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 1.3 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
110 @c Status as of November 2009:
111 @c Unfortunately texi2pdf and texi2html treat the trailing "@c"
112 @c differently, and faulty output is produced by one or the other
113 @c depending on whether the "@c" is present or absent.
114 @c As a result, the @ovar macro is not used, and all invocations
115 @c of the @ovar macro have been expanded inline.
118 @settitle @value{EDITION} User's Guide @value{PLATFORM}
119 @dircategory GNU Ada tools
121 * @value{EDITION} User's Guide: (gnat_ugn). @value{PLATFORM}
124 @include gcc-common.texi
126 @setchapternewpage odd
131 @title @value{EDITION} User's Guide
135 @titlefont{@i{@value{PLATFORM}}}
141 @subtitle GNAT, The GNU Ada Compiler
146 @vskip 0pt plus 1filll
153 @node Top, About This Guide, (dir), (dir)
154 @top @value{EDITION} User's Guide
157 @value{EDITION} User's Guide @value{PLATFORM}
160 GNAT, The GNU Ada Compiler@*
161 GCC version @value{version-GCC}@*
168 * Getting Started with GNAT::
169 * The GNAT Compilation Model::
170 * Compiling Using gcc::
171 * Binding Using gnatbind::
172 * Linking Using gnatlink::
173 * The GNAT Make Program gnatmake::
174 * Improving Performance::
175 * Renaming Files Using gnatchop::
176 * Configuration Pragmas::
177 * Handling Arbitrary File Naming Conventions Using gnatname::
178 * GNAT Project Manager::
179 * Tools Supporting Project Files::
180 * The Cross-Referencing Tools gnatxref and gnatfind::
181 * The GNAT Pretty-Printer gnatpp::
182 * The GNAT Metric Tool gnatmetric::
183 * File Name Krunching Using gnatkr::
184 * Preprocessing Using gnatprep::
186 * The GNAT Run-Time Library Builder gnatlbr::
188 * The GNAT Library Browser gnatls::
189 * Cleaning Up Using gnatclean::
191 * GNAT and Libraries::
192 * Using the GNU make Utility::
194 * Memory Management Issues::
195 * Stack Related Facilities::
196 * Verifying Properties Using gnatcheck::
197 * Creating Sample Bodies Using gnatstub::
198 * Generating Ada Bindings for C and C++ headers::
199 * Other Utility Programs::
200 * Running and Debugging Ada Programs::
202 * Code Coverage and Profiling::
205 * Compatibility with HP Ada::
207 * Platform-Specific Information for the Run-Time Libraries::
208 * Example of Binder Output File::
209 * Elaboration Order Handling in GNAT::
210 * Conditional Compilation::
212 * Compatibility and Porting Guide::
214 * Microsoft Windows Topics::
216 * GNU Free Documentation License::
219 --- The Detailed Node Listing ---
223 * What This Guide Contains::
224 * What You Should Know before Reading This Guide::
225 * Related Information::
228 Getting Started with GNAT
231 * Running a Simple Ada Program::
232 * Running a Program with Multiple Units::
233 * Using the gnatmake Utility::
235 * Editing with Emacs::
238 * Introduction to GPS::
241 The GNAT Compilation Model
243 * Source Representation::
244 * Foreign Language Representation::
245 * File Naming Rules::
246 * Using Other File Names::
247 * Alternative File Naming Schemes::
248 * Generating Object Files::
249 * Source Dependencies::
250 * The Ada Library Information Files::
251 * Binding an Ada Program::
252 * Mixed Language Programming::
254 * Building Mixed Ada & C++ Programs::
255 * Comparison between GNAT and C/C++ Compilation Models::
257 * Comparison between GNAT and Conventional Ada Library Models::
259 * Placement of temporary files::
262 Foreign Language Representation
265 * Other 8-Bit Codes::
266 * Wide Character Encodings::
268 Compiling Ada Programs With gcc
270 * Compiling Programs::
272 * Search Paths and the Run-Time Library (RTL)::
273 * Order of Compilation Issues::
278 * Output and Error Message Control::
279 * Warning Message Control::
280 * Debugging and Assertion Control::
281 * Validity Checking::
284 * Using gcc for Syntax Checking::
285 * Using gcc for Semantic Checking::
286 * Compiling Different Versions of Ada::
287 * Character Set Control::
288 * File Naming Control::
289 * Subprogram Inlining Control::
290 * Auxiliary Output Control::
291 * Debugging Control::
292 * Exception Handling Control::
293 * Units to Sources Mapping Files::
294 * Integrated Preprocessing::
299 Binding Ada Programs With gnatbind
302 * Switches for gnatbind::
303 * Command-Line Access::
304 * Search Paths for gnatbind::
305 * Examples of gnatbind Usage::
307 Switches for gnatbind
309 * Consistency-Checking Modes::
310 * Binder Error Message Control::
311 * Elaboration Control::
313 * Binding with Non-Ada Main Programs::
314 * Binding Programs with No Main Subprogram::
316 Linking Using gnatlink
319 * Switches for gnatlink::
321 The GNAT Make Program gnatmake
324 * Switches for gnatmake::
325 * Mode Switches for gnatmake::
326 * Notes on the Command Line::
327 * How gnatmake Works::
328 * Examples of gnatmake Usage::
330 Improving Performance
331 * Performance Considerations::
332 * Text_IO Suggestions::
333 * Reducing Size of Ada Executables with gnatelim::
334 * Reducing Size of Executables with unused subprogram/data elimination::
336 Performance Considerations
337 * Controlling Run-Time Checks::
338 * Use of Restrictions::
339 * Optimization Levels::
340 * Debugging Optimized Code::
341 * Inlining of Subprograms::
342 * Other Optimization Switches::
343 * Optimization and Strict Aliasing::
345 * Coverage Analysis::
348 Reducing Size of Ada Executables with gnatelim
351 * Processing Precompiled Libraries::
352 * Correcting the List of Eliminate Pragmas::
353 * Making Your Executables Smaller::
354 * Summary of the gnatelim Usage Cycle::
356 Reducing Size of Executables with unused subprogram/data elimination
357 * About unused subprogram/data elimination::
358 * Compilation options::
360 Renaming Files Using gnatchop
362 * Handling Files with Multiple Units::
363 * Operating gnatchop in Compilation Mode::
364 * Command Line for gnatchop::
365 * Switches for gnatchop::
366 * Examples of gnatchop Usage::
368 Configuration Pragmas
370 * Handling of Configuration Pragmas::
371 * The Configuration Pragmas Files::
373 Handling Arbitrary File Naming Conventions Using gnatname
375 * Arbitrary File Naming Conventions::
377 * Switches for gnatname::
378 * Examples of gnatname Usage::
380 The Cross-Referencing Tools gnatxref and gnatfind
382 * Switches for gnatxref::
383 * Switches for gnatfind::
384 * Project Files for gnatxref and gnatfind::
385 * Regular Expressions in gnatfind and gnatxref::
386 * Examples of gnatxref Usage::
387 * Examples of gnatfind Usage::
389 The GNAT Pretty-Printer gnatpp
391 * Switches for gnatpp::
394 The GNAT Metrics Tool gnatmetric
396 * Switches for gnatmetric::
398 File Name Krunching Using gnatkr
403 * Examples of gnatkr Usage::
405 Preprocessing Using gnatprep
406 * Preprocessing Symbols::
408 * Switches for gnatprep::
409 * Form of Definitions File::
410 * Form of Input Text for gnatprep::
413 The GNAT Run-Time Library Builder gnatlbr
416 * Switches for gnatlbr::
417 * Examples of gnatlbr Usage::
420 The GNAT Library Browser gnatls
423 * Switches for gnatls::
424 * Examples of gnatls Usage::
426 Cleaning Up Using gnatclean
428 * Running gnatclean::
429 * Switches for gnatclean::
430 @c * Examples of gnatclean Usage::
436 * Introduction to Libraries in GNAT::
437 * General Ada Libraries::
438 * Stand-alone Ada Libraries::
439 * Rebuilding the GNAT Run-Time Library::
441 Using the GNU make Utility
443 * Using gnatmake in a Makefile::
444 * Automatically Creating a List of Directories::
445 * Generating the Command Line Switches::
446 * Overcoming Command Line Length Limits::
449 Memory Management Issues
451 * Some Useful Memory Pools::
452 * The GNAT Debug Pool Facility::
457 Stack Related Facilities
459 * Stack Overflow Checking::
460 * Static Stack Usage Analysis::
461 * Dynamic Stack Usage Analysis::
463 Some Useful Memory Pools
465 The GNAT Debug Pool Facility
471 * Switches for gnatmem::
472 * Example of gnatmem Usage::
475 Verifying Properties Using gnatcheck
477 * Format of the Report File::
478 * General gnatcheck Switches::
479 * gnatcheck Rule Options::
480 * Adding the Results of Compiler Checks to gnatcheck Output::
481 * Project-Wide Checks::
484 * Example of gnatcheck Usage::
486 Sample Bodies Using gnatstub
489 * Switches for gnatstub::
491 Other Utility Programs
493 * Using Other Utility Programs with GNAT::
494 * The External Symbol Naming Scheme of GNAT::
495 * Converting Ada Files to html with gnathtml::
498 Code Coverage and Profiling
500 * Code Coverage of Ada Programs using gcov::
501 * Profiling an Ada Program using gprof::
504 Running and Debugging Ada Programs
506 * The GNAT Debugger GDB::
508 * Introduction to GDB Commands::
509 * Using Ada Expressions::
510 * Calling User-Defined Subprograms::
511 * Using the Next Command in a Function::
514 * Debugging Generic Units::
515 * Remote Debugging using gdbserver::
516 * GNAT Abnormal Termination or Failure to Terminate::
517 * Naming Conventions for GNAT Source Files::
518 * Getting Internal Debugging Information::
526 Compatibility with HP Ada
528 * Ada Language Compatibility::
529 * Differences in the Definition of Package System::
530 * Language-Related Features::
531 * The Package STANDARD::
532 * The Package SYSTEM::
533 * Tasking and Task-Related Features::
534 * Pragmas and Pragma-Related Features::
535 * Library of Predefined Units::
537 * Main Program Definition::
538 * Implementation-Defined Attributes::
539 * Compiler and Run-Time Interfacing::
540 * Program Compilation and Library Management::
542 * Implementation Limits::
543 * Tools and Utilities::
545 Language-Related Features
547 * Integer Types and Representations::
548 * Floating-Point Types and Representations::
549 * Pragmas Float_Representation and Long_Float::
550 * Fixed-Point Types and Representations::
551 * Record and Array Component Alignment::
553 * Other Representation Clauses::
555 Tasking and Task-Related Features
557 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
558 * Assigning Task IDs::
559 * Task IDs and Delays::
560 * Task-Related Pragmas::
561 * Scheduling and Task Priority::
563 * External Interrupts::
565 Pragmas and Pragma-Related Features
567 * Restrictions on the Pragma INLINE::
568 * Restrictions on the Pragma INTERFACE::
569 * Restrictions on the Pragma SYSTEM_NAME::
571 Library of Predefined Units
573 * Changes to DECLIB::
577 * Shared Libraries and Options Files::
581 Platform-Specific Information for the Run-Time Libraries
583 * Summary of Run-Time Configurations::
584 * Specifying a Run-Time Library::
585 * Choosing the Scheduling Policy::
586 * Solaris-Specific Considerations::
587 * Linux-Specific Considerations::
588 * AIX-Specific Considerations::
589 * Irix-Specific Considerations::
590 * RTX-Specific Considerations::
591 * HP-UX-Specific Considerations::
593 Example of Binder Output File
595 Elaboration Order Handling in GNAT
598 * Checking the Elaboration Order::
599 * Controlling the Elaboration Order::
600 * Controlling Elaboration in GNAT - Internal Calls::
601 * Controlling Elaboration in GNAT - External Calls::
602 * Default Behavior in GNAT - Ensuring Safety::
603 * Treatment of Pragma Elaborate::
604 * Elaboration Issues for Library Tasks::
605 * Mixing Elaboration Models::
606 * What to Do If the Default Elaboration Behavior Fails::
607 * Elaboration for Access-to-Subprogram Values::
608 * Summary of Procedures for Elaboration Control::
609 * Other Elaboration Order Considerations::
611 Conditional Compilation
612 * Use of Boolean Constants::
613 * Debugging - A Special Case::
614 * Conditionalizing Declarations::
615 * Use of Alternative Implementations::
620 * Basic Assembler Syntax::
621 * A Simple Example of Inline Assembler::
622 * Output Variables in Inline Assembler::
623 * Input Variables in Inline Assembler::
624 * Inlining Inline Assembler Code::
625 * Other Asm Functionality::
627 Compatibility and Porting Guide
629 * Compatibility with Ada 83::
630 * Compatibility between Ada 95 and Ada 2005::
631 * Implementation-dependent characteristics::
633 @c This brief section is only in the non-VMS version
634 @c The complete chapter on HP Ada issues is in the VMS version
635 * Compatibility with HP Ada 83::
637 * Compatibility with Other Ada Systems::
638 * Representation Clauses::
640 * Transitioning to 64-Bit GNAT for OpenVMS::
644 Microsoft Windows Topics
646 * Using GNAT on Windows::
647 * CONSOLE and WINDOWS subsystems::
649 * Mixed-Language Programming on Windows::
650 * Windows Calling Conventions::
651 * Introduction to Dynamic Link Libraries (DLLs)::
652 * Using DLLs with GNAT::
653 * Building DLLs with GNAT::
654 * GNAT and Windows Resources::
656 * Setting Stack Size from gnatlink::
657 * Setting Heap Size from gnatlink::
664 @node About This Guide
665 @unnumbered About This Guide
669 This guide describes the use of @value{EDITION},
670 a compiler and software development toolset for the full Ada
671 programming language, implemented on OpenVMS for HP's Alpha and
672 Integrity server (I64) platforms.
675 This guide describes the use of @value{EDITION},
676 a compiler and software development
677 toolset for the full Ada programming language.
679 It documents the features of the compiler and tools, and explains
680 how to use them to build Ada applications.
682 @value{EDITION} implements Ada 95 and Ada 2005, and it may also be invoked in
683 Ada 83 compatibility mode.
684 By default, @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
685 but you can override with a compiler switch
686 (@pxref{Compiling Different Versions of Ada})
687 to explicitly specify the language version.
688 Throughout this manual, references to ``Ada'' without a year suffix
689 apply to both the Ada 95 and Ada 2005 versions of the language.
693 For ease of exposition, ``@value{EDITION}'' will be referred to simply as
694 ``GNAT'' in the remainder of this document.
701 * What This Guide Contains::
702 * What You Should Know before Reading This Guide::
703 * Related Information::
707 @node What This Guide Contains
708 @unnumberedsec What This Guide Contains
711 This guide contains the following chapters:
715 @ref{Getting Started with GNAT}, describes how to get started compiling
716 and running Ada programs with the GNAT Ada programming environment.
718 @ref{The GNAT Compilation Model}, describes the compilation model used
722 @ref{Compiling Using gcc}, describes how to compile
723 Ada programs with @command{gcc}, the Ada compiler.
726 @ref{Binding Using gnatbind}, describes how to
727 perform binding of Ada programs with @code{gnatbind}, the GNAT binding
731 @ref{Linking Using gnatlink},
732 describes @command{gnatlink}, a
733 program that provides for linking using the GNAT run-time library to
734 construct a program. @command{gnatlink} can also incorporate foreign language
735 object units into the executable.
738 @ref{The GNAT Make Program gnatmake}, describes @command{gnatmake}, a
739 utility that automatically determines the set of sources
740 needed by an Ada compilation unit, and executes the necessary compilations
744 @ref{Improving Performance}, shows various techniques for making your
745 Ada program run faster or take less space.
746 It discusses the effect of the compiler's optimization switch and
747 also describes the @command{gnatelim} tool and unused subprogram/data
751 @ref{Renaming Files Using gnatchop}, describes
752 @code{gnatchop}, a utility that allows you to preprocess a file that
753 contains Ada source code, and split it into one or more new files, one
754 for each compilation unit.
757 @ref{Configuration Pragmas}, describes the configuration pragmas
761 @ref{Handling Arbitrary File Naming Conventions Using gnatname},
762 shows how to override the default GNAT file naming conventions,
763 either for an individual unit or globally.
766 @ref{GNAT Project Manager}, describes how to use project files
767 to organize large projects.
770 @ref{The Cross-Referencing Tools gnatxref and gnatfind}, discusses
771 @code{gnatxref} and @code{gnatfind}, two tools that provide an easy
772 way to navigate through sources.
775 @ref{The GNAT Pretty-Printer gnatpp}, shows how to produce a reformatted
776 version of an Ada source file with control over casing, indentation,
777 comment placement, and other elements of program presentation style.
780 @ref{The GNAT Metric Tool gnatmetric}, shows how to compute various
781 metrics for an Ada source file, such as the number of types and subprograms,
782 and assorted complexity measures.
785 @ref{File Name Krunching Using gnatkr}, describes the @code{gnatkr}
786 file name krunching utility, used to handle shortened
787 file names on operating systems with a limit on the length of names.
790 @ref{Preprocessing Using gnatprep}, describes @code{gnatprep}, a
791 preprocessor utility that allows a single source file to be used to
792 generate multiple or parameterized source files by means of macro
797 @ref{The GNAT Run-Time Library Builder gnatlbr}, describes @command{gnatlbr},
798 a tool for rebuilding the GNAT run time with user-supplied
799 configuration pragmas.
803 @ref{The GNAT Library Browser gnatls}, describes @code{gnatls}, a
804 utility that displays information about compiled units, including dependences
805 on the corresponding sources files, and consistency of compilations.
808 @ref{Cleaning Up Using gnatclean}, describes @code{gnatclean}, a utility
809 to delete files that are produced by the compiler, binder and linker.
813 @ref{GNAT and Libraries}, describes the process of creating and using
814 Libraries with GNAT. It also describes how to recompile the GNAT run-time
818 @ref{Using the GNU make Utility}, describes some techniques for using
819 the GNAT toolset in Makefiles.
823 @ref{Memory Management Issues}, describes some useful predefined storage pools
824 and in particular the GNAT Debug Pool facility, which helps detect incorrect
827 It also describes @command{gnatmem}, a utility that monitors dynamic
828 allocation and deallocation and helps detect ``memory leaks''.
832 @ref{Stack Related Facilities}, describes some useful tools associated with
833 stack checking and analysis.
836 @ref{Verifying Properties Using gnatcheck}, discusses @code{gnatcheck},
837 a utility that checks Ada code against a set of rules.
840 @ref{Creating Sample Bodies Using gnatstub}, discusses @code{gnatstub},
841 a utility that generates empty but compilable bodies for library units.
844 @ref{Generating Ada Bindings for C and C++ headers}, describes how to
845 generate automatically Ada bindings from C and C++ headers.
848 @ref{Other Utility Programs}, discusses several other GNAT utilities,
849 including @code{gnathtml}.
853 @ref{Code Coverage and Profiling}, describes how to perform a structural
854 coverage and profile the execution of Ada programs.
858 @ref{Running and Debugging Ada Programs}, describes how to run and debug
863 @ref{Compatibility with HP Ada}, details the compatibility of GNAT with
864 HP Ada 83 @footnote{``HP Ada'' refers to the legacy product originally
865 developed by Digital Equipment Corporation and currently supported by HP.}
866 for OpenVMS Alpha. This product was formerly known as DEC Ada,
869 historical compatibility reasons, the relevant libraries still use the
874 @ref{Platform-Specific Information for the Run-Time Libraries},
875 describes the various run-time
876 libraries supported by GNAT on various platforms and explains how to
877 choose a particular library.
880 @ref{Example of Binder Output File}, shows the source code for the binder
881 output file for a sample program.
884 @ref{Elaboration Order Handling in GNAT}, describes how GNAT helps
885 you deal with elaboration order issues.
888 @ref{Conditional Compilation}, describes how to model conditional compilation,
889 both with Ada in general and with GNAT facilities in particular.
892 @ref{Inline Assembler}, shows how to use the inline assembly facility
896 @ref{Compatibility and Porting Guide}, contains sections on compatibility
897 of GNAT with other Ada development environments (including Ada 83 systems),
898 to assist in porting code from those environments.
902 @ref{Microsoft Windows Topics}, presents information relevant to the
903 Microsoft Windows platform.
907 @c *************************************************
908 @node What You Should Know before Reading This Guide
909 @c *************************************************
910 @unnumberedsec What You Should Know before Reading This Guide
912 @cindex Ada 95 Language Reference Manual
913 @cindex Ada 2005 Language Reference Manual
915 This guide assumes a basic familiarity with the Ada 95 language, as
916 described in the International Standard ANSI/ISO/IEC-8652:1995, January
918 It does not require knowledge of the new features introduced by Ada 2005,
919 (officially known as ISO/IEC 8652:1995 with Technical Corrigendum 1
921 Both reference manuals are included in the GNAT documentation
924 @node Related Information
925 @unnumberedsec Related Information
928 For further information about related tools, refer to the following
933 @xref{Top, GNAT Reference Manual, About This Guide, gnat_rm, GNAT
934 Reference Manual}, which contains all reference material for the GNAT
935 implementation of Ada.
939 @cite{Using the GNAT Programming Studio}, which describes the GPS
940 Integrated Development Environment.
943 @cite{GNAT Programming Studio Tutorial}, which introduces the
944 main GPS features through examples.
948 @cite{Ada 95 Reference Manual}, which contains reference
949 material for the Ada 95 programming language.
952 @cite{Ada 2005 Reference Manual}, which contains reference
953 material for the Ada 2005 programming language.
956 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
958 in the GNU:[DOCS] directory,
960 for all details on the use of the GNU source-level debugger.
963 @xref{Top,, The extensible self-documenting text editor, emacs,
966 located in the GNU:[DOCS] directory if the EMACS kit is installed,
968 for full information on the extensible editor and programming
975 @unnumberedsec Conventions
977 @cindex Typographical conventions
980 Following are examples of the typographical and graphic conventions used
985 @code{Functions}, @command{utility program names}, @code{standard names},
989 @option{Option flags}
992 @file{File names}, @samp{button names}, and @samp{field names}.
995 @code{Variables}, @env{environment variables}, and @var{metasyntactic
1002 @r{[}optional information or parameters@r{]}
1005 Examples are described by text
1007 and then shown this way.
1012 Commands that are entered by the user are preceded in this manual by the
1013 characters @w{``@code{$ }''} (dollar sign followed by space). If your system
1014 uses this sequence as a prompt, then the commands will appear exactly as
1015 you see them in the manual. If your system uses some other prompt, then
1016 the command will appear with the @code{$} replaced by whatever prompt
1017 character you are using.
1020 Full file names are shown with the ``@code{/}'' character
1021 as the directory separator; e.g., @file{parent-dir/subdir/myfile.adb}.
1022 If you are using GNAT on a Windows platform, please note that
1023 the ``@code{\}'' character should be used instead.
1026 @c ****************************
1027 @node Getting Started with GNAT
1028 @chapter Getting Started with GNAT
1031 This chapter describes some simple ways of using GNAT to build
1032 executable Ada programs.
1034 @ref{Running GNAT}, through @ref{Using the gnatmake Utility},
1035 show how to use the command line environment.
1036 @ref{Introduction to GPS}, provides a brief
1037 introduction to the GNAT Programming Studio, a visually-oriented
1038 Integrated Development Environment for GNAT.
1039 GPS offers a graphical ``look and feel'', support for development in
1040 other programming languages, comprehensive browsing features, and
1041 many other capabilities.
1042 For information on GPS please refer to
1043 @cite{Using the GNAT Programming Studio}.
1048 * Running a Simple Ada Program::
1049 * Running a Program with Multiple Units::
1050 * Using the gnatmake Utility::
1052 * Editing with Emacs::
1055 * Introduction to GPS::
1060 @section Running GNAT
1063 Three steps are needed to create an executable file from an Ada source
1068 The source file(s) must be compiled.
1070 The file(s) must be bound using the GNAT binder.
1072 All appropriate object files must be linked to produce an executable.
1076 All three steps are most commonly handled by using the @command{gnatmake}
1077 utility program that, given the name of the main program, automatically
1078 performs the necessary compilation, binding and linking steps.
1080 @node Running a Simple Ada Program
1081 @section Running a Simple Ada Program
1084 Any text editor may be used to prepare an Ada program.
1086 used, the optional Ada mode may be helpful in laying out the program.)
1088 program text is a normal text file. We will assume in our initial
1089 example that you have used your editor to prepare the following
1090 standard format text file:
1092 @smallexample @c ada
1094 with Ada.Text_IO; use Ada.Text_IO;
1097 Put_Line ("Hello WORLD!");
1103 This file should be named @file{hello.adb}.
1104 With the normal default file naming conventions, GNAT requires
1106 contain a single compilation unit whose file name is the
1108 with periods replaced by hyphens; the
1109 extension is @file{ads} for a
1110 spec and @file{adb} for a body.
1111 You can override this default file naming convention by use of the
1112 special pragma @code{Source_File_Name} (@pxref{Using Other File Names}).
1113 Alternatively, if you want to rename your files according to this default
1114 convention, which is probably more convenient if you will be using GNAT
1115 for all your compilations, then the @code{gnatchop} utility
1116 can be used to generate correctly-named source files
1117 (@pxref{Renaming Files Using gnatchop}).
1119 You can compile the program using the following command (@code{$} is used
1120 as the command prompt in the examples in this document):
1127 @command{gcc} is the command used to run the compiler. This compiler is
1128 capable of compiling programs in several languages, including Ada and
1129 C. It assumes that you have given it an Ada program if the file extension is
1130 either @file{.ads} or @file{.adb}, and it will then call
1131 the GNAT compiler to compile the specified file.
1134 The @option{-c} switch is required. It tells @command{gcc} to only do a
1135 compilation. (For C programs, @command{gcc} can also do linking, but this
1136 capability is not used directly for Ada programs, so the @option{-c}
1137 switch must always be present.)
1140 This compile command generates a file
1141 @file{hello.o}, which is the object
1142 file corresponding to your Ada program. It also generates
1143 an ``Ada Library Information'' file @file{hello.ali},
1144 which contains additional information used to check
1145 that an Ada program is consistent.
1146 To build an executable file,
1147 use @code{gnatbind} to bind the program
1148 and @command{gnatlink} to link it. The
1149 argument to both @code{gnatbind} and @command{gnatlink} is the name of the
1150 @file{ALI} file, but the default extension of @file{.ali} can
1151 be omitted. This means that in the most common case, the argument
1152 is simply the name of the main program:
1160 A simpler method of carrying out these steps is to use
1162 a master program that invokes all the required
1163 compilation, binding and linking tools in the correct order. In particular,
1164 @command{gnatmake} automatically recompiles any sources that have been
1165 modified since they were last compiled, or sources that depend
1166 on such modified sources, so that ``version skew'' is avoided.
1167 @cindex Version skew (avoided by @command{gnatmake})
1170 $ gnatmake hello.adb
1174 The result is an executable program called @file{hello}, which can be
1182 assuming that the current directory is on the search path
1183 for executable programs.
1186 and, if all has gone well, you will see
1193 appear in response to this command.
1195 @c ****************************************
1196 @node Running a Program with Multiple Units
1197 @section Running a Program with Multiple Units
1200 Consider a slightly more complicated example that has three files: a
1201 main program, and the spec and body of a package:
1203 @smallexample @c ada
1206 package Greetings is
1211 with Ada.Text_IO; use Ada.Text_IO;
1212 package body Greetings is
1215 Put_Line ("Hello WORLD!");
1218 procedure Goodbye is
1220 Put_Line ("Goodbye WORLD!");
1237 Following the one-unit-per-file rule, place this program in the
1238 following three separate files:
1242 spec of package @code{Greetings}
1245 body of package @code{Greetings}
1248 body of main program
1252 To build an executable version of
1253 this program, we could use four separate steps to compile, bind, and link
1254 the program, as follows:
1258 $ gcc -c greetings.adb
1264 Note that there is no required order of compilation when using GNAT.
1265 In particular it is perfectly fine to compile the main program first.
1266 Also, it is not necessary to compile package specs in the case where
1267 there is an accompanying body; you only need to compile the body. If you want
1268 to submit these files to the compiler for semantic checking and not code
1269 generation, then use the
1270 @option{-gnatc} switch:
1273 $ gcc -c greetings.ads -gnatc
1277 Although the compilation can be done in separate steps as in the
1278 above example, in practice it is almost always more convenient
1279 to use the @command{gnatmake} tool. All you need to know in this case
1280 is the name of the main program's source file. The effect of the above four
1281 commands can be achieved with a single one:
1284 $ gnatmake gmain.adb
1288 In the next section we discuss the advantages of using @command{gnatmake} in
1291 @c *****************************
1292 @node Using the gnatmake Utility
1293 @section Using the @command{gnatmake} Utility
1296 If you work on a program by compiling single components at a time using
1297 @command{gcc}, you typically keep track of the units you modify. In order to
1298 build a consistent system, you compile not only these units, but also any
1299 units that depend on the units you have modified.
1300 For example, in the preceding case,
1301 if you edit @file{gmain.adb}, you only need to recompile that file. But if
1302 you edit @file{greetings.ads}, you must recompile both
1303 @file{greetings.adb} and @file{gmain.adb}, because both files contain
1304 units that depend on @file{greetings.ads}.
1306 @code{gnatbind} will warn you if you forget one of these compilation
1307 steps, so that it is impossible to generate an inconsistent program as a
1308 result of forgetting to do a compilation. Nevertheless it is tedious and
1309 error-prone to keep track of dependencies among units.
1310 One approach to handle the dependency-bookkeeping is to use a
1311 makefile. However, makefiles present maintenance problems of their own:
1312 if the dependencies change as you change the program, you must make
1313 sure that the makefile is kept up-to-date manually, which is also an
1314 error-prone process.
1316 The @command{gnatmake} utility takes care of these details automatically.
1317 Invoke it using either one of the following forms:
1320 $ gnatmake gmain.adb
1321 $ gnatmake ^gmain^GMAIN^
1325 The argument is the name of the file containing the main program;
1326 you may omit the extension. @command{gnatmake}
1327 examines the environment, automatically recompiles any files that need
1328 recompiling, and binds and links the resulting set of object files,
1329 generating the executable file, @file{^gmain^GMAIN.EXE^}.
1330 In a large program, it
1331 can be extremely helpful to use @command{gnatmake}, because working out by hand
1332 what needs to be recompiled can be difficult.
1334 Note that @command{gnatmake}
1335 takes into account all the Ada rules that
1336 establish dependencies among units. These include dependencies that result
1337 from inlining subprogram bodies, and from
1338 generic instantiation. Unlike some other
1339 Ada make tools, @command{gnatmake} does not rely on the dependencies that were
1340 found by the compiler on a previous compilation, which may possibly
1341 be wrong when sources change. @command{gnatmake} determines the exact set of
1342 dependencies from scratch each time it is run.
1345 @node Editing with Emacs
1346 @section Editing with Emacs
1350 Emacs is an extensible self-documenting text editor that is available in a
1351 separate VMSINSTAL kit.
1353 Invoke Emacs by typing @kbd{Emacs} at the command prompt. To get started,
1354 click on the Emacs Help menu and run the Emacs Tutorial.
1355 In a character cell terminal, Emacs help is invoked with @kbd{Ctrl-h} (also
1356 written as @kbd{C-h}), and the tutorial by @kbd{C-h t}.
1358 Documentation on Emacs and other tools is available in Emacs under the
1359 pull-down menu button: @code{Help - Info}. After selecting @code{Info},
1360 use the middle mouse button to select a topic (e.g.@: Emacs).
1362 In a character cell terminal, do @kbd{C-h i} to invoke info, and then @kbd{m}
1363 (stands for menu) followed by the menu item desired, as in @kbd{m Emacs}, to
1364 get to the Emacs manual.
1365 Help on Emacs is also available by typing @kbd{HELP EMACS} at the DCL command
1368 The tutorial is highly recommended in order to learn the intricacies of Emacs,
1369 which is sufficiently extensible to provide for a complete programming
1370 environment and shell for the sophisticated user.
1374 @node Introduction to GPS
1375 @section Introduction to GPS
1376 @cindex GPS (GNAT Programming Studio)
1377 @cindex GNAT Programming Studio (GPS)
1379 Although the command line interface (@command{gnatmake}, etc.) alone
1380 is sufficient, a graphical Interactive Development
1381 Environment can make it easier for you to compose, navigate, and debug
1382 programs. This section describes the main features of GPS
1383 (``GNAT Programming Studio''), the GNAT graphical IDE.
1384 You will see how to use GPS to build and debug an executable, and
1385 you will also learn some of the basics of the GNAT ``project'' facility.
1387 GPS enables you to do much more than is presented here;
1388 e.g., you can produce a call graph, interface to a third-party
1389 Version Control System, and inspect the generated assembly language
1391 Indeed, GPS also supports languages other than Ada.
1392 Such additional information, and an explanation of all of the GPS menu
1393 items. may be found in the on-line help, which includes
1394 a user's guide and a tutorial (these are also accessible from the GNAT
1398 * Building a New Program with GPS::
1399 * Simple Debugging with GPS::
1402 @node Building a New Program with GPS
1403 @subsection Building a New Program with GPS
1405 GPS invokes the GNAT compilation tools using information
1406 contained in a @emph{project} (also known as a @emph{project file}):
1407 a collection of properties such
1408 as source directories, identities of main subprograms, tool switches, etc.,
1409 and their associated values.
1410 See @ref{GNAT Project Manager} for details.
1411 In order to run GPS, you will need to either create a new project
1412 or else open an existing one.
1414 This section will explain how you can use GPS to create a project,
1415 to associate Ada source files with a project, and to build and run
1419 @item @emph{Creating a project}
1421 Invoke GPS, either from the command line or the platform's IDE.
1422 After it starts, GPS will display a ``Welcome'' screen with three
1427 @code{Start with default project in directory}
1430 @code{Create new project with wizard}
1433 @code{Open existing project}
1437 Select @code{Create new project with wizard} and press @code{OK}.
1438 A new window will appear. In the text box labeled with
1439 @code{Enter the name of the project to create}, type @file{sample}
1440 as the project name.
1441 In the next box, browse to choose the directory in which you
1442 would like to create the project file.
1443 After selecting an appropriate directory, press @code{Forward}.
1445 A window will appear with the title
1446 @code{Version Control System Configuration}.
1447 Simply press @code{Forward}.
1449 A window will appear with the title
1450 @code{Please select the source directories for this project}.
1451 The directory that you specified for the project file will be selected
1452 by default as the one to use for sources; simply press @code{Forward}.
1454 A window will appear with the title
1455 @code{Please select the build directory for this project}.
1456 The directory that you specified for the project file will be selected
1457 by default for object files and executables;
1458 simply press @code{Forward}.
1460 A window will appear with the title
1461 @code{Please select the main units for this project}.
1462 You will supply this information later, after creating the source file.
1463 Simply press @code{Forward} for now.
1465 A window will appear with the title
1466 @code{Please select the switches to build the project}.
1467 Press @code{Apply}. This will create a project file named
1468 @file{sample.prj} in the directory that you had specified.
1470 @item @emph{Creating and saving the source file}
1472 After you create the new project, a GPS window will appear, which is
1473 partitioned into two main sections:
1477 A @emph{Workspace area}, initially greyed out, which you will use for
1478 creating and editing source files
1481 Directly below, a @emph{Messages area}, which initially displays a
1482 ``Welcome'' message.
1483 (If the Messages area is not visible, drag its border upward to expand it.)
1487 Select @code{File} on the menu bar, and then the @code{New} command.
1488 The Workspace area will become white, and you can now
1489 enter the source program explicitly.
1490 Type the following text
1492 @smallexample @c ada
1494 with Ada.Text_IO; use Ada.Text_IO;
1497 Put_Line("Hello from GPS!");
1503 Select @code{File}, then @code{Save As}, and enter the source file name
1505 The file will be saved in the same directory you specified as the
1506 location of the default project file.
1508 @item @emph{Updating the project file}
1510 You need to add the new source file to the project.
1512 the @code{Project} menu and then @code{Edit project properties}.
1513 Click the @code{Main files} tab on the left, and then the
1515 Choose @file{hello.adb} from the list, and press @code{Open}.
1516 The project settings window will reflect this action.
1519 @item @emph{Building and running the program}
1521 In the main GPS window, now choose the @code{Build} menu, then @code{Make},
1522 and select @file{hello.adb}.
1523 The Messages window will display the resulting invocations of @command{gcc},
1524 @command{gnatbind}, and @command{gnatlink}
1525 (reflecting the default switch settings from the
1526 project file that you created) and then a ``successful compilation/build''
1529 To run the program, choose the @code{Build} menu, then @code{Run}, and
1530 select @command{hello}.
1531 An @emph{Arguments Selection} window will appear.
1532 There are no command line arguments, so just click @code{OK}.
1534 The Messages window will now display the program's output (the string
1535 @code{Hello from GPS}), and at the bottom of the GPS window a status
1536 update is displayed (@code{Run: hello}).
1537 Close the GPS window (or select @code{File}, then @code{Exit}) to
1538 terminate this GPS session.
1541 @node Simple Debugging with GPS
1542 @subsection Simple Debugging with GPS
1544 This section illustrates basic debugging techniques (setting breakpoints,
1545 examining/modifying variables, single stepping).
1548 @item @emph{Opening a project}
1550 Start GPS and select @code{Open existing project}; browse to
1551 specify the project file @file{sample.prj} that you had created in the
1554 @item @emph{Creating a source file}
1556 Select @code{File}, then @code{New}, and type in the following program:
1558 @smallexample @c ada
1560 with Ada.Text_IO; use Ada.Text_IO;
1561 procedure Example is
1562 Line : String (1..80);
1565 Put_Line("Type a line of text at each prompt; an empty line to exit");
1569 Put_Line (Line (1..N) );
1577 Select @code{File}, then @code{Save as}, and enter the file name
1580 @item @emph{Updating the project file}
1582 Add @code{Example} as a new main unit for the project:
1585 Select @code{Project}, then @code{Edit Project Properties}.
1588 Select the @code{Main files} tab, click @code{Add}, then
1589 select the file @file{example.adb} from the list, and
1591 You will see the file name appear in the list of main units
1597 @item @emph{Building/running the executable}
1599 To build the executable
1600 select @code{Build}, then @code{Make}, and then choose @file{example.adb}.
1602 Run the program to see its effect (in the Messages area).
1603 Each line that you enter is displayed; an empty line will
1604 cause the loop to exit and the program to terminate.
1606 @item @emph{Debugging the program}
1608 Note that the @option{-g} switches to @command{gcc} and @command{gnatlink},
1609 which are required for debugging, are on by default when you create
1611 Thus unless you intentionally remove these settings, you will be able
1612 to debug any program that you develop using GPS.
1615 @item @emph{Initializing}
1617 Select @code{Debug}, then @code{Initialize}, then @file{example}
1619 @item @emph{Setting a breakpoint}
1621 After performing the initialization step, you will observe a small
1622 icon to the right of each line number.
1623 This serves as a toggle for breakpoints; clicking the icon will
1624 set a breakpoint at the corresponding line (the icon will change to
1625 a red circle with an ``x''), and clicking it again
1626 will remove the breakpoint / reset the icon.
1628 For purposes of this example, set a breakpoint at line 10 (the
1629 statement @code{Put_Line@ (Line@ (1..N));}
1631 @item @emph{Starting program execution}
1633 Select @code{Debug}, then @code{Run}. When the
1634 @code{Program Arguments} window appears, click @code{OK}.
1635 A console window will appear; enter some line of text,
1636 e.g.@: @code{abcde}, at the prompt.
1637 The program will pause execution when it gets to the
1638 breakpoint, and the corresponding line is highlighted.
1640 @item @emph{Examining a variable}
1642 Move the mouse over one of the occurrences of the variable @code{N}.
1643 You will see the value (5) displayed, in ``tool tip'' fashion.
1644 Right click on @code{N}, select @code{Debug}, then select @code{Display N}.
1645 You will see information about @code{N} appear in the @code{Debugger Data}
1646 pane, showing the value as 5.
1648 @item @emph{Assigning a new value to a variable}
1650 Right click on the @code{N} in the @code{Debugger Data} pane, and
1651 select @code{Set value of N}.
1652 When the input window appears, enter the value @code{4} and click
1654 This value does not automatically appear in the @code{Debugger Data}
1655 pane; to see it, right click again on the @code{N} in the
1656 @code{Debugger Data} pane and select @code{Update value}.
1657 The new value, 4, will appear in red.
1659 @item @emph{Single stepping}
1661 Select @code{Debug}, then @code{Next}.
1662 This will cause the next statement to be executed, in this case the
1663 call of @code{Put_Line} with the string slice.
1664 Notice in the console window that the displayed string is simply
1665 @code{abcd} and not @code{abcde} which you had entered.
1666 This is because the upper bound of the slice is now 4 rather than 5.
1668 @item @emph{Removing a breakpoint}
1670 Toggle the breakpoint icon at line 10.
1672 @item @emph{Resuming execution from a breakpoint}
1674 Select @code{Debug}, then @code{Continue}.
1675 The program will reach the next iteration of the loop, and
1676 wait for input after displaying the prompt.
1677 This time, just hit the @kbd{Enter} key.
1678 The value of @code{N} will be 0, and the program will terminate.
1679 The console window will disappear.
1684 @node The GNAT Compilation Model
1685 @chapter The GNAT Compilation Model
1686 @cindex GNAT compilation model
1687 @cindex Compilation model
1690 * Source Representation::
1691 * Foreign Language Representation::
1692 * File Naming Rules::
1693 * Using Other File Names::
1694 * Alternative File Naming Schemes::
1695 * Generating Object Files::
1696 * Source Dependencies::
1697 * The Ada Library Information Files::
1698 * Binding an Ada Program::
1699 * Mixed Language Programming::
1701 * Building Mixed Ada & C++ Programs::
1702 * Comparison between GNAT and C/C++ Compilation Models::
1704 * Comparison between GNAT and Conventional Ada Library Models::
1706 * Placement of temporary files::
1711 This chapter describes the compilation model used by GNAT. Although
1712 similar to that used by other languages, such as C and C++, this model
1713 is substantially different from the traditional Ada compilation models,
1714 which are based on a library. The model is initially described without
1715 reference to the library-based model. If you have not previously used an
1716 Ada compiler, you need only read the first part of this chapter. The
1717 last section describes and discusses the differences between the GNAT
1718 model and the traditional Ada compiler models. If you have used other
1719 Ada compilers, this section will help you to understand those
1720 differences, and the advantages of the GNAT model.
1722 @node Source Representation
1723 @section Source Representation
1727 Ada source programs are represented in standard text files, using
1728 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1729 7-bit ASCII set, plus additional characters used for
1730 representing foreign languages (@pxref{Foreign Language Representation}
1731 for support of non-USA character sets). The format effector characters
1732 are represented using their standard ASCII encodings, as follows:
1737 Vertical tab, @code{16#0B#}
1741 Horizontal tab, @code{16#09#}
1745 Carriage return, @code{16#0D#}
1749 Line feed, @code{16#0A#}
1753 Form feed, @code{16#0C#}
1757 Source files are in standard text file format. In addition, GNAT will
1758 recognize a wide variety of stream formats, in which the end of
1759 physical lines is marked by any of the following sequences:
1760 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1761 in accommodating files that are imported from other operating systems.
1763 @cindex End of source file
1764 @cindex Source file, end
1766 The end of a source file is normally represented by the physical end of
1767 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1768 recognized as signalling the end of the source file. Again, this is
1769 provided for compatibility with other operating systems where this
1770 code is used to represent the end of file.
1772 Each file contains a single Ada compilation unit, including any pragmas
1773 associated with the unit. For example, this means you must place a
1774 package declaration (a package @dfn{spec}) and the corresponding body in
1775 separate files. An Ada @dfn{compilation} (which is a sequence of
1776 compilation units) is represented using a sequence of files. Similarly,
1777 you will place each subunit or child unit in a separate file.
1779 @node Foreign Language Representation
1780 @section Foreign Language Representation
1783 GNAT supports the standard character sets defined in Ada as well as
1784 several other non-standard character sets for use in localized versions
1785 of the compiler (@pxref{Character Set Control}).
1788 * Other 8-Bit Codes::
1789 * Wide Character Encodings::
1797 The basic character set is Latin-1. This character set is defined by ISO
1798 standard 8859, part 1. The lower half (character codes @code{16#00#}
1799 @dots{} @code{16#7F#)} is identical to standard ASCII coding, but the upper half
1800 is used to represent additional characters. These include extended letters
1801 used by European languages, such as French accents, the vowels with umlauts
1802 used in German, and the extra letter A-ring used in Swedish.
1804 @findex Ada.Characters.Latin_1
1805 For a complete list of Latin-1 codes and their encodings, see the source
1806 file of library unit @code{Ada.Characters.Latin_1} in file
1807 @file{a-chlat1.ads}.
1808 You may use any of these extended characters freely in character or
1809 string literals. In addition, the extended characters that represent
1810 letters can be used in identifiers.
1812 @node Other 8-Bit Codes
1813 @subsection Other 8-Bit Codes
1816 GNAT also supports several other 8-bit coding schemes:
1819 @item ISO 8859-2 (Latin-2)
1822 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1825 @item ISO 8859-3 (Latin-3)
1828 Latin-3 letters allowed in identifiers, with uppercase and lowercase
1831 @item ISO 8859-4 (Latin-4)
1834 Latin-4 letters allowed in identifiers, with uppercase and lowercase
1837 @item ISO 8859-5 (Cyrillic)
1840 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
1841 lowercase equivalence.
1843 @item ISO 8859-15 (Latin-9)
1846 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
1847 lowercase equivalence
1849 @item IBM PC (code page 437)
1850 @cindex code page 437
1851 This code page is the normal default for PCs in the U.S. It corresponds
1852 to the original IBM PC character set. This set has some, but not all, of
1853 the extended Latin-1 letters, but these letters do not have the same
1854 encoding as Latin-1. In this mode, these letters are allowed in
1855 identifiers with uppercase and lowercase equivalence.
1857 @item IBM PC (code page 850)
1858 @cindex code page 850
1859 This code page is a modification of 437 extended to include all the
1860 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
1861 mode, all these letters are allowed in identifiers with uppercase and
1862 lowercase equivalence.
1864 @item Full Upper 8-bit
1865 Any character in the range 80-FF allowed in identifiers, and all are
1866 considered distinct. In other words, there are no uppercase and lowercase
1867 equivalences in this range. This is useful in conjunction with
1868 certain encoding schemes used for some foreign character sets (e.g.,
1869 the typical method of representing Chinese characters on the PC).
1872 No upper-half characters in the range 80-FF are allowed in identifiers.
1873 This gives Ada 83 compatibility for identifier names.
1877 For precise data on the encodings permitted, and the uppercase and lowercase
1878 equivalences that are recognized, see the file @file{csets.adb} in
1879 the GNAT compiler sources. You will need to obtain a full source release
1880 of GNAT to obtain this file.
1882 @node Wide Character Encodings
1883 @subsection Wide Character Encodings
1886 GNAT allows wide character codes to appear in character and string
1887 literals, and also optionally in identifiers, by means of the following
1888 possible encoding schemes:
1893 In this encoding, a wide character is represented by the following five
1901 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1902 characters (using uppercase letters) of the wide character code. For
1903 example, ESC A345 is used to represent the wide character with code
1905 This scheme is compatible with use of the full Wide_Character set.
1907 @item Upper-Half Coding
1908 @cindex Upper-Half Coding
1909 The wide character with encoding @code{16#abcd#} where the upper bit is on
1910 (in other words, ``a'' is in the range 8-F) is represented as two bytes,
1911 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
1912 character, but is not required to be in the upper half. This method can
1913 be also used for shift-JIS or EUC, where the internal coding matches the
1916 @item Shift JIS Coding
1917 @cindex Shift JIS Coding
1918 A wide character is represented by a two-character sequence,
1920 @code{16#cd#}, with the restrictions described for upper-half encoding as
1921 described above. The internal character code is the corresponding JIS
1922 character according to the standard algorithm for Shift-JIS
1923 conversion. Only characters defined in the JIS code set table can be
1924 used with this encoding method.
1928 A wide character is represented by a two-character sequence
1930 @code{16#cd#}, with both characters being in the upper half. The internal
1931 character code is the corresponding JIS character according to the EUC
1932 encoding algorithm. Only characters defined in the JIS code set table
1933 can be used with this encoding method.
1936 A wide character is represented using
1937 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1938 10646-1/Am.2. Depending on the character value, the representation
1939 is a one, two, or three byte sequence:
1944 16#0000#-16#007f#: 2#0@var{xxxxxxx}#
1945 16#0080#-16#07ff#: 2#110@var{xxxxx}# 2#10@var{xxxxxx}#
1946 16#0800#-16#ffff#: 2#1110@var{xxxx}# 2#10@var{xxxxxx}# 2#10@var{xxxxxx}#
1951 where the @var{xxx} bits correspond to the left-padded bits of the
1952 16-bit character value. Note that all lower half ASCII characters
1953 are represented as ASCII bytes and all upper half characters and
1954 other wide characters are represented as sequences of upper-half
1955 (The full UTF-8 scheme allows for encoding 31-bit characters as
1956 6-byte sequences, but in this implementation, all UTF-8 sequences
1957 of four or more bytes length will be treated as illegal).
1958 @item Brackets Coding
1959 In this encoding, a wide character is represented by the following eight
1967 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1968 characters (using uppercase letters) of the wide character code. For
1969 example, [``A345''] is used to represent the wide character with code
1970 @code{16#A345#}. It is also possible (though not required) to use the
1971 Brackets coding for upper half characters. For example, the code
1972 @code{16#A3#} can be represented as @code{[``A3'']}.
1974 This scheme is compatible with use of the full Wide_Character set,
1975 and is also the method used for wide character encoding in the standard
1976 ACVC (Ada Compiler Validation Capability) test suite distributions.
1981 Note: Some of these coding schemes do not permit the full use of the
1982 Ada character set. For example, neither Shift JIS, nor EUC allow the
1983 use of the upper half of the Latin-1 set.
1985 @node File Naming Rules
1986 @section File Naming Rules
1989 The default file name is determined by the name of the unit that the
1990 file contains. The name is formed by taking the full expanded name of
1991 the unit and replacing the separating dots with hyphens and using
1992 ^lowercase^uppercase^ for all letters.
1994 An exception arises if the file name generated by the above rules starts
1995 with one of the characters
1997 @samp{A}, @samp{G}, @samp{I}, or @samp{S},
2000 @samp{a}, @samp{g}, @samp{i}, or @samp{s},
2002 and the second character is a
2003 minus. In this case, the character ^tilde^dollar sign^ is used in place
2004 of the minus. The reason for this special rule is to avoid clashes with
2005 the standard names for child units of the packages System, Ada,
2006 Interfaces, and GNAT, which use the prefixes
2008 @samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},
2011 @samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},
2015 The file extension is @file{.ads} for a spec and
2016 @file{.adb} for a body. The following list shows some
2017 examples of these rules.
2024 @item arith_functions.ads
2025 Arith_Functions (package spec)
2026 @item arith_functions.adb
2027 Arith_Functions (package body)
2029 Func.Spec (child package spec)
2031 Func.Spec (child package body)
2033 Sub (subunit of Main)
2034 @item ^a~bad.adb^A$BAD.ADB^
2035 A.Bad (child package body)
2039 Following these rules can result in excessively long
2040 file names if corresponding
2041 unit names are long (for example, if child units or subunits are
2042 heavily nested). An option is available to shorten such long file names
2043 (called file name ``krunching''). This may be particularly useful when
2044 programs being developed with GNAT are to be used on operating systems
2045 with limited file name lengths. @xref{Using gnatkr}.
2047 Of course, no file shortening algorithm can guarantee uniqueness over
2048 all possible unit names; if file name krunching is used, it is your
2049 responsibility to ensure no name clashes occur. Alternatively you
2050 can specify the exact file names that you want used, as described
2051 in the next section. Finally, if your Ada programs are migrating from a
2052 compiler with a different naming convention, you can use the gnatchop
2053 utility to produce source files that follow the GNAT naming conventions.
2054 (For details @pxref{Renaming Files Using gnatchop}.)
2056 Note: in the case of @code{Windows NT/XP} or @code{OpenVMS} operating
2057 systems, case is not significant. So for example on @code{Windows XP}
2058 if the canonical name is @code{main-sub.adb}, you can use the file name
2059 @code{Main-Sub.adb} instead. However, case is significant for other
2060 operating systems, so for example, if you want to use other than
2061 canonically cased file names on a Unix system, you need to follow
2062 the procedures described in the next section.
2064 @node Using Other File Names
2065 @section Using Other File Names
2069 In the previous section, we have described the default rules used by
2070 GNAT to determine the file name in which a given unit resides. It is
2071 often convenient to follow these default rules, and if you follow them,
2072 the compiler knows without being explicitly told where to find all
2075 However, in some cases, particularly when a program is imported from
2076 another Ada compiler environment, it may be more convenient for the
2077 programmer to specify which file names contain which units. GNAT allows
2078 arbitrary file names to be used by means of the Source_File_Name pragma.
2079 The form of this pragma is as shown in the following examples:
2080 @cindex Source_File_Name pragma
2082 @smallexample @c ada
2084 pragma Source_File_Name (My_Utilities.Stacks,
2085 Spec_File_Name => "myutilst_a.ada");
2086 pragma Source_File_name (My_Utilities.Stacks,
2087 Body_File_Name => "myutilst.ada");
2092 As shown in this example, the first argument for the pragma is the unit
2093 name (in this example a child unit). The second argument has the form
2094 of a named association. The identifier
2095 indicates whether the file name is for a spec or a body;
2096 the file name itself is given by a string literal.
2098 The source file name pragma is a configuration pragma, which means that
2099 normally it will be placed in the @file{gnat.adc}
2100 file used to hold configuration
2101 pragmas that apply to a complete compilation environment.
2102 For more details on how the @file{gnat.adc} file is created and used
2103 see @ref{Handling of Configuration Pragmas}.
2104 @cindex @file{gnat.adc}
2107 GNAT allows completely arbitrary file names to be specified using the
2108 source file name pragma. However, if the file name specified has an
2109 extension other than @file{.ads} or @file{.adb} it is necessary to use
2110 a special syntax when compiling the file. The name in this case must be
2111 preceded by the special sequence @option{-x} followed by a space and the name
2112 of the language, here @code{ada}, as in:
2115 $ gcc -c -x ada peculiar_file_name.sim
2120 @command{gnatmake} handles non-standard file names in the usual manner (the
2121 non-standard file name for the main program is simply used as the
2122 argument to gnatmake). Note that if the extension is also non-standard,
2123 then it must be included in the @command{gnatmake} command, it may not
2126 @node Alternative File Naming Schemes
2127 @section Alternative File Naming Schemes
2128 @cindex File naming schemes, alternative
2131 In the previous section, we described the use of the @code{Source_File_Name}
2132 pragma to allow arbitrary names to be assigned to individual source files.
2133 However, this approach requires one pragma for each file, and especially in
2134 large systems can result in very long @file{gnat.adc} files, and also create
2135 a maintenance problem.
2137 GNAT also provides a facility for specifying systematic file naming schemes
2138 other than the standard default naming scheme previously described. An
2139 alternative scheme for naming is specified by the use of
2140 @code{Source_File_Name} pragmas having the following format:
2141 @cindex Source_File_Name pragma
2143 @smallexample @c ada
2144 pragma Source_File_Name (
2145 Spec_File_Name => FILE_NAME_PATTERN
2146 @r{[},Casing => CASING_SPEC@r{]}
2147 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2149 pragma Source_File_Name (
2150 Body_File_Name => FILE_NAME_PATTERN
2151 @r{[},Casing => CASING_SPEC@r{]}
2152 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2154 pragma Source_File_Name (
2155 Subunit_File_Name => FILE_NAME_PATTERN
2156 @r{[},Casing => CASING_SPEC@r{]}
2157 @r{[},Dot_Replacement => STRING_LITERAL@r{]});
2159 FILE_NAME_PATTERN ::= STRING_LITERAL
2160 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
2164 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
2165 It contains a single asterisk character, and the unit name is substituted
2166 systematically for this asterisk. The optional parameter
2167 @code{Casing} indicates
2168 whether the unit name is to be all upper-case letters, all lower-case letters,
2169 or mixed-case. If no
2170 @code{Casing} parameter is used, then the default is all
2171 ^lower-case^upper-case^.
2173 The optional @code{Dot_Replacement} string is used to replace any periods
2174 that occur in subunit or child unit names. If no @code{Dot_Replacement}
2175 argument is used then separating dots appear unchanged in the resulting
2177 Although the above syntax indicates that the
2178 @code{Casing} argument must appear
2179 before the @code{Dot_Replacement} argument, but it
2180 is also permissible to write these arguments in the opposite order.
2182 As indicated, it is possible to specify different naming schemes for
2183 bodies, specs, and subunits. Quite often the rule for subunits is the
2184 same as the rule for bodies, in which case, there is no need to give
2185 a separate @code{Subunit_File_Name} rule, and in this case the
2186 @code{Body_File_name} rule is used for subunits as well.
2188 The separate rule for subunits can also be used to implement the rather
2189 unusual case of a compilation environment (e.g.@: a single directory) which
2190 contains a subunit and a child unit with the same unit name. Although
2191 both units cannot appear in the same partition, the Ada Reference Manual
2192 allows (but does not require) the possibility of the two units coexisting
2193 in the same environment.
2195 The file name translation works in the following steps:
2200 If there is a specific @code{Source_File_Name} pragma for the given unit,
2201 then this is always used, and any general pattern rules are ignored.
2204 If there is a pattern type @code{Source_File_Name} pragma that applies to
2205 the unit, then the resulting file name will be used if the file exists. If
2206 more than one pattern matches, the latest one will be tried first, and the
2207 first attempt resulting in a reference to a file that exists will be used.
2210 If no pattern type @code{Source_File_Name} pragma that applies to the unit
2211 for which the corresponding file exists, then the standard GNAT default
2212 naming rules are used.
2217 As an example of the use of this mechanism, consider a commonly used scheme
2218 in which file names are all lower case, with separating periods copied
2219 unchanged to the resulting file name, and specs end with @file{.1.ada}, and
2220 bodies end with @file{.2.ada}. GNAT will follow this scheme if the following
2223 @smallexample @c ada
2224 pragma Source_File_Name
2225 (Spec_File_Name => "*.1.ada");
2226 pragma Source_File_Name
2227 (Body_File_Name => "*.2.ada");
2231 The default GNAT scheme is actually implemented by providing the following
2232 default pragmas internally:
2234 @smallexample @c ada
2235 pragma Source_File_Name
2236 (Spec_File_Name => "*.ads", Dot_Replacement => "-");
2237 pragma Source_File_Name
2238 (Body_File_Name => "*.adb", Dot_Replacement => "-");
2242 Our final example implements a scheme typically used with one of the
2243 Ada 83 compilers, where the separator character for subunits was ``__''
2244 (two underscores), specs were identified by adding @file{_.ADA}, bodies
2245 by adding @file{.ADA}, and subunits by
2246 adding @file{.SEP}. All file names were
2247 upper case. Child units were not present of course since this was an
2248 Ada 83 compiler, but it seems reasonable to extend this scheme to use
2249 the same double underscore separator for child units.
2251 @smallexample @c ada
2252 pragma Source_File_Name
2253 (Spec_File_Name => "*_.ADA",
2254 Dot_Replacement => "__",
2255 Casing = Uppercase);
2256 pragma Source_File_Name
2257 (Body_File_Name => "*.ADA",
2258 Dot_Replacement => "__",
2259 Casing = Uppercase);
2260 pragma Source_File_Name
2261 (Subunit_File_Name => "*.SEP",
2262 Dot_Replacement => "__",
2263 Casing = Uppercase);
2266 @node Generating Object Files
2267 @section Generating Object Files
2270 An Ada program consists of a set of source files, and the first step in
2271 compiling the program is to generate the corresponding object files.
2272 These are generated by compiling a subset of these source files.
2273 The files you need to compile are the following:
2277 If a package spec has no body, compile the package spec to produce the
2278 object file for the package.
2281 If a package has both a spec and a body, compile the body to produce the
2282 object file for the package. The source file for the package spec need
2283 not be compiled in this case because there is only one object file, which
2284 contains the code for both the spec and body of the package.
2287 For a subprogram, compile the subprogram body to produce the object file
2288 for the subprogram. The spec, if one is present, is as usual in a
2289 separate file, and need not be compiled.
2293 In the case of subunits, only compile the parent unit. A single object
2294 file is generated for the entire subunit tree, which includes all the
2298 Compile child units independently of their parent units
2299 (though, of course, the spec of all the ancestor unit must be present in order
2300 to compile a child unit).
2304 Compile generic units in the same manner as any other units. The object
2305 files in this case are small dummy files that contain at most the
2306 flag used for elaboration checking. This is because GNAT always handles generic
2307 instantiation by means of macro expansion. However, it is still necessary to
2308 compile generic units, for dependency checking and elaboration purposes.
2312 The preceding rules describe the set of files that must be compiled to
2313 generate the object files for a program. Each object file has the same
2314 name as the corresponding source file, except that the extension is
2317 You may wish to compile other files for the purpose of checking their
2318 syntactic and semantic correctness. For example, in the case where a
2319 package has a separate spec and body, you would not normally compile the
2320 spec. However, it is convenient in practice to compile the spec to make
2321 sure it is error-free before compiling clients of this spec, because such
2322 compilations will fail if there is an error in the spec.
2324 GNAT provides an option for compiling such files purely for the
2325 purposes of checking correctness; such compilations are not required as
2326 part of the process of building a program. To compile a file in this
2327 checking mode, use the @option{-gnatc} switch.
2329 @node Source Dependencies
2330 @section Source Dependencies
2333 A given object file clearly depends on the source file which is compiled
2334 to produce it. Here we are using @dfn{depends} in the sense of a typical
2335 @code{make} utility; in other words, an object file depends on a source
2336 file if changes to the source file require the object file to be
2338 In addition to this basic dependency, a given object may depend on
2339 additional source files as follows:
2343 If a file being compiled @code{with}'s a unit @var{X}, the object file
2344 depends on the file containing the spec of unit @var{X}. This includes
2345 files that are @code{with}'ed implicitly either because they are parents
2346 of @code{with}'ed child units or they are run-time units required by the
2347 language constructs used in a particular unit.
2350 If a file being compiled instantiates a library level generic unit, the
2351 object file depends on both the spec and body files for this generic
2355 If a file being compiled instantiates a generic unit defined within a
2356 package, the object file depends on the body file for the package as
2357 well as the spec file.
2361 @cindex @option{-gnatn} switch
2362 If a file being compiled contains a call to a subprogram for which
2363 pragma @code{Inline} applies and inlining is activated with the
2364 @option{-gnatn} switch, the object file depends on the file containing the
2365 body of this subprogram as well as on the file containing the spec. Note
2366 that for inlining to actually occur as a result of the use of this switch,
2367 it is necessary to compile in optimizing mode.
2369 @cindex @option{-gnatN} switch
2370 The use of @option{-gnatN} activates inlining optimization
2371 that is performed by the front end of the compiler. This inlining does
2372 not require that the code generation be optimized. Like @option{-gnatn},
2373 the use of this switch generates additional dependencies.
2375 When using a gcc-based back end (in practice this means using any version
2376 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
2377 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
2378 Historically front end inlining was more extensive than the gcc back end
2379 inlining, but that is no longer the case.
2382 If an object file @file{O} depends on the proper body of a subunit through
2383 inlining or instantiation, it depends on the parent unit of the subunit.
2384 This means that any modification of the parent unit or one of its subunits
2385 affects the compilation of @file{O}.
2388 The object file for a parent unit depends on all its subunit body files.
2391 The previous two rules meant that for purposes of computing dependencies and
2392 recompilation, a body and all its subunits are treated as an indivisible whole.
2395 These rules are applied transitively: if unit @code{A} @code{with}'s
2396 unit @code{B}, whose elaboration calls an inlined procedure in package
2397 @code{C}, the object file for unit @code{A} will depend on the body of
2398 @code{C}, in file @file{c.adb}.
2400 The set of dependent files described by these rules includes all the
2401 files on which the unit is semantically dependent, as dictated by the
2402 Ada language standard. However, it is a superset of what the
2403 standard describes, because it includes generic, inline, and subunit
2406 An object file must be recreated by recompiling the corresponding source
2407 file if any of the source files on which it depends are modified. For
2408 example, if the @code{make} utility is used to control compilation,
2409 the rule for an Ada object file must mention all the source files on
2410 which the object file depends, according to the above definition.
2411 The determination of the necessary
2412 recompilations is done automatically when one uses @command{gnatmake}.
2415 @node The Ada Library Information Files
2416 @section The Ada Library Information Files
2417 @cindex Ada Library Information files
2418 @cindex @file{ALI} files
2421 Each compilation actually generates two output files. The first of these
2422 is the normal object file that has a @file{.o} extension. The second is a
2423 text file containing full dependency information. It has the same
2424 name as the source file, but an @file{.ali} extension.
2425 This file is known as the Ada Library Information (@file{ALI}) file.
2426 The following information is contained in the @file{ALI} file.
2430 Version information (indicates which version of GNAT was used to compile
2431 the unit(s) in question)
2434 Main program information (including priority and time slice settings,
2435 as well as the wide character encoding used during compilation).
2438 List of arguments used in the @command{gcc} command for the compilation
2441 Attributes of the unit, including configuration pragmas used, an indication
2442 of whether the compilation was successful, exception model used etc.
2445 A list of relevant restrictions applying to the unit (used for consistency)
2449 Categorization information (e.g.@: use of pragma @code{Pure}).
2452 Information on all @code{with}'ed units, including presence of
2453 @code{Elaborate} or @code{Elaborate_All} pragmas.
2456 Information from any @code{Linker_Options} pragmas used in the unit
2459 Information on the use of @code{Body_Version} or @code{Version}
2460 attributes in the unit.
2463 Dependency information. This is a list of files, together with
2464 time stamp and checksum information. These are files on which
2465 the unit depends in the sense that recompilation is required
2466 if any of these units are modified.
2469 Cross-reference data. Contains information on all entities referenced
2470 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
2471 provide cross-reference information.
2476 For a full detailed description of the format of the @file{ALI} file,
2477 see the source of the body of unit @code{Lib.Writ}, contained in file
2478 @file{lib-writ.adb} in the GNAT compiler sources.
2480 @node Binding an Ada Program
2481 @section Binding an Ada Program
2484 When using languages such as C and C++, once the source files have been
2485 compiled the only remaining step in building an executable program
2486 is linking the object modules together. This means that it is possible to
2487 link an inconsistent version of a program, in which two units have
2488 included different versions of the same header.
2490 The rules of Ada do not permit such an inconsistent program to be built.
2491 For example, if two clients have different versions of the same package,
2492 it is illegal to build a program containing these two clients.
2493 These rules are enforced by the GNAT binder, which also determines an
2494 elaboration order consistent with the Ada rules.
2496 The GNAT binder is run after all the object files for a program have
2497 been created. It is given the name of the main program unit, and from
2498 this it determines the set of units required by the program, by reading the
2499 corresponding ALI files. It generates error messages if the program is
2500 inconsistent or if no valid order of elaboration exists.
2502 If no errors are detected, the binder produces a main program, in Ada by
2503 default, that contains calls to the elaboration procedures of those
2504 compilation unit that require them, followed by
2505 a call to the main program. This Ada program is compiled to generate the
2506 object file for the main program. The name of
2507 the Ada file is @file{b~@var{xxx}.adb} (with the corresponding spec
2508 @file{b~@var{xxx}.ads}) where @var{xxx} is the name of the
2511 Finally, the linker is used to build the resulting executable program,
2512 using the object from the main program from the bind step as well as the
2513 object files for the Ada units of the program.
2515 @node Mixed Language Programming
2516 @section Mixed Language Programming
2517 @cindex Mixed Language Programming
2520 This section describes how to develop a mixed-language program,
2521 specifically one that comprises units in both Ada and C.
2524 * Interfacing to C::
2525 * Calling Conventions::
2528 @node Interfacing to C
2529 @subsection Interfacing to C
2531 Interfacing Ada with a foreign language such as C involves using
2532 compiler directives to import and/or export entity definitions in each
2533 language---using @code{extern} statements in C, for instance, and the
2534 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada.
2535 A full treatment of these topics is provided in Appendix B, section 1
2536 of the Ada Reference Manual.
2538 There are two ways to build a program using GNAT that contains some Ada
2539 sources and some foreign language sources, depending on whether or not
2540 the main subprogram is written in Ada. Here is a source example with
2541 the main subprogram in Ada:
2547 void print_num (int num)
2549 printf ("num is %d.\n", num);
2555 /* num_from_Ada is declared in my_main.adb */
2556 extern int num_from_Ada;
2560 return num_from_Ada;
2564 @smallexample @c ada
2566 procedure My_Main is
2568 -- Declare then export an Integer entity called num_from_Ada
2569 My_Num : Integer := 10;
2570 pragma Export (C, My_Num, "num_from_Ada");
2572 -- Declare an Ada function spec for Get_Num, then use
2573 -- C function get_num for the implementation.
2574 function Get_Num return Integer;
2575 pragma Import (C, Get_Num, "get_num");
2577 -- Declare an Ada procedure spec for Print_Num, then use
2578 -- C function print_num for the implementation.
2579 procedure Print_Num (Num : Integer);
2580 pragma Import (C, Print_Num, "print_num");
2583 Print_Num (Get_Num);
2589 To build this example, first compile the foreign language files to
2590 generate object files:
2592 ^gcc -c file1.c^gcc -c FILE1.C^
2593 ^gcc -c file2.c^gcc -c FILE2.C^
2597 Then, compile the Ada units to produce a set of object files and ALI
2600 gnatmake ^-c^/ACTIONS=COMPILE^ my_main.adb
2604 Run the Ada binder on the Ada main program:
2606 gnatbind my_main.ali
2610 Link the Ada main program, the Ada objects and the other language
2613 gnatlink my_main.ali file1.o file2.o
2617 The last three steps can be grouped in a single command:
2619 gnatmake my_main.adb -largs file1.o file2.o
2622 @cindex Binder output file
2624 If the main program is in a language other than Ada, then you may have
2625 more than one entry point into the Ada subsystem. You must use a special
2626 binder option to generate callable routines that initialize and
2627 finalize the Ada units (@pxref{Binding with Non-Ada Main Programs}).
2628 Calls to the initialization and finalization routines must be inserted
2629 in the main program, or some other appropriate point in the code. The
2630 call to initialize the Ada units must occur before the first Ada
2631 subprogram is called, and the call to finalize the Ada units must occur
2632 after the last Ada subprogram returns. The binder will place the
2633 initialization and finalization subprograms into the
2634 @file{b~@var{xxx}.adb} file where they can be accessed by your C
2635 sources. To illustrate, we have the following example:
2639 extern void adainit (void);
2640 extern void adafinal (void);
2641 extern int add (int, int);
2642 extern int sub (int, int);
2644 int main (int argc, char *argv[])
2650 /* Should print "21 + 7 = 28" */
2651 printf ("%d + %d = %d\n", a, b, add (a, b));
2652 /* Should print "21 - 7 = 14" */
2653 printf ("%d - %d = %d\n", a, b, sub (a, b));
2659 @smallexample @c ada
2662 function Add (A, B : Integer) return Integer;
2663 pragma Export (C, Add, "add");
2667 package body Unit1 is
2668 function Add (A, B : Integer) return Integer is
2676 function Sub (A, B : Integer) return Integer;
2677 pragma Export (C, Sub, "sub");
2681 package body Unit2 is
2682 function Sub (A, B : Integer) return Integer is
2691 The build procedure for this application is similar to the last
2692 example's. First, compile the foreign language files to generate object
2695 ^gcc -c main.c^gcc -c main.c^
2699 Next, compile the Ada units to produce a set of object files and ALI
2702 gnatmake ^-c^/ACTIONS=COMPILE^ unit1.adb
2703 gnatmake ^-c^/ACTIONS=COMPILE^ unit2.adb
2707 Run the Ada binder on every generated ALI file. Make sure to use the
2708 @option{-n} option to specify a foreign main program:
2710 gnatbind ^-n^/NOMAIN^ unit1.ali unit2.ali
2714 Link the Ada main program, the Ada objects and the foreign language
2715 objects. You need only list the last ALI file here:
2717 gnatlink unit2.ali main.o -o exec_file
2720 This procedure yields a binary executable called @file{exec_file}.
2724 Depending on the circumstances (for example when your non-Ada main object
2725 does not provide symbol @code{main}), you may also need to instruct the
2726 GNAT linker not to include the standard startup objects by passing the
2727 @option{^-nostartfiles^/NOSTART_FILES^} switch to @command{gnatlink}.
