.
31746 Resolve all warnings related to size mismatches in
31747 unchecked conversions. Failing to do so causes
31748 erroneous execution if the source object is outside
31749 the 32-bit address space.
31752 (optional) Explicitly use the 32-bit storage pool
31753 for access types used in a 32-bit context, or use
31754 generic access types where possible
31755 (@pxref{Restrictions on use of 64 bit objects}).
31759 If these rules are followed, the compiler will automatically insert
31760 any necessary checks to ensure that no addresses or access values
31761 passed to 32-bit code ever refer to objects outside the 32-bit
31763 Any attempt to do this will raise @code{Constraint_Error}.
31765 @node Allocating memory from the 64 bit storage pool
31766 @subsubsection Allocating memory from the 64-bit storage pool
31769 For any access type @code{T} that potentially requires memory allocations
31770 beyond the 32-bit address space,
31771 use the following representation clause:
31773 @smallexample @c ada
31774 for T'Storage_Pool use System.Pool_64;
31777 @node Restrictions on use of 64 bit objects
31778 @subsubsection Restrictions on use of 64-bit objects
31781 Taking the address of an object allocated from a 64-bit storage pool,
31782 and then passing this address to a subprogram expecting
31783 @code{System.Short_Address},
31784 or assigning it to a variable of type @code{Short_Address}, will cause
31785 @code{Constraint_Error} to be raised. In case the code is not 64-bit clean
31786 (@pxref{Making code 64 bit clean}), or checks are suppressed,
31787 no exception is raised and execution
31788 will become erroneous.
31790 @node Using 64 bit storage pools by default
31791 @subsubsection Using 64-bit storage pools by default
31794 In some cases it may be desirable to have the compiler allocate
31795 from 64-bit storage pools by default. This may be the case for
31796 libraries that are 64-bit clean, but may be used in both 32-bit
31797 and 64-bit contexts. For these cases the following configuration
31798 pragma may be specified:
31800 @smallexample @c ada
31801 pragma Pool_64_Default;
31805 Any code compiled in the context of this pragma will by default
31806 use the @code{System.Pool_64} storage pool. This default may be overridden
31807 for a specific access type @code{T} by the representation clause:
31809 @smallexample @c ada
31810 for T'Storage_Pool use System.Pool_32;
31814 Any object whose address may be passed to a subprogram with a
31815 @code{Short_Address} argument, or assigned to a variable of type
31816 @code{Short_Address}, needs to be allocated from this pool.
31818 @node General access types
31819 @subsubsection General access types
31822 Objects designated by access values from a
31823 general access type (declared with @code{access all}) are never allocated
31824 from a 64-bit storage pool. Code that uses general access types will
31825 accept objects allocated in either 32-bit or 64-bit address spaces,
31826 but never allocate objects outside the 32-bit address space.
31827 Using general access types ensures maximum compatibility with both
31828 32-bit and 64-bit code.
31830 @node STARLET and other predefined libraries
31831 @subsubsection STARLET and other predefined libraries
31834 All code that comes as part of GNAT is 64-bit clean, but the
31835 restrictions given in @ref{Restrictions on use of 64 bit objects},
31836 still apply. Look at the package
31837 specs to see in which contexts objects allocated
31838 in 64-bit address space are acceptable.
31840 @node Technical details
31841 @subsection Technical details
31844 64-bit @value{EDITION} for Open VMS takes advantage of the freedom given in the
31845 Ada standard with respect to the type of @code{System.Address}. Previous
31846 versions of GNAT Pro have defined this type as private and implemented it as a
31849 In order to allow defining @code{System.Short_Address} as a proper subtype,
31850 and to match the implicit sign extension in parameter passing,
31851 in 64-bit @value{EDITION} for Open VMS, @code{System.Address} is defined as a
31852 visible (i.e., non-private) integer type.
31853 Standard operations on the type, such as the binary operators ``+'', ``-'',
31854 etc., that take @code{Address} operands and return an @code{Address} result,
31855 have been hidden by declaring these
31856 @code{abstract}, a feature introduced in Ada 95 that helps avoid the potential
31857 ambiguities that would otherwise result from overloading.
31858 (Note that, although @code{Address} is a visible integer type,
31859 good programming practice dictates against exploiting the type's
31860 integer properties such as literals, since this will compromise
31863 Defining @code{Address} as a visible integer type helps achieve
31864 maximum compatibility for existing Ada code,
31865 without sacrificing the capabilities of the 64-bit architecture.
31868 @c ************************************************
31870 @node Microsoft Windows Topics
31871 @appendix Microsoft Windows Topics
31877 This chapter describes topics that are specific to the Microsoft Windows
31878 platforms (NT, 2000, and XP Professional).
31881 * Using GNAT on Windows::
31882 * Using a network installation of GNAT::
31883 * CONSOLE and WINDOWS subsystems::
31884 * Temporary Files::
31885 * Mixed-Language Programming on Windows::
31886 * Windows Calling Conventions::
31887 * Introduction to Dynamic Link Libraries (DLLs)::
31888 * Using DLLs with GNAT::
31889 * Building DLLs with GNAT::
31890 * Building DLLs with GNAT Project files::
31891 * Building DLLs with gnatdll::
31892 * GNAT and Windows Resources::
31893 * Debugging a DLL::
31894 * Setting Stack Size from gnatlink::
31895 * Setting Heap Size from gnatlink::
31898 @node Using GNAT on Windows
31899 @section Using GNAT on Windows
31902 One of the strengths of the GNAT technology is that its tool set
31903 (@command{gcc}, @command{gnatbind}, @command{gnatlink}, @command{gnatmake}, the
31904 @code{gdb} debugger, etc.) is used in the same way regardless of the
31907 On Windows this tool set is complemented by a number of Microsoft-specific
31908 tools that have been provided to facilitate interoperability with Windows
31909 when this is required. With these tools:
31914 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
31918 You can use any Dynamically Linked Library (DLL) in your Ada code (both
31919 relocatable and non-relocatable DLLs are supported).
31922 You can build Ada DLLs for use in other applications. These applications
31923 can be written in a language other than Ada (e.g., C, C++, etc). Again both
31924 relocatable and non-relocatable Ada DLLs are supported.
31927 You can include Windows resources in your Ada application.
31930 You can use or create COM/DCOM objects.
31934 Immediately below are listed all known general GNAT-for-Windows restrictions.
31935 Other restrictions about specific features like Windows Resources and DLLs
31936 are listed in separate sections below.
31941 It is not possible to use @code{GetLastError} and @code{SetLastError}
31942 when tasking, protected records, or exceptions are used. In these
31943 cases, in order to implement Ada semantics, the GNAT run-time system
31944 calls certain Win32 routines that set the last error variable to 0 upon
31945 success. It should be possible to use @code{GetLastError} and
31946 @code{SetLastError} when tasking, protected record, and exception
31947 features are not used, but it is not guaranteed to work.
31950 It is not possible to link against Microsoft libraries except for
31951 import libraries. The library must be built to be compatible with
31952 @file{MSVCRT.LIB} (/MD Microsoft compiler option), @file{LIBC.LIB} and
31953 @file{LIBCMT.LIB} (/ML or /MT Microsoft compiler options) are known to
31954 not be compatible with the GNAT runtime. Even if the library is
31955 compatible with @file{MSVCRT.LIB} it is not guaranteed to work.
31958 When the compilation environment is located on FAT32 drives, users may
31959 experience recompilations of the source files that have not changed if
31960 Daylight Saving Time (DST) state has changed since the last time files
31961 were compiled. NTFS drives do not have this problem.
31964 No components of the GNAT toolset use any entries in the Windows
31965 registry. The only entries that can be created are file associations and
31966 PATH settings, provided the user has chosen to create them at installation
31967 time, as well as some minimal book-keeping information needed to correctly
31968 uninstall or integrate different GNAT products.
31971 @node Using a network installation of GNAT
31972 @section Using a network installation of GNAT
31975 Make sure the system on which GNAT is installed is accessible from the
31976 current machine, i.e., the install location is shared over the network.
31977 Shared resources are accessed on Windows by means of UNC paths, which
31978 have the format @code{\\server\sharename\path}
31980 In order to use such a network installation, simply add the UNC path of the
31981 @file{bin} directory of your GNAT installation in front of your PATH. For
31982 example, if GNAT is installed in @file{\GNAT} directory of a share location
31983 called @file{c-drive} on a machine @file{LOKI}, the following command will
31986 @code{@ @ @ path \\loki\c-drive\gnat\bin;%path%}
31988 Be aware that every compilation using the network installation results in the
31989 transfer of large amounts of data across the network and will likely cause
31990 serious performance penalty.
31992 @node CONSOLE and WINDOWS subsystems
31993 @section CONSOLE and WINDOWS subsystems
31994 @cindex CONSOLE Subsystem
31995 @cindex WINDOWS Subsystem
31999 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
32000 (which is the default subsystem) will always create a console when
32001 launching the application. This is not something desirable when the
32002 application has a Windows GUI. To get rid of this console the
32003 application must be using the @code{WINDOWS} subsystem. To do so
32004 the @option{-mwindows} linker option must be specified.
32007 $ gnatmake winprog -largs -mwindows
32010 @node Temporary Files
32011 @section Temporary Files
32012 @cindex Temporary files
32015 It is possible to control where temporary files gets created by setting
32016 the @env{TMP} environment variable. The file will be created:
32019 @item Under the directory pointed to by the @env{TMP} environment variable if
32020 this directory exists.
32022 @item Under @file{c:\temp}, if the @env{TMP} environment variable is not
32023 set (or not pointing to a directory) and if this directory exists.
32025 @item Under the current working directory otherwise.
32029 This allows you to determine exactly where the temporary
32030 file will be created. This is particularly useful in networked
32031 environments where you may not have write access to some
32034 @node Mixed-Language Programming on Windows
32035 @section Mixed-Language Programming on Windows
32038 Developing pure Ada applications on Windows is no different than on
32039 other GNAT-supported platforms. However, when developing or porting an
32040 application that contains a mix of Ada and C/C++, the choice of your
32041 Windows C/C++ development environment conditions your overall
32042 interoperability strategy.
32044 If you use @command{gcc} to compile the non-Ada part of your application,
32045 there are no Windows-specific restrictions that affect the overall
32046 interoperability with your Ada code. If you plan to use
32047 Microsoft tools (e.g.@: Microsoft Visual C/C++), you should be aware of
32048 the following limitations:
32052 You cannot link your Ada code with an object or library generated with
32053 Microsoft tools if these use the @code{.tls} section (Thread Local
32054 Storage section) since the GNAT linker does not yet support this section.
32057 You cannot link your Ada code with an object or library generated with
32058 Microsoft tools if these use I/O routines other than those provided in
32059 the Microsoft DLL: @code{msvcrt.dll}. This is because the GNAT run time
32060 uses the services of @code{msvcrt.dll} for its I/Os. Use of other I/O
32061 libraries can cause a conflict with @code{msvcrt.dll} services. For
32062 instance Visual C++ I/O stream routines conflict with those in
32067 If you do want to use the Microsoft tools for your non-Ada code and hit one
32068 of the above limitations, you have two choices:
32072 Encapsulate your non-Ada code in a DLL to be linked with your Ada
32073 application. In this case, use the Microsoft or whatever environment to
32074 build the DLL and use GNAT to build your executable
32075 (@pxref{Using DLLs with GNAT}).
32078 Or you can encapsulate your Ada code in a DLL to be linked with the
32079 other part of your application. In this case, use GNAT to build the DLL
32080 (@pxref{Building DLLs with GNAT}) and use the Microsoft or whatever
32081 environment to build your executable.
32084 @node Windows Calling Conventions
32085 @section Windows Calling Conventions
32090 * C Calling Convention::
32091 * Stdcall Calling Convention::
32092 * Win32 Calling Convention::
32093 * DLL Calling Convention::
32097 When a subprogram @code{F} (caller) calls a subprogram @code{G}
32098 (callee), there are several ways to push @code{G}'s parameters on the
32099 stack and there are several possible scenarios to clean up the stack
32100 upon @code{G}'s return. A calling convention is an agreed upon software
32101 protocol whereby the responsibilities between the caller (@code{F}) and
32102 the callee (@code{G}) are clearly defined. Several calling conventions
32103 are available for Windows:
32107 @code{C} (Microsoft defined)
32110 @code{Stdcall} (Microsoft defined)
32113 @code{Win32} (GNAT specific)
32116 @code{DLL} (GNAT specific)
32119 @node C Calling Convention
32120 @subsection @code{C} Calling Convention
32123 This is the default calling convention used when interfacing to C/C++
32124 routines compiled with either @command{gcc} or Microsoft Visual C++.
32126 In the @code{C} calling convention subprogram parameters are pushed on the
32127 stack by the caller from right to left. The caller itself is in charge of
32128 cleaning up the stack after the call. In addition, the name of a routine
32129 with @code{C} calling convention is mangled by adding a leading underscore.
32131 The name to use on the Ada side when importing (or exporting) a routine
32132 with @code{C} calling convention is the name of the routine. For
32133 instance the C function:
32136 int get_val (long);
32140 should be imported from Ada as follows:
32142 @smallexample @c ada
32144 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
32145 pragma Import (C, Get_Val, External_Name => "get_val");
32150 Note that in this particular case the @code{External_Name} parameter could
32151 have been omitted since, when missing, this parameter is taken to be the
32152 name of the Ada entity in lower case. When the @code{Link_Name} parameter
32153 is missing, as in the above example, this parameter is set to be the
32154 @code{External_Name} with a leading underscore.
32156 When importing a variable defined in C, you should always use the @code{C}
32157 calling convention unless the object containing the variable is part of a
32158 DLL (in which case you should use the @code{Stdcall} calling
32159 convention, @pxref{Stdcall Calling Convention}).
32161 @node Stdcall Calling Convention
32162 @subsection @code{Stdcall} Calling Convention
32165 This convention, which was the calling convention used for Pascal
32166 programs, is used by Microsoft for all the routines in the Win32 API for
32167 efficiency reasons. It must be used to import any routine for which this
32168 convention was specified.