2729 @node Calling Conventions
2730 @subsection Calling Conventions
2731 @cindex Foreign Languages
2732 @cindex Calling Conventions
2733 GNAT follows standard calling sequence conventions and will thus interface
2734 to any other language that also follows these conventions. The following
2735 Convention identifiers are recognized by GNAT:
2738 @cindex Interfacing to Ada
2739 @cindex Other Ada compilers
2740 @cindex Convention Ada
2742 This indicates that the standard Ada calling sequence will be
2743 used and all Ada data items may be passed without any limitations in the
2744 case where GNAT is used to generate both the caller and callee. It is also
2745 possible to mix GNAT generated code and code generated by another Ada
2746 compiler. In this case, the data types should be restricted to simple
2747 cases, including primitive types. Whether complex data types can be passed
2748 depends on the situation. Probably it is safe to pass simple arrays, such
2749 as arrays of integers or floats. Records may or may not work, depending
2750 on whether both compilers lay them out identically. Complex structures
2751 involving variant records, access parameters, tasks, or protected types,
2752 are unlikely to be able to be passed.
2754 Note that in the case of GNAT running
2755 on a platform that supports HP Ada 83, a higher degree of compatibility
2756 can be guaranteed, and in particular records are layed out in an identical
2757 manner in the two compilers. Note also that if output from two different
2758 compilers is mixed, the program is responsible for dealing with elaboration
2759 issues. Probably the safest approach is to write the main program in the
2760 version of Ada other than GNAT, so that it takes care of its own elaboration
2761 requirements, and then call the GNAT-generated adainit procedure to ensure
2762 elaboration of the GNAT components. Consult the documentation of the other
2763 Ada compiler for further details on elaboration.
2765 However, it is not possible to mix the tasking run time of GNAT and
2766 HP Ada 83, All the tasking operations must either be entirely within
2767 GNAT compiled sections of the program, or entirely within HP Ada 83
2768 compiled sections of the program.
2770 @cindex Interfacing to Assembly
2771 @cindex Convention Assembler
2773 Specifies assembler as the convention. In practice this has the
2774 same effect as convention Ada (but is not equivalent in the sense of being
2775 considered the same convention).
2777 @cindex Convention Asm
2780 Equivalent to Assembler.
2782 @cindex Interfacing to COBOL
2783 @cindex Convention COBOL
2786 Data will be passed according to the conventions described
2787 in section B.4 of the Ada Reference Manual.
2790 @cindex Interfacing to C
2791 @cindex Convention C
2793 Data will be passed according to the conventions described
2794 in section B.3 of the Ada Reference Manual.
2796 A note on interfacing to a C ``varargs'' function:
2797 @findex C varargs function
2798 @cindex Interfacing to C varargs function
2799 @cindex varargs function interfaces
2803 In C, @code{varargs} allows a function to take a variable number of
2804 arguments. There is no direct equivalent in this to Ada. One
2805 approach that can be used is to create a C wrapper for each
2806 different profile and then interface to this C wrapper. For
2807 example, to print an @code{int} value using @code{printf},
2808 create a C function @code{printfi} that takes two arguments, a
2809 pointer to a string and an int, and calls @code{printf}.
2810 Then in the Ada program, use pragma @code{Import} to
2811 interface to @code{printfi}.
2814 It may work on some platforms to directly interface to
2815 a @code{varargs} function by providing a specific Ada profile
2816 for a particular call. However, this does not work on
2817 all platforms, since there is no guarantee that the
2818 calling sequence for a two argument normal C function
2819 is the same as for calling a @code{varargs} C function with
2820 the same two arguments.
2823 @cindex Convention Default
2828 @cindex Convention External
2835 @cindex Interfacing to C++
2836 @cindex Convention C++
2837 @item C_Plus_Plus (or CPP)
2838 This stands for C++. For most purposes this is identical to C.
2839 See the separate description of the specialized GNAT pragmas relating to
2840 C++ interfacing for further details.
2844 @cindex Interfacing to Fortran
2845 @cindex Convention Fortran
2847 Data will be passed according to the conventions described
2848 in section B.5 of the Ada Reference Manual.
2851 This applies to an intrinsic operation, as defined in the Ada
2852 Reference Manual. If a pragma Import (Intrinsic) applies to a subprogram,
2853 this means that the body of the subprogram is provided by the compiler itself,
2854 usually by means of an efficient code sequence, and that the user does not
2855 supply an explicit body for it. In an application program, the pragma may
2856 be applied to the following sets of names:
2860 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right,
2861 Shift_Right_Arithmetic. The corresponding subprogram declaration must have
2862 two formal parameters. The
2863 first one must be a signed integer type or a modular type with a binary
2864 modulus, and the second parameter must be of type Natural.
2865 The return type must be the same as the type of the first argument. The size
2866 of this type can only be 8, 16, 32, or 64.
2869 Binary arithmetic operators: ``+'', ``-'', ``*'', ``/''
2870 The corresponding operator declaration must have parameters and result type
2871 that have the same root numeric type (for example, all three are long_float
2872 types). This simplifies the definition of operations that use type checking
2873 to perform dimensional checks:
2875 @smallexample @c ada
2876 type Distance is new Long_Float;
2877 type Time is new Long_Float;
2878 type Velocity is new Long_Float;
2879 function "/" (D : Distance; T : Time)
2881 pragma Import (Intrinsic, "/");
2885 This common idiom is often programmed with a generic definition and an
2886 explicit body. The pragma makes it simpler to introduce such declarations.
2887 It incurs no overhead in compilation time or code size, because it is
2888 implemented as a single machine instruction.
2891 General subprogram entities, to bind an Ada subprogram declaration to
2892 a compiler builtin by name with back-ends where such interfaces are
2893 available. A typical example is the set of ``__builtin'' functions
2894 exposed by the GCC back-end, as in the following example:
2896 @smallexample @c ada
2897 function builtin_sqrt (F : Float) return Float;
2898 pragma Import (Intrinsic, builtin_sqrt, "__builtin_sqrtf");
2901 Most of the GCC builtins are accessible this way, and as for other
2902 import conventions (e.g. C), it is the user's responsibility to ensure
2903 that the Ada subprogram profile matches the underlying builtin
2911 @cindex Convention Stdcall
2913 This is relevant only to Windows XP/2000/NT implementations of GNAT,
2914 and specifies that the @code{Stdcall} calling sequence will be used,
2915 as defined by the NT API. Nevertheless, to ease building
2916 cross-platform bindings this convention will be handled as a @code{C} calling
2917 convention on non-Windows platforms.
2920 @cindex Convention DLL
2922 This is equivalent to @code{Stdcall}.
2925 @cindex Convention Win32
2927 This is equivalent to @code{Stdcall}.
2931 @cindex Convention Stubbed
2933 This is a special convention that indicates that the compiler
2934 should provide a stub body that raises @code{Program_Error}.
2938 GNAT additionally provides a useful pragma @code{Convention_Identifier}
2939 that can be used to parametrize conventions and allow additional synonyms
2940 to be specified. For example if you have legacy code in which the convention
2941 identifier Fortran77 was used for Fortran, you can use the configuration
2944 @smallexample @c ada
2945 pragma Convention_Identifier (Fortran77, Fortran);
2949 And from now on the identifier Fortran77 may be used as a convention
2950 identifier (for example in an @code{Import} pragma) with the same
2954 @node Building Mixed Ada & C++ Programs
2955 @section Building Mixed Ada and C++ Programs
2958 A programmer inexperienced with mixed-language development may find that
2959 building an application containing both Ada and C++ code can be a
2960 challenge. This section gives a few
2961 hints that should make this task easier. The first section addresses
2962 the differences between interfacing with C and interfacing with C++.
2964 looks into the delicate problem of linking the complete application from
2965 its Ada and C++ parts. The last section gives some hints on how the GNAT
2966 run-time library can be adapted in order to allow inter-language dispatching
2967 with a new C++ compiler.
2970 * Interfacing to C++::
2971 * Linking a Mixed C++ & Ada Program::
2972 * A Simple Example::
2973 * Interfacing with C++ constructors::
2974 * Interfacing with C++ at the Class Level::
2977 @node Interfacing to C++
2978 @subsection Interfacing to C++
2981 GNAT supports interfacing with the G++ compiler (or any C++ compiler
2982 generating code that is compatible with the G++ Application Binary
2983 Interface ---see http://www.codesourcery.com/archives/cxx-abi).
2986 Interfacing can be done at 3 levels: simple data, subprograms, and
2987 classes. In the first two cases, GNAT offers a specific @code{Convention
2988 C_Plus_Plus} (or @code{CPP}) that behaves exactly like @code{Convention C}.
2989 Usually, C++ mangles the names of subprograms. To generate proper mangled
2990 names automatically, see @ref{Generating Ada Bindings for C and C++ headers}).
2991 This problem can also be addressed manually in two ways:
2995 by modifying the C++ code in order to force a C convention using
2996 the @code{extern "C"} syntax.
2999 by figuring out the mangled name (using e.g. @command{nm}) and using it as the
3000 Link_Name argument of the pragma import.
3004 Interfacing at the class level can be achieved by using the GNAT specific
3005 pragmas such as @code{CPP_Constructor}. @xref{Interfacing to C++,,,
3006 gnat_rm, GNAT Reference Manual}, for additional information.
3008 @node Linking a Mixed C++ & Ada Program
3009 @subsection Linking a Mixed C++ & Ada Program
3012 Usually the linker of the C++ development system must be used to link
3013 mixed applications because most C++ systems will resolve elaboration
3014 issues (such as calling constructors on global class instances)
3015 transparently during the link phase. GNAT has been adapted to ease the
3016 use of a foreign linker for the last phase. Three cases can be
3021 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
3022 The C++ linker can simply be called by using the C++ specific driver
3025 Note that if the C++ code uses inline functions, you will need to
3026 compile your C++ code with the @code{-fkeep-inline-functions} switch in
3027 order to provide an existing function implementation that the Ada code can
3031 $ g++ -c -fkeep-inline-functions file1.C
3032 $ g++ -c -fkeep-inline-functions file2.C
3033 $ gnatmake ada_unit -largs file1.o file2.o --LINK=g++
3037 Using GNAT and G++ from two different GCC installations: If both
3038 compilers are on the @env{PATH}, the previous method may be used. It is
3039 important to note that environment variables such as
3040 @env{C_INCLUDE_PATH}, @env{GCC_EXEC_PREFIX}, @env{BINUTILS_ROOT}, and
3041 @env{GCC_ROOT} will affect both compilers
3042 at the same time and may make one of the two compilers operate
3043 improperly if set during invocation of the wrong compiler. It is also
3044 very important that the linker uses the proper @file{libgcc.a} GCC
3045 library -- that is, the one from the C++ compiler installation. The
3046 implicit link command as suggested in the @command{gnatmake} command
3047 from the former example can be replaced by an explicit link command with
3048 the full-verbosity option in order to verify which library is used:
3051 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
3053 If there is a problem due to interfering environment variables, it can
3054 be worked around by using an intermediate script. The following example
3055 shows the proper script to use when GNAT has not been installed at its
3056 default location and g++ has been installed at its default location:
3064 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
3068 Using a non-GNU C++ compiler: The commands previously described can be
3069 used to insure that the C++ linker is used. Nonetheless, you need to add
3070 a few more parameters to the link command line, depending on the exception
3073 If the @code{setjmp/longjmp} exception mechanism is used, only the paths
3074 to the libgcc libraries are required:
3079 CC $* `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a`
3080 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3083 Where CC is the name of the non-GNU C++ compiler.
3085 If the @code{zero cost} exception mechanism is used, and the platform
3086 supports automatic registration of exception tables (e.g.@: Solaris or IRIX),
3087 paths to more objects are required:
3092 CC `gcc -print-file-name=crtbegin.o` $* \
3093 `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a` \
3094 `gcc -print-file-name=crtend.o`
3095 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3098 If the @code{zero cost} exception mechanism is used, and the platform
3099 doesn't support automatic registration of exception tables (e.g.@: HP-UX,
3100 Tru64 or AIX), the simple approach described above will not work and
3101 a pre-linking phase using GNAT will be necessary.
3105 Another alternative is to use the @command{gprbuild} multi-language builder
3106 which has a large knowledge base and knows how to link Ada and C++ code
3107 together automatically in most cases.
3109 @node A Simple Example
3110 @subsection A Simple Example
3112 The following example, provided as part of the GNAT examples, shows how
3113 to achieve procedural interfacing between Ada and C++ in both
3114 directions. The C++ class A has two methods. The first method is exported
3115 to Ada by the means of an extern C wrapper function. The second method
3116 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
3117 a limited record with a layout comparable to the C++ class. The Ada
3118 subprogram, in turn, calls the C++ method. So, starting from the C++
3119 main program, the process passes back and forth between the two
3123 Here are the compilation commands:
3125 $ gnatmake -c simple_cpp_interface
3128 $ gnatbind -n simple_cpp_interface
3129 $ gnatlink simple_cpp_interface -o cpp_main --LINK=g++
3130 -lstdc++ ex7.o cpp_main.o
3134 Here are the corresponding sources:
3142 void adainit (void);
3143 void adafinal (void);
3144 void method1 (A *t);
3166 class A : public Origin @{
3168 void method1 (void);
3169 void method2 (int v);
3179 extern "C" @{ void ada_method2 (A *t, int v);@}
3181 void A::method1 (void)
3184 printf ("in A::method1, a_value = %d \n",a_value);
3188 void A::method2 (int v)
3190 ada_method2 (this, v);
3191 printf ("in A::method2, a_value = %d \n",a_value);
3198 printf ("in A::A, a_value = %d \n",a_value);
3202 @smallexample @c ada
3204 package body Simple_Cpp_Interface is
3206 procedure Ada_Method2 (This : in out A; V : Integer) is
3212 end Simple_Cpp_Interface;
3215 package Simple_Cpp_Interface is
3218 Vptr : System.Address;
3222 pragma Convention (C, A);
3224 procedure Method1 (This : in out A);
3225 pragma Import (C, Method1);
3227 procedure Ada_Method2 (This : in out A; V : Integer);
3228 pragma Export (C, Ada_Method2);
3230 end Simple_Cpp_Interface;
3233 @node Interfacing with C++ constructors
3234 @subsection Interfacing with C++ constructors
3237 In order to interface with C++ constructors GNAT provides the
3238 @code{pragma CPP_Constructor} (@xref{Interfacing to C++,,,
3239 gnat_rm, GNAT Reference Manual}, for additional information).
3240 In this section we present some common uses of C++ constructors
3241 in mixed-languages programs in GNAT.
3243 Let us assume that we need to interface with the following
3251 @b{virtual} int Get_Value ();
3252 Root(); // Default constructor
3253 Root(int v); // 1st non-default constructor
3254 Root(int v, int w); // 2nd non-default constructor
3258 For this purpose we can write the following package spec (further
3259 information on how to build this spec is available in
3260 @ref{Interfacing with C++ at the Class Level} and
3261 @ref{Generating Ada Bindings for C and C++ headers}).
3263 @smallexample @c ada
3264 with Interfaces.C; use Interfaces.C;
3266 type Root is tagged limited record
3270 pragma Import (CPP, Root);
3272 function Get_Value (Obj : Root) return int;
3273 pragma Import (CPP, Get_Value);
3275 function Constructor return Root;
3276 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ev");
3278 function Constructor (v : Integer) return Root;
3279 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Ei");
3281 function Constructor (v, w : Integer) return Root;
3282 pragma Cpp_Constructor (Constructor, "_ZN4RootC1Eii");
3286 On the Ada side the constructor is represented by a function (whose
3287 name is arbitrary) that returns the classwide type corresponding to
3288 the imported C++ class. Although the constructor is described as a
3289 function, it is typically a procedure with an extra implicit argument
3290 (the object being initialized) at the implementation level. GNAT
3291 issues the appropriate call, whatever it is, to get the object
3292 properly initialized.
3294 Constructors can only appear in the following contexts:
3298 On the right side of an initialization of an object of type @var{T}.
3300 On the right side of an initialization of a record component of type @var{T}.
3302 In an Ada 2005 limited aggregate.
3304 In an Ada 2005 nested limited aggregate.
3306 In an Ada 2005 limited aggregate that initializes an object built in
3307 place by an extended return statement.
3311 In a declaration of an object whose type is a class imported from C++,
3312 either the default C++ constructor is implicitly called by GNAT, or
3313 else the required C++ constructor must be explicitly called in the
3314 expression that initializes the object. For example:
3316 @smallexample @c ada
3318 Obj2 : Root := Constructor;
3319 Obj3 : Root := Constructor (v => 10);
3320 Obj4 : Root := Constructor (30, 40);
3323 The first two declarations are equivalent: in both cases the default C++
3324 constructor is invoked (in the former case the call to the constructor is
3325 implicit, and in the latter case the call is explicit in the object
3326 declaration). @code{Obj3} is initialized by the C++ non-default constructor
3327 that takes an integer argument, and @code{Obj4} is initialized by the
3328 non-default C++ constructor that takes two integers.
3330 Let us derive the imported C++ class in the Ada side. For example:
3332 @smallexample @c ada
3333 type DT is new Root with record
3334 C_Value : Natural := 2009;
3338 In this case the components DT inherited from the C++ side must be
3339 initialized by a C++ constructor, and the additional Ada components
3340 of type DT are initialized by GNAT. The initialization of such an
3341 object is done either by default, or by means of a function returning
3342 an aggregate of type DT, or by means of an extension aggregate.
3344 @smallexample @c ada
3346 Obj6 : DT := Function_Returning_DT (50);
3347 Obj7 : DT := (Constructor (30,40) with C_Value => 50);
3350 The declaration of @code{Obj5} invokes the default constructors: the
3351 C++ default constructor of the parent type takes care of the initialization
3352 of the components inherited from Root, and GNAT takes care of the default
3353 initialization of the additional Ada components of type DT (that is,
3354 @code{C_Value} is initialized to value 2009). The order of invocation of
3355 the constructors is consistent with the order of elaboration required by
3356 Ada and C++. That is, the constructor of the parent type is always called
3357 before the constructor of the derived type.
3359 Let us now consider a record that has components whose type is imported
3360 from C++. For example:
3362 @smallexample @c ada
3363 type Rec1 is limited record
3364 Data1 : Root := Constructor (10);
3365 Value : Natural := 1000;
3368 type Rec2 (D : Integer := 20) is limited record
3370 Data2 : Root := Constructor (D, 30);
3374 The initialization of an object of type @code{Rec2} will call the
3375 non-default C++ constructors specified for the imported components.
3378 @smallexample @c ada
3382 Using Ada 2005 we can use limited aggregates to initialize an object
3383 invoking C++ constructors that differ from those specified in the type
3384 declarations. For example:
3386 @smallexample @c ada
3387 Obj9 : Rec2 := (Rec => (Data1 => Constructor (15, 16),
3392 The above declaration uses an Ada 2005 limited aggregate to
3393 initialize @code{Obj9}, and the C++ constructor that has two integer
3394 arguments is invoked to initialize the @code{Data1} component instead
3395 of the constructor specified in the declaration of type @code{Rec1}. In
3396 Ada 2005 the box in the aggregate indicates that unspecified components
3397 are initialized using the expression (if any) available in the component
3398 declaration. That is, in this case discriminant @code{D} is initialized
3399 to value @code{20}, @code{Value} is initialized to value 1000, and the
3400 non-default C++ constructor that handles two integers takes care of
3401 initializing component @code{Data2} with values @code{20,30}.
3403 In Ada 2005 we can use the extended return statement to build the Ada
3404 equivalent to C++ non-default constructors. For example:
3406 @smallexample @c ada
3407 function Constructor (V : Integer) return Rec2 is
3409 return Obj : Rec2 := (Rec => (Data1 => Constructor (V, 20),
3412 -- Further actions required for construction of
3413 -- objects of type Rec2
3419 In this example the extended return statement construct is used to
3420 build in place the returned object whose components are initialized
3421 by means of a limited aggregate. Any further action associated with
3422 the constructor can be placed inside the construct.
3424 @node Interfacing with C++ at the Class Level
3425 @subsection Interfacing with C++ at the Class Level
3427 In this section we demonstrate the GNAT features for interfacing with
3428 C++ by means of an example making use of Ada 2005 abstract interface
3429 types. This example consists of a classification of animals; classes
3430 have been used to model our main classification of animals, and
3431 interfaces provide support for the management of secondary
3432 classifications. We first demonstrate a case in which the types and
3433 constructors are defined on the C++ side and imported from the Ada
3434 side, and latter the reverse case.
3436 The root of our derivation will be the @code{Animal} class, with a
3437 single private attribute (the @code{Age} of the animal) and two public
3438 primitives to set and get the value of this attribute.
3443 @b{virtual} void Set_Age (int New_Age);
3444 @b{virtual} int Age ();
3450 Abstract interface types are defined in C++ by means of classes with pure
3451 virtual functions and no data members. In our example we will use two
3452 interfaces that provide support for the common management of @code{Carnivore}
3453 and @code{Domestic} animals:
3456 @b{class} Carnivore @{
3458 @b{virtual} int Number_Of_Teeth () = 0;
3461 @b{class} Domestic @{
3463 @b{virtual void} Set_Owner (char* Name) = 0;
3467 Using these declarations, we can now say that a @code{Dog} is an animal that is
3468 both Carnivore and Domestic, that is:
3471 @b{class} Dog : Animal, Carnivore, Domestic @{
3473 @b{virtual} int Number_Of_Teeth ();
3474 @b{virtual} void Set_Owner (char* Name);
3476 Dog(); // Constructor
3483 In the following examples we will assume that the previous declarations are
3484 located in a file named @code{animals.h}. The following package demonstrates
3485 how to import these C++ declarations from the Ada side:
3487 @smallexample @c ada
3488 with Interfaces.C.Strings; use Interfaces.C.Strings;
3490 type Carnivore is interface;
3491 pragma Convention (C_Plus_Plus, Carnivore);
3492 function Number_Of_Teeth (X : Carnivore)
3493 return Natural is abstract;
3495 type Domestic is interface;
3496 pragma Convention (C_Plus_Plus, Set_Owner);
3498 (X : in out Domestic;
3499 Name : Chars_Ptr) is abstract;
3501 type Animal is tagged record
3504 pragma Import (C_Plus_Plus, Animal);
3506 procedure Set_Age (X : in out Animal; Age : Integer);
3507 pragma Import (C_Plus_Plus, Set_Age);
3509 function Age (X : Animal) return Integer;
3510 pragma Import (C_Plus_Plus, Age);
3512 type Dog is new Animal and Carnivore and Domestic with record
3513 Tooth_Count : Natural;
3514 Owner : String (1 .. 30);
3516 pragma Import (C_Plus_Plus, Dog);
3518 function Number_Of_Teeth (A : Dog) return Integer;
3519 pragma Import (C_Plus_Plus, Number_Of_Teeth);
3521 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3522 pragma Import (C_Plus_Plus, Set_Owner);
3524 function New_Dog return Dog;
3525 pragma CPP_Constructor (New_Dog);
3526 pragma Import (CPP, New_Dog, "_ZN3DogC2Ev");
3530 Thanks to the compatibility between GNAT run-time structures and the C++ ABI,
3531 interfacing with these C++ classes is easy. The only requirement is that all
3532 the primitives and components must be declared exactly in the same order in
3535 Regarding the abstract interfaces, we must indicate to the GNAT compiler by
3536 means of a @code{pragma Convention (C_Plus_Plus)}, the convention used to pass
3537 the arguments to the called primitives will be the same as for C++. For the
3538 imported classes we use @code{pragma Import} with convention @code{C_Plus_Plus}
3539 to indicate that they have been defined on the C++ side; this is required
3540 because the dispatch table associated with these tagged types will be built
3541 in the C++ side and therefore will not contain the predefined Ada primitives
3542 which Ada would otherwise expect.
3544 As the reader can see there is no need to indicate the C++ mangled names
3545 associated with each subprogram because it is assumed that all the calls to
3546 these primitives will be dispatching calls. The only exception is the
3547 constructor, which must be registered with the compiler by means of
3548 @code{pragma CPP_Constructor} and needs to provide its associated C++
3549 mangled name because the Ada compiler generates direct calls to it.
3551 With the above packages we can now declare objects of type Dog on the Ada side
3552 and dispatch calls to the corresponding subprograms on the C++ side. We can
3553 also extend the tagged type Dog with further fields and primitives, and
3554 override some of its C++ primitives on the Ada side. For example, here we have
3555 a type derivation defined on the Ada side that inherits all the dispatching
3556 primitives of the ancestor from the C++ side.
3559 @b{with} Animals; @b{use} Animals;
3560 @b{package} Vaccinated_Animals @b{is}
3561 @b{type} Vaccinated_Dog @b{is new} Dog @b{with null record};
3562 @b{function} Vaccination_Expired (A : Vaccinated_Dog) @b{return} Boolean;
3563 @b{end} Vaccinated_Animals;
3566 It is important to note that, because of the ABI compatibility, the programmer
3567 does not need to add any further information to indicate either the object
3568 layout or the dispatch table entry associated with each dispatching operation.
3570 Now let us define all the types and constructors on the Ada side and export
3571 them to C++, using the same hierarchy of our previous example:
3573 @smallexample @c ada
3574 with Interfaces.C.Strings;
3575 use Interfaces.C.Strings;
3577 type Carnivore is interface;
3578 pragma Convention (C_Plus_Plus, Carnivore);
3579 function Number_Of_Teeth (X : Carnivore)
3580 return Natural is abstract;
3582 type Domestic is interface;
3583 pragma Convention (C_Plus_Plus, Set_Owner);
3585 (X : in out Domestic;
3586 Name : Chars_Ptr) is abstract;
3588 type Animal is tagged record
3591 pragma Convention (C_Plus_Plus, Animal);
3593 procedure Set_Age (X : in out Animal; Age : Integer);
3594 pragma Export (C_Plus_Plus, Set_Age);
3596 function Age (X : Animal) return Integer;
3597 pragma Export (C_Plus_Plus, Age);
3599 type Dog is new Animal and Carnivore and Domestic with record
3600 Tooth_Count : Natural;
3601 Owner : String (1 .. 30);
3603 pragma Convention (C_Plus_Plus, Dog);
3605 function Number_Of_Teeth (A : Dog) return Integer;
3606 pragma Export (C_Plus_Plus, Number_Of_Teeth);
3608 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3609 pragma Export (C_Plus_Plus, Set_Owner);
3611 function New_Dog return Dog'Class;
3612 pragma Export (C_Plus_Plus, New_Dog);
3616 Compared with our previous example the only difference is the use of
3617 @code{pragma Export} to indicate to the GNAT compiler that the primitives will
3618 be available to C++. Thanks to the ABI compatibility, on the C++ side there is
3619 nothing else to be done; as explained above, the only requirement is that all
3620 the primitives and components are declared in exactly the same order.
3622 For completeness, let us see a brief C++ main program that uses the
3623 declarations available in @code{animals.h} (presented in our first example) to
3624 import and use the declarations from the Ada side, properly initializing and
3625 finalizing the Ada run-time system along the way:
3628 @b{#include} "animals.h"
3629 @b{#include} <iostream>
3630 @b{using namespace} std;
3632 void Check_Carnivore (Carnivore *obj) @{@dots{}@}
3633 void Check_Domestic (Domestic *obj) @{@dots{}@}
3634 void Check_Animal (Animal *obj) @{@dots{}@}
3635 void Check_Dog (Dog *obj) @{@dots{}@}
3638 void adainit (void);
3639 void adafinal (void);
3645 Dog *obj = new_dog(); // Ada constructor
3646 Check_Carnivore (obj); // Check secondary DT
3647 Check_Domestic (obj); // Check secondary DT
3648 Check_Animal (obj); // Check primary DT
3649 Check_Dog (obj); // Check primary DT
3654 adainit (); test(); adafinal ();
3659 @node Comparison between GNAT and C/C++ Compilation Models
3660 @section Comparison between GNAT and C/C++ Compilation Models
3663 The GNAT model of compilation is close to the C and C++ models. You can
3664 think of Ada specs as corresponding to header files in C. As in C, you
3665 don't need to compile specs; they are compiled when they are used. The
3666 Ada @code{with} is similar in effect to the @code{#include} of a C
3669 One notable difference is that, in Ada, you may compile specs separately
3670 to check them for semantic and syntactic accuracy. This is not always
3671 possible with C headers because they are fragments of programs that have
3672 less specific syntactic or semantic rules.
3674 The other major difference is the requirement for running the binder,
3675 which performs two important functions. First, it checks for
3676 consistency. In C or C++, the only defense against assembling
3677 inconsistent programs lies outside the compiler, in a makefile, for
3678 example. The binder satisfies the Ada requirement that it be impossible
3679 to construct an inconsistent program when the compiler is used in normal
3682 @cindex Elaboration order control
3683 The other important function of the binder is to deal with elaboration
3684 issues. There are also elaboration issues in C++ that are handled
3685 automatically. This automatic handling has the advantage of being
3686 simpler to use, but the C++ programmer has no control over elaboration.
3687 Where @code{gnatbind} might complain there was no valid order of
3688 elaboration, a C++ compiler would simply construct a program that
3689 malfunctioned at run time.
3692 @node Comparison between GNAT and Conventional Ada Library Models
3693 @section Comparison between GNAT and Conventional Ada Library Models
3696 This section is intended for Ada programmers who have
3697 used an Ada compiler implementing the traditional Ada library
3698 model, as described in the Ada Reference Manual.
3700 @cindex GNAT library
3701 In GNAT, there is no ``library'' in the normal sense. Instead, the set of
3702 source files themselves acts as the library. Compiling Ada programs does
3703 not generate any centralized information, but rather an object file and
3704 a ALI file, which are of interest only to the binder and linker.
3705 In a traditional system, the compiler reads information not only from
3706 the source file being compiled, but also from the centralized library.
3707 This means that the effect of a compilation depends on what has been
3708 previously compiled. In particular:
3712 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3713 to the version of the unit most recently compiled into the library.
3716 Inlining is effective only if the necessary body has already been
3717 compiled into the library.
3720 Compiling a unit may obsolete other units in the library.
3724 In GNAT, compiling one unit never affects the compilation of any other
3725 units because the compiler reads only source files. Only changes to source
3726 files can affect the results of a compilation. In particular:
3730 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3731 to the source version of the unit that is currently accessible to the
3736 Inlining requires the appropriate source files for the package or
3737 subprogram bodies to be available to the compiler. Inlining is always
3738 effective, independent of the order in which units are complied.
3741 Compiling a unit never affects any other compilations. The editing of
3742 sources may cause previous compilations to be out of date if they
3743 depended on the source file being modified.
3747 The most important result of these differences is that order of compilation
3748 is never significant in GNAT. There is no situation in which one is
3749 required to do one compilation before another. What shows up as order of
3750 compilation requirements in the traditional Ada library becomes, in
3751 GNAT, simple source dependencies; in other words, there is only a set
3752 of rules saying what source files must be present when a file is
3756 @node Placement of temporary files
3757 @section Placement of temporary files
3758 @cindex Temporary files (user control over placement)
3761 GNAT creates temporary files in the directory designated by the environment
3762 variable @env{TMPDIR}.
3763 (See the HP @emph{C RTL Reference Manual} on the function @code{getenv()}
3764 for detailed information on how environment variables are resolved.
3765 For most users the easiest way to make use of this feature is to simply
3766 define @env{TMPDIR} as a job level logical name).
3767 For example, if you wish to use a Ramdisk (assuming DECRAM is installed)
3768 for compiler temporary files, then you can include something like the
3769 following command in your @file{LOGIN.COM} file:
3772 $ define/job TMPDIR "/disk$scratchram/000000/temp/"
3776 If @env{TMPDIR} is not defined, then GNAT uses the directory designated by
3777 @env{TMP}; if @env{TMP} is not defined, then GNAT uses the directory
3778 designated by @env{TEMP}.
3779 If none of these environment variables are defined then GNAT uses the
3780 directory designated by the logical name @code{SYS$SCRATCH:}
3781 (by default the user's home directory). If all else fails
3782 GNAT uses the current directory for temporary files.
3785 @c *************************
3786 @node Compiling Using gcc
3787 @chapter Compiling Using @command{gcc}
3790 This chapter discusses how to compile Ada programs using the @command{gcc}
3791 command. It also describes the set of switches
3792 that can be used to control the behavior of the compiler.
3794 * Compiling Programs::
3795 * Switches for gcc::
3796 * Search Paths and the Run-Time Library (RTL)::
3797 * Order of Compilation Issues::
3801 @node Compiling Programs
3802 @section Compiling Programs
3805 The first step in creating an executable program is to compile the units
3806 of the program using the @command{gcc} command. You must compile the
3811 the body file (@file{.adb}) for a library level subprogram or generic
3815 the spec file (@file{.ads}) for a library level package or generic
3816 package that has no body
3819 the body file (@file{.adb}) for a library level package
3820 or generic package that has a body
3825 You need @emph{not} compile the following files
3830 the spec of a library unit which has a body
3837 because they are compiled as part of compiling related units. GNAT
3839 when the corresponding body is compiled, and subunits when the parent is
3842 @cindex cannot generate code
3843 If you attempt to compile any of these files, you will get one of the
3844 following error messages (where @var{fff} is the name of the file you compiled):
3847 cannot generate code for file @var{fff} (package spec)
3848 to check package spec, use -gnatc
3850 cannot generate code for file @var{fff} (missing subunits)
3851 to check parent unit, use -gnatc
3853 cannot generate code for file @var{fff} (subprogram spec)
3854 to check subprogram spec, use -gnatc
3856 cannot generate code for file @var{fff} (subunit)
3857 to check subunit, use -gnatc
3861 As indicated by the above error messages, if you want to submit
3862 one of these files to the compiler to check for correct semantics
3863 without generating code, then use the @option{-gnatc} switch.
3865 The basic command for compiling a file containing an Ada unit is
3868 @c $ gcc -c @ovar{switches} @file{file name}
3869 @c Expanding @ovar macro inline (explanation in macro def comments)
3870 $ gcc -c @r{[}@var{switches}@r{]} @file{file name}
3874 where @var{file name} is the name of the Ada file (usually
3876 @file{.ads} for a spec or @file{.adb} for a body).
3879 @option{-c} switch to tell @command{gcc} to compile, but not link, the file.
3881 The result of a successful compilation is an object file, which has the
3882 same name as the source file but an extension of @file{.o} and an Ada
3883 Library Information (ALI) file, which also has the same name as the
3884 source file, but with @file{.ali} as the extension. GNAT creates these
3885 two output files in the current directory, but you may specify a source
3886 file in any directory using an absolute or relative path specification
3887 containing the directory information.
3890 @command{gcc} is actually a driver program that looks at the extensions of
3891 the file arguments and loads the appropriate compiler. For example, the
3892 GNU C compiler is @file{cc1}, and the Ada compiler is @file{gnat1}.
3893 These programs are in directories known to the driver program (in some
3894 configurations via environment variables you set), but need not be in
3895 your path. The @command{gcc} driver also calls the assembler and any other
3896 utilities needed to complete the generation of the required object
3899 It is possible to supply several file names on the same @command{gcc}
3900 command. This causes @command{gcc} to call the appropriate compiler for
3901 each file. For example, the following command lists three separate
3902 files to be compiled:
3905 $ gcc -c x.adb y.adb z.c
3909 calls @code{gnat1} (the Ada compiler) twice to compile @file{x.adb} and
3910 @file{y.adb}, and @code{cc1} (the C compiler) once to compile @file{z.c}.
3911 The compiler generates three object files @file{x.o}, @file{y.o} and
3912 @file{z.o} and the two ALI files @file{x.ali} and @file{y.ali} from the
3913 Ada compilations. Any switches apply to all the files ^listed,^listed.^
3916 @option{-gnat@var{x}} switches, which apply only to Ada compilations.
3919 @node Switches for gcc
3920 @section Switches for @command{gcc}
3923 The @command{gcc} command accepts switches that control the
3924 compilation process. These switches are fully described in this section.
3925 First we briefly list all the switches, in alphabetical order, then we
3926 describe the switches in more detail in functionally grouped sections.
3928 More switches exist for GCC than those documented here, especially
3929 for specific targets. However, their use is not recommended as
3930 they may change code generation in ways that are incompatible with
3931 the Ada run-time library, or can cause inconsistencies between
3935 * Output and Error Message Control::
3936 * Warning Message Control::
3937 * Debugging and Assertion Control::
3938 * Validity Checking::
3941 * Using gcc for Syntax Checking::
3942 * Using gcc for Semantic Checking::
3943 * Compiling Different Versions of Ada::
3944 * Character Set Control::
3945 * File Naming Control::
3946 * Subprogram Inlining Control::
3947 * Auxiliary Output Control::
3948 * Debugging Control::
3949 * Exception Handling Control::
3950 * Units to Sources Mapping Files::
3951 * Integrated Preprocessing::
3952 * Code Generation Control::
3961 @cindex @option{-b} (@command{gcc})
3962 @item -b @var{target}
3963 Compile your program to run on @var{target}, which is the name of a
3964 system configuration. You must have a GNAT cross-compiler built if
3965 @var{target} is not the same as your host system.
3968 @cindex @option{-B} (@command{gcc})
3969 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
3970 from @var{dir} instead of the default location. Only use this switch
3971 when multiple versions of the GNAT compiler are available.
3972 @xref{Directory Options,, Options for Directory Search, gcc, Using the
3973 GNU Compiler Collection (GCC)}, for further details. You would normally
3974 use the @option{-b} or @option{-V} switch instead.
3977 @cindex @option{-c} (@command{gcc})
3978 Compile. Always use this switch when compiling Ada programs.
3980 Note: for some other languages when using @command{gcc}, notably in
3981 the case of C and C++, it is possible to use
3982 use @command{gcc} without a @option{-c} switch to
3983 compile and link in one step. In the case of GNAT, you
3984 cannot use this approach, because the binder must be run
3985 and @command{gcc} cannot be used to run the GNAT binder.
3989 @cindex @option{-fno-inline} (@command{gcc})
3990 Suppresses all back-end inlining, even if other optimization or inlining
3992 This includes suppression of inlining that results
3993 from the use of the pragma @code{Inline_Always}.
3994 Any occurrences of pragma @code{Inline} or @code{Inline_Always}
3995 are ignored, and @option{-gnatn} and @option{-gnatN} have no
3996 effect if this switch is present.
3998 @item -fno-inline-functions
3999 @cindex @option{-fno-inline-functions} (@command{gcc})
4000 Suppresses automatic inlining of simple subprograms, which is enabled
4001 if @option{-O3} is used.
4003 @item -fno-inline-small-functions
4004 @cindex @option{-fno-inline-small-functions} (@command{gcc})
4005 Suppresses automatic inlining of small subprograms, which is enabled
4006 if @option{-O2} is used.
4008 @item -fno-inline-functions-called-once
4009 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
4010 Suppresses inlining of subprograms local to the unit and called once
4011 from within it, which is enabled if @option{-O1} is used.
4014 @cindex @option{-fno-ivopts} (@command{gcc})
4015 Suppresses high-level loop induction variable optimizations, which are
4016 enabled if @option{-O1} is used. These optimizations are generally
4017 profitable but, for some specific cases of loops with numerous uses
4018 of the iteration variable that follow a common pattern, they may end
4019 up destroying the regularity that could be exploited at a lower level
4020 and thus producing inferior code.
4022 @item -fno-strict-aliasing
4023 @cindex @option{-fno-strict-aliasing} (@command{gcc})
4024 Causes the compiler to avoid assumptions regarding non-aliasing
4025 of objects of different types. See
4026 @ref{Optimization and Strict Aliasing} for details.
4029 @cindex @option{-fstack-check} (@command{gcc})
4030 Activates stack checking.
4031 See @ref{Stack Overflow Checking} for details.
4034 @cindex @option{-fstack-usage} (@command{gcc})
4035 Makes the compiler output stack usage information for the program, on a
4036 per-function basis. See @ref{Static Stack Usage Analysis} for details.
4038 @item -fcallgraph-info@r{[}=su@r{]}
4039 @cindex @option{-fcallgraph-info} (@command{gcc})
4040 Makes the compiler output callgraph information for the program, on a
4041 per-file basis. The information is generated in the VCG format. It can
4042 be decorated with stack-usage per-node information.
4045 @cindex @option{^-g^/DEBUG^} (@command{gcc})
4046 Generate debugging information. This information is stored in the object
4047 file and copied from there to the final executable file by the linker,
4048 where it can be read by the debugger. You must use the
4049 @option{^-g^/DEBUG^} switch if you plan on using the debugger.
4052 @cindex @option{-gnat83} (@command{gcc})
4053 Enforce Ada 83 restrictions.
4056 @cindex @option{-gnat95} (@command{gcc})
4057 Enforce Ada 95 restrictions.
4060 @cindex @option{-gnat05} (@command{gcc})
4061 Allow full Ada 2005 features.
4064 @cindex @option{-gnat2005} (@command{gcc})
4065 Allow full Ada 2005 features (same as @option{-gnat05}
4068 @cindex @option{-gnat12} (@command{gcc})
4071 @cindex @option{-gnat2012} (@command{gcc})
4072 Allow full Ada 2012 features (same as @option{-gnat12}
4075 @cindex @option{-gnata} (@command{gcc})
4076 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
4077 activated. Note that these pragmas can also be controlled using the
4078 configuration pragmas @code{Assertion_Policy} and @code{Debug_Policy}.
4079 It also activates pragmas @code{Check}, @code{Precondition}, and
4080 @code{Postcondition}. Note that these pragmas can also be controlled
4081 using the configuration pragma @code{Check_Policy}.
4084 @cindex @option{-gnatA} (@command{gcc})
4085 Avoid processing @file{gnat.adc}. If a @file{gnat.adc} file is present,
4089 @cindex @option{-gnatb} (@command{gcc})
4090 Generate brief messages to @file{stderr} even if verbose mode set.
4093 @cindex @option{-gnatB} (@command{gcc})
4094 Assume no invalid (bad) values except for 'Valid attribute use
4095 (@pxref{Validity Checking}).
4098 @cindex @option{-gnatc} (@command{gcc})
4099 Check syntax and semantics only (no code generation attempted).
4102 @cindex @option{-gnatC} (@command{gcc})
4103 Generate CodePeer information (no code generation attempted).
4104 This switch will generate an intermediate representation suitable for
4105 use by CodePeer (@file{.scil} files). This switch is not compatible with
4106 code generation (it will, among other things, disable some switches such
4107 as -gnatn, and enable others such as -gnata).
4110 @cindex @option{-gnatd} (@command{gcc})
4111 Specify debug options for the compiler. The string of characters after
4112 the @option{-gnatd} specify the specific debug options. The possible
4113 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
4114 compiler source file @file{debug.adb} for details of the implemented
4115 debug options. Certain debug options are relevant to applications
4116 programmers, and these are documented at appropriate points in this
4121 @cindex @option{-gnatD[nn]} (@command{gcc})
4124 @item /XDEBUG /LXDEBUG=nnn
4126 Create expanded source files for source level debugging. This switch
4127 also suppress generation of cross-reference information
4128 (see @option{-gnatx}).
4130 @item -gnatec=@var{path}
4131 @cindex @option{-gnatec} (@command{gcc})
4132 Specify a configuration pragma file
4134 (the equal sign is optional)
4136 (@pxref{The Configuration Pragmas Files}).
4138 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=@var{value}@r{]}
4139 @cindex @option{-gnateD} (@command{gcc})
4140 Defines a symbol, associated with @var{value}, for preprocessing.
4141 (@pxref{Integrated Preprocessing}).
4144 @cindex @option{-gnatef} (@command{gcc})
4145 Display full source path name in brief error messages.
4148 @cindex @option{-gnateG} (@command{gcc})
4149 Save result of preprocessing in a text file.
4151 @item -gnatem=@var{path}
4152 @cindex @option{-gnatem} (@command{gcc})
4153 Specify a mapping file
4155 (the equal sign is optional)
4157 (@pxref{Units to Sources Mapping Files}).
4159 @item -gnatep=@var{file}
4160 @cindex @option{-gnatep} (@command{gcc})
4161 Specify a preprocessing data file
4163 (the equal sign is optional)
4165 (@pxref{Integrated Preprocessing}).
4168 @cindex @option{-gnateS} (@command{gcc})
4169 Generate SCO (Source Coverage Obligation) information in the ALI
4170 file. This information is used by advanced coverage tools. See
4171 unit @file{SCOs} in the compiler sources for details in files
4172 @file{scos.ads} and @file{scos.adb}.
4175 @cindex @option{-gnatE} (@command{gcc})
4176 Full dynamic elaboration checks.
4179 @cindex @option{-gnatf} (@command{gcc})
4180 Full errors. Multiple errors per line, all undefined references, do not
4181 attempt to suppress cascaded errors.
4184 @cindex @option{-gnatF} (@command{gcc})
4185 Externals names are folded to all uppercase.
4187 @item ^-gnatg^/GNAT_INTERNAL^
4188 @cindex @option{^-gnatg^/GNAT_INTERNAL^} (@command{gcc})
4189 Internal GNAT implementation mode. This should not be used for
4190 applications programs, it is intended only for use by the compiler
4191 and its run-time library. For documentation, see the GNAT sources.
4192 Note that @option{^-gnatg^/GNAT_INTERNAL^} implies
4193 @option{^-gnatwae^/WARNINGS=ALL,ERRORS^} and
4194 @option{^-gnatyg^/STYLE_CHECKS=GNAT^}
4195 so that all standard warnings and all standard style options are turned on.
4196 All warnings and style messages are treated as errors.
4200 @cindex @option{-gnatG[nn]} (@command{gcc})
4203 @item /EXPAND_SOURCE, /LEXPAND_SOURCE=nnn
4205 List generated expanded code in source form.
4207 @item ^-gnath^/HELP^
4208 @cindex @option{^-gnath^/HELP^} (@command{gcc})
4209 Output usage information. The output is written to @file{stdout}.
4211 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
4212 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
4213 Identifier character set
4215 (@var{c}=1/2/3/4/8/9/p/f/n/w).
4217 For details of the possible selections for @var{c},
4218 see @ref{Character Set Control}.
4220 @item ^-gnatI^/IGNORE_REP_CLAUSES^
4221 @cindex @option{^-gnatI^IGNORE_REP_CLAUSES^} (@command{gcc})
4222 Ignore representation clauses. When this switch is used,
4223 representation clauses are treated as comments. This is useful
4224 when initially porting code where you want to ignore rep clause
4225 problems, and also for compiling foreign code (particularly
4226 for use with ASIS). The representation clauses that are ignored
4227 are: enumeration_representation_clause, record_representation_clause,
4228 and attribute_definition_clause for the following attributes:
4229 Address, Alignment, Bit_Order, Component_Size, Machine_Radix,
4230 Object_Size, Size, Small, Stream_Size, and Value_Size.
4231 Note that this option should be used only for compiling -- the
4232 code is likely to malfunction at run time.
4235 @cindex @option{-gnatjnn} (@command{gcc})
4236 Reformat error messages to fit on nn character lines
4238 @item -gnatk=@var{n}
4239 @cindex @option{-gnatk} (@command{gcc})
4240 Limit file names to @var{n} (1-999) characters ^(@code{k} = krunch)^^.
4243 @cindex @option{-gnatl} (@command{gcc})
4244 Output full source listing with embedded error messages.
4247 @cindex @option{-gnatL} (@command{gcc})
4248 Used in conjunction with -gnatG or -gnatD to intersperse original
4249 source lines (as comment lines with line numbers) in the expanded
4252 @item -gnatm=@var{n}
4253 @cindex @option{-gnatm} (@command{gcc})
4254 Limit number of detected error or warning messages to @var{n}
4255 where @var{n} is in the range 1..999999. The default setting if
4256 no switch is given is 9999. If the number of warnings reaches this
4257 limit, then a message is output and further warnings are suppressed,
4258 but the compilation is continued. If the number of error messages
4259 reaches this limit, then a message is output and the compilation
4260 is abandoned. The equal sign here is optional. A value of zero
4261 means that no limit applies.
4264 @cindex @option{-gnatn} (@command{gcc})
4265 Activate inlining for subprograms for which
4266 pragma @code{inline} is specified. This inlining is performed
4267 by the GCC back-end.
4270 @cindex @option{-gnatN} (@command{gcc})
4271 Activate front end inlining for subprograms for which
4272 pragma @code{Inline} is specified. This inlining is performed
4273 by the front end and will be visible in the
4274 @option{-gnatG} output.
4276 When using a gcc-based back end (in practice this means using any version
4277 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
4278 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
4279 Historically front end inlining was more extensive than the gcc back end
4280 inlining, but that is no longer the case.
4283 @cindex @option{-gnato} (@command{gcc})
4284 Enable numeric overflow checking (which is not normally enabled by
4285 default). Note that division by zero is a separate check that is not
4286 controlled by this switch (division by zero checking is on by default).
4289 @cindex @option{-gnatp} (@command{gcc})
4290 Suppress all checks. See @ref{Run-Time Checks} for details. This switch
4291 has no effect if cancelled by a subsequent @option{-gnat-p} switch.
4294 @cindex @option{-gnat-p} (@command{gcc})
4295 Cancel effect of previous @option{-gnatp} switch.
4298 @cindex @option{-gnatP} (@command{gcc})
4299 Enable polling. This is required on some systems (notably Windows NT) to
4300 obtain asynchronous abort and asynchronous transfer of control capability.
4301 @xref{Pragma Polling,,, gnat_rm, GNAT Reference Manual}, for full
4305 @cindex @option{-gnatq} (@command{gcc})
4306 Don't quit. Try semantics, even if parse errors.
4309 @cindex @option{-gnatQ} (@command{gcc})
4310 Don't quit. Generate @file{ALI} and tree files even if illegalities.
4313 @cindex @option{-gnatr} (@command{gcc})
4314 Treat pragma Restrictions as Restriction_Warnings.
4316 @item ^-gnatR@r{[}0@r{/}1@r{/}2@r{/}3@r{[}s@r{]]}^/REPRESENTATION_INFO^
4317 @cindex @option{-gnatR} (@command{gcc})
4318 Output representation information for declared types and objects.
4321 @cindex @option{-gnats} (@command{gcc})
4325 @cindex @option{-gnatS} (@command{gcc})
4326 Print package Standard.
4329 @cindex @option{-gnatt} (@command{gcc})
4330 Generate tree output file.
4332 @item ^-gnatT^/TABLE_MULTIPLIER=^@var{nnn}
4333 @cindex @option{^-gnatT^/TABLE_MULTIPLIER^} (@command{gcc})
4334 All compiler tables start at @var{nnn} times usual starting size.
4337 @cindex @option{-gnatu} (@command{gcc})
4338 List units for this compilation.
4341 @cindex @option{-gnatU} (@command{gcc})
4342 Tag all error messages with the unique string ``error:''
4345 @cindex @option{-gnatv} (@command{gcc})
4346 Verbose mode. Full error output with source lines to @file{stdout}.
4349 @cindex @option{-gnatV} (@command{gcc})
4350 Control level of validity checking (@pxref{Validity Checking}).
4352 @item ^-gnatw@var{xxx}^/WARNINGS=(@var{option}@r{[},@dots{}@r{]})^
4353 @cindex @option{^-gnatw^/WARNINGS^} (@command{gcc})
4355 ^@var{xxx} is a string of option letters that^the list of options^ denotes
4356 the exact warnings that
4357 are enabled or disabled (@pxref{Warning Message Control}).
4359 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
4360 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
4361 Wide character encoding method
4363 (@var{e}=n/h/u/s/e/8).
4366 (@var{e}=@code{BRACKETS, NONE, HEX, UPPER, SHIFT_JIS, EUC, UTF8})
4370 @cindex @option{-gnatx} (@command{gcc})
4371 Suppress generation of cross-reference information.
4374 @cindex @option{-gnatX} (@command{gcc})
4375 Enable GNAT implementation extensions and latest Ada version.
4377 @item ^-gnaty^/STYLE_CHECKS=(option,option@dots{})^
4378 @cindex @option{^-gnaty^/STYLE_CHECKS^} (@command{gcc})
4379 Enable built-in style checks (@pxref{Style Checking}).
4381 @item ^-gnatz^/DISTRIBUTION_STUBS=^@var{m}
4382 @cindex @option{^-gnatz^/DISTRIBUTION_STUBS^} (@command{gcc})
4383 Distribution stub generation and compilation
4385 (@var{m}=r/c for receiver/caller stubs).
4388 (@var{m}=@code{RECEIVER} or @code{CALLER} to specify the type of stubs
4389 to be generated and compiled).
4392 @item ^-I^/SEARCH=^@var{dir}
4393 @cindex @option{^-I^/SEARCH^} (@command{gcc})
4395 Direct GNAT to search the @var{dir} directory for source files needed by
4396 the current compilation
4397 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4399 @item ^-I-^/NOCURRENT_DIRECTORY^
4400 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gcc})
4402 Except for the source file named in the command line, do not look for source
4403 files in the directory containing the source file named in the command line
4404 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4408 @cindex @option{-mbig-switch} (@command{gcc})
4409 @cindex @code{case} statement (effect of @option{-mbig-switch} option)
4410 This standard gcc switch causes the compiler to use larger offsets in its
4411 jump table representation for @code{case} statements.
4412 This may result in less efficient code, but is sometimes necessary
4413 (for example on HP-UX targets)
4414 @cindex HP-UX and @option{-mbig-switch} option
4415 in order to compile large and/or nested @code{case} statements.
4418 @cindex @option{-o} (@command{gcc})
4419 This switch is used in @command{gcc} to redirect the generated object file
4420 and its associated ALI file. Beware of this switch with GNAT, because it may
4421 cause the object file and ALI file to have different names which in turn
4422 may confuse the binder and the linker.
4426 @cindex @option{-nostdinc} (@command{gcc})
4427 Inhibit the search of the default location for the GNAT Run Time
4428 Library (RTL) source files.
4431 @cindex @option{-nostdlib} (@command{gcc})
4432 Inhibit the search of the default location for the GNAT Run Time
4433 Library (RTL) ALI files.
4437 @c Expanding @ovar macro inline (explanation in macro def comments)
4438 @item -O@r{[}@var{n}@r{]}
4439 @cindex @option{-O} (@command{gcc})
4440 @var{n} controls the optimization level.
4444 No optimization, the default setting if no @option{-O} appears
4447 Normal optimization, the default if you specify @option{-O} without
4448 an operand. A good compromise between code quality and compilation
4452 Extensive optimization, may improve execution time, possibly at the cost of
4453 substantially increased compilation time.
4456 Same as @option{-O2}, and also includes inline expansion for small subprograms
4460 Optimize space usage
4464 See also @ref{Optimization Levels}.
4469 @cindex @option{/NOOPTIMIZE} (@code{GNAT COMPILE})
4470 Equivalent to @option{/OPTIMIZE=NONE}.
4471 This is the default behavior in the absence of an @option{/OPTIMIZE}
4474 @item /OPTIMIZE@r{[}=(keyword@r{[},@dots{}@r{]})@r{]}
4475 @cindex @option{/OPTIMIZE} (@code{GNAT COMPILE})
4476 Selects the level of optimization for your program. The supported
4477 keywords are as follows:
4480 Perform most optimizations, including those that
4482 This is the default if the @option{/OPTIMIZE} qualifier is supplied
4483 without keyword options.
4486 Do not do any optimizations. Same as @code{/NOOPTIMIZE}.
4489 Perform some optimizations, but omit ones that are costly.
4492 Same as @code{SOME}.
4495 Full optimization as in @option{/OPTIMIZE=ALL}, and also attempts
4496 automatic inlining of small subprograms within a unit
4499 Try to unroll loops. This keyword may be specified together with
4500 any keyword above other than @code{NONE}. Loop unrolling
4501 usually, but not always, improves the performance of programs.
4504 Optimize space usage
4508 See also @ref{Optimization Levels}.
4512 @item -pass-exit-codes
4513 @cindex @option{-pass-exit-codes} (@command{gcc})
4514 Catch exit codes from the compiler and use the most meaningful as
4518 @item --RTS=@var{rts-path}
4519 @cindex @option{--RTS} (@command{gcc})
4520 Specifies the default location of the runtime library. Same meaning as the
4521 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
4524 @cindex @option{^-S^/ASM^} (@command{gcc})
4525 ^Used in place of @option{-c} to^Used to^
4526 cause the assembler source file to be
4527 generated, using @file{^.s^.S^} as the extension,
4528 instead of the object file.
4529 This may be useful if you need to examine the generated assembly code.
4531 @item ^-fverbose-asm^/VERBOSE_ASM^
4532 @cindex @option{^-fverbose-asm^/VERBOSE_ASM^} (@command{gcc})
4533 ^Used in conjunction with @option{-S}^Used in place of @option{/ASM}^
4534 to cause the generated assembly code file to be annotated with variable
4535 names, making it significantly easier to follow.
4538 @cindex @option{^-v^/VERBOSE^} (@command{gcc})
4539 Show commands generated by the @command{gcc} driver. Normally used only for
4540 debugging purposes or if you need to be sure what version of the
4541 compiler you are executing.
4545 @cindex @option{-V} (@command{gcc})
4546 Execute @var{ver} version of the compiler. This is the @command{gcc}
4547 version, not the GNAT version.
4550 @item ^-w^/NO_BACK_END_WARNINGS^
4551 @cindex @option{-w} (@command{gcc})
4552 Turn off warnings generated by the back end of the compiler. Use of
4553 this switch also causes the default for front end warnings to be set
4554 to suppress (as though @option{-gnatws} had appeared at the start of
4560 @c Combining qualifiers does not work on VMS
4561 You may combine a sequence of GNAT switches into a single switch. For
4562 example, the combined switch
4564 @cindex Combining GNAT switches
4570 is equivalent to specifying the following sequence of switches:
4573 -gnato -gnatf -gnati3
4578 The following restrictions apply to the combination of switches
4583 The switch @option{-gnatc} if combined with other switches must come
4584 first in the string.
4587 The switch @option{-gnats} if combined with other switches must come
4588 first in the string.
4592 ^^@option{/DISTRIBUTION_STUBS=},^
4593 @option{-gnatzc} and @option{-gnatzr} may not be combined with any other
4594 switches, and only one of them may appear in the command line.
4597 The switch @option{-gnat-p} may not be combined with any other switch.
4601 Once a ``y'' appears in the string (that is a use of the @option{-gnaty}
4602 switch), then all further characters in the switch are interpreted
4603 as style modifiers (see description of @option{-gnaty}).
4606 Once a ``d'' appears in the string (that is a use of the @option{-gnatd}
4607 switch), then all further characters in the switch are interpreted
4608 as debug flags (see description of @option{-gnatd}).
4611 Once a ``w'' appears in the string (that is a use of the @option{-gnatw}
4612 switch), then all further characters in the switch are interpreted
4613 as warning mode modifiers (see description of @option{-gnatw}).
4616 Once a ``V'' appears in the string (that is a use of the @option{-gnatV}
4617 switch), then all further characters in the switch are interpreted
4618 as validity checking options (@pxref{Validity Checking}).
4621 Option ``em'', ``ec'', ``ep'', ``l='' and ``R'' must be the last options in
4622 a combined list of options.
4626 @node Output and Error Message Control
4627 @subsection Output and Error Message Control
4631 The standard default format for error messages is called ``brief format''.
4632 Brief format messages are written to @file{stderr} (the standard error
4633 file) and have the following form:
4636 e.adb:3:04: Incorrect spelling of keyword "function"
4637 e.adb:4:20: ";" should be "is"
4641 The first integer after the file name is the line number in the file,
4642 and the second integer is the column number within the line.
4644 @code{GPS} can parse the error messages
4645 and point to the referenced character.
4647 The following switches provide control over the error message
4653 @cindex @option{-gnatv} (@command{gcc})
4656 The v stands for verbose.
4658 The effect of this setting is to write long-format error
4659 messages to @file{stdout} (the standard output file.
4660 The same program compiled with the
4661 @option{-gnatv} switch would generate:
4665 3. funcion X (Q : Integer)
4667 >>> Incorrect spelling of keyword "function"
4670 >>> ";" should be "is"
4675 The vertical bar indicates the location of the error, and the @samp{>>>}
4676 prefix can be used to search for error messages. When this switch is
4677 used the only source lines output are those with errors.
4680 @cindex @option{-gnatl} (@command{gcc})
4682 The @code{l} stands for list.
4684 This switch causes a full listing of
4685 the file to be generated. In the case where a body is
4686 compiled, the corresponding spec is also listed, along
4687 with any subunits. Typical output from compiling a package
4688 body @file{p.adb} might look like:
4690 @smallexample @c ada
4694 1. package body p is
4696 3. procedure a is separate;
4707 2. pragma Elaborate_Body
4731 When you specify the @option{-gnatv} or @option{-gnatl} switches and
4732 standard output is redirected, a brief summary is written to
4733 @file{stderr} (standard error) giving the number of error messages and
4734 warning messages generated.
4736 @item -^gnatl^OUTPUT_FILE^=file
4737 @cindex @option{^-gnatl^OUTPUT_FILE^=fname} (@command{gcc})
4738 This has the same effect as @option{-gnatl} except that the output is
4739 written to a file instead of to standard output. If the given name
4740 @file{fname} does not start with a period, then it is the full name
4741 of the file to be written. If @file{fname} is an extension, it is
4742 appended to the name of the file being compiled. For example, if
4743 file @file{xyz.adb} is compiled with @option{^-gnatl^OUTPUT_FILE^=.lst},
4744 then the output is written to file ^xyz.adb.lst^xyz.adb_lst^.
4747 @cindex @option{-gnatU} (@command{gcc})
4748 This switch forces all error messages to be preceded by the unique
4749 string ``error:''. This means that error messages take a few more
4750 characters in space, but allows easy searching for and identification
4754 @cindex @option{-gnatb} (@command{gcc})
4756 The @code{b} stands for brief.
4758 This switch causes GNAT to generate the
4759 brief format error messages to @file{stderr} (the standard error
4760 file) as well as the verbose
4761 format message or full listing (which as usual is written to
4762 @file{stdout} (the standard output file).
4764 @item -gnatm=@var{n}
4765 @cindex @option{-gnatm} (@command{gcc})
4767 The @code{m} stands for maximum.
4769 @var{n} is a decimal integer in the
4770 range of 1 to 999999 and limits the number of error or warning
4771 messages to be generated. For example, using
4772 @option{-gnatm2} might yield
4775 e.adb:3:04: Incorrect spelling of keyword "function"
4776 e.adb:5:35: missing ".."
4777 fatal error: maximum number of errors detected
4778 compilation abandoned
4782 The default setting if
4783 no switch is given is 9999. If the number of warnings reaches this
4784 limit, then a message is output and further warnings are suppressed,
4785 but the compilation is continued. If the number of error messages
4786 reaches this limit, then a message is output and the compilation
4787 is abandoned. A value of zero means that no limit applies.
4790 Note that the equal sign is optional, so the switches
4791 @option{-gnatm2} and @option{-gnatm=2} are equivalent.
4794 @cindex @option{-gnatf} (@command{gcc})
4795 @cindex Error messages, suppressing
4797 The @code{f} stands for full.
4799 Normally, the compiler suppresses error messages that are likely to be
4800 redundant. This switch causes all error
4801 messages to be generated. In particular, in the case of
4802 references to undefined variables. If a given variable is referenced
4803 several times, the normal format of messages is
4805 e.adb:7:07: "V" is undefined (more references follow)
4809 where the parenthetical comment warns that there are additional
4810 references to the variable @code{V}. Compiling the same program with the
4811 @option{-gnatf} switch yields
4814 e.adb:7:07: "V" is undefined
4815 e.adb:8:07: "V" is undefined
4816 e.adb:8:12: "V" is undefined
4817 e.adb:8:16: "V" is undefined
4818 e.adb:9:07: "V" is undefined
4819 e.adb:9:12: "V" is undefined
4823 The @option{-gnatf} switch also generates additional information for
4824 some error messages. Some examples are:
4828 Details on possibly non-portable unchecked conversion
4830 List possible interpretations for ambiguous calls
4832 Additional details on incorrect parameters
4836 @cindex @option{-gnatjnn} (@command{gcc})
4837 In normal operation mode (or if @option{-gnatj0} is used, then error messages
4838 with continuation lines are treated as though the continuation lines were
4839 separate messages (and so a warning with two continuation lines counts as
4840 three warnings, and is listed as three separate messages).
4842 If the @option{-gnatjnn} switch is used with a positive value for nn, then
4843 messages are output in a different manner. A message and all its continuation
4844 lines are treated as a unit, and count as only one warning or message in the
4845 statistics totals. Furthermore, the message is reformatted so that no line
4846 is longer than nn characters.
4849 @cindex @option{-gnatq} (@command{gcc})
4851 The @code{q} stands for quit (really ``don't quit'').
4853 In normal operation mode, the compiler first parses the program and
4854 determines if there are any syntax errors. If there are, appropriate
4855 error messages are generated and compilation is immediately terminated.
4857 GNAT to continue with semantic analysis even if syntax errors have been
4858 found. This may enable the detection of more errors in a single run. On
4859 the other hand, the semantic analyzer is more likely to encounter some
4860 internal fatal error when given a syntactically invalid tree.
4863 @cindex @option{-gnatQ} (@command{gcc})
4864 In normal operation mode, the @file{ALI} file is not generated if any
4865 illegalities are detected in the program. The use of @option{-gnatQ} forces
4866 generation of the @file{ALI} file. This file is marked as being in
4867 error, so it cannot be used for binding purposes, but it does contain
4868 reasonably complete cross-reference information, and thus may be useful
4869 for use by tools (e.g., semantic browsing tools or integrated development
4870 environments) that are driven from the @file{ALI} file. This switch
4871 implies @option{-gnatq}, since the semantic phase must be run to get a
4872 meaningful ALI file.
4874 In addition, if @option{-gnatt} is also specified, then the tree file is
4875 generated even if there are illegalities. It may be useful in this case
4876 to also specify @option{-gnatq} to ensure that full semantic processing
4877 occurs. The resulting tree file can be processed by ASIS, for the purpose
4878 of providing partial information about illegal units, but if the error
4879 causes the tree to be badly malformed, then ASIS may crash during the
4882 When @option{-gnatQ} is used and the generated @file{ALI} file is marked as
4883 being in error, @command{gnatmake} will attempt to recompile the source when it
4884 finds such an @file{ALI} file, including with switch @option{-gnatc}.
4886 Note that @option{-gnatQ} has no effect if @option{-gnats} is specified,
4887 since ALI files are never generated if @option{-gnats} is set.
4891 @node Warning Message Control
4892 @subsection Warning Message Control
4893 @cindex Warning messages
4895 In addition to error messages, which correspond to illegalities as defined
4896 in the Ada Reference Manual, the compiler detects two kinds of warning
4899 First, the compiler considers some constructs suspicious and generates a
4900 warning message to alert you to a possible error. Second, if the
4901 compiler detects a situation that is sure to raise an exception at
4902 run time, it generates a warning message. The following shows an example
4903 of warning messages:
4905 e.adb:4:24: warning: creation of object may raise Storage_Error
4906 e.adb:10:17: warning: static value out of range
4907 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
4911 GNAT considers a large number of situations as appropriate
4912 for the generation of warning messages. As always, warnings are not
4913 definite indications of errors. For example, if you do an out-of-range
4914 assignment with the deliberate intention of raising a
4915 @code{Constraint_Error} exception, then the warning that may be
4916 issued does not indicate an error. Some of the situations for which GNAT
4917 issues warnings (at least some of the time) are given in the following
4918 list. This list is not complete, and new warnings are often added to
4919 subsequent versions of GNAT. The list is intended to give a general idea
4920 of the kinds of warnings that are generated.
4924 Possible infinitely recursive calls
4927 Out-of-range values being assigned
4930 Possible order of elaboration problems
4933 Assertions (pragma Assert) that are sure to fail
4939 Address clauses with possibly unaligned values, or where an attempt is
4940 made to overlay a smaller variable with a larger one.
4943 Fixed-point type declarations with a null range
4946 Direct_IO or Sequential_IO instantiated with a type that has access values
4949 Variables that are never assigned a value
4952 Variables that are referenced before being initialized
4955 Task entries with no corresponding @code{accept} statement
4958 Duplicate accepts for the same task entry in a @code{select}
4961 Objects that take too much storage
4964 Unchecked conversion between types of differing sizes
4967 Missing @code{return} statement along some execution path in a function
4970 Incorrect (unrecognized) pragmas
4973 Incorrect external names
4976 Allocation from empty storage pool
4979 Potentially blocking operation in protected type
4982 Suspicious parenthesization of expressions
4985 Mismatching bounds in an aggregate
4988 Attempt to return local value by reference
4991 Premature instantiation of a generic body
4994 Attempt to pack aliased components
4997 Out of bounds array subscripts
5000 Wrong length on string assignment
5003 Violations of style rules if style checking is enabled
5006 Unused @code{with} clauses
5009 @code{Bit_Order} usage that does not have any effect
5012 @code{Standard.Duration} used to resolve universal fixed expression
5015 Dereference of possibly null value
5018 Declaration that is likely to cause storage error
5021 Internal GNAT unit @code{with}'ed by application unit
5024 Values known to be out of range at compile time
5027 Unreferenced labels and variables
5030 Address overlays that could clobber memory
5033 Unexpected initialization when address clause present
5036 Bad alignment for address clause
5039 Useless type conversions
5042 Redundant assignment statements and other redundant constructs
5045 Useless exception handlers
5048 Accidental hiding of name by child unit
5051 Access before elaboration detected at compile time
5054 A range in a @code{for} loop that is known to be null or might be null
5059 The following section lists compiler switches that are available
5060 to control the handling of warning messages. It is also possible
5061 to exercise much finer control over what warnings are issued and
5062 suppressed using the GNAT pragma Warnings, @xref{Pragma Warnings,,,
5063 gnat_rm, GNAT Reference manual}.
5068 @emph{Activate all optional errors.}
5069 @cindex @option{-gnatwa} (@command{gcc})
5070 This switch activates most optional warning messages, see remaining list
5071 in this section for details on optional warning messages that can be
5072 individually controlled. The warnings that are not turned on by this
5074 @option{-gnatwd} (implicit dereferencing),
5075 @option{-gnatwh} (hiding),
5076 @option{-gnatwl} (elaboration warnings),
5077 @option{-gnatw.o} (warn on values set by out parameters ignored)
5078 and @option{-gnatwt} (tracking of deleted conditional code).
5079 All other optional warnings are turned on.
5082 @emph{Suppress all optional errors.}
5083 @cindex @option{-gnatwA} (@command{gcc})
5084 This switch suppresses all optional warning messages, see remaining list
5085 in this section for details on optional warning messages that can be
5086 individually controlled.
5089 @emph{Activate warnings on failing assertions.}
5090 @cindex @option{-gnatw.a} (@command{gcc})
5091 @cindex Assert failures
5092 This switch activates warnings for assertions where the compiler can tell at
5093 compile time that the assertion will fail. Note that this warning is given
5094 even if assertions are disabled. The default is that such warnings are
5098 @emph{Suppress warnings on failing assertions.}
5099 @cindex @option{-gnatw.A} (@command{gcc})
5100 @cindex Assert failures
5101 This switch suppresses warnings for assertions where the compiler can tell at
5102 compile time that the assertion will fail.