32170 In the @code{Stdcall} calling convention subprogram parameters are pushed
32171 on the stack by the caller from right to left. The callee (and not the
32172 caller) is in charge of cleaning the stack on routine exit. In addition,
32173 the name of a routine with @code{Stdcall} calling convention is mangled by
32174 adding a leading underscore (as for the @code{C} calling convention) and a
32175 trailing @code{@@}@code{@var{nn}}, where @var{nn} is the overall size (in
32176 bytes) of the parameters passed to the routine.
32178 The name to use on the Ada side when importing a C routine with a
32179 @code{Stdcall} calling convention is the name of the C routine. The leading
32180 underscore and trailing @code{@@}@code{@var{nn}} are added automatically by
32181 the compiler. For instance the Win32 function:
32184 @b{APIENTRY} int get_val (long);
32188 should be imported from Ada as follows:
32190 @smallexample @c ada
32192 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
32193 pragma Import (Stdcall, Get_Val);
32194 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
32199 As for the @code{C} calling convention, when the @code{External_Name}
32200 parameter is missing, it is taken to be the name of the Ada entity in lower
32201 case. If instead of writing the above import pragma you write:
32203 @smallexample @c ada
32205 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
32206 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
32211 then the imported routine is @code{_retrieve_val@@4}. However, if instead
32212 of specifying the @code{External_Name} parameter you specify the
32213 @code{Link_Name} as in the following example:
32215 @smallexample @c ada
32217 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
32218 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
32223 then the imported routine is @code{retrieve_val}, that is, there is no
32224 decoration at all. No leading underscore and no Stdcall suffix
32225 @code{@@}@code{@var{nn}}.
32228 This is especially important as in some special cases a DLL's entry
32229 point name lacks a trailing @code{@@}@code{@var{nn}} while the exported
32230 name generated for a call has it.
32233 It is also possible to import variables defined in a DLL by using an
32234 import pragma for a variable. As an example, if a DLL contains a
32235 variable defined as:
32242 then, to access this variable from Ada you should write:
32244 @smallexample @c ada
32246 My_Var : Interfaces.C.int;
32247 pragma Import (Stdcall, My_Var);
32252 Note that to ease building cross-platform bindings this convention
32253 will be handled as a @code{C} calling convention on non-Windows platforms.
32255 @node Win32 Calling Convention
32256 @subsection @code{Win32} Calling Convention
32259 This convention, which is GNAT-specific is fully equivalent to the
32260 @code{Stdcall} calling convention described above.
32262 @node DLL Calling Convention
32263 @subsection @code{DLL} Calling Convention
32266 This convention, which is GNAT-specific is fully equivalent to the
32267 @code{Stdcall} calling convention described above.
32269 @node Introduction to Dynamic Link Libraries (DLLs)
32270 @section Introduction to Dynamic Link Libraries (DLLs)
32274 A Dynamically Linked Library (DLL) is a library that can be shared by
32275 several applications running under Windows. A DLL can contain any number of
32276 routines and variables.
32278 One advantage of DLLs is that you can change and enhance them without
32279 forcing all the applications that depend on them to be relinked or
32280 recompiled. However, you should be aware than all calls to DLL routines are
32281 slower since, as you will understand below, such calls are indirect.
32283 To illustrate the remainder of this section, suppose that an application
32284 wants to use the services of a DLL @file{API.dll}. To use the services
32285 provided by @file{API.dll} you must statically link against the DLL or
32286 an import library which contains a jump table with an entry for each
32287 routine and variable exported by the DLL. In the Microsoft world this
32288 import library is called @file{API.lib}. When using GNAT this import
32289 library is called either @file{libAPI.dll.a}, @file{libapi.dll.a},
32290 @file{libAPI.a} or @file{libapi.a} (names are case insensitive).
32292 After you have linked your application with the DLL or the import library
32293 and you run your application, here is what happens:
32297 Your application is loaded into memory.
32300 The DLL @file{API.dll} is mapped into the address space of your
32301 application. This means that:
32305 The DLL will use the stack of the calling thread.
32308 The DLL will use the virtual address space of the calling process.
32311 The DLL will allocate memory from the virtual address space of the calling
32315 Handles (pointers) can be safely exchanged between routines in the DLL
32316 routines and routines in the application using the DLL.
32320 The entries in the jump table (from the import library @file{libAPI.dll.a}
32321 or @file{API.lib} or automatically created when linking against a DLL)
32322 which is part of your application are initialized with the addresses
32323 of the routines and variables in @file{API.dll}.
32326 If present in @file{API.dll}, routines @code{DllMain} or
32327 @code{DllMainCRTStartup} are invoked. These routines typically contain
32328 the initialization code needed for the well-being of the routines and
32329 variables exported by the DLL.
32333 There is an additional point which is worth mentioning. In the Windows
32334 world there are two kind of DLLs: relocatable and non-relocatable
32335 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
32336 in the target application address space. If the addresses of two
32337 non-relocatable DLLs overlap and these happen to be used by the same
32338 application, a conflict will occur and the application will run
32339 incorrectly. Hence, when possible, it is always preferable to use and
32340 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
32341 supported by GNAT. Note that the @option{-s} linker option (see GNU Linker
32342 User's Guide) removes the debugging symbols from the DLL but the DLL can
32343 still be relocated.
32345 As a side note, an interesting difference between Microsoft DLLs and
32346 Unix shared libraries, is the fact that on most Unix systems all public
32347 routines are exported by default in a Unix shared library, while under
32348 Windows it is possible (but not required) to list exported routines in
32349 a definition file (@pxref{The Definition File}).
32351 @node Using DLLs with GNAT
32352 @section Using DLLs with GNAT
32355 * Creating an Ada Spec for the DLL Services::
32356 * Creating an Import Library::
32360 To use the services of a DLL, say @file{API.dll}, in your Ada application
32365 The Ada spec for the routines and/or variables you want to access in
32366 @file{API.dll}. If not available this Ada spec must be built from the C/C++
32367 header files provided with the DLL.
32370 The import library (@file{libAPI.dll.a} or @file{API.lib}). As previously
32371 mentioned an import library is a statically linked library containing the
32372 import table which will be filled at load time to point to the actual
32373 @file{API.dll} routines. Sometimes you don't have an import library for the
32374 DLL you want to use. The following sections will explain how to build
32375 one. Note that this is optional.
32378 The actual DLL, @file{API.dll}.
32382 Once you have all the above, to compile an Ada application that uses the
32383 services of @file{API.dll} and whose main subprogram is @code{My_Ada_App},
32384 you simply issue the command
32387 $ gnatmake my_ada_app -largs -lAPI
32391 The argument @option{-largs -lAPI} at the end of the @command{gnatmake} command
32392 tells the GNAT linker to look first for a library named @file{API.lib}
32393 (Microsoft-style name) and if not found for a libraries named
32394 @file{libAPI.dll.a}, @file{API.dll.a} or @file{libAPI.a}.
32395 (GNAT-style name). Note that if the Ada package spec for @file{API.dll}
32396 contains the following pragma
32398 @smallexample @c ada
32399 pragma Linker_Options ("-lAPI");
32403 you do not have to add @option{-largs -lAPI} at the end of the
32404 @command{gnatmake} command.
32406 If any one of the items above is missing you will have to create it
32407 yourself. The following sections explain how to do so using as an
32408 example a fictitious DLL called @file{API.dll}.
32410 @node Creating an Ada Spec for the DLL Services
32411 @subsection Creating an Ada Spec for the DLL Services
32414 A DLL typically comes with a C/C++ header file which provides the
32415 definitions of the routines and variables exported by the DLL. The Ada
32416 equivalent of this header file is a package spec that contains definitions
32417 for the imported entities. If the DLL you intend to use does not come with
32418 an Ada spec you have to generate one such spec yourself. For example if
32419 the header file of @file{API.dll} is a file @file{api.h} containing the
32420 following two definitions:
32432 then the equivalent Ada spec could be:
32434 @smallexample @c ada
32437 with Interfaces.C.Strings;
32442 function Get (Str : C.Strings.Chars_Ptr) return C.int;
32445 pragma Import (C, Get);
32446 pragma Import (DLL, Some_Var);
32453 Note that a variable is
32454 @strong{always imported with a Stdcall convention}. A function
32455 can have @code{C} or @code{Stdcall} convention.
32456 (@pxref{Windows Calling Conventions}).
32458 @node Creating an Import Library
32459 @subsection Creating an Import Library
32460 @cindex Import library
32463 * The Definition File::
32464 * GNAT-Style Import Library::
32465 * Microsoft-Style Import Library::
32469 If a Microsoft-style import library @file{API.lib} or a GNAT-style
32470 import library @file{libAPI.dll.a} or @file{libAPI.a} is available
32471 with @file{API.dll} you can skip this section. You can also skip this
32472 section if @file{API.dll} or @file{libAPI.dll} is built with GNU tools
32473 as in this case it is possible to link directly against the
32474 DLL. Otherwise read on.
32476 @node The Definition File
32477 @subsubsection The Definition File
32478 @cindex Definition file
32482 As previously mentioned, and unlike Unix systems, the list of symbols
32483 that are exported from a DLL must be provided explicitly in Windows.
32484 The main goal of a definition file is precisely that: list the symbols
32485 exported by a DLL. A definition file (usually a file with a @code{.def}
32486 suffix) has the following structure:
32491 @r{[}LIBRARY @var{name}@r{]}
32492 @r{[}DESCRIPTION @var{string}@r{]}
32502 @item LIBRARY @var{name}
32503 This section, which is optional, gives the name of the DLL.
32505 @item DESCRIPTION @var{string}
32506 This section, which is optional, gives a description string that will be
32507 embedded in the import library.
32510 This section gives the list of exported symbols (procedures, functions or
32511 variables). For instance in the case of @file{API.dll} the @code{EXPORTS}
32512 section of @file{API.def} looks like:
32526 Note that you must specify the correct suffix (@code{@@}@code{@var{nn}})
32527 (@pxref{Windows Calling Conventions}) for a Stdcall
32528 calling convention function in the exported symbols list.
32531 There can actually be other sections in a definition file, but these
32532 sections are not relevant to the discussion at hand.
32534 @node GNAT-Style Import Library
32535 @subsubsection GNAT-Style Import Library
32538 To create a static import library from @file{API.dll} with the GNAT tools
32539 you should proceed as follows:
32543 Create the definition file @file{API.def} (@pxref{The Definition File}).
32544 For that use the @code{dll2def} tool as follows:
32547 $ dll2def API.dll > API.def
32551 @code{dll2def} is a very simple tool: it takes as input a DLL and prints
32552 to standard output the list of entry points in the DLL. Note that if
32553 some routines in the DLL have the @code{Stdcall} convention
32554 (@pxref{Windows Calling Conventions}) with stripped @code{@@}@var{nn}
32555 suffix then you'll have to edit @file{api.def} to add it, and specify
32556 @option{-k} to @command{gnatdll} when creating the import library.
32559 Here are some hints to find the right @code{@@}@var{nn} suffix.
32563 If you have the Microsoft import library (.lib), it is possible to get
32564 the right symbols by using Microsoft @code{dumpbin} tool (see the
32565 corresponding Microsoft documentation for further details).
32568 $ dumpbin /exports api.lib
32572 If you have a message about a missing symbol at link time the compiler
32573 tells you what symbol is expected. You just have to go back to the
32574 definition file and add the right suffix.
32578 Build the import library @code{libAPI.dll.a}, using @code{gnatdll}
32579 (@pxref{Using gnatdll}) as follows:
32582 $ gnatdll -e API.def -d API.dll
32586 @code{gnatdll} takes as input a definition file @file{API.def} and the
32587 name of the DLL containing the services listed in the definition file
32588 @file{API.dll}. The name of the static import library generated is
32589 computed from the name of the definition file as follows: if the
32590 definition file name is @var{xyz}@code{.def}, the import library name will
32591 be @code{lib}@var{xyz}@code{.a}. Note that in the previous example option
32592 @option{-e} could have been removed because the name of the definition
32593 file (before the ``@code{.def}'' suffix) is the same as the name of the
32594 DLL (@pxref{Using gnatdll} for more information about @code{gnatdll}).
32597 @node Microsoft-Style Import Library
32598 @subsubsection Microsoft-Style Import Library
32601 With GNAT you can either use a GNAT-style or Microsoft-style import
32602 library. A Microsoft import library is needed only if you plan to make an
32603 Ada DLL available to applications developed with Microsoft
32604 tools (@pxref{Mixed-Language Programming on Windows}).
32606 To create a Microsoft-style import library for @file{API.dll} you
32607 should proceed as follows:
32611 Create the definition file @file{API.def} from the DLL. For this use either
32612 the @code{dll2def} tool as described above or the Microsoft @code{dumpbin}
32613 tool (see the corresponding Microsoft documentation for further details).
32616 Build the actual import library using Microsoft's @code{lib} utility:
32619 $ lib -machine:IX86 -def:API.def -out:API.lib
32623 If you use the above command the definition file @file{API.def} must
32624 contain a line giving the name of the DLL:
32631 See the Microsoft documentation for further details about the usage of
32635 @node Building DLLs with GNAT
32636 @section Building DLLs with GNAT
32637 @cindex DLLs, building
32640 This section explain how to build DLLs using the GNAT built-in DLL
32641 support. With the following procedure it is straight forward to build
32642 and use DLLs with GNAT.
32646 @item building object files
32648 The first step is to build all objects files that are to be included
32649 into the DLL. This is done by using the standard @command{gnatmake} tool.
32651 @item building the DLL
32653 To build the DLL you must use @command{gcc}'s @option{-shared}
32654 option. It is quite simple to use this method:
32657 $ gcc -shared -o api.dll obj1.o obj2.o @dots{}
32660 It is important to note that in this case all symbols found in the
32661 object files are automatically exported. It is possible to restrict
32662 the set of symbols to export by passing to @command{gcc} a definition
32663 file, @pxref{The Definition File}. For example:
32666 $ gcc -shared -o api.dll api.def obj1.o obj2.o @dots{}
32669 If you use a definition file you must export the elaboration procedures
32670 for every package that required one. Elaboration procedures are named
32671 using the package name followed by "_E".