5105 @emph{Activate warnings on bad fixed values.}
5106 @cindex @option{-gnatwb} (@command{gcc})
5107 @cindex Bad fixed values
5108 @cindex Fixed-point Small value
5110 This switch activates warnings for static fixed-point expressions whose
5111 value is not an exact multiple of Small. Such values are implementation
5112 dependent, since an implementation is free to choose either of the multiples
5113 that surround the value. GNAT always chooses the closer one, but this is not
5114 required behavior, and it is better to specify a value that is an exact
5115 multiple, ensuring predictable execution. The default is that such warnings
5119 @emph{Suppress warnings on bad fixed values.}
5120 @cindex @option{-gnatwB} (@command{gcc})
5121 This switch suppresses warnings for static fixed-point expressions whose
5122 value is not an exact multiple of Small.
5125 @emph{Activate warnings on biased representation.}
5126 @cindex @option{-gnatw.b} (@command{gcc})
5127 @cindex Biased representation
5128 This switch activates warnings when a size clause, value size clause, component
5129 clause, or component size clause forces the use of biased representation for an
5130 integer type (e.g. representing a range of 10..11 in a single bit by using 0/1
5131 to represent 10/11). The default is that such warnings are generated.
5134 @emph{Suppress warnings on biased representation.}
5135 @cindex @option{-gnatwB} (@command{gcc})
5136 This switch suppresses warnings for representation clauses that force the use
5137 of biased representation.
5140 @emph{Activate warnings on conditionals.}
5141 @cindex @option{-gnatwc} (@command{gcc})
5142 @cindex Conditionals, constant
5143 This switch activates warnings for conditional expressions used in
5144 tests that are known to be True or False at compile time. The default
5145 is that such warnings are not generated.
5146 Note that this warning does
5147 not get issued for the use of boolean variables or constants whose
5148 values are known at compile time, since this is a standard technique
5149 for conditional compilation in Ada, and this would generate too many
5150 false positive warnings.
5152 This warning option also activates a special test for comparisons using
5153 the operators ``>='' and`` <=''.
5154 If the compiler can tell that only the equality condition is possible,
5155 then it will warn that the ``>'' or ``<'' part of the test
5156 is useless and that the operator could be replaced by ``=''.
5157 An example would be comparing a @code{Natural} variable <= 0.
5159 This warning option also generates warnings if
5160 one or both tests is optimized away in a membership test for integer
5161 values if the result can be determined at compile time. Range tests on
5162 enumeration types are not included, since it is common for such tests
5163 to include an end point.
5165 This warning can also be turned on using @option{-gnatwa}.
5168 @emph{Suppress warnings on conditionals.}
5169 @cindex @option{-gnatwC} (@command{gcc})
5170 This switch suppresses warnings for conditional expressions used in
5171 tests that are known to be True or False at compile time.
5174 @emph{Activate warnings on missing component clauses.}
5175 @cindex @option{-gnatw.c} (@command{gcc})
5176 @cindex Component clause, missing
5177 This switch activates warnings for record components where a record
5178 representation clause is present and has component clauses for the
5179 majority, but not all, of the components. A warning is given for each
5180 component for which no component clause is present.
5182 This warning can also be turned on using @option{-gnatwa}.
5185 @emph{Suppress warnings on missing component clauses.}
5186 @cindex @option{-gnatwC} (@command{gcc})
5187 This switch suppresses warnings for record components that are
5188 missing a component clause in the situation described above.
5191 @emph{Activate warnings on implicit dereferencing.}
5192 @cindex @option{-gnatwd} (@command{gcc})
5193 If this switch is set, then the use of a prefix of an access type
5194 in an indexed component, slice, or selected component without an
5195 explicit @code{.all} will generate a warning. With this warning
5196 enabled, access checks occur only at points where an explicit
5197 @code{.all} appears in the source code (assuming no warnings are
5198 generated as a result of this switch). The default is that such
5199 warnings are not generated.
5200 Note that @option{-gnatwa} does not affect the setting of
5201 this warning option.
5204 @emph{Suppress warnings on implicit dereferencing.}
5205 @cindex @option{-gnatwD} (@command{gcc})
5206 @cindex Implicit dereferencing
5207 @cindex Dereferencing, implicit
5208 This switch suppresses warnings for implicit dereferences in
5209 indexed components, slices, and selected components.
5212 @emph{Treat warnings and style checks as errors.}
5213 @cindex @option{-gnatwe} (@command{gcc})
5214 @cindex Warnings, treat as error
5215 This switch causes warning messages and style check messages to be
5217 The warning string still appears, but the warning messages are counted
5218 as errors, and prevent the generation of an object file. Note that this
5219 is the only -gnatw switch that affects the handling of style check messages.
5222 @emph{Activate every optional warning}
5223 @cindex @option{-gnatw.e} (@command{gcc})
5224 @cindex Warnings, activate every optional warning
5225 This switch activates all optional warnings, including those which
5226 are not activated by @code{-gnatwa}.
5229 @emph{Activate warnings on unreferenced formals.}
5230 @cindex @option{-gnatwf} (@command{gcc})
5231 @cindex Formals, unreferenced
5232 This switch causes a warning to be generated if a formal parameter
5233 is not referenced in the body of the subprogram. This warning can
5234 also be turned on using @option{-gnatwa} or @option{-gnatwu}. The
5235 default is that these warnings are not generated.
5238 @emph{Suppress warnings on unreferenced formals.}
5239 @cindex @option{-gnatwF} (@command{gcc})
5240 This switch suppresses warnings for unreferenced formal
5241 parameters. Note that the
5242 combination @option{-gnatwu} followed by @option{-gnatwF} has the
5243 effect of warning on unreferenced entities other than subprogram
5247 @emph{Activate warnings on unrecognized pragmas.}
5248 @cindex @option{-gnatwg} (@command{gcc})
5249 @cindex Pragmas, unrecognized
5250 This switch causes a warning to be generated if an unrecognized
5251 pragma is encountered. Apart from issuing this warning, the
5252 pragma is ignored and has no effect. This warning can
5253 also be turned on using @option{-gnatwa}. The default
5254 is that such warnings are issued (satisfying the Ada Reference
5255 Manual requirement that such warnings appear).
5258 @emph{Suppress warnings on unrecognized pragmas.}
5259 @cindex @option{-gnatwG} (@command{gcc})
5260 This switch suppresses warnings for unrecognized pragmas.
5263 @emph{Activate warnings on hiding.}
5264 @cindex @option{-gnatwh} (@command{gcc})
5265 @cindex Hiding of Declarations
5266 This switch activates warnings on hiding declarations.
5267 A declaration is considered hiding
5268 if it is for a non-overloadable entity, and it declares an entity with the
5269 same name as some other entity that is directly or use-visible. The default
5270 is that such warnings are not generated.
5271 Note that @option{-gnatwa} does not affect the setting of this warning option.
5274 @emph{Suppress warnings on hiding.}
5275 @cindex @option{-gnatwH} (@command{gcc})
5276 This switch suppresses warnings on hiding declarations.
5279 @emph{Activate warnings on implementation units.}
5280 @cindex @option{-gnatwi} (@command{gcc})
5281 This switch activates warnings for a @code{with} of an internal GNAT
5282 implementation unit, defined as any unit from the @code{Ada},
5283 @code{Interfaces}, @code{GNAT},
5284 ^^@code{DEC},^ or @code{System}
5285 hierarchies that is not
5286 documented in either the Ada Reference Manual or the GNAT
5287 Programmer's Reference Manual. Such units are intended only
5288 for internal implementation purposes and should not be @code{with}'ed
5289 by user programs. The default is that such warnings are generated
5290 This warning can also be turned on using @option{-gnatwa}.
5293 @emph{Disable warnings on implementation units.}
5294 @cindex @option{-gnatwI} (@command{gcc})
5295 This switch disables warnings for a @code{with} of an internal GNAT
5296 implementation unit.
5299 @emph{Activate warnings on overlapping actuals.}
5300 @cindex @option{-gnatw.i} (@command{gcc})
5301 This switch enables a warning on statically detectable overlapping actuals in
5302 a subprogram call, when one of the actuals is an in-out parameter, and the
5303 types of the actuals are not by-copy types. The warning is off by default,
5304 and is not included under -gnatwa.
5307 @emph{Disable warnings on overlapping actuals.}
5308 @cindex @option{-gnatw.I} (@command{gcc})
5309 This switch disables warnings on overlapping actuals in a call..
5312 @emph{Activate warnings on obsolescent features (Annex J).}
5313 @cindex @option{-gnatwj} (@command{gcc})
5314 @cindex Features, obsolescent
5315 @cindex Obsolescent features
5316 If this warning option is activated, then warnings are generated for
5317 calls to subprograms marked with @code{pragma Obsolescent} and
5318 for use of features in Annex J of the Ada Reference Manual. In the
5319 case of Annex J, not all features are flagged. In particular use
5320 of the renamed packages (like @code{Text_IO}) and use of package
5321 @code{ASCII} are not flagged, since these are very common and
5322 would generate many annoying positive warnings. The default is that
5323 such warnings are not generated. This warning is also turned on by
5324 the use of @option{-gnatwa}.
5326 In addition to the above cases, warnings are also generated for
5327 GNAT features that have been provided in past versions but which
5328 have been superseded (typically by features in the new Ada standard).
5329 For example, @code{pragma Ravenscar} will be flagged since its
5330 function is replaced by @code{pragma Profile(Ravenscar)}.
5332 Note that this warning option functions differently from the
5333 restriction @code{No_Obsolescent_Features} in two respects.
5334 First, the restriction applies only to annex J features.
5335 Second, the restriction does flag uses of package @code{ASCII}.
5338 @emph{Suppress warnings on obsolescent features (Annex J).}
5339 @cindex @option{-gnatwJ} (@command{gcc})
5340 This switch disables warnings on use of obsolescent features.
5343 @emph{Activate warnings on variables that could be constants.}
5344 @cindex @option{-gnatwk} (@command{gcc})
5345 This switch activates warnings for variables that are initialized but
5346 never modified, and then could be declared constants. The default is that
5347 such warnings are not given.
5348 This warning can also be turned on using @option{-gnatwa}.
5351 @emph{Suppress warnings on variables that could be constants.}
5352 @cindex @option{-gnatwK} (@command{gcc})
5353 This switch disables warnings on variables that could be declared constants.
5356 @emph{Activate warnings for elaboration pragmas.}
5357 @cindex @option{-gnatwl} (@command{gcc})
5358 @cindex Elaboration, warnings
5359 This switch activates warnings on missing
5360 @code{Elaborate_All} and @code{Elaborate} pragmas.
5361 See the section in this guide on elaboration checking for details on
5362 when such pragmas should be used. In dynamic elaboration mode, this switch
5363 generations warnings about the need to add elaboration pragmas. Note however,
5364 that if you blindly follow these warnings, and add @code{Elaborate_All}
5365 warnings wherever they are recommended, you basically end up with the
5366 equivalent of the static elaboration model, which may not be what you want for
5367 legacy code for which the static model does not work.
5369 For the static model, the messages generated are labeled "info:" (for
5370 information messages). They are not warnings to add elaboration pragmas,
5371 merely informational messages showing what implicit elaboration pragmas
5372 have been added, for use in analyzing elaboration circularity problems.
5374 Warnings are also generated if you
5375 are using the static mode of elaboration, and a @code{pragma Elaborate}
5376 is encountered. The default is that such warnings
5378 This warning is not automatically turned on by the use of @option{-gnatwa}.
5381 @emph{Suppress warnings for elaboration pragmas.}
5382 @cindex @option{-gnatwL} (@command{gcc})
5383 This switch suppresses warnings on missing Elaborate and Elaborate_All pragmas.
5384 See the section in this guide on elaboration checking for details on
5385 when such pragmas should be used.
5388 @emph{Activate warnings on modified but unreferenced variables.}
5389 @cindex @option{-gnatwm} (@command{gcc})
5390 This switch activates warnings for variables that are assigned (using
5391 an initialization value or with one or more assignment statements) but
5392 whose value is never read. The warning is suppressed for volatile
5393 variables and also for variables that are renamings of other variables
5394 or for which an address clause is given.
5395 This warning can also be turned on using @option{-gnatwa}.
5396 The default is that these warnings are not given.
5399 @emph{Disable warnings on modified but unreferenced variables.}
5400 @cindex @option{-gnatwM} (@command{gcc})
5401 This switch disables warnings for variables that are assigned or
5402 initialized, but never read.
5405 @emph{Activate warnings on suspicious modulus values.}
5406 @cindex @option{-gnatw.m} (@command{gcc})
5407 This switch activates warnings for modulus values that seem suspicious.
5408 The cases caught are where the size is the same as the modulus (e.g.
5409 a modulus of 7 with a size of 7 bits), and modulus values of 32 or 64
5410 with no size clause. The guess in both cases is that 2**x was intended
5411 rather than x. The default is that these warnings are given.
5414 @emph{Disable warnings on suspicious modulus values.}
5415 @cindex @option{-gnatw.M} (@command{gcc})
5416 This switch disables warnings for suspicious modulus values.
5419 @emph{Set normal warnings mode.}
5420 @cindex @option{-gnatwn} (@command{gcc})
5421 This switch sets normal warning mode, in which enabled warnings are
5422 issued and treated as warnings rather than errors. This is the default
5423 mode. the switch @option{-gnatwn} can be used to cancel the effect of
5424 an explicit @option{-gnatws} or
5425 @option{-gnatwe}. It also cancels the effect of the
5426 implicit @option{-gnatwe} that is activated by the
5427 use of @option{-gnatg}.
5430 @emph{Activate warnings on address clause overlays.}
5431 @cindex @option{-gnatwo} (@command{gcc})
5432 @cindex Address Clauses, warnings
5433 This switch activates warnings for possibly unintended initialization
5434 effects of defining address clauses that cause one variable to overlap
5435 another. The default is that such warnings are generated.
5436 This warning can also be turned on using @option{-gnatwa}.
5439 @emph{Suppress warnings on address clause overlays.}
5440 @cindex @option{-gnatwO} (@command{gcc})
5441 This switch suppresses warnings on possibly unintended initialization
5442 effects of defining address clauses that cause one variable to overlap
5446 @emph{Activate warnings on modified but unreferenced out parameters.}
5447 @cindex @option{-gnatw.o} (@command{gcc})
5448 This switch activates warnings for variables that are modified by using
5449 them as actuals for a call to a procedure with an out mode formal, where
5450 the resulting assigned value is never read. It is applicable in the case
5451 where there is more than one out mode formal. If there is only one out
5452 mode formal, the warning is issued by default (controlled by -gnatwu).
5453 The warning is suppressed for volatile
5454 variables and also for variables that are renamings of other variables
5455 or for which an address clause is given.
5456 The default is that these warnings are not given. Note that this warning
5457 is not included in -gnatwa, it must be activated explicitly.
5460 @emph{Disable warnings on modified but unreferenced out parameters.}
5461 @cindex @option{-gnatw.O} (@command{gcc})
5462 This switch suppresses warnings for variables that are modified by using
5463 them as actuals for a call to a procedure with an out mode formal, where
5464 the resulting assigned value is never read.
5467 @emph{Activate warnings on ineffective pragma Inlines.}
5468 @cindex @option{-gnatwp} (@command{gcc})
5469 @cindex Inlining, warnings
5470 This switch activates warnings for failure of front end inlining
5471 (activated by @option{-gnatN}) to inline a particular call. There are
5472 many reasons for not being able to inline a call, including most
5473 commonly that the call is too complex to inline. The default is
5474 that such warnings are not given.
5475 This warning can also be turned on using @option{-gnatwa}.
5476 Warnings on ineffective inlining by the gcc back-end can be activated
5477 separately, using the gcc switch -Winline.
5480 @emph{Suppress warnings on ineffective pragma Inlines.}
5481 @cindex @option{-gnatwP} (@command{gcc})
5482 This switch suppresses warnings on ineffective pragma Inlines. If the
5483 inlining mechanism cannot inline a call, it will simply ignore the
5487 @emph{Activate warnings on parameter ordering.}
5488 @cindex @option{-gnatw.p} (@command{gcc})
5489 @cindex Parameter order, warnings
5490 This switch activates warnings for cases of suspicious parameter
5491 ordering when the list of arguments are all simple identifiers that
5492 match the names of the formals, but are in a different order. The
5493 warning is suppressed if any use of named parameter notation is used,
5494 so this is the appropriate way to suppress a false positive (and
5495 serves to emphasize that the "misordering" is deliberate). The
5497 that such warnings are not given.
5498 This warning can also be turned on using @option{-gnatwa}.
5501 @emph{Suppress warnings on parameter ordering.}
5502 @cindex @option{-gnatw.P} (@command{gcc})
5503 This switch suppresses warnings on cases of suspicious parameter
5507 @emph{Activate warnings on questionable missing parentheses.}
5508 @cindex @option{-gnatwq} (@command{gcc})
5509 @cindex Parentheses, warnings
5510 This switch activates warnings for cases where parentheses are not used and
5511 the result is potential ambiguity from a readers point of view. For example
5512 (not a > b) when a and b are modular means ((not a) > b) and very likely the
5513 programmer intended (not (a > b)). Similarly (-x mod 5) means (-(x mod 5)) and
5514 quite likely ((-x) mod 5) was intended. In such situations it seems best to
5515 follow the rule of always parenthesizing to make the association clear, and
5516 this warning switch warns if such parentheses are not present. The default
5517 is that these warnings are given.
5518 This warning can also be turned on using @option{-gnatwa}.
5521 @emph{Suppress warnings on questionable missing parentheses.}
5522 @cindex @option{-gnatwQ} (@command{gcc})
5523 This switch suppresses warnings for cases where the association is not
5524 clear and the use of parentheses is preferred.
5527 @emph{Activate warnings on redundant constructs.}
5528 @cindex @option{-gnatwr} (@command{gcc})
5529 This switch activates warnings for redundant constructs. The following
5530 is the current list of constructs regarded as redundant:
5534 Assignment of an item to itself.
5536 Type conversion that converts an expression to its own type.
5538 Use of the attribute @code{Base} where @code{typ'Base} is the same
5541 Use of pragma @code{Pack} when all components are placed by a record
5542 representation clause.
5544 Exception handler containing only a reraise statement (raise with no
5545 operand) which has no effect.
5547 Use of the operator abs on an operand that is known at compile time
5550 Comparison of boolean expressions to an explicit True value.
5553 This warning can also be turned on using @option{-gnatwa}.
5554 The default is that warnings for redundant constructs are not given.
5557 @emph{Suppress warnings on redundant constructs.}
5558 @cindex @option{-gnatwR} (@command{gcc})
5559 This switch suppresses warnings for redundant constructs.
5562 @emph{Activate warnings for object renaming function.}
5563 @cindex @option{-gnatw.r} (@command{gcc})
5564 This switch activates warnings for an object renaming that renames a
5565 function call, which is equivalent to a constant declaration (as
5566 opposed to renaming the function itself). The default is that these
5567 warnings are given. This warning can also be turned on using
5571 @emph{Suppress warnings for object renaming function.}
5572 @cindex @option{-gnatwT} (@command{gcc})
5573 This switch suppresses warnings for object renaming function.
5576 @emph{Suppress all warnings.}
5577 @cindex @option{-gnatws} (@command{gcc})
5578 This switch completely suppresses the
5579 output of all warning messages from the GNAT front end.
5580 Note that it does not suppress warnings from the @command{gcc} back end.
5581 To suppress these back end warnings as well, use the switch @option{-w}
5582 in addition to @option{-gnatws}. Also this switch has no effect on the
5583 handling of style check messages.
5586 @emph{Activate warnings for tracking of deleted conditional code.}
5587 @cindex @option{-gnatwt} (@command{gcc})
5588 @cindex Deactivated code, warnings
5589 @cindex Deleted code, warnings
5590 This switch activates warnings for tracking of code in conditionals (IF and
5591 CASE statements) that is detected to be dead code which cannot be executed, and
5592 which is removed by the front end. This warning is off by default, and is not
5593 turned on by @option{-gnatwa}, it has to be turned on explicitly. This may be
5594 useful for detecting deactivated code in certified applications.
5597 @emph{Suppress warnings for tracking of deleted conditional code.}
5598 @cindex @option{-gnatwT} (@command{gcc})
5599 This switch suppresses warnings for tracking of deleted conditional code.
5602 @emph{Activate warnings on unused entities.}
5603 @cindex @option{-gnatwu} (@command{gcc})
5604 This switch activates warnings to be generated for entities that
5605 are declared but not referenced, and for units that are @code{with}'ed
5607 referenced. In the case of packages, a warning is also generated if
5608 no entities in the package are referenced. This means that if the package
5609 is referenced but the only references are in @code{use}
5610 clauses or @code{renames}
5611 declarations, a warning is still generated. A warning is also generated
5612 for a generic package that is @code{with}'ed but never instantiated.
5613 In the case where a package or subprogram body is compiled, and there
5614 is a @code{with} on the corresponding spec
5615 that is only referenced in the body,
5616 a warning is also generated, noting that the
5617 @code{with} can be moved to the body. The default is that
5618 such warnings are not generated.
5619 This switch also activates warnings on unreferenced formals
5620 (it includes the effect of @option{-gnatwf}).
5621 This warning can also be turned on using @option{-gnatwa}.
5624 @emph{Suppress warnings on unused entities.}
5625 @cindex @option{-gnatwU} (@command{gcc})
5626 This switch suppresses warnings for unused entities and packages.
5627 It also turns off warnings on unreferenced formals (and thus includes
5628 the effect of @option{-gnatwF}).
5631 @emph{Activate warnings on unassigned variables.}
5632 @cindex @option{-gnatwv} (@command{gcc})
5633 @cindex Unassigned variable warnings
5634 This switch activates warnings for access to variables which
5635 may not be properly initialized. The default is that
5636 such warnings are generated.
5637 This warning can also be turned on using @option{-gnatwa}.
5640 @emph{Suppress warnings on unassigned variables.}
5641 @cindex @option{-gnatwV} (@command{gcc})
5642 This switch suppresses warnings for access to variables which
5643 may not be properly initialized.
5644 For variables of a composite type, the warning can also be suppressed in
5645 Ada 2005 by using a default initialization with a box. For example, if
5646 Table is an array of records whose components are only partially uninitialized,
5647 then the following code:
5649 @smallexample @c ada
5650 Tab : Table := (others => <>);
5653 will suppress warnings on subsequent statements that access components
5657 @emph{Activate warnings on wrong low bound assumption.}
5658 @cindex @option{-gnatww} (@command{gcc})
5659 @cindex String indexing warnings
5660 This switch activates warnings for indexing an unconstrained string parameter
5661 with a literal or S'Length. This is a case where the code is assuming that the
5662 low bound is one, which is in general not true (for example when a slice is
5663 passed). The default is that such warnings are generated.
5664 This warning can also be turned on using @option{-gnatwa}.
5667 @emph{Suppress warnings on wrong low bound assumption.}
5668 @cindex @option{-gnatwW} (@command{gcc})
5669 This switch suppresses warnings for indexing an unconstrained string parameter
5670 with a literal or S'Length. Note that this warning can also be suppressed
5671 in a particular case by adding an
5672 assertion that the lower bound is 1,
5673 as shown in the following example.
5675 @smallexample @c ada
5676 procedure K (S : String) is
5677 pragma Assert (S'First = 1);
5682 @emph{Activate warnings on unnecessary Warnings Off pragmas}
5683 @cindex @option{-gnatw.w} (@command{gcc})
5684 @cindex Warnings Off control
5685 This switch activates warnings for use of @code{pragma Warnings (Off, entity}
5686 where either the pragma is entirely useless (because it suppresses no
5687 warnings), or it could be replaced by @code{pragma Unreferenced} or
5688 @code{pragma Unmodified}.The default is that these warnings are not given.
5689 Note that this warning is not included in -gnatwa, it must be
5690 activated explicitly.
5693 @emph{Suppress warnings on unnecessary Warnings Off pragmas}
5694 @cindex @option{-gnatw.W} (@command{gcc})
5695 This switch suppresses warnings for use of @code{pragma Warnings (Off, entity}.
5698 @emph{Activate warnings on Export/Import pragmas.}
5699 @cindex @option{-gnatwx} (@command{gcc})
5700 @cindex Export/Import pragma warnings
5701 This switch activates warnings on Export/Import pragmas when
5702 the compiler detects a possible conflict between the Ada and
5703 foreign language calling sequences. For example, the use of
5704 default parameters in a convention C procedure is dubious
5705 because the C compiler cannot supply the proper default, so
5706 a warning is issued. The default is that such warnings are
5708 This warning can also be turned on using @option{-gnatwa}.
5711 @emph{Suppress warnings on Export/Import pragmas.}
5712 @cindex @option{-gnatwX} (@command{gcc})
5713 This switch suppresses warnings on Export/Import pragmas.
5714 The sense of this is that you are telling the compiler that
5715 you know what you are doing in writing the pragma, and it
5716 should not complain at you.
5719 @emph{Activate warnings for No_Exception_Propagation mode.}
5720 @cindex @option{-gnatwm} (@command{gcc})
5721 This switch activates warnings for exception usage when pragma Restrictions
5722 (No_Exception_Propagation) is in effect. Warnings are given for implicit or
5723 explicit exception raises which are not covered by a local handler, and for
5724 exception handlers which do not cover a local raise. The default is that these
5725 warnings are not given.
5728 @emph{Disable warnings for No_Exception_Propagation mode.}
5729 This switch disables warnings for exception usage when pragma Restrictions
5730 (No_Exception_Propagation) is in effect.
5733 @emph{Activate warnings for Ada 2005 compatibility issues.}
5734 @cindex @option{-gnatwy} (@command{gcc})
5735 @cindex Ada 2005 compatibility issues warnings
5736 For the most part Ada 2005 is upwards compatible with Ada 95,
5737 but there are some exceptions (for example the fact that
5738 @code{interface} is now a reserved word in Ada 2005). This
5739 switch activates several warnings to help in identifying
5740 and correcting such incompatibilities. The default is that
5741 these warnings are generated. Note that at one point Ada 2005
5742 was called Ada 0Y, hence the choice of character.
5743 This warning can also be turned on using @option{-gnatwa}.
5746 @emph{Disable warnings for Ada 2005 compatibility issues.}
5747 @cindex @option{-gnatwY} (@command{gcc})
5748 @cindex Ada 2005 compatibility issues warnings
5749 This switch suppresses several warnings intended to help in identifying
5750 incompatibilities between Ada 95 and Ada 2005.
5753 @emph{Activate warnings on unchecked conversions.}
5754 @cindex @option{-gnatwz} (@command{gcc})
5755 @cindex Unchecked_Conversion warnings
5756 This switch activates warnings for unchecked conversions
5757 where the types are known at compile time to have different
5759 is that such warnings are generated. Warnings are also
5760 generated for subprogram pointers with different conventions,
5761 and, on VMS only, for data pointers with different conventions.
5762 This warning can also be turned on using @option{-gnatwa}.
5765 @emph{Suppress warnings on unchecked conversions.}
5766 @cindex @option{-gnatwZ} (@command{gcc})
5767 This switch suppresses warnings for unchecked conversions
5768 where the types are known at compile time to have different
5769 sizes or conventions.
5771 @item ^-Wunused^WARNINGS=UNUSED^
5772 @cindex @option{-Wunused}
5773 The warnings controlled by the @option{-gnatw} switch are generated by
5774 the front end of the compiler. The @option{GCC} back end can provide
5775 additional warnings and they are controlled by the @option{-W} switch.
5776 For example, @option{^-Wunused^WARNINGS=UNUSED^} activates back end
5777 warnings for entities that are declared but not referenced.
5779 @item ^-Wuninitialized^WARNINGS=UNINITIALIZED^
5780 @cindex @option{-Wuninitialized}
5781 Similarly, @option{^-Wuninitialized^WARNINGS=UNINITIALIZED^} activates
5782 the back end warning for uninitialized variables. This switch must be
5783 used in conjunction with an optimization level greater than zero.
5785 @item ^-Wall^/ALL_BACK_END_WARNINGS^
5786 @cindex @option{-Wall}
5787 This switch enables all the above warnings from the @option{GCC} back end.
5788 The code generator detects a number of warning situations that are missed
5789 by the @option{GNAT} front end, and this switch can be used to activate them.
5790 The use of this switch also sets the default front end warning mode to
5791 @option{-gnatwa}, that is, most front end warnings activated as well.
5793 @item ^-w^/NO_BACK_END_WARNINGS^
5795 Conversely, this switch suppresses warnings from the @option{GCC} back end.
5796 The use of this switch also sets the default front end warning mode to
5797 @option{-gnatws}, that is, front end warnings suppressed as well.
5803 A string of warning parameters can be used in the same parameter. For example:
5810 will turn on all optional warnings except for elaboration pragma warnings,
5811 and also specify that warnings should be treated as errors.
5813 When no switch @option{^-gnatw^/WARNINGS^} is used, this is equivalent to:
5838 @node Debugging and Assertion Control
5839 @subsection Debugging and Assertion Control
5843 @cindex @option{-gnata} (@command{gcc})
5849 The pragmas @code{Assert} and @code{Debug} normally have no effect and
5850 are ignored. This switch, where @samp{a} stands for assert, causes
5851 @code{Assert} and @code{Debug} pragmas to be activated.
5853 The pragmas have the form:
5857 @b{pragma} Assert (@var{Boolean-expression} @r{[},
5858 @var{static-string-expression}@r{]})
5859 @b{pragma} Debug (@var{procedure call})
5864 The @code{Assert} pragma causes @var{Boolean-expression} to be tested.
5865 If the result is @code{True}, the pragma has no effect (other than
5866 possible side effects from evaluating the expression). If the result is
5867 @code{False}, the exception @code{Assert_Failure} declared in the package
5868 @code{System.Assertions} is
5869 raised (passing @var{static-string-expression}, if present, as the
5870 message associated with the exception). If no string expression is
5871 given the default is a string giving the file name and line number
5874 The @code{Debug} pragma causes @var{procedure} to be called. Note that
5875 @code{pragma Debug} may appear within a declaration sequence, allowing
5876 debugging procedures to be called between declarations.
5879 @item /DEBUG@r{[}=debug-level@r{]}
5881 Specifies how much debugging information is to be included in
5882 the resulting object file where 'debug-level' is one of the following:
5885 Include both debugger symbol records and traceback
5887 This is the default setting.
5889 Include both debugger symbol records and traceback in
5892 Excludes both debugger symbol records and traceback
5893 the object file. Same as /NODEBUG.
5895 Includes only debugger symbol records in the object
5896 file. Note that this doesn't include traceback information.
5901 @node Validity Checking
5902 @subsection Validity Checking
5903 @findex Validity Checking
5906 The Ada Reference Manual defines the concept of invalid values (see
5907 RM 13.9.1). The primary source of invalid values is uninitialized
5908 variables. A scalar variable that is left uninitialized may contain
5909 an invalid value; the concept of invalid does not apply to access or
5912 It is an error to read an invalid value, but the RM does not require
5913 run-time checks to detect such errors, except for some minimal
5914 checking to prevent erroneous execution (i.e. unpredictable
5915 behavior). This corresponds to the @option{-gnatVd} switch below,
5916 which is the default. For example, by default, if the expression of a
5917 case statement is invalid, it will raise Constraint_Error rather than
5918 causing a wild jump, and if an array index on the left-hand side of an
5919 assignment is invalid, it will raise Constraint_Error rather than
5920 overwriting an arbitrary memory location.
5922 The @option{-gnatVa} may be used to enable additional validity checks,
5923 which are not required by the RM. These checks are often very
5924 expensive (which is why the RM does not require them). These checks
5925 are useful in tracking down uninitialized variables, but they are
5926 not usually recommended for production builds.
5928 The other @option{-gnatV^@var{x}^^} switches below allow finer-grained
5929 control; you can enable whichever validity checks you desire. However,
5930 for most debugging purposes, @option{-gnatVa} is sufficient, and the
5931 default @option{-gnatVd} (i.e. standard Ada behavior) is usually
5932 sufficient for non-debugging use.
5934 The @option{-gnatB} switch tells the compiler to assume that all
5935 values are valid (that is, within their declared subtype range)
5936 except in the context of a use of the Valid attribute. This means
5937 the compiler can generate more efficient code, since the range
5938 of values is better known at compile time. However, an uninitialized
5939 variable can cause wild jumps and memory corruption in this mode.
5941 The @option{-gnatV^@var{x}^^} switch allows control over the validity
5942 checking mode as described below.
5944 The @code{x} argument is a string of letters that
5945 indicate validity checks that are performed or not performed in addition
5946 to the default checks required by Ada as described above.
5949 The options allowed for this qualifier
5950 indicate validity checks that are performed or not performed in addition
5951 to the default checks required by Ada as described above.
5957 @emph{All validity checks.}
5958 @cindex @option{-gnatVa} (@command{gcc})
5959 All validity checks are turned on.
5961 That is, @option{-gnatVa} is
5962 equivalent to @option{gnatVcdfimorst}.
5966 @emph{Validity checks for copies.}
5967 @cindex @option{-gnatVc} (@command{gcc})
5968 The right hand side of assignments, and the initializing values of
5969 object declarations are validity checked.
5972 @emph{Default (RM) validity checks.}
5973 @cindex @option{-gnatVd} (@command{gcc})
5974 Some validity checks are done by default following normal Ada semantics
5976 A check is done in case statements that the expression is within the range
5977 of the subtype. If it is not, Constraint_Error is raised.
5978 For assignments to array components, a check is done that the expression used
5979 as index is within the range. If it is not, Constraint_Error is raised.
5980 Both these validity checks may be turned off using switch @option{-gnatVD}.
5981 They are turned on by default. If @option{-gnatVD} is specified, a subsequent
5982 switch @option{-gnatVd} will leave the checks turned on.
5983 Switch @option{-gnatVD} should be used only if you are sure that all such
5984 expressions have valid values. If you use this switch and invalid values
5985 are present, then the program is erroneous, and wild jumps or memory
5986 overwriting may occur.
5989 @emph{Validity checks for elementary components.}
5990 @cindex @option{-gnatVe} (@command{gcc})
5991 In the absence of this switch, assignments to record or array components are
5992 not validity checked, even if validity checks for assignments generally
5993 (@option{-gnatVc}) are turned on. In Ada, assignment of composite values do not
5994 require valid data, but assignment of individual components does. So for
5995 example, there is a difference between copying the elements of an array with a
5996 slice assignment, compared to assigning element by element in a loop. This
5997 switch allows you to turn off validity checking for components, even when they
5998 are assigned component by component.
6001 @emph{Validity checks for floating-point values.}
6002 @cindex @option{-gnatVf} (@command{gcc})
6003 In the absence of this switch, validity checking occurs only for discrete
6004 values. If @option{-gnatVf} is specified, then validity checking also applies
6005 for floating-point values, and NaNs and infinities are considered invalid,
6006 as well as out of range values for constrained types. Note that this means
6007 that standard IEEE infinity mode is not allowed. The exact contexts
6008 in which floating-point values are checked depends on the setting of other
6009 options. For example,
6010 @option{^-gnatVif^VALIDITY_CHECKING=(IN_PARAMS,FLOATS)^} or
6011 @option{^-gnatVfi^VALIDITY_CHECKING=(FLOATS,IN_PARAMS)^}
6012 (the order does not matter) specifies that floating-point parameters of mode
6013 @code{in} should be validity checked.
6016 @emph{Validity checks for @code{in} mode parameters}
6017 @cindex @option{-gnatVi} (@command{gcc})
6018 Arguments for parameters of mode @code{in} are validity checked in function
6019 and procedure calls at the point of call.
6022 @emph{Validity checks for @code{in out} mode parameters.}
6023 @cindex @option{-gnatVm} (@command{gcc})
6024 Arguments for parameters of mode @code{in out} are validity checked in
6025 procedure calls at the point of call. The @code{'m'} here stands for
6026 modify, since this concerns parameters that can be modified by the call.
6027 Note that there is no specific option to test @code{out} parameters,
6028 but any reference within the subprogram will be tested in the usual
6029 manner, and if an invalid value is copied back, any reference to it
6030 will be subject to validity checking.
6033 @emph{No validity checks.}
6034 @cindex @option{-gnatVn} (@command{gcc})
6035 This switch turns off all validity checking, including the default checking
6036 for case statements and left hand side subscripts. Note that the use of
6037 the switch @option{-gnatp} suppresses all run-time checks, including
6038 validity checks, and thus implies @option{-gnatVn}. When this switch
6039 is used, it cancels any other @option{-gnatV} previously issued.
6042 @emph{Validity checks for operator and attribute operands.}
6043 @cindex @option{-gnatVo} (@command{gcc})
6044 Arguments for predefined operators and attributes are validity checked.
6045 This includes all operators in package @code{Standard},
6046 the shift operators defined as intrinsic in package @code{Interfaces}
6047 and operands for attributes such as @code{Pos}. Checks are also made
6048 on individual component values for composite comparisons, and on the
6049 expressions in type conversions and qualified expressions. Checks are
6050 also made on explicit ranges using @samp{..} (e.g.@: slices, loops etc).
6053 @emph{Validity checks for parameters.}
6054 @cindex @option{-gnatVp} (@command{gcc})
6055 This controls the treatment of parameters within a subprogram (as opposed
6056 to @option{-gnatVi} and @option{-gnatVm} which control validity testing
6057 of parameters on a call. If either of these call options is used, then
6058 normally an assumption is made within a subprogram that the input arguments
6059 have been validity checking at the point of call, and do not need checking
6060 again within a subprogram). If @option{-gnatVp} is set, then this assumption
6061 is not made, and parameters are not assumed to be valid, so their validity
6062 will be checked (or rechecked) within the subprogram.
6065 @emph{Validity checks for function returns.}
6066 @cindex @option{-gnatVr} (@command{gcc})
6067 The expression in @code{return} statements in functions is validity
6071 @emph{Validity checks for subscripts.}
6072 @cindex @option{-gnatVs} (@command{gcc})
6073 All subscripts expressions are checked for validity, whether they appear
6074 on the right side or left side (in default mode only left side subscripts
6075 are validity checked).
6078 @emph{Validity checks for tests.}
6079 @cindex @option{-gnatVt} (@command{gcc})
6080 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
6081 statements are checked, as well as guard expressions in entry calls.
6086 The @option{-gnatV} switch may be followed by
6087 ^a string of letters^a list of options^
6088 to turn on a series of validity checking options.
6090 @option{^-gnatVcr^/VALIDITY_CHECKING=(COPIES, RETURNS)^}
6091 specifies that in addition to the default validity checking, copies and
6092 function return expressions are to be validity checked.
6093 In order to make it easier
6094 to specify the desired combination of effects,
6096 the upper case letters @code{CDFIMORST} may
6097 be used to turn off the corresponding lower case option.
6100 the prefix @code{NO} on an option turns off the corresponding validity
6103 @item @code{NOCOPIES}
6104 @item @code{NODEFAULT}
6105 @item @code{NOFLOATS}
6106 @item @code{NOIN_PARAMS}
6107 @item @code{NOMOD_PARAMS}
6108 @item @code{NOOPERANDS}
6109 @item @code{NORETURNS}
6110 @item @code{NOSUBSCRIPTS}
6111 @item @code{NOTESTS}
6115 @option{^-gnatVaM^/VALIDITY_CHECKING=(ALL, NOMOD_PARAMS)^}
6116 turns on all validity checking options except for
6117 checking of @code{@b{in out}} procedure arguments.
6119 The specification of additional validity checking generates extra code (and
6120 in the case of @option{-gnatVa} the code expansion can be substantial).
6121 However, these additional checks can be very useful in detecting
6122 uninitialized variables, incorrect use of unchecked conversion, and other
6123 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
6124 is useful in conjunction with the extra validity checking, since this
6125 ensures that wherever possible uninitialized variables have invalid values.
6127 See also the pragma @code{Validity_Checks} which allows modification of
6128 the validity checking mode at the program source level, and also allows for
6129 temporary disabling of validity checks.
6131 @node Style Checking
6132 @subsection Style Checking
6133 @findex Style checking
6136 The @option{-gnaty^x^(option,option,@dots{})^} switch
6137 @cindex @option{-gnaty} (@command{gcc})
6138 causes the compiler to
6139 enforce specified style rules. A limited set of style rules has been used
6140 in writing the GNAT sources themselves. This switch allows user programs
6141 to activate all or some of these checks. If the source program fails a
6142 specified style check, an appropriate message is given, preceded by
6143 the character sequence ``(style)''. This message does not prevent
6144 successful compilation (unless the @option{-gnatwe} switch is used).
6147 @code{(option,option,@dots{})} is a sequence of keywords
6150 The string @var{x} is a sequence of letters or digits
6152 indicating the particular style
6153 checks to be performed. The following checks are defined:
6158 @emph{Specify indentation level.}
6159 If a digit from 1-9 appears
6160 ^in the string after @option{-gnaty}^as an option for /STYLE_CHECKS^
6161 then proper indentation is checked, with the digit indicating the
6162 indentation level required. A value of zero turns off this style check.
6163 The general style of required indentation is as specified by
6164 the examples in the Ada Reference Manual. Full line comments must be
6165 aligned with the @code{--} starting on a column that is a multiple of
6166 the alignment level, or they may be aligned the same way as the following
6167 non-blank line (this is useful when full line comments appear in the middle
6171 @emph{Check attribute casing.}
6172 Attribute names, including the case of keywords such as @code{digits}
6173 used as attributes names, must be written in mixed case, that is, the
6174 initial letter and any letter following an underscore must be uppercase.
6175 All other letters must be lowercase.
6177 @item ^A^ARRAY_INDEXES^
6178 @emph{Use of array index numbers in array attributes.}
6179 When using the array attributes First, Last, Range,
6180 or Length, the index number must be omitted for one-dimensional arrays
6181 and is required for multi-dimensional arrays.
6184 @emph{Blanks not allowed at statement end.}
6185 Trailing blanks are not allowed at the end of statements. The purpose of this
6186 rule, together with h (no horizontal tabs), is to enforce a canonical format
6187 for the use of blanks to separate source tokens.
6189 @item ^B^BOOLEAN_OPERATORS^
6190 @emph{Check Boolean operators.}
6191 The use of AND/OR operators is not permitted except in the cases of modular
6192 operands, array operands, and simple stand-alone boolean variables or
6193 boolean constants. In all other cases AND THEN/OR ELSE are required.
6196 @emph{Check comments.}
6197 Comments must meet the following set of rules:
6202 The ``@code{--}'' that starts the column must either start in column one,
6203 or else at least one blank must precede this sequence.
6206 Comments that follow other tokens on a line must have at least one blank
6207 following the ``@code{--}'' at the start of the comment.
6210 Full line comments must have two blanks following the ``@code{--}'' that
6211 starts the comment, with the following exceptions.
6214 A line consisting only of the ``@code{--}'' characters, possibly preceded
6215 by blanks is permitted.
6218 A comment starting with ``@code{--x}'' where @code{x} is a special character
6220 This allows proper processing of the output generated by specialized tools
6221 including @command{gnatprep} (where ``@code{--!}'' is used) and the SPARK
6223 language (where ``@code{--#}'' is used). For the purposes of this rule, a
6224 special character is defined as being in one of the ASCII ranges
6225 @code{16#21#@dots{}16#2F#} or @code{16#3A#@dots{}16#3F#}.
6226 Note that this usage is not permitted
6227 in GNAT implementation units (i.e., when @option{-gnatg} is used).
6230 A line consisting entirely of minus signs, possibly preceded by blanks, is
6231 permitted. This allows the construction of box comments where lines of minus
6232 signs are used to form the top and bottom of the box.
6235 A comment that starts and ends with ``@code{--}'' is permitted as long as at
6236 least one blank follows the initial ``@code{--}''. Together with the preceding
6237 rule, this allows the construction of box comments, as shown in the following
6240 ---------------------------
6241 -- This is a box comment --
6242 -- with two text lines. --
6243 ---------------------------
6247 @item ^d^DOS_LINE_ENDINGS^
6248 @emph{Check no DOS line terminators present.}
6249 All lines must be terminated by a single ASCII.LF
6250 character (in particular the DOS line terminator sequence CR/LF is not
6254 @emph{Check end/exit labels.}
6255 Optional labels on @code{end} statements ending subprograms and on
6256 @code{exit} statements exiting named loops, are required to be present.
6259 @emph{No form feeds or vertical tabs.}
6260 Neither form feeds nor vertical tab characters are permitted
6264 @emph{GNAT style mode}
6265 The set of style check switches is set to match that used by the GNAT sources.
6266 This may be useful when developing code that is eventually intended to be
6267 incorporated into GNAT. For further details, see GNAT sources.
6270 @emph{No horizontal tabs.}
6271 Horizontal tab characters are not permitted in the source text.
6272 Together with the b (no blanks at end of line) check, this
6273 enforces a canonical form for the use of blanks to separate
6277 @emph{Check if-then layout.}
6278 The keyword @code{then} must appear either on the same
6279 line as corresponding @code{if}, or on a line on its own, lined
6280 up under the @code{if} with at least one non-blank line in between
6281 containing all or part of the condition to be tested.
6284 @emph{check mode IN keywords}
6285 Mode @code{in} (the default mode) is not
6286 allowed to be given explicitly. @code{in out} is fine,
6287 but not @code{in} on its own.
6290 @emph{Check keyword casing.}
6291 All keywords must be in lower case (with the exception of keywords
6292 such as @code{digits} used as attribute names to which this check
6296 @emph{Check layout.}
6297 Layout of statement and declaration constructs must follow the
6298 recommendations in the Ada Reference Manual, as indicated by the
6299 form of the syntax rules. For example an @code{else} keyword must
6300 be lined up with the corresponding @code{if} keyword.
6302 There are two respects in which the style rule enforced by this check
6303 option are more liberal than those in the Ada Reference Manual. First
6304 in the case of record declarations, it is permissible to put the
6305 @code{record} keyword on the same line as the @code{type} keyword, and
6306 then the @code{end} in @code{end record} must line up under @code{type}.
6307 This is also permitted when the type declaration is split on two lines.
6308 For example, any of the following three layouts is acceptable:
6310 @smallexample @c ada
6333 Second, in the case of a block statement, a permitted alternative
6334 is to put the block label on the same line as the @code{declare} or
6335 @code{begin} keyword, and then line the @code{end} keyword up under
6336 the block label. For example both the following are permitted:
6338 @smallexample @c ada
6356 The same alternative format is allowed for loops. For example, both of
6357 the following are permitted:
6359 @smallexample @c ada
6361 Clear : while J < 10 loop
6372 @item ^Lnnn^MAX_NESTING=nnn^
6373 @emph{Set maximum nesting level}
6374 The maximum level of nesting of constructs (including subprograms, loops,
6375 blocks, packages, and conditionals) may not exceed the given value
6376 @option{nnn}. A value of zero disconnects this style check.
6378 @item ^m^LINE_LENGTH^
6379 @emph{Check maximum line length.}
6380 The length of source lines must not exceed 79 characters, including
6381 any trailing blanks. The value of 79 allows convenient display on an
6382 80 character wide device or window, allowing for possible special
6383 treatment of 80 character lines. Note that this count is of
6384 characters in the source text. This means that a tab character counts
6385 as one character in this count but a wide character sequence counts as
6386 a single character (however many bytes are needed in the encoding).
6388 @item ^Mnnn^MAX_LENGTH=nnn^
6389 @emph{Set maximum line length.}
6390 The length of lines must not exceed the
6391 given value @option{nnn}. The maximum value that can be specified is 32767.
6393 @item ^n^STANDARD_CASING^
6394 @emph{Check casing of entities in Standard.}
6395 Any identifier from Standard must be cased
6396 to match the presentation in the Ada Reference Manual (for example,
6397 @code{Integer} and @code{ASCII.NUL}).
6400 @emph{Turn off all style checks}
6401 All style check options are turned off.
6403 @item ^o^ORDERED_SUBPROGRAMS^
6404 @emph{Check order of subprogram bodies.}
6405 All subprogram bodies in a given scope
6406 (e.g.@: a package body) must be in alphabetical order. The ordering
6407 rule uses normal Ada rules for comparing strings, ignoring casing
6408 of letters, except that if there is a trailing numeric suffix, then
6409 the value of this suffix is used in the ordering (e.g.@: Junk2 comes
6412 @item ^O^OVERRIDING_INDICATORS^
6413 @emph{Check that overriding subprograms are explicitly marked as such.}
6414 The declaration of a primitive operation of a type extension that overrides
6415 an inherited operation must carry an overriding indicator.
6418 @emph{Check pragma casing.}
6419 Pragma names must be written in mixed case, that is, the
6420 initial letter and any letter following an underscore must be uppercase.
6421 All other letters must be lowercase.
6423 @item ^r^REFERENCES^
6424 @emph{Check references.}
6425 All identifier references must be cased in the same way as the
6426 corresponding declaration. No specific casing style is imposed on
6427 identifiers. The only requirement is for consistency of references
6430 @item ^S^STATEMENTS_AFTER_THEN_ELSE^
6431 @emph{Check no statements after THEN/ELSE.}
6432 No statements are allowed
6433 on the same line as a THEN or ELSE keyword following the
6434 keyword in an IF statement. OR ELSE and AND THEN are not affected,
6435 and a special exception allows a pragma to appear after ELSE.
6438 @emph{Check separate specs.}
6439 Separate declarations (``specs'') are required for subprograms (a
6440 body is not allowed to serve as its own declaration). The only
6441 exception is that parameterless library level procedures are
6442 not required to have a separate declaration. This exception covers
6443 the most frequent form of main program procedures.
6446 @emph{Check token spacing.}
6447 The following token spacing rules are enforced:
6452 The keywords @code{@b{abs}} and @code{@b{not}} must be followed by a space.
6455 The token @code{=>} must be surrounded by spaces.
6458 The token @code{<>} must be preceded by a space or a left parenthesis.
6461 Binary operators other than @code{**} must be surrounded by spaces.
6462 There is no restriction on the layout of the @code{**} binary operator.
6465 Colon must be surrounded by spaces.
6468 Colon-equal (assignment, initialization) must be surrounded by spaces.
6471 Comma must be the first non-blank character on the line, or be
6472 immediately preceded by a non-blank character, and must be followed
6476 If the token preceding a left parenthesis ends with a letter or digit, then
6477 a space must separate the two tokens.
6480 if the token following a right parenthesis starts with a letter or digit, then
6481 a space must separate the two tokens.
6484 A right parenthesis must either be the first non-blank character on
6485 a line, or it must be preceded by a non-blank character.
6488 A semicolon must not be preceded by a space, and must not be followed by
6489 a non-blank character.
6492 A unary plus or minus may not be followed by a space.
6495 A vertical bar must be surrounded by spaces.
6498 @item ^u^UNNECESSARY_BLANK_LINES^
6499 @emph{Check unnecessary blank lines.}
6500 Unnecessary blank lines are not allowed. A blank line is considered
6501 unnecessary if it appears at the end of the file, or if more than
6502 one blank line occurs in sequence.
6504 @item ^x^XTRA_PARENS^
6505 @emph{Check extra parentheses.}
6506 Unnecessary extra level of parentheses (C-style) are not allowed
6507 around conditions in @code{if} statements, @code{while} statements and
6508 @code{exit} statements.
6510 @item ^y^ALL_BUILTIN^
6511 @emph{Set all standard style check options}
6512 This is equivalent to @code{gnaty3aAbcefhiklmnprst}, that is all checking
6513 options enabled with the exception of @option{-gnatyo}, @option{-gnatyI},
6514 @option{-gnatyS}, @option{-gnatyLnnn},
6515 @option{-gnatyd}, @option{-gnatyu}, and @option{-gnatyx}.
6519 @emph{Remove style check options}
6520 This causes any subsequent options in the string to act as canceling the
6521 corresponding style check option. To cancel maximum nesting level control,
6522 use @option{L} parameter witout any integer value after that, because any
6523 digit following @option{-} in the parameter string of the @option{-gnaty}
6524 option will be threated as canceling indentation check. The same is true
6525 for @option{M} parameter. @option{y} and @option{N} parameters are not
6526 allowed after @option{-}.
6529 This causes any subsequent options in the string to enable the corresponding
6530 style check option. That is, it cancels the effect of a previous ^-^REMOVE^,
6536 @emph{Removing style check options}
6537 If the name of a style check is preceded by @option{NO} then the corresponding
6538 style check is turned off. For example @option{NOCOMMENTS} turns off style
6539 checking for comments.
6544 In the above rules, appearing in column one is always permitted, that is,
6545 counts as meeting either a requirement for a required preceding space,
6546 or as meeting a requirement for no preceding space.
6548 Appearing at the end of a line is also always permitted, that is, counts
6549 as meeting either a requirement for a following space, or as meeting
6550 a requirement for no following space.
6553 If any of these style rules is violated, a message is generated giving
6554 details on the violation. The initial characters of such messages are
6555 always ``@code{(style)}''. Note that these messages are treated as warning
6556 messages, so they normally do not prevent the generation of an object
6557 file. The @option{-gnatwe} switch can be used to treat warning messages,
6558 including style messages, as fatal errors.
6562 @option{-gnaty} on its own (that is not
6563 followed by any letters or digits), then the effect is equivalent
6564 to the use of @option{-gnatyy}, as described above, that is all
6565 built-in standard style check options are enabled.
6569 /STYLE_CHECKS=ALL_BUILTIN enables all checking options with
6570 the exception of ORDERED_SUBPROGRAMS, UNNECESSARY_BLANK_LINES,
6571 XTRA_PARENS, and DOS_LINE_ENDINGS. In addition
6581 clears any previously set style checks.
6583 @node Run-Time Checks
6584 @subsection Run-Time Checks
6585 @cindex Division by zero
6586 @cindex Access before elaboration
6587 @cindex Checks, division by zero
6588 @cindex Checks, access before elaboration
6589 @cindex Checks, stack overflow checking
6592 By default, the following checks are suppressed: integer overflow
6593 checks, stack overflow checks, and checks for access before
6594 elaboration on subprogram calls. All other checks, including range
6595 checks and array bounds checks, are turned on by default. The
6596 following @command{gcc} switches refine this default behavior.
6601 @cindex @option{-gnatp} (@command{gcc})
6602 @cindex Suppressing checks
6603 @cindex Checks, suppressing
6605 This switch causes the unit to be compiled
6606 as though @code{pragma Suppress (All_checks)}
6607 had been present in the source. Validity checks are also eliminated (in
6608 other words @option{-gnatp} also implies @option{-gnatVn}.
6609 Use this switch to improve the performance
6610 of the code at the expense of safety in the presence of invalid data or
6613 Note that when checks are suppressed, the compiler is allowed, but not
6614 required, to omit the checking code. If the run-time cost of the
6615 checking code is zero or near-zero, the compiler will generate it even
6616 if checks are suppressed. In particular, if the compiler can prove
6617 that a certain check will necessarily fail, it will generate code to
6618 do an unconditional ``raise'', even if checks are suppressed. The
6619 compiler warns in this case. Another case in which checks may not be
6620 eliminated is when they are embedded in certain run time routines such
6621 as math library routines.
6623 Of course, run-time checks are omitted whenever the compiler can prove
6624 that they will not fail, whether or not checks are suppressed.
6626 Note that if you suppress a check that would have failed, program
6627 execution is erroneous, which means the behavior is totally
6628 unpredictable. The program might crash, or print wrong answers, or
6629 do anything else. It might even do exactly what you wanted it to do
6630 (and then it might start failing mysteriously next week or next
6631 year). The compiler will generate code based on the assumption that
6632 the condition being checked is true, which can result in disaster if
6633 that assumption is wrong.
6635 The @option{-gnatp} switch has no effect if a subsequent
6636 @option{-gnat-p} switch appears.
6639 @cindex @option{-gnat-p} (@command{gcc})
6640 @cindex Suppressing checks
6641 @cindex Checks, suppressing
6643 This switch cancels the effect of a previous @option{gnatp} switch.
6646 @cindex @option{-gnato} (@command{gcc})
6647 @cindex Overflow checks
6648 @cindex Check, overflow
6649 Enables overflow checking for integer operations.
6650 This causes GNAT to generate slower and larger executable
6651 programs by adding code to check for overflow (resulting in raising
6652 @code{Constraint_Error} as required by standard Ada
6653 semantics). These overflow checks correspond to situations in which
6654 the true value of the result of an operation may be outside the base
6655 range of the result type. The following example shows the distinction:
6657 @smallexample @c ada
6658 X1 : Integer := "Integer'Last";
6659 X2 : Integer range 1 .. 5 := "5";
6660 X3 : Integer := "Integer'Last";
6661 X4 : Integer range 1 .. 5 := "5";
6662 F : Float := "2.0E+20";
6671 Note that if explicit values are assigned at compile time, the
6672 compiler may be able to detect overflow at compile time, in which case
6673 no actual run-time checking code is required, and Constraint_Error
6674 will be raised unconditionally, with or without
6675 @option{-gnato}. That's why the assigned values in the above fragment
6676 are in quotes, the meaning is "assign a value not known to the
6677 compiler that happens to be equal to ...". The remaining discussion
6678 assumes that the compiler cannot detect the values at compile time.
6680 Here the first addition results in a value that is outside the base range
6681 of Integer, and hence requires an overflow check for detection of the
6682 constraint error. Thus the first assignment to @code{X1} raises a
6683 @code{Constraint_Error} exception only if @option{-gnato} is set.
6685 The second increment operation results in a violation of the explicit
6686 range constraint; such range checks are performed by default, and are
6687 unaffected by @option{-gnato}.
6689 The two conversions of @code{F} both result in values that are outside
6690 the base range of type @code{Integer} and thus will raise
6691 @code{Constraint_Error} exceptions only if @option{-gnato} is used.
6692 The fact that the result of the second conversion is assigned to
6693 variable @code{X4} with a restricted range is irrelevant, since the problem
6694 is in the conversion, not the assignment.
6696 Basically the rule is that in the default mode (@option{-gnato} not
6697 used), the generated code assures that all integer variables stay
6698 within their declared ranges, or within the base range if there is
6699 no declared range. This prevents any serious problems like indexes
6700 out of range for array operations.
6702 What is not checked in default mode is an overflow that results in
6703 an in-range, but incorrect value. In the above example, the assignments
6704 to @code{X1}, @code{X2}, @code{X3} all give results that are within the
6705 range of the target variable, but the result is wrong in the sense that
6706 it is too large to be represented correctly. Typically the assignment
6707 to @code{X1} will result in wrap around to the largest negative number.
6708 The conversions of @code{F} will result in some @code{Integer} value
6709 and if that integer value is out of the @code{X4} range then the
6710 subsequent assignment would generate an exception.
6712 @findex Machine_Overflows
6713 Note that the @option{-gnato} switch does not affect the code generated
6714 for any floating-point operations; it applies only to integer
6716 For floating-point, GNAT has the @code{Machine_Overflows}
6717 attribute set to @code{False} and the normal mode of operation is to
6718 generate IEEE NaN and infinite values on overflow or invalid operations
6719 (such as dividing 0.0 by 0.0).
6721 The reason that we distinguish overflow checking from other kinds of
6722 range constraint checking is that a failure of an overflow check, unlike
6723 for example the failure of a range check, can result in an incorrect
6724 value, but cannot cause random memory destruction (like an out of range
6725 subscript), or a wild jump (from an out of range case value). Overflow
6726 checking is also quite expensive in time and space, since in general it
6727 requires the use of double length arithmetic.
6729 Note again that @option{-gnato} is off by default, so overflow checking is
6730 not performed in default mode. This means that out of the box, with the
6731 default settings, GNAT does not do all the checks expected from the
6732 language description in the Ada Reference Manual. If you want all constraint
6733 checks to be performed, as described in this Manual, then you must
6734 explicitly use the -gnato switch either on the @command{gnatmake} or
6735 @command{gcc} command.
6738 @cindex @option{-gnatE} (@command{gcc})
6739 @cindex Elaboration checks
6740 @cindex Check, elaboration
6741 Enables dynamic checks for access-before-elaboration
6742 on subprogram calls and generic instantiations.
6743 Note that @option{-gnatE} is not necessary for safety, because in the
6744 default mode, GNAT ensures statically that the checks would not fail.
6745 For full details of the effect and use of this switch,
6746 @xref{Compiling Using gcc}.
6749 @cindex @option{-fstack-check} (@command{gcc})
6750 @cindex Stack Overflow Checking
6751 @cindex Checks, stack overflow checking
6752 Activates stack overflow checking. For full details of the effect and use of
6753 this switch see @ref{Stack Overflow Checking}.
6758 The setting of these switches only controls the default setting of the
6759 checks. You may modify them using either @code{Suppress} (to remove
6760 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
6763 @node Using gcc for Syntax Checking
6764 @subsection Using @command{gcc} for Syntax Checking
6767 @cindex @option{-gnats} (@command{gcc})
6771 The @code{s} stands for ``syntax''.
6774 Run GNAT in syntax checking only mode. For
6775 example, the command
6778 $ gcc -c -gnats x.adb
6782 compiles file @file{x.adb} in syntax-check-only mode. You can check a
6783 series of files in a single command
6785 , and can use wild cards to specify such a group of files.
6786 Note that you must specify the @option{-c} (compile
6787 only) flag in addition to the @option{-gnats} flag.
6790 You may use other switches in conjunction with @option{-gnats}. In
6791 particular, @option{-gnatl} and @option{-gnatv} are useful to control the
6792 format of any generated error messages.
6794 When the source file is empty or contains only empty lines and/or comments,
6795 the output is a warning:
6798 $ gcc -c -gnats -x ada toto.txt
6799 toto.txt:1:01: warning: empty file, contains no compilation units
6803 Otherwise, the output is simply the error messages, if any. No object file or
6804 ALI file is generated by a syntax-only compilation. Also, no units other
6805 than the one specified are accessed. For example, if a unit @code{X}
6806 @code{with}'s a unit @code{Y}, compiling unit @code{X} in syntax
6807 check only mode does not access the source file containing unit
6810 @cindex Multiple units, syntax checking
6811 Normally, GNAT allows only a single unit in a source file. However, this
6812 restriction does not apply in syntax-check-only mode, and it is possible
6813 to check a file containing multiple compilation units concatenated
6814 together. This is primarily used by the @code{gnatchop} utility
6815 (@pxref{Renaming Files Using gnatchop}).
6818 @node Using gcc for Semantic Checking
6819 @subsection Using @command{gcc} for Semantic Checking
6822 @cindex @option{-gnatc} (@command{gcc})
6826 The @code{c} stands for ``check''.
6828 Causes the compiler to operate in semantic check mode,
6829 with full checking for all illegalities specified in the
6830 Ada Reference Manual, but without generation of any object code
6831 (no object file is generated).
6833 Because dependent files must be accessed, you must follow the GNAT
6834 semantic restrictions on file structuring to operate in this mode:
6838 The needed source files must be accessible
6839 (@pxref{Search Paths and the Run-Time Library (RTL)}).
6842 Each file must contain only one compilation unit.
6845 The file name and unit name must match (@pxref{File Naming Rules}).
6848 The output consists of error messages as appropriate. No object file is
6849 generated. An @file{ALI} file is generated for use in the context of
6850 cross-reference tools, but this file is marked as not being suitable
6851 for binding (since no object file is generated).
6852 The checking corresponds exactly to the notion of
6853 legality in the Ada Reference Manual.
6855 Any unit can be compiled in semantics-checking-only mode, including
6856 units that would not normally be compiled (subunits,
6857 and specifications where a separate body is present).
6860 @node Compiling Different Versions of Ada
6861 @subsection Compiling Different Versions of Ada
6864 The switches described in this section allow you to explicitly specify
6865 the version of the Ada language that your programs are written in.
6866 By default @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
6867 but you can also specify @value{NONDEFAULTLANGUAGEVERSION} or
6868 indicate Ada 83 compatibility mode.
6871 @cindex Compatibility with Ada 83
6873 @item -gnat83 (Ada 83 Compatibility Mode)
6874 @cindex @option{-gnat83} (@command{gcc})
6875 @cindex ACVC, Ada 83 tests
6879 Although GNAT is primarily an Ada 95 / Ada 2005 compiler, this switch
6880 specifies that the program is to be compiled in Ada 83 mode. With
6881 @option{-gnat83}, GNAT rejects most post-Ada 83 extensions and applies Ada 83
6882 semantics where this can be done easily.
6883 It is not possible to guarantee this switch does a perfect
6884 job; some subtle tests, such as are
6885 found in earlier ACVC tests (and that have been removed from the ACATS suite
6886 for Ada 95), might not compile correctly.
6887 Nevertheless, this switch may be useful in some circumstances, for example
6888 where, due to contractual reasons, existing code needs to be maintained
6889 using only Ada 83 features.
6891 With few exceptions (most notably the need to use @code{<>} on
6892 @cindex Generic formal parameters
6893 unconstrained generic formal parameters, the use of the new Ada 95 / Ada 2005
6894 reserved words, and the use of packages
6895 with optional bodies), it is not necessary to specify the
6896 @option{-gnat83} switch when compiling Ada 83 programs, because, with rare
6897 exceptions, Ada 95 and Ada 2005 are upwardly compatible with Ada 83. Thus
6898 a correct Ada 83 program is usually also a correct program
6899 in these later versions of the language standard.
6900 For further information, please refer to @ref{Compatibility and Porting Guide}.
6902 @item -gnat95 (Ada 95 mode)
6903 @cindex @option{-gnat95} (@command{gcc})
6907 This switch directs the compiler to implement the Ada 95 version of the
6909 Since Ada 95 is almost completely upwards
6910 compatible with Ada 83, Ada 83 programs may generally be compiled using
6911 this switch (see the description of the @option{-gnat83} switch for further
6912 information about Ada 83 mode).
6913 If an Ada 2005 program is compiled in Ada 95 mode,
6914 uses of the new Ada 2005 features will cause error
6915 messages or warnings.
6917 This switch also can be used to cancel the effect of a previous
6918 @option{-gnat83}, @option{-gnat05/2005}, or @option{-gnat12/2012}
6919 switch earlier in the command line.
6921 @item -gnat05 or -gnat2005 (Ada 2005 mode)
6922 @cindex @option{-gnat05} (@command{gcc})
6923 @cindex @option{-gnat2005} (@command{gcc})
6924 @cindex Ada 2005 mode
6927 This switch directs the compiler to implement the Ada 2005 version of the
6928 language, as documented in the official Ada standards document.
6929 Since Ada 2005 is almost completely upwards
6930 compatible with Ada 95 (and thus also with Ada 83), Ada 83 and Ada 95 programs
6931 may generally be compiled using this switch (see the description of the
6932 @option{-gnat83} and @option{-gnat95} switches for further
6935 Note that even though Ada 2005 is the current official version of the
6936 language, GNAT still compiles in Ada 95 mode by default, so if you are
6937 using Ada 2005 features in your program, you must use this switch (or
6938 the equivalent Ada_05 or Ada_2005 configuration pragmas).
6940 @item -gnat12 or -gnat2012 (Ada 2012 mode)
6941 @cindex @option{-gnat12} (@command{gcc})
6942 @cindex @option{-gnat2012} (@command{gcc})
6943 @cindex Ada 2012 mode
6946 This switch directs the compiler to implement the Ada 2012 version of the
6948 Since Ada 2012 is almost completely upwards
6949 compatible with Ada 2005 (and thus also with Ada 83, and Ada 95),
6950 Ada 83 and Ada 95 programs
6951 may generally be compiled using this switch (see the description of the
6952 @option{-gnat83}, @option{-gnat95}, and @option{-gnat05/2005} switches
6953 for further information).
6955 For information about the approved ``Ada Issues'' that have been incorporated
6956 into Ada 2012, see @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs}.
6957 Included with GNAT releases is a file @file{features-ada12} that describes
6958 the set of implemented Ada 2012 features.
6960 @item -gnatX (Enable GNAT Extensions)
6961 @cindex @option{-gnatX} (@command{gcc})
6962 @cindex Ada language extensions
6963 @cindex GNAT extensions
6966 This switch directs the compiler to implement the latest version of the
6967 language (currently Ada 2012) and also to enable certain GNAT implementation
6968 extensions that are not part of any Ada standard. For a full list of these
6969 extensions, see the GNAT reference manual.
6973 @node Character Set Control
6974 @subsection Character Set Control
6976 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
6977 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
6980 Normally GNAT recognizes the Latin-1 character set in source program
6981 identifiers, as described in the Ada Reference Manual.
6983 GNAT to recognize alternate character sets in identifiers. @var{c} is a
6984 single character ^^or word^ indicating the character set, as follows:
6988 ISO 8859-1 (Latin-1) identifiers
6991 ISO 8859-2 (Latin-2) letters allowed in identifiers
6994 ISO 8859-3 (Latin-3) letters allowed in identifiers
6997 ISO 8859-4 (Latin-4) letters allowed in identifiers
7000 ISO 8859-5 (Cyrillic) letters allowed in identifiers
7003 ISO 8859-15 (Latin-9) letters allowed in identifiers
7006 IBM PC letters (code page 437) allowed in identifiers
7009 IBM PC letters (code page 850) allowed in identifiers
7011 @item ^f^FULL_UPPER^
7012 Full upper-half codes allowed in identifiers
7015 No upper-half codes allowed in identifiers
7018 Wide-character codes (that is, codes greater than 255)
7019 allowed in identifiers
7022 @xref{Foreign Language Representation}, for full details on the
7023 implementation of these character sets.
7025 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
7026 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
7027 Specify the method of encoding for wide characters.
7028 @var{e} is one of the following:
7033 Hex encoding (brackets coding also recognized)
7036 Upper half encoding (brackets encoding also recognized)
7039 Shift/JIS encoding (brackets encoding also recognized)
7042 EUC encoding (brackets encoding also recognized)
7045 UTF-8 encoding (brackets encoding also recognized)
7048 Brackets encoding only (default value)
7050 For full details on these encoding
7051 methods see @ref{Wide Character Encodings}.
7052 Note that brackets coding is always accepted, even if one of the other
7053 options is specified, so for example @option{-gnatW8} specifies that both
7054 brackets and UTF-8 encodings will be recognized. The units that are
7055 with'ed directly or indirectly will be scanned using the specified
7056 representation scheme, and so if one of the non-brackets scheme is
7057 used, it must be used consistently throughout the program. However,
7058 since brackets encoding is always recognized, it may be conveniently
7059 used in standard libraries, allowing these libraries to be used with
7060 any of the available coding schemes.
7063 If no @option{-gnatW?} parameter is present, then the default
7064 representation is normally Brackets encoding only. However, if the
7065 first three characters of the file are 16#EF# 16#BB# 16#BF# (the standard
7066 byte order mark or BOM for UTF-8), then these three characters are
7067 skipped and the default representation for the file is set to UTF-8.
7069 Note that the wide character representation that is specified (explicitly
7070 or by default) for the main program also acts as the default encoding used
7071 for Wide_Text_IO files if not specifically overridden by a WCEM form
7075 @node File Naming Control
7076 @subsection File Naming Control
7079 @item ^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{n}
7080 @cindex @option{-gnatk} (@command{gcc})
7081 Activates file name ``krunching''. @var{n}, a decimal integer in the range
7082 1-999, indicates the maximum allowable length of a file name (not
7083 including the @file{.ads} or @file{.adb} extension). The default is not
7084 to enable file name krunching.
7086 For the source file naming rules, @xref{File Naming Rules}.
7089 @node Subprogram Inlining Control
7090 @subsection Subprogram Inlining Control
7095 @cindex @option{-gnatn} (@command{gcc})
7097 The @code{n} here is intended to suggest the first syllable of the
7100 GNAT recognizes and processes @code{Inline} pragmas. However, for the
7101 inlining to actually occur, optimization must be enabled. To enable
7102 inlining of subprograms specified by pragma @code{Inline},
7103 you must also specify this switch.
7104 In the absence of this switch, GNAT does not attempt
7105 inlining and does not need to access the bodies of
7106 subprograms for which @code{pragma Inline} is specified if they are not
7107 in the current unit.
7109 If you specify this switch the compiler will access these bodies,
7110 creating an extra source dependency for the resulting object file, and
7111 where possible, the call will be inlined.
7112 For further details on when inlining is possible
7113 see @ref{Inlining of Subprograms}.
7116 @cindex @option{-gnatN} (@command{gcc})
7117 This switch activates front-end inlining which also
7118 generates additional dependencies.
7120 When using a gcc-based back end (in practice this means using any version
7121 of GNAT other than the JGNAT, .NET or GNAAMP versions), then the use of
7122 @option{-gnatN} is deprecated, and the use of @option{-gnatn} is preferred.
7123 Historically front end inlining was more extensive than the gcc back end
7124 inlining, but that is no longer the case.