32673 @item preparing DLL to be used
32675 For the DLL to be used by client programs the bodies must be hidden
32676 from it and the .ali set with read-only attribute. This is very important
32677 otherwise GNAT will recompile all packages and will not actually use
32678 the code in the DLL. For example:
32682 $ copy *.ads *.ali api.dll apilib
32683 $ attrib +R apilib\*.ali
32688 At this point it is possible to use the DLL by directly linking
32689 against it. Note that you must use the GNAT shared runtime when using
32690 GNAT shared libraries. This is achieved by using @option{-shared} binder's
32694 $ gnatmake main -Iapilib -bargs -shared -largs -Lapilib -lAPI
32697 @node Building DLLs with GNAT Project files
32698 @section Building DLLs with GNAT Project files
32699 @cindex DLLs, building
32702 There is nothing specific to Windows in the build process.
32703 @pxref{Library Projects}.
32706 Due to a system limitation, it is not possible under Windows to create threads
32707 when inside the @code{DllMain} routine which is used for auto-initialization
32708 of shared libraries, so it is not possible to have library level tasks in SALs.
32710 @node Building DLLs with gnatdll
32711 @section Building DLLs with gnatdll
32712 @cindex DLLs, building
32715 * Limitations When Using Ada DLLs from Ada::
32716 * Exporting Ada Entities::
32717 * Ada DLLs and Elaboration::
32718 * Ada DLLs and Finalization::
32719 * Creating a Spec for Ada DLLs::
32720 * Creating the Definition File::
32725 Note that it is preferred to use the built-in GNAT DLL support
32726 (@pxref{Building DLLs with GNAT}) or GNAT Project files
32727 (@pxref{Building DLLs with GNAT Project files}) to build DLLs.
32729 This section explains how to build DLLs containing Ada code using
32730 @code{gnatdll}. These DLLs will be referred to as Ada DLLs in the
32731 remainder of this section.
32733 The steps required to build an Ada DLL that is to be used by Ada as well as
32734 non-Ada applications are as follows:
32738 You need to mark each Ada @i{entity} exported by the DLL with a @code{C} or
32739 @code{Stdcall} calling convention to avoid any Ada name mangling for the
32740 entities exported by the DLL (@pxref{Exporting Ada Entities}). You can
32741 skip this step if you plan to use the Ada DLL only from Ada applications.
32744 Your Ada code must export an initialization routine which calls the routine
32745 @code{adainit} generated by @command{gnatbind} to perform the elaboration of
32746 the Ada code in the DLL (@pxref{Ada DLLs and Elaboration}). The initialization
32747 routine exported by the Ada DLL must be invoked by the clients of the DLL
32748 to initialize the DLL.
32751 When useful, the DLL should also export a finalization routine which calls
32752 routine @code{adafinal} generated by @command{gnatbind} to perform the
32753 finalization of the Ada code in the DLL (@pxref{Ada DLLs and Finalization}).
32754 The finalization routine exported by the Ada DLL must be invoked by the
32755 clients of the DLL when the DLL services are no further needed.
32758 You must provide a spec for the services exported by the Ada DLL in each
32759 of the programming languages to which you plan to make the DLL available.
32762 You must provide a definition file listing the exported entities
32763 (@pxref{The Definition File}).
32766 Finally you must use @code{gnatdll} to produce the DLL and the import
32767 library (@pxref{Using gnatdll}).
32771 Note that a relocatable DLL stripped using the @code{strip}
32772 binutils tool will not be relocatable anymore. To build a DLL without
32773 debug information pass @code{-largs -s} to @code{gnatdll}. This
32774 restriction does not apply to a DLL built using a Library Project.
32775 @pxref{Library Projects}.
32777 @node Limitations When Using Ada DLLs from Ada
32778 @subsection Limitations When Using Ada DLLs from Ada
32781 When using Ada DLLs from Ada applications there is a limitation users
32782 should be aware of. Because on Windows the GNAT run time is not in a DLL of
32783 its own, each Ada DLL includes a part of the GNAT run time. Specifically,
32784 each Ada DLL includes the services of the GNAT run time that are necessary
32785 to the Ada code inside the DLL. As a result, when an Ada program uses an
32786 Ada DLL there are two independent GNAT run times: one in the Ada DLL and
32787 one in the main program.
32789 It is therefore not possible to exchange GNAT run-time objects between the
32790 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
32791 handles (e.g.@: @code{Text_IO.File_Type}), tasks types, protected objects
32794 It is completely safe to exchange plain elementary, array or record types,
32795 Windows object handles, etc.
32797 @node Exporting Ada Entities
32798 @subsection Exporting Ada Entities
32799 @cindex Export table
32802 Building a DLL is a way to encapsulate a set of services usable from any
32803 application. As a result, the Ada entities exported by a DLL should be
32804 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
32805 any Ada name mangling. As an example here is an Ada package
32806 @code{API}, spec and body, exporting two procedures, a function, and a
32809 @smallexample @c ada
32812 with Interfaces.C; use Interfaces;
32814 Count : C.int := 0;
32815 function Factorial (Val : C.int) return C.int;
32817 procedure Initialize_API;
32818 procedure Finalize_API;
32819 -- Initialization & Finalization routines. More in the next section.
32821 pragma Export (C, Initialize_API);
32822 pragma Export (C, Finalize_API);
32823 pragma Export (C, Count);
32824 pragma Export (C, Factorial);
32830 @smallexample @c ada
32833 package body API is
32834 function Factorial (Val : C.int) return C.int is
32837 Count := Count + 1;
32838 for K in 1 .. Val loop
32844 procedure Initialize_API is
32846 pragma Import (C, Adainit);
32849 end Initialize_API;
32851 procedure Finalize_API is
32852 procedure Adafinal;
32853 pragma Import (C, Adafinal);
32863 If the Ada DLL you are building will only be used by Ada applications
32864 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
32865 convention. As an example, the previous package could be written as
32868 @smallexample @c ada
32872 Count : Integer := 0;
32873 function Factorial (Val : Integer) return Integer;
32875 procedure Initialize_API;
32876 procedure Finalize_API;
32877 -- Initialization and Finalization routines.
32883 @smallexample @c ada
32886 package body API is
32887 function Factorial (Val : Integer) return Integer is
32888 Fact : Integer := 1;
32890 Count := Count + 1;
32891 for K in 1 .. Val loop
32898 -- The remainder of this package body is unchanged.
32905 Note that if you do not export the Ada entities with a @code{C} or
32906 @code{Stdcall} convention you will have to provide the mangled Ada names
32907 in the definition file of the Ada DLL
32908 (@pxref{Creating the Definition File}).
32910 @node Ada DLLs and Elaboration
32911 @subsection Ada DLLs and Elaboration
32912 @cindex DLLs and elaboration
32915 The DLL that you are building contains your Ada code as well as all the
32916 routines in the Ada library that are needed by it. The first thing a
32917 user of your DLL must do is elaborate the Ada code
32918 (@pxref{Elaboration Order Handling in GNAT}).
32920 To achieve this you must export an initialization routine
32921 (@code{Initialize_API} in the previous example), which must be invoked
32922 before using any of the DLL services. This elaboration routine must call
32923 the Ada elaboration routine @code{adainit} generated by the GNAT binder
32924 (@pxref{Binding with Non-Ada Main Programs}). See the body of
32925 @code{Initialize_Api} for an example. Note that the GNAT binder is
32926 automatically invoked during the DLL build process by the @code{gnatdll}
32927 tool (@pxref{Using gnatdll}).
32929 When a DLL is loaded, Windows systematically invokes a routine called
32930 @code{DllMain}. It would therefore be possible to call @code{adainit}
32931 directly from @code{DllMain} without having to provide an explicit
32932 initialization routine. Unfortunately, it is not possible to call
32933 @code{adainit} from the @code{DllMain} if your program has library level
32934 tasks because access to the @code{DllMain} entry point is serialized by
32935 the system (that is, only a single thread can execute ``through'' it at a
32936 time), which means that the GNAT run time will deadlock waiting for the
32937 newly created task to complete its initialization.
32939 @node Ada DLLs and Finalization
32940 @subsection Ada DLLs and Finalization
32941 @cindex DLLs and finalization
32944 When the services of an Ada DLL are no longer needed, the client code should
32945 invoke the DLL finalization routine, if available. The DLL finalization
32946 routine is in charge of releasing all resources acquired by the DLL. In the
32947 case of the Ada code contained in the DLL, this is achieved by calling
32948 routine @code{adafinal} generated by the GNAT binder
32949 (@pxref{Binding with Non-Ada Main Programs}).
32950 See the body of @code{Finalize_Api} for an
32951 example. As already pointed out the GNAT binder is automatically invoked
32952 during the DLL build process by the @code{gnatdll} tool
32953 (@pxref{Using gnatdll}).
32955 @node Creating a Spec for Ada DLLs
32956 @subsection Creating a Spec for Ada DLLs
32959 To use the services exported by the Ada DLL from another programming
32960 language (e.g.@: C), you have to translate the specs of the exported Ada
32961 entities in that language. For instance in the case of @code{API.dll},
32962 the corresponding C header file could look like:
32967 extern int *_imp__count;
32968 #define count (*_imp__count)
32969 int factorial (int);
32975 It is important to understand that when building an Ada DLL to be used by
32976 other Ada applications, you need two different specs for the packages
32977 contained in the DLL: one for building the DLL and the other for using
32978 the DLL. This is because the @code{DLL} calling convention is needed to
32979 use a variable defined in a DLL, but when building the DLL, the variable
32980 must have either the @code{Ada} or @code{C} calling convention. As an
32981 example consider a DLL comprising the following package @code{API}:
32983 @smallexample @c ada
32987 Count : Integer := 0;
32989 -- Remainder of the package omitted.
32996 After producing a DLL containing package @code{API}, the spec that
32997 must be used to import @code{API.Count} from Ada code outside of the
33000 @smallexample @c ada
33005 pragma Import (DLL, Count);
33011 @node Creating the Definition File
33012 @subsection Creating the Definition File
33015 The definition file is the last file needed to build the DLL. It lists
33016 the exported symbols. As an example, the definition file for a DLL
33017 containing only package @code{API} (where all the entities are exported
33018 with a @code{C} calling convention) is:
33033 If the @code{C} calling convention is missing from package @code{API},
33034 then the definition file contains the mangled Ada names of the above
33035 entities, which in this case are:
33044 api__initialize_api
33049 @node Using gnatdll
33050 @subsection Using @code{gnatdll}
33054 * gnatdll Example::
33055 * gnatdll behind the Scenes::
33060 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
33061 and non-Ada sources that make up your DLL have been compiled.
33062 @code{gnatdll} is actually in charge of two distinct tasks: build the
33063 static import library for the DLL and the actual DLL. The form of the
33064 @code{gnatdll} command is
33068 $ gnatdll @ovar{switches} @var{list-of-files} @r{[}-largs @var{opts}@r{]}
33073 where @var{list-of-files} is a list of ALI and object files. The object
33074 file list must be the exact list of objects corresponding to the non-Ada
33075 sources whose services are to be included in the DLL. The ALI file list
33076 must be the exact list of ALI files for the corresponding Ada sources
33077 whose services are to be included in the DLL. If @var{list-of-files} is
33078 missing, only the static import library is generated.
33081 You may specify any of the following switches to @code{gnatdll}:
33084 @item -a@ovar{address}
33085 @cindex @option{-a} (@code{gnatdll})
33086 Build a non-relocatable DLL at @var{address}. If @var{address} is not
33087 specified the default address @var{0x11000000} will be used. By default,
33088 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
33089 advise the reader to build relocatable DLL.
33091 @item -b @var{address}
33092 @cindex @option{-b} (@code{gnatdll})
33093 Set the relocatable DLL base address. By default the address is
33096 @item -bargs @var{opts}
33097 @cindex @option{-bargs} (@code{gnatdll})
33098 Binder options. Pass @var{opts} to the binder.
33100 @item -d @var{dllfile}
33101 @cindex @option{-d} (@code{gnatdll})
33102 @var{dllfile} is the name of the DLL. This switch must be present for
33103 @code{gnatdll} to do anything. The name of the generated import library is
33104 obtained algorithmically from @var{dllfile} as shown in the following
33105 example: if @var{dllfile} is @code{xyz.dll}, the import library name is
33106 @code{libxyz.dll.a}. The name of the definition file to use (if not specified
33107 by option @option{-e}) is obtained algorithmically from @var{dllfile}
33108 as shown in the following example:
33109 if @var{dllfile} is @code{xyz.dll}, the definition
33110 file used is @code{xyz.def}.
33112 @item -e @var{deffile}
33113 @cindex @option{-e} (@code{gnatdll})
33114 @var{deffile} is the name of the definition file.
33117 @cindex @option{-g} (@code{gnatdll})
33118 Generate debugging information. This information is stored in the object
33119 file and copied from there to the final DLL file by the linker,
33120 where it can be read by the debugger. You must use the
33121 @option{-g} switch if you plan on using the debugger or the symbolic
33125 @cindex @option{-h} (@code{gnatdll})
33126 Help mode. Displays @code{gnatdll} switch usage information.
33129 @cindex @option{-I} (@code{gnatdll})
33130 Direct @code{gnatdll} to search the @var{dir} directory for source and
33131 object files needed to build the DLL.
33132 (@pxref{Search Paths and the Run-Time Library (RTL)}).
33135 @cindex @option{-k} (@code{gnatdll})
33136 Removes the @code{@@}@var{nn} suffix from the import library's exported
33137 names, but keeps them for the link names. You must specify this
33138 option if you want to use a @code{Stdcall} function in a DLL for which
33139 the @code{@@}@var{nn} suffix has been removed. This is the case for most
33140 of the Windows NT DLL for example. This option has no effect when
33141 @option{-n} option is specified.
33143 @item -l @var{file}
33144 @cindex @option{-l} (@code{gnatdll})
33145 The list of ALI and object files used to build the DLL are listed in
33146 @var{file}, instead of being given in the command line. Each line in
33147 @var{file} contains the name of an ALI or object file.