7127 @node Auxiliary Output Control
7128 @subsection Auxiliary Output Control
7132 @cindex @option{-gnatt} (@command{gcc})
7133 @cindex Writing internal trees
7134 @cindex Internal trees, writing to file
7135 Causes GNAT to write the internal tree for a unit to a file (with the
7136 extension @file{.adt}.
7137 This not normally required, but is used by separate analysis tools.
7139 these tools do the necessary compilations automatically, so you should
7140 not have to specify this switch in normal operation.
7141 Note that the combination of switches @option{-gnatct}
7142 generates a tree in the form required by ASIS applications.
7145 @cindex @option{-gnatu} (@command{gcc})
7146 Print a list of units required by this compilation on @file{stdout}.
7147 The listing includes all units on which the unit being compiled depends
7148 either directly or indirectly.
7151 @item -pass-exit-codes
7152 @cindex @option{-pass-exit-codes} (@command{gcc})
7153 If this switch is not used, the exit code returned by @command{gcc} when
7154 compiling multiple files indicates whether all source files have
7155 been successfully used to generate object files or not.
7157 When @option{-pass-exit-codes} is used, @command{gcc} exits with an extended
7158 exit status and allows an integrated development environment to better
7159 react to a compilation failure. Those exit status are:
7163 There was an error in at least one source file.
7165 At least one source file did not generate an object file.
7167 The compiler died unexpectedly (internal error for example).
7169 An object file has been generated for every source file.
7174 @node Debugging Control
7175 @subsection Debugging Control
7179 @cindex Debugging options
7182 @cindex @option{-gnatd} (@command{gcc})
7183 Activate internal debugging switches. @var{x} is a letter or digit, or
7184 string of letters or digits, which specifies the type of debugging
7185 outputs desired. Normally these are used only for internal development
7186 or system debugging purposes. You can find full documentation for these
7187 switches in the body of the @code{Debug} unit in the compiler source
7188 file @file{debug.adb}.
7192 @cindex @option{-gnatG} (@command{gcc})
7193 This switch causes the compiler to generate auxiliary output containing
7194 a pseudo-source listing of the generated expanded code. Like most Ada
7195 compilers, GNAT works by first transforming the high level Ada code into
7196 lower level constructs. For example, tasking operations are transformed
7197 into calls to the tasking run-time routines. A unique capability of GNAT
7198 is to list this expanded code in a form very close to normal Ada source.
7199 This is very useful in understanding the implications of various Ada
7200 usage on the efficiency of the generated code. There are many cases in
7201 Ada (e.g.@: the use of controlled types), where simple Ada statements can
7202 generate a lot of run-time code. By using @option{-gnatG} you can identify
7203 these cases, and consider whether it may be desirable to modify the coding
7204 approach to improve efficiency.
7206 The optional parameter @code{nn} if present after -gnatG specifies an
7207 alternative maximum line length that overrides the normal default of 72.
7208 This value is in the range 40-999999, values less than 40 being silently
7209 reset to 40. The equal sign is optional.
7211 The format of the output is very similar to standard Ada source, and is
7212 easily understood by an Ada programmer. The following special syntactic
7213 additions correspond to low level features used in the generated code that
7214 do not have any exact analogies in pure Ada source form. The following
7215 is a partial list of these special constructions. See the spec
7216 of package @code{Sprint} in file @file{sprint.ads} for a full list.
7218 If the switch @option{-gnatL} is used in conjunction with
7219 @cindex @option{-gnatL} (@command{gcc})
7220 @option{-gnatG}, then the original source lines are interspersed
7221 in the expanded source (as comment lines with the original line number).
7224 @item new @var{xxx} @r{[}storage_pool = @var{yyy}@r{]}
7225 Shows the storage pool being used for an allocator.
7227 @item at end @var{procedure-name};
7228 Shows the finalization (cleanup) procedure for a scope.
7230 @item (if @var{expr} then @var{expr} else @var{expr})
7231 Conditional expression equivalent to the @code{x?y:z} construction in C.
7233 @item @var{target}^^^(@var{source})
7234 A conversion with floating-point truncation instead of rounding.
7236 @item @var{target}?(@var{source})
7237 A conversion that bypasses normal Ada semantic checking. In particular
7238 enumeration types and fixed-point types are treated simply as integers.
7240 @item @var{target}?^^^(@var{source})
7241 Combines the above two cases.
7243 @item @var{x} #/ @var{y}
7244 @itemx @var{x} #mod @var{y}
7245 @itemx @var{x} #* @var{y}
7246 @itemx @var{x} #rem @var{y}
7247 A division or multiplication of fixed-point values which are treated as
7248 integers without any kind of scaling.
7250 @item free @var{expr} @r{[}storage_pool = @var{xxx}@r{]}
7251 Shows the storage pool associated with a @code{free} statement.
7253 @item [subtype or type declaration]
7254 Used to list an equivalent declaration for an internally generated
7255 type that is referenced elsewhere in the listing.
7257 @c @item freeze @var{type-name} @ovar{actions}
7258 @c Expanding @ovar macro inline (explanation in macro def comments)
7259 @item freeze @var{type-name} @r{[}@var{actions}@r{]}
7260 Shows the point at which @var{type-name} is frozen, with possible
7261 associated actions to be performed at the freeze point.
7263 @item reference @var{itype}
7264 Reference (and hence definition) to internal type @var{itype}.
7266 @item @var{function-name}! (@var{arg}, @var{arg}, @var{arg})
7267 Intrinsic function call.
7269 @item @var{label-name} : label
7270 Declaration of label @var{labelname}.
7272 @item #$ @var{subprogram-name}
7273 An implicit call to a run-time support routine
7274 (to meet the requirement of H.3.1(9) in a
7277 @item @var{expr} && @var{expr} && @var{expr} @dots{} && @var{expr}
7278 A multiple concatenation (same effect as @var{expr} & @var{expr} &
7279 @var{expr}, but handled more efficiently).
7281 @item [constraint_error]
7282 Raise the @code{Constraint_Error} exception.
7284 @item @var{expression}'reference
7285 A pointer to the result of evaluating @var{expression}.
7287 @item @var{target-type}!(@var{source-expression})
7288 An unchecked conversion of @var{source-expression} to @var{target-type}.
7290 @item [@var{numerator}/@var{denominator}]
7291 Used to represent internal real literals (that) have no exact
7292 representation in base 2-16 (for example, the result of compile time
7293 evaluation of the expression 1.0/27.0).
7297 @cindex @option{-gnatD} (@command{gcc})
7298 When used in conjunction with @option{-gnatG}, this switch causes
7299 the expanded source, as described above for
7300 @option{-gnatG} to be written to files with names
7301 @file{^xxx.dg^XXX_DG^}, where @file{xxx} is the normal file name,
7302 instead of to the standard output file. For
7303 example, if the source file name is @file{hello.adb}, then a file
7304 @file{^hello.adb.dg^HELLO.ADB_DG^} will be written. The debugging
7305 information generated by the @command{gcc} @option{^-g^/DEBUG^} switch
7306 will refer to the generated @file{^xxx.dg^XXX_DG^} file. This allows
7307 you to do source level debugging using the generated code which is
7308 sometimes useful for complex code, for example to find out exactly
7309 which part of a complex construction raised an exception. This switch
7310 also suppress generation of cross-reference information (see
7311 @option{-gnatx}) since otherwise the cross-reference information
7312 would refer to the @file{^.dg^.DG^} file, which would cause
7313 confusion since this is not the original source file.
7315 Note that @option{-gnatD} actually implies @option{-gnatG}
7316 automatically, so it is not necessary to give both options.
7317 In other words @option{-gnatD} is equivalent to @option{-gnatDG}).
7319 If the switch @option{-gnatL} is used in conjunction with
7320 @cindex @option{-gnatL} (@command{gcc})
7321 @option{-gnatDG}, then the original source lines are interspersed
7322 in the expanded source (as comment lines with the original line number).
7324 The optional parameter @code{nn} if present after -gnatD specifies an
7325 alternative maximum line length that overrides the normal default of 72.
7326 This value is in the range 40-999999, values less than 40 being silently
7327 reset to 40. The equal sign is optional.
7330 @cindex @option{-gnatr} (@command{gcc})
7331 @cindex pragma Restrictions
7332 This switch causes pragma Restrictions to be treated as Restriction_Warnings
7333 so that violation of restrictions causes warnings rather than illegalities.
7334 This is useful during the development process when new restrictions are added
7335 or investigated. The switch also causes pragma Profile to be treated as
7336 Profile_Warnings, and pragma Restricted_Run_Time and pragma Ravenscar set
7337 restriction warnings rather than restrictions.
7340 @item -gnatR@r{[}0@r{|}1@r{|}2@r{|}3@r{[}s@r{]]}
7341 @cindex @option{-gnatR} (@command{gcc})
7342 This switch controls output from the compiler of a listing showing
7343 representation information for declared types and objects. For
7344 @option{-gnatR0}, no information is output (equivalent to omitting
7345 the @option{-gnatR} switch). For @option{-gnatR1} (which is the default,
7346 so @option{-gnatR} with no parameter has the same effect), size and alignment
7347 information is listed for declared array and record types. For
7348 @option{-gnatR2}, size and alignment information is listed for all
7349 declared types and objects. Finally @option{-gnatR3} includes symbolic
7350 expressions for values that are computed at run time for
7351 variant records. These symbolic expressions have a mostly obvious
7352 format with #n being used to represent the value of the n'th
7353 discriminant. See source files @file{repinfo.ads/adb} in the
7354 @code{GNAT} sources for full details on the format of @option{-gnatR3}
7355 output. If the switch is followed by an s (e.g.@: @option{-gnatR2s}), then
7356 the output is to a file with the name @file{^file.rep^file_REP^} where
7357 file is the name of the corresponding source file.
7360 @item /REPRESENTATION_INFO
7361 @cindex @option{/REPRESENTATION_INFO} (@command{gcc})
7362 This qualifier controls output from the compiler of a listing showing
7363 representation information for declared types and objects. For
7364 @option{/REPRESENTATION_INFO=NONE}, no information is output
7365 (equivalent to omitting the @option{/REPRESENTATION_INFO} qualifier).
7366 @option{/REPRESENTATION_INFO} without option is equivalent to
7367 @option{/REPRESENTATION_INFO=ARRAYS}.
7368 For @option{/REPRESENTATION_INFO=ARRAYS}, size and alignment
7369 information is listed for declared array and record types. For
7370 @option{/REPRESENTATION_INFO=OBJECTS}, size and alignment information
7371 is listed for all expression information for values that are computed
7372 at run time for variant records. These symbolic expressions have a mostly
7373 obvious format with #n being used to represent the value of the n'th
7374 discriminant. See source files @file{REPINFO.ADS/ADB} in the
7375 @code{GNAT} sources for full details on the format of
7376 @option{/REPRESENTATION_INFO=SYMBOLIC} output.
7377 If _FILE is added at the end of an option
7378 (e.g.@: @option{/REPRESENTATION_INFO=ARRAYS_FILE}),
7379 then the output is to a file with the name @file{file_REP} where
7380 file is the name of the corresponding source file.
7382 Note that it is possible for record components to have zero size. In
7383 this case, the component clause uses an obvious extension of permitted
7384 Ada syntax, for example @code{at 0 range 0 .. -1}.
7386 Representation information requires that code be generated (since it is the
7387 code generator that lays out complex data structures). If an attempt is made
7388 to output representation information when no code is generated, for example
7389 when a subunit is compiled on its own, then no information can be generated
7390 and the compiler outputs a message to this effect.
7393 @cindex @option{-gnatS} (@command{gcc})
7394 The use of the switch @option{-gnatS} for an
7395 Ada compilation will cause the compiler to output a
7396 representation of package Standard in a form very
7397 close to standard Ada. It is not quite possible to
7398 do this entirely in standard Ada (since new
7399 numeric base types cannot be created in standard
7400 Ada), but the output is easily
7401 readable to any Ada programmer, and is useful to
7402 determine the characteristics of target dependent
7403 types in package Standard.
7406 @cindex @option{-gnatx} (@command{gcc})
7407 Normally the compiler generates full cross-referencing information in
7408 the @file{ALI} file. This information is used by a number of tools,
7409 including @code{gnatfind} and @code{gnatxref}. The @option{-gnatx} switch
7410 suppresses this information. This saves some space and may slightly
7411 speed up compilation, but means that these tools cannot be used.
7414 @node Exception Handling Control
7415 @subsection Exception Handling Control
7418 GNAT uses two methods for handling exceptions at run-time. The
7419 @code{setjmp/longjmp} method saves the context when entering
7420 a frame with an exception handler. Then when an exception is
7421 raised, the context can be restored immediately, without the
7422 need for tracing stack frames. This method provides very fast
7423 exception propagation, but introduces significant overhead for
7424 the use of exception handlers, even if no exception is raised.
7426 The other approach is called ``zero cost'' exception handling.
7427 With this method, the compiler builds static tables to describe
7428 the exception ranges. No dynamic code is required when entering
7429 a frame containing an exception handler. When an exception is
7430 raised, the tables are used to control a back trace of the
7431 subprogram invocation stack to locate the required exception
7432 handler. This method has considerably poorer performance for
7433 the propagation of exceptions, but there is no overhead for
7434 exception handlers if no exception is raised. Note that in this
7435 mode and in the context of mixed Ada and C/C++ programming,
7436 to propagate an exception through a C/C++ code, the C/C++ code
7437 must be compiled with the @option{-funwind-tables} GCC's
7440 The following switches may be used to control which of the
7441 two exception handling methods is used.
7447 @cindex @option{--RTS=sjlj} (@command{gnatmake})
7448 This switch causes the setjmp/longjmp run-time (when available) to be used
7449 for exception handling. If the default
7450 mechanism for the target is zero cost exceptions, then
7451 this switch can be used to modify this default, and must be
7452 used for all units in the partition.
7453 This option is rarely used. One case in which it may be
7454 advantageous is if you have an application where exception
7455 raising is common and the overall performance of the
7456 application is improved by favoring exception propagation.
7459 @cindex @option{--RTS=zcx} (@command{gnatmake})
7460 @cindex Zero Cost Exceptions
7461 This switch causes the zero cost approach to be used
7462 for exception handling. If this is the default mechanism for the
7463 target (see below), then this switch is unneeded. If the default
7464 mechanism for the target is setjmp/longjmp exceptions, then
7465 this switch can be used to modify this default, and must be
7466 used for all units in the partition.
7467 This option can only be used if the zero cost approach
7468 is available for the target in use, otherwise it will generate an error.
7472 The same option @option{--RTS} must be used both for @command{gcc}
7473 and @command{gnatbind}. Passing this option to @command{gnatmake}
7474 (@pxref{Switches for gnatmake}) will ensure the required consistency
7475 through the compilation and binding steps.
7477 @node Units to Sources Mapping Files
7478 @subsection Units to Sources Mapping Files
7482 @item -gnatem=@var{path}
7483 @cindex @option{-gnatem} (@command{gcc})
7484 A mapping file is a way to communicate to the compiler two mappings:
7485 from unit names to file names (without any directory information) and from
7486 file names to path names (with full directory information). These mappings
7487 are used by the compiler to short-circuit the path search.
7489 The use of mapping files is not required for correct operation of the
7490 compiler, but mapping files can improve efficiency, particularly when
7491 sources are read over a slow network connection. In normal operation,
7492 you need not be concerned with the format or use of mapping files,
7493 and the @option{-gnatem} switch is not a switch that you would use
7494 explicitly. It is intended primarily for use by automatic tools such as
7495 @command{gnatmake} running under the project file facility. The
7496 description here of the format of mapping files is provided
7497 for completeness and for possible use by other tools.
7499 A mapping file is a sequence of sets of three lines. In each set, the
7500 first line is the unit name, in lower case, with @code{%s} appended
7501 for specs and @code{%b} appended for bodies; the second line is the
7502 file name; and the third line is the path name.
7508 /gnat/project1/sources/main.2.ada
7511 When the switch @option{-gnatem} is specified, the compiler will
7512 create in memory the two mappings from the specified file. If there is
7513 any problem (nonexistent file, truncated file or duplicate entries),
7514 no mapping will be created.
7516 Several @option{-gnatem} switches may be specified; however, only the
7517 last one on the command line will be taken into account.
7519 When using a project file, @command{gnatmake} creates a temporary
7520 mapping file and communicates it to the compiler using this switch.
7524 @node Integrated Preprocessing
7525 @subsection Integrated Preprocessing
7528 GNAT sources may be preprocessed immediately before compilation.
7529 In this case, the actual
7530 text of the source is not the text of the source file, but is derived from it
7531 through a process called preprocessing. Integrated preprocessing is specified
7532 through switches @option{-gnatep} and/or @option{-gnateD}. @option{-gnatep}
7533 indicates, through a text file, the preprocessing data to be used.
7534 @option{-gnateD} specifies or modifies the values of preprocessing symbol.
7537 Note that when integrated preprocessing is used, the output from the
7538 preprocessor is not written to any external file. Instead it is passed
7539 internally to the compiler. If you need to preserve the result of
7540 preprocessing in a file, then you should use @command{gnatprep}
7541 to perform the desired preprocessing in stand-alone mode.
7544 It is recommended that @command{gnatmake} switch ^-s^/SWITCH_CHECK^ should be
7545 used when Integrated Preprocessing is used. The reason is that preprocessing
7546 with another Preprocessing Data file without changing the sources will
7547 not trigger recompilation without this switch.
7550 Note that @command{gnatmake} switch ^-m^/MINIMAL_RECOMPILATION^ will almost
7551 always trigger recompilation for sources that are preprocessed,
7552 because @command{gnatmake} cannot compute the checksum of the source after
7556 The actual preprocessing function is described in details in section
7557 @ref{Preprocessing Using gnatprep}. This section only describes how integrated
7558 preprocessing is triggered and parameterized.
7562 @item -gnatep=@var{file}
7563 @cindex @option{-gnatep} (@command{gcc})
7564 This switch indicates to the compiler the file name (without directory
7565 information) of the preprocessor data file to use. The preprocessor data file
7566 should be found in the source directories.
7569 A preprocessing data file is a text file with significant lines indicating
7570 how should be preprocessed either a specific source or all sources not
7571 mentioned in other lines. A significant line is a nonempty, non-comment line.
7572 Comments are similar to Ada comments.
7575 Each significant line starts with either a literal string or the character '*'.
7576 A literal string is the file name (without directory information) of the source
7577 to preprocess. A character '*' indicates the preprocessing for all the sources
7578 that are not specified explicitly on other lines (order of the lines is not
7579 significant). It is an error to have two lines with the same file name or two
7580 lines starting with the character '*'.
7583 After the file name or the character '*', another optional literal string
7584 indicating the file name of the definition file to be used for preprocessing
7585 (@pxref{Form of Definitions File}). The definition files are found by the
7586 compiler in one of the source directories. In some cases, when compiling
7587 a source in a directory other than the current directory, if the definition
7588 file is in the current directory, it may be necessary to add the current
7589 directory as a source directory through switch ^-I.^/SEARCH=[]^, otherwise
7590 the compiler would not find the definition file.
7593 Then, optionally, ^switches^switches^ similar to those of @code{gnatprep} may
7594 be found. Those ^switches^switches^ are:
7599 Causes both preprocessor lines and the lines deleted by
7600 preprocessing to be replaced by blank lines, preserving the line number.
7601 This ^switch^switch^ is always implied; however, if specified after @option{-c}
7602 it cancels the effect of @option{-c}.
7605 Causes both preprocessor lines and the lines deleted
7606 by preprocessing to be retained as comments marked
7607 with the special string ``@code{--! }''.
7609 @item -Dsymbol=value
7610 Define or redefine a symbol, associated with value. A symbol is an Ada
7611 identifier, or an Ada reserved word, with the exception of @code{if},
7612 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7613 @code{value} is either a literal string, an Ada identifier or any Ada reserved
7614 word. A symbol declared with this ^switch^switch^ replaces a symbol with the
7615 same name defined in a definition file.
7618 Causes a sorted list of symbol names and values to be
7619 listed on the standard output file.
7622 Causes undefined symbols to be treated as having the value @code{FALSE}
7624 of a preprocessor test. In the absence of this option, an undefined symbol in
7625 a @code{#if} or @code{#elsif} test will be treated as an error.
7630 Examples of valid lines in a preprocessor data file:
7633 "toto.adb" "prep.def" -u
7634 -- preprocess "toto.adb", using definition file "prep.def",
7635 -- undefined symbol are False.
7638 -- preprocess all other sources without a definition file;
7639 -- suppressed lined are commented; symbol VERSION has the value V101.
7641 "titi.adb" "prep2.def" -s
7642 -- preprocess "titi.adb", using definition file "prep2.def";
7643 -- list all symbols with their values.
7646 @item ^-gnateD^/DATA_PREPROCESSING=^symbol@r{[}=value@r{]}
7647 @cindex @option{-gnateD} (@command{gcc})
7648 Define or redefine a preprocessing symbol, associated with value. If no value
7649 is given on the command line, then the value of the symbol is @code{True}.
7650 A symbol is an identifier, following normal Ada (case-insensitive)
7651 rules for its syntax, and value is any sequence (including an empty sequence)
7652 of characters from the set (letters, digits, period, underline).
7653 Ada reserved words may be used as symbols, with the exceptions of @code{if},
7654 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
7657 A symbol declared with this ^switch^switch^ on the command line replaces a
7658 symbol with the same name either in a definition file or specified with a
7659 ^switch^switch^ -D in the preprocessor data file.
7662 This switch is similar to switch @option{^-D^/ASSOCIATE^} of @code{gnatprep}.
7665 When integrated preprocessing is performed and the preprocessor modifies
7666 the source text, write the result of this preprocessing into a file
7667 <source>^.prep^_prep^.
7671 @node Code Generation Control
7672 @subsection Code Generation Control
7676 The GCC technology provides a wide range of target dependent
7677 @option{-m} switches for controlling
7678 details of code generation with respect to different versions of
7679 architectures. This includes variations in instruction sets (e.g.@:
7680 different members of the power pc family), and different requirements
7681 for optimal arrangement of instructions (e.g.@: different members of
7682 the x86 family). The list of available @option{-m} switches may be
7683 found in the GCC documentation.
7685 Use of these @option{-m} switches may in some cases result in improved
7688 The GNAT Pro technology is tested and qualified without any
7689 @option{-m} switches,
7690 so generally the most reliable approach is to avoid the use of these
7691 switches. However, we generally expect most of these switches to work
7692 successfully with GNAT Pro, and many customers have reported successful
7693 use of these options.
7695 Our general advice is to avoid the use of @option{-m} switches unless
7696 special needs lead to requirements in this area. In particular,
7697 there is no point in using @option{-m} switches to improve performance
7698 unless you actually see a performance improvement.
7702 @subsection Return Codes
7703 @cindex Return Codes
7704 @cindex @option{/RETURN_CODES=VMS}
7707 On VMS, GNAT compiled programs return POSIX-style codes by default,
7708 e.g.@: @option{/RETURN_CODES=POSIX}.
7710 To enable VMS style return codes, use GNAT BIND and LINK with the option
7711 @option{/RETURN_CODES=VMS}. For example:
7714 GNAT BIND MYMAIN.ALI /RETURN_CODES=VMS
7715 GNAT LINK MYMAIN.ALI /RETURN_CODES=VMS
7719 Programs built with /RETURN_CODES=VMS are suitable to be called in
7720 VMS DCL scripts. Programs compiled with the default /RETURN_CODES=POSIX
7721 are suitable for spawning with appropriate GNAT RTL routines.
7725 @node Search Paths and the Run-Time Library (RTL)
7726 @section Search Paths and the Run-Time Library (RTL)
7729 With the GNAT source-based library system, the compiler must be able to
7730 find source files for units that are needed by the unit being compiled.
7731 Search paths are used to guide this process.
7733 The compiler compiles one source file whose name must be given
7734 explicitly on the command line. In other words, no searching is done
7735 for this file. To find all other source files that are needed (the most
7736 common being the specs of units), the compiler examines the following
7737 directories, in the following order:
7741 The directory containing the source file of the main unit being compiled
7742 (the file name on the command line).
7745 Each directory named by an @option{^-I^/SOURCE_SEARCH^} switch given on the
7746 @command{gcc} command line, in the order given.
7749 @findex ADA_PRJ_INCLUDE_FILE
7750 Each of the directories listed in the text file whose name is given
7751 by the @env{ADA_PRJ_INCLUDE_FILE} ^environment variable^logical name^.
7754 @env{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
7755 driver when project files are used. It should not normally be set
7759 @findex ADA_INCLUDE_PATH
7760 Each of the directories listed in the value of the
7761 @env{ADA_INCLUDE_PATH} ^environment variable^logical name^.
7763 Construct this value
7764 exactly as the @env{PATH} environment variable: a list of directory
7765 names separated by colons (semicolons when working with the NT version).
7768 Normally, define this value as a logical name containing a comma separated
7769 list of directory names.
7771 This variable can also be defined by means of an environment string
7772 (an argument to the HP C exec* set of functions).
7776 DEFINE ANOTHER_PATH FOO:[BAG]
7777 DEFINE ADA_INCLUDE_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
7780 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
7781 first, followed by the standard Ada
7782 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADAINCLUDE].
7783 If this is not redefined, the user will obtain the HP Ada 83 IO packages
7784 (Text_IO, Sequential_IO, etc)
7785 instead of the standard Ada packages. Thus, in order to get the standard Ada
7786 packages by default, ADA_INCLUDE_PATH must be redefined.
7790 The content of the @file{ada_source_path} file which is part of the GNAT
7791 installation tree and is used to store standard libraries such as the
7792 GNAT Run Time Library (RTL) source files.
7794 @ref{Installing a library}
7799 Specifying the switch @option{^-I-^/NOCURRENT_DIRECTORY^}
7800 inhibits the use of the directory
7801 containing the source file named in the command line. You can still
7802 have this directory on your search path, but in this case it must be
7803 explicitly requested with a @option{^-I^/SOURCE_SEARCH^} switch.
7805 Specifying the switch @option{-nostdinc}
7806 inhibits the search of the default location for the GNAT Run Time
7807 Library (RTL) source files.
7809 The compiler outputs its object files and ALI files in the current
7812 Caution: The object file can be redirected with the @option{-o} switch;
7813 however, @command{gcc} and @code{gnat1} have not been coordinated on this
7814 so the @file{ALI} file will not go to the right place. Therefore, you should
7815 avoid using the @option{-o} switch.
7819 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
7820 children make up the GNAT RTL, together with the simple @code{System.IO}
7821 package used in the @code{"Hello World"} example. The sources for these units
7822 are needed by the compiler and are kept together in one directory. Not
7823 all of the bodies are needed, but all of the sources are kept together
7824 anyway. In a normal installation, you need not specify these directory
7825 names when compiling or binding. Either the environment variables or
7826 the built-in defaults cause these files to be found.
7828 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
7829 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
7830 consisting of child units of @code{GNAT}. This is a collection of generally
7831 useful types, subprograms, etc. @xref{Top, GNAT Reference Manual, About
7832 This Guid, gnat_rm, GNAT Reference Manual}, for further details.
7834 Besides simplifying access to the RTL, a major use of search paths is
7835 in compiling sources from multiple directories. This can make
7836 development environments much more flexible.
7838 @node Order of Compilation Issues
7839 @section Order of Compilation Issues
7842 If, in our earlier example, there was a spec for the @code{hello}
7843 procedure, it would be contained in the file @file{hello.ads}; yet this
7844 file would not have to be explicitly compiled. This is the result of the
7845 model we chose to implement library management. Some of the consequences
7846 of this model are as follows:
7850 There is no point in compiling specs (except for package
7851 specs with no bodies) because these are compiled as needed by clients. If
7852 you attempt a useless compilation, you will receive an error message.
7853 It is also useless to compile subunits because they are compiled as needed
7857 There are no order of compilation requirements: performing a
7858 compilation never obsoletes anything. The only way you can obsolete
7859 something and require recompilations is to modify one of the
7860 source files on which it depends.
7863 There is no library as such, apart from the ALI files
7864 (@pxref{The Ada Library Information Files}, for information on the format
7865 of these files). For now we find it convenient to create separate ALI files,
7866 but eventually the information therein may be incorporated into the object
7870 When you compile a unit, the source files for the specs of all units
7871 that it @code{with}'s, all its subunits, and the bodies of any generics it
7872 instantiates must be available (reachable by the search-paths mechanism
7873 described above), or you will receive a fatal error message.
7880 The following are some typical Ada compilation command line examples:
7883 @item $ gcc -c xyz.adb
7884 Compile body in file @file{xyz.adb} with all default options.
7887 @item $ gcc -c -O2 -gnata xyz-def.adb
7890 @item $ GNAT COMPILE /OPTIMIZE=ALL -gnata xyz-def.adb
7893 Compile the child unit package in file @file{xyz-def.adb} with extensive
7894 optimizations, and pragma @code{Assert}/@code{Debug} statements
7897 @item $ gcc -c -gnatc abc-def.adb
7898 Compile the subunit in file @file{abc-def.adb} in semantic-checking-only
7902 @node Binding Using gnatbind
7903 @chapter Binding Using @code{gnatbind}
7907 * Running gnatbind::
7908 * Switches for gnatbind::
7909 * Command-Line Access::
7910 * Search Paths for gnatbind::
7911 * Examples of gnatbind Usage::
7915 This chapter describes the GNAT binder, @code{gnatbind}, which is used
7916 to bind compiled GNAT objects.
7918 Note: to invoke @code{gnatbind} with a project file, use the @code{gnat}
7919 driver (see @ref{The GNAT Driver and Project Files}).
7921 The @code{gnatbind} program performs four separate functions:
7925 Checks that a program is consistent, in accordance with the rules in
7926 Chapter 10 of the Ada Reference Manual. In particular, error
7927 messages are generated if a program uses inconsistent versions of a
7931 Checks that an acceptable order of elaboration exists for the program
7932 and issues an error message if it cannot find an order of elaboration
7933 that satisfies the rules in Chapter 10 of the Ada Language Manual.
7936 Generates a main program incorporating the given elaboration order.
7937 This program is a small Ada package (body and spec) that
7938 must be subsequently compiled
7939 using the GNAT compiler. The necessary compilation step is usually
7940 performed automatically by @command{gnatlink}. The two most important
7941 functions of this program
7942 are to call the elaboration routines of units in an appropriate order
7943 and to call the main program.
7946 Determines the set of object files required by the given main program.
7947 This information is output in the forms of comments in the generated program,
7948 to be read by the @command{gnatlink} utility used to link the Ada application.
7951 @node Running gnatbind
7952 @section Running @code{gnatbind}
7955 The form of the @code{gnatbind} command is
7958 @c $ gnatbind @ovar{switches} @var{mainprog}@r{[}.ali@r{]} @ovar{switches}
7959 @c Expanding @ovar macro inline (explanation in macro def comments)
7960 $ gnatbind @r{[}@var{switches}@r{]} @var{mainprog}@r{[}.ali@r{]} @r{[}@var{switches}@r{]}
7964 where @file{@var{mainprog}.adb} is the Ada file containing the main program
7965 unit body. @code{gnatbind} constructs an Ada
7966 package in two files whose names are
7967 @file{b~@var{mainprog}.ads}, and @file{b~@var{mainprog}.adb}.
7968 For example, if given the
7969 parameter @file{hello.ali}, for a main program contained in file
7970 @file{hello.adb}, the binder output files would be @file{b~hello.ads}
7971 and @file{b~hello.adb}.
7973 When doing consistency checking, the binder takes into consideration
7974 any source files it can locate. For example, if the binder determines
7975 that the given main program requires the package @code{Pack}, whose
7977 file is @file{pack.ali} and whose corresponding source spec file is
7978 @file{pack.ads}, it attempts to locate the source file @file{pack.ads}
7979 (using the same search path conventions as previously described for the
7980 @command{gcc} command). If it can locate this source file, it checks that
7982 or source checksums of the source and its references to in @file{ALI} files
7983 match. In other words, any @file{ALI} files that mentions this spec must have
7984 resulted from compiling this version of the source file (or in the case
7985 where the source checksums match, a version close enough that the
7986 difference does not matter).
7988 @cindex Source files, use by binder
7989 The effect of this consistency checking, which includes source files, is
7990 that the binder ensures that the program is consistent with the latest
7991 version of the source files that can be located at bind time. Editing a
7992 source file without compiling files that depend on the source file cause
7993 error messages to be generated by the binder.
7995 For example, suppose you have a main program @file{hello.adb} and a
7996 package @code{P}, from file @file{p.ads} and you perform the following
8001 Enter @code{gcc -c hello.adb} to compile the main program.
8004 Enter @code{gcc -c p.ads} to compile package @code{P}.
8007 Edit file @file{p.ads}.
8010 Enter @code{gnatbind hello}.
8014 At this point, the file @file{p.ali} contains an out-of-date time stamp
8015 because the file @file{p.ads} has been edited. The attempt at binding
8016 fails, and the binder generates the following error messages:
8019 error: "hello.adb" must be recompiled ("p.ads" has been modified)
8020 error: "p.ads" has been modified and must be recompiled
8024 Now both files must be recompiled as indicated, and then the bind can
8025 succeed, generating a main program. You need not normally be concerned
8026 with the contents of this file, but for reference purposes a sample
8027 binder output file is given in @ref{Example of Binder Output File}.
8029 In most normal usage, the default mode of @command{gnatbind} which is to
8030 generate the main package in Ada, as described in the previous section.
8031 In particular, this means that any Ada programmer can read and understand
8032 the generated main program. It can also be debugged just like any other
8033 Ada code provided the @option{^-g^/DEBUG^} switch is used for
8034 @command{gnatbind} and @command{gnatlink}.
8036 @node Switches for gnatbind
8037 @section Switches for @command{gnatbind}
8040 The following switches are available with @code{gnatbind}; details will
8041 be presented in subsequent sections.
8044 * Consistency-Checking Modes::
8045 * Binder Error Message Control::
8046 * Elaboration Control::
8048 * Binding with Non-Ada Main Programs::
8049 * Binding Programs with No Main Subprogram::
8056 @cindex @option{--version} @command{gnatbind}
8057 Display Copyright and version, then exit disregarding all other options.
8060 @cindex @option{--help} @command{gnatbind}
8061 If @option{--version} was not used, display usage, then exit disregarding
8065 @cindex @option{-a} @command{gnatbind}
8066 Indicates that, if supported by the platform, the adainit procedure should
8067 be treated as an initialisation routine by the linker (a constructor). This
8068 is intended to be used by the Project Manager to automatically initialize
8069 shared Stand-Alone Libraries.
8071 @item ^-aO^/OBJECT_SEARCH^
8072 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatbind})
8073 Specify directory to be searched for ALI files.
8075 @item ^-aI^/SOURCE_SEARCH^
8076 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8077 Specify directory to be searched for source file.
8079 @item ^-A^/ALI_LIST^@r{[=}@var{filename}@r{]}
8080 @cindex @option{^-A^/ALI_LIST^} (@command{gnatbind})
8081 Output ALI list (to standard output or to the named file).
8083 @item ^-b^/REPORT_ERRORS=BRIEF^
8084 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@command{gnatbind})
8085 Generate brief messages to @file{stderr} even if verbose mode set.
8087 @item ^-c^/NOOUTPUT^
8088 @cindex @option{^-c^/NOOUTPUT^} (@command{gnatbind})
8089 Check only, no generation of binder output file.
8091 @item ^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8092 @cindex @option{^-d^/DEFAULT_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}} (@command{gnatbind})
8093 This switch can be used to change the default task stack size value
8094 to a specified size @var{nn}, which is expressed in bytes by default, or
8095 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8097 In the absence of a @samp{@r{[}k@r{|}m@r{]}} suffix, this switch is equivalent,
8098 in effect, to completing all task specs with
8099 @smallexample @c ada
8100 pragma Storage_Size (nn);
8102 When they do not already have such a pragma.
8104 @item ^-D^/DEFAULT_SECONDARY_STACK_SIZE=^@var{nn}@r{[}k@r{|}m@r{]}
8105 @cindex @option{^-D^/DEFAULT_SECONDARY_STACK_SIZE=nnnnn^} (@command{gnatbind})
8106 This switch can be used to change the default secondary stack size value
8107 to a specified size @var{nn}, which is expressed in bytes by default, or
8108 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
8111 The secondary stack is used to deal with functions that return a variable
8112 sized result, for example a function returning an unconstrained
8113 String. There are two ways in which this secondary stack is allocated.
8115 For most targets, the secondary stack is growing on demand and is allocated
8116 as a chain of blocks in the heap. The -D option is not very
8117 relevant. It only give some control over the size of the allocated
8118 blocks (whose size is the minimum of the default secondary stack size value,
8119 and the actual size needed for the current allocation request).
8121 For certain targets, notably VxWorks 653,
8122 the secondary stack is allocated by carving off a fixed ratio chunk of the
8123 primary task stack. The -D option is used to define the
8124 size of the environment task's secondary stack.
8126 @item ^-e^/ELABORATION_DEPENDENCIES^
8127 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@command{gnatbind})
8128 Output complete list of elaboration-order dependencies.
8130 @item ^-E^/STORE_TRACEBACKS^
8131 @cindex @option{^-E^/STORE_TRACEBACKS^} (@command{gnatbind})
8132 Store tracebacks in exception occurrences when the target supports it.
8134 @c The following may get moved to an appendix
8135 This option is currently supported on the following targets:
8136 all x86 ports, Solaris, Windows, HP-UX, AIX, PowerPC VxWorks and Alpha VxWorks.
8138 See also the packages @code{GNAT.Traceback} and
8139 @code{GNAT.Traceback.Symbolic} for more information.
8141 Note that on x86 ports, you must not use @option{-fomit-frame-pointer}
8142 @command{gcc} option.
8145 @item ^-F^/FORCE_ELABS_FLAGS^
8146 @cindex @option{^-F^/FORCE_ELABS_FLAGS^} (@command{gnatbind})
8147 Force the checks of elaboration flags. @command{gnatbind} does not normally
8148 generate checks of elaboration flags for the main executable, except when
8149 a Stand-Alone Library is used. However, there are cases when this cannot be
8150 detected by gnatbind. An example is importing an interface of a Stand-Alone
8151 Library through a pragma Import and only specifying through a linker switch
8152 this Stand-Alone Library. This switch is used to guarantee that elaboration
8153 flag checks are generated.
8156 @cindex @option{^-h^/HELP^} (@command{gnatbind})
8157 Output usage (help) information
8160 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8161 Specify directory to be searched for source and ALI files.
8163 @item ^-I-^/NOCURRENT_DIRECTORY^
8164 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatbind})
8165 Do not look for sources in the current directory where @code{gnatbind} was
8166 invoked, and do not look for ALI files in the directory containing the
8167 ALI file named in the @code{gnatbind} command line.
8169 @item ^-l^/ORDER_OF_ELABORATION^
8170 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@command{gnatbind})
8171 Output chosen elaboration order.
8173 @item ^-L@var{xxx}^/BUILD_LIBRARY=@var{xxx}^
8174 @cindex @option{^-L^/BUILD_LIBRARY^} (@command{gnatbind})
8175 Bind the units for library building. In this case the adainit and
8176 adafinal procedures (@pxref{Binding with Non-Ada Main Programs})
8177 are renamed to ^@var{xxx}init^@var{XXX}INIT^ and
8178 ^@var{xxx}final^@var{XXX}FINAL^.
8179 Implies ^-n^/NOCOMPILE^.
8181 (@xref{GNAT and Libraries}, for more details.)
8184 On OpenVMS, these init and final procedures are exported in uppercase
8185 letters. For example if /BUILD_LIBRARY=toto is used, the exported name of
8186 the init procedure will be "TOTOINIT" and the exported name of the final
8187 procedure will be "TOTOFINAL".
8190 @item ^-Mxyz^/RENAME_MAIN=xyz^
8191 @cindex @option{^-M^/RENAME_MAIN^} (@command{gnatbind})
8192 Rename generated main program from main to xyz. This option is
8193 supported on cross environments only.
8195 @item ^-m^/ERROR_LIMIT=^@var{n}
8196 @cindex @option{^-m^/ERROR_LIMIT^} (@command{gnatbind})
8197 Limit number of detected errors or warnings to @var{n}, where @var{n} is
8198 in the range 1..999999. The default value if no switch is
8199 given is 9999. If the number of warnings reaches this limit, then a
8200 message is output and further warnings are suppressed, the bind
8201 continues in this case. If the number of errors reaches this
8202 limit, then a message is output and the bind is abandoned.
8203 A value of zero means that no limit is enforced. The equal
8207 Furthermore, under Windows, the sources pointed to by the libraries path
8208 set in the registry are not searched for.
8212 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8216 @cindex @option{-nostdinc} (@command{gnatbind})
8217 Do not look for sources in the system default directory.
8220 @cindex @option{-nostdlib} (@command{gnatbind})
8221 Do not look for library files in the system default directory.
8223 @item --RTS=@var{rts-path}
8224 @cindex @option{--RTS} (@code{gnatbind})
8225 Specifies the default location of the runtime library. Same meaning as the
8226 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
8228 @item ^-o ^/OUTPUT=^@var{file}
8229 @cindex @option{^-o ^/OUTPUT^} (@command{gnatbind})
8230 Name the output file @var{file} (default is @file{b~@var{xxx}.adb}).
8231 Note that if this option is used, then linking must be done manually,
8232 gnatlink cannot be used.
8234 @item ^-O^/OBJECT_LIST^@r{[=}@var{filename}@r{]}
8235 @cindex @option{^-O^/OBJECT_LIST^} (@command{gnatbind})
8236 Output object list (to standard output or to the named file).
8238 @item ^-p^/PESSIMISTIC_ELABORATION^
8239 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@command{gnatbind})
8240 Pessimistic (worst-case) elaboration order
8243 @cindex @option{^-R^-R^} (@command{gnatbind})
8244 Output closure source list.
8246 @item ^-s^/READ_SOURCES=ALL^
8247 @cindex @option{^-s^/READ_SOURCES=ALL^} (@command{gnatbind})
8248 Require all source files to be present.
8250 @item ^-S@var{xxx}^/INITIALIZE_SCALARS=@var{xxx}^
8251 @cindex @option{^-S^/INITIALIZE_SCALARS^} (@command{gnatbind})
8252 Specifies the value to be used when detecting uninitialized scalar
8253 objects with pragma Initialize_Scalars.
8254 The @var{xxx} ^string specified with the switch^option^ may be either
8256 @item ``@option{^in^INVALID^}'' requesting an invalid value where possible
8257 @item ``@option{^lo^LOW^}'' for the lowest possible value
8258 @item ``@option{^hi^HIGH^}'' for the highest possible value
8259 @item ``@option{@var{xx}}'' for a value consisting of repeated bytes with the
8260 value @code{16#@var{xx}#} (i.e., @var{xx} is a string of two hexadecimal digits).
8263 In addition, you can specify @option{-Sev} to indicate that the value is
8264 to be set at run time. In this case, the program will look for an environment
8265 @cindex GNAT_INIT_SCALARS
8266 variable of the form @env{GNAT_INIT_SCALARS=@var{xx}}, where @var{xx} is one
8267 of @option{in/lo/hi/@var{xx}} with the same meanings as above.
8268 If no environment variable is found, or if it does not have a valid value,
8269 then the default is @option{in} (invalid values).
8273 @cindex @option{-static} (@code{gnatbind})
8274 Link against a static GNAT run time.
8277 @cindex @option{-shared} (@code{gnatbind})
8278 Link against a shared GNAT run time when available.
8281 @item ^-t^/NOTIME_STAMP_CHECK^
8282 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8283 Tolerate time stamp and other consistency errors
8285 @item ^-T@var{n}^/TIME_SLICE=@var{n}^
8286 @cindex @option{^-T^/TIME_SLICE^} (@code{gnatbind})
8287 Set the time slice value to @var{n} milliseconds. If the system supports
8288 the specification of a specific time slice value, then the indicated value
8289 is used. If the system does not support specific time slice values, but
8290 does support some general notion of round-robin scheduling, then any
8291 nonzero value will activate round-robin scheduling.
8293 A value of zero is treated specially. It turns off time
8294 slicing, and in addition, indicates to the tasking run time that the
8295 semantics should match as closely as possible the Annex D
8296 requirements of the Ada RM, and in particular sets the default
8297 scheduling policy to @code{FIFO_Within_Priorities}.
8299 @item ^-u@var{n}^/DYNAMIC_STACK_USAGE=@var{n}^
8300 @cindex @option{^-u^/DYNAMIC_STACK_USAGE^} (@code{gnatbind})
8301 Enable dynamic stack usage, with @var{n} results stored and displayed
8302 at program termination. A result is generated when a task
8303 terminates. Results that can't be stored are displayed on the fly, at
8304 task termination. This option is currently not supported on Itanium
8305 platforms. (See @ref{Dynamic Stack Usage Analysis} for details.)
8307 @item ^-v^/REPORT_ERRORS=VERBOSE^
8308 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8309 Verbose mode. Write error messages, header, summary output to
8314 @cindex @option{-w} (@code{gnatbind})
8315 Warning mode (@var{x}=s/e for suppress/treat as error)
8319 @item /WARNINGS=NORMAL
8320 @cindex @option{/WARNINGS} (@code{gnatbind})
8321 Normal warnings mode. Warnings are issued but ignored
8323 @item /WARNINGS=SUPPRESS
8324 @cindex @option{/WARNINGS} (@code{gnatbind})
8325 All warning messages are suppressed
8327 @item /WARNINGS=ERROR
8328 @cindex @option{/WARNINGS} (@code{gnatbind})
8329 Warning messages are treated as fatal errors
8332 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8333 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8334 Override default wide character encoding for standard Text_IO files.
8336 @item ^-x^/READ_SOURCES=NONE^
8337 @cindex @option{^-x^/READ_SOURCES^} (@code{gnatbind})
8338 Exclude source files (check object consistency only).
8341 @item /READ_SOURCES=AVAILABLE
8342 @cindex @option{/READ_SOURCES} (@code{gnatbind})
8343 Default mode, in which sources are checked for consistency only if
8347 @item ^-y^/ENABLE_LEAP_SECONDS^
8348 @cindex @option{^-y^/ENABLE_LEAP_SECONDS^} (@code{gnatbind})
8349 Enable leap seconds support in @code{Ada.Calendar} and its children.
8351 @item ^-z^/ZERO_MAIN^
8352 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8358 You may obtain this listing of switches by running @code{gnatbind} with
8362 @node Consistency-Checking Modes
8363 @subsection Consistency-Checking Modes
8366 As described earlier, by default @code{gnatbind} checks
8367 that object files are consistent with one another and are consistent
8368 with any source files it can locate. The following switches control binder
8373 @item ^-s^/READ_SOURCES=ALL^
8374 @cindex @option{^-s^/READ_SOURCES=ALL^} (@code{gnatbind})
8375 Require source files to be present. In this mode, the binder must be
8376 able to locate all source files that are referenced, in order to check
8377 their consistency. In normal mode, if a source file cannot be located it
8378 is simply ignored. If you specify this switch, a missing source
8381 @item ^-Wx^/WIDE_CHARACTER_ENCODING=^@var{e}
8382 @cindex @option{^-Wx^/WIDE_CHARACTER_ENCODING^} (@code{gnatbind})
8383 Override default wide character encoding for standard Text_IO files.
8384 Normally the default wide character encoding method used for standard
8385 [Wide_[Wide_]]Text_IO files is taken from the encoding specified for
8386 the main source input (see description of switch
8387 @option{^-gnatWx^/WIDE_CHARACTER_ENCODING^} for the compiler). The
8388 use of this switch for the binder (which has the same set of
8389 possible arguments) overrides this default as specified.
8391 @item ^-x^/READ_SOURCES=NONE^
8392 @cindex @option{^-x^/READ_SOURCES=NONE^} (@code{gnatbind})
8393 Exclude source files. In this mode, the binder only checks that ALI
8394 files are consistent with one another. Source files are not accessed.
8395 The binder runs faster in this mode, and there is still a guarantee that
8396 the resulting program is self-consistent.
8397 If a source file has been edited since it was last compiled, and you
8398 specify this switch, the binder will not detect that the object
8399 file is out of date with respect to the source file. Note that this is the
8400 mode that is automatically used by @command{gnatmake} because in this
8401 case the checking against sources has already been performed by
8402 @command{gnatmake} in the course of compilation (i.e.@: before binding).
8405 @item /READ_SOURCES=AVAILABLE
8406 @cindex @code{/READ_SOURCES=AVAILABLE} (@code{gnatbind})
8407 This is the default mode in which source files are checked if they are
8408 available, and ignored if they are not available.
8412 @node Binder Error Message Control
8413 @subsection Binder Error Message Control
8416 The following switches provide control over the generation of error
8417 messages from the binder:
8421 @item ^-v^/REPORT_ERRORS=VERBOSE^
8422 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
8423 Verbose mode. In the normal mode, brief error messages are generated to
8424 @file{stderr}. If this switch is present, a header is written
8425 to @file{stdout} and any error messages are directed to @file{stdout}.
8426 All that is written to @file{stderr} is a brief summary message.
8428 @item ^-b^/REPORT_ERRORS=BRIEF^
8429 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@code{gnatbind})
8430 Generate brief error messages to @file{stderr} even if verbose mode is
8431 specified. This is relevant only when used with the
8432 @option{^-v^/REPORT_ERRORS=VERBOSE^} switch.
8436 @cindex @option{-m} (@code{gnatbind})
8437 Limits the number of error messages to @var{n}, a decimal integer in the
8438 range 1-999. The binder terminates immediately if this limit is reached.
8441 @cindex @option{-M} (@code{gnatbind})
8442 Renames the generated main program from @code{main} to @code{xxx}.
8443 This is useful in the case of some cross-building environments, where
8444 the actual main program is separate from the one generated
8448 @item ^-ws^/WARNINGS=SUPPRESS^
8449 @cindex @option{^-ws^/WARNINGS=SUPPRESS^} (@code{gnatbind})
8451 Suppress all warning messages.
8453 @item ^-we^/WARNINGS=ERROR^
8454 @cindex @option{^-we^/WARNINGS=ERROR^} (@code{gnatbind})
8455 Treat any warning messages as fatal errors.
8458 @item /WARNINGS=NORMAL
8459 Standard mode with warnings generated, but warnings do not get treated
8463 @item ^-t^/NOTIME_STAMP_CHECK^
8464 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
8465 @cindex Time stamp checks, in binder
8466 @cindex Binder consistency checks
8467 @cindex Consistency checks, in binder
8468 The binder performs a number of consistency checks including:
8472 Check that time stamps of a given source unit are consistent
8474 Check that checksums of a given source unit are consistent
8476 Check that consistent versions of @code{GNAT} were used for compilation
8478 Check consistency of configuration pragmas as required
8482 Normally failure of such checks, in accordance with the consistency
8483 requirements of the Ada Reference Manual, causes error messages to be
8484 generated which abort the binder and prevent the output of a binder
8485 file and subsequent link to obtain an executable.
8487 The @option{^-t^/NOTIME_STAMP_CHECK^} switch converts these error messages
8488 into warnings, so that
8489 binding and linking can continue to completion even in the presence of such
8490 errors. The result may be a failed link (due to missing symbols), or a
8491 non-functional executable which has undefined semantics.
8492 @emph{This means that
8493 @option{^-t^/NOTIME_STAMP_CHECK^} should be used only in unusual situations,
8497 @node Elaboration Control
8498 @subsection Elaboration Control
8501 The following switches provide additional control over the elaboration
8502 order. For full details see @ref{Elaboration Order Handling in GNAT}.
8505 @item ^-p^/PESSIMISTIC_ELABORATION^
8506 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@code{gnatbind})
8507 Normally the binder attempts to choose an elaboration order that is
8508 likely to minimize the likelihood of an elaboration order error resulting
8509 in raising a @code{Program_Error} exception. This switch reverses the
8510 action of the binder, and requests that it deliberately choose an order
8511 that is likely to maximize the likelihood of an elaboration error.
8512 This is useful in ensuring portability and avoiding dependence on
8513 accidental fortuitous elaboration ordering.
8515 Normally it only makes sense to use the @option{^-p^/PESSIMISTIC_ELABORATION^}
8517 elaboration checking is used (@option{-gnatE} switch used for compilation).
8518 This is because in the default static elaboration mode, all necessary
8519 @code{Elaborate} and @code{Elaborate_All} pragmas are implicitly inserted.
8520 These implicit pragmas are still respected by the binder in
8521 @option{^-p^/PESSIMISTIC_ELABORATION^} mode, so a
8522 safe elaboration order is assured.
8525 @node Output Control
8526 @subsection Output Control
8529 The following switches allow additional control over the output
8530 generated by the binder.
8535 @item ^-c^/NOOUTPUT^
8536 @cindex @option{^-c^/NOOUTPUT^} (@code{gnatbind})
8537 Check only. Do not generate the binder output file. In this mode the
8538 binder performs all error checks but does not generate an output file.
8540 @item ^-e^/ELABORATION_DEPENDENCIES^
8541 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@code{gnatbind})
8542 Output complete list of elaboration-order dependencies, showing the
8543 reason for each dependency. This output can be rather extensive but may
8544 be useful in diagnosing problems with elaboration order. The output is
8545 written to @file{stdout}.
8548 @cindex @option{^-h^/HELP^} (@code{gnatbind})
8549 Output usage information. The output is written to @file{stdout}.
8551 @item ^-K^/LINKER_OPTION_LIST^
8552 @cindex @option{^-K^/LINKER_OPTION_LIST^} (@code{gnatbind})
8553 Output linker options to @file{stdout}. Includes library search paths,
8554 contents of pragmas Ident and Linker_Options, and libraries added
8557 @item ^-l^/ORDER_OF_ELABORATION^
8558 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@code{gnatbind})
8559 Output chosen elaboration order. The output is written to @file{stdout}.
8561 @item ^-O^/OBJECT_LIST^
8562 @cindex @option{^-O^/OBJECT_LIST^} (@code{gnatbind})
8563 Output full names of all the object files that must be linked to provide
8564 the Ada component of the program. The output is written to @file{stdout}.
8565 This list includes the files explicitly supplied and referenced by the user
8566 as well as implicitly referenced run-time unit files. The latter are
8567 omitted if the corresponding units reside in shared libraries. The
8568 directory names for the run-time units depend on the system configuration.
8570 @item ^-o ^/OUTPUT=^@var{file}
8571 @cindex @option{^-o^/OUTPUT^} (@code{gnatbind})
8572 Set name of output file to @var{file} instead of the normal
8573 @file{b~@var{mainprog}.adb} default. Note that @var{file} denote the Ada
8574 binder generated body filename.
8575 Note that if this option is used, then linking must be done manually.
8576 It is not possible to use gnatlink in this case, since it cannot locate
8579 @item ^-r^/RESTRICTION_LIST^
8580 @cindex @option{^-r^/RESTRICTION_LIST^} (@code{gnatbind})
8581 Generate list of @code{pragma Restrictions} that could be applied to
8582 the current unit. This is useful for code audit purposes, and also may
8583 be used to improve code generation in some cases.
8587 @node Binding with Non-Ada Main Programs
8588 @subsection Binding with Non-Ada Main Programs
8591 In our description so far we have assumed that the main
8592 program is in Ada, and that the task of the binder is to generate a
8593 corresponding function @code{main} that invokes this Ada main
8594 program. GNAT also supports the building of executable programs where
8595 the main program is not in Ada, but some of the called routines are
8596 written in Ada and compiled using GNAT (@pxref{Mixed Language Programming}).
8597 The following switch is used in this situation:
8601 @cindex @option{^-n^/NOMAIN^} (@code{gnatbind})
8602 No main program. The main program is not in Ada.
8606 In this case, most of the functions of the binder are still required,
8607 but instead of generating a main program, the binder generates a file
8608 containing the following callable routines:
8613 You must call this routine to initialize the Ada part of the program by
8614 calling the necessary elaboration routines. A call to @code{adainit} is
8615 required before the first call to an Ada subprogram.
8617 Note that it is assumed that the basic execution environment must be setup
8618 to be appropriate for Ada execution at the point where the first Ada
8619 subprogram is called. In particular, if the Ada code will do any
8620 floating-point operations, then the FPU must be setup in an appropriate
8621 manner. For the case of the x86, for example, full precision mode is
8622 required. The procedure GNAT.Float_Control.Reset may be used to ensure
8623 that the FPU is in the right state.
8627 You must call this routine to perform any library-level finalization
8628 required by the Ada subprograms. A call to @code{adafinal} is required
8629 after the last call to an Ada subprogram, and before the program
8634 If the @option{^-n^/NOMAIN^} switch
8635 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
8636 @cindex Binder, multiple input files
8637 is given, more than one ALI file may appear on
8638 the command line for @code{gnatbind}. The normal @dfn{closure}
8639 calculation is performed for each of the specified units. Calculating
8640 the closure means finding out the set of units involved by tracing
8641 @code{with} references. The reason it is necessary to be able to
8642 specify more than one ALI file is that a given program may invoke two or
8643 more quite separate groups of Ada units.
8645 The binder takes the name of its output file from the last specified ALI
8646 file, unless overridden by the use of the @option{^-o file^/OUTPUT=file^}.
8647 @cindex @option{^-o^/OUTPUT^} (@command{gnatbind})
8648 The output is an Ada unit in source form that can be compiled with GNAT.
8649 This compilation occurs automatically as part of the @command{gnatlink}
8652 Currently the GNAT run time requires a FPU using 80 bits mode
8653 precision. Under targets where this is not the default it is required to
8654 call GNAT.Float_Control.Reset before using floating point numbers (this
8655 include float computation, float input and output) in the Ada code. A
8656 side effect is that this could be the wrong mode for the foreign code
8657 where floating point computation could be broken after this call.
8659 @node Binding Programs with No Main Subprogram
8660 @subsection Binding Programs with No Main Subprogram
8663 It is possible to have an Ada program which does not have a main
8664 subprogram. This program will call the elaboration routines of all the
8665 packages, then the finalization routines.
8667 The following switch is used to bind programs organized in this manner:
8670 @item ^-z^/ZERO_MAIN^
8671 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
8672 Normally the binder checks that the unit name given on the command line
8673 corresponds to a suitable main subprogram. When this switch is used,
8674 a list of ALI files can be given, and the execution of the program
8675 consists of elaboration of these units in an appropriate order. Note
8676 that the default wide character encoding method for standard Text_IO
8677 files is always set to Brackets if this switch is set (you can use
8679 @option{^-Wx^WIDE_CHARACTER_ENCODING^} to override this default).
8682 @node Command-Line Access
8683 @section Command-Line Access
8686 The package @code{Ada.Command_Line} provides access to the command-line
8687 arguments and program name. In order for this interface to operate
8688 correctly, the two variables
8700 are declared in one of the GNAT library routines. These variables must
8701 be set from the actual @code{argc} and @code{argv} values passed to the
8702 main program. With no @option{^n^/NOMAIN^} present, @code{gnatbind}
8703 generates the C main program to automatically set these variables.
8704 If the @option{^n^/NOMAIN^} switch is used, there is no automatic way to
8705 set these variables. If they are not set, the procedures in
8706 @code{Ada.Command_Line} will not be available, and any attempt to use
8707 them will raise @code{Constraint_Error}. If command line access is
8708 required, your main program must set @code{gnat_argc} and
8709 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
8712 @node Search Paths for gnatbind
8713 @section Search Paths for @code{gnatbind}
8716 The binder takes the name of an ALI file as its argument and needs to
8717 locate source files as well as other ALI files to verify object consistency.
8719 For source files, it follows exactly the same search rules as @command{gcc}
8720 (@pxref{Search Paths and the Run-Time Library (RTL)}). For ALI files the
8721 directories searched are:
8725 The directory containing the ALI file named in the command line, unless
8726 the switch @option{^-I-^/NOCURRENT_DIRECTORY^} is specified.
8729 All directories specified by @option{^-I^/SEARCH^}
8730 switches on the @code{gnatbind}
8731 command line, in the order given.
8734 @findex ADA_PRJ_OBJECTS_FILE
8735 Each of the directories listed in the text file whose name is given
8736 by the @env{ADA_PRJ_OBJECTS_FILE} ^environment variable^logical name^.
8739 @env{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
8740 driver when project files are used. It should not normally be set
8744 @findex ADA_OBJECTS_PATH
8745 Each of the directories listed in the value of the
8746 @env{ADA_OBJECTS_PATH} ^environment variable^logical name^.
8748 Construct this value
8749 exactly as the @env{PATH} environment variable: a list of directory
8750 names separated by colons (semicolons when working with the NT version
8754 Normally, define this value as a logical name containing a comma separated
8755 list of directory names.
8757 This variable can also be defined by means of an environment string
8758 (an argument to the HP C exec* set of functions).
8762 DEFINE ANOTHER_PATH FOO:[BAG]
8763 DEFINE ADA_OBJECTS_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
8766 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
8767 first, followed by the standard Ada
8768 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB].
8769 If this is not redefined, the user will obtain the HP Ada 83 IO packages
8770 (Text_IO, Sequential_IO, etc)
8771 instead of the standard Ada packages. Thus, in order to get the standard Ada
8772 packages by default, ADA_OBJECTS_PATH must be redefined.
8776 The content of the @file{ada_object_path} file which is part of the GNAT
8777 installation tree and is used to store standard libraries such as the
8778 GNAT Run Time Library (RTL) unless the switch @option{-nostdlib} is
8781 @ref{Installing a library}
8786 In the binder the switch @option{^-I^/SEARCH^}
8787 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8788 is used to specify both source and
8789 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
8790 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8791 instead if you want to specify
8792 source paths only, and @option{^-aO^/LIBRARY_SEARCH^}
8793 @cindex @option{^-aO^/LIBRARY_SEARCH^} (@command{gnatbind})
8794 if you want to specify library paths
8795 only. This means that for the binder
8796 @option{^-I^/SEARCH=^}@var{dir} is equivalent to
8797 @option{^-aI^/SOURCE_SEARCH=^}@var{dir}
8798 @option{^-aO^/OBJECT_SEARCH=^}@var{dir}.
8799 The binder generates the bind file (a C language source file) in the
8800 current working directory.
8806 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
8807 children make up the GNAT Run-Time Library, together with the package
8808 GNAT and its children, which contain a set of useful additional
8809 library functions provided by GNAT. The sources for these units are
8810 needed by the compiler and are kept together in one directory. The ALI
8811 files and object files generated by compiling the RTL are needed by the
8812 binder and the linker and are kept together in one directory, typically
8813 different from the directory containing the sources. In a normal
8814 installation, you need not specify these directory names when compiling
8815 or binding. Either the environment variables or the built-in defaults
8816 cause these files to be found.
8818 Besides simplifying access to the RTL, a major use of search paths is
8819 in compiling sources from multiple directories. This can make
8820 development environments much more flexible.
8822 @node Examples of gnatbind Usage
8823 @section Examples of @code{gnatbind} Usage
8826 This section contains a number of examples of using the GNAT binding
8827 utility @code{gnatbind}.
8830 @item gnatbind hello
8831 The main program @code{Hello} (source program in @file{hello.adb}) is
8832 bound using the standard switch settings. The generated main program is
8833 @file{b~hello.adb}. This is the normal, default use of the binder.
8836 @item gnatbind hello -o mainprog.adb
8839 @item gnatbind HELLO.ALI /OUTPUT=Mainprog.ADB
8841 The main program @code{Hello} (source program in @file{hello.adb}) is
8842 bound using the standard switch settings. The generated main program is
8843 @file{mainprog.adb} with the associated spec in
8844 @file{mainprog.ads}. Note that you must specify the body here not the
8845 spec. Note that if this option is used, then linking must be done manually,
8846 since gnatlink will not be able to find the generated file.
8849 @c ------------------------------------
8850 @node Linking Using gnatlink
8851 @chapter Linking Using @command{gnatlink}
8852 @c ------------------------------------
8856 This chapter discusses @command{gnatlink}, a tool that links
8857 an Ada program and builds an executable file. This utility
8858 invokes the system linker ^(via the @command{gcc} command)^^
8859 with a correct list of object files and library references.
8860 @command{gnatlink} automatically determines the list of files and
8861 references for the Ada part of a program. It uses the binder file
8862 generated by the @command{gnatbind} to determine this list.
8864 Note: to invoke @code{gnatlink} with a project file, use the @code{gnat}
8865 driver (see @ref{The GNAT Driver and Project Files}).
8868 * Running gnatlink::
8869 * Switches for gnatlink::
8872 @node Running gnatlink
8873 @section Running @command{gnatlink}
8876 The form of the @command{gnatlink} command is
8879 @c $ gnatlink @ovar{switches} @var{mainprog}@r{[}.ali@r{]}
8880 @c @ovar{non-Ada objects} @ovar{linker options}
8881 @c Expanding @ovar macro inline (explanation in macro def comments)
8882 $ gnatlink @r{[}@var{switches}@r{]} @var{mainprog}@r{[}.ali@r{]}
8883 @r{[}@var{non-Ada objects}@r{]} @r{[}@var{linker options}@r{]}
8888 The arguments of @command{gnatlink} (switches, main @file{ALI} file,
8890 or linker options) may be in any order, provided that no non-Ada object may
8891 be mistaken for a main @file{ALI} file.
8892 Any file name @file{F} without the @file{.ali}
8893 extension will be taken as the main @file{ALI} file if a file exists
8894 whose name is the concatenation of @file{F} and @file{.ali}.
8897 @file{@var{mainprog}.ali} references the ALI file of the main program.
8898 The @file{.ali} extension of this file can be omitted. From this
8899 reference, @command{gnatlink} locates the corresponding binder file
8900 @file{b~@var{mainprog}.adb} and, using the information in this file along
8901 with the list of non-Ada objects and linker options, constructs a
8902 linker command file to create the executable.
8904 The arguments other than the @command{gnatlink} switches and the main
8905 @file{ALI} file are passed to the linker uninterpreted.
8906 They typically include the names of
8907 object files for units written in other languages than Ada and any library
8908 references required to resolve references in any of these foreign language
8909 units, or in @code{Import} pragmas in any Ada units.
8911 @var{linker options} is an optional list of linker specific
8913 The default linker called by gnatlink is @command{gcc} which in
8914 turn calls the appropriate system linker.
8915 Standard options for the linker such as @option{-lmy_lib} or
8916 @option{-Ldir} can be added as is.
8917 For options that are not recognized by
8918 @command{gcc} as linker options, use the @command{gcc} switches
8919 @option{-Xlinker} or @option{-Wl,}.
8920 Refer to the GCC documentation for
8921 details. Here is an example showing how to generate a linker map:
8924 $ ^gnatlink my_prog -Wl,-Map,MAPFILE^GNAT LINK my_prog.ali /MAP^
8927 Using @var{linker options} it is possible to set the program stack and
8930 See @ref{Setting Stack Size from gnatlink} and
8931 @ref{Setting Heap Size from gnatlink}.
8934 @command{gnatlink} determines the list of objects required by the Ada
8935 program and prepends them to the list of objects passed to the linker.
8936 @command{gnatlink} also gathers any arguments set by the use of
8937 @code{pragma Linker_Options} and adds them to the list of arguments
8938 presented to the linker.
8941 @command{gnatlink} accepts the following types of extra files on the command
8942 line: objects (@file{.OBJ}), libraries (@file{.OLB}), sharable images
8943 (@file{.EXE}), and options files (@file{.OPT}). These are recognized and
8944 handled according to their extension.
8947 @node Switches for gnatlink
8948 @section Switches for @command{gnatlink}
8951 The following switches are available with the @command{gnatlink} utility:
8957 @cindex @option{--version} @command{gnatlink}
8958 Display Copyright and version, then exit disregarding all other options.
8961 @cindex @option{--help} @command{gnatlink}
8962 If @option{--version} was not used, display usage, then exit disregarding
8965 @item ^-f^/FORCE_OBJECT_FILE_LIST^
8966 @cindex Command line length
8967 @cindex @option{^-f^/FORCE_OBJECT_FILE_LIST^} (@command{gnatlink})
8968 On some targets, the command line length is limited, and @command{gnatlink}
8969 will generate a separate file for the linker if the list of object files
8971 The @option{^-f^/FORCE_OBJECT_FILE_LIST^} switch forces this file
8972 to be generated even if
8973 the limit is not exceeded. This is useful in some cases to deal with
8974 special situations where the command line length is exceeded.
8977 @cindex Debugging information, including
8978 @cindex @option{^-g^/DEBUG^} (@command{gnatlink})
8979 The option to include debugging information causes the Ada bind file (in
8980 other words, @file{b~@var{mainprog}.adb}) to be compiled with
8981 @option{^-g^/DEBUG^}.
8982 In addition, the binder does not delete the @file{b~@var{mainprog}.adb},
8983 @file{b~@var{mainprog}.o} and @file{b~@var{mainprog}.ali} files.
8984 Without @option{^-g^/DEBUG^}, the binder removes these files by
8985 default. The same procedure apply if a C bind file was generated using
8986 @option{^-C^/BIND_FILE=C^} @code{gnatbind} option, in this case the filenames
8987 are @file{b_@var{mainprog}.c} and @file{b_@var{mainprog}.o}.
8989 @item ^-n^/NOCOMPILE^
8990 @cindex @option{^-n^/NOCOMPILE^} (@command{gnatlink})
8991 Do not compile the file generated by the binder. This may be used when
8992 a link is rerun with different options, but there is no need to recompile
8996 @cindex @option{^-v^/VERBOSE^} (@command{gnatlink})
8997 Causes additional information to be output, including a full list of the
8998 included object files. This switch option is most useful when you want
8999 to see what set of object files are being used in the link step.
9001 @item ^-v -v^/VERBOSE/VERBOSE^
9002 @cindex @option{^-v -v^/VERBOSE/VERBOSE^} (@command{gnatlink})
9003 Very verbose mode. Requests that the compiler operate in verbose mode when
9004 it compiles the binder file, and that the system linker run in verbose mode.
9006 @item ^-o ^/EXECUTABLE=^@var{exec-name}
9007 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatlink})
9008 @var{exec-name} specifies an alternate name for the generated
9009 executable program. If this switch is omitted, the executable has the same
9010 name as the main unit. For example, @code{gnatlink try.ali} creates
9011 an executable called @file{^try^TRY.EXE^}.
9014 @item -b @var{target}
9015 @cindex @option{-b} (@command{gnatlink})
9016 Compile your program to run on @var{target}, which is the name of a
9017 system configuration. You must have a GNAT cross-compiler built if
9018 @var{target} is not the same as your host system.
9021 @cindex @option{-B} (@command{gnatlink})
9022 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
9023 from @var{dir} instead of the default location. Only use this switch
9024 when multiple versions of the GNAT compiler are available.
9025 @xref{Directory Options,,, gcc, The GNU Compiler Collection},
9026 for further details. You would normally use the @option{-b} or
9027 @option{-V} switch instead.
9029 @item --GCC=@var{compiler_name}
9030 @cindex @option{--GCC=compiler_name} (@command{gnatlink})
9031 Program used for compiling the binder file. The default is
9032 @command{gcc}. You need to use quotes around @var{compiler_name} if
9033 @code{compiler_name} contains spaces or other separator characters.
9034 As an example @option{--GCC="foo -x -y"} will instruct @command{gnatlink} to
9035 use @code{foo -x -y} as your compiler. Note that switch @option{-c} is always
9036 inserted after your command name. Thus in the above example the compiler
9037 command that will be used by @command{gnatlink} will be @code{foo -c -x -y}.
9038 A limitation of this syntax is that the name and path name of the executable
9039 itself must not include any embedded spaces. If the compiler executable is
9040 different from the default one (gcc or <prefix>-gcc), then the back-end
9041 switches in the ALI file are not used to compile the binder generated source.
9042 For example, this is the case with @option{--GCC="foo -x -y"}. But the back end
9043 switches will be used for @option{--GCC="gcc -gnatv"}. If several
9044 @option{--GCC=compiler_name} are used, only the last @var{compiler_name}
9045 is taken into account. However, all the additional switches are also taken
9047 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9048 @option{--GCC="bar -x -y -z -t"}.
9050 @item --LINK=@var{name}
9051 @cindex @option{--LINK=} (@command{gnatlink})
9052 @var{name} is the name of the linker to be invoked. This is especially
9053 useful in mixed language programs since languages such as C++ require
9054 their own linker to be used. When this switch is omitted, the default
9055 name for the linker is @command{gcc}. When this switch is used, the
9056 specified linker is called instead of @command{gcc} with exactly the same
9057 parameters that would have been passed to @command{gcc} so if the desired
9058 linker requires different parameters it is necessary to use a wrapper
9059 script that massages the parameters before invoking the real linker. It
9060 may be useful to control the exact invocation by using the verbose
9066 @item /DEBUG=TRACEBACK
9067 @cindex @code{/DEBUG=TRACEBACK} (@command{gnatlink})
9068 This qualifier causes sufficient information to be included in the
9069 executable file to allow a traceback, but does not include the full
9070 symbol information needed by the debugger.
9072 @item /IDENTIFICATION="<string>"
9073 @code{"<string>"} specifies the string to be stored in the image file
9074 identification field in the image header.
9075 It overrides any pragma @code{Ident} specified string.
9077 @item /NOINHIBIT-EXEC
9078 Generate the executable file even if there are linker warnings.
9080 @item /NOSTART_FILES
9081 Don't link in the object file containing the ``main'' transfer address.
9082 Used when linking with a foreign language main program compiled with an
9086 Prefer linking with object libraries over sharable images, even without
9092 @node The GNAT Make Program gnatmake
9093 @chapter The GNAT Make Program @command{gnatmake}
9097 * Running gnatmake::
9098 * Switches for gnatmake::
9099 * Mode Switches for gnatmake::
9100 * Notes on the Command Line::
9101 * How gnatmake Works::
9102 * Examples of gnatmake Usage::
9105 A typical development cycle when working on an Ada program consists of
9106 the following steps:
9110 Edit some sources to fix bugs.
9116 Compile all sources affected.
9126 The third step can be tricky, because not only do the modified files
9127 @cindex Dependency rules
9128 have to be compiled, but any files depending on these files must also be
9129 recompiled. The dependency rules in Ada can be quite complex, especially
9130 in the presence of overloading, @code{use} clauses, generics and inlined
9133 @command{gnatmake} automatically takes care of the third and fourth steps
9134 of this process. It determines which sources need to be compiled,
9135 compiles them, and binds and links the resulting object files.
9137 Unlike some other Ada make programs, the dependencies are always
9138 accurately recomputed from the new sources. The source based approach of
9139 the GNAT compilation model makes this possible. This means that if
9140 changes to the source program cause corresponding changes in
9141 dependencies, they will always be tracked exactly correctly by
9144 @node Running gnatmake
9145 @section Running @command{gnatmake}
9148 The usual form of the @command{gnatmake} command is
9151 @c $ gnatmake @ovar{switches} @var{file_name}
9152 @c @ovar{file_names} @ovar{mode_switches}
9153 @c Expanding @ovar macro inline (explanation in macro def comments)
9154 $ gnatmake @r{[}@var{switches}@r{]} @var{file_name}
9155 @r{[}@var{file_names}@r{]} @r{[}@var{mode_switches}@r{]}
9159 The only required argument is one @var{file_name}, which specifies
9160 a compilation unit that is a main program. Several @var{file_names} can be
9161 specified: this will result in several executables being built.
9162 If @code{switches} are present, they can be placed before the first
9163 @var{file_name}, between @var{file_names} or after the last @var{file_name}.
9164 If @var{mode_switches} are present, they must always be placed after
9165 the last @var{file_name} and all @code{switches}.
9167 If you are using standard file extensions (@file{.adb} and @file{.ads}), then the
9168 extension may be omitted from the @var{file_name} arguments. However, if
9169 you are using non-standard extensions, then it is required that the
9170 extension be given. A relative or absolute directory path can be
9171 specified in a @var{file_name}, in which case, the input source file will
9172 be searched for in the specified directory only. Otherwise, the input
9173 source file will first be searched in the directory where
9174 @command{gnatmake} was invoked and if it is not found, it will be search on
9175 the source path of the compiler as described in
9176 @ref{Search Paths and the Run-Time Library (RTL)}.
9178 All @command{gnatmake} output (except when you specify
9179 @option{^-M^/DEPENDENCIES_LIST^}) is to
9180 @file{stderr}. The output produced by the
9181 @option{^-M^/DEPENDENCIES_LIST^} switch is send to
9184 @node Switches for gnatmake
9185 @section Switches for @command{gnatmake}
9188 You may specify any of the following switches to @command{gnatmake}:
9194 @cindex @option{--version} @command{gnatmake}
9195 Display Copyright and version, then exit disregarding all other options.
9198 @cindex @option{--help} @command{gnatmake}
9199 If @option{--version} was not used, display usage, then exit disregarding
9203 @item --GCC=@var{compiler_name}
9204 @cindex @option{--GCC=compiler_name} (@command{gnatmake})
9205 Program used for compiling. The default is `@command{gcc}'. You need to use
9206 quotes around @var{compiler_name} if @code{compiler_name} contains
9207 spaces or other separator characters. As an example @option{--GCC="foo -x
9208 -y"} will instruct @command{gnatmake} to use @code{foo -x -y} as your
9209 compiler. A limitation of this syntax is that the name and path name of
9210 the executable itself must not include any embedded spaces. Note that
9211 switch @option{-c} is always inserted after your command name. Thus in the
9212 above example the compiler command that will be used by @command{gnatmake}
9213 will be @code{foo -c -x -y}. If several @option{--GCC=compiler_name} are
9214 used, only the last @var{compiler_name} is taken into account. However,
9215 all the additional switches are also taken into account. Thus,
9216 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
9217 @option{--GCC="bar -x -y -z -t"}.
9219 @item --GNATBIND=@var{binder_name}
9220 @cindex @option{--GNATBIND=binder_name} (@command{gnatmake})
9221 Program used for binding. The default is `@code{gnatbind}'. You need to
9222 use quotes around @var{binder_name} if @var{binder_name} contains spaces
9223 or other separator characters. As an example @option{--GNATBIND="bar -x
9224 -y"} will instruct @command{gnatmake} to use @code{bar -x -y} as your
9225 binder. Binder switches that are normally appended by @command{gnatmake}
9226 to `@code{gnatbind}' are now appended to the end of @code{bar -x -y}.
9227 A limitation of this syntax is that the name and path name of the executable
9228 itself must not include any embedded spaces.
9230 @item --GNATLINK=@var{linker_name}
9231 @cindex @option{--GNATLINK=linker_name} (@command{gnatmake})
9232 Program used for linking. The default is `@command{gnatlink}'. You need to
9233 use quotes around @var{linker_name} if @var{linker_name} contains spaces
9234 or other separator characters. As an example @option{--GNATLINK="lan -x
9235 -y"} will instruct @command{gnatmake} to use @code{lan -x -y} as your
9236 linker. Linker switches that are normally appended by @command{gnatmake} to
9237 `@command{gnatlink}' are now appended to the end of @code{lan -x -y}.
9238 A limitation of this syntax is that the name and path name of the executable
9239 itself must not include any embedded spaces.
9243 @item ^--subdirs^/SUBDIRS^=subdir
9244 Actual object directory of each project file is the subdirectory subdir of the
9245 object directory specified or defauted in the project file.
9247 @item ^--unchecked-shared-lib-imports^/UNCHECKED_SHARED_LIB_IMPORTS^
9248 By default, shared library projects are not allowed to import static library
9249 projects. When this switch is used on the command line, this restriction is
9252 @item ^-a^/ALL_FILES^
9253 @cindex @option{^-a^/ALL_FILES^} (@command{gnatmake})
9254 Consider all files in the make process, even the GNAT internal system
9255 files (for example, the predefined Ada library files), as well as any
9256 locked files. Locked files are files whose ALI file is write-protected.
9258 @command{gnatmake} does not check these files,
9259 because the assumption is that the GNAT internal files are properly up
9260 to date, and also that any write protected ALI files have been properly
9261 installed. Note that if there is an installation problem, such that one
9262 of these files is not up to date, it will be properly caught by the
9264 You may have to specify this switch if you are working on GNAT
9265 itself. The switch @option{^-a^/ALL_FILES^} is also useful
9266 in conjunction with @option{^-f^/FORCE_COMPILE^}
9267 if you need to recompile an entire application,
9268 including run-time files, using special configuration pragmas,
9269 such as a @code{Normalize_Scalars} pragma.
9272 @code{gnatmake ^-a^/ALL_FILES^} compiles all GNAT
9275 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
9278 the @code{/CHECKS=SUPPRESS_ALL /STYLE_CHECKS=GNAT} switch.
9281 @item ^-b^/ACTIONS=BIND^
9282 @cindex @option{^-b^/ACTIONS=BIND^} (@command{gnatmake})
9283 Bind only. Can be combined with @option{^-c^/ACTIONS=COMPILE^} to do
9284 compilation and binding, but no link.
9285 Can be combined with @option{^-l^/ACTIONS=LINK^}
9286 to do binding and linking. When not combined with
9287 @option{^-c^/ACTIONS=COMPILE^}
9288 all the units in the closure of the main program must have been previously
9289 compiled and must be up to date. The root unit specified by @var{file_name}
9290 may be given without extension, with the source extension or, if no GNAT
9291 Project File is specified, with the ALI file extension.
9293 @item ^-c^/ACTIONS=COMPILE^
9294 @cindex @option{^-c^/ACTIONS=COMPILE^} (@command{gnatmake})
9295 Compile only. Do not perform binding, except when @option{^-b^/ACTIONS=BIND^}
9296 is also specified. Do not perform linking, except if both
9297 @option{^-b^/ACTIONS=BIND^} and
9298 @option{^-l^/ACTIONS=LINK^} are also specified.
9299 If the root unit specified by @var{file_name} is not a main unit, this is the
9300 default. Otherwise @command{gnatmake} will attempt binding and linking
9301 unless all objects are up to date and the executable is more recent than
9305 @cindex @option{^-C^/MAPPING^} (@command{gnatmake})
9306 Use a temporary mapping file. A mapping file is a way to communicate
9307 to the compiler two mappings: from unit names to file names (without
9308 any directory information) and from file names to path names (with
9309 full directory information). A mapping file can make the compiler's
9310 file searches faster, especially if there are many source directories,
9311 or the sources are read over a slow network connection. If
9312 @option{^-P^/PROJECT_FILE^} is used, a mapping file is always used, so
9313 @option{^-C^/MAPPING^} is unnecessary; in this case the mapping file
9314 is initially populated based on the project file. If
9315 @option{^-C^/MAPPING^} is used without
9316 @option{^-P^/PROJECT_FILE^},
9317 the mapping file is initially empty. Each invocation of the compiler
9318 will add any newly accessed sources to the mapping file.
9320 @item ^-C=^/USE_MAPPING_FILE=^@var{file}
9321 @cindex @option{^-C=^/USE_MAPPING^} (@command{gnatmake})
9322 Use a specific mapping file. The file, specified as a path name (absolute or
9323 relative) by this switch, should already exist, otherwise the switch is
9324 ineffective. The specified mapping file will be communicated to the compiler.
9325 This switch is not compatible with a project file
9326 (^-P^/PROJECT_FILE=^@var{file}) or with multiple compiling processes
9327 (^-j^/PROCESSES=^nnn, when nnn is greater than 1).
9329 @item ^-d^/DISPLAY_PROGRESS^
9330 @cindex @option{^-d^/DISPLAY_PROGRESS^} (@command{gnatmake})
9331 Display progress for each source, up to date or not, as a single line
9334 completed x out of y (zz%)
9337 If the file needs to be compiled this is displayed after the invocation of
9338 the compiler. These lines are displayed even in quiet output mode.
9340 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
9341 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@command{gnatmake})
9342 Put all object files and ALI file in directory @var{dir}.
9343 If the @option{^-D^/DIRECTORY_OBJECTS^} switch is not used, all object files
9344 and ALI files go in the current working directory.
9346 This switch cannot be used when using a project file.
9350 @cindex @option{-eL} (@command{gnatmake})
9351 @cindex symbolic links
9352 Follow all symbolic links when processing project files.
9353 This should be used if your project uses symbolic links for files or
9354 directories, but is not needed in other cases.
9356 @cindex naming scheme
9357 This also assumes that no directory matches the naming scheme for files (for
9358 instance that you do not have a directory called "sources.ads" when using the
9359 default GNAT naming scheme).
9361 When you do not have to use this switch (ie by default), gnatmake is able to
9362 save a lot of system calls (several per source file and object file), which
9363 can result in a significant speed up to load and manipulate a project file,
9364 especially when using source files from a remote system.
9368 @item ^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^
9369 @cindex @option{^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^} (@command{gnatmake})
9370 Output the commands for the compiler, the binder and the linker
9371 on ^standard output^SYS$OUTPUT^,
9372 instead of ^standard error^SYS$ERROR^.
9374 @item ^-f^/FORCE_COMPILE^
9375 @cindex @option{^-f^/FORCE_COMPILE^} (@command{gnatmake})
9376 Force recompilations. Recompile all sources, even though some object
9377 files may be up to date, but don't recompile predefined or GNAT internal
9378 files or locked files (files with a write-protected ALI file),
9379 unless the @option{^-a^/ALL_FILES^} switch is also specified.
9381 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
9382 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@command{gnatmake})
9383 When using project files, if some errors or warnings are detected during
9384 parsing and verbose mode is not in effect (no use of switch
9385 ^-v^/VERBOSE^), then error lines start with the full path name of the project
9386 file, rather than its simple file name.
9389 @cindex @option{^-g^/DEBUG^} (@command{gnatmake})
9390 Enable debugging. This switch is simply passed to the compiler and to the
9393 @item ^-i^/IN_PLACE^
9394 @cindex @option{^-i^/IN_PLACE^} (@command{gnatmake})
9395 In normal mode, @command{gnatmake} compiles all object files and ALI files
9396 into the current directory. If the @option{^-i^/IN_PLACE^} switch is used,
9397 then instead object files and ALI files that already exist are overwritten
9398 in place. This means that once a large project is organized into separate
9399 directories in the desired manner, then @command{gnatmake} will automatically
9400 maintain and update this organization. If no ALI files are found on the
9401 Ada object path (@ref{Search Paths and the Run-Time Library (RTL)}),
9402 the new object and ALI files are created in the
9403 directory containing the source being compiled. If another organization
9404 is desired, where objects and sources are kept in different directories,
9405 a useful technique is to create dummy ALI files in the desired directories.
9406 When detecting such a dummy file, @command{gnatmake} will be forced to
9407 recompile the corresponding source file, and it will be put the resulting
9408 object and ALI files in the directory where it found the dummy file.
9410 @item ^-j^/PROCESSES=^@var{n}
9411 @cindex @option{^-j^/PROCESSES^} (@command{gnatmake})
9412 @cindex Parallel make
9413 Use @var{n} processes to carry out the (re)compilations. On a
9414 multiprocessor machine compilations will occur in parallel. In the
9415 event of compilation errors, messages from various compilations might
9416 get interspersed (but @command{gnatmake} will give you the full ordered
9417 list of failing compiles at the end). If this is problematic, rerun
9418 the make process with n set to 1 to get a clean list of messages.
9420 @item ^-k^/CONTINUE_ON_ERROR^
9421 @cindex @option{^-k^/CONTINUE_ON_ERROR^} (@command{gnatmake})
9422 Keep going. Continue as much as possible after a compilation error. To
9423 ease the programmer's task in case of compilation errors, the list of
9424 sources for which the compile fails is given when @command{gnatmake}
9427 If @command{gnatmake} is invoked with several @file{file_names} and with this
9428 switch, if there are compilation errors when building an executable,
9429 @command{gnatmake} will not attempt to build the following executables.
9431 @item ^-l^/ACTIONS=LINK^
9432 @cindex @option{^-l^/ACTIONS=LINK^} (@command{gnatmake})
9433 Link only. Can be combined with @option{^-b^/ACTIONS=BIND^} to binding
9434 and linking. Linking will not be performed if combined with
9435 @option{^-c^/ACTIONS=COMPILE^}
9436 but not with @option{^-b^/ACTIONS=BIND^}.
9437 When not combined with @option{^-b^/ACTIONS=BIND^}
9438 all the units in the closure of the main program must have been previously
9439 compiled and must be up to date, and the main program needs to have been bound.
9440 The root unit specified by @var{file_name}
9441 may be given without extension, with the source extension or, if no GNAT
9442 Project File is specified, with the ALI file extension.
9444 @item ^-m^/MINIMAL_RECOMPILATION^
9445 @cindex @option{^-m^/MINIMAL_RECOMPILATION^} (@command{gnatmake})
9446 Specify that the minimum necessary amount of recompilations
9447 be performed. In this mode @command{gnatmake} ignores time
9448 stamp differences when the only
9449 modifications to a source file consist in adding/removing comments,
9450 empty lines, spaces or tabs. This means that if you have changed the
9451 comments in a source file or have simply reformatted it, using this
9452 switch will tell @command{gnatmake} not to recompile files that depend on it
9453 (provided other sources on which these files depend have undergone no
9454 semantic modifications). Note that the debugging information may be
9455 out of date with respect to the sources if the @option{-m} switch causes
9456 a compilation to be switched, so the use of this switch represents a
9457 trade-off between compilation time and accurate debugging information.
9459 @item ^-M^/DEPENDENCIES_LIST^
9460 @cindex Dependencies, producing list
9461 @cindex @option{^-M^/DEPENDENCIES_LIST^} (@command{gnatmake})
9462 Check if all objects are up to date. If they are, output the object
9463 dependences to @file{stdout} in a form that can be directly exploited in
9464 a @file{Makefile}. By default, each source file is prefixed with its
9465 (relative or absolute) directory name. This name is whatever you
9466 specified in the various @option{^-aI^/SOURCE_SEARCH^}
9467 and @option{^-I^/SEARCH^} switches. If you use
9468 @code{gnatmake ^-M^/DEPENDENCIES_LIST^}
9469 @option{^-q^/QUIET^}
9470 (see below), only the source file names,
9471 without relative paths, are output. If you just specify the
9472 @option{^-M^/DEPENDENCIES_LIST^}
9473 switch, dependencies of the GNAT internal system files are omitted. This
9474 is typically what you want. If you also specify
9475 the @option{^-a^/ALL_FILES^} switch,
9476 dependencies of the GNAT internal files are also listed. Note that
9477 dependencies of the objects in external Ada libraries (see switch
9478 @option{^-aL^/SKIP_MISSING=^}@var{dir} in the following list)
9481 @item ^-n^/DO_OBJECT_CHECK^
9482 @cindex @option{^-n^/DO_OBJECT_CHECK^} (@command{gnatmake})
9483 Don't compile, bind, or link. Checks if all objects are up to date.
9484 If they are not, the full name of the first file that needs to be
9485 recompiled is printed.
9486 Repeated use of this option, followed by compiling the indicated source
9487 file, will eventually result in recompiling all required units.
9489 @item ^-o ^/EXECUTABLE=^@var{exec_name}
9490 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatmake})
9491 Output executable name. The name of the final executable program will be
9492 @var{exec_name}. If the @option{^-o^/EXECUTABLE^} switch is omitted the default
9493 name for the executable will be the name of the input file in appropriate form
9494 for an executable file on the host system.
9496 This switch cannot be used when invoking @command{gnatmake} with several
9499 @item ^-p or --create-missing-dirs^/CREATE_MISSING_DIRS^
9500 @cindex @option{^-p^/CREATE_MISSING_DIRS^} (@command{gnatmake})
9501 When using project files (^-P^/PROJECT_FILE=^@var{project}), create
9502 automatically missing object directories, library directories and exec
9505 @item ^-P^/PROJECT_FILE=^@var{project}
9506 @cindex @option{^-P^/PROJECT_FILE^} (@command{gnatmake})
9507 Use project file @var{project}. Only one such switch can be used.
9508 @xref{gnatmake and Project Files}.
9511 @cindex @option{^-q^/QUIET^} (@command{gnatmake})
9512 Quiet. When this flag is not set, the commands carried out by
9513 @command{gnatmake} are displayed.
9515 @item ^-s^/SWITCH_CHECK/^
9516 @cindex @option{^-s^/SWITCH_CHECK^} (@command{gnatmake})
9517 Recompile if compiler switches have changed since last compilation.
9518 All compiler switches but -I and -o are taken into account in the
9520 orders between different ``first letter'' switches are ignored, but
9521 orders between same switches are taken into account. For example,
9522 @option{-O -O2} is different than @option{-O2 -O}, but @option{-g -O}
9523 is equivalent to @option{-O -g}.
9525 This switch is recommended when Integrated Preprocessing is used.
9528 @cindex @option{^-u^/UNIQUE^} (@command{gnatmake})
9529 Unique. Recompile at most the main files. It implies -c. Combined with
9530 -f, it is equivalent to calling the compiler directly. Note that using
9531 ^-u^/UNIQUE^ with a project file and no main has a special meaning
9532 (@pxref{Project Files and Main Subprograms}).
9534 @item ^-U^/ALL_PROJECTS^
9535 @cindex @option{^-U^/ALL_PROJECTS^} (@command{gnatmake})
9536 When used without a project file or with one or several mains on the command
9537 line, is equivalent to ^-u^/UNIQUE^. When used with a project file and no main
9538 on the command line, all sources of all project files are checked and compiled
9539 if not up to date, and libraries are rebuilt, if necessary.
9542 @cindex @option{^-v^/REASONS^} (@command{gnatmake})
9543 Verbose. Display the reason for all recompilations @command{gnatmake}
9544 decides are necessary, with the highest verbosity level.
9546 @item ^-vl^/LOW_VERBOSITY^
9547 @cindex @option{^-vl^/LOW_VERBOSITY^} (@command{gnatmake})
9548 Verbosity level Low. Display fewer lines than in verbosity Medium.
9550 @item ^-vm^/MEDIUM_VERBOSITY^
9551 @cindex @option{^-vm^/MEDIUM_VERBOSITY^} (@command{gnatmake})
9552 Verbosity level Medium. Potentially display fewer lines than in verbosity High.
9554 @item ^-vh^/HIGH_VERBOSITY^
9555 @cindex @option{^-vm^/HIGH_VERBOSITY^} (@command{gnatmake})
9556 Verbosity level High. Equivalent to ^-v^/REASONS^.
9558 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
9559 Indicate the verbosity of the parsing of GNAT project files.
9560 @xref{Switches Related to Project Files}.
9562 @item ^-x^/NON_PROJECT_UNIT_COMPILATION^
9563 @cindex @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} (@command{gnatmake})
9564 Indicate that sources that are not part of any Project File may be compiled.
9565 Normally, when using Project Files, only sources that are part of a Project
9566 File may be compile. When this switch is used, a source outside of all Project
9567 Files may be compiled. The ALI file and the object file will be put in the
9568 object directory of the main Project. The compilation switches used will only
9569 be those specified on the command line. Even when
9570 @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} is used, mains specified on the
9571 command line need to be sources of a project file.
9573 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
9574 Indicate that external variable @var{name} has the value @var{value}.
9575 The Project Manager will use this value for occurrences of
9576 @code{external(name)} when parsing the project file.
9577 @xref{Switches Related to Project Files}.
9580 @cindex @option{^-z^/NOMAIN^} (@command{gnatmake})
9581 No main subprogram. Bind and link the program even if the unit name
9582 given on the command line is a package name. The resulting executable
9583 will execute the elaboration routines of the package and its closure,
9584 then the finalization routines.
9589 @item @command{gcc} @asis{switches}
9591 Any uppercase or multi-character switch that is not a @command{gnatmake} switch
9592 is passed to @command{gcc} (e.g.@: @option{-O}, @option{-gnato,} etc.)
9595 Any qualifier that cannot be recognized as a qualifier for @code{GNAT MAKE}
9596 but is recognizable as a valid qualifier for @code{GNAT COMPILE} is
9597 automatically treated as a compiler switch, and passed on to all
9598 compilations that are carried out.
9603 Source and library search path switches:
9607 @item ^-aI^/SOURCE_SEARCH=^@var{dir}
9608 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatmake})
9609 When looking for source files also look in directory @var{dir}.
9610 The order in which source files search is undertaken is
9611 described in @ref{Search Paths and the Run-Time Library (RTL)}.
9613 @item ^-aL^/SKIP_MISSING=^@var{dir}
9614 @cindex @option{^-aL^/SKIP_MISSING^} (@command{gnatmake})
9615 Consider @var{dir} as being an externally provided Ada library.
9616 Instructs @command{gnatmake} to skip compilation units whose @file{.ALI}
9617 files have been located in directory @var{dir}. This allows you to have
9618 missing bodies for the units in @var{dir} and to ignore out of date bodies
9619 for the same units. You still need to specify
9620 the location of the specs for these units by using the switches
9621 @option{^-aI^/SOURCE_SEARCH=^@var{dir}}
9622 or @option{^-I^/SEARCH=^@var{dir}}.
9623 Note: this switch is provided for compatibility with previous versions
9624 of @command{gnatmake}. The easier method of causing standard libraries
9625 to be excluded from consideration is to write-protect the corresponding
9628 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
9629 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatmake})
9630 When searching for library and object files, look in directory
9631 @var{dir}. The order in which library files are searched is described in
9632 @ref{Search Paths for gnatbind}.
9634 @item ^-A^/CONDITIONAL_SOURCE_SEARCH=^@var{dir}
9635 @cindex Search paths, for @command{gnatmake}
9636 @cindex @option{^-A^/CONDITIONAL_SOURCE_SEARCH^} (@command{gnatmake})
9637 Equivalent to @option{^-aL^/SKIP_MISSING=^@var{dir}
9638 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9640 @item ^-I^/SEARCH=^@var{dir}
9641 @cindex @option{^-I^/SEARCH^} (@command{gnatmake})
9642 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}
9643 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
9645 @item ^-I-^/NOCURRENT_DIRECTORY^
9646 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatmake})
9647 @cindex Source files, suppressing search
9648 Do not look for source files in the directory containing the source
9649 file named in the command line.
9650 Do not look for ALI or object files in the directory
9651 where @command{gnatmake} was invoked.
9653 @item ^-L^/LIBRARY_SEARCH=^@var{dir}
9654 @cindex @option{^-L^/LIBRARY_SEARCH^} (@command{gnatmake})
9655 @cindex Linker libraries
9656 Add directory @var{dir} to the list of directories in which the linker
9657 will search for libraries. This is equivalent to
9658 @option{-largs ^-L^/LIBRARY_SEARCH=^}@var{dir}.
9660 Furthermore, under Windows, the sources pointed to by the libraries path
9661 set in the registry are not searched for.
9665 @cindex @option{-nostdinc} (@command{gnatmake})
9666 Do not look for source files in the system default directory.
9669 @cindex @option{-nostdlib} (@command{gnatmake})
9670 Do not look for library files in the system default directory.
9672 @item --RTS=@var{rts-path}
9673 @cindex @option{--RTS} (@command{gnatmake})
9674 Specifies the default location of the runtime library. GNAT looks for the
9676 in the following directories, and stops as soon as a valid runtime is found
9677 (@file{adainclude} or @file{ada_source_path}, and @file{adalib} or
9678 @file{ada_object_path} present):
9681 @item <current directory>/$rts_path
9683 @item <default-search-dir>/$rts_path
9685 @item <default-search-dir>/rts-$rts_path
9689 The selected path is handled like a normal RTS path.
9693 @node Mode Switches for gnatmake
9694 @section Mode Switches for @command{gnatmake}
9697 The mode switches (referred to as @code{mode_switches}) allow the
9698 inclusion of switches that are to be passed to the compiler itself, the
9699 binder or the linker. The effect of a mode switch is to cause all
9700 subsequent switches up to the end of the switch list, or up to the next
9701 mode switch, to be interpreted as switches to be passed on to the
9702 designated component of GNAT.
9706 @item -cargs @var{switches}
9707 @cindex @option{-cargs} (@command{gnatmake})
9708 Compiler switches. Here @var{switches} is a list of switches
9709 that are valid switches for @command{gcc}. They will be passed on to
9710 all compile steps performed by @command{gnatmake}.
9712 @item -bargs @var{switches}
9713 @cindex @option{-bargs} (@command{gnatmake})
9714 Binder switches. Here @var{switches} is a list of switches
9715 that are valid switches for @code{gnatbind}. They will be passed on to
9716 all bind steps performed by @command{gnatmake}.
9718 @item -largs @var{switches}
9719 @cindex @option{-largs} (@command{gnatmake})
9720 Linker switches. Here @var{switches} is a list of switches
9721 that are valid switches for @command{gnatlink}. They will be passed on to
9722 all link steps performed by @command{gnatmake}.
9724 @item -margs @var{switches}
9725 @cindex @option{-margs} (@command{gnatmake})
9726 Make switches. The switches are directly interpreted by @command{gnatmake},
9727 regardless of any previous occurrence of @option{-cargs}, @option{-bargs}
9731 @node Notes on the Command Line
9732 @section Notes on the Command Line
9735 This section contains some additional useful notes on the operation
9736 of the @command{gnatmake} command.
9740 @cindex Recompilation, by @command{gnatmake}
9741 If @command{gnatmake} finds no ALI files, it recompiles the main program
9742 and all other units required by the main program.
9743 This means that @command{gnatmake}
9744 can be used for the initial compile, as well as during subsequent steps of
9745 the development cycle.
9748 If you enter @code{gnatmake @var{file}.adb}, where @file{@var{file}.adb}
9749 is a subunit or body of a generic unit, @command{gnatmake} recompiles
9750 @file{@var{file}.adb} (because it finds no ALI) and stops, issuing a
9754 In @command{gnatmake} the switch @option{^-I^/SEARCH^}
9755 is used to specify both source and
9756 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
9757 instead if you just want to specify
9758 source paths only and @option{^-aO^/OBJECT_SEARCH^}
9759 if you want to specify library paths
9763 @command{gnatmake} will ignore any files whose ALI file is write-protected.
9764 This may conveniently be used to exclude standard libraries from
9765 consideration and in particular it means that the use of the
9766 @option{^-f^/FORCE_COMPILE^} switch will not recompile these files
9767 unless @option{^-a^/ALL_FILES^} is also specified.
9770 @command{gnatmake} has been designed to make the use of Ada libraries
9771 particularly convenient. Assume you have an Ada library organized
9772 as follows: @i{^obj-dir^[OBJ_DIR]^} contains the objects and ALI files for
9773 of your Ada compilation units,
9774 whereas @i{^include-dir^[INCLUDE_DIR]^} contains the
9775 specs of these units, but no bodies. Then to compile a unit
9776 stored in @code{main.adb}, which uses this Ada library you would just type
9780 $ gnatmake -aI@var{include-dir} -aL@var{obj-dir} main
9783 $ gnatmake /SOURCE_SEARCH=@i{[INCLUDE_DIR]}
9784 /SKIP_MISSING=@i{[OBJ_DIR]} main
9789 Using @command{gnatmake} along with the
9790 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}
9791 switch provides a mechanism for avoiding unnecessary recompilations. Using
9793 you can update the comments/format of your
9794 source files without having to recompile everything. Note, however, that
9795 adding or deleting lines in a source files may render its debugging
9796 info obsolete. If the file in question is a spec, the impact is rather
9797 limited, as that debugging info will only be useful during the
9798 elaboration phase of your program. For bodies the impact can be more
9799 significant. In all events, your debugger will warn you if a source file
9800 is more recent than the corresponding object, and alert you to the fact
9801 that the debugging information may be out of date.
9804 @node How gnatmake Works
9805 @section How @command{gnatmake} Works
9808 Generally @command{gnatmake} automatically performs all necessary
9809 recompilations and you don't need to worry about how it works. However,
9810 it may be useful to have some basic understanding of the @command{gnatmake}
9811 approach and in particular to understand how it uses the results of
9812 previous compilations without incorrectly depending on them.
9814 First a definition: an object file is considered @dfn{up to date} if the
9815 corresponding ALI file exists and if all the source files listed in the
9816 dependency section of this ALI file have time stamps matching those in
9817 the ALI file. This means that neither the source file itself nor any
9818 files that it depends on have been modified, and hence there is no need
9819 to recompile this file.
9821 @command{gnatmake} works by first checking if the specified main unit is up
9822 to date. If so, no compilations are required for the main unit. If not,
9823 @command{gnatmake} compiles the main program to build a new ALI file that
9824 reflects the latest sources. Then the ALI file of the main unit is
9825 examined to find all the source files on which the main program depends,
9826 and @command{gnatmake} recursively applies the above procedure on all these
9829 This process ensures that @command{gnatmake} only trusts the dependencies
9830 in an existing ALI file if they are known to be correct. Otherwise it
9831 always recompiles to determine a new, guaranteed accurate set of
9832 dependencies. As a result the program is compiled ``upside down'' from what may
9833 be more familiar as the required order of compilation in some other Ada
9834 systems. In particular, clients are compiled before the units on which
9835 they depend. The ability of GNAT to compile in any order is critical in
9836 allowing an order of compilation to be chosen that guarantees that
9837 @command{gnatmake} will recompute a correct set of new dependencies if
9840 When invoking @command{gnatmake} with several @var{file_names}, if a unit is
9841 imported by several of the executables, it will be recompiled at most once.
9843 Note: when using non-standard naming conventions
9844 (@pxref{Using Other File Names}), changing through a configuration pragmas
9845 file the version of a source and invoking @command{gnatmake} to recompile may
9846 have no effect, if the previous version of the source is still accessible
9847 by @command{gnatmake}. It may be necessary to use the switch
9848 ^-f^/FORCE_COMPILE^.
9850 @node Examples of gnatmake Usage
9851 @section Examples of @command{gnatmake} Usage
9854 @item gnatmake hello.adb
9855 Compile all files necessary to bind and link the main program
9856 @file{hello.adb} (containing unit @code{Hello}) and bind and link the
9857 resulting object files to generate an executable file @file{^hello^HELLO.EXE^}.
9859 @item gnatmake main1 main2 main3
9860 Compile all files necessary to bind and link the main programs
9861 @file{main1.adb} (containing unit @code{Main1}), @file{main2.adb}
9862 (containing unit @code{Main2}) and @file{main3.adb}
9863 (containing unit @code{Main3}) and bind and link the resulting object files
9864 to generate three executable files @file{^main1^MAIN1.EXE^},
9865 @file{^main2^MAIN2.EXE^}
9866 and @file{^main3^MAIN3.EXE^}.
9869 @item gnatmake -q Main_Unit -cargs -O2 -bargs -l
9873 @item gnatmake Main_Unit /QUIET
9874 /COMPILER_QUALIFIERS /OPTIMIZE=ALL
9875 /BINDER_QUALIFIERS /ORDER_OF_ELABORATION
9877 Compile all files necessary to bind and link the main program unit
9878 @code{Main_Unit} (from file @file{main_unit.adb}). All compilations will
9879 be done with optimization level 2 and the order of elaboration will be
9880 listed by the binder. @command{gnatmake} will operate in quiet mode, not
9881 displaying commands it is executing.
9884 @c *************************
9885 @node Improving Performance
9886 @chapter Improving Performance
9887 @cindex Improving performance
9890 This chapter presents several topics related to program performance.
9891 It first describes some of the tradeoffs that need to be considered
9892 and some of the techniques for making your program run faster.
9893 It then documents the @command{gnatelim} tool and unused subprogram/data
9894 elimination feature, which can reduce the size of program executables.
9896 Note: to invoke @command{gnatelim} with a project file, use the @code{gnat}
9897 driver (see @ref{The GNAT Driver and Project Files}).
9901 * Performance Considerations::
9902 * Text_IO Suggestions::
9903 * Reducing Size of Ada Executables with gnatelim::
9904 * Reducing Size of Executables with unused subprogram/data elimination::
9908 @c *****************************
9909 @node Performance Considerations
9910 @section Performance Considerations
9913 The GNAT system provides a number of options that allow a trade-off
9918 performance of the generated code
9921 speed of compilation
9924 minimization of dependences and recompilation
9927 the degree of run-time checking.
9931 The defaults (if no options are selected) aim at improving the speed
9932 of compilation and minimizing dependences, at the expense of performance
9933 of the generated code:
9940 no inlining of subprogram calls
9943 all run-time checks enabled except overflow and elaboration checks
9947 These options are suitable for most program development purposes. This
9948 chapter describes how you can modify these choices, and also provides
9949 some guidelines on debugging optimized code.
9952 * Controlling Run-Time Checks::
9953 * Use of Restrictions::
9954 * Optimization Levels::
9955 * Debugging Optimized Code::
9956 * Inlining of Subprograms::
9957 * Other Optimization Switches::
9958 * Optimization and Strict Aliasing::
9961 * Coverage Analysis::
9965 @node Controlling Run-Time Checks
9966 @subsection Controlling Run-Time Checks
9969 By default, GNAT generates all run-time checks, except integer overflow
9970 checks, stack overflow checks, and checks for access before elaboration on
9971 subprogram calls. The latter are not required in default mode, because all
9972 necessary checking is done at compile time.
9973 @cindex @option{-gnatp} (@command{gcc})
9974 @cindex @option{-gnato} (@command{gcc})
9975 Two gnat switches, @option{-gnatp} and @option{-gnato} allow this default to
9976 be modified. @xref{Run-Time Checks}.
9978 Our experience is that the default is suitable for most development
9981 We treat integer overflow specially because these
9982 are quite expensive and in our experience are not as important as other
9983 run-time checks in the development process. Note that division by zero
9984 is not considered an overflow check, and divide by zero checks are
9985 generated where required by default.
9987 Elaboration checks are off by default, and also not needed by default, since
9988 GNAT uses a static elaboration analysis approach that avoids the need for
9989 run-time checking. This manual contains a full chapter discussing the issue
9990 of elaboration checks, and if the default is not satisfactory for your use,
9991 you should read this chapter.
9993 For validity checks, the minimal checks required by the Ada Reference
9994 Manual (for case statements and assignments to array elements) are on
9995 by default. These can be suppressed by use of the @option{-gnatVn} switch.
9996 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
9997 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
9998 it may be reasonable to routinely use @option{-gnatVn}. Validity checks
9999 are also suppressed entirely if @option{-gnatp} is used.
10001 @cindex Overflow checks
10002 @cindex Checks, overflow
10005 @cindex pragma Suppress
10006 @cindex pragma Unsuppress
10007 Note that the setting of the switches controls the default setting of
10008 the checks. They may be modified using either @code{pragma Suppress} (to
10009 remove checks) or @code{pragma Unsuppress} (to add back suppressed
10010 checks) in the program source.
10012 @node Use of Restrictions
10013 @subsection Use of Restrictions
10016 The use of pragma Restrictions allows you to control which features are
10017 permitted in your program. Apart from the obvious point that if you avoid
10018 relatively expensive features like finalization (enforceable by the use
10019 of pragma Restrictions (No_Finalization), the use of this pragma does not
10020 affect the generated code in most cases.
10022 One notable exception to this rule is that the possibility of task abort
10023 results in some distributed overhead, particularly if finalization or
10024 exception handlers are used. The reason is that certain sections of code
10025 have to be marked as non-abortable.
10027 If you use neither the @code{abort} statement, nor asynchronous transfer
10028 of control (@code{select @dots{} then abort}), then this distributed overhead
10029 is removed, which may have a general positive effect in improving
10030 overall performance. Especially code involving frequent use of tasking
10031 constructs and controlled types will show much improved performance.
10032 The relevant restrictions pragmas are
10034 @smallexample @c ada
10035 pragma Restrictions (No_Abort_Statements);
10036 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
10040 It is recommended that these restriction pragmas be used if possible. Note
10041 that this also means that you can write code without worrying about the
10042 possibility of an immediate abort at any point.
10044 @node Optimization Levels
10045 @subsection Optimization Levels
10046 @cindex @option{^-O^/OPTIMIZE^} (@command{gcc})
10049 Without any optimization ^option,^qualifier,^
10050 the compiler's goal is to reduce the cost of
10051 compilation and to make debugging produce the expected results.
10052 Statements are independent: if you stop the program with a breakpoint between
10053 statements, you can then assign a new value to any variable or change
10054 the program counter to any other statement in the subprogram and get exactly
10055 the results you would expect from the source code.
10057 Turning on optimization makes the compiler attempt to improve the
10058 performance and/or code size at the expense of compilation time and
10059 possibly the ability to debug the program.
10061 If you use multiple
10062 ^-O options, with or without level numbers,^/OPTIMIZE qualifiers,^
10063 the last such option is the one that is effective.
10066 The default is optimization off. This results in the fastest compile
10067 times, but GNAT makes absolutely no attempt to optimize, and the
10068 generated programs are considerably larger and slower than when
10069 optimization is enabled. You can use the
10071 @option{-O} switch (the permitted forms are @option{-O0}, @option{-O1}
10072 @option{-O2}, @option{-O3}, and @option{-Os})
10075 @code{OPTIMIZE} qualifier
10077 to @command{gcc} to control the optimization level:
10080 @item ^-O0^/OPTIMIZE=NONE^
10081 No optimization (the default);
10082 generates unoptimized code but has
10083 the fastest compilation time.
10085 Note that many other compilers do fairly extensive optimization
10086 even if ``no optimization'' is specified. With gcc, it is
10087 very unusual to use ^-O0^/OPTIMIZE=NONE^ for production if
10088 execution time is of any concern, since ^-O0^/OPTIMIZE=NONE^
10089 really does mean no optimization at all. This difference between
10090 gcc and other compilers should be kept in mind when doing
10091 performance comparisons.
10093 @item ^-O1^/OPTIMIZE=SOME^
10094 Moderate optimization;
10095 optimizes reasonably well but does not
10096 degrade compilation time significantly.
10098 @item ^-O2^/OPTIMIZE=ALL^
10100 @itemx /OPTIMIZE=DEVELOPMENT
10103 generates highly optimized code and has
10104 the slowest compilation time.
10106 @item ^-O3^/OPTIMIZE=INLINING^
10107 Full optimization as in @option{-O2},
10108 and also attempts automatic inlining of small
10109 subprograms within a unit (@pxref{Inlining of Subprograms}).
10111 @item ^-Os^/OPTIMIZE=SPACE^
10112 Optimize space usage of resulting program.
10116 Higher optimization levels perform more global transformations on the
10117 program and apply more expensive analysis algorithms in order to generate
10118 faster and more compact code. The price in compilation time, and the
10119 resulting improvement in execution time,
10120 both depend on the particular application and the hardware environment.
10121 You should experiment to find the best level for your application.
10123 Since the precise set of optimizations done at each level will vary from
10124 release to release (and sometime from target to target), it is best to think
10125 of the optimization settings in general terms.
10126 @xref{Optimize Options,, Options That Control Optimization, gcc, Using
10127 the GNU Compiler Collection (GCC)}, for details about
10128 ^the @option{-O} settings and a number of @option{-f} options that^how to^
10129 individually enable or disable specific optimizations.
10131 Unlike some other compilation systems, ^@command{gcc}^GNAT^ has
10132 been tested extensively at all optimization levels. There are some bugs
10133 which appear only with optimization turned on, but there have also been
10134 bugs which show up only in @emph{unoptimized} code. Selecting a lower
10135 level of optimization does not improve the reliability of the code
10136 generator, which in practice is highly reliable at all optimization
10139 Note regarding the use of @option{-O3}: The use of this optimization level
10140 is generally discouraged with GNAT, since it often results in larger
10141 executables which run more slowly. See further discussion of this point
10142 in @ref{Inlining of Subprograms}.
10144 @node Debugging Optimized Code
10145 @subsection Debugging Optimized Code
10146 @cindex Debugging optimized code
10147 @cindex Optimization and debugging
10150 Although it is possible to do a reasonable amount of debugging at
10152 nonzero optimization levels,
10153 the higher the level the more likely that
10156 @option{/OPTIMIZE} settings other than @code{NONE},
10157 such settings will make it more likely that
10159 source-level constructs will have been eliminated by optimization.
10160 For example, if a loop is strength-reduced, the loop
10161 control variable may be completely eliminated and thus cannot be
10162 displayed in the debugger.
10163 This can only happen at @option{-O2} or @option{-O3}.
10164 Explicit temporary variables that you code might be eliminated at
10165 ^level^setting^ @option{-O1} or higher.
10167 The use of the @option{^-g^/DEBUG^} switch,
10168 @cindex @option{^-g^/DEBUG^} (@command{gcc})
10169 which is needed for source-level debugging,
10170 affects the size of the program executable on disk,
10171 and indeed the debugging information can be quite large.
10172 However, it has no effect on the generated code (and thus does not
10173 degrade performance)
10175 Since the compiler generates debugging tables for a compilation unit before
10176 it performs optimizations, the optimizing transformations may invalidate some
10177 of the debugging data. You therefore need to anticipate certain
10178 anomalous situations that may arise while debugging optimized code.
10179 These are the most common cases:
10183 @i{The ``hopping Program Counter'':} Repeated @code{step} or @code{next}
10185 the PC bouncing back and forth in the code. This may result from any of
10186 the following optimizations:
10190 @i{Common subexpression elimination:} using a single instance of code for a
10191 quantity that the source computes several times. As a result you
10192 may not be able to stop on what looks like a statement.
10195 @i{Invariant code motion:} moving an expression that does not change within a
10196 loop, to the beginning of the loop.
10199 @i{Instruction scheduling:} moving instructions so as to
10200 overlap loads and stores (typically) with other code, or in
10201 general to move computations of values closer to their uses. Often
10202 this causes you to pass an assignment statement without the assignment
10203 happening and then later bounce back to the statement when the
10204 value is actually needed. Placing a breakpoint on a line of code
10205 and then stepping over it may, therefore, not always cause all the
10206 expected side-effects.
10210 @i{The ``big leap'':} More commonly known as @emph{cross-jumping}, in which
10211 two identical pieces of code are merged and the program counter suddenly
10212 jumps to a statement that is not supposed to be executed, simply because
10213 it (and the code following) translates to the same thing as the code
10214 that @emph{was} supposed to be executed. This effect is typically seen in
10215 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
10216 a @code{break} in a C @code{^switch^switch^} statement.
10219 @i{The ``roving variable'':} The symptom is an unexpected value in a variable.
10220 There are various reasons for this effect:
10224 In a subprogram prologue, a parameter may not yet have been moved to its
10228 A variable may be dead, and its register re-used. This is
10229 probably the most common cause.
10232 As mentioned above, the assignment of a value to a variable may
10236 A variable may be eliminated entirely by value propagation or
10237 other means. In this case, GCC may incorrectly generate debugging
10238 information for the variable
10242 In general, when an unexpected value appears for a local variable or parameter
10243 you should first ascertain if that value was actually computed by
10244 your program, as opposed to being incorrectly reported by the debugger.
10246 array elements in an object designated by an access value
10247 are generally less of a problem, once you have ascertained that the access
10249 Typically, this means checking variables in the preceding code and in the
10250 calling subprogram to verify that the value observed is explainable from other
10251 values (one must apply the procedure recursively to those
10252 other values); or re-running the code and stopping a little earlier
10253 (perhaps before the call) and stepping to better see how the variable obtained
10254 the value in question; or continuing to step @emph{from} the point of the
10255 strange value to see if code motion had simply moved the variable's
10260 In light of such anomalies, a recommended technique is to use @option{-O0}
10261 early in the software development cycle, when extensive debugging capabilities
10262 are most needed, and then move to @option{-O1} and later @option{-O2} as
10263 the debugger becomes less critical.
10264 Whether to use the @option{^-g^/DEBUG^} switch in the release version is
10265 a release management issue.
10267 Note that if you use @option{-g} you can then use the @command{strip} program
10268 on the resulting executable,
10269 which removes both debugging information and global symbols.
10272 @node Inlining of Subprograms
10273 @subsection Inlining of Subprograms
10276 A call to a subprogram in the current unit is inlined if all the
10277 following conditions are met:
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 @cindex pragma Inline
10291 Either @code{pragma Inline} applies to the subprogram, or it is local
10292 to the unit and called once from within it, or it is small and automatic
10293 inlining (optimization level @option{-O3}) is specified.
10297 Calls to subprograms in @code{with}'ed units are normally not inlined.
10298 To achieve actual inlining (that is, replacement of the call by the code
10299 in the body of the subprogram), the following conditions must all be true.
10303 The optimization level is at least @option{-O1}.
10306 The called subprogram is suitable for inlining: It must be small enough
10307 and not contain something that @command{gcc} cannot support in inlined
10311 The call appears in a body (not in a package spec).
10314 There is a @code{pragma Inline} for the subprogram.
10317 @cindex @option{-gnatn} (@command{gcc})
10318 The @option{^-gnatn^/INLINE^} switch
10319 is used in the @command{gcc} command line
10322 Even if all these conditions are met, it may not be possible for
10323 the compiler to inline the call, due to the length of the body,
10324 or features in the body that make it impossible for the compiler
10325 to do the inlining.
10327 Note that specifying the @option{-gnatn} switch causes additional
10328 compilation dependencies. Consider the following:
10330 @smallexample @c ada
10350 With the default behavior (no @option{-gnatn} switch specified), the
10351 compilation of the @code{Main} procedure depends only on its own source,
10352 @file{main.adb}, and the spec of the package in file @file{r.ads}. This
10353 means that editing the body of @code{R} does not require recompiling
10356 On the other hand, the call @code{R.Q} is not inlined under these
10357 circumstances. If the @option{-gnatn} switch is present when @code{Main}
10358 is compiled, the call will be inlined if the body of @code{Q} is small
10359 enough, but now @code{Main} depends on the body of @code{R} in
10360 @file{r.adb} as well as on the spec. This means that if this body is edited,
10361 the main program must be recompiled. Note that this extra dependency
10362 occurs whether or not the call is in fact inlined by @command{gcc}.
10364 The use of front end inlining with @option{-gnatN} generates similar
10365 additional dependencies.
10367 @cindex @option{^-fno-inline^/INLINE=SUPPRESS^} (@command{gcc})
10368 Note: The @option{^-fno-inline^/INLINE=SUPPRESS^} switch
10369 can be used to prevent
10370 all inlining. This switch overrides all other conditions and ensures
10371 that no inlining occurs. The extra dependences resulting from
10372 @option{-gnatn} will still be active, even if
10373 this switch is used to suppress the resulting inlining actions.
10375 @cindex @option{-fno-inline-functions} (@command{gcc})
10376 Note: The @option{-fno-inline-functions} switch can be used to prevent
10377 automatic inlining of small subprograms if @option{-O3} is used.
10379 @cindex @option{-fno-inline-functions-called-once} (@command{gcc})
10380 Note: The @option{-fno-inline-functions-called-once} switch
10381 can be used to prevent inlining of subprograms local to the unit
10382 and called once from within it if @option{-O1} is used.
10384 Note regarding the use of @option{-O3}: There is no difference in inlining
10385 behavior between @option{-O2} and @option{-O3} for subprograms with an explicit
10386 pragma @code{Inline} assuming the use of @option{-gnatn}
10387 or @option{-gnatN} (the switches that activate inlining). If you have used
10388 pragma @code{Inline} in appropriate cases, then it is usually much better
10389 to use @option{-O2} and @option{-gnatn} and avoid the use of @option{-O3} which
10390 in this case only has the effect of inlining subprograms you did not
10391 think should be inlined. We often find that the use of @option{-O3} slows
10392 down code by performing excessive inlining, leading to increased instruction
10393 cache pressure from the increased code size. So the bottom line here is
10394 that you should not automatically assume that @option{-O3} is better than
10395 @option{-O2}, and indeed you should use @option{-O3} only if tests show that
10396 it actually improves performance.
10398 @node Other Optimization Switches
10399 @subsection Other Optimization Switches
10400 @cindex Optimization Switches
10402 Since @code{GNAT} uses the @command{gcc} back end, all the specialized
10403 @command{gcc} optimization switches are potentially usable. These switches
10404 have not been extensively tested with GNAT but can generally be expected
10405 to work. Examples of switches in this category are
10406 @option{-funroll-loops} and
10407 the various target-specific @option{-m} options (in particular, it has been
10408 observed that @option{-march=pentium4} can significantly improve performance
10409 on appropriate machines). For full details of these switches, see
10410 @ref{Submodel Options,, Hardware Models and Configurations, gcc, Using
10411 the GNU Compiler Collection (GCC)}.
10413 @node Optimization and Strict Aliasing
10414 @subsection Optimization and Strict Aliasing
10416 @cindex Strict Aliasing
10417 @cindex No_Strict_Aliasing
10420 The strong typing capabilities of Ada allow an optimizer to generate
10421 efficient code in situations where other languages would be forced to
10422 make worst case assumptions preventing such optimizations. Consider
10423 the following example:
10425 @smallexample @c ada
10428 type Int1 is new Integer;
10429 type Int2 is new Integer;
10430 type Int1A is access Int1;
10431 type Int2A is access Int2;
10438 for J in Data'Range loop
10439 if Data (J) = Int1V.all then
10440 Int2V.all := Int2V.all + 1;
10449 In this example, since the variable @code{Int1V} can only access objects
10450 of type @code{Int1}, and @code{Int2V} can only access objects of type
10451 @code{Int2}, there is no possibility that the assignment to
10452 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
10453 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
10454 for all iterations of the loop and avoid the extra memory reference
10455 required to dereference it each time through the loop.
10457 This kind of optimization, called strict aliasing analysis, is
10458 triggered by specifying an optimization level of @option{-O2} or
10459 higher or @option{-Os} and allows @code{GNAT} to generate more efficient code
10460 when access values are involved.
10462 However, although this optimization is always correct in terms of
10463 the formal semantics of the Ada Reference Manual, difficulties can
10464 arise if features like @code{Unchecked_Conversion} are used to break
10465 the typing system. Consider the following complete program example:
10467 @smallexample @c ada
10470 type int1 is new integer;
10471 type int2 is new integer;
10472 type a1 is access int1;
10473 type a2 is access int2;
10478 function to_a2 (Input : a1) return a2;
10481 with Unchecked_Conversion;
10483 function to_a2 (Input : a1) return a2 is
10485 new Unchecked_Conversion (a1, a2);
10487 return to_a2u (Input);
10493 with Text_IO; use Text_IO;
10495 v1 : a1 := new int1;
10496 v2 : a2 := to_a2 (v1);
10500 put_line (int1'image (v1.all));
10506 This program prints out 0 in @option{-O0} or @option{-O1}
10507 mode, but it prints out 1 in @option{-O2} mode. That's
10508 because in strict aliasing mode, the compiler can and
10509 does assume that the assignment to @code{v2.all} could not
10510 affect the value of @code{v1.all}, since different types
10513 This behavior is not a case of non-conformance with the standard, since
10514 the Ada RM specifies that an unchecked conversion where the resulting
10515 bit pattern is not a correct value of the target type can result in an
10516 abnormal value and attempting to reference an abnormal value makes the
10517 execution of a program erroneous. That's the case here since the result
10518 does not point to an object of type @code{int2}. This means that the
10519 effect is entirely unpredictable.
10521 However, although that explanation may satisfy a language
10522 lawyer, in practice an applications programmer expects an
10523 unchecked conversion involving pointers to create true
10524 aliases and the behavior of printing 1 seems plain wrong.
10525 In this case, the strict aliasing optimization is unwelcome.
10527 Indeed the compiler recognizes this possibility, and the
10528 unchecked conversion generates a warning:
10531 p2.adb:5:07: warning: possible aliasing problem with type "a2"
10532 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
10533 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
10537 Unfortunately the problem is recognized when compiling the body of
10538 package @code{p2}, but the actual "bad" code is generated while
10539 compiling the body of @code{m} and this latter compilation does not see
10540 the suspicious @code{Unchecked_Conversion}.
10542 As implied by the warning message, there are approaches you can use to
10543 avoid the unwanted strict aliasing optimization in a case like this.
10545 One possibility is to simply avoid the use of @option{-O2}, but
10546 that is a bit drastic, since it throws away a number of useful
10547 optimizations that do not involve strict aliasing assumptions.
10549 A less drastic approach is to compile the program using the
10550 option @option{-fno-strict-aliasing}. Actually it is only the
10551 unit containing the dereferencing of the suspicious pointer
10552 that needs to be compiled. So in this case, if we compile
10553 unit @code{m} with this switch, then we get the expected
10554 value of zero printed. Analyzing which units might need
10555 the switch can be painful, so a more reasonable approach
10556 is to compile the entire program with options @option{-O2}
10557 and @option{-fno-strict-aliasing}. If the performance is
10558 satisfactory with this combination of options, then the
10559 advantage is that the entire issue of possible "wrong"
10560 optimization due to strict aliasing is avoided.
10562 To avoid the use of compiler switches, the configuration
10563 pragma @code{No_Strict_Aliasing} with no parameters may be
10564 used to specify that for all access types, the strict
10565 aliasing optimization should be suppressed.
10567 However, these approaches are still overkill, in that they causes
10568 all manipulations of all access values to be deoptimized. A more
10569 refined approach is to concentrate attention on the specific
10570 access type identified as problematic.
10572 First, if a careful analysis of uses of the pointer shows
10573 that there are no possible problematic references, then
10574 the warning can be suppressed by bracketing the
10575 instantiation of @code{Unchecked_Conversion} to turn
10578 @smallexample @c ada
10579 pragma Warnings (Off);
10581 new Unchecked_Conversion (a1, a2);
10582 pragma Warnings (On);
10586 Of course that approach is not appropriate for this particular
10587 example, since indeed there is a problematic reference. In this
10588 case we can take one of two other approaches.
10590 The first possibility is to move the instantiation of unchecked
10591 conversion to the unit in which the type is declared. In
10592 this example, we would move the instantiation of
10593 @code{Unchecked_Conversion} from the body of package
10594 @code{p2} to the spec of package @code{p1}. Now the
10595 warning disappears. That's because any use of the
10596 access type knows there is a suspicious unchecked
10597 conversion, and the strict aliasing optimization
10598 is automatically suppressed for the type.
10600 If it is not practical to move the unchecked conversion to the same unit
10601 in which the destination access type is declared (perhaps because the
10602 source type is not visible in that unit), you may use pragma
10603 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
10604 same declarative sequence as the declaration of the access type:
10606 @smallexample @c ada
10607 type a2 is access int2;
10608 pragma No_Strict_Aliasing (a2);
10612 Here again, the compiler now knows that the strict aliasing optimization
10613 should be suppressed for any reference to type @code{a2} and the
10614 expected behavior is obtained.
10616 Finally, note that although the compiler can generate warnings for
10617 simple cases of unchecked conversions, there are tricker and more
10618 indirect ways of creating type incorrect aliases which the compiler
10619 cannot detect. Examples are the use of address overlays and unchecked
10620 conversions involving composite types containing access types as
10621 components. In such cases, no warnings are generated, but there can
10622 still be aliasing problems. One safe coding practice is to forbid the
10623 use of address clauses for type overlaying, and to allow unchecked
10624 conversion only for primitive types. This is not really a significant
10625 restriction since any possible desired effect can be achieved by
10626 unchecked conversion of access values.
10628 The aliasing analysis done in strict aliasing mode can certainly
10629 have significant benefits. We have seen cases of large scale
10630 application code where the time is increased by up to 5% by turning
10631 this optimization off. If you have code that includes significant
10632 usage of unchecked conversion, you might want to just stick with
10633 @option{-O1} and avoid the entire issue. If you get adequate
10634 performance at this level of optimization level, that's probably
10635 the safest approach. If tests show that you really need higher
10636 levels of optimization, then you can experiment with @option{-O2}
10637 and @option{-O2 -fno-strict-aliasing} to see how much effect this
10638 has on size and speed of the code. If you really need to use
10639 @option{-O2} with strict aliasing in effect, then you should
10640 review any uses of unchecked conversion of access types,
10641 particularly if you are getting the warnings described above.
10644 @node Coverage Analysis
10645 @subsection Coverage Analysis
10648 GNAT supports the HP Performance Coverage Analyzer (PCA), which allows
10649 the user to determine the distribution of execution time across a program,
10650 @pxref{Profiling} for details of usage.
10654 @node Text_IO Suggestions
10655 @section @code{Text_IO} Suggestions
10656 @cindex @code{Text_IO} and performance
10659 The @code{Ada.Text_IO} package has fairly high overheads due in part to
10660 the requirement of maintaining page and line counts. If performance
10661 is critical, a recommendation is to use @code{Stream_IO} instead of
10662 @code{Text_IO} for volume output, since this package has less overhead.
10664 If @code{Text_IO} must be used, note that by default output to the standard
10665 output and standard error files is unbuffered (this provides better
10666 behavior when output statements are used for debugging, or if the
10667 progress of a program is observed by tracking the output, e.g. by
10668 using the Unix @command{tail -f} command to watch redirected output.
10670 If you are generating large volumes of output with @code{Text_IO} and
10671 performance is an important factor, use a designated file instead
10672 of the standard output file, or change the standard output file to
10673 be buffered using @code{Interfaces.C_Streams.setvbuf}.
10677 @node Reducing Size of Ada Executables with gnatelim
10678 @section Reducing Size of Ada Executables with @code{gnatelim}
10682 This section describes @command{gnatelim}, a tool which detects unused
10683 subprograms and helps the compiler to create a smaller executable for your
10688 * Running gnatelim::
10689 * Processing Precompiled Libraries::
10690 * Correcting the List of Eliminate Pragmas::
10691 * Making Your Executables Smaller::
10692 * Summary of the gnatelim Usage Cycle::
10695 @node About gnatelim
10696 @subsection About @code{gnatelim}
10699 When a program shares a set of Ada
10700 packages with other programs, it may happen that this program uses
10701 only a fraction of the subprograms defined in these packages. The code
10702 created for these unused subprograms increases the size of the executable.
10704 @code{gnatelim} tracks unused subprograms in an Ada program and
10705 outputs a list of GNAT-specific pragmas @code{Eliminate} marking all the
10706 subprograms that are declared but never called. By placing the list of
10707 @code{Eliminate} pragmas in the GNAT configuration file @file{gnat.adc} and
10708 recompiling your program, you may decrease the size of its executable,
10709 because the compiler will not generate the code for 'eliminated' subprograms.
10710 @xref{Pragma Eliminate,,, gnat_rm, GNAT Reference Manual}, for more
10711 information about this pragma.
10713 @code{gnatelim} needs as its input data the name of the main subprogram.
10715 If a set of source files is specified as @code{gnatelim} arguments, it
10716 treats these files as a complete set of sources making up a program to
10717 analyse, and analyses only these sources.
10719 After a full successful build of the main subprogram @code{gnatelim} can be
10720 called without specifying sources to analyse, in this case it computes
10721 the source closure of the main unit from the @file{ALI} files.
10723 The following command will create the set of @file{ALI} files needed for
10727 $ gnatmake ^-c Main_Prog^/ACTIONS=COMPILE MAIN_PROG^
10730 Note that @code{gnatelim} does not need object files.
10732 @node Running gnatelim
10733 @subsection Running @code{gnatelim}
10736 @code{gnatelim} has the following command-line interface:
10739 $ gnatelim [@var{switches}] ^-main^?MAIN^=@var{main_unit_name} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
10743 @var{main_unit_name} should be a name of a source file that contains the main
10744 subprogram of a program (partition).
10746 Each @var{filename} is the name (including the extension) of a source
10747 file to process. ``Wildcards'' are allowed, and
10748 the file name may contain path information.
10750 @samp{@var{gcc_switches}} is a list of switches for
10751 @command{gcc}. They will be passed on to all compiler invocations made by
10752 @command{gnatelim} to generate the ASIS trees. Here you can provide
10753 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
10754 use the @option{-gnatec} switch to set the configuration file etc.
10756 @code{gnatelim} has the following switches:
10760 @item ^-files^/FILES^=@var{filename}
10761 @cindex @option{^-files^/FILES^} (@code{gnatelim})
10762 Take the argument source files from the specified file. This file should be an
10763 ordinary text file containing file names separated by spaces or
10764 line breaks. You can use this switch more than once in the same call to
10765 @command{gnatelim}. You also can combine this switch with
10766 an explicit list of files.
10769 @cindex @option{^-log^/LOG^} (@command{gnatelim})
10770 Duplicate all the output sent to @file{stderr} into a log file. The log file
10771 is named @file{gnatelim.log} and is located in the current directory.
10773 @item ^-log^/LOGFILE^=@var{filename}
10774 @cindex @option{^-log^/LOGFILE^} (@command{gnatelim})
10775 Duplicate all the output sent to @file{stderr} into a specified log file.
10777 @cindex @option{^--no-elim-dispatch^/NO_DISPATCH^} (@command{gnatelim})
10778 @item ^--no-elim-dispatch^/NO_DISPATCH^
10779 Do not generate pragmas for dispatching operations.
10781 @cindex @option{^-o^/OUTPUT^} (@command{gnatelim})
10782 @item ^-o^/OUTPUT^=@var{report_file}
10783 Put @command{gnatelim} output into a specified file. If this file already exists,
10784 it is overridden. If this switch is not used, @command{gnatelim} outputs its results
10788 @cindex @option{^-q^/QUIET^} (@command{gnatelim})
10789 Quiet mode: by default @code{gnatelim} outputs to the standard error
10790 stream the number of program units left to be processed. This option turns
10793 @cindex @option{^-t^/TIME^} (@command{gnatelim})
10795 Print out execution time.
10797 @item ^-v^/VERBOSE^
10798 @cindex @option{^-v^/VERBOSE^} (@command{gnatelim})
10799 Verbose mode: @code{gnatelim} version information is printed as Ada
10800 comments to the standard output stream. Also, in addition to the number of
10801 program units left @code{gnatelim} will output the name of the current unit
10804 @item ^-wq^/WARNINGS=QUIET^
10805 @cindex @option{^-wq^/WARNINGS=QUIET^} (@command{gnatelim})
10806 Quet warning mode - some warnings are suppressed. In particular warnings that
10807 indicate that the analysed set of sources is incomplete to make up a
10808 partition and that some subprogram bodies are missing are not generated.
10811 @node Processing Precompiled Libraries
10812 @subsection Processing Precompiled Libraries
10815 If some program uses a precompiled Ada library, it can be processed by
10816 @code{gnatelim} in a usual way. @code{gnatelim} will newer generate an
10817 Eliminate pragma for a subprogram if the body of this subprogram has not
10818 been analysed, this is a typical case for subprograms from precompiled
10819 libraries. Switch @option{^-wq^/WARNINGS=QUIET^} may be used to suppress
10820 warnings about missing source files and non-analyzed subprogram bodies
10821 that can be generated when processing precompiled Ada libraries.
10823 @node Correcting the List of Eliminate Pragmas
10824 @subsection Correcting the List of Eliminate Pragmas
10827 In some rare cases @code{gnatelim} may try to eliminate
10828 subprograms that are actually called in the program. In this case, the
10829 compiler will generate an error message of the form:
10832 main.adb:4:08: cannot reference subprogram "P" eliminated at elim.out:5
10836 You will need to manually remove the wrong @code{Eliminate} pragmas from
10837 the configuration file indicated in the error message. You should recompile
10838 your program from scratch after that, because you need a consistent
10839 configuration file(s) during the entire compilation.
10841 @node Making Your Executables Smaller
10842 @subsection Making Your Executables Smaller
10845 In order to get a smaller executable for your program you now have to
10846 recompile the program completely with the configuration file containing
10847 pragmas Eliminate generated by gnatelim. If these pragmas are placed in
10848 @file{gnat.adc} file located in your current directory, just do:
10851 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10855 (Use the @option{^-f^/FORCE_COMPILE^} option for @command{gnatmake} to
10856 recompile everything
10857 with the set of pragmas @code{Eliminate} that you have obtained with
10858 @command{gnatelim}).
10860 Be aware that the set of @code{Eliminate} pragmas is specific to each
10861 program. It is not recommended to merge sets of @code{Eliminate}
10862 pragmas created for different programs in one configuration file.
10864 @node Summary of the gnatelim Usage Cycle
10865 @subsection Summary of the @code{gnatelim} Usage Cycle
10868 Here is a quick summary of the steps to be taken in order to reduce
10869 the size of your executables with @code{gnatelim}. You may use
10870 other GNAT options to control the optimization level,
10871 to produce the debugging information, to set search path, etc.
10875 Create a complete set of @file{ALI} files (if the program has not been
10879 $ gnatmake ^-c main_prog^/ACTIONS=COMPILE MAIN_PROG^
10883 Generate a list of @code{Eliminate} pragmas in default configuration file
10884 @file{gnat.adc} in the current directory
10887 $ PIPE GNAT ELIM MAIN_PROG > GNAT.ADC
10890 $ gnatelim main_prog >@r{[}>@r{]} gnat.adc
10895 Recompile the application
10898 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10903 @node Reducing Size of Executables with unused subprogram/data elimination
10904 @section Reducing Size of Executables with Unused Subprogram/Data Elimination
10905 @findex unused subprogram/data elimination
10908 This section describes how you can eliminate unused subprograms and data from
10909 your executable just by setting options at compilation time.
10912 * About unused subprogram/data elimination::
10913 * Compilation options::
10914 * Example of unused subprogram/data elimination::
10917 @node About unused subprogram/data elimination
10918 @subsection About unused subprogram/data elimination
10921 By default, an executable contains all code and data of its composing objects
10922 (directly linked or coming from statically linked libraries), even data or code
10923 never used by this executable.
10925 This feature will allow you to eliminate such unused code from your
10926 executable, making it smaller (in disk and in memory).
10928 This functionality is available on all Linux platforms except for the IA-64
10929 architecture and on all cross platforms using the ELF binary file format.
10930 In both cases GNU binutils version 2.16 or later are required to enable it.
10932 @node Compilation options
10933 @subsection Compilation options
10936 The operation of eliminating the unused code and data from the final executable
10937 is directly performed by the linker.
10939 In order to do this, it has to work with objects compiled with the
10941 @option{-ffunction-sections} @option{-fdata-sections}.
10942 @cindex @option{-ffunction-sections} (@command{gcc})
10943 @cindex @option{-fdata-sections} (@command{gcc})
10944 These options are usable with C and Ada files.
10945 They will place respectively each
10946 function or data in a separate section in the resulting object file.
10948 Once the objects and static libraries are created with these options, the
10949 linker can perform the dead code elimination. You can do this by setting
10950 the @option{-Wl,--gc-sections} option to gcc command or in the
10951 @option{-largs} section of @command{gnatmake}. This will perform a
10952 garbage collection of code and data never referenced.
10954 If the linker performs a partial link (@option{-r} ld linker option), then you
10955 will need to provide one or several entry point using the
10956 @option{-e} / @option{--entry} ld option.
10958 Note that objects compiled without the @option{-ffunction-sections} and
10959 @option{-fdata-sections} options can still be linked with the executable.
10960 However, no dead code elimination will be performed on those objects (they will
10963 The GNAT static library is now compiled with -ffunction-sections and
10964 -fdata-sections on some platforms. This allows you to eliminate the unused code
10965 and data of the GNAT library from your executable.
10967 @node Example of unused subprogram/data elimination
10968 @subsection Example of unused subprogram/data elimination
10971 Here is a simple example:
10973 @smallexample @c ada
10982 Used_Data : Integer;
10983 Unused_Data : Integer;
10985 procedure Used (Data : Integer);
10986 procedure Unused (Data : Integer);
10989 package body Aux is
10990 procedure Used (Data : Integer) is
10995 procedure Unused (Data : Integer) is
10997 Unused_Data := Data;
11003 @code{Unused} and @code{Unused_Data} are never referenced in this code
11004 excerpt, and hence they may be safely removed from the final executable.
11009 $ nm test | grep used
11010 020015f0 T aux__unused
11011 02005d88 B aux__unused_data
11012 020015cc T aux__used
11013 02005d84 B aux__used_data
11015 $ gnatmake test -cargs -fdata-sections -ffunction-sections \
11016 -largs -Wl,--gc-sections
11018 $ nm test | grep used
11019 02005350 T aux__used
11020 0201ffe0 B aux__used_data
11024 It can be observed that the procedure @code{Unused} and the object
11025 @code{Unused_Data} are removed by the linker when using the
11026 appropriate options.
11028 @c ********************************
11029 @node Renaming Files Using gnatchop
11030 @chapter Renaming Files Using @code{gnatchop}
11034 This chapter discusses how to handle files with multiple units by using
11035 the @code{gnatchop} utility. This utility is also useful in renaming
11036 files to meet the standard GNAT default file naming conventions.
11039 * Handling Files with Multiple Units::
11040 * Operating gnatchop in Compilation Mode::
11041 * Command Line for gnatchop::
11042 * Switches for gnatchop::
11043 * Examples of gnatchop Usage::
11046 @node Handling Files with Multiple Units
11047 @section Handling Files with Multiple Units
11050 The basic compilation model of GNAT requires that a file submitted to the
11051 compiler have only one unit and there be a strict correspondence
11052 between the file name and the unit name.
11054 The @code{gnatchop} utility allows both of these rules to be relaxed,
11055 allowing GNAT to process files which contain multiple compilation units
11056 and files with arbitrary file names. @code{gnatchop}
11057 reads the specified file and generates one or more output files,
11058 containing one unit per file. The unit and the file name correspond,
11059 as required by GNAT.
11061 If you want to permanently restructure a set of ``foreign'' files so that
11062 they match the GNAT rules, and do the remaining development using the
11063 GNAT structure, you can simply use @command{gnatchop} once, generate the
11064 new set of files and work with them from that point on.
11066 Alternatively, if you want to keep your files in the ``foreign'' format,
11067 perhaps to maintain compatibility with some other Ada compilation
11068 system, you can set up a procedure where you use @command{gnatchop} each
11069 time you compile, regarding the source files that it writes as temporary
11070 files that you throw away.
11072 Note that if your file containing multiple units starts with a byte order
11073 mark (BOM) specifying UTF-8 encoding, then the files generated by gnatchop
11074 will each start with a copy of this BOM, meaning that they can be compiled
11075 automatically in UTF-8 mode without needing to specify an explicit encoding.
11077 @node Operating gnatchop in Compilation Mode
11078 @section Operating gnatchop in Compilation Mode
11081 The basic function of @code{gnatchop} is to take a file with multiple units
11082 and split it into separate files. The boundary between files is reasonably
11083 clear, except for the issue of comments and pragmas. In default mode, the
11084 rule is that any pragmas between units belong to the previous unit, except
11085 that configuration pragmas always belong to the following unit. Any comments
11086 belong to the following unit. These rules
11087 almost always result in the right choice of
11088 the split point without needing to mark it explicitly and most users will
11089 find this default to be what they want. In this default mode it is incorrect to
11090 submit a file containing only configuration pragmas, or one that ends in
11091 configuration pragmas, to @code{gnatchop}.
11093 However, using a special option to activate ``compilation mode'',
11095 can perform another function, which is to provide exactly the semantics
11096 required by the RM for handling of configuration pragmas in a compilation.
11097 In the absence of configuration pragmas (at the main file level), this
11098 option has no effect, but it causes such configuration pragmas to be handled
11099 in a quite different manner.
11101 First, in compilation mode, if @code{gnatchop} is given a file that consists of
11102 only configuration pragmas, then this file is appended to the
11103 @file{gnat.adc} file in the current directory. This behavior provides
11104 the required behavior described in the RM for the actions to be taken
11105 on submitting such a file to the compiler, namely that these pragmas
11106 should apply to all subsequent compilations in the same compilation
11107 environment. Using GNAT, the current directory, possibly containing a
11108 @file{gnat.adc} file is the representation
11109 of a compilation environment. For more information on the
11110 @file{gnat.adc} file, see @ref{Handling of Configuration Pragmas}.
11112 Second, in compilation mode, if @code{gnatchop}
11113 is given a file that starts with
11114 configuration pragmas, and contains one or more units, then these
11115 configuration pragmas are prepended to each of the chopped files. This
11116 behavior provides the required behavior described in the RM for the
11117 actions to be taken on compiling such a file, namely that the pragmas
11118 apply to all units in the compilation, but not to subsequently compiled
11121 Finally, if configuration pragmas appear between units, they are appended
11122 to the previous unit. This results in the previous unit being illegal,
11123 since the compiler does not accept configuration pragmas that follow
11124 a unit. This provides the required RM behavior that forbids configuration
11125 pragmas other than those preceding the first compilation unit of a
11128 For most purposes, @code{gnatchop} will be used in default mode. The
11129 compilation mode described above is used only if you need exactly
11130 accurate behavior with respect to compilations, and you have files
11131 that contain multiple units and configuration pragmas. In this
11132 circumstance the use of @code{gnatchop} with the compilation mode
11133 switch provides the required behavior, and is for example the mode
11134 in which GNAT processes the ACVC tests.
11136 @node Command Line for gnatchop
11137 @section Command Line for @code{gnatchop}
11140 The @code{gnatchop} command has the form:
11143 @c $ gnatchop switches @var{file name} @r{[}@var{file name} @dots{}@r{]}
11144 @c @ovar{directory}
11145 @c Expanding @ovar macro inline (explanation in macro def comments)
11146 $ gnatchop switches @var{file name} @r{[}@var{file name} @dots{}@r{]}
11147 @r{[}@var{directory}@r{]}
11151 The only required argument is the file name of the file to be chopped.
11152 There are no restrictions on the form of this file name. The file itself
11153 contains one or more Ada units, in normal GNAT format, concatenated
11154 together. As shown, more than one file may be presented to be chopped.
11156 When run in default mode, @code{gnatchop} generates one output file in
11157 the current directory for each unit in each of the files.
11159 @var{directory}, if specified, gives the name of the directory to which
11160 the output files will be written. If it is not specified, all files are
11161 written to the current directory.
11163 For example, given a
11164 file called @file{hellofiles} containing
11166 @smallexample @c ada
11171 with Text_IO; use Text_IO;
11174 Put_Line ("Hello");
11184 $ gnatchop ^hellofiles^HELLOFILES.^
11188 generates two files in the current directory, one called
11189 @file{hello.ads} containing the single line that is the procedure spec,
11190 and the other called @file{hello.adb} containing the remaining text. The
11191 original file is not affected. The generated files can be compiled in
11195 When gnatchop is invoked on a file that is empty or that contains only empty
11196 lines and/or comments, gnatchop will not fail, but will not produce any
11199 For example, given a
11200 file called @file{toto.txt} containing
11202 @smallexample @c ada
11214 $ gnatchop ^toto.txt^TOT.TXT^
11218 will not produce any new file and will result in the following warnings:
11221 toto.txt:1:01: warning: empty file, contains no compilation units
11222 no compilation units found
11223 no source files written
11226 @node Switches for gnatchop
11227 @section Switches for @code{gnatchop}
11230 @command{gnatchop} recognizes the following switches:
11236 @cindex @option{--version} @command{gnatchop}
11237 Display Copyright and version, then exit disregarding all other options.
11240 @cindex @option{--help} @command{gnatchop}
11241 If @option{--version} was not used, display usage, then exit disregarding
11244 @item ^-c^/COMPILATION^
11245 @cindex @option{^-c^/COMPILATION^} (@code{gnatchop})
11246 Causes @code{gnatchop} to operate in compilation mode, in which
11247 configuration pragmas are handled according to strict RM rules. See
11248 previous section for a full description of this mode.
11251 @item -gnat@var{xxx}
11252 This passes the given @option{-gnat@var{xxx}} switch to @code{gnat} which is
11253 used to parse the given file. Not all @var{xxx} options make sense,
11254 but for example, the use of @option{-gnati2} allows @code{gnatchop} to
11255 process a source file that uses Latin-2 coding for identifiers.
11259 Causes @code{gnatchop} to generate a brief help summary to the standard
11260 output file showing usage information.
11262 @item ^-k@var{mm}^/FILE_NAME_MAX_LENGTH=@var{mm}^
11263 @cindex @option{^-k^/FILE_NAME_MAX_LENGTH^} (@code{gnatchop})
11264 Limit generated file names to the specified number @code{mm}
11266 This is useful if the
11267 resulting set of files is required to be interoperable with systems
11268 which limit the length of file names.
11270 If no value is given, or
11271 if no @code{/FILE_NAME_MAX_LENGTH} qualifier is given,
11272 a default of 39, suitable for OpenVMS Alpha
11273 Systems, is assumed
11276 No space is allowed between the @option{-k} and the numeric value. The numeric
11277 value may be omitted in which case a default of @option{-k8},
11279 with DOS-like file systems, is used. If no @option{-k} switch
11281 there is no limit on the length of file names.
11284 @item ^-p^/PRESERVE^
11285 @cindex @option{^-p^/PRESERVE^} (@code{gnatchop})
11286 Causes the file ^modification^creation^ time stamp of the input file to be
11287 preserved and used for the time stamp of the output file(s). This may be
11288 useful for preserving coherency of time stamps in an environment where
11289 @code{gnatchop} is used as part of a standard build process.
11292 @cindex @option{^-q^/QUIET^} (@code{gnatchop})
11293 Causes output of informational messages indicating the set of generated
11294 files to be suppressed. Warnings and error messages are unaffected.
11296 @item ^-r^/REFERENCE^
11297 @cindex @option{^-r^/REFERENCE^} (@code{gnatchop})
11298 @findex Source_Reference
11299 Generate @code{Source_Reference} pragmas. Use this switch if the output
11300 files are regarded as temporary and development is to be done in terms
11301 of the original unchopped file. This switch causes
11302 @code{Source_Reference} pragmas to be inserted into each of the
11303 generated files to refers back to the original file name and line number.
11304 The result is that all error messages refer back to the original
11306 In addition, the debugging information placed into the object file (when
11307 the @option{^-g^/DEBUG^} switch of @command{gcc} or @command{gnatmake} is
11309 also refers back to this original file so that tools like profilers and
11310 debuggers will give information in terms of the original unchopped file.
11312 If the original file to be chopped itself contains
11313 a @code{Source_Reference}
11314 pragma referencing a third file, then gnatchop respects
11315 this pragma, and the generated @code{Source_Reference} pragmas
11316 in the chopped file refer to the original file, with appropriate
11317 line numbers. This is particularly useful when @code{gnatchop}
11318 is used in conjunction with @code{gnatprep} to compile files that
11319 contain preprocessing statements and multiple units.
11321 @item ^-v^/VERBOSE^
11322 @cindex @option{^-v^/VERBOSE^} (@code{gnatchop})
11323 Causes @code{gnatchop} to operate in verbose mode. The version
11324 number and copyright notice are output, as well as exact copies of
11325 the gnat1 commands spawned to obtain the chop control information.
11327 @item ^-w^/OVERWRITE^
11328 @cindex @option{^-w^/OVERWRITE^} (@code{gnatchop})
11329 Overwrite existing file names. Normally @code{gnatchop} regards it as a
11330 fatal error if there is already a file with the same name as a
11331 file it would otherwise output, in other words if the files to be
11332 chopped contain duplicated units. This switch bypasses this
11333 check, and causes all but the last instance of such duplicated
11334 units to be skipped.
11337 @item --GCC=@var{xxxx}
11338 @cindex @option{--GCC=} (@code{gnatchop})
11339 Specify the path of the GNAT parser to be used. When this switch is used,
11340 no attempt is made to add the prefix to the GNAT parser executable.
11344 @node Examples of gnatchop Usage
11345 @section Examples of @code{gnatchop} Usage
11349 @item gnatchop /OVERWRITE HELLO_S.ADA [PRERELEASE.FILES]
11352 @item gnatchop -w hello_s.ada prerelease/files
11355 Chops the source file @file{hello_s.ada}. The output files will be
11356 placed in the directory @file{^prerelease/files^[PRERELEASE.FILES]^},
11358 files with matching names in that directory (no files in the current
11359 directory are modified).
11361 @item gnatchop ^archive^ARCHIVE.^
11362 Chops the source file @file{^archive^ARCHIVE.^}
11363 into the current directory. One
11364 useful application of @code{gnatchop} is in sending sets of sources
11365 around, for example in email messages. The required sources are simply
11366 concatenated (for example, using a ^Unix @code{cat}^VMS @code{APPEND/NEW}^
11368 @command{gnatchop} is used at the other end to reconstitute the original
11371 @item gnatchop file1 file2 file3 direc
11372 Chops all units in files @file{file1}, @file{file2}, @file{file3}, placing
11373 the resulting files in the directory @file{direc}. Note that if any units
11374 occur more than once anywhere within this set of files, an error message
11375 is generated, and no files are written. To override this check, use the
11376 @option{^-w^/OVERWRITE^} switch,
11377 in which case the last occurrence in the last file will
11378 be the one that is output, and earlier duplicate occurrences for a given
11379 unit will be skipped.
11382 @node Configuration Pragmas
11383 @chapter Configuration Pragmas
11384 @cindex Configuration pragmas
11385 @cindex Pragmas, configuration
11388 Configuration pragmas include those pragmas described as
11389 such in the Ada Reference Manual, as well as
11390 implementation-dependent pragmas that are configuration pragmas.
11391 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
11392 for details on these additional GNAT-specific configuration pragmas.
11393 Most notably, the pragma @code{Source_File_Name}, which allows
11394 specifying non-default names for source files, is a configuration
11395 pragma. The following is a complete list of configuration pragmas
11396 recognized by GNAT:
11406 Assume_No_Invalid_Values
11411 Compile_Time_Warning
11413 Component_Alignment
11414 Convention_Identifier
11422 External_Name_Casing
11425 Float_Representation
11438 Priority_Specific_Dispatching
11441 Propagate_Exceptions
11444 Restricted_Run_Time
11446 Restrictions_Warnings
11448 Short_Circuit_And_Or
11450 Source_File_Name_Project
11453 Suppress_Exception_Locations
11454 Task_Dispatching_Policy
11460 Wide_Character_Encoding
11465 * Handling of Configuration Pragmas::
11466 * The Configuration Pragmas Files::
11469 @node Handling of Configuration Pragmas
11470 @section Handling of Configuration Pragmas
11472 Configuration pragmas may either appear at the start of a compilation
11473 unit, in which case they apply only to that unit, or they may apply to
11474 all compilations performed in a given compilation environment.
11476 GNAT also provides the @code{gnatchop} utility to provide an automatic
11477 way to handle configuration pragmas following the semantics for
11478 compilations (that is, files with multiple units), described in the RM.
11479 See @ref{Operating gnatchop in Compilation Mode} for details.
11480 However, for most purposes, it will be more convenient to edit the
11481 @file{gnat.adc} file that contains configuration pragmas directly,
11482 as described in the following section.
11484 @node The Configuration Pragmas Files
11485 @section The Configuration Pragmas Files
11486 @cindex @file{gnat.adc}
11489 In GNAT a compilation environment is defined by the current
11490 directory at the time that a compile command is given. This current
11491 directory is searched for a file whose name is @file{gnat.adc}. If
11492 this file is present, it is expected to contain one or more
11493 configuration pragmas that will be applied to the current compilation.
11494 However, if the switch @option{-gnatA} is used, @file{gnat.adc} is not
11497 Configuration pragmas may be entered into the @file{gnat.adc} file
11498 either by running @code{gnatchop} on a source file that consists only of
11499 configuration pragmas, or more conveniently by
11500 direct editing of the @file{gnat.adc} file, which is a standard format
11503 In addition to @file{gnat.adc}, additional files containing configuration
11504 pragmas may be applied to the current compilation using the switch
11505 @option{-gnatec}@var{path}. @var{path} must designate an existing file that
11506 contains only configuration pragmas. These configuration pragmas are
11507 in addition to those found in @file{gnat.adc} (provided @file{gnat.adc}
11508 is present and switch @option{-gnatA} is not used).
11510 It is allowed to specify several switches @option{-gnatec}, all of which
11511 will be taken into account.
11513 If you are using project file, a separate mechanism is provided using
11514 project attributes, see @ref{Specifying Configuration Pragmas} for more
11518 Of special interest to GNAT OpenVMS Alpha is the following
11519 configuration pragma:
11521 @smallexample @c ada
11523 pragma Extend_System (Aux_DEC);
11528 In the presence of this pragma, GNAT adds to the definition of the
11529 predefined package SYSTEM all the additional types and subprograms that are
11530 defined in HP Ada. See @ref{Compatibility with HP Ada} for details.
11533 @node Handling Arbitrary File Naming Conventions Using gnatname
11534 @chapter Handling Arbitrary File Naming Conventions Using @code{gnatname}
11535 @cindex Arbitrary File Naming Conventions
11538 * Arbitrary File Naming Conventions::
11539 * Running gnatname::
11540 * Switches for gnatname::
11541 * Examples of gnatname Usage::
11544 @node Arbitrary File Naming Conventions
11545 @section Arbitrary File Naming Conventions
11548 The GNAT compiler must be able to know the source file name of a compilation
11549 unit. When using the standard GNAT default file naming conventions
11550 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
11551 does not need additional information.
11554 When the source file names do not follow the standard GNAT default file naming
11555 conventions, the GNAT compiler must be given additional information through
11556 a configuration pragmas file (@pxref{Configuration Pragmas})
11558 When the non-standard file naming conventions are well-defined,
11559 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
11560 (@pxref{Alternative File Naming Schemes}) may be sufficient. However,
11561 if the file naming conventions are irregular or arbitrary, a number
11562 of pragma @code{Source_File_Name} for individual compilation units
11564 To help maintain the correspondence between compilation unit names and
11565 source file names within the compiler,
11566 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
11569 @node Running gnatname
11570 @section Running @code{gnatname}
11573 The usual form of the @code{gnatname} command is
11576 @c $ gnatname @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}
11577 @c @r{[}--and @ovar{switches} @var{naming_pattern} @ovar{naming_patterns}@r{]}
11578 @c Expanding @ovar macro inline (explanation in macro def comments)
11579 $ gnatname @r{[}@var{switches}@r{]} @var{naming_pattern} @r{[}@var{naming_patterns}@r{]}
11580 @r{[}--and @r{[}@var{switches}@r{]} @var{naming_pattern} @r{[}@var{naming_patterns}@r{]}@r{]}
11584 All of the arguments are optional. If invoked without any argument,
11585 @code{gnatname} will display its usage.
11588 When used with at least one naming pattern, @code{gnatname} will attempt to
11589 find all the compilation units in files that follow at least one of the
11590 naming patterns. To find these compilation units,
11591 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
11595 One or several Naming Patterns may be given as arguments to @code{gnatname}.
11596 Each Naming Pattern is enclosed between double quotes (or single
11597 quotes on Windows).
11598 A Naming Pattern is a regular expression similar to the wildcard patterns
11599 used in file names by the Unix shells or the DOS prompt.
11602 @code{gnatname} may be called with several sections of directories/patterns.
11603 Sections are separated by switch @code{--and}. In each section, there must be
11604 at least one pattern. If no directory is specified in a section, the current
11605 directory (or the project directory is @code{-P} is used) is implied.
11606 The options other that the directory switches and the patterns apply globally
11607 even if they are in different sections.
11610 Examples of Naming Patterns are
11619 For a more complete description of the syntax of Naming Patterns,
11620 see the second kind of regular expressions described in @file{g-regexp.ads}
11621 (the ``Glob'' regular expressions).
11624 When invoked with no switch @code{-P}, @code{gnatname} will create a
11625 configuration pragmas file @file{gnat.adc} in the current working directory,
11626 with pragmas @code{Source_File_Name} for each file that contains a valid Ada
11629 @node Switches for gnatname
11630 @section Switches for @code{gnatname}
11633 Switches for @code{gnatname} must precede any specified Naming Pattern.
11636 You may specify any of the following switches to @code{gnatname}:
11642 @cindex @option{--version} @command{gnatname}
11643 Display Copyright and version, then exit disregarding all other options.
11646 @cindex @option{--help} @command{gnatname}
11647 If @option{--version} was not used, display usage, then exit disregarding
11651 Start another section of directories/patterns.
11653 @item ^-c^/CONFIG_FILE=^@file{file}
11654 @cindex @option{^-c^/CONFIG_FILE^} (@code{gnatname})
11655 Create a configuration pragmas file @file{file} (instead of the default
11658 There may be zero, one or more space between @option{-c} and
11661 @file{file} may include directory information. @file{file} must be
11662 writable. There may be only one switch @option{^-c^/CONFIG_FILE^}.
11663 When a switch @option{^-c^/CONFIG_FILE^} is
11664 specified, no switch @option{^-P^/PROJECT_FILE^} may be specified (see below).
11666 @item ^-d^/SOURCE_DIRS=^@file{dir}
11667 @cindex @option{^-d^/SOURCE_DIRS^} (@code{gnatname})
11668 Look for source files in directory @file{dir}. There may be zero, one or more
11669 spaces between @option{^-d^/SOURCE_DIRS=^} and @file{dir}.
11670 When a switch @option{^-d^/SOURCE_DIRS^}
11671 is specified, the current working directory will not be searched for source
11672 files, unless it is explicitly specified with a @option{^-d^/SOURCE_DIRS^}
11673 or @option{^-D^/DIR_FILES^} switch.
11674 Several switches @option{^-d^/SOURCE_DIRS^} may be specified.
11675 If @file{dir} is a relative path, it is relative to the directory of
11676 the configuration pragmas file specified with switch
11677 @option{^-c^/CONFIG_FILE^},
11678 or to the directory of the project file specified with switch
11679 @option{^-P^/PROJECT_FILE^} or,
11680 if neither switch @option{^-c^/CONFIG_FILE^}
11681 nor switch @option{^-P^/PROJECT_FILE^} are specified, it is relative to the
11682 current working directory. The directory
11683 specified with switch @option{^-d^/SOURCE_DIRS^} must exist and be readable.
11685 @item ^-D^/DIRS_FILE=^@file{file}
11686 @cindex @option{^-D^/DIRS_FILE^} (@code{gnatname})
11687 Look for source files in all directories listed in text file @file{file}.
11688 There may be zero, one or more spaces between @option{^-D^/DIRS_FILE=^}
11690 @file{file} must be an existing, readable text file.
11691 Each nonempty line in @file{file} must be a directory.
11692 Specifying switch @option{^-D^/DIRS_FILE^} is equivalent to specifying as many
11693 switches @option{^-d^/SOURCE_DIRS^} as there are nonempty lines in
11696 @item ^-f^/FOREIGN_PATTERN=^@file{pattern}
11697 @cindex @option{^-f^/FOREIGN_PATTERN^} (@code{gnatname})
11698 Foreign patterns. Using this switch, it is possible to add sources of languages
11699 other than Ada to the list of sources of a project file.
11700 It is only useful if a ^-P^/PROJECT_FILE^ switch is used.
11703 gnatname ^-Pprj -f"*.c"^/PROJECT_FILE=PRJ /FOREIGN_PATTERN=*.C^ "*.ada"
11706 will look for Ada units in all files with the @file{.ada} extension,
11707 and will add to the list of file for project @file{prj.gpr} the C files
11708 with extension @file{.^c^C^}.
11711 @cindex @option{^-h^/HELP^} (@code{gnatname})
11712 Output usage (help) information. The output is written to @file{stdout}.
11714 @item ^-P^/PROJECT_FILE=^@file{proj}
11715 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatname})
11716 Create or update project file @file{proj}. There may be zero, one or more space
11717 between @option{-P} and @file{proj}. @file{proj} may include directory
11718 information. @file{proj} must be writable.
11719 There may be only one switch @option{^-P^/PROJECT_FILE^}.
11720 When a switch @option{^-P^/PROJECT_FILE^} is specified,
11721 no switch @option{^-c^/CONFIG_FILE^} may be specified.
11723 @item ^-v^/VERBOSE^
11724 @cindex @option{^-v^/VERBOSE^} (@code{gnatname})
11725 Verbose mode. Output detailed explanation of behavior to @file{stdout}.
11726 This includes name of the file written, the name of the directories to search
11727 and, for each file in those directories whose name matches at least one of
11728 the Naming Patterns, an indication of whether the file contains a unit,
11729 and if so the name of the unit.
11731 @item ^-v -v^/VERBOSE /VERBOSE^
11732 @cindex @option{^-v -v^/VERBOSE /VERBOSE^} (@code{gnatname})
11733 Very Verbose mode. In addition to the output produced in verbose mode,
11734 for each file in the searched directories whose name matches none of
11735 the Naming Patterns, an indication is given that there is no match.
11737 @item ^-x^/EXCLUDED_PATTERN=^@file{pattern}
11738 @cindex @option{^-x^/EXCLUDED_PATTERN^} (@code{gnatname})
11739 Excluded patterns. Using this switch, it is possible to exclude some files
11740 that would match the name patterns. For example,
11742 gnatname ^-x "*_nt.ada"^/EXCLUDED_PATTERN=*_nt.ada^ "*.ada"
11745 will look for Ada units in all files with the @file{.ada} extension,
11746 except those whose names end with @file{_nt.ada}.
11750 @node Examples of gnatname Usage
11751 @section Examples of @code{gnatname} Usage
11755 $ gnatname /CONFIG_FILE=[HOME.ME]NAMES.ADC /SOURCE_DIRS=SOURCES "[a-z]*.ada*"
11761 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
11766 In this example, the directory @file{^/home/me^[HOME.ME]^} must already exist
11767 and be writable. In addition, the directory
11768 @file{^/home/me/sources^[HOME.ME.SOURCES]^} (specified by
11769 @option{^-d sources^/SOURCE_DIRS=SOURCES^}) must exist and be readable.
11772 Note the optional spaces after @option{-c} and @option{-d}.
11777 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
11778 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
11781 $ gnatname /PROJECT_FILE=[HOME.ME]PROJ
11782 /EXCLUDED_PATTERN=*_nt_body.ada
11783 /SOURCE_DIRS=(SOURCES,[SOURCES.PLUS])
11784 /DIRS_FILE=COMMON_DIRS.TXT "body_*" "spec_*"
11788 Note that several switches @option{^-d^/SOURCE_DIRS^} may be used,
11789 even in conjunction with one or several switches
11790 @option{^-D^/DIRS_FILE^}. Several Naming Patterns and one excluded pattern
11791 are used in this example.
11793 @c *****************************************
11794 @c * G N A T P r o j e c t M a n a g e r *
11795 @c *****************************************
11797 @c ------ macros for projects.texi
11798 @c These macros are needed when building the gprbuild documentation, but
11799 @c should have no effect in the gnat user's guide
11801 @macro CODESAMPLE{TXT}
11809 @macro PROJECTFILE{TXT}
11813 @c simulates a newline when in a @CODESAMPLE
11824 @macro TIPHTML{TXT}
11828 @macro IMPORTANT{TXT}
11843 @include projects.texi
11845 @c *****************************************
11846 @c * Cross-referencing tools
11847 @c *****************************************
11849 @node The Cross-Referencing Tools gnatxref and gnatfind
11850 @chapter The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
11855 The compiler generates cross-referencing information (unless
11856 you set the @samp{-gnatx} switch), which are saved in the @file{.ali} files.
11857 This information indicates where in the source each entity is declared and
11858 referenced. Note that entities in package Standard are not included, but
11859 entities in all other predefined units are included in the output.
11861 Before using any of these two tools, you need to compile successfully your
11862 application, so that GNAT gets a chance to generate the cross-referencing
11865 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
11866 information to provide the user with the capability to easily locate the
11867 declaration and references to an entity. These tools are quite similar,
11868 the difference being that @code{gnatfind} is intended for locating
11869 definitions and/or references to a specified entity or entities, whereas
11870 @code{gnatxref} is oriented to generating a full report of all
11873 To use these tools, you must not compile your application using the
11874 @option{-gnatx} switch on the @command{gnatmake} command line
11875 (@pxref{The GNAT Make Program gnatmake}). Otherwise, cross-referencing
11876 information will not be generated.
11878 Note: to invoke @code{gnatxref} or @code{gnatfind} with a project file,
11879 use the @code{gnat} driver (see @ref{The GNAT Driver and Project Files}).
11882 * Switches for gnatxref::
11883 * Switches for gnatfind::
11884 * Project Files for gnatxref and gnatfind::
11885 * Regular Expressions in gnatfind and gnatxref::
11886 * Examples of gnatxref Usage::
11887 * Examples of gnatfind Usage::
11890 @node Switches for gnatxref
11891 @section @code{gnatxref} Switches
11894 The command invocation for @code{gnatxref} is:
11896 @c $ gnatxref @ovar{switches} @var{sourcefile1} @r{[}@var{sourcefile2} @dots{}@r{]}
11897 @c Expanding @ovar macro inline (explanation in macro def comments)
11898 $ gnatxref @r{[}@var{switches}@r{]} @var{sourcefile1} @r{[}@var{sourcefile2} @dots{}@r{]}
11907 identifies the source files for which a report is to be generated. The
11908 ``with''ed units will be processed too. You must provide at least one file.
11910 These file names are considered to be regular expressions, so for instance
11911 specifying @file{source*.adb} is the same as giving every file in the current
11912 directory whose name starts with @file{source} and whose extension is
11915 You shouldn't specify any directory name, just base names. @command{gnatxref}
11916 and @command{gnatfind} will be able to locate these files by themselves using
11917 the source path. If you specify directories, no result is produced.
11922 The switches can be:
11926 @cindex @option{--version} @command{gnatxref}
11927 Display Copyright and version, then exit disregarding all other options.
11930 @cindex @option{--help} @command{gnatxref}
11931 If @option{--version} was not used, display usage, then exit disregarding
11934 @item ^-a^/ALL_FILES^
11935 @cindex @option{^-a^/ALL_FILES^} (@command{gnatxref})
11936 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
11937 the read-only files found in the library search path. Otherwise, these files
11938 will be ignored. This option can be used to protect Gnat sources or your own
11939 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
11940 much faster, and their output much smaller. Read-only here refers to access
11941 or permissions status in the file system for the current user.
11944 @cindex @option{-aIDIR} (@command{gnatxref})
11945 When looking for source files also look in directory DIR. The order in which
11946 source file search is undertaken is the same as for @command{gnatmake}.
11949 @cindex @option{-aODIR} (@command{gnatxref})
11950 When searching for library and object files, look in directory
11951 DIR. The order in which library files are searched is the same as for
11952 @command{gnatmake}.
11955 @cindex @option{-nostdinc} (@command{gnatxref})
11956 Do not look for sources in the system default directory.
11959 @cindex @option{-nostdlib} (@command{gnatxref})
11960 Do not look for library files in the system default directory.
11962 @item --ext=@var{extension}
11963 @cindex @option{--ext} (@command{gnatxref})
11964 Specify an alternate ali file extension. The default is @code{ali} and other
11965 extensions (e.g. @code{sli} for SPARK library files) may be specified via this
11966 switch. Note that if this switch overrides the default, which means that only
11967 the new extension will be considered.
11969 @item --RTS=@var{rts-path}
11970 @cindex @option{--RTS} (@command{gnatxref})
11971 Specifies the default location of the runtime library. Same meaning as the
11972 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
11974 @item ^-d^/DERIVED_TYPES^
11975 @cindex @option{^-d^/DERIVED_TYPES^} (@command{gnatxref})
11976 If this switch is set @code{gnatxref} will output the parent type
11977 reference for each matching derived types.
11979 @item ^-f^/FULL_PATHNAME^
11980 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatxref})
11981 If this switch is set, the output file names will be preceded by their
11982 directory (if the file was found in the search path). If this switch is
11983 not set, the directory will not be printed.
11985 @item ^-g^/IGNORE_LOCALS^
11986 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatxref})
11987 If this switch is set, information is output only for library-level
11988 entities, ignoring local entities. The use of this switch may accelerate
11989 @code{gnatfind} and @code{gnatxref}.
11992 @cindex @option{-IDIR} (@command{gnatxref})
11993 Equivalent to @samp{-aODIR -aIDIR}.
11996 @cindex @option{-pFILE} (@command{gnatxref})
11997 Specify a project file to use @xref{GNAT Project Manager}.
11998 If you need to use the @file{.gpr}
11999 project files, you should use gnatxref through the GNAT driver
12000 (@command{gnat xref -Pproject}).
12002 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
12003 project file in the current directory.
12005 If a project file is either specified or found by the tools, then the content
12006 of the source directory and object directory lines are added as if they
12007 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^}
12008 and @samp{^-aO^OBJECT_SEARCH^}.
12010 Output only unused symbols. This may be really useful if you give your
12011 main compilation unit on the command line, as @code{gnatxref} will then
12012 display every unused entity and 'with'ed package.
12016 Instead of producing the default output, @code{gnatxref} will generate a
12017 @file{tags} file that can be used by vi. For examples how to use this
12018 feature, see @ref{Examples of gnatxref Usage}. The tags file is output
12019 to the standard output, thus you will have to redirect it to a file.
12025 All these switches may be in any order on the command line, and may even
12026 appear after the file names. They need not be separated by spaces, thus
12027 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
12028 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
12030 @node Switches for gnatfind
12031 @section @code{gnatfind} Switches
12034 The command line for @code{gnatfind} is:
12037 @c $ gnatfind @ovar{switches} @var{pattern}@r{[}:@var{sourcefile}@r{[}:@var{line}@r{[}:@var{column}@r{]]]}
12038 @c @r{[}@var{file1} @var{file2} @dots{}]
12039 @c Expanding @ovar macro inline (explanation in macro def comments)
12040 $ gnatfind @r{[}@var{switches}@r{]} @var{pattern}@r{[}:@var{sourcefile}@r{[}:@var{line}@r{[}:@var{column}@r{]]]}
12041 @r{[}@var{file1} @var{file2} @dots{}@r{]}
12049 An entity will be output only if it matches the regular expression found
12050 in @var{pattern}, see @ref{Regular Expressions in gnatfind and gnatxref}.
12052 Omitting the pattern is equivalent to specifying @samp{*}, which
12053 will match any entity. Note that if you do not provide a pattern, you
12054 have to provide both a sourcefile and a line.
12056 Entity names are given in Latin-1, with uppercase/lowercase equivalence
12057 for matching purposes. At the current time there is no support for
12058 8-bit codes other than Latin-1, or for wide characters in identifiers.
12061 @code{gnatfind} will look for references, bodies or declarations
12062 of symbols referenced in @file{@var{sourcefile}}, at line @var{line}
12063 and column @var{column}. See @ref{Examples of gnatfind Usage}
12064 for syntax examples.
12067 is a decimal integer identifying the line number containing
12068 the reference to the entity (or entities) to be located.
12071 is a decimal integer identifying the exact location on the
12072 line of the first character of the identifier for the
12073 entity reference. Columns are numbered from 1.
12075 @item file1 file2 @dots{}
12076 The search will be restricted to these source files. If none are given, then
12077 the search will be done for every library file in the search path.
12078 These file must appear only after the pattern or sourcefile.
12080 These file names are considered to be regular expressions, so for instance
12081 specifying @file{source*.adb} is the same as giving every file in the current
12082 directory whose name starts with @file{source} and whose extension is
12085 The location of the spec of the entity will always be displayed, even if it
12086 isn't in one of @file{@var{file1}}, @file{@var{file2}},@enddots{} The
12087 occurrences of the entity in the separate units of the ones given on the
12088 command line will also be displayed.
12090 Note that if you specify at least one file in this part, @code{gnatfind} may
12091 sometimes not be able to find the body of the subprograms.
12096 At least one of 'sourcefile' or 'pattern' has to be present on
12099 The following switches are available:
12103 @cindex @option{--version} @command{gnatfind}
12104 Display Copyright and version, then exit disregarding all other options.
12107 @cindex @option{--help} @command{gnatfind}
12108 If @option{--version} was not used, display usage, then exit disregarding
12111 @item ^-a^/ALL_FILES^
12112 @cindex @option{^-a^/ALL_FILES^} (@command{gnatfind})
12113 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
12114 the read-only files found in the library search path. Otherwise, these files
12115 will be ignored. This option can be used to protect Gnat sources or your own
12116 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
12117 much faster, and their output much smaller. Read-only here refers to access
12118 or permission status in the file system for the current user.
12121 @cindex @option{-aIDIR} (@command{gnatfind})
12122 When looking for source files also look in directory DIR. The order in which
12123 source file search is undertaken is the same as for @command{gnatmake}.
12126 @cindex @option{-aODIR} (@command{gnatfind})
12127 When searching for library and object files, look in directory
12128 DIR. The order in which library files are searched is the same as for
12129 @command{gnatmake}.
12132 @cindex @option{-nostdinc} (@command{gnatfind})
12133 Do not look for sources in the system default directory.
12136 @cindex @option{-nostdlib} (@command{gnatfind})
12137 Do not look for library files in the system default directory.
12139 @item --ext=@var{extension}
12140 @cindex @option{--ext} (@command{gnatfind})
12141 Specify an alternate ali file extension. The default is @code{ali} and other
12142 extensions (e.g. @code{sli} for SPARK library files) may be specified via this
12143 switch. Note that if this switch overrides the default, which means that only
12144 the new extension will be considered.
12146 @item --RTS=@var{rts-path}
12147 @cindex @option{--RTS} (@command{gnatfind})
12148 Specifies the default location of the runtime library. Same meaning as the
12149 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
12151 @item ^-d^/DERIVED_TYPE_INFORMATION^
12152 @cindex @option{^-d^/DERIVED_TYPE_INFORMATION^} (@code{gnatfind})
12153 If this switch is set, then @code{gnatfind} will output the parent type
12154 reference for each matching derived types.
12156 @item ^-e^/EXPRESSIONS^
12157 @cindex @option{^-e^/EXPRESSIONS^} (@command{gnatfind})
12158 By default, @code{gnatfind} accept the simple regular expression set for
12159 @samp{pattern}. If this switch is set, then the pattern will be
12160 considered as full Unix-style regular expression.
12162 @item ^-f^/FULL_PATHNAME^
12163 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatfind})
12164 If this switch is set, the output file names will be preceded by their
12165 directory (if the file was found in the search path). If this switch is
12166 not set, the directory will not be printed.
12168 @item ^-g^/IGNORE_LOCALS^
12169 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatfind})
12170 If this switch is set, information is output only for library-level
12171 entities, ignoring local entities. The use of this switch may accelerate
12172 @code{gnatfind} and @code{gnatxref}.
12175 @cindex @option{-IDIR} (@command{gnatfind})
12176 Equivalent to @samp{-aODIR -aIDIR}.
12179 @cindex @option{-pFILE} (@command{gnatfind})
12180 Specify a project file (@pxref{GNAT Project Manager}) to use.
12181 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
12182 project file in the current directory.
12184 If a project file is either specified or found by the tools, then the content
12185 of the source directory and object directory lines are added as if they
12186 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^} and
12187 @samp{^-aO^/OBJECT_SEARCH^}.
12189 @item ^-r^/REFERENCES^
12190 @cindex @option{^-r^/REFERENCES^} (@command{gnatfind})
12191 By default, @code{gnatfind} will output only the information about the
12192 declaration, body or type completion of the entities. If this switch is
12193 set, the @code{gnatfind} will locate every reference to the entities in
12194 the files specified on the command line (or in every file in the search
12195 path if no file is given on the command line).
12197 @item ^-s^/PRINT_LINES^
12198 @cindex @option{^-s^/PRINT_LINES^} (@command{gnatfind})
12199 If this switch is set, then @code{gnatfind} will output the content
12200 of the Ada source file lines were the entity was found.
12202 @item ^-t^/TYPE_HIERARCHY^
12203 @cindex @option{^-t^/TYPE_HIERARCHY^} (@command{gnatfind})
12204 If this switch is set, then @code{gnatfind} will output the type hierarchy for
12205 the specified type. It act like -d option but recursively from parent
12206 type to parent type. When this switch is set it is not possible to
12207 specify more than one file.
12212 All these switches may be in any order on the command line, and may even
12213 appear after the file names. They need not be separated by spaces, thus
12214 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
12215 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
12217 As stated previously, gnatfind will search in every directory in the
12218 search path. You can force it to look only in the current directory if
12219 you specify @code{*} at the end of the command line.
12221 @node Project Files for gnatxref and gnatfind
12222 @section Project Files for @command{gnatxref} and @command{gnatfind}
12225 Project files allow a programmer to specify how to compile its
12226 application, where to find sources, etc. These files are used
12228 primarily by GPS, but they can also be used
12231 @code{gnatxref} and @code{gnatfind}.
12233 A project file name must end with @file{.gpr}. If a single one is
12234 present in the current directory, then @code{gnatxref} and @code{gnatfind} will
12235 extract the information from it. If multiple project files are found, none of
12236 them is read, and you have to use the @samp{-p} switch to specify the one
12239 The following lines can be included, even though most of them have default
12240 values which can be used in most cases.
12241 The lines can be entered in any order in the file.
12242 Except for @file{src_dir} and @file{obj_dir}, you can only have one instance of
12243 each line. If you have multiple instances, only the last one is taken into
12248 [default: @code{"^./^[]^"}]
12249 specifies a directory where to look for source files. Multiple @code{src_dir}
12250 lines can be specified and they will be searched in the order they
12254 [default: @code{"^./^[]^"}]
12255 specifies a directory where to look for object and library files. Multiple
12256 @code{obj_dir} lines can be specified, and they will be searched in the order
12259 @item comp_opt=SWITCHES
12260 [default: @code{""}]
12261 creates a variable which can be referred to subsequently by using
12262 the @code{$@{comp_opt@}} notation. This is intended to store the default
12263 switches given to @command{gnatmake} and @command{gcc}.
12265 @item bind_opt=SWITCHES
12266 [default: @code{""}]
12267 creates a variable which can be referred to subsequently by using
12268 the @samp{$@{bind_opt@}} notation. This is intended to store the default
12269 switches given to @command{gnatbind}.
12271 @item link_opt=SWITCHES
12272 [default: @code{""}]
12273 creates a variable which can be referred to subsequently by using
12274 the @samp{$@{link_opt@}} notation. This is intended to store the default
12275 switches given to @command{gnatlink}.
12277 @item main=EXECUTABLE
12278 [default: @code{""}]
12279 specifies the name of the executable for the application. This variable can
12280 be referred to in the following lines by using the @samp{$@{main@}} notation.
12283 @item comp_cmd=COMMAND
12284 [default: @code{"GNAT COMPILE /SEARCH=$@{src_dir@} /DEBUG /TRY_SEMANTICS"}]
12287 @item comp_cmd=COMMAND
12288 [default: @code{"gcc -c -I$@{src_dir@} -g -gnatq"}]
12290 specifies the command used to compile a single file in the application.
12293 @item make_cmd=COMMAND
12294 [default: @code{"GNAT MAKE $@{main@}
12295 /SOURCE_SEARCH=$@{src_dir@} /OBJECT_SEARCH=$@{obj_dir@}
12296 /DEBUG /TRY_SEMANTICS /COMPILER_QUALIFIERS $@{comp_opt@}
12297 /BINDER_QUALIFIERS $@{bind_opt@} /LINKER_QUALIFIERS $@{link_opt@}"}]
12300 @item make_cmd=COMMAND
12301 [default: @code{"gnatmake $@{main@} -aI$@{src_dir@}
12302 -aO$@{obj_dir@} -g -gnatq -cargs $@{comp_opt@}
12303 -bargs $@{bind_opt@} -largs $@{link_opt@}"}]
12305 specifies the command used to recompile the whole application.
12307 @item run_cmd=COMMAND
12308 [default: @code{"$@{main@}"}]
12309 specifies the command used to run the application.
12311 @item debug_cmd=COMMAND
12312 [default: @code{"gdb $@{main@}"}]
12313 specifies the command used to debug the application
12318 @command{gnatxref} and @command{gnatfind} only take into account the
12319 @code{src_dir} and @code{obj_dir} lines, and ignore the others.
12321 @node Regular Expressions in gnatfind and gnatxref
12322 @section Regular Expressions in @code{gnatfind} and @code{gnatxref}
12325 As specified in the section about @command{gnatfind}, the pattern can be a
12326 regular expression. Actually, there are to set of regular expressions
12327 which are recognized by the program:
12330 @item globbing patterns
12331 These are the most usual regular expression. They are the same that you
12332 generally used in a Unix shell command line, or in a DOS session.
12334 Here is a more formal grammar:
12341 term ::= elmt -- matches elmt
12342 term ::= elmt elmt -- concatenation (elmt then elmt)
12343 term ::= * -- any string of 0 or more characters
12344 term ::= ? -- matches any character
12345 term ::= [char @{char@}] -- matches any character listed
12346 term ::= [char - char] -- matches any character in range
12350 @item full regular expression
12351 The second set of regular expressions is much more powerful. This is the
12352 type of regular expressions recognized by utilities such a @file{grep}.
12354 The following is the form of a regular expression, expressed in Ada
12355 reference manual style BNF is as follows
12362 regexp ::= term @{| term@} -- alternation (term or term @dots{})
12364 term ::= item @{item@} -- concatenation (item then item)
12366 item ::= elmt -- match elmt
12367 item ::= elmt * -- zero or more elmt's
12368 item ::= elmt + -- one or more elmt's
12369 item ::= elmt ? -- matches elmt or nothing
12372 elmt ::= nschar -- matches given character
12373 elmt ::= [nschar @{nschar@}] -- matches any character listed
12374 elmt ::= [^^^ nschar @{nschar@}] -- matches any character not listed
12375 elmt ::= [char - char] -- matches chars in given range
12376 elmt ::= \ char -- matches given character
12377 elmt ::= . -- matches any single character
12378 elmt ::= ( regexp ) -- parens used for grouping
12380 char ::= any character, including special characters
12381 nschar ::= any character except ()[].*+?^^^
12385 Following are a few examples:
12389 will match any of the two strings @samp{abcde} and @samp{fghi},
12392 will match any string like @samp{abd}, @samp{abcd}, @samp{abccd},
12393 @samp{abcccd}, and so on,
12396 will match any string which has only lowercase characters in it (and at
12397 least one character.
12402 @node Examples of gnatxref Usage
12403 @section Examples of @code{gnatxref} Usage
12405 @subsection General Usage
12408 For the following examples, we will consider the following units:
12410 @smallexample @c ada
12416 3: procedure Foo (B : in Integer);
12423 1: package body Main is
12424 2: procedure Foo (B : in Integer) is
12435 2: procedure Print (B : Integer);
12444 The first thing to do is to recompile your application (for instance, in
12445 that case just by doing a @samp{gnatmake main}, so that GNAT generates
12446 the cross-referencing information.
12447 You can then issue any of the following commands:
12449 @item gnatxref main.adb
12450 @code{gnatxref} generates cross-reference information for main.adb
12451 and every unit 'with'ed by main.adb.
12453 The output would be:
12461 Decl: main.ads 3:20
12462 Body: main.adb 2:20
12463 Ref: main.adb 4:13 5:13 6:19
12466 Ref: main.adb 6:8 7:8
12476 Decl: main.ads 3:15
12477 Body: main.adb 2:15
12480 Body: main.adb 1:14
12483 Ref: main.adb 6:12 7:12
12487 that is the entity @code{Main} is declared in main.ads, line 2, column 9,
12488 its body is in main.adb, line 1, column 14 and is not referenced any where.
12490 The entity @code{Print} is declared in bar.ads, line 2, column 15 and it
12491 it referenced in main.adb, line 6 column 12 and line 7 column 12.
12493 @item gnatxref package1.adb package2.ads
12494 @code{gnatxref} will generates cross-reference information for
12495 package1.adb, package2.ads and any other package 'with'ed by any
12501 @subsection Using gnatxref with vi
12503 @code{gnatxref} can generate a tags file output, which can be used
12504 directly from @command{vi}. Note that the standard version of @command{vi}
12505 will not work properly with overloaded symbols. Consider using another
12506 free implementation of @command{vi}, such as @command{vim}.
12509 $ gnatxref -v gnatfind.adb > tags
12513 will generate the tags file for @code{gnatfind} itself (if the sources
12514 are in the search path!).
12516 From @command{vi}, you can then use the command @samp{:tag @var{entity}}
12517 (replacing @var{entity} by whatever you are looking for), and vi will
12518 display a new file with the corresponding declaration of entity.
12521 @node Examples of gnatfind Usage
12522 @section Examples of @code{gnatfind} Usage
12526 @item gnatfind ^-f^/FULL_PATHNAME^ xyz:main.adb
12527 Find declarations for all entities xyz referenced at least once in
12528 main.adb. The references are search in every library file in the search
12531 The directories will be printed as well (as the @samp{^-f^/FULL_PATHNAME^}
12534 The output will look like:
12536 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
12537 ^directory/^[directory]^main.adb:24:10: xyz <= body
12538 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
12542 that is to say, one of the entities xyz found in main.adb is declared at
12543 line 12 of main.ads (and its body is in main.adb), and another one is
12544 declared at line 45 of foo.ads
12546 @item gnatfind ^-fs^/FULL_PATHNAME/SOURCE_LINE^ xyz:main.adb
12547 This is the same command as the previous one, instead @code{gnatfind} will
12548 display the content of the Ada source file lines.
12550 The output will look like:
12553 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
12555 ^directory/^[directory]^main.adb:24:10: xyz <= body
12557 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
12562 This can make it easier to find exactly the location your are looking
12565 @item gnatfind ^-r^/REFERENCES^ "*x*":main.ads:123 foo.adb
12566 Find references to all entities containing an x that are
12567 referenced on line 123 of main.ads.
12568 The references will be searched only in main.ads and foo.adb.
12570 @item gnatfind main.ads:123
12571 Find declarations and bodies for all entities that are referenced on
12572 line 123 of main.ads.
12574 This is the same as @code{gnatfind "*":main.adb:123}.
12576 @item gnatfind ^mydir/^[mydir]^main.adb:123:45
12577 Find the declaration for the entity referenced at column 45 in
12578 line 123 of file main.adb in directory mydir. Note that it
12579 is usual to omit the identifier name when the column is given,
12580 since the column position identifies a unique reference.
12582 The column has to be the beginning of the identifier, and should not
12583 point to any character in the middle of the identifier.
12587 @c *********************************
12588 @node The GNAT Pretty-Printer gnatpp
12589 @chapter The GNAT Pretty-Printer @command{gnatpp}
12591 @cindex Pretty-Printer
12594 ^The @command{gnatpp} tool^GNAT PRETTY^ is an ASIS-based utility
12595 for source reformatting / pretty-printing.
12596 It takes an Ada source file as input and generates a reformatted
12598 You can specify various style directives via switches; e.g.,
12599 identifier case conventions, rules of indentation, and comment layout.
12601 To produce a reformatted file, @command{gnatpp} generates and uses the ASIS
12602 tree for the input source and thus requires the input to be syntactically and
12603 semantically legal.
12604 If this condition is not met, @command{gnatpp} will terminate with an
12605 error message; no output file will be generated.
12607 If the source files presented to @command{gnatpp} contain
12608 preprocessing directives, then the output file will
12609 correspond to the generated source after all
12610 preprocessing is carried out. There is no way
12611 using @command{gnatpp} to obtain pretty printed files that
12612 include the preprocessing directives.
12614 If the compilation unit
12615 contained in the input source depends semantically upon units located
12616 outside the current directory, you have to provide the source search path
12617 when invoking @command{gnatpp}, if these units are contained in files with
12618 names that do not follow the GNAT file naming rules, you have to provide
12619 the configuration file describing the corresponding naming scheme;
12620 see the description of the @command{gnatpp}
12621 switches below. Another possibility is to use a project file and to
12622 call @command{gnatpp} through the @command{gnat} driver
12624 The @command{gnatpp} command has the form
12627 @c $ gnatpp @ovar{switches} @var{filename}
12628 @c Expanding @ovar macro inline (explanation in macro def comments)
12629 $ gnatpp @r{[}@var{switches}@r{]} @var{filename} @r{[}-cargs @var{gcc_switches}@r{]}
12636 @var{switches} is an optional sequence of switches defining such properties as
12637 the formatting rules, the source search path, and the destination for the
12641 @var{filename} is the name (including the extension) of the source file to
12642 reformat; ``wildcards'' or several file names on the same gnatpp command are
12643 allowed. The file name may contain path information; it does not have to
12644 follow the GNAT file naming rules
12647 @samp{@var{gcc_switches}} is a list of switches for
12648 @command{gcc}. They will be passed on to all compiler invocations made by
12649 @command{gnatelim} to generate the ASIS trees. Here you can provide
12650 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
12651 use the @option{-gnatec} switch to set the configuration file etc.
12655 * Switches for gnatpp::
12656 * Formatting Rules::
12659 @node Switches for gnatpp
12660 @section Switches for @command{gnatpp}
12663 The following subsections describe the various switches accepted by
12664 @command{gnatpp}, organized by category.
12667 You specify a switch by supplying a name and generally also a value.
12668 In many cases the values for a switch with a given name are incompatible with
12670 (for example the switch that controls the casing of a reserved word may have
12671 exactly one value: upper case, lower case, or
12672 mixed case) and thus exactly one such switch can be in effect for an
12673 invocation of @command{gnatpp}.
12674 If more than one is supplied, the last one is used.
12675 However, some values for the same switch are mutually compatible.
12676 You may supply several such switches to @command{gnatpp}, but then
12677 each must be specified in full, with both the name and the value.
12678 Abbreviated forms (the name appearing once, followed by each value) are
12680 For example, to set
12681 the alignment of the assignment delimiter both in declarations and in
12682 assignment statements, you must write @option{-A2A3}
12683 (or @option{-A2 -A3}), but not @option{-A23}.
12687 In many cases the set of options for a given qualifier are incompatible with
12688 each other (for example the qualifier that controls the casing of a reserved
12689 word may have exactly one option, which specifies either upper case, lower
12690 case, or mixed case), and thus exactly one such option can be in effect for
12691 an invocation of @command{gnatpp}.
12692 If more than one is supplied, the last one is used.
12693 However, some qualifiers have options that are mutually compatible,
12694 and then you may then supply several such options when invoking
12698 In most cases, it is obvious whether or not the
12699 ^values for a switch with a given name^options for a given qualifier^
12700 are compatible with each other.
12701 When the semantics might not be evident, the summaries below explicitly
12702 indicate the effect.
12705 * Alignment Control::
12707 * Construct Layout Control::
12708 * General Text Layout Control::
12709 * Other Formatting Options::
12710 * Setting the Source Search Path::
12711 * Output File Control::
12712 * Other gnatpp Switches::
12715 @node Alignment Control
12716 @subsection Alignment Control
12717 @cindex Alignment control in @command{gnatpp}
12720 Programs can be easier to read if certain constructs are vertically aligned.
12721 By default all alignments are set ON.
12722 Through the @option{^-A0^/ALIGN=OFF^} switch you may reset the default to
12723 OFF, and then use one or more of the other
12724 ^@option{-A@var{n}} switches^@option{/ALIGN} options^
12725 to activate alignment for specific constructs.
12728 @cindex @option{^-A@var{n}^/ALIGN^} (@command{gnatpp})
12732 Set all alignments to ON
12735 @item ^-A0^/ALIGN=OFF^
12736 Set all alignments to OFF
12738 @item ^-A1^/ALIGN=COLONS^
12739 Align @code{:} in declarations
12741 @item ^-A2^/ALIGN=DECLARATIONS^
12742 Align @code{:=} in initializations in declarations
12744 @item ^-A3^/ALIGN=STATEMENTS^
12745 Align @code{:=} in assignment statements
12747 @item ^-A4^/ALIGN=ARROWS^
12748 Align @code{=>} in associations
12750 @item ^-A5^/ALIGN=COMPONENT_CLAUSES^
12751 Align @code{at} keywords in the component clauses in record
12752 representation clauses
12756 The @option{^-A^/ALIGN^} switches are mutually compatible; any combination
12759 @node Casing Control
12760 @subsection Casing Control
12761 @cindex Casing control in @command{gnatpp}
12764 @command{gnatpp} allows you to specify the casing for reserved words,
12765 pragma names, attribute designators and identifiers.
12766 For identifiers you may define a
12767 general rule for name casing but also override this rule
12768 via a set of dictionary files.
12770 Three types of casing are supported: lower case, upper case, and mixed case.
12771 Lower and upper case are self-explanatory (but since some letters in
12772 Latin1 and other GNAT-supported character sets
12773 exist only in lower-case form, an upper case conversion will have no
12775 ``Mixed case'' means that the first letter, and also each letter immediately
12776 following an underscore, are converted to their uppercase forms;
12777 all the other letters are converted to their lowercase forms.
12780 @cindex @option{^-a@var{x}^/ATTRIBUTE^} (@command{gnatpp})
12781 @item ^-aL^/ATTRIBUTE_CASING=LOWER_CASE^
12782 Attribute designators are lower case
12784 @item ^-aU^/ATTRIBUTE_CASING=UPPER_CASE^
12785 Attribute designators are upper case
12787 @item ^-aM^/ATTRIBUTE_CASING=MIXED_CASE^
12788 Attribute designators are mixed case (this is the default)
12790 @cindex @option{^-k@var{x}^/KEYWORD_CASING^} (@command{gnatpp})
12791 @item ^-kL^/KEYWORD_CASING=LOWER_CASE^
12792 Keywords (technically, these are known in Ada as @emph{reserved words}) are
12793 lower case (this is the default)
12795 @item ^-kU^/KEYWORD_CASING=UPPER_CASE^
12796 Keywords are upper case
12798 @cindex @option{^-n@var{x}^/NAME_CASING^} (@command{gnatpp})
12799 @item ^-nD^/NAME_CASING=AS_DECLARED^
12800 Name casing for defining occurrences are as they appear in the source file
12801 (this is the default)
12803 @item ^-nU^/NAME_CASING=UPPER_CASE^
12804 Names are in upper case
12806 @item ^-nL^/NAME_CASING=LOWER_CASE^
12807 Names are in lower case
12809 @item ^-nM^/NAME_CASING=MIXED_CASE^
12810 Names are in mixed case
12812 @cindex @option{^-p@var{x}^/PRAGMA_CASING^} (@command{gnatpp})
12813 @item ^-pL^/PRAGMA_CASING=LOWER_CASE^
12814 Pragma names are lower case
12816 @item ^-pU^/PRAGMA_CASING=UPPER_CASE^
12817 Pragma names are upper case
12819 @item ^-pM^/PRAGMA_CASING=MIXED_CASE^
12820 Pragma names are mixed case (this is the default)
12822 @item ^-D@var{file}^/DICTIONARY=@var{file}^
12823 @cindex @option{^-D^/DICTIONARY^} (@command{gnatpp})
12824 Use @var{file} as a @emph{dictionary file} that defines
12825 the casing for a set of specified names,
12826 thereby overriding the effect on these names by
12827 any explicit or implicit
12828 ^-n^/NAME_CASING^ switch.
12829 To supply more than one dictionary file,
12830 use ^several @option{-D} switches^a list of files as options^.
12833 @option{gnatpp} implicitly uses a @emph{default dictionary file}
12834 to define the casing for the Ada predefined names and
12835 the names declared in the GNAT libraries.
12837 @item ^-D-^/SPECIFIC_CASING^
12838 @cindex @option{^-D-^/SPECIFIC_CASING^} (@command{gnatpp})
12839 Do not use the default dictionary file;
12840 instead, use the casing
12841 defined by a @option{^-n^/NAME_CASING^} switch and any explicit
12846 The structure of a dictionary file, and details on the conventions
12847 used in the default dictionary file, are defined in @ref{Name Casing}.
12849 The @option{^-D-^/SPECIFIC_CASING^} and
12850 @option{^-D@var{file}^/DICTIONARY=@var{file}^} switches are mutually
12853 @node Construct Layout Control
12854 @subsection Construct Layout Control
12855 @cindex Layout control in @command{gnatpp}
12858 This group of @command{gnatpp} switches controls the layout of comments and
12859 complex syntactic constructs. See @ref{Formatting Comments} for details
12863 @cindex @option{^-c@var{n}^/COMMENTS_LAYOUT^} (@command{gnatpp})
12864 @item ^-c0^/COMMENTS_LAYOUT=UNTOUCHED^
12865 All the comments remain unchanged
12867 @item ^-c1^/COMMENTS_LAYOUT=DEFAULT^
12868 GNAT-style comment line indentation (this is the default).
12870 @item ^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^
12871 Reference-manual comment line indentation.
12873 @item ^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^
12874 GNAT-style comment beginning
12876 @item ^-c4^/COMMENTS_LAYOUT=REFORMAT^
12877 Reformat comment blocks
12879 @item ^-c5^/COMMENTS_LAYOUT=KEEP_SPECIAL^
12880 Keep unchanged special form comments
12882 Reformat comment blocks
12884 @cindex @option{^-l@var{n}^/CONSTRUCT_LAYOUT^} (@command{gnatpp})
12885 @item ^-l1^/CONSTRUCT_LAYOUT=GNAT^
12886 GNAT-style layout (this is the default)
12888 @item ^-l2^/CONSTRUCT_LAYOUT=COMPACT^
12891 @item ^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^
12894 @cindex @option{^-N^/NOTABS^} (@command{gnatpp})
12896 All the VT characters are removed from the comment text. All the HT characters
12897 are expanded with the sequences of space characters to get to the next tab
12900 @cindex @option{^--no-separate-is^/NO_SEPARATE_IS^} (@command{gnatpp})
12901 @item ^--no-separate-is^/NO_SEPARATE_IS^
12902 Do not place the keyword @code{is} on a separate line in a subprogram body in
12903 case if the spec occupies more then one line.
12905 @cindex @option{^--separate-label^/SEPARATE_LABEL^} (@command{gnatpp})
12906 @item ^--separate-label^/SEPARATE_LABEL^
12907 Place statement label(s) on a separate line, with the following statement
12910 @cindex @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} (@command{gnatpp})
12911 @item ^--separate-loop-then^/SEPARATE_LOOP_THEN^
12912 Place the keyword @code{loop} in FOR and WHILE loop statements and the
12913 keyword @code{then} in IF statements on a separate line.
12915 @cindex @option{^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^} (@command{gnatpp})
12916 @item ^--no-separate-loop-then^/NO_SEPARATE_LOOP_THEN^
12917 Do not place the keyword @code{loop} in FOR and WHILE loop statements and the
12918 keyword @code{then} in IF statements on a separate line. This option is
12919 incompatible with @option{^--separate-loop-then^/SEPARATE_LOOP_THEN^} option.
12921 @cindex @option{^--use-on-new-line^/USE_ON_NEW_LINE^} (@command{gnatpp})
12922 @item ^--use-on-new-line^/USE_ON_NEW_LINE^
12923 Start each USE clause in a context clause from a separate line.
12925 @cindex @option{^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^} (@command{gnatpp})
12926 @item ^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^
12927 Use a separate line for a loop or block statement name, but do not use an extra
12928 indentation level for the statement itself.
12934 The @option{-c1} and @option{-c2} switches are incompatible.
12935 The @option{-c3} and @option{-c4} switches are compatible with each other and
12936 also with @option{-c1} and @option{-c2}. The @option{-c0} switch disables all
12937 the other comment formatting switches.
12939 The @option{-l1}, @option{-l2}, and @option{-l3} switches are incompatible.
12944 For the @option{/COMMENTS_LAYOUT} qualifier:
12947 The @option{DEFAULT} and @option{STANDARD_INDENT} options are incompatible.
12949 The @option{GNAT_BEGINNING} and @option{REFORMAT} options are compatible with
12950 each other and also with @option{DEFAULT} and @option{STANDARD_INDENT}.
12954 The @option{GNAT}, @option{COMPACT}, and @option{UNCOMPACT} options for the
12955 @option{/CONSTRUCT_LAYOUT} qualifier are incompatible.
12958 @node General Text Layout Control
12959 @subsection General Text Layout Control
12962 These switches allow control over line length and indentation.
12965 @item ^-M@var{nnn}^/LINE_LENGTH_MAX=@var{nnn}^
12966 @cindex @option{^-M^/LINE_LENGTH^} (@command{gnatpp})
12967 Maximum line length, @var{nnn} from 32@dots{}256, the default value is 79
12969 @item ^-i@var{nnn}^/INDENTATION_LEVEL=@var{nnn}^
12970 @cindex @option{^-i^/INDENTATION_LEVEL^} (@command{gnatpp})
12971 Indentation level, @var{nnn} from 1@dots{}9, the default value is 3
12973 @item ^-cl@var{nnn}^/CONTINUATION_INDENT=@var{nnn}^
12974 @cindex @option{^-cl^/CONTINUATION_INDENT^} (@command{gnatpp})
12975 Indentation level for continuation lines (relative to the line being
12976 continued), @var{nnn} from 1@dots{}9.
12978 value is one less then the (normal) indentation level, unless the
12979 indentation is set to 1 (in which case the default value for continuation
12980 line indentation is also 1)
12983 @node Other Formatting Options
12984 @subsection Other Formatting Options
12987 These switches control the inclusion of missing end/exit labels, and
12988 the indentation level in @b{case} statements.
12991 @item ^-e^/NO_MISSED_LABELS^
12992 @cindex @option{^-e^/NO_MISSED_LABELS^} (@command{gnatpp})
12993 Do not insert missing end/exit labels. An end label is the name of
12994 a construct that may optionally be repeated at the end of the
12995 construct's declaration;
12996 e.g., the names of packages, subprograms, and tasks.
12997 An exit label is the name of a loop that may appear as target
12998 of an exit statement within the loop.
12999 By default, @command{gnatpp} inserts these end/exit labels when
13000 they are absent from the original source. This option suppresses such
13001 insertion, so that the formatted source reflects the original.
13003 @item ^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^
13004 @cindex @option{^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^} (@command{gnatpp})
13005 Insert a Form Feed character after a pragma Page.
13007 @item ^-T@var{nnn}^/MAX_INDENT=@var{nnn}^
13008 @cindex @option{^-T^/MAX_INDENT^} (@command{gnatpp})
13009 Do not use an additional indentation level for @b{case} alternatives
13010 and variants if there are @var{nnn} or more (the default
13012 If @var{nnn} is 0, an additional indentation level is
13013 used for @b{case} alternatives and variants regardless of their number.
13016 @node Setting the Source Search Path
13017 @subsection Setting the Source Search Path
13020 To define the search path for the input source file, @command{gnatpp}
13021 uses the same switches as the GNAT compiler, with the same effects.
13024 @item ^-I^/SEARCH=^@var{dir}
13025 @cindex @option{^-I^/SEARCH^} (@code{gnatpp})
13026 The same as the corresponding gcc switch
13028 @item ^-I-^/NOCURRENT_DIRECTORY^
13029 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatpp})
13030 The same as the corresponding gcc switch
13032 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE^=@var{path}
13033 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@code{gnatpp})
13034 The same as the corresponding gcc switch
13036 @item ^--RTS^/RUNTIME_SYSTEM^=@var{path}
13037 @cindex @option{^--RTS^/RUNTIME_SYSTEM^} (@code{gnatpp})
13038 The same as the corresponding gcc switch
13042 @node Output File Control
13043 @subsection Output File Control
13046 By default the output is sent to the file whose name is obtained by appending
13047 the ^@file{.pp}^@file{$PP}^ suffix to the name of the input file
13048 (if the file with this name already exists, it is unconditionally overwritten).
13049 Thus if the input file is @file{^my_ada_proc.adb^MY_ADA_PROC.ADB^} then
13050 @command{gnatpp} will produce @file{^my_ada_proc.adb.pp^MY_ADA_PROC.ADB$PP^}
13052 The output may be redirected by the following switches:
13055 @item ^-pipe^/STANDARD_OUTPUT^
13056 @cindex @option{^-pipe^/STANDARD_OUTPUT^} (@code{gnatpp})
13057 Send the output to @code{Standard_Output}
13059 @item ^-o @var{output_file}^/OUTPUT=@var{output_file}^
13060 @cindex @option{^-o^/OUTPUT^} (@code{gnatpp})
13061 Write the output into @var{output_file}.
13062 If @var{output_file} already exists, @command{gnatpp} terminates without
13063 reading or processing the input file.
13065 @item ^-of ^/FORCED_OUTPUT=^@var{output_file}
13066 @cindex @option{^-of^/FORCED_OUTPUT^} (@code{gnatpp})
13067 Write the output into @var{output_file}, overwriting the existing file
13068 (if one is present).
13070 @item ^-r^/REPLACE^
13071 @cindex @option{^-r^/REPLACE^} (@code{gnatpp})
13072 Replace the input source file with the reformatted output, and copy the
13073 original input source into the file whose name is obtained by appending the
13074 ^@file{.npp}^@file{$NPP}^ suffix to the name of the input file.
13075 If a file with this name already exists, @command{gnatpp} terminates without
13076 reading or processing the input file.
13078 @item ^-rf^/OVERRIDING_REPLACE^
13079 @cindex @option{^-rf^/OVERRIDING_REPLACE^} (@code{gnatpp})
13080 Like @option{^-r^/REPLACE^} except that if the file with the specified name
13081 already exists, it is overwritten.
13083 @item ^-rnb^/REPLACE_NO_BACKUP^
13084 @cindex @option{^-rnb^/REPLACE_NO_BACKUP^} (@code{gnatpp})
13085 Replace the input source file with the reformatted output without
13086 creating any backup copy of the input source.
13088 @item ^--eol=@var{xxx}^/END_OF_LINE=@var{xxx}^
13089 @cindex @option{^--eol^/END_OF_LINE^} (@code{gnatpp})
13090 Specifies the format of the reformatted output file. The @var{xxx}
13091 ^string specified with the switch^option^ may be either
13093 @item ``@option{^dos^DOS^}'' MS DOS style, lines end with CR LF characters
13094 @item ``@option{^crlf^CRLF^}''
13095 the same as @option{^crlf^CRLF^}
13096 @item ``@option{^unix^UNIX^}'' UNIX style, lines end with LF character
13097 @item ``@option{^lf^LF^}''
13098 the same as @option{^unix^UNIX^}
13101 @item ^-W^/RESULT_ENCODING=^@var{e}
13102 @cindex @option{^-W^/RESULT_ENCODING=^} (@command{gnatpp})
13103 Specify the wide character encoding method used to write the code in the
13105 @var{e} is one of the following:
13113 Upper half encoding
13115 @item ^s^SHIFT_JIS^
13125 Brackets encoding (default value)
13131 Options @option{^-pipe^/STANDARD_OUTPUT^},
13132 @option{^-o^/OUTPUT^} and
13133 @option{^-of^/FORCED_OUTPUT^} are allowed only if the call to gnatpp
13134 contains only one file to reformat.
13136 @option{^--eol^/END_OF_LINE^}
13138 @option{^-W^/RESULT_ENCODING^}
13139 cannot be used together
13140 with @option{^-pipe^/STANDARD_OUTPUT^} option.
13142 @node Other gnatpp Switches
13143 @subsection Other @code{gnatpp} Switches
13146 The additional @command{gnatpp} switches are defined in this subsection.
13149 @item ^-files @var{filename}^/FILES=@var{filename}^
13150 @cindex @option{^-files^/FILES^} (@code{gnatpp})
13151 Take the argument source files from the specified file. This file should be an
13152 ordinary text file containing file names separated by spaces or
13153 line breaks. You can use this switch more than once in the same call to
13154 @command{gnatpp}. You also can combine this switch with an explicit list of
13157 @item ^-v^/VERBOSE^
13158 @cindex @option{^-v^/VERBOSE^} (@code{gnatpp})
13160 @command{gnatpp} generates version information and then
13161 a trace of the actions it takes to produce or obtain the ASIS tree.
13163 @item ^-w^/WARNINGS^
13164 @cindex @option{^-w^/WARNINGS^} (@code{gnatpp})
13166 @command{gnatpp} generates a warning whenever it cannot provide
13167 a required layout in the result source.
13170 @node Formatting Rules
13171 @section Formatting Rules
13174 The following subsections show how @command{gnatpp} treats ``white space'',
13175 comments, program layout, and name casing.
13176 They provide the detailed descriptions of the switches shown above.
13179 * White Space and Empty Lines::
13180 * Formatting Comments::
13181 * Construct Layout::
13185 @node White Space and Empty Lines
13186 @subsection White Space and Empty Lines
13189 @command{gnatpp} does not have an option to control space characters.
13190 It will add or remove spaces according to the style illustrated by the
13191 examples in the @cite{Ada Reference Manual}.
13193 The only format effectors
13194 (see @cite{Ada Reference Manual}, paragraph 2.1(13))
13195 that will appear in the output file are platform-specific line breaks,
13196 and also format effectors within (but not at the end of) comments.
13197 In particular, each horizontal tab character that is not inside
13198 a comment will be treated as a space and thus will appear in the
13199 output file as zero or more spaces depending on
13200 the reformatting of the line in which it appears.
13201 The only exception is a Form Feed character, which is inserted after a
13202 pragma @code{Page} when @option{-ff} is set.
13204 The output file will contain no lines with trailing ``white space'' (spaces,
13207 Empty lines in the original source are preserved
13208 only if they separate declarations or statements.
13209 In such contexts, a
13210 sequence of two or more empty lines is replaced by exactly one empty line.
13211 Note that a blank line will be removed if it separates two ``comment blocks''
13212 (a comment block is a sequence of whole-line comments).
13213 In order to preserve a visual separation between comment blocks, use an
13214 ``empty comment'' (a line comprising only hyphens) rather than an empty line.
13215 Likewise, if for some reason you wish to have a sequence of empty lines,
13216 use a sequence of empty comments instead.
13218 @node Formatting Comments
13219 @subsection Formatting Comments
13222 Comments in Ada code are of two kinds:
13225 a @emph{whole-line comment}, which appears by itself (possibly preceded by
13226 ``white space'') on a line
13229 an @emph{end-of-line comment}, which follows some other Ada lexical element
13234 The indentation of a whole-line comment is that of either
13235 the preceding or following line in
13236 the formatted source, depending on switch settings as will be described below.
13238 For an end-of-line comment, @command{gnatpp} leaves the same number of spaces
13239 between the end of the preceding Ada lexical element and the beginning
13240 of the comment as appear in the original source,
13241 unless either the comment has to be split to
13242 satisfy the line length limitation, or else the next line contains a
13243 whole line comment that is considered a continuation of this end-of-line
13244 comment (because it starts at the same position).
13246 cases, the start of the end-of-line comment is moved right to the nearest
13247 multiple of the indentation level.
13248 This may result in a ``line overflow'' (the right-shifted comment extending
13249 beyond the maximum line length), in which case the comment is split as
13252 There is a difference between @option{^-c1^/COMMENTS_LAYOUT=DEFAULT^}
13253 (GNAT-style comment line indentation)
13254 and @option{^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^}
13255 (reference-manual comment line indentation).
13256 With reference-manual style, a whole-line comment is indented as if it
13257 were a declaration or statement at the same place
13258 (i.e., according to the indentation of the preceding line(s)).
13259 With GNAT style, a whole-line comment that is immediately followed by an
13260 @b{if} or @b{case} statement alternative, a record variant, or the reserved
13261 word @b{begin}, is indented based on the construct that follows it.
13264 @smallexample @c ada
13276 Reference-manual indentation produces:
13278 @smallexample @c ada
13290 while GNAT-style indentation produces:
13292 @smallexample @c ada
13304 The @option{^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^} switch
13305 (GNAT style comment beginning) has the following
13310 For each whole-line comment that does not end with two hyphens,
13311 @command{gnatpp} inserts spaces if necessary after the starting two hyphens
13312 to ensure that there are at least two spaces between these hyphens and the
13313 first non-blank character of the comment.
13317 For an end-of-line comment, if in the original source the next line is a
13318 whole-line comment that starts at the same position
13319 as the end-of-line comment,
13320 then the whole-line comment (and all whole-line comments
13321 that follow it and that start at the same position)
13322 will start at this position in the output file.
13325 That is, if in the original source we have:
13327 @smallexample @c ada
13330 A := B + C; -- B must be in the range Low1..High1
13331 -- C must be in the range Low2..High2
13332 --B+C will be in the range Low1+Low2..High1+High2
13338 Then in the formatted source we get
13340 @smallexample @c ada
13343 A := B + C; -- B must be in the range Low1..High1
13344 -- C must be in the range Low2..High2
13345 -- B+C will be in the range Low1+Low2..High1+High2
13351 A comment that exceeds the line length limit will be split.
13353 @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} (reformat comment blocks) is set and
13354 the line belongs to a reformattable block, splitting the line generates a
13355 @command{gnatpp} warning.
13356 The @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} switch specifies that whole-line
13357 comments may be reformatted in typical
13358 word processor style (that is, moving words between lines and putting as
13359 many words in a line as possible).
13362 The @option{^-c5^/COMMENTS_LAYOUT=KEEP_SPECIAL^} switch specifies, that comments
13363 that has a special format (that is, a character that is neither a letter nor digit
13364 not white space nor line break immediately following the leading @code{--} of
13365 the comment) should be without any change moved from the argument source
13366 into reformatted source. This switch allows to preserve comments that are used
13367 as a special marks in the code (e.g.@: SPARK annotation).
13369 @node Construct Layout
13370 @subsection Construct Layout
13373 In several cases the suggested layout in the Ada Reference Manual includes
13374 an extra level of indentation that many programmers prefer to avoid. The
13375 affected cases include:
13379 @item Record type declaration (RM 3.8)
13381 @item Record representation clause (RM 13.5.1)
13383 @item Loop statement in case if a loop has a statement identifier (RM 5.6)
13385 @item Block statement in case if a block has a statement identifier (RM 5.6)
13389 In compact mode (when GNAT style layout or compact layout is set),
13390 the pretty printer uses one level of indentation instead
13391 of two. This is achieved in the record definition and record representation
13392 clause cases by putting the @code{record} keyword on the same line as the
13393 start of the declaration or representation clause, and in the block and loop
13394 case by putting the block or loop header on the same line as the statement
13398 The difference between GNAT style @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^}
13399 and compact @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^}
13400 layout on the one hand, and uncompact layout
13401 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} on the other hand,
13402 can be illustrated by the following examples:
13406 @multitable @columnfractions .5 .5
13407 @item @i{GNAT style, compact layout} @tab @i{Uncompact layout}
13410 @smallexample @c ada
13417 @smallexample @c ada
13426 @smallexample @c ada
13428 a at 0 range 0 .. 31;
13429 b at 4 range 0 .. 31;
13433 @smallexample @c ada
13436 a at 0 range 0 .. 31;
13437 b at 4 range 0 .. 31;
13442 @smallexample @c ada
13450 @smallexample @c ada
13460 @smallexample @c ada
13461 Clear : for J in 1 .. 10 loop
13466 @smallexample @c ada
13468 for J in 1 .. 10 loop
13479 GNAT style, compact layout Uncompact layout
13481 type q is record type q is
13482 a : integer; record
13483 b : integer; a : integer;
13484 end record; b : integer;
13487 for q use record for q use
13488 a at 0 range 0 .. 31; record
13489 b at 4 range 0 .. 31; a at 0 range 0 .. 31;
13490 end record; b at 4 range 0 .. 31;
13493 Block : declare Block :
13494 A : Integer := 3; declare
13495 begin A : Integer := 3;
13497 end Block; Proc (A, A);
13500 Clear : for J in 1 .. 10 loop Clear :
13501 A (J) := 0; for J in 1 .. 10 loop
13502 end loop Clear; A (J) := 0;
13509 A further difference between GNAT style layout and compact layout is that
13510 GNAT style layout inserts empty lines as separation for
13511 compound statements, return statements and bodies.
13513 Note that the layout specified by
13514 @option{^--separate-stmt-name^/STMT_NAME_ON_NEW_LINE^}
13515 for named block and loop statements overrides the layout defined by these
13516 constructs by @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^},
13517 @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^} or
13518 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} option.
13521 @subsection Name Casing
13524 @command{gnatpp} always converts the usage occurrence of a (simple) name to
13525 the same casing as the corresponding defining identifier.
13527 You control the casing for defining occurrences via the
13528 @option{^-n^/NAME_CASING^} switch.
13530 With @option{-nD} (``as declared'', which is the default),
13533 With @option{/NAME_CASING=AS_DECLARED}, which is the default,
13535 defining occurrences appear exactly as in the source file
13536 where they are declared.
13537 The other ^values for this switch^options for this qualifier^ ---
13538 @option{^-nU^UPPER_CASE^},
13539 @option{^-nL^LOWER_CASE^},
13540 @option{^-nM^MIXED_CASE^} ---
13542 ^upper, lower, or mixed case, respectively^the corresponding casing^.
13543 If @command{gnatpp} changes the casing of a defining
13544 occurrence, it analogously changes the casing of all the
13545 usage occurrences of this name.
13547 If the defining occurrence of a name is not in the source compilation unit
13548 currently being processed by @command{gnatpp}, the casing of each reference to
13549 this name is changed according to the value of the @option{^-n^/NAME_CASING^}
13550 switch (subject to the dictionary file mechanism described below).
13551 Thus @command{gnatpp} acts as though the @option{^-n^/NAME_CASING^} switch
13553 casing for the defining occurrence of the name.
13555 Some names may need to be spelled with casing conventions that are not
13556 covered by the upper-, lower-, and mixed-case transformations.
13557 You can arrange correct casing by placing such names in a
13558 @emph{dictionary file},
13559 and then supplying a @option{^-D^/DICTIONARY^} switch.
13560 The casing of names from dictionary files overrides
13561 any @option{^-n^/NAME_CASING^} switch.
13563 To handle the casing of Ada predefined names and the names from GNAT libraries,
13564 @command{gnatpp} assumes a default dictionary file.
13565 The name of each predefined entity is spelled with the same casing as is used
13566 for the entity in the @cite{Ada Reference Manual}.
13567 The name of each entity in the GNAT libraries is spelled with the same casing
13568 as is used in the declaration of that entity.
13570 The @w{@option{^-D-^/SPECIFIC_CASING^}} switch suppresses the use of the
13571 default dictionary file.
13572 Instead, the casing for predefined and GNAT-defined names will be established
13573 by the @option{^-n^/NAME_CASING^} switch or explicit dictionary files.
13574 For example, by default the names @code{Ada.Text_IO} and @code{GNAT.OS_Lib}
13575 will appear as just shown,
13576 even in the presence of a @option{^-nU^/NAME_CASING=UPPER_CASE^} switch.
13577 To ensure that even such names are rendered in uppercase,
13578 additionally supply the @w{@option{^-D-^/SPECIFIC_CASING^}} switch
13579 (or else, less conveniently, place these names in upper case in a dictionary
13582 A dictionary file is
13583 a plain text file; each line in this file can be either a blank line
13584 (containing only space characters and ASCII.HT characters), an Ada comment
13585 line, or the specification of exactly one @emph{casing schema}.
13587 A casing schema is a string that has the following syntax:
13591 @var{casing_schema} ::= @var{identifier} | *@var{simple_identifier}*
13593 @var{simple_identifier} ::= @var{letter}@{@var{letter_or_digit}@}
13598 (See @cite{Ada Reference Manual}, Section 2.3) for the definition of the
13599 @var{identifier} lexical element and the @var{letter_or_digit} category.)
13601 The casing schema string can be followed by white space and/or an Ada-style
13602 comment; any amount of white space is allowed before the string.
13604 If a dictionary file is passed as
13606 the value of a @option{-D@var{file}} switch
13609 an option to the @option{/DICTIONARY} qualifier
13612 simple name and every identifier, @command{gnatpp} checks if the dictionary
13613 defines the casing for the name or for some of its parts (the term ``subword''
13614 is used below to denote the part of a name which is delimited by ``_'' or by
13615 the beginning or end of the word and which does not contain any ``_'' inside):
13619 if the whole name is in the dictionary, @command{gnatpp} uses for this name
13620 the casing defined by the dictionary; no subwords are checked for this word
13623 for every subword @command{gnatpp} checks if the dictionary contains the
13624 corresponding string of the form @code{*@var{simple_identifier}*},
13625 and if it does, the casing of this @var{simple_identifier} is used
13629 if the whole name does not contain any ``_'' inside, and if for this name
13630 the dictionary contains two entries - one of the form @var{identifier},
13631 and another - of the form *@var{simple_identifier}*, then the first one
13632 is applied to define the casing of this name
13635 if more than one dictionary file is passed as @command{gnatpp} switches, each
13636 dictionary adds new casing exceptions and overrides all the existing casing
13637 exceptions set by the previous dictionaries
13640 when @command{gnatpp} checks if the word or subword is in the dictionary,
13641 this check is not case sensitive
13645 For example, suppose we have the following source to reformat:
13647 @smallexample @c ada
13650 name1 : integer := 1;
13651 name4_name3_name2 : integer := 2;
13652 name2_name3_name4 : Boolean;
13655 name2_name3_name4 := name4_name3_name2 > name1;
13661 And suppose we have two dictionaries:
13678 If @command{gnatpp} is called with the following switches:
13682 @command{gnatpp -nM -D dict1 -D dict2 test.adb}
13685 @command{gnatpp test.adb /NAME_CASING=MIXED_CASE /DICTIONARY=(dict1, dict2)}
13690 then we will get the following name casing in the @command{gnatpp} output:
13692 @smallexample @c ada
13695 NAME1 : Integer := 1;
13696 Name4_NAME3_Name2 : Integer := 2;
13697 Name2_NAME3_Name4 : Boolean;
13700 Name2_NAME3_Name4 := Name4_NAME3_Name2 > NAME1;
13705 @c *********************************
13706 @node The GNAT Metric Tool gnatmetric
13707 @chapter The GNAT Metric Tool @command{gnatmetric}
13709 @cindex Metric tool
13712 ^The @command{gnatmetric} tool^@command{GNAT METRIC}^ is an ASIS-based utility
13713 for computing various program metrics.
13714 It takes an Ada source file as input and generates a file containing the
13715 metrics data as output. Various switches control which
13716 metrics are computed and output.
13718 @command{gnatmetric} generates and uses the ASIS
13719 tree for the input source and thus requires the input to be syntactically and
13720 semantically legal.
13721 If this condition is not met, @command{gnatmetric} will generate
13722 an error message; no metric information for this file will be
13723 computed and reported.
13725 If the compilation unit contained in the input source depends semantically
13726 upon units in files located outside the current directory, you have to provide
13727 the source search path when invoking @command{gnatmetric}.
13728 If it depends semantically upon units that are contained
13729 in files with names that do not follow the GNAT file naming rules, you have to
13730 provide the configuration file describing the corresponding naming scheme (see
13731 the description of the @command{gnatmetric} switches below.)
13732 Alternatively, you may use a project file and invoke @command{gnatmetric}
13733 through the @command{gnat} driver.
13735 The @command{gnatmetric} command has the form
13738 @c $ gnatmetric @ovar{switches} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
13739 @c Expanding @ovar macro inline (explanation in macro def comments)
13740 $ gnatmetric @r{[}@var{switches}@r{]} @{@var{filename}@} @r{[}-cargs @var{gcc_switches}@r{]}
13747 @var{switches} specify the metrics to compute and define the destination for
13751 Each @var{filename} is the name (including the extension) of a source
13752 file to process. ``Wildcards'' are allowed, and
13753 the file name may contain path information.
13754 If no @var{filename} is supplied, then the @var{switches} list must contain
13756 @option{-files} switch (@pxref{Other gnatmetric Switches}).
13757 Including both a @option{-files} switch and one or more
13758 @var{filename} arguments is permitted.
13761 @samp{@var{gcc_switches}} is a list of switches for
13762 @command{gcc}. They will be passed on to all compiler invocations made by
13763 @command{gnatmetric} to generate the ASIS trees. Here you can provide
13764 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
13765 and use the @option{-gnatec} switch to set the configuration file.
13769 * Switches for gnatmetric::
13772 @node Switches for gnatmetric
13773 @section Switches for @command{gnatmetric}
13776 The following subsections describe the various switches accepted by
13777 @command{gnatmetric}, organized by category.
13780 * Output Files Control::
13781 * Disable Metrics For Local Units::
13782 * Specifying a set of metrics to compute::
13783 * Other gnatmetric Switches::
13784 * Generate project-wide metrics::
13787 @node Output Files Control
13788 @subsection Output File Control
13789 @cindex Output file control in @command{gnatmetric}
13792 @command{gnatmetric} has two output formats. It can generate a
13793 textual (human-readable) form, and also XML. By default only textual
13794 output is generated.
13796 When generating the output in textual form, @command{gnatmetric} creates
13797 for each Ada source file a corresponding text file
13798 containing the computed metrics, except for the case when the set of metrics
13799 specified by gnatmetric parameters consists only of metrics that are computed
13800 for the whole set of analyzed sources, but not for each Ada source.
13801 By default, this file is placed in the same directory as where the source
13802 file is located, and its name is obtained
13803 by appending the ^@file{.metrix}^@file{$METRIX}^ suffix to the name of the
13806 All the output information generated in XML format is placed in a single
13807 file. By default this file is placed in the current directory and has the
13808 name ^@file{metrix.xml}^@file{METRIX$XML}^.
13810 Some of the computed metrics are summed over the units passed to
13811 @command{gnatmetric}; for example, the total number of lines of code.
13812 By default this information is sent to @file{stdout}, but a file
13813 can be specified with the @option{-og} switch.
13815 The following switches control the @command{gnatmetric} output:
13818 @cindex @option{^-x^/XML^} (@command{gnatmetric})
13820 Generate the XML output
13822 @cindex @option{^-xs^/XSD^} (@command{gnatmetric})
13824 Generate the XML output and the XML schema file that describes the structure
13825 of the XML metric report, this schema is assigned to the XML file. The schema
13826 file has the same name as the XML output file with @file{.xml} suffix replaced
13829 @cindex @option{^-nt^/NO_TEXT^} (@command{gnatmetric})
13830 @item ^-nt^/NO_TEXT^
13831 Do not generate the output in text form (implies @option{^-x^/XML^})
13833 @cindex @option{^-d^/DIRECTORY^} (@command{gnatmetric})
13834 @item ^-d @var{output_dir}^/DIRECTORY=@var{output_dir}^
13835 Put text files with detailed metrics into @var{output_dir}
13837 @cindex @option{^-o^/SUFFIX_DETAILS^} (@command{gnatmetric})
13838 @item ^-o @var{file_suffix}^/SUFFIX_DETAILS=@var{file_suffix}^
13839 Use @var{file_suffix}, instead of ^@file{.metrix}^@file{$METRIX}^
13840 in the name of the output file.
13842 @cindex @option{^-og^/GLOBAL_OUTPUT^} (@command{gnatmetric})
13843 @item ^-og @var{file_name}^/GLOBAL_OUTPUT=@var{file_name}^
13844 Put global metrics into @var{file_name}
13846 @cindex @option{^-ox^/XML_OUTPUT^} (@command{gnatmetric})
13847 @item ^-ox @var{file_name}^/XML_OUTPUT=@var{file_name}^
13848 Put the XML output into @var{file_name} (also implies @option{^-x^/XML^})
13850 @cindex @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} (@command{gnatmetric})
13851 @item ^-sfn^/SHORT_SOURCE_FILE_NAME^
13852 Use ``short'' source file names in the output. (The @command{gnatmetric}
13853 output includes the name(s) of the Ada source file(s) from which the metrics
13854 are computed. By default each name includes the absolute path. The
13855 @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} switch causes @command{gnatmetric}
13856 to exclude all directory information from the file names that are output.)
13860 @node Disable Metrics For Local Units
13861 @subsection Disable Metrics For Local Units
13862 @cindex Disable Metrics For Local Units in @command{gnatmetric}
13865 @command{gnatmetric} relies on the GNAT compilation model @minus{}
13867 unit per one source file. It computes line metrics for the whole source
13868 file, and it also computes syntax
13869 and complexity metrics for the file's outermost unit.
13871 By default, @command{gnatmetric} will also compute all metrics for certain
13872 kinds of locally declared program units:
13876 subprogram (and generic subprogram) bodies;
13879 package (and generic package) specs and bodies;
13882 task object and type specifications and bodies;
13885 protected object and type specifications and bodies.
13889 These kinds of entities will be referred to as
13890 @emph{eligible local program units}, or simply @emph{eligible local units},
13891 @cindex Eligible local unit (for @command{gnatmetric})
13892 in the discussion below.
13894 Note that a subprogram declaration, generic instantiation,
13895 or renaming declaration only receives metrics
13896 computation when it appear as the outermost entity
13899 Suppression of metrics computation for eligible local units can be
13900 obtained via the following switch:
13903 @cindex @option{^-n@var{x}^/SUPPRESS^} (@command{gnatmetric})
13904 @item ^-nolocal^/SUPPRESS=LOCAL_DETAILS^
13905 Do not compute detailed metrics for eligible local program units
13909 @node Specifying a set of metrics to compute
13910 @subsection Specifying a set of metrics to compute
13913 By default all the metrics are computed and reported. The switches
13914 described in this subsection allow you to control, on an individual
13915 basis, whether metrics are computed and
13916 reported. If at least one positive metric
13917 switch is specified (that is, a switch that defines that a given
13918 metric or set of metrics is to be computed), then only
13919 explicitly specified metrics are reported.
13922 * Line Metrics Control::
13923 * Syntax Metrics Control::
13924 * Complexity Metrics Control::
13925 * Object-Oriented Metrics Control::
13928 @node Line Metrics Control
13929 @subsubsection Line Metrics Control
13930 @cindex Line metrics control in @command{gnatmetric}
13933 For any (legal) source file, and for each of its
13934 eligible local program units, @command{gnatmetric} computes the following
13939 the total number of lines;
13942 the total number of code lines (i.e., non-blank lines that are not comments)
13945 the number of comment lines
13948 the number of code lines containing end-of-line comments;
13951 the comment percentage: the ratio between the number of lines that contain
13952 comments and the number of all non-blank lines, expressed as a percentage;
13955 the number of empty lines and lines containing only space characters and/or
13956 format effectors (blank lines)
13959 the average number of code lines in subprogram bodies, task bodies, entry
13960 bodies and statement sequences in package bodies (this metric is only computed
13961 across the whole set of the analyzed units)
13966 @command{gnatmetric} sums the values of the line metrics for all the
13967 files being processed and then generates the cumulative results. The tool
13968 also computes for all the files being processed the average number of code
13971 You can use the following switches to select the specific line metrics
13972 to be computed and reported.
13975 @cindex @option{^--lines@var{x}^/LINE_COUNT_METRICS^} (@command{gnatmetric})
13978 @cindex @option{--no-lines@var{x}}
13981 @item ^--lines-all^/LINE_COUNT_METRICS=ALL^
13982 Report all the line metrics
13984 @item ^--no-lines-all^/LINE_COUNT_METRICS=NONE^
13985 Do not report any of line metrics
13987 @item ^--lines^/LINE_COUNT_METRICS=ALL_LINES^
13988 Report the number of all lines
13990 @item ^--no-lines^/LINE_COUNT_METRICS=NOALL_LINES^
13991 Do not report the number of all lines
13993 @item ^--lines-code^/LINE_COUNT_METRICS=CODE_LINES^
13994 Report the number of code lines
13996 @item ^--no-lines-code^/LINE_COUNT_METRICS=NOCODE_LINES^
13997 Do not report the number of code lines
13999 @item ^--lines-comment^/LINE_COUNT_METRICS=COMMENT_LINES^
14000 Report the number of comment lines
14002 @item ^--no-lines-comment^/LINE_COUNT_METRICS=NOCOMMENT_LINES^
14003 Do not report the number of comment lines
14005 @item ^--lines-eol-comment^/LINE_COUNT_METRICS=CODE_COMMENT_LINES^
14006 Report the number of code lines containing
14007 end-of-line comments
14009 @item ^--no-lines-eol-comment^/LINE_COUNT_METRICS=NOCODE_COMMENT_LINES^
14010 Do not report the number of code lines containing
14011 end-of-line comments
14013 @item ^--lines-ratio^/LINE_COUNT_METRICS=COMMENT_PERCENTAGE^
14014 Report the comment percentage in the program text
14016 @item ^--no-lines-ratio^/LINE_COUNT_METRICS=NOCOMMENT_PERCENTAGE^
14017 Do not report the comment percentage in the program text
14019 @item ^--lines-blank^/LINE_COUNT_METRICS=BLANK_LINES^
14020 Report the number of blank lines
14022 @item ^--no-lines-blank^/LINE_COUNT_METRICS=NOBLANK_LINES^
14023 Do not report the number of blank lines
14025 @item ^--lines-average^/LINE_COUNT_METRICS=AVERAGE_BODY_LINES^
14026 Report the average number of code lines in subprogram bodies, task bodies,
14027 entry bodies and statement sequences in package bodies. The metric is computed
14028 and reported for the whole set of processed Ada sources only.
14030 @item ^--no-lines-average^/LINE_COUNT_METRICS=NOAVERAGE_BODY_LINES^
14031 Do not report the average number of code lines in subprogram bodies,
14032 task bodies, entry bodies and statement sequences in package bodies.
14036 @node Syntax Metrics Control
14037 @subsubsection Syntax Metrics Control
14038 @cindex Syntax metrics control in @command{gnatmetric}
14041 @command{gnatmetric} computes various syntactic metrics for the
14042 outermost unit and for each eligible local unit:
14045 @item LSLOC (``Logical Source Lines Of Code'')
14046 The total number of declarations and the total number of statements
14048 @item Maximal static nesting level of inner program units
14050 @cite{Ada Reference Manual}, 10.1(1), ``A program unit is either a
14051 package, a task unit, a protected unit, a
14052 protected entry, a generic unit, or an explicitly declared subprogram other
14053 than an enumeration literal.''
14055 @item Maximal nesting level of composite syntactic constructs
14056 This corresponds to the notion of the
14057 maximum nesting level in the GNAT built-in style checks
14058 (@pxref{Style Checking})
14062 For the outermost unit in the file, @command{gnatmetric} additionally computes
14063 the following metrics:
14066 @item Public subprograms
14067 This metric is computed for package specs. It is the
14068 number of subprograms and generic subprograms declared in the visible
14069 part (including the visible part of nested packages, protected objects, and
14072 @item All subprograms
14073 This metric is computed for bodies and subunits. The
14074 metric is equal to a total number of subprogram bodies in the compilation
14076 Neither generic instantiations nor renamings-as-a-body nor body stubs
14077 are counted. Any subprogram body is counted, independently of its nesting
14078 level and enclosing constructs. Generic bodies and bodies of protected
14079 subprograms are counted in the same way as ``usual'' subprogram bodies.
14082 This metric is computed for package specs and
14083 generic package declarations. It is the total number of types
14084 that can be referenced from outside this compilation unit, plus the
14085 number of types from all the visible parts of all the visible generic
14086 packages. Generic formal types are not counted. Only types, not subtypes,
14090 Along with the total number of public types, the following
14091 types are counted and reported separately:
14098 Root tagged types (abstract, non-abstract, private, non-private). Type
14099 extensions are @emph{not} counted
14102 Private types (including private extensions)
14113 This metric is computed for any compilation unit. It is equal to the total
14114 number of the declarations of different types given in the compilation unit.
14115 The private and the corresponding full type declaration are counted as one
14116 type declaration. Incomplete type declarations and generic formal types
14118 No distinction is made among different kinds of types (abstract,
14119 private etc.); the total number of types is computed and reported.
14124 By default, all the syntax metrics are computed and reported. You can use the
14125 following switches to select specific syntax metrics.
14129 @cindex @option{^--syntax@var{x}^/SYNTAX_METRICS^} (@command{gnatmetric})
14132 @cindex @option{--no-syntax@var{x}} (@command{gnatmetric})
14135 @item ^--syntax-all^/SYNTAX_METRICS=ALL^
14136 Report all the syntax metrics
14138 @item ^--no-syntax-all^/SYNTAX_METRICS=NONE^
14139 Do not report any of syntax metrics
14141 @item ^--declarations^/SYNTAX_METRICS=DECLARATIONS^
14142 Report the total number of declarations
14144 @item ^--no-declarations^/SYNTAX_METRICS=NODECLARATIONS^
14145 Do not report the total number of declarations
14147 @item ^--statements^/SYNTAX_METRICS=STATEMENTS^
14148 Report the total number of statements
14150 @item ^--no-statements^/SYNTAX_METRICS=NOSTATEMENTS^
14151 Do not report the total number of statements
14153 @item ^--public-subprograms^/SYNTAX_METRICS=PUBLIC_SUBPROGRAMS^
14154 Report the number of public subprograms in a compilation unit
14156 @item ^--no-public-subprograms^/SYNTAX_METRICS=NOPUBLIC_SUBPROGRAMS^
14157 Do not report the number of public subprograms in a compilation unit
14159 @item ^--all-subprograms^/SYNTAX_METRICS=ALL_SUBPROGRAMS^
14160 Report the number of all the subprograms in a compilation unit
14162 @item ^--no-all-subprograms^/SYNTAX_METRICS=NOALL_SUBPROGRAMS^
14163 Do not report the number of all the subprograms in a compilation unit
14165 @item ^--public-types^/SYNTAX_METRICS=PUBLIC_TYPES^
14166 Report the number of public types in a compilation unit
14168 @item ^--no-public-types^/SYNTAX_METRICS=NOPUBLIC_TYPES^
14169 Do not report the number of public types in a compilation unit
14171 @item ^--all-types^/SYNTAX_METRICS=ALL_TYPES^
14172 Report the number of all the types in a compilation unit
14174 @item ^--no-all-types^/SYNTAX_METRICS=NOALL_TYPES^
14175 Do not report the number of all the types in a compilation unit
14177 @item ^--unit-nesting^/SYNTAX_METRICS=UNIT_NESTING^
14178 Report the maximal program unit nesting level
14180 @item ^--no-unit-nesting^/SYNTAX_METRICS=UNIT_NESTING_OFF^
14181 Do not report the maximal program unit nesting level
14183 @item ^--construct-nesting^/SYNTAX_METRICS=CONSTRUCT_NESTING^
14184 Report the maximal construct nesting level
14186 @item ^--no-construct-nesting^/SYNTAX_METRICS=NOCONSTRUCT_NESTING^
14187 Do not report the maximal construct nesting level
14191 @node Complexity Metrics Control
14192 @subsubsection Complexity Metrics Control
14193 @cindex Complexity metrics control in @command{gnatmetric}
14196 For a program unit that is an executable body (a subprogram body (including
14197 generic bodies), task body, entry body or a package body containing
14198 its own statement sequence) @command{gnatmetric} computes the following
14199 complexity metrics:
14203 McCabe cyclomatic complexity;
14206 McCabe essential complexity;
14209 maximal loop nesting level
14214 The McCabe complexity metrics are defined
14215 in @url{http://www.mccabe.com/pdf/nist235r.pdf}
14217 According to McCabe, both control statements and short-circuit control forms
14218 should be taken into account when computing cyclomatic complexity. For each
14219 body, we compute three metric values:
14223 the complexity introduced by control
14224 statements only, without taking into account short-circuit forms,
14227 the complexity introduced by short-circuit control forms only, and
14231 cyclomatic complexity, which is the sum of these two values.
14235 When computing cyclomatic and essential complexity, @command{gnatmetric} skips
14236 the code in the exception handlers and in all the nested program units.
14238 By default, all the complexity metrics are computed and reported.
14239 For more fine-grained control you can use
14240 the following switches:
14243 @cindex @option{^-complexity@var{x}^/COMPLEXITY_METRICS^} (@command{gnatmetric})
14246 @cindex @option{--no-complexity@var{x}}
14249 @item ^--complexity-all^/COMPLEXITY_METRICS=ALL^
14250 Report all the complexity metrics
14252 @item ^--no-complexity-all^/COMPLEXITY_METRICS=NONE^
14253 Do not report any of complexity metrics
14255 @item ^--complexity-cyclomatic^/COMPLEXITY_METRICS=CYCLOMATIC^
14256 Report the McCabe Cyclomatic Complexity
14258 @item ^--no-complexity-cyclomatic^/COMPLEXITY_METRICS=NOCYCLOMATIC^
14259 Do not report the McCabe Cyclomatic Complexity
14261 @item ^--complexity-essential^/COMPLEXITY_METRICS=ESSENTIAL^
14262 Report the Essential Complexity
14264 @item ^--no-complexity-essential^/COMPLEXITY_METRICS=NOESSENTIAL^
14265 Do not report the Essential Complexity
14267 @item ^--loop-nesting^/COMPLEXITY_METRICS=LOOP_NESTING_ON^
14268 Report maximal loop nesting level
14270 @item ^--no-loop-nesting^/COMPLEXITY_METRICS=NOLOOP_NESTING^
14271 Do not report maximal loop nesting level
14273 @item ^--complexity-average^/COMPLEXITY_METRICS=AVERAGE_COMPLEXITY^
14274 Report the average McCabe Cyclomatic Complexity for all the subprogram bodies,
14275 task bodies, entry bodies and statement sequences in package bodies.
14276 The metric is computed and reported for whole set of processed Ada sources
14279 @item ^--no-complexity-average^/COMPLEXITY_METRICS=NOAVERAGE_COMPLEXITY^
14280 Do not report the average McCabe Cyclomatic Complexity for all the subprogram
14281 bodies, task bodies, entry bodies and statement sequences in package bodies
14283 @cindex @option{^-ne^/NO_EXITS_AS_GOTOS^} (@command{gnatmetric})
14284 @item ^-ne^/NO_EXITS_AS_GOTOS^
14285 Do not consider @code{exit} statements as @code{goto}s when
14286 computing Essential Complexity
14288 @item ^--extra-exit-points^/EXTRA_EXIT_POINTS^
14289 Report the extra exit points for subprogram bodies. As an exit point, this
14290 metric counts @code{return} statements and raise statements in case when the
14291 raised exception is not handled in the same body. In case of a function this
14292 metric subtracts 1 from the number of exit points, because a function body
14293 must contain at least one @code{return} statement.
14295 @item ^--no-extra-exit-points^/NOEXTRA_EXIT_POINTS^
14296 Do not report the extra exit points for subprogram bodies
14300 @node Object-Oriented Metrics Control
14301 @subsubsection Object-Oriented Metrics Control
14302 @cindex Object-Oriented metrics control in @command{gnatmetric}
14305 @cindex Coupling metrics (in in @command{gnatmetric})
14306 Coupling metrics are object-oriented metrics that measure the
14307 dependencies between a given class (or a group of classes) and the
14308 ``external world'' (that is, the other classes in the program). In this
14309 subsection the term ``class'' is used in its
14310 traditional object-oriented programming sense
14311 (an instantiable module that contains data and/or method members).
14312 A @emph{category} (of classes)
14313 is a group of closely related classes that are reused and/or
14316 A class @code{K}'s @emph{efferent coupling} is the number of classes
14317 that @code{K} depends upon.
14318 A category's efferent coupling is the number of classes outside the
14319 category that the classes inside the category depend upon.
14321 A class @code{K}'s @emph{afferent coupling} is the number of classes
14322 that depend upon @code{K}.
14323 A category's afferent coupling is the number of classes outside the
14324 category that depend on classes belonging to the category.
14326 Ada's implementation of the object-oriented paradigm does not use the
14327 traditional class notion, so the definition of the coupling
14328 metrics for Ada maps the class and class category notions
14329 onto Ada constructs.
14331 For the coupling metrics, several kinds of modules -- a library package,
14332 a library generic package, and a library generic package instantiation --
14333 that define a tagged type or an interface type are
14334 considered to be a class. A category consists of a library package (or
14335 a library generic package) that defines a tagged or an interface type,
14336 together with all its descendant (generic) packages that define tagged
14337 or interface types. For any package counted as a class,
14338 its body and subunits (if any) are considered
14339 together with its spec when counting the dependencies, and coupling
14340 metrics are reported for spec units only. For dependencies
14341 between classes, the Ada semantic dependencies are considered.
14342 For coupling metrics, only dependencies on units that are considered as
14343 classes, are considered.
14345 When computing coupling metrics, @command{gnatmetric} counts only
14346 dependencies between units that are arguments of the gnatmetric call.
14347 Coupling metrics are program-wide (or project-wide) metrics, so to
14348 get a valid result, you should call @command{gnatmetric} for
14349 the whole set of sources that make up your program. It can be done
14350 by calling @command{gnatmetric} from the GNAT driver with @option{-U}
14351 option (see See @ref{The GNAT Driver and Project Files} for details.
14353 By default, all the coupling metrics are disabled. You can use the following
14354 switches to specify the coupling metrics to be computed and reported:
14359 @cindex @option{--package@var{x}} (@command{gnatmetric})
14360 @cindex @option{--no-package@var{x}} (@command{gnatmetric})
14361 @cindex @option{--category@var{x}} (@command{gnatmetric})
14362 @cindex @option{--no-category@var{x}} (@command{gnatmetric})
14366 @cindex @option{/COUPLING_METRICS} (@command{gnatmetric})
14369 @item ^--coupling-all^/COUPLING_METRICS=ALL^
14370 Report all the coupling metrics
14372 @item ^--no-coupling-all^/COUPLING_METRICS=NONE^
14373 Do not report any of metrics
14375 @item ^--package-efferent-coupling^/COUPLING_METRICS=PACKAGE_EFFERENT^
14376 Report package efferent coupling
14378 @item ^--no-package-efferent-coupling^/COUPLING_METRICS=NOPACKAGE_EFFERENT^
14379 Do not report package efferent coupling
14381 @item ^--package-afferent-coupling^/COUPLING_METRICS=PACKAGE_AFFERENT^
14382 Report package afferent coupling
14384 @item ^--no-package-afferent-coupling^/COUPLING_METRICS=NOPACKAGE_AFFERENT^
14385 Do not report package afferent coupling
14387 @item ^--category-efferent-coupling^/COUPLING_METRICS=CATEGORY_EFFERENT^
14388 Report category efferent coupling
14390 @item ^--no-category-efferent-coupling^/COUPLING_METRICS=NOCATEGORY_EFFERENT^
14391 Do not report category efferent coupling
14393 @item ^--category-afferent-coupling^/COUPLING_METRICS=CATEGORY_AFFERENT^
14394 Report category afferent coupling
14396 @item ^--no-category-afferent-coupling^/COUPLING_METRICS=NOCATEGORY_AFFERENT^
14397 Do not report category afferent coupling
14401 @node Other gnatmetric Switches
14402 @subsection Other @code{gnatmetric} Switches
14405 Additional @command{gnatmetric} switches are as follows:
14408 @item ^-files @var{filename}^/FILES=@var{filename}^
14409 @cindex @option{^-files^/FILES^} (@code{gnatmetric})
14410 Take the argument source files from the specified file. This file should be an
14411 ordinary text file containing file names separated by spaces or
14412 line breaks. You can use this switch more than once in the same call to
14413 @command{gnatmetric}. You also can combine this switch with
14414 an explicit list of files.
14416 @item ^-v^/VERBOSE^
14417 @cindex @option{^-v^/VERBOSE^} (@code{gnatmetric})
14419 @command{gnatmetric} generates version information and then
14420 a trace of sources being processed.
14422 @item ^-dv^/DEBUG_OUTPUT^
14423 @cindex @option{^-dv^/DEBUG_OUTPUT^} (@code{gnatmetric})
14425 @command{gnatmetric} generates various messages useful to understand what
14426 happens during the metrics computation
14429 @cindex @option{^-q^/QUIET^} (@code{gnatmetric})
14433 @node Generate project-wide metrics
14434 @subsection Generate project-wide metrics
14436 In order to compute metrics on all units of a given project, you can use
14437 the @command{gnat} driver along with the @option{-P} option:
14443 If the project @code{proj} depends upon other projects, you can compute
14444 the metrics on the project closure using the @option{-U} option:
14446 gnat metric -Pproj -U
14450 Finally, if not all the units are relevant to a particular main
14451 program in the project closure, you can generate metrics for the set
14452 of units needed to create a given main program (unit closure) using
14453 the @option{-U} option followed by the name of the main unit:
14455 gnat metric -Pproj -U main
14459 @c ***********************************
14460 @node File Name Krunching Using gnatkr
14461 @chapter File Name Krunching Using @code{gnatkr}
14465 This chapter discusses the method used by the compiler to shorten
14466 the default file names chosen for Ada units so that they do not
14467 exceed the maximum length permitted. It also describes the
14468 @code{gnatkr} utility that can be used to determine the result of
14469 applying this shortening.
14473 * Krunching Method::
14474 * Examples of gnatkr Usage::
14478 @section About @code{gnatkr}
14481 The default file naming rule in GNAT
14482 is that the file name must be derived from
14483 the unit name. The exact default rule is as follows:
14486 Take the unit name and replace all dots by hyphens.
14488 If such a replacement occurs in the
14489 second character position of a name, and the first character is
14490 ^@samp{a}, @samp{g}, @samp{s}, or @samp{i}, ^@samp{A}, @samp{G}, @samp{S}, or @samp{I},^
14491 then replace the dot by the character
14492 ^@samp{~} (tilde)^@samp{$} (dollar sign)^
14493 instead of a minus.
14495 The reason for this exception is to avoid clashes
14496 with the standard names for children of System, Ada, Interfaces,
14497 and GNAT, which use the prefixes
14498 ^@samp{s-}, @samp{a-}, @samp{i-}, and @samp{g-},^@samp{S-}, @samp{A-}, @samp{I-}, and @samp{G-},^
14501 The @option{^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{nn}}
14502 switch of the compiler activates a ``krunching''
14503 circuit that limits file names to nn characters (where nn is a decimal
14504 integer). For example, using OpenVMS,
14505 where the maximum file name length is
14506 39, the value of nn is usually set to 39, but if you want to generate
14507 a set of files that would be usable if ported to a system with some
14508 different maximum file length, then a different value can be specified.
14509 The default value of 39 for OpenVMS need not be specified.
14511 The @code{gnatkr} utility can be used to determine the krunched name for
14512 a given file, when krunched to a specified maximum length.
14515 @section Using @code{gnatkr}
14518 The @code{gnatkr} command has the form
14522 @c $ gnatkr @var{name} @ovar{length}
14523 @c Expanding @ovar macro inline (explanation in macro def comments)
14524 $ gnatkr @var{name} @r{[}@var{length}@r{]}
14530 $ gnatkr @var{name} /COUNT=nn
14535 @var{name} is the uncrunched file name, derived from the name of the unit
14536 in the standard manner described in the previous section (i.e., in particular
14537 all dots are replaced by hyphens). The file name may or may not have an
14538 extension (defined as a suffix of the form period followed by arbitrary
14539 characters other than period). If an extension is present then it will
14540 be preserved in the output. For example, when krunching @file{hellofile.ads}
14541 to eight characters, the result will be hellofil.ads.
14543 Note: for compatibility with previous versions of @code{gnatkr} dots may
14544 appear in the name instead of hyphens, but the last dot will always be
14545 taken as the start of an extension. So if @code{gnatkr} is given an argument
14546 such as @file{Hello.World.adb} it will be treated exactly as if the first
14547 period had been a hyphen, and for example krunching to eight characters
14548 gives the result @file{hellworl.adb}.
14550 Note that the result is always all lower case (except on OpenVMS where it is
14551 all upper case). Characters of the other case are folded as required.
14553 @var{length} represents the length of the krunched name. The default
14554 when no argument is given is ^8^39^ characters. A length of zero stands for
14555 unlimited, in other words do not chop except for system files where the
14556 implied crunching length is always eight characters.
14559 The output is the krunched name. The output has an extension only if the
14560 original argument was a file name with an extension.
14562 @node Krunching Method
14563 @section Krunching Method
14566 The initial file name is determined by the name of the unit that the file
14567 contains. The name is formed by taking the full expanded name of the
14568 unit and replacing the separating dots with hyphens and
14569 using ^lowercase^uppercase^
14570 for all letters, except that a hyphen in the second character position is
14571 replaced by a ^tilde^dollar sign^ if the first character is
14572 ^@samp{a}, @samp{i}, @samp{g}, or @samp{s}^@samp{A}, @samp{I}, @samp{G}, or @samp{S}^.
14573 The extension is @code{.ads} for a
14574 spec and @code{.adb} for a body.
14575 Krunching does not affect the extension, but the file name is shortened to
14576 the specified length by following these rules:
14580 The name is divided into segments separated by hyphens, tildes or
14581 underscores and all hyphens, tildes, and underscores are
14582 eliminated. If this leaves the name short enough, we are done.
14585 If the name is too long, the longest segment is located (left-most
14586 if there are two of equal length), and shortened by dropping
14587 its last character. This is repeated until the name is short enough.
14589 As an example, consider the krunching of @*@file{our-strings-wide_fixed.adb}
14590 to fit the name into 8 characters as required by some operating systems.
14593 our-strings-wide_fixed 22
14594 our strings wide fixed 19
14595 our string wide fixed 18
14596 our strin wide fixed 17
14597 our stri wide fixed 16
14598 our stri wide fixe 15
14599 our str wide fixe 14
14600 our str wid fixe 13
14606 Final file name: oustwifi.adb
14610 The file names for all predefined units are always krunched to eight
14611 characters. The krunching of these predefined units uses the following
14612 special prefix replacements:
14616 replaced by @file{^a^A^-}
14619 replaced by @file{^g^G^-}
14622 replaced by @file{^i^I^-}
14625 replaced by @file{^s^S^-}
14628 These system files have a hyphen in the second character position. That
14629 is why normal user files replace such a character with a
14630 ^tilde^dollar sign^, to
14631 avoid confusion with system file names.
14633 As an example of this special rule, consider
14634 @*@file{ada-strings-wide_fixed.adb}, which gets krunched as follows:
14637 ada-strings-wide_fixed 22
14638 a- strings wide fixed 18
14639 a- string wide fixed 17
14640 a- strin wide fixed 16
14641 a- stri wide fixed 15
14642 a- stri wide fixe 14
14643 a- str wide fixe 13
14649 Final file name: a-stwifi.adb
14653 Of course no file shortening algorithm can guarantee uniqueness over all
14654 possible unit names, and if file name krunching is used then it is your
14655 responsibility to ensure that no name clashes occur. The utility
14656 program @code{gnatkr} is supplied for conveniently determining the
14657 krunched name of a file.
14659 @node Examples of gnatkr Usage
14660 @section Examples of @code{gnatkr} Usage
14667 $ gnatkr very_long_unit_name.ads --> velounna.ads
14668 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
14669 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
14670 $ gnatkr grandparent-parent-child --> grparchi
14672 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
14673 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
14676 @node Preprocessing Using gnatprep
14677 @chapter Preprocessing Using @code{gnatprep}
14681 This chapter discusses how to use GNAT's @code{gnatprep} utility for simple
14683 Although designed for use with GNAT, @code{gnatprep} does not depend on any
14684 special GNAT features.
14685 For further discussion of conditional compilation in general, see
14686 @ref{Conditional Compilation}.
14689 * Preprocessing Symbols::
14691 * Switches for gnatprep::
14692 * Form of Definitions File::
14693 * Form of Input Text for gnatprep::
14696 @node Preprocessing Symbols
14697 @section Preprocessing Symbols
14700 Preprocessing symbols are defined in definition files and referred to in
14701 sources to be preprocessed. A Preprocessing symbol is an identifier, following
14702 normal Ada (case-insensitive) rules for its syntax, with the restriction that
14703 all characters need to be in the ASCII set (no accented letters).
14705 @node Using gnatprep
14706 @section Using @code{gnatprep}
14709 To call @code{gnatprep} use
14712 @c $ gnatprep @ovar{switches} @var{infile} @var{outfile} @ovar{deffile}
14713 @c Expanding @ovar macro inline (explanation in macro def comments)
14714 $ gnatprep @r{[}@var{switches}@r{]} @var{infile} @var{outfile} @r{[}@var{deffile}@r{]}
14721 is an optional sequence of switches as described in the next section.
14724 is the full name of the input file, which is an Ada source
14725 file containing preprocessor directives.
14728 is the full name of the output file, which is an Ada source
14729 in standard Ada form. When used with GNAT, this file name will
14730 normally have an ads or adb suffix.
14733 is the full name of a text file containing definitions of
14734 preprocessing symbols to be referenced by the preprocessor. This argument is
14735 optional, and can be replaced by the use of the @option{-D} switch.
14739 @node Switches for gnatprep
14740 @section Switches for @code{gnatprep}
14745 @item ^-b^/BLANK_LINES^
14746 @cindex @option{^-b^/BLANK_LINES^} (@command{gnatprep})
14747 Causes both preprocessor lines and the lines deleted by
14748 preprocessing to be replaced by blank lines in the output source file,
14749 preserving line numbers in the output file.
14751 @item ^-c^/COMMENTS^
14752 @cindex @option{^-c^/COMMENTS^} (@command{gnatprep})
14753 Causes both preprocessor lines and the lines deleted
14754 by preprocessing to be retained in the output source as comments marked
14755 with the special string @code{"--! "}. This option will result in line numbers
14756 being preserved in the output file.
14758 @item ^-C^/REPLACE_IN_COMMENTS^
14759 @cindex @option{^-C^/REPLACE_IN_COMMENTS^} (@command{gnatprep})
14760 Causes comments to be scanned. Normally comments are ignored by gnatprep.
14761 If this option is specified, then comments are scanned and any $symbol
14762 substitutions performed as in program text. This is particularly useful
14763 when structured comments are used (e.g., when writing programs in the
14764 SPARK dialect of Ada). Note that this switch is not available when
14765 doing integrated preprocessing (it would be useless in this context
14766 since comments are ignored by the compiler in any case).
14768 @item ^-Dsymbol=value^/ASSOCIATE="symbol=value"^
14769 @cindex @option{^-D^/ASSOCIATE^} (@command{gnatprep})
14770 Defines a new preprocessing symbol, associated with value. If no value is given
14771 on the command line, then symbol is considered to be @code{True}. This switch
14772 can be used in place of a definition file.
14776 @cindex @option{/REMOVE} (@command{gnatprep})
14777 This is the default setting which causes lines deleted by preprocessing
14778 to be entirely removed from the output file.
14781 @item ^-r^/REFERENCE^
14782 @cindex @option{^-r^/REFERENCE^} (@command{gnatprep})
14783 Causes a @code{Source_Reference} pragma to be generated that
14784 references the original input file, so that error messages will use
14785 the file name of this original file. The use of this switch implies
14786 that preprocessor lines are not to be removed from the file, so its
14787 use will force @option{^-b^/BLANK_LINES^} mode if
14788 @option{^-c^/COMMENTS^}
14789 has not been specified explicitly.
14791 Note that if the file to be preprocessed contains multiple units, then
14792 it will be necessary to @code{gnatchop} the output file from
14793 @code{gnatprep}. If a @code{Source_Reference} pragma is present
14794 in the preprocessed file, it will be respected by
14795 @code{gnatchop ^-r^/REFERENCE^}
14796 so that the final chopped files will correctly refer to the original
14797 input source file for @code{gnatprep}.
14799 @item ^-s^/SYMBOLS^
14800 @cindex @option{^-s^/SYMBOLS^} (@command{gnatprep})
14801 Causes a sorted list of symbol names and values to be
14802 listed on the standard output file.
14804 @item ^-u^/UNDEFINED^
14805 @cindex @option{^-u^/UNDEFINED^} (@command{gnatprep})
14806 Causes undefined symbols to be treated as having the value FALSE in the context
14807 of a preprocessor test. In the absence of this option, an undefined symbol in
14808 a @code{#if} or @code{#elsif} test will be treated as an error.
14814 Note: if neither @option{-b} nor @option{-c} is present,
14815 then preprocessor lines and
14816 deleted lines are completely removed from the output, unless -r is
14817 specified, in which case -b is assumed.
14820 @node Form of Definitions File
14821 @section Form of Definitions File
14824 The definitions file contains lines of the form
14831 where symbol is a preprocessing symbol, and value is one of the following:
14835 Empty, corresponding to a null substitution
14837 A string literal using normal Ada syntax
14839 Any sequence of characters from the set
14840 (letters, digits, period, underline).
14844 Comment lines may also appear in the definitions file, starting with
14845 the usual @code{--},
14846 and comments may be added to the definitions lines.
14848 @node Form of Input Text for gnatprep
14849 @section Form of Input Text for @code{gnatprep}
14852 The input text may contain preprocessor conditional inclusion lines,
14853 as well as general symbol substitution sequences.
14855 The preprocessor conditional inclusion commands have the form
14860 #if @i{expression} @r{[}then@r{]}
14862 #elsif @i{expression} @r{[}then@r{]}
14864 #elsif @i{expression} @r{[}then@r{]}
14875 In this example, @i{expression} is defined by the following grammar:
14877 @i{expression} ::= <symbol>
14878 @i{expression} ::= <symbol> = "<value>"
14879 @i{expression} ::= <symbol> = <symbol>
14880 @i{expression} ::= <symbol> 'Defined
14881 @i{expression} ::= not @i{expression}
14882 @i{expression} ::= @i{expression} and @i{expression}
14883 @i{expression} ::= @i{expression} or @i{expression}
14884 @i{expression} ::= @i{expression} and then @i{expression}
14885 @i{expression} ::= @i{expression} or else @i{expression}
14886 @i{expression} ::= ( @i{expression} )
14889 The following restriction exists: it is not allowed to have "and" or "or"
14890 following "not" in the same expression without parentheses. For example, this
14897 This should be one of the following:
14905 For the first test (@i{expression} ::= <symbol>) the symbol must have
14906 either the value true or false, that is to say the right-hand of the
14907 symbol definition must be one of the (case-insensitive) literals
14908 @code{True} or @code{False}. If the value is true, then the
14909 corresponding lines are included, and if the value is false, they are
14912 The test (@i{expression} ::= <symbol> @code{'Defined}) is true only if
14913 the symbol has been defined in the definition file or by a @option{-D}
14914 switch on the command line. Otherwise, the test is false.
14916 The equality tests are case insensitive, as are all the preprocessor lines.
14918 If the symbol referenced is not defined in the symbol definitions file,
14919 then the effect depends on whether or not switch @option{-u}
14920 is specified. If so, then the symbol is treated as if it had the value
14921 false and the test fails. If this switch is not specified, then
14922 it is an error to reference an undefined symbol. It is also an error to
14923 reference a symbol that is defined with a value other than @code{True}
14926 The use of the @code{not} operator inverts the sense of this logical test.
14927 The @code{not} operator cannot be combined with the @code{or} or @code{and}
14928 operators, without parentheses. For example, "if not X or Y then" is not
14929 allowed, but "if (not X) or Y then" and "if not (X or Y) then" are.
14931 The @code{then} keyword is optional as shown
14933 The @code{#} must be the first non-blank character on a line, but
14934 otherwise the format is free form. Spaces or tabs may appear between
14935 the @code{#} and the keyword. The keywords and the symbols are case
14936 insensitive as in normal Ada code. Comments may be used on a
14937 preprocessor line, but other than that, no other tokens may appear on a
14938 preprocessor line. Any number of @code{elsif} clauses can be present,
14939 including none at all. The @code{else} is optional, as in Ada.
14941 The @code{#} marking the start of a preprocessor line must be the first
14942 non-blank character on the line, i.e., it must be preceded only by
14943 spaces or horizontal tabs.
14945 Symbol substitution outside of preprocessor lines is obtained by using
14953 anywhere within a source line, except in a comment or within a
14954 string literal. The identifier
14955 following the @code{$} must match one of the symbols defined in the symbol
14956 definition file, and the result is to substitute the value of the
14957 symbol in place of @code{$symbol} in the output file.
14959 Note that although the substitution of strings within a string literal
14960 is not possible, it is possible to have a symbol whose defined value is
14961 a string literal. So instead of setting XYZ to @code{hello} and writing:
14964 Header : String := "$XYZ";
14968 you should set XYZ to @code{"hello"} and write:
14971 Header : String := $XYZ;
14975 and then the substitution will occur as desired.
14978 @node The GNAT Run-Time Library Builder gnatlbr
14979 @chapter The GNAT Run-Time Library Builder @code{gnatlbr}
14981 @cindex Library builder
14984 @code{gnatlbr} is a tool for rebuilding the GNAT run time with user
14985 supplied configuration pragmas.
14988 * Running gnatlbr::
14989 * Switches for gnatlbr::
14990 * Examples of gnatlbr Usage::
14993 @node Running gnatlbr
14994 @section Running @code{gnatlbr}
14997 The @code{gnatlbr} command has the form
15000 $ GNAT LIBRARY /@r{[}CREATE@r{|}SET@r{|}DELETE@r{]}=directory @r{[}/CONFIG=file@r{]}
15003 @node Switches for gnatlbr
15004 @section Switches for @code{gnatlbr}
15007 @code{gnatlbr} recognizes the following switches:
15011 @item /CREATE=directory
15012 @cindex @code{/CREATE} (@code{gnatlbr})
15013 Create the new run-time library in the specified directory.
15015 @item /SET=directory
15016 @cindex @code{/SET} (@code{gnatlbr})
15017 Make the library in the specified directory the current run-time library.
15019 @item /DELETE=directory
15020 @cindex @code{/DELETE} (@code{gnatlbr})
15021 Delete the run-time library in the specified directory.
15024 @cindex @code{/CONFIG} (@code{gnatlbr})
15025 With /CREATE: Use the configuration pragmas in the specified file when
15026 building the library.
15028 With /SET: Use the configuration pragmas in the specified file when
15033 @node Examples of gnatlbr Usage
15034 @section Example of @code{gnatlbr} Usage
15037 Contents of VAXFLOAT.ADC:
15038 pragma Float_Representation (VAX_Float);
15040 $ GNAT LIBRARY /CREATE=[.VAXFLOAT] /CONFIG=VAXFLOAT.ADC
15042 GNAT LIBRARY rebuilds the run-time library in directory [.VAXFLOAT]
15047 @node The GNAT Library Browser gnatls
15048 @chapter The GNAT Library Browser @code{gnatls}
15050 @cindex Library browser
15053 @code{gnatls} is a tool that outputs information about compiled
15054 units. It gives the relationship between objects, unit names and source
15055 files. It can also be used to check the source dependencies of a unit
15056 as well as various characteristics.
15058 Note: to invoke @code{gnatls} with a project file, use the @code{gnat}
15059 driver (see @ref{The GNAT Driver and Project Files}).
15063 * Switches for gnatls::
15064 * Examples of gnatls Usage::
15067 @node Running gnatls
15068 @section Running @code{gnatls}
15071 The @code{gnatls} command has the form
15074 $ gnatls switches @var{object_or_ali_file}
15078 The main argument is the list of object or @file{ali} files
15079 (@pxref{The Ada Library Information Files})
15080 for which information is requested.
15082 In normal mode, without additional option, @code{gnatls} produces a
15083 four-column listing. Each line represents information for a specific
15084 object. The first column gives the full path of the object, the second
15085 column gives the name of the principal unit in this object, the third
15086 column gives the status of the source and the fourth column gives the
15087 full path of the source representing this unit.
15088 Here is a simple example of use:
15092 ^./^[]^demo1.o demo1 DIF demo1.adb
15093 ^./^[]^demo2.o demo2 OK demo2.adb
15094 ^./^[]^hello.o h1 OK hello.adb
15095 ^./^[]^instr-child.o instr.child MOK instr-child.adb
15096 ^./^[]^instr.o instr OK instr.adb
15097 ^./^[]^tef.o tef DIF tef.adb
15098 ^./^[]^text_io_example.o text_io_example OK text_io_example.adb
15099 ^./^[]^tgef.o tgef DIF tgef.adb
15103 The first line can be interpreted as follows: the main unit which is
15105 object file @file{demo1.o} is demo1, whose main source is in
15106 @file{demo1.adb}. Furthermore, the version of the source used for the
15107 compilation of demo1 has been modified (DIF). Each source file has a status
15108 qualifier which can be:
15111 @item OK (unchanged)
15112 The version of the source file used for the compilation of the
15113 specified unit corresponds exactly to the actual source file.
15115 @item MOK (slightly modified)
15116 The version of the source file used for the compilation of the
15117 specified unit differs from the actual source file but not enough to
15118 require recompilation. If you use gnatmake with the qualifier
15119 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}, a file marked
15120 MOK will not be recompiled.
15122 @item DIF (modified)
15123 No version of the source found on the path corresponds to the source
15124 used to build this object.
15126 @item ??? (file not found)
15127 No source file was found for this unit.
15129 @item HID (hidden, unchanged version not first on PATH)
15130 The version of the source that corresponds exactly to the source used
15131 for compilation has been found on the path but it is hidden by another
15132 version of the same source that has been modified.
15136 @node Switches for gnatls
15137 @section Switches for @code{gnatls}
15140 @code{gnatls} recognizes the following switches:
15144 @cindex @option{--version} @command{gnatls}
15145 Display Copyright and version, then exit disregarding all other options.
15148 @cindex @option{--help} @command{gnatls}
15149 If @option{--version} was not used, display usage, then exit disregarding
15152 @item ^-a^/ALL_UNITS^
15153 @cindex @option{^-a^/ALL_UNITS^} (@code{gnatls})
15154 Consider all units, including those of the predefined Ada library.
15155 Especially useful with @option{^-d^/DEPENDENCIES^}.
15157 @item ^-d^/DEPENDENCIES^
15158 @cindex @option{^-d^/DEPENDENCIES^} (@code{gnatls})
15159 List sources from which specified units depend on.
15161 @item ^-h^/OUTPUT=OPTIONS^
15162 @cindex @option{^-h^/OUTPUT=OPTIONS^} (@code{gnatls})
15163 Output the list of options.
15165 @item ^-o^/OUTPUT=OBJECTS^
15166 @cindex @option{^-o^/OUTPUT=OBJECTS^} (@code{gnatls})
15167 Only output information about object files.
15169 @item ^-s^/OUTPUT=SOURCES^
15170 @cindex @option{^-s^/OUTPUT=SOURCES^} (@code{gnatls})
15171 Only output information about source files.
15173 @item ^-u^/OUTPUT=UNITS^
15174 @cindex @option{^-u^/OUTPUT=UNITS^} (@code{gnatls})
15175 Only output information about compilation units.
15177 @item ^-files^/FILES^=@var{file}
15178 @cindex @option{^-files^/FILES^} (@code{gnatls})
15179 Take as arguments the files listed in text file @var{file}.
15180 Text file @var{file} may contain empty lines that are ignored.
15181 Each nonempty line should contain the name of an existing file.
15182 Several such switches may be specified simultaneously.
15184 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
15185 @itemx ^-aI^/SOURCE_SEARCH=^@var{dir}
15186 @itemx ^-I^/SEARCH=^@var{dir}
15187 @itemx ^-I-^/NOCURRENT_DIRECTORY^
15189 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatls})
15190 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatls})
15191 @cindex @option{^-I^/SEARCH^} (@code{gnatls})
15192 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatls})
15193 Source path manipulation. Same meaning as the equivalent @command{gnatmake}
15194 flags (@pxref{Switches for gnatmake}).
15196 @item --RTS=@var{rts-path}
15197 @cindex @option{--RTS} (@code{gnatls})
15198 Specifies the default location of the runtime library. Same meaning as the
15199 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
15201 @item ^-v^/OUTPUT=VERBOSE^
15202 @cindex @option{^-v^/OUTPUT=VERBOSE^} (@code{gnatls})
15203 Verbose mode. Output the complete source, object and project paths. Do not use
15204 the default column layout but instead use long format giving as much as
15205 information possible on each requested units, including special
15206 characteristics such as:
15209 @item Preelaborable
15210 The unit is preelaborable in the Ada sense.
15213 No elaboration code has been produced by the compiler for this unit.
15216 The unit is pure in the Ada sense.
15218 @item Elaborate_Body
15219 The unit contains a pragma Elaborate_Body.
15222 The unit contains a pragma Remote_Types.
15224 @item Shared_Passive
15225 The unit contains a pragma Shared_Passive.
15228 This unit is part of the predefined environment and cannot be modified
15231 @item Remote_Call_Interface
15232 The unit contains a pragma Remote_Call_Interface.
15238 @node Examples of gnatls Usage
15239 @section Example of @code{gnatls} Usage
15243 Example of using the verbose switch. Note how the source and
15244 object paths are affected by the -I switch.
15247 $ gnatls -v -I.. demo1.o
15249 GNATLS 5.03w (20041123-34)
15250 Copyright 1997-2004 Free Software Foundation, Inc.
15252 Source Search Path:
15253 <Current_Directory>
15255 /home/comar/local/adainclude/
15257 Object Search Path:
15258 <Current_Directory>
15260 /home/comar/local/lib/gcc-lib/x86-linux/3.4.3/adalib/
15262 Project Search Path:
15263 <Current_Directory>
15264 /home/comar/local/lib/gnat/
15269 Kind => subprogram body
15270 Flags => No_Elab_Code
15271 Source => demo1.adb modified
15275 The following is an example of use of the dependency list.
15276 Note the use of the -s switch
15277 which gives a straight list of source files. This can be useful for
15278 building specialized scripts.
15281 $ gnatls -d demo2.o
15282 ./demo2.o demo2 OK demo2.adb
15288 $ gnatls -d -s -a demo1.o
15290 /home/comar/local/adainclude/ada.ads
15291 /home/comar/local/adainclude/a-finali.ads
15292 /home/comar/local/adainclude/a-filico.ads
15293 /home/comar/local/adainclude/a-stream.ads
15294 /home/comar/local/adainclude/a-tags.ads
15297 /home/comar/local/adainclude/gnat.ads
15298 /home/comar/local/adainclude/g-io.ads
15300 /home/comar/local/adainclude/system.ads
15301 /home/comar/local/adainclude/s-exctab.ads
15302 /home/comar/local/adainclude/s-finimp.ads
15303 /home/comar/local/adainclude/s-finroo.ads
15304 /home/comar/local/adainclude/s-secsta.ads
15305 /home/comar/local/adainclude/s-stalib.ads
15306 /home/comar/local/adainclude/s-stoele.ads
15307 /home/comar/local/adainclude/s-stratt.ads
15308 /home/comar/local/adainclude/s-tasoli.ads
15309 /home/comar/local/adainclude/s-unstyp.ads
15310 /home/comar/local/adainclude/unchconv.ads
15316 GNAT LIST /DEPENDENCIES /OUTPUT=SOURCES /ALL_UNITS DEMO1.ADB
15318 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]ada.ads
15319 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-finali.ads
15320 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-filico.ads
15321 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-stream.ads
15322 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-tags.ads
15326 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]gnat.ads
15327 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]g-io.ads
15329 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]system.ads
15330 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-exctab.ads
15331 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finimp.ads
15332 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finroo.ads
15333 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-secsta.ads
15334 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stalib.ads
15335 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stoele.ads
15336 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stratt.ads
15337 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-tasoli.ads
15338 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-unstyp.ads
15339 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]unchconv.ads
15343 @node Cleaning Up Using gnatclean
15344 @chapter Cleaning Up Using @code{gnatclean}
15346 @cindex Cleaning tool
15349 @code{gnatclean} is a tool that allows the deletion of files produced by the
15350 compiler, binder and linker, including ALI files, object files, tree files,
15351 expanded source files, library files, interface copy source files, binder
15352 generated files and executable files.
15355 * Running gnatclean::
15356 * Switches for gnatclean::
15357 @c * Examples of gnatclean Usage::
15360 @node Running gnatclean
15361 @section Running @code{gnatclean}
15364 The @code{gnatclean} command has the form:
15367 $ gnatclean switches @var{names}
15371 @var{names} is a list of source file names. Suffixes @code{.^ads^ADS^} and
15372 @code{^adb^ADB^} may be omitted. If a project file is specified using switch
15373 @code{^-P^/PROJECT_FILE=^}, then @var{names} may be completely omitted.
15376 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
15377 if switch @code{^-c^/COMPILER_FILES_ONLY^} is not specified, by the binder and
15378 the linker. In informative-only mode, specified by switch
15379 @code{^-n^/NODELETE^}, the list of files that would have been deleted in
15380 normal mode is listed, but no file is actually deleted.
15382 @node Switches for gnatclean
15383 @section Switches for @code{gnatclean}
15386 @code{gnatclean} recognizes the following switches:
15390 @cindex @option{--version} @command{gnatclean}
15391 Display Copyright and version, then exit disregarding all other options.
15394 @cindex @option{--help} @command{gnatclean}
15395 If @option{--version} was not used, display usage, then exit disregarding
15398 @item ^--subdirs^/SUBDIRS^=subdir
15399 Actual object directory of each project file is the subdirectory subdir of the
15400 object directory specified or defauted in the project file.
15402 @item ^--unchecked-shared-lib-imports^/UNCHECKED_SHARED_LIB_IMPORTS^
15403 By default, shared library projects are not allowed to import static library
15404 projects. When this switch is used on the command line, this restriction is
15407 @item ^-c^/COMPILER_FILES_ONLY^
15408 @cindex @option{^-c^/COMPILER_FILES_ONLY^} (@code{gnatclean})
15409 Only attempt to delete the files produced by the compiler, not those produced
15410 by the binder or the linker. The files that are not to be deleted are library
15411 files, interface copy files, binder generated files and executable files.
15413 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
15414 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatclean})
15415 Indicate that ALI and object files should normally be found in directory
15418 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
15419 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatclean})
15420 When using project files, if some errors or warnings are detected during
15421 parsing and verbose mode is not in effect (no use of switch
15422 ^-v^/VERBOSE^), then error lines start with the full path name of the project
15423 file, rather than its simple file name.
15426 @cindex @option{^-h^/HELP^} (@code{gnatclean})
15427 Output a message explaining the usage of @code{^gnatclean^gnatclean^}.
15429 @item ^-n^/NODELETE^
15430 @cindex @option{^-n^/NODELETE^} (@code{gnatclean})
15431 Informative-only mode. Do not delete any files. Output the list of the files
15432 that would have been deleted if this switch was not specified.
15434 @item ^-P^/PROJECT_FILE=^@var{project}
15435 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatclean})
15436 Use project file @var{project}. Only one such switch can be used.
15437 When cleaning a project file, the files produced by the compilation of the
15438 immediate sources or inherited sources of the project files are to be
15439 deleted. This is not depending on the presence or not of executable names
15440 on the command line.
15443 @cindex @option{^-q^/QUIET^} (@code{gnatclean})
15444 Quiet output. If there are no errors, do not output anything, except in
15445 verbose mode (switch ^-v^/VERBOSE^) or in informative-only mode
15446 (switch ^-n^/NODELETE^).
15448 @item ^-r^/RECURSIVE^
15449 @cindex @option{^-r^/RECURSIVE^} (@code{gnatclean})
15450 When a project file is specified (using switch ^-P^/PROJECT_FILE=^),
15451 clean all imported and extended project files, recursively. If this switch
15452 is not specified, only the files related to the main project file are to be
15453 deleted. This switch has no effect if no project file is specified.
15455 @item ^-v^/VERBOSE^
15456 @cindex @option{^-v^/VERBOSE^} (@code{gnatclean})
15459 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
15460 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (@code{gnatclean})
15461 Indicates the verbosity of the parsing of GNAT project files.
15462 @xref{Switches Related to Project Files}.
15464 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
15465 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (@code{gnatclean})
15466 Indicates that external variable @var{name} has the value @var{value}.
15467 The Project Manager will use this value for occurrences of
15468 @code{external(name)} when parsing the project file.
15469 @xref{Switches Related to Project Files}.
15471 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
15472 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatclean})
15473 When searching for ALI and object files, look in directory
15476 @item ^-I^/SEARCH=^@var{dir}
15477 @cindex @option{^-I^/SEARCH^} (@code{gnatclean})
15478 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}}.
15480 @item ^-I-^/NOCURRENT_DIRECTORY^
15481 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatclean})
15482 @cindex Source files, suppressing search
15483 Do not look for ALI or object files in the directory
15484 where @code{gnatclean} was invoked.
15488 @c @node Examples of gnatclean Usage
15489 @c @section Examples of @code{gnatclean} Usage
15492 @node GNAT and Libraries
15493 @chapter GNAT and Libraries
15494 @cindex Library, building, installing, using
15497 This chapter describes how to build and use libraries with GNAT, and also shows
15498 how to recompile the GNAT run-time library. You should be familiar with the
15499 Project Manager facility (@pxref{GNAT Project Manager}) before reading this
15503 * Introduction to Libraries in GNAT::
15504 * General Ada Libraries::
15505 * Stand-alone Ada Libraries::
15506 * Rebuilding the GNAT Run-Time Library::
15509 @node Introduction to Libraries in GNAT
15510 @section Introduction to Libraries in GNAT
15513 A library is, conceptually, a collection of objects which does not have its
15514 own main thread of execution, but rather provides certain services to the
15515 applications that use it. A library can be either statically linked with the
15516 application, in which case its code is directly included in the application,
15517 or, on platforms that support it, be dynamically linked, in which case
15518 its code is shared by all applications making use of this library.
15520 GNAT supports both types of libraries.
15521 In the static case, the compiled code can be provided in different ways. The
15522 simplest approach is to provide directly the set of objects resulting from
15523 compilation of the library source files. Alternatively, you can group the
15524 objects into an archive using whatever commands are provided by the operating
15525 system. For the latter case, the objects are grouped into a shared library.
15527 In the GNAT environment, a library has three types of components:
15533 @xref{The Ada Library Information Files}.
15535 Object files, an archive or a shared library.
15539 A GNAT library may expose all its source files, which is useful for
15540 documentation purposes. Alternatively, it may expose only the units needed by
15541 an external user to make use of the library. That is to say, the specs
15542 reflecting the library services along with all the units needed to compile
15543 those specs, which can include generic bodies or any body implementing an
15544 inlined routine. In the case of @emph{stand-alone libraries} those exposed
15545 units are called @emph{interface units} (@pxref{Stand-alone Ada Libraries}).
15547 All compilation units comprising an application, including those in a library,
15548 need to be elaborated in an order partially defined by Ada's semantics. GNAT
15549 computes the elaboration order from the @file{ALI} files and this is why they
15550 constitute a mandatory part of GNAT libraries.
15551 @emph{Stand-alone libraries} are the exception to this rule because a specific
15552 library elaboration routine is produced independently of the application(s)
15555 @node General Ada Libraries
15556 @section General Ada Libraries
15559 * Building a library::
15560 * Installing a library::
15561 * Using a library::
15564 @node Building a library
15565 @subsection Building a library
15568 The easiest way to build a library is to use the Project Manager,
15569 which supports a special type of project called a @emph{Library Project}
15570 (@pxref{Library Projects}).
15572 A project is considered a library project, when two project-level attributes
15573 are defined in it: @code{Library_Name} and @code{Library_Dir}. In order to
15574 control different aspects of library configuration, additional optional
15575 project-level attributes can be specified:
15578 This attribute controls whether the library is to be static or dynamic
15580 @item Library_Version
15581 This attribute specifies the library version; this value is used
15582 during dynamic linking of shared libraries to determine if the currently
15583 installed versions of the binaries are compatible.
15585 @item Library_Options
15587 These attributes specify additional low-level options to be used during
15588 library generation, and redefine the actual application used to generate
15593 The GNAT Project Manager takes full care of the library maintenance task,
15594 including recompilation of the source files for which objects do not exist
15595 or are not up to date, assembly of the library archive, and installation of
15596 the library (i.e., copying associated source, object and @file{ALI} files
15597 to the specified location).
15599 Here is a simple library project file:
15600 @smallexample @c ada
15602 for Source_Dirs use ("src1", "src2");
15603 for Object_Dir use "obj";
15604 for Library_Name use "mylib";
15605 for Library_Dir use "lib";
15606 for Library_Kind use "dynamic";
15611 and the compilation command to build and install the library:
15613 @smallexample @c ada
15614 $ gnatmake -Pmy_lib
15618 It is not entirely trivial to perform manually all the steps required to
15619 produce a library. We recommend that you use the GNAT Project Manager
15620 for this task. In special cases where this is not desired, the necessary
15621 steps are discussed below.
15623 There are various possibilities for compiling the units that make up the
15624 library: for example with a Makefile (@pxref{Using the GNU make Utility}) or
15625 with a conventional script. For simple libraries, it is also possible to create
15626 a dummy main program which depends upon all the packages that comprise the
15627 interface of the library. This dummy main program can then be given to
15628 @command{gnatmake}, which will ensure that all necessary objects are built.
15630 After this task is accomplished, you should follow the standard procedure
15631 of the underlying operating system to produce the static or shared library.
15633 Here is an example of such a dummy program:
15634 @smallexample @c ada
15636 with My_Lib.Service1;
15637 with My_Lib.Service2;
15638 with My_Lib.Service3;
15639 procedure My_Lib_Dummy is
15647 Here are the generic commands that will build an archive or a shared library.
15650 # compiling the library
15651 $ gnatmake -c my_lib_dummy.adb
15653 # we don't need the dummy object itself
15654 $ rm my_lib_dummy.o my_lib_dummy.ali
15656 # create an archive with the remaining objects
15657 $ ar rc libmy_lib.a *.o
15658 # some systems may require "ranlib" to be run as well
15660 # or create a shared library
15661 $ gcc -shared -o libmy_lib.so *.o
15662 # some systems may require the code to have been compiled with -fPIC
15664 # remove the object files that are now in the library
15667 # Make the ALI files read-only so that gnatmake will not try to
15668 # regenerate the objects that are in the library
15673 Please note that the library must have a name of the form @file{lib@var{xxx}.a}
15674 or @file{lib@var{xxx}.so} (or @file{lib@var{xxx}.dll} on Windows) in order to
15675 be accessed by the directive @option{-l@var{xxx}} at link time.
15677 @node Installing a library
15678 @subsection Installing a library
15679 @cindex @code{ADA_PROJECT_PATH}
15680 @cindex @code{GPR_PROJECT_PATH}
15683 If you use project files, library installation is part of the library build
15684 process (@pxref{Installing a library with project files}).
15686 When project files are not an option, it is also possible, but not recommended,
15687 to install the library so that the sources needed to use the library are on the
15688 Ada source path and the ALI files & libraries be on the Ada Object path (see
15689 @ref{Search Paths and the Run-Time Library (RTL)}. Alternatively, the system
15690 administrator can place general-purpose libraries in the default compiler
15691 paths, by specifying the libraries' location in the configuration files
15692 @file{ada_source_path} and @file{ada_object_path}. These configuration files
15693 must be located in the GNAT installation tree at the same place as the gcc spec
15694 file. The location of the gcc spec file can be determined as follows:
15700 The configuration files mentioned above have a simple format: each line
15701 must contain one unique directory name.
15702 Those names are added to the corresponding path
15703 in their order of appearance in the file. The names can be either absolute
15704 or relative; in the latter case, they are relative to where theses files
15707 The files @file{ada_source_path} and @file{ada_object_path} might not be
15709 GNAT installation, in which case, GNAT will look for its run-time library in
15710 the directories @file{adainclude} (for the sources) and @file{adalib} (for the
15711 objects and @file{ALI} files). When the files exist, the compiler does not
15712 look in @file{adainclude} and @file{adalib}, and thus the
15713 @file{ada_source_path} file
15714 must contain the location for the GNAT run-time sources (which can simply
15715 be @file{adainclude}). In the same way, the @file{ada_object_path} file must
15716 contain the location for the GNAT run-time objects (which can simply
15719 You can also specify a new default path to the run-time library at compilation
15720 time with the switch @option{--RTS=rts-path}. You can thus choose / change
15721 the run-time library you want your program to be compiled with. This switch is
15722 recognized by @command{gcc}, @command{gnatmake}, @command{gnatbind},
15723 @command{gnatls}, @command{gnatfind} and @command{gnatxref}.
15725 It is possible to install a library before or after the standard GNAT
15726 library, by reordering the lines in the configuration files. In general, a
15727 library must be installed before the GNAT library if it redefines
15730 @node Using a library
15731 @subsection Using a library
15733 @noindent Once again, the project facility greatly simplifies the use of
15734 libraries. In this context, using a library is just a matter of adding a
15735 @code{with} clause in the user project. For instance, to make use of the
15736 library @code{My_Lib} shown in examples in earlier sections, you can
15739 @smallexample @c projectfile
15746 Even if you have a third-party, non-Ada library, you can still use GNAT's
15747 Project Manager facility to provide a wrapper for it. For example, the
15748 following project, when @code{with}ed by your main project, will link with the
15749 third-party library @file{liba.a}:
15751 @smallexample @c projectfile
15754 for Externally_Built use "true";
15755 for Source_Files use ();
15756 for Library_Dir use "lib";
15757 for Library_Name use "a";
15758 for Library_Kind use "static";
15762 This is an alternative to the use of @code{pragma Linker_Options}. It is
15763 especially interesting in the context of systems with several interdependent
15764 static libraries where finding a proper linker order is not easy and best be
15765 left to the tools having visibility over project dependence information.
15768 In order to use an Ada library manually, you need to make sure that this
15769 library is on both your source and object path
15770 (see @ref{Search Paths and the Run-Time Library (RTL)}
15771 and @ref{Search Paths for gnatbind}). Furthermore, when the objects are grouped
15772 in an archive or a shared library, you need to specify the desired
15773 library at link time.
15775 For example, you can use the library @file{mylib} installed in
15776 @file{/dir/my_lib_src} and @file{/dir/my_lib_obj} with the following commands:
15779 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \
15784 This can be expressed more simply:
15789 when the following conditions are met:
15792 @file{/dir/my_lib_src} has been added by the user to the environment
15793 variable @env{ADA_INCLUDE_PATH}, or by the administrator to the file
15794 @file{ada_source_path}
15796 @file{/dir/my_lib_obj} has been added by the user to the environment
15797 variable @env{ADA_OBJECTS_PATH}, or by the administrator to the file
15798 @file{ada_object_path}
15800 a pragma @code{Linker_Options} has been added to one of the sources.
15803 @smallexample @c ada
15804 pragma Linker_Options ("-lmy_lib");
15808 @node Stand-alone Ada Libraries
15809 @section Stand-alone Ada Libraries
15810 @cindex Stand-alone library, building, using
15813 * Introduction to Stand-alone Libraries::
15814 * Building a Stand-alone Library::
15815 * Creating a Stand-alone Library to be used in a non-Ada context::
15816 * Restrictions in Stand-alone Libraries::
15819 @node Introduction to Stand-alone Libraries
15820 @subsection Introduction to Stand-alone Libraries
15823 A Stand-alone Library (abbreviated ``SAL'') is a library that contains the
15825 elaborate the Ada units that are included in the library. In contrast with
15826 an ordinary library, which consists of all sources, objects and @file{ALI}
15828 library, a SAL may specify a restricted subset of compilation units
15829 to serve as a library interface. In this case, the fully
15830 self-sufficient set of files will normally consist of an objects
15831 archive, the sources of interface units' specs, and the @file{ALI}
15832 files of interface units.
15833 If an interface spec contains a generic unit or an inlined subprogram,
15835 source must also be provided; if the units that must be provided in the source
15836 form depend on other units, the source and @file{ALI} files of those must
15839 The main purpose of a SAL is to minimize the recompilation overhead of client
15840 applications when a new version of the library is installed. Specifically,
15841 if the interface sources have not changed, client applications do not need to
15842 be recompiled. If, furthermore, a SAL is provided in the shared form and its
15843 version, controlled by @code{Library_Version} attribute, is not changed,
15844 then the clients do not need to be relinked.
15846 SALs also allow the library providers to minimize the amount of library source
15847 text exposed to the clients. Such ``information hiding'' might be useful or
15848 necessary for various reasons.
15850 Stand-alone libraries are also well suited to be used in an executable whose
15851 main routine is not written in Ada.
15853 @node Building a Stand-alone Library
15854 @subsection Building a Stand-alone Library
15857 GNAT's Project facility provides a simple way of building and installing
15858 stand-alone libraries; see @ref{Stand-alone Library Projects}.
15859 To be a Stand-alone Library Project, in addition to the two attributes
15860 that make a project a Library Project (@code{Library_Name} and
15861 @code{Library_Dir}; see @ref{Library Projects}), the attribute
15862 @code{Library_Interface} must be defined. For example:
15864 @smallexample @c projectfile
15866 for Library_Dir use "lib_dir";
15867 for Library_Name use "dummy";
15868 for Library_Interface use ("int1", "int1.child");
15873 Attribute @code{Library_Interface} has a non-empty string list value,
15874 each string in the list designating a unit contained in an immediate source
15875 of the project file.
15877 When a Stand-alone Library is built, first the binder is invoked to build
15878 a package whose name depends on the library name
15879 (@file{^b~dummy.ads/b^B$DUMMY.ADS/B^} in the example above).
15880 This binder-generated package includes initialization and
15881 finalization procedures whose
15882 names depend on the library name (@code{dummyinit} and @code{dummyfinal}
15884 above). The object corresponding to this package is included in the library.
15886 You must ensure timely (e.g., prior to any use of interfaces in the SAL)
15887 calling of these procedures if a static SAL is built, or if a shared SAL
15889 with the project-level attribute @code{Library_Auto_Init} set to
15892 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
15893 (those that are listed in attribute @code{Library_Interface}) are copied to
15894 the Library Directory. As a consequence, only the Interface Units may be
15895 imported from Ada units outside of the library. If other units are imported,
15896 the binding phase will fail.
15898 The attribute @code{Library_Src_Dir} may be specified for a
15899 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
15900 single string value. Its value must be the path (absolute or relative to the
15901 project directory) of an existing directory. This directory cannot be the
15902 object directory or one of the source directories, but it can be the same as
15903 the library directory. The sources of the Interface
15904 Units of the library that are needed by an Ada client of the library will be
15905 copied to the designated directory, called the Interface Copy directory.
15906 These sources include the specs of the Interface Units, but they may also
15907 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
15908 are used, or when there is a generic unit in the spec. Before the sources
15909 are copied to the Interface Copy directory, an attempt is made to delete all
15910 files in the Interface Copy directory.
15912 Building stand-alone libraries by hand is somewhat tedious, but for those
15913 occasions when it is necessary here are the steps that you need to perform:
15916 Compile all library sources.
15919 Invoke the binder with the switch @option{-n} (No Ada main program),
15920 with all the @file{ALI} files of the interfaces, and
15921 with the switch @option{-L} to give specific names to the @code{init}
15922 and @code{final} procedures. For example:
15924 gnatbind -n int1.ali int2.ali -Lsal1
15928 Compile the binder generated file:
15934 Link the dynamic library with all the necessary object files,
15935 indicating to the linker the names of the @code{init} (and possibly
15936 @code{final}) procedures for automatic initialization (and finalization).
15937 The built library should be placed in a directory different from
15938 the object directory.
15941 Copy the @code{ALI} files of the interface to the library directory,
15942 add in this copy an indication that it is an interface to a SAL
15943 (i.e., add a word @option{SL} on the line in the @file{ALI} file that starts
15944 with letter ``P'') and make the modified copy of the @file{ALI} file
15949 Using SALs is not different from using other libraries
15950 (see @ref{Using a library}).
15952 @node Creating a Stand-alone Library to be used in a non-Ada context
15953 @subsection Creating a Stand-alone Library to be used in a non-Ada context
15956 It is easy to adapt the SAL build procedure discussed above for use of a SAL in
15959 The only extra step required is to ensure that library interface subprograms
15960 are compatible with the main program, by means of @code{pragma Export}
15961 or @code{pragma Convention}.
15963 Here is an example of simple library interface for use with C main program:
15965 @smallexample @c ada
15966 package My_Package is
15968 procedure Do_Something;
15969 pragma Export (C, Do_Something, "do_something");
15971 procedure Do_Something_Else;
15972 pragma Export (C, Do_Something_Else, "do_something_else");
15978 On the foreign language side, you must provide a ``foreign'' view of the
15979 library interface; remember that it should contain elaboration routines in
15980 addition to interface subprograms.
15982 The example below shows the content of @code{mylib_interface.h} (note
15983 that there is no rule for the naming of this file, any name can be used)
15985 /* the library elaboration procedure */
15986 extern void mylibinit (void);
15988 /* the library finalization procedure */
15989 extern void mylibfinal (void);
15991 /* the interface exported by the library */
15992 extern void do_something (void);
15993 extern void do_something_else (void);
15997 Libraries built as explained above can be used from any program, provided
15998 that the elaboration procedures (named @code{mylibinit} in the previous
15999 example) are called before the library services are used. Any number of
16000 libraries can be used simultaneously, as long as the elaboration
16001 procedure of each library is called.
16003 Below is an example of a C program that uses the @code{mylib} library.
16006 #include "mylib_interface.h"
16011 /* First, elaborate the library before using it */
16014 /* Main program, using the library exported entities */
16016 do_something_else ();
16018 /* Library finalization at the end of the program */
16025 Note that invoking any library finalization procedure generated by
16026 @code{gnatbind} shuts down the Ada run-time environment.
16028 finalization of all Ada libraries must be performed at the end of the program.
16029 No call to these libraries or to the Ada run-time library should be made
16030 after the finalization phase.
16032 @node Restrictions in Stand-alone Libraries
16033 @subsection Restrictions in Stand-alone Libraries
16036 The pragmas listed below should be used with caution inside libraries,
16037 as they can create incompatibilities with other Ada libraries:
16039 @item pragma @code{Locking_Policy}
16040 @item pragma @code{Queuing_Policy}
16041 @item pragma @code{Task_Dispatching_Policy}
16042 @item pragma @code{Unreserve_All_Interrupts}
16046 When using a library that contains such pragmas, the user must make sure
16047 that all libraries use the same pragmas with the same values. Otherwise,
16048 @code{Program_Error} will
16049 be raised during the elaboration of the conflicting
16050 libraries. The usage of these pragmas and its consequences for the user
16051 should therefore be well documented.
16053 Similarly, the traceback in the exception occurrence mechanism should be
16054 enabled or disabled in a consistent manner across all libraries.
16055 Otherwise, Program_Error will be raised during the elaboration of the
16056 conflicting libraries.
16058 If the @code{Version} or @code{Body_Version}
16059 attributes are used inside a library, then you need to
16060 perform a @code{gnatbind} step that specifies all @file{ALI} files in all
16061 libraries, so that version identifiers can be properly computed.
16062 In practice these attributes are rarely used, so this is unlikely
16063 to be a consideration.
16065 @node Rebuilding the GNAT Run-Time Library
16066 @section Rebuilding the GNAT Run-Time Library
16067 @cindex GNAT Run-Time Library, rebuilding
16068 @cindex Building the GNAT Run-Time Library
16069 @cindex Rebuilding the GNAT Run-Time Library
16070 @cindex Run-Time Library, rebuilding
16073 It may be useful to recompile the GNAT library in various contexts, the
16074 most important one being the use of partition-wide configuration pragmas
16075 such as @code{Normalize_Scalars}. A special Makefile called
16076 @code{Makefile.adalib} is provided to that effect and can be found in
16077 the directory containing the GNAT library. The location of this
16078 directory depends on the way the GNAT environment has been installed and can
16079 be determined by means of the command:
16086 The last entry in the object search path usually contains the
16087 gnat library. This Makefile contains its own documentation and in
16088 particular the set of instructions needed to rebuild a new library and
16091 @node Using the GNU make Utility
16092 @chapter Using the GNU @code{make} Utility
16096 This chapter offers some examples of makefiles that solve specific
16097 problems. It does not explain how to write a makefile (@pxref{Top,, GNU
16098 make, make, GNU @code{make}}), nor does it try to replace the
16099 @command{gnatmake} utility (@pxref{The GNAT Make Program gnatmake}).
16101 All the examples in this section are specific to the GNU version of
16102 make. Although @command{make} is a standard utility, and the basic language
16103 is the same, these examples use some advanced features found only in
16107 * Using gnatmake in a Makefile::
16108 * Automatically Creating a List of Directories::
16109 * Generating the Command Line Switches::
16110 * Overcoming Command Line Length Limits::
16113 @node Using gnatmake in a Makefile
16114 @section Using gnatmake in a Makefile
16119 Complex project organizations can be handled in a very powerful way by
16120 using GNU make combined with gnatmake. For instance, here is a Makefile
16121 which allows you to build each subsystem of a big project into a separate
16122 shared library. Such a makefile allows you to significantly reduce the link
16123 time of very big applications while maintaining full coherence at
16124 each step of the build process.
16126 The list of dependencies are handled automatically by
16127 @command{gnatmake}. The Makefile is simply used to call gnatmake in each of
16128 the appropriate directories.
16130 Note that you should also read the example on how to automatically
16131 create the list of directories
16132 (@pxref{Automatically Creating a List of Directories})
16133 which might help you in case your project has a lot of subdirectories.
16138 @font@heightrm=cmr8
16141 ## This Makefile is intended to be used with the following directory
16143 ## - The sources are split into a series of csc (computer software components)
16144 ## Each of these csc is put in its own directory.
16145 ## Their name are referenced by the directory names.
16146 ## They will be compiled into shared library (although this would also work
16147 ## with static libraries
16148 ## - The main program (and possibly other packages that do not belong to any
16149 ## csc is put in the top level directory (where the Makefile is).
16150 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
16151 ## \_ second_csc (sources) __ lib (will contain the library)
16153 ## Although this Makefile is build for shared library, it is easy to modify
16154 ## to build partial link objects instead (modify the lines with -shared and
16157 ## With this makefile, you can change any file in the system or add any new
16158 ## file, and everything will be recompiled correctly (only the relevant shared
16159 ## objects will be recompiled, and the main program will be re-linked).
16161 # The list of computer software component for your project. This might be
16162 # generated automatically.
16165 # Name of the main program (no extension)
16168 # If we need to build objects with -fPIC, uncomment the following line
16171 # The following variable should give the directory containing libgnat.so
16172 # You can get this directory through 'gnatls -v'. This is usually the last
16173 # directory in the Object_Path.
16176 # The directories for the libraries
16177 # (This macro expands the list of CSC to the list of shared libraries, you
16178 # could simply use the expanded form:
16179 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
16180 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
16182 $@{MAIN@}: objects $@{LIB_DIR@}
16183 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
16184 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
16187 # recompile the sources
16188 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
16190 # Note: In a future version of GNAT, the following commands will be simplified
16191 # by a new tool, gnatmlib
16193 mkdir -p $@{dir $@@ @}
16194 cd $@{dir $@@ @} && gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
16195 cd $@{dir $@@ @} && cp -f ../*.ali .
16197 # The dependencies for the modules
16198 # Note that we have to force the expansion of *.o, since in some cases
16199 # make won't be able to do it itself.
16200 aa/lib/libaa.so: $@{wildcard aa/*.o@}
16201 bb/lib/libbb.so: $@{wildcard bb/*.o@}
16202 cc/lib/libcc.so: $@{wildcard cc/*.o@}
16204 # Make sure all of the shared libraries are in the path before starting the
16207 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
16210 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
16211 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
16212 $@{RM@} $@{CSC_LIST:%=%/*.o@}
16213 $@{RM@} *.o *.ali $@{MAIN@}
16216 @node Automatically Creating a List of Directories
16217 @section Automatically Creating a List of Directories
16220 In most makefiles, you will have to specify a list of directories, and
16221 store it in a variable. For small projects, it is often easier to
16222 specify each of them by hand, since you then have full control over what
16223 is the proper order for these directories, which ones should be
16226 However, in larger projects, which might involve hundreds of
16227 subdirectories, it might be more convenient to generate this list
16230 The example below presents two methods. The first one, although less
16231 general, gives you more control over the list. It involves wildcard
16232 characters, that are automatically expanded by @command{make}. Its
16233 shortcoming is that you need to explicitly specify some of the
16234 organization of your project, such as for instance the directory tree
16235 depth, whether some directories are found in a separate tree, @enddots{}
16237 The second method is the most general one. It requires an external
16238 program, called @command{find}, which is standard on all Unix systems. All
16239 the directories found under a given root directory will be added to the
16245 @font@heightrm=cmr8
16248 # The examples below are based on the following directory hierarchy:
16249 # All the directories can contain any number of files
16250 # ROOT_DIRECTORY -> a -> aa -> aaa
16253 # -> b -> ba -> baa
16256 # This Makefile creates a variable called DIRS, that can be reused any time
16257 # you need this list (see the other examples in this section)
16259 # The root of your project's directory hierarchy
16263 # First method: specify explicitly the list of directories
16264 # This allows you to specify any subset of all the directories you need.
16267 DIRS := a/aa/ a/ab/ b/ba/
16270 # Second method: use wildcards
16271 # Note that the argument(s) to wildcard below should end with a '/'.
16272 # Since wildcards also return file names, we have to filter them out
16273 # to avoid duplicate directory names.
16274 # We thus use make's @code{dir} and @code{sort} functions.
16275 # It sets DIRs to the following value (note that the directories aaa and baa
16276 # are not given, unless you change the arguments to wildcard).
16277 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
16280 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
16281 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
16284 # Third method: use an external program
16285 # This command is much faster if run on local disks, avoiding NFS slowdowns.
16286 # This is the most complete command: it sets DIRs to the following value:
16287 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
16290 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
16294 @node Generating the Command Line Switches
16295 @section Generating the Command Line Switches
16298 Once you have created the list of directories as explained in the
16299 previous section (@pxref{Automatically Creating a List of Directories}),
16300 you can easily generate the command line arguments to pass to gnatmake.
16302 For the sake of completeness, this example assumes that the source path
16303 is not the same as the object path, and that you have two separate lists
16307 # see "Automatically creating a list of directories" to create
16312 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
16313 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
16316 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
16319 @node Overcoming Command Line Length Limits
16320 @section Overcoming Command Line Length Limits
16323 One problem that might be encountered on big projects is that many
16324 operating systems limit the length of the command line. It is thus hard to give
16325 gnatmake the list of source and object directories.
16327 This example shows how you can set up environment variables, which will
16328 make @command{gnatmake} behave exactly as if the directories had been
16329 specified on the command line, but have a much higher length limit (or
16330 even none on most systems).
16332 It assumes that you have created a list of directories in your Makefile,
16333 using one of the methods presented in
16334 @ref{Automatically Creating a List of Directories}.
16335 For the sake of completeness, we assume that the object
16336 path (where the ALI files are found) is different from the sources patch.
16338 Note a small trick in the Makefile below: for efficiency reasons, we
16339 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
16340 expanded immediately by @code{make}. This way we overcome the standard
16341 make behavior which is to expand the variables only when they are
16344 On Windows, if you are using the standard Windows command shell, you must
16345 replace colons with semicolons in the assignments to these variables.
16350 @font@heightrm=cmr8
16353 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECT_PATH.
16354 # This is the same thing as putting the -I arguments on the command line.
16355 # (the equivalent of using -aI on the command line would be to define
16356 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECT_PATH).
16357 # You can of course have different values for these variables.
16359 # Note also that we need to keep the previous values of these variables, since
16360 # they might have been set before running 'make' to specify where the GNAT
16361 # library is installed.
16363 # see "Automatically creating a list of directories" to create these
16369 space:=$@{empty@} $@{empty@}
16370 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
16371 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
16372 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
16373 ADA_OBJECT_PATH += $@{OBJECT_LIST@}
16374 export ADA_INCLUDE_PATH
16375 export ADA_OBJECT_PATH
16382 @node Memory Management Issues
16383 @chapter Memory Management Issues
16386 This chapter describes some useful memory pools provided in the GNAT library
16387 and in particular the GNAT Debug Pool facility, which can be used to detect
16388 incorrect uses of access values (including ``dangling references'').
16390 It also describes the @command{gnatmem} tool, which can be used to track down
16395 * Some Useful Memory Pools::
16396 * The GNAT Debug Pool Facility::
16398 * The gnatmem Tool::
16402 @node Some Useful Memory Pools
16403 @section Some Useful Memory Pools
16404 @findex Memory Pool
16405 @cindex storage, pool
16408 The @code{System.Pool_Global} package offers the Unbounded_No_Reclaim_Pool
16409 storage pool. Allocations use the standard system call @code{malloc} while
16410 deallocations use the standard system call @code{free}. No reclamation is
16411 performed when the pool goes out of scope. For performance reasons, the
16412 standard default Ada allocators/deallocators do not use any explicit storage
16413 pools but if they did, they could use this storage pool without any change in
16414 behavior. That is why this storage pool is used when the user
16415 manages to make the default implicit allocator explicit as in this example:
16416 @smallexample @c ada
16417 type T1 is access Something;
16418 -- no Storage pool is defined for T2
16419 type T2 is access Something_Else;
16420 for T2'Storage_Pool use T1'Storage_Pool;
16421 -- the above is equivalent to
16422 for T2'Storage_Pool use System.Pool_Global.Global_Pool_Object;
16426 The @code{System.Pool_Local} package offers the Unbounded_Reclaim_Pool storage
16427 pool. The allocation strategy is similar to @code{Pool_Local}'s
16428 except that the all
16429 storage allocated with this pool is reclaimed when the pool object goes out of
16430 scope. This pool provides a explicit mechanism similar to the implicit one
16431 provided by several Ada 83 compilers for allocations performed through a local
16432 access type and whose purpose was to reclaim memory when exiting the
16433 scope of a given local access. As an example, the following program does not
16434 leak memory even though it does not perform explicit deallocation:
16436 @smallexample @c ada
16437 with System.Pool_Local;
16438 procedure Pooloc1 is
16439 procedure Internal is
16440 type A is access Integer;
16441 X : System.Pool_Local.Unbounded_Reclaim_Pool;
16442 for A'Storage_Pool use X;
16445 for I in 1 .. 50 loop
16450 for I in 1 .. 100 loop
16457 The @code{System.Pool_Size} package implements the Stack_Bounded_Pool used when
16458 @code{Storage_Size} is specified for an access type.
16459 The whole storage for the pool is
16460 allocated at once, usually on the stack at the point where the access type is
16461 elaborated. It is automatically reclaimed when exiting the scope where the
16462 access type is defined. This package is not intended to be used directly by the
16463 user and it is implicitly used for each such declaration:
16465 @smallexample @c ada
16466 type T1 is access Something;
16467 for T1'Storage_Size use 10_000;
16470 @node The GNAT Debug Pool Facility
16471 @section The GNAT Debug Pool Facility
16473 @cindex storage, pool, memory corruption
16476 The use of unchecked deallocation and unchecked conversion can easily
16477 lead to incorrect memory references. The problems generated by such
16478 references are usually difficult to tackle because the symptoms can be
16479 very remote from the origin of the problem. In such cases, it is
16480 very helpful to detect the problem as early as possible. This is the
16481 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
16483 In order to use the GNAT specific debugging pool, the user must
16484 associate a debug pool object with each of the access types that may be
16485 related to suspected memory problems. See Ada Reference Manual 13.11.
16486 @smallexample @c ada
16487 type Ptr is access Some_Type;
16488 Pool : GNAT.Debug_Pools.Debug_Pool;
16489 for Ptr'Storage_Pool use Pool;
16493 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
16494 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
16495 allow the user to redefine allocation and deallocation strategies. They
16496 also provide a checkpoint for each dereference, through the use of
16497 the primitive operation @code{Dereference} which is implicitly called at
16498 each dereference of an access value.
16500 Once an access type has been associated with a debug pool, operations on
16501 values of the type may raise four distinct exceptions,
16502 which correspond to four potential kinds of memory corruption:
16505 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
16507 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
16509 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
16511 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage }
16515 For types associated with a Debug_Pool, dynamic allocation is performed using
16516 the standard GNAT allocation routine. References to all allocated chunks of
16517 memory are kept in an internal dictionary. Several deallocation strategies are
16518 provided, whereupon the user can choose to release the memory to the system,
16519 keep it allocated for further invalid access checks, or fill it with an easily
16520 recognizable pattern for debug sessions. The memory pattern is the old IBM
16521 hexadecimal convention: @code{16#DEADBEEF#}.
16523 See the documentation in the file g-debpoo.ads for more information on the
16524 various strategies.
16526 Upon each dereference, a check is made that the access value denotes a
16527 properly allocated memory location. Here is a complete example of use of
16528 @code{Debug_Pools}, that includes typical instances of memory corruption:
16529 @smallexample @c ada
16533 with Gnat.Io; use Gnat.Io;
16534 with Unchecked_Deallocation;
16535 with Unchecked_Conversion;
16536 with GNAT.Debug_Pools;
16537 with System.Storage_Elements;
16538 with Ada.Exceptions; use Ada.Exceptions;
16539 procedure Debug_Pool_Test is
16541 type T is access Integer;
16542 type U is access all T;
16544 P : GNAT.Debug_Pools.Debug_Pool;
16545 for T'Storage_Pool use P;
16547 procedure Free is new Unchecked_Deallocation (Integer, T);
16548 function UC is new Unchecked_Conversion (U, T);
16551 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
16561 Put_Line (Integer'Image(B.all));
16563 when E : others => Put_Line ("raised: " & Exception_Name (E));
16568 when E : others => Put_Line ("raised: " & Exception_Name (E));
16572 Put_Line (Integer'Image(B.all));
16574 when E : others => Put_Line ("raised: " & Exception_Name (E));
16579 when E : others => Put_Line ("raised: " & Exception_Name (E));
16582 end Debug_Pool_Test;
16586 The debug pool mechanism provides the following precise diagnostics on the
16587 execution of this erroneous program:
16590 Total allocated bytes : 0
16591 Total deallocated bytes : 0
16592 Current Water Mark: 0
16596 Total allocated bytes : 8
16597 Total deallocated bytes : 0
16598 Current Water Mark: 8
16601 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
16602 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
16603 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
16604 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
16606 Total allocated bytes : 8
16607 Total deallocated bytes : 4
16608 Current Water Mark: 4
16613 @node The gnatmem Tool
16614 @section The @command{gnatmem} Tool
16618 The @code{gnatmem} utility monitors dynamic allocation and
16619 deallocation activity in a program, and displays information about
16620 incorrect deallocations and possible sources of memory leaks.
16621 It is designed to work in association with a static runtime library
16622 only and in this context provides three types of information:
16625 General information concerning memory management, such as the total
16626 number of allocations and deallocations, the amount of allocated
16627 memory and the high water mark, i.e.@: the largest amount of allocated
16628 memory in the course of program execution.
16631 Backtraces for all incorrect deallocations, that is to say deallocations
16632 which do not correspond to a valid allocation.
16635 Information on each allocation that is potentially the origin of a memory
16640 * Running gnatmem::
16641 * Switches for gnatmem::
16642 * Example of gnatmem Usage::
16645 @node Running gnatmem
16646 @subsection Running @code{gnatmem}
16649 @code{gnatmem} makes use of the output created by the special version of
16650 allocation and deallocation routines that record call information. This
16651 allows to obtain accurate dynamic memory usage history at a minimal cost to
16652 the execution speed. Note however, that @code{gnatmem} is not supported on
16653 all platforms (currently, it is supported on AIX, HP-UX, GNU/Linux,
16654 Solaris and Windows NT/2000/XP (x86).
16657 The @code{gnatmem} command has the form
16660 @c $ gnatmem @ovar{switches} user_program
16661 @c Expanding @ovar macro inline (explanation in macro def comments)
16662 $ gnatmem @r{[}@var{switches}@r{]} @var{user_program}
16666 The program must have been linked with the instrumented version of the
16667 allocation and deallocation routines. This is done by linking with the
16668 @file{libgmem.a} library. For correct symbolic backtrace information,
16669 the user program should be compiled with debugging options
16670 (see @ref{Switches for gcc}). For example to build @file{my_program}:
16673 $ gnatmake -g my_program -largs -lgmem
16677 As library @file{libgmem.a} contains an alternate body for package
16678 @code{System.Memory}, @file{s-memory.adb} should not be compiled and linked
16679 when an executable is linked with library @file{libgmem.a}. It is then not
16680 recommended to use @command{gnatmake} with switch @option{^-a^/ALL_FILES^}.
16683 When @file{my_program} is executed, the file @file{gmem.out} is produced.
16684 This file contains information about all allocations and deallocations
16685 performed by the program. It is produced by the instrumented allocations and
16686 deallocations routines and will be used by @code{gnatmem}.
16688 In order to produce symbolic backtrace information for allocations and
16689 deallocations performed by the GNAT run-time library, you need to use a
16690 version of that library that has been compiled with the @option{-g} switch
16691 (see @ref{Rebuilding the GNAT Run-Time Library}).
16693 Gnatmem must be supplied with the @file{gmem.out} file and the executable to
16694 examine. If the location of @file{gmem.out} file was not explicitly supplied by
16695 @option{-i} switch, gnatmem will assume that this file can be found in the
16696 current directory. For example, after you have executed @file{my_program},
16697 @file{gmem.out} can be analyzed by @code{gnatmem} using the command:
16700 $ gnatmem my_program
16704 This will produce the output with the following format:
16706 *************** debut cc
16708 $ gnatmem my_program
16712 Total number of allocations : 45
16713 Total number of deallocations : 6
16714 Final Water Mark (non freed mem) : 11.29 Kilobytes
16715 High Water Mark : 11.40 Kilobytes
16720 Allocation Root # 2
16721 -------------------
16722 Number of non freed allocations : 11
16723 Final Water Mark (non freed mem) : 1.16 Kilobytes
16724 High Water Mark : 1.27 Kilobytes
16726 my_program.adb:23 my_program.alloc
16732 The first block of output gives general information. In this case, the
16733 Ada construct ``@code{@b{new}}'' was executed 45 times, and only 6 calls to an
16734 Unchecked_Deallocation routine occurred.
16737 Subsequent paragraphs display information on all allocation roots.
16738 An allocation root is a specific point in the execution of the program
16739 that generates some dynamic allocation, such as a ``@code{@b{new}}''
16740 construct. This root is represented by an execution backtrace (or subprogram
16741 call stack). By default the backtrace depth for allocations roots is 1, so
16742 that a root corresponds exactly to a source location. The backtrace can
16743 be made deeper, to make the root more specific.
16745 @node Switches for gnatmem
16746 @subsection Switches for @code{gnatmem}
16749 @code{gnatmem} recognizes the following switches:
16754 @cindex @option{-q} (@code{gnatmem})
16755 Quiet. Gives the minimum output needed to identify the origin of the
16756 memory leaks. Omits statistical information.
16759 @cindex @var{N} (@code{gnatmem})
16760 N is an integer literal (usually between 1 and 10) which controls the
16761 depth of the backtraces defining allocation root. The default value for
16762 N is 1. The deeper the backtrace, the more precise the localization of
16763 the root. Note that the total number of roots can depend on this
16764 parameter. This parameter must be specified @emph{before} the name of the
16765 executable to be analyzed, to avoid ambiguity.
16768 @cindex @option{-b} (@code{gnatmem})
16769 This switch has the same effect as just depth parameter.
16771 @item -i @var{file}
16772 @cindex @option{-i} (@code{gnatmem})
16773 Do the @code{gnatmem} processing starting from @file{file}, rather than
16774 @file{gmem.out} in the current directory.
16777 @cindex @option{-m} (@code{gnatmem})
16778 This switch causes @code{gnatmem} to mask the allocation roots that have less
16779 than n leaks. The default value is 1. Specifying the value of 0 will allow to
16780 examine even the roots that didn't result in leaks.
16783 @cindex @option{-s} (@code{gnatmem})
16784 This switch causes @code{gnatmem} to sort the allocation roots according to the
16785 specified order of sort criteria, each identified by a single letter. The
16786 currently supported criteria are @code{n, h, w} standing respectively for
16787 number of unfreed allocations, high watermark, and final watermark
16788 corresponding to a specific root. The default order is @code{nwh}.
16792 @node Example of gnatmem Usage
16793 @subsection Example of @code{gnatmem} Usage
16796 The following example shows the use of @code{gnatmem}
16797 on a simple memory-leaking program.
16798 Suppose that we have the following Ada program:
16800 @smallexample @c ada
16803 with Unchecked_Deallocation;
16804 procedure Test_Gm is
16806 type T is array (1..1000) of Integer;
16807 type Ptr is access T;
16808 procedure Free is new Unchecked_Deallocation (T, Ptr);
16811 procedure My_Alloc is
16816 procedure My_DeAlloc is
16824 for I in 1 .. 5 loop
16825 for J in I .. 5 loop
16836 The program needs to be compiled with debugging option and linked with
16837 @code{gmem} library:
16840 $ gnatmake -g test_gm -largs -lgmem
16844 Then we execute the program as usual:
16851 Then @code{gnatmem} is invoked simply with
16857 which produces the following output (result may vary on different platforms):
16862 Total number of allocations : 18
16863 Total number of deallocations : 5
16864 Final Water Mark (non freed mem) : 53.00 Kilobytes
16865 High Water Mark : 56.90 Kilobytes
16867 Allocation Root # 1
16868 -------------------
16869 Number of non freed allocations : 11
16870 Final Water Mark (non freed mem) : 42.97 Kilobytes
16871 High Water Mark : 46.88 Kilobytes
16873 test_gm.adb:11 test_gm.my_alloc
16875 Allocation Root # 2
16876 -------------------
16877 Number of non freed allocations : 1
16878 Final Water Mark (non freed mem) : 10.02 Kilobytes
16879 High Water Mark : 10.02 Kilobytes
16881 s-secsta.adb:81 system.secondary_stack.ss_init
16883 Allocation Root # 3
16884 -------------------
16885 Number of non freed allocations : 1
16886 Final Water Mark (non freed mem) : 12 Bytes
16887 High Water Mark : 12 Bytes
16889 s-secsta.adb:181 system.secondary_stack.ss_init
16893 Note that the GNAT run time contains itself a certain number of
16894 allocations that have no corresponding deallocation,
16895 as shown here for root #2 and root
16896 #3. This is a normal behavior when the number of non-freed allocations
16897 is one, it allocates dynamic data structures that the run time needs for
16898 the complete lifetime of the program. Note also that there is only one
16899 allocation root in the user program with a single line back trace:
16900 test_gm.adb:11 test_gm.my_alloc, whereas a careful analysis of the
16901 program shows that 'My_Alloc' is called at 2 different points in the
16902 source (line 21 and line 24). If those two allocation roots need to be
16903 distinguished, the backtrace depth parameter can be used:
16906 $ gnatmem 3 test_gm
16910 which will give the following output:
16915 Total number of allocations : 18
16916 Total number of deallocations : 5
16917 Final Water Mark (non freed mem) : 53.00 Kilobytes
16918 High Water Mark : 56.90 Kilobytes
16920 Allocation Root # 1
16921 -------------------
16922 Number of non freed allocations : 10
16923 Final Water Mark (non freed mem) : 39.06 Kilobytes
16924 High Water Mark : 42.97 Kilobytes
16926 test_gm.adb:11 test_gm.my_alloc
16927 test_gm.adb:24 test_gm
16928 b_test_gm.c:52 main
16930 Allocation Root # 2
16931 -------------------
16932 Number of non freed allocations : 1
16933 Final Water Mark (non freed mem) : 10.02 Kilobytes
16934 High Water Mark : 10.02 Kilobytes
16936 s-secsta.adb:81 system.secondary_stack.ss_init
16937 s-secsta.adb:283 <system__secondary_stack___elabb>
16938 b_test_gm.c:33 adainit
16940 Allocation Root # 3
16941 -------------------
16942 Number of non freed allocations : 1
16943 Final Water Mark (non freed mem) : 3.91 Kilobytes
16944 High Water Mark : 3.91 Kilobytes
16946 test_gm.adb:11 test_gm.my_alloc
16947 test_gm.adb:21 test_gm
16948 b_test_gm.c:52 main
16950 Allocation Root # 4
16951 -------------------
16952 Number of non freed allocations : 1
16953 Final Water Mark (non freed mem) : 12 Bytes
16954 High Water Mark : 12 Bytes
16956 s-secsta.adb:181 system.secondary_stack.ss_init
16957 s-secsta.adb:283 <system__secondary_stack___elabb>
16958 b_test_gm.c:33 adainit
16962 The allocation root #1 of the first example has been split in 2 roots #1
16963 and #3 thanks to the more precise associated backtrace.
16967 @node Stack Related Facilities
16968 @chapter Stack Related Facilities
16971 This chapter describes some useful tools associated with stack
16972 checking and analysis. In
16973 particular, it deals with dynamic and static stack usage measurements.
16976 * Stack Overflow Checking::
16977 * Static Stack Usage Analysis::
16978 * Dynamic Stack Usage Analysis::
16981 @node Stack Overflow Checking
16982 @section Stack Overflow Checking
16983 @cindex Stack Overflow Checking
16984 @cindex -fstack-check
16987 For most operating systems, @command{gcc} does not perform stack overflow
16988 checking by default. This means that if the main environment task or
16989 some other task exceeds the available stack space, then unpredictable
16990 behavior will occur. Most native systems offer some level of protection by
16991 adding a guard page at the end of each task stack. This mechanism is usually
16992 not enough for dealing properly with stack overflow situations because
16993 a large local variable could ``jump'' above the guard page.
16994 Furthermore, when the
16995 guard page is hit, there may not be any space left on the stack for executing
16996 the exception propagation code. Enabling stack checking avoids
16999 To activate stack checking, compile all units with the gcc option
17000 @option{-fstack-check}. For example:
17003 gcc -c -fstack-check package1.adb
17007 Units compiled with this option will generate extra instructions to check
17008 that any use of the stack (for procedure calls or for declaring local
17009 variables in declare blocks) does not exceed the available stack space.
17010 If the space is exceeded, then a @code{Storage_Error} exception is raised.
17012 For declared tasks, the stack size is controlled by the size
17013 given in an applicable @code{Storage_Size} pragma or by the value specified
17014 at bind time with @option{-d} (@pxref{Switches for gnatbind}) or is set to
17015 the default size as defined in the GNAT runtime otherwise.
17017 For the environment task, the stack size depends on
17018 system defaults and is unknown to the compiler. Stack checking
17019 may still work correctly if a fixed
17020 size stack is allocated, but this cannot be guaranteed.
17022 To ensure that a clean exception is signalled for stack
17023 overflow, set the environment variable
17024 @env{GNAT_STACK_LIMIT} to indicate the maximum
17025 stack area that can be used, as in:
17026 @cindex GNAT_STACK_LIMIT
17029 SET GNAT_STACK_LIMIT 1600
17033 The limit is given in kilobytes, so the above declaration would
17034 set the stack limit of the environment task to 1.6 megabytes.
17035 Note that the only purpose of this usage is to limit the amount
17036 of stack used by the environment task. If it is necessary to
17037 increase the amount of stack for the environment task, then this
17038 is an operating systems issue, and must be addressed with the
17039 appropriate operating systems commands.
17042 To have a fixed size stack in the environment task, the stack must be put
17043 in the P0 address space and its size specified. Use these switches to
17047 gnatmake my_progs -largs "-Wl,--opt=STACK=4000,/p0image"
17051 The quotes are required to keep case. The number after @samp{STACK=} is the
17052 size of the environmental task stack in pagelets (512 bytes). In this example
17053 the stack size is about 2 megabytes.
17056 A consequence of the @option{/p0image} qualifier is also to makes RMS buffers
17057 be placed in P0 space. Refer to @cite{HP OpenVMS Linker Utility Manual} for
17058 more details about the @option{/p0image} qualifier and the @option{stack}
17062 @node Static Stack Usage Analysis
17063 @section Static Stack Usage Analysis
17064 @cindex Static Stack Usage Analysis
17065 @cindex -fstack-usage
17068 A unit compiled with @option{-fstack-usage} will generate an extra file
17070 the maximum amount of stack used, on a per-function basis.
17071 The file has the same
17072 basename as the target object file with a @file{.su} extension.
17073 Each line of this file is made up of three fields:
17077 The name of the function.
17081 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
17084 The second field corresponds to the size of the known part of the function
17087 The qualifier @code{static} means that the function frame size
17089 It usually means that all local variables have a static size.
17090 In this case, the second field is a reliable measure of the function stack
17093 The qualifier @code{dynamic} means that the function frame size is not static.
17094 It happens mainly when some local variables have a dynamic size. When this
17095 qualifier appears alone, the second field is not a reliable measure
17096 of the function stack analysis. When it is qualified with @code{bounded}, it
17097 means that the second field is a reliable maximum of the function stack
17100 @node Dynamic Stack Usage Analysis
17101 @section Dynamic Stack Usage Analysis
17104 It is possible to measure the maximum amount of stack used by a task, by
17105 adding a switch to @command{gnatbind}, as:
17108 $ gnatbind -u0 file
17112 With this option, at each task termination, its stack usage is output on
17114 It is not always convenient to output the stack usage when the program
17115 is still running. Hence, it is possible to delay this output until program
17116 termination. for a given number of tasks specified as the argument of the
17117 @option{-u} option. For instance:
17120 $ gnatbind -u100 file
17124 will buffer the stack usage information of the first 100 tasks to terminate and
17125 output this info at program termination. Results are displayed in four
17129 Index | Task Name | Stack Size | Stack Usage [Value +/- Variation]
17136 is a number associated with each task.
17139 is the name of the task analyzed.
17142 is the maximum size for the stack.
17145 is the measure done by the stack analyzer. In order to prevent overflow, the stack
17146 is not entirely analyzed, and it's not possible to know exactly how
17147 much has actually been used. The report thus contains the theoretical stack usage
17148 (Value) and the possible variation (Variation) around this value.
17153 The environment task stack, e.g., the stack that contains the main unit, is
17154 only processed when the environment variable GNAT_STACK_LIMIT is set.
17157 @c *********************************
17159 @c *********************************
17160 @node Verifying Properties Using gnatcheck
17161 @chapter Verifying Properties Using @command{gnatcheck}
17163 @cindex @command{gnatcheck}
17166 The @command{gnatcheck} tool is an ASIS-based utility that checks properties
17167 of Ada source files according to a given set of semantic rules.
17170 In order to check compliance with a given rule, @command{gnatcheck} has to
17171 semantically analyze the Ada sources.
17172 Therefore, checks can only be performed on
17173 legal Ada units. Moreover, when a unit depends semantically upon units located
17174 outside the current directory, the source search path has to be provided when
17175 calling @command{gnatcheck}, either through a specified project file or
17176 through @command{gnatcheck} switches as described below.
17178 A number of rules are predefined in @command{gnatcheck} and are described
17179 later in this chapter.
17180 You can also add new rules, by modifying the @command{gnatcheck} code and
17181 rebuilding the tool. In order to add a simple rule making some local checks,
17182 a small amount of straightforward ASIS-based programming is usually needed.
17184 Project support for @command{gnatcheck} is provided by the GNAT
17185 driver (see @ref{The GNAT Driver and Project Files}).
17187 Invoking @command{gnatcheck} on the command line has the form:
17190 @c $ gnatcheck @ovar{switches} @{@var{filename}@}
17191 @c @r{[}^-files^/FILES^=@{@var{arg_list_filename}@}@r{]}
17192 @c @r{[}-cargs @var{gcc_switches}@r{]} -rules @var{rule_options}
17193 @c Expanding @ovar macro inline (explanation in macro def comments)
17194 $ gnatcheck @r{[}@var{switches}@r{]} @{@var{filename}@}
17195 @r{[}^-files^/FILES^=@{@var{arg_list_filename}@}@r{]}
17196 @r{[}-cargs @var{gcc_switches}@r{]} -rules @var{rule_options}
17203 @var{switches} specify the general tool options
17206 Each @var{filename} is the name (including the extension) of a source
17207 file to process. ``Wildcards'' are allowed, and
17208 the file name may contain path information.
17211 Each @var{arg_list_filename} is the name (including the extension) of a text
17212 file containing the names of the source files to process, separated by spaces
17216 @var{gcc_switches} is a list of switches for
17217 @command{gcc}. They will be passed on to all compiler invocations made by
17218 @command{gnatcheck} to generate the ASIS trees. Here you can provide
17219 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
17220 and use the @option{-gnatec} switch to set the configuration file.
17223 @var{rule_options} is a list of options for controlling a set of
17224 rules to be checked by @command{gnatcheck} (@pxref{gnatcheck Rule Options}).
17228 Either a @file{@var{filename}} or an @file{@var{arg_list_filename}} must be
17232 * Format of the Report File::
17233 * General gnatcheck Switches::
17234 * gnatcheck Rule Options::
17235 * Adding the Results of Compiler Checks to gnatcheck Output::
17236 * Project-Wide Checks::
17238 * Predefined Rules::
17239 * Example of gnatcheck Usage::
17242 @node Format of the Report File
17243 @section Format of the Report File
17244 @cindex Report file (for @code{gnatcheck})
17247 The @command{gnatcheck} tool outputs on @file{stdout} all messages concerning
17249 It also creates a text file that
17250 contains the complete report of the last gnatcheck run. By default this file
17251 is named named @file{^gnatcheck.out^GNATCHECK.OUT^} and it is located in the
17252 current directory; the @option{^-o^/OUTPUT^} option can be used to change the
17253 name and/or location of the report file. This report contains:
17255 @item date and time of @command{gnatcheck} run, the version of
17256 the tool that has generated this report and the full parameters
17257 of the @command{gnatcheck} invocation;
17258 @item list of enabled rules;
17259 @item total number of detected violations;
17260 @item list of source files where rule violations have been detected;
17261 @item list of source files where no violations have been detected.
17264 @node General gnatcheck Switches
17265 @section General @command{gnatcheck} Switches
17268 The following switches control the general @command{gnatcheck} behavior
17272 @cindex @option{^-a^/ALL^} (@command{gnatcheck})
17274 Process all units including those with read-only ALI files such as
17275 those from the GNAT Run-Time library.
17279 @cindex @option{-d} (@command{gnatcheck})
17284 @cindex @option{-dd} (@command{gnatcheck})
17286 Progress indicator mode (for use in GPS).
17289 @cindex @option{^-h^/HELP^} (@command{gnatcheck})
17291 List the predefined and user-defined rules. For more details see
17292 @ref{Predefined Rules}.
17294 @cindex @option{^-l^/LOCS^} (@command{gnatcheck})
17296 Use full source locations references in the report file. For a construct from
17297 a generic instantiation a full source location is a chain from the location
17298 of this construct in the generic unit to the place where this unit is
17301 @cindex @option{^-log^/LOG^} (@command{gnatcheck})
17303 Duplicate all the output sent to @file{stderr} into a log file. The log file
17304 is named @file{gnatcheck.log} and is located in the current directory.
17306 @cindex @option{^-m^/DIAGNOSTIC_LIMIT^} (@command{gnatcheck})
17307 @item ^-m@i{nnnn}^/DIAGNOSTIC_LIMIT=@i{nnnn}^
17308 Maximum number of diagnostics to be sent to @file{stdout}, where @i{nnnn} is in
17309 the range 0@dots{}1000;
17310 the default value is 500. Zero means that there is no limitation on
17311 the number of diagnostic messages to be output.
17313 @cindex @option{^-q^/QUIET^} (@command{gnatcheck})
17315 Quiet mode. All the diagnostics about rule violations are placed in the
17316 @command{gnatcheck} report file only, without duplication on @file{stdout}.
17318 @cindex @option{^-s^/SHORT^} (@command{gnatcheck})
17320 Short format of the report file (no version information, no list of applied
17321 rules, no list of checked sources is included)
17323 @cindex @option{^--include-file=@var{file}^/INCLUDE_FILE=@var{file}^} (@command{gnatcheck})
17324 @item ^--include-file^/INCLUDE_FILE^
17325 Append the content of the specified text file to the report file
17327 @cindex @option{^-t^/TIME^} (@command{gnatcheck})
17329 Print out execution time.
17331 @cindex @option{^-v^/VERBOSE^} (@command{gnatcheck})
17332 @item ^-v^/VERBOSE^
17333 Verbose mode; @command{gnatcheck} generates version information and then
17334 a trace of sources being processed.
17336 @cindex @option{^-o ^/OUTPUT^} (@command{gnatcheck})
17337 @item ^-o ^/OUTPUT=^@var{report_file}
17338 Set name of report file file to @var{report_file} .
17342 @node gnatcheck Rule Options
17343 @section @command{gnatcheck} Rule Options
17346 The following options control the processing performed by
17347 @command{gnatcheck}.
17350 @cindex @option{+ALL} (@command{gnatcheck})
17352 Turn all the rule checks ON.
17354 @cindex @option{-ALL} (@command{gnatcheck})
17356 Turn all the rule checks OFF.
17358 @cindex @option{+R} (@command{gnatcheck})
17359 @item +R@var{rule_id}@r{[}:@var{param}@r{]}
17360 Turn on the check for a specified rule with the specified parameter, if any.
17361 @var{rule_id} must be the identifier of one of the currently implemented rules
17362 (use @option{^-h^/HELP^} for the list of implemented rules). Rule identifiers
17363 are not case-sensitive. The @var{param} item must
17364 be a string representing a valid parameter(s) for the specified rule.
17365 If it contains any space characters then this string must be enclosed in
17368 @cindex @option{-R} (@command{gnatcheck})
17369 @item -R@var{rule_id}@r{[}:@var{param}@r{]}
17370 Turn off the check for a specified rule with the specified parameter, if any.
17372 @cindex @option{-from} (@command{gnatcheck})
17373 @item -from=@var{rule_option_filename}
17374 Read the rule options from the text file @var{rule_option_filename}, referred
17375 to as a ``coding standard file'' below.
17380 The default behavior is that all the rule checks are disabled.
17382 A coding standard file is a text file that contains a set of rule options
17384 @cindex Coding standard file (for @code{gnatcheck})
17385 The file may contain empty lines and Ada-style comments (comment
17386 lines and end-of-line comments). There can be several rule options on a
17387 single line (separated by a space).
17389 A coding standard file may reference other coding standard files by including
17390 more @option{-from=@var{rule_option_filename}}
17391 options, each such option being replaced with the content of the
17392 corresponding coding standard file during processing. In case a
17393 cycle is detected (that is, @file{@var{rule_file_1}} reads rule options
17394 from @file{@var{rule_file_2}}, and @file{@var{rule_file_2}} reads
17395 (directly or indirectly) rule options from @file{@var{rule_file_1}}),
17396 processing fails with an error message.
17399 @node Adding the Results of Compiler Checks to gnatcheck Output
17400 @section Adding the Results of Compiler Checks to @command{gnatcheck} Output
17403 The @command{gnatcheck} tool can include in the generated diagnostic messages
17405 the report file the results of the checks performed by the compiler. Though
17406 disabled by default, this effect may be obtained by using @option{+R} with
17407 the following rule identifiers and parameters:
17411 To record restrictions violations (which are performed by the compiler if the
17412 pragma @code{Restrictions} or @code{Restriction_Warnings} are given),
17413 use the @code{Restrictions} rule
17414 with the same parameters as pragma
17415 @code{Restrictions} or @code{Restriction_Warnings}.
17418 To record compiler style checks (@pxref{Style Checking}), use the
17419 @code{Style_Checks} rule.
17420 This rule takes a parameter in one of the following forms:
17424 which enables the standard style checks corresponding to the @option{-gnatyy}
17425 GNAT style check option, or
17428 a string with the same
17429 structure and semantics as the @code{string_LITERAL} parameter of the
17430 GNAT pragma @code{Style_Checks}
17431 (for further information about this pragma,
17432 @pxref{Pragma Style_Checks,,, gnat_rm, GNAT Reference Manual}).
17437 @code{+RStyle_Checks:O} rule option activates
17438 the compiler style check that corresponds to
17439 @code{-gnatyO} style check option.
17442 To record compiler warnings (@pxref{Warning Message Control}), use the
17443 @code{Warnings} rule with a parameter that is a valid
17444 @i{static_string_expression} argument of the GNAT pragma @code{Warnings}
17445 (for further information about this pragma,
17446 @pxref{Pragma Warnings,,,gnat_rm, GNAT Reference Manual}).
17447 Note that in case of gnatcheck
17448 's' parameter, that corresponds to the GNAT @option{-gnatws} option, disables
17449 all the specific warnings, but not suppresses the warning mode,
17450 and 'e' parameter, corresponding to @option{-gnatwe} that means
17451 "treat warnings as errors", does not have any effect.
17455 To disable a specific restriction check, use @code{-RStyle_Checks} gnatcheck
17456 option with the corresponding restriction name as a parameter. @code{-R} is
17457 not available for @code{Style_Checks} and @code{Warnings} options, to disable
17458 warnings and style checks, use the corresponding warning and style options.
17460 @node Project-Wide Checks
17461 @section Project-Wide Checks
17462 @cindex Project-wide checks (for @command{gnatcheck})
17465 In order to perform checks on all units of a given project, you can use
17466 the GNAT driver along with the @option{-P} option:
17468 gnat check -Pproj -rules -from=my_rules
17472 If the project @code{proj} depends upon other projects, you can perform
17473 checks on the project closure using the @option{-U} option:
17475 gnat check -Pproj -U -rules -from=my_rules
17479 Finally, if not all the units are relevant to a particular main
17480 program in the project closure, you can perform checks for the set
17481 of units needed to create a given main program (unit closure) using
17482 the @option{-U} option followed by the name of the main unit:
17484 gnat check -Pproj -U main -rules -from=my_rules
17488 @node Rule exemption
17489 @section Rule exemption
17490 @cindex Rule exemption (for @command{gnatcheck})
17493 One of the most useful applications of @command{gnatcheck} is to
17494 automate the enforcement of project-specific coding standards,
17495 for example in safety-critical systems where particular features
17496 must be restricted in order to simplify the certification effort.
17497 However, it may sometimes be appropriate to violate a coding standard rule,
17498 and in such cases the rationale for the violation should be provided
17499 in the source program itself so that the individuals
17500 reviewing or maintaining the program can immediately understand the intent.
17502 The @command{gnatcheck} tool supports this practice with the notion of
17503 a ``rule exemption'' covering a specific source code section. Normally
17504 rule violation messages are issued both on @file{stderr}
17505 and in a report file. In contrast, exempted violations are not listed on
17506 @file{stderr}; thus users invoking @command{gnatcheck} interactively
17507 (e.g. in its GPS interface) do not need to pay attention to known and
17508 justified violations. However, exempted violations along with their
17509 justification are documented in a special section of the report file that
17510 @command{gnatcheck} generates.
17513 * Using pragma Annotate to Control Rule Exemption::
17514 * gnatcheck Annotations Rules::
17517 @node Using pragma Annotate to Control Rule Exemption
17518 @subsection Using pragma @code{Annotate} to Control Rule Exemption
17519 @cindex Using pragma Annotate to control rule exemption
17522 Rule exemption is controlled by pragma @code{Annotate} when its first
17523 argument is ``gnatcheck''. The syntax of @command{gnatcheck}'s
17524 exemption control annotations is as follows:
17526 @smallexample @c ada
17528 pragma Annotate (gnatcheck, @i{exemption_control}, @i{Rule_Name}, [@i{justification}]);
17530 @i{exemption_control} ::= Exempt_On | Exempt_Off
17532 @i{Rule_Name} ::= string_literal
17534 @i{justification} ::= string_literal
17539 When a @command{gnatcheck} annotation has more then four arguments,
17540 @command{gnatcheck} issues a warning and ignores the additional arguments.
17541 If the additional arguments do not follow the syntax above,
17542 @command{gnatcheck} emits a warning and ignores the annotation.
17544 The @i{@code{Rule_Name}} argument should be the name of some existing
17545 @command{gnatcheck} rule.
17546 Otherwise a warning message is generated and the pragma is
17547 ignored. If @code{Rule_Name} denotes a rule that is not activated by the given
17548 @command{gnatcheck} call, the pragma is ignored and no warning is issued.
17550 A source code section where an exemption is active for a given rule is
17551 delimited by an @code{exempt_on} and @code{exempt_off} annotation pair:
17553 @smallexample @c ada
17554 pragma Annotate (gnatcheck, Exempt_On, Rule_Name, "justification");
17555 -- source code section
17556 pragma Annotate (gnatcheck, Exempt_Off, Rule_Name);
17560 @node gnatcheck Annotations Rules
17561 @subsection @command{gnatcheck} Annotations Rules
17562 @cindex @command{gnatcheck} annotations rules
17567 An ``Exempt_Off'' annotation can only appear after a corresponding
17568 ``Exempt_On'' annotation.
17571 Exempted source code sections are only based on the source location of the
17572 annotations. Any source construct between the two
17573 annotations is part of the exempted source code section.
17576 Exempted source code sections for different rules are independent. They can
17577 be nested or intersect with one another without limitation.
17578 Creating nested or intersecting source code sections for the same rule is
17582 Malformed exempted source code sections are reported by a warning, and
17583 the corresponding rule exemptions are ignored.
17586 When an exempted source code section does not contain at least one violation
17587 of the exempted rule, a warning is emitted on @file{stderr}.
17590 If an ``Exempt_On'' annotation pragma does not have a matching
17591 ``Exempt_Off'' annotation pragma in the same compilation unit, then the
17592 exemption for the given rule is ignored and a warning is issued.
17596 @node Predefined Rules
17597 @section Predefined Rules
17598 @cindex Predefined rules (for @command{gnatcheck})
17601 @c (Jan 2007) Since the global rules are still under development and are not
17602 @c documented, there is no point in explaining the difference between
17603 @c global and local rules
17605 A rule in @command{gnatcheck} is either local or global.
17606 A @emph{local rule} is a rule that applies to a well-defined section
17607 of a program and that can be checked by analyzing only this section.
17608 A @emph{global rule} requires analysis of some global properties of the
17609 whole program (mostly related to the program call graph).
17610 As of @value{NOW}, the implementation of global rules should be
17611 considered to be at a preliminary stage. You can use the
17612 @option{+GLOBAL} option to enable all the global rules, and the
17613 @option{-GLOBAL} rule option to disable all the global rules.
17615 All the global rules in the list below are
17616 so indicated by marking them ``GLOBAL''.
17617 This +GLOBAL and -GLOBAL options are not
17618 included in the list of gnatcheck options above, because at the moment they
17619 are considered as a temporary debug options.
17621 @command{gnatcheck} performs rule checks for generic
17622 instances only for global rules. This limitation may be relaxed in a later
17627 The predefined rules implemented in @command{gnatcheck}
17628 are described in a companion document,
17629 @cite{GNATcheck Reference Manual -- Predefined Rules}.
17630 The rule identifier is
17631 used as a parameter of @command{gnatcheck}'s @option{+R} or @option{-R}
17635 @node Example of gnatcheck Usage
17636 @section Example of @command{gnatcheck} Usage
17639 Here is a simple example. Suppose that in the current directory we have a
17640 project file named @file{gnatcheck_example.gpr} with the following content:
17642 @smallexample @c projectfile
17643 project Gnatcheck_Example is
17645 for Source_Dirs use ("src");
17646 for Object_Dir use "obj";
17647 for Main use ("main.adb");
17650 for Default_Switches ("ada") use ("-rules", "-from=coding_standard");
17653 end Gnatcheck_Example;
17657 And the file named @file{coding_standard} is also located in the current
17658 directory and has the following content:
17661 -----------------------------------------------------
17662 -- This is a sample gnatcheck coding standard file --
17663 -----------------------------------------------------
17665 -- First, turning on rules, that are directly implemented in gnatcheck
17666 +RAbstract_Type_Declarations
17669 +RFloat_Equality_Checks
17670 +REXIT_Statements_With_No_Loop_Name
17672 -- Then, activating compiler checks of interest:
17674 -- This style check checks if a unit name is present on END keyword that
17675 -- is the end of the unit declaration
17679 And the subdirectory @file{src} contains the following Ada sources:
17683 @smallexample @c ada
17685 type T is abstract tagged private;
17686 procedure P (X : T) is abstract;
17689 type My_Float is digits 8;
17690 function Is_Equal (L, R : My_Float) return Boolean;
17693 type T is abstract tagged null record;
17700 @smallexample @c ada
17701 package body Pack is
17702 package body Inner is
17703 function Is_Equal (L, R : My_Float) return Boolean is
17712 and @file{main.adb}
17714 @smallexample @c ada
17715 with Pack; use Pack;
17719 (gnatcheck, Exempt_On, "Anonymous_Arrays", "this one is fine");
17720 Float_Array : array (1 .. 10) of Inner.My_Float;
17721 pragma Annotate (gnatcheck, Exempt_Off, "Anonymous_Arrays");
17723 Another_Float_Array : array (1 .. 10) of Inner.My_Float;
17727 B : Boolean := False;
17730 for J in Float_Array'Range loop
17731 if Is_Equal (Float_Array (J), Another_Float_Array (J)) then
17740 And suppose we call @command{gnatcheck} from the current directory using
17741 the @command{gnat} driver:
17744 gnat check -Pgnatcheck_example.gpr
17748 As a result, @command{gnatcheck} is called to check all the files from the
17749 project @file{gnatcheck_example.gpr} using the coding standard defined by
17750 the file @file{coding_standard}. As the result, the @command{gnatcheck}
17751 report file named @file{gnatcheck.out} will be created in the current
17752 directory, and it will have the following content:
17755 RULE CHECKING REPORT
17759 Date and time of execution: 2009.10.28 14:17
17760 Tool version: GNATCHECK (built with ASIS 2.0.R for GNAT Pro 6.3.0w (20091016))
17763 gnatcheck -files=.../GNAT-TEMP-000004.TMP -cargs -gnatec=.../GNAT-TEMP-000003.TMP -rules -from=coding_standard
17765 Coding standard (applied rules):
17766 Abstract_Type_Declarations
17768 EXIT_Statements_With_No_Loop_Name
17769 Float_Equality_Checks
17772 Compiler style checks: -gnatye
17774 Number of coding standard violations: 6
17775 Number of exempted coding standard violations: 1
17777 2. DETECTED RULE VIOLATIONS
17779 2.1. NON-EXEMPTED VIOLATIONS
17781 Source files with non-exempted violations
17786 List of violations grouped by files, and ordered by increasing source location:
17788 pack.ads:2:4: declaration of abstract type
17789 pack.ads:5:4: declaration of local package
17790 pack.ads:10:30: declaration of abstract type
17791 pack.ads:11:1: (style) "end Pack" required
17792 pack.adb:5:19: use of equality operation for float values
17793 pack.adb:6:7: (style) "end Is_Equal" required
17794 main.adb:9:26: anonymous array type
17795 main.adb:19:10: exit statement with no loop name
17797 2.2. EXEMPTED VIOLATIONS
17799 Source files with exempted violations
17802 List of violations grouped by files, and ordered by increasing source location:
17804 main.adb:6:18: anonymous array type
17807 2.3. SOURCE FILES WITH NO VIOLATION
17809 No files without violations
17815 @c *********************************
17816 @node Creating Sample Bodies Using gnatstub
17817 @chapter Creating Sample Bodies Using @command{gnatstub}
17821 @command{gnatstub} creates body stubs, that is, empty but compilable bodies
17822 for library unit declarations.
17824 Note: to invoke @code{gnatstub} with a project file, use the @code{gnat}
17825 driver (see @ref{The GNAT Driver and Project Files}).
17827 To create a body stub, @command{gnatstub} has to compile the library
17828 unit declaration. Therefore, bodies can be created only for legal
17829 library units. Moreover, if a library unit depends semantically upon
17830 units located outside the current directory, you have to provide
17831 the source search path when calling @command{gnatstub}, see the description
17832 of @command{gnatstub} switches below.
17834 By default, all the program unit body stubs generated by @code{gnatstub}
17835 raise the predefined @code{Program_Error} exception, which will catch
17836 accidental calls of generated stubs. This behavior can be changed with
17837 option @option{^--no-exception^/NO_EXCEPTION^} (see below).
17840 * Running gnatstub::
17841 * Switches for gnatstub::
17844 @node Running gnatstub
17845 @section Running @command{gnatstub}
17848 @command{gnatstub} has the command-line interface of the form
17851 @c $ gnatstub @ovar{switches} @var{filename} @ovar{directory}
17852 @c Expanding @ovar macro inline (explanation in macro def comments)
17853 $ gnatstub @r{[}@var{switches}@r{]} @var{filename} @r{[}@var{directory}@r{]} @r{[}-cargs @var{gcc_switches}@r{]}
17860 is the name of the source file that contains a library unit declaration
17861 for which a body must be created. The file name may contain the path
17863 The file name does not have to follow the GNAT file name conventions. If the
17865 does not follow GNAT file naming conventions, the name of the body file must
17867 explicitly as the value of the @option{^-o^/BODY=^@var{body-name}} option.
17868 If the file name follows the GNAT file naming
17869 conventions and the name of the body file is not provided,
17872 of the body file from the argument file name by replacing the @file{.ads}
17874 with the @file{.adb} suffix.
17877 indicates the directory in which the body stub is to be placed (the default
17881 @item @samp{@var{gcc_switches}} is a list of switches for
17882 @command{gcc}. They will be passed on to all compiler invocations made by
17883 @command{gnatelim} to generate the ASIS trees. Here you can provide
17884 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
17885 use the @option{-gnatec} switch to set the configuration file etc.
17888 is an optional sequence of switches as described in the next section
17891 @node Switches for gnatstub
17892 @section Switches for @command{gnatstub}
17898 @cindex @option{^-f^/FULL^} (@command{gnatstub})
17899 If the destination directory already contains a file with the name of the
17901 for the argument spec file, replace it with the generated body stub.
17903 @item ^-hs^/HEADER=SPEC^
17904 @cindex @option{^-hs^/HEADER=SPEC^} (@command{gnatstub})
17905 Put the comment header (i.e., all the comments preceding the
17906 compilation unit) from the source of the library unit declaration
17907 into the body stub.
17909 @item ^-hg^/HEADER=GENERAL^
17910 @cindex @option{^-hg^/HEADER=GENERAL^} (@command{gnatstub})
17911 Put a sample comment header into the body stub.
17913 @item ^--header-file=@var{filename}^/FROM_HEADER_FILE=@var{filename}^
17914 @cindex @option{^--header-file^/FROM_HEADER_FILE=^} (@command{gnatstub})
17915 Use the content of the file as the comment header for a generated body stub.
17919 @cindex @option{-IDIR} (@command{gnatstub