33150 @cindex @option{-n} (@code{gnatdll})
33151 No Import. Do not create the import library.
33154 @cindex @option{-q} (@code{gnatdll})
33155 Quiet mode. Do not display unnecessary messages.
33158 @cindex @option{-v} (@code{gnatdll})
33159 Verbose mode. Display extra information.
33161 @item -largs @var{opts}
33162 @cindex @option{-largs} (@code{gnatdll})
33163 Linker options. Pass @var{opts} to the linker.
33166 @node gnatdll Example
33167 @subsubsection @code{gnatdll} Example
33170 As an example the command to build a relocatable DLL from @file{api.adb}
33171 once @file{api.adb} has been compiled and @file{api.def} created is
33174 $ gnatdll -d api.dll api.ali
33178 The above command creates two files: @file{libapi.dll.a} (the import
33179 library) and @file{api.dll} (the actual DLL). If you want to create
33180 only the DLL, just type:
33183 $ gnatdll -d api.dll -n api.ali
33187 Alternatively if you want to create just the import library, type:
33190 $ gnatdll -d api.dll
33193 @node gnatdll behind the Scenes
33194 @subsubsection @code{gnatdll} behind the Scenes
33197 This section details the steps involved in creating a DLL. @code{gnatdll}
33198 does these steps for you. Unless you are interested in understanding what
33199 goes on behind the scenes, you should skip this section.
33201 We use the previous example of a DLL containing the Ada package @code{API},
33202 to illustrate the steps necessary to build a DLL. The starting point is a
33203 set of objects that will make up the DLL and the corresponding ALI
33204 files. In the case of this example this means that @file{api.o} and
33205 @file{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
33210 @code{gnatdll} builds the base file (@file{api.base}). A base file gives
33211 the information necessary to generate relocation information for the
33217 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
33222 In addition to the base file, the @command{gnatlink} command generates an
33223 output file @file{api.jnk} which can be discarded. The @option{-mdll} switch
33224 asks @command{gnatlink} to generate the routines @code{DllMain} and
33225 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
33226 is loaded into memory.
33229 @code{gnatdll} uses @code{dlltool} (@pxref{Using dlltool}) to build the
33230 export table (@file{api.exp}). The export table contains the relocation
33231 information in a form which can be used during the final link to ensure
33232 that the Windows loader is able to place the DLL anywhere in memory.
33236 $ dlltool --dllname api.dll --def api.def --base-file api.base \
33237 --output-exp api.exp
33242 @code{gnatdll} builds the base file using the new export table. Note that
33243 @command{gnatbind} must be called once again since the binder generated file
33244 has been deleted during the previous call to @command{gnatlink}.
33249 $ gnatlink api -o api.jnk api.exp -mdll
33250 -Wl,--base-file,api.base
33255 @code{gnatdll} builds the new export table using the new base file and
33256 generates the DLL import library @file{libAPI.dll.a}.
33260 $ dlltool --dllname api.dll --def api.def --base-file api.base \
33261 --output-exp api.exp --output-lib libAPI.a
33266 Finally @code{gnatdll} builds the relocatable DLL using the final export
33272 $ gnatlink api api.exp -o api.dll -mdll
33277 @node Using dlltool
33278 @subsubsection Using @code{dlltool}
33281 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
33282 DLLs and static import libraries. This section summarizes the most
33283 common @code{dlltool} switches. The form of the @code{dlltool} command
33287 $ dlltool @ovar{switches}
33291 @code{dlltool} switches include:
33294 @item --base-file @var{basefile}
33295 @cindex @option{--base-file} (@command{dlltool})
33296 Read the base file @var{basefile} generated by the linker. This switch
33297 is used to create a relocatable DLL.
33299 @item --def @var{deffile}
33300 @cindex @option{--def} (@command{dlltool})
33301 Read the definition file.
33303 @item --dllname @var{name}
33304 @cindex @option{--dllname} (@command{dlltool})
33305 Gives the name of the DLL. This switch is used to embed the name of the
33306 DLL in the static import library generated by @code{dlltool} with switch
33307 @option{--output-lib}.
33310 @cindex @option{-k} (@command{dlltool})
33311 Kill @code{@@}@var{nn} from exported names
33312 (@pxref{Windows Calling Conventions}
33313 for a discussion about @code{Stdcall}-style symbols.
33316 @cindex @option{--help} (@command{dlltool})
33317 Prints the @code{dlltool} switches with a concise description.
33319 @item --output-exp @var{exportfile}
33320 @cindex @option{--output-exp} (@command{dlltool})
33321 Generate an export file @var{exportfile}. The export file contains the
33322 export table (list of symbols in the DLL) and is used to create the DLL.
33324 @item --output-lib @var{libfile}
33325 @cindex @option{--output-lib} (@command{dlltool})
33326 Generate a static import library @var{libfile}.
33329 @cindex @option{-v} (@command{dlltool})
33332 @item --as @var{assembler-name}
33333 @cindex @option{--as} (@command{dlltool})
33334 Use @var{assembler-name} as the assembler. The default is @code{as}.
33337 @node GNAT and Windows Resources
33338 @section GNAT and Windows Resources
33339 @cindex Resources, windows
33342 * Building Resources::
33343 * Compiling Resources::
33344 * Using Resources::
33348 Resources are an easy way to add Windows specific objects to your
33349 application. The objects that can be added as resources include:
33378 This section explains how to build, compile and use resources.
33380 @node Building Resources
33381 @subsection Building Resources
33382 @cindex Resources, building
33385 A resource file is an ASCII file. By convention resource files have an
33386 @file{.rc} extension.
33387 The easiest way to build a resource file is to use Microsoft tools
33388 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
33389 @code{dlgedit.exe} to build dialogs.
33390 It is always possible to build an @file{.rc} file yourself by writing a
33393 It is not our objective to explain how to write a resource file. A
33394 complete description of the resource script language can be found in the
33395 Microsoft documentation.
33397 @node Compiling Resources
33398 @subsection Compiling Resources
33401 @cindex Resources, compiling
33404 This section describes how to build a GNAT-compatible (COFF) object file
33405 containing the resources. This is done using the Resource Compiler
33406 @code{windres} as follows:
33409 $ windres -i myres.rc -o myres.o
33413 By default @code{windres} will run @command{gcc} to preprocess the @file{.rc}
33414 file. You can specify an alternate preprocessor (usually named
33415 @file{cpp.exe}) using the @code{windres} @option{--preprocessor}
33416 parameter. A list of all possible options may be obtained by entering
33417 the command @code{windres} @option{--help}.
33419 It is also possible to use the Microsoft resource compiler @code{rc.exe}
33420 to produce a @file{.res} file (binary resource file). See the
33421 corresponding Microsoft documentation for further details. In this case
33422 you need to use @code{windres} to translate the @file{.res} file to a
33423 GNAT-compatible object file as follows:
33426 $ windres -i myres.res -o myres.o
33429 @node Using Resources
33430 @subsection Using Resources
33431 @cindex Resources, using
33434 To include the resource file in your program just add the
33435 GNAT-compatible object file for the resource(s) to the linker
33436 arguments. With @command{gnatmake} this is done by using the @option{-largs}
33440 $ gnatmake myprog -largs myres.o
33443 @node Debugging a DLL
33444 @section Debugging a DLL
33445 @cindex DLL debugging
33448 * Program and DLL Both Built with GCC/GNAT::
33449 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
33453 Debugging a DLL is similar to debugging a standard program. But
33454 we have to deal with two different executable parts: the DLL and the
33455 program that uses it. We have the following four possibilities:
33459 The program and the DLL are built with @code{GCC/GNAT}.
33461 The program is built with foreign tools and the DLL is built with
33464 The program is built with @code{GCC/GNAT} and the DLL is built with
33470 In this section we address only cases one and two above.
33471 There is no point in trying to debug
33472 a DLL with @code{GNU/GDB}, if there is no GDB-compatible debugging
33473 information in it. To do so you must use a debugger compatible with the
33474 tools suite used to build the DLL.
33476 @node Program and DLL Both Built with GCC/GNAT
33477 @subsection Program and DLL Both Built with GCC/GNAT
33480 This is the simplest case. Both the DLL and the program have @code{GDB}
33481 compatible debugging information. It is then possible to break anywhere in
33482 the process. Let's suppose here that the main procedure is named
33483 @code{ada_main} and that in the DLL there is an entry point named
33487 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) and
33488 program must have been built with the debugging information (see GNAT -g
33489 switch). Here are the step-by-step instructions for debugging it:
33492 @item Launch @code{GDB} on the main program.
33498 @item Start the program and stop at the beginning of the main procedure
33505 This step is required to be able to set a breakpoint inside the DLL. As long
33506 as the program is not run, the DLL is not loaded. This has the
33507 consequence that the DLL debugging information is also not loaded, so it is not
33508 possible to set a breakpoint in the DLL.
33510 @item Set a breakpoint inside the DLL
33513 (gdb) break ada_dll
33520 At this stage a breakpoint is set inside the DLL. From there on
33521 you can use the standard approach to debug the whole program
33522 (@pxref{Running and Debugging Ada Programs}).
33525 @c This used to work, probably because the DLLs were non-relocatable
33526 @c keep this section around until the problem is sorted out.
33528 To break on the @code{DllMain} routine it is not possible to follow
33529 the procedure above. At the time the program stop on @code{ada_main}
33530 the @code{DllMain} routine as already been called. Either you can use
33531 the procedure below @pxref{Debugging the DLL Directly} or this procedure:
33534 @item Launch @code{GDB} on the main program.
33540 @item Load DLL symbols
33543 (gdb) add-sym api.dll
33546 @item Set a breakpoint inside the DLL
33549 (gdb) break ada_dll.adb:45
33552 Note that at this point it is not possible to break using the routine symbol
33553 directly as the program is not yet running. The solution is to break
33554 on the proper line (break in @file{ada_dll.adb} line 45).
33556 @item Start the program
33565 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT
33566 @subsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
33569 * Debugging the DLL Directly::
33570 * Attaching to a Running Process::
33574 In this case things are slightly more complex because it is not possible to
33575 start the main program and then break at the beginning to load the DLL and the
33576 associated DLL debugging information. It is not possible to break at the
33577 beginning of the program because there is no @code{GDB} debugging information,
33578 and therefore there is no direct way of getting initial control. This
33579 section addresses this issue by describing some methods that can be used
33580 to break somewhere in the DLL to debug it.
33583 First suppose that the main procedure is named @code{main} (this is for
33584 example some C code built with Microsoft Visual C) and that there is a
33585 DLL named @code{test.dll} containing an Ada entry point named
33589 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) must have
33590 been built with debugging information (see GNAT -g option).
33592 @node Debugging the DLL Directly
33593 @subsubsection Debugging the DLL Directly
33597 Find out the executable starting address
33600 $ objdump --file-header main.exe
33603 The starting address is reported on the last line. For example:
33606 main.exe: file format pei-i386
33607 architecture: i386, flags 0x0000010a:
33608 EXEC_P, HAS_DEBUG, D_PAGED
33609 start address 0x00401010
33613 Launch the debugger on the executable.
33620 Set a breakpoint at the starting address, and launch the program.
33623 $ (gdb) break *0x00401010
33627 The program will stop at the given address.
33630 Set a breakpoint on a DLL subroutine.
33633 (gdb) break ada_dll.adb:45
33636 Or if you want to break using a symbol on the DLL, you need first to
33637 select the Ada language (language used by the DLL).
33640 (gdb) set language ada
33641 (gdb) break ada_dll
33645 Continue the program.
33652 This will run the program until it reaches the breakpoint that has been
33653 set. From that point you can use the standard way to debug a program
33654 as described in (@pxref{Running and Debugging Ada Programs}).
33659 It is also possible to debug the DLL by attaching to a running process.
33661 @node Attaching to a Running Process
33662 @subsubsection Attaching to a Running Process
33663 @cindex DLL debugging, attach to process
33666 With @code{GDB} it is always possible to debug a running process by
33667 attaching to it. It is possible to debug a DLL this way. The limitation
33668 of this approach is that the DLL must run long enough to perform the
33669 attach operation. It may be useful for instance to insert a time wasting
33670 loop in the code of the DLL to meet this criterion.
33674 @item Launch the main program @file{main.exe}.
33680 @item Use the Windows @i{Task Manager} to find the process ID. Let's say
33681 that the process PID for @file{main.exe} is 208.
33689 @item Attach to the running process to be debugged.
33695 @item Load the process debugging information.
33698 (gdb) symbol-file main.exe
33701 @item Break somewhere in the DLL.
33704 (gdb) break ada_dll
33707 @item Continue process execution.
33716 This last step will resume the process execution, and stop at
33717 the breakpoint we have set. From there you can use the standard
33718 approach to debug a program as described in
33719 (@pxref{Running and Debugging Ada Programs}).
33721 @node Setting Stack Size from gnatlink
33722 @section Setting Stack Size from @command{gnatlink}
33725 It is possible to specify the program stack size at link time. On modern
33726 versions of Windows, starting with XP, this is mostly useful to set the size of
33727 the main stack (environment task). The other task stacks are set with pragma
33728 Storage_Size or with the @command{gnatbind -d} command.
33730 Since older versions of Windows (2000, NT4, etc.) do not allow setting the
33731 reserve size of individual tasks, the link-time stack size applies to all
33732 tasks, and pragma Storage_Size has no effect.
33733 In particular, Stack Overflow checks are made against this
33734 link-time specified size.
33736 This setting can be done with
33737 @command{gnatlink} using either:
33741 @item using @option{-Xlinker} linker option
33744 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
33747 This sets the stack reserve size to 0x10000 bytes and the stack commit
33748 size to 0x1000 bytes.
33750 @item using @option{-Wl} linker option
33753 $ gnatlink hello -Wl,--stack=0x1000000
33756 This sets the stack reserve size to 0x1000000 bytes. Note that with
33757 @option{-Wl} option it is not possible to set the stack commit size
33758 because the coma is a separator for this option.
33762 @node Setting Heap Size from gnatlink
33763 @section Setting Heap Size from @command{gnatlink}
33766 Under Windows systems, it is possible to specify the program heap size from
33767 @command{gnatlink} using either:
33771 @item using @option{-Xlinker} linker option
33774 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
33777 This sets the heap reserve size to 0x10000 bytes and the heap commit
33778 size to 0x1000 bytes.
33780 @item using @option{-Wl} linker option
33783 $ gnatlink hello -Wl,--heap=0x1000000
33786 This sets the heap reserve size to 0x1000000 bytes. Note that with
33787 @option{-Wl} option it is not possible to set the heap commit size
33788 because the coma is a separator for this option.
33794 @c **********************************
33795 @c * GNU Free Documentation License *
33796 @c **********************************
33798 @c GNU Free Documentation License
33800 @node Index,,GNU Free Documentation License, Top
33806 @c Put table of contents at end, otherwise it precedes the "title page" in
33807 @c the .txt version
33808 @c Edit the pdf file to move the contents to the beginning, after the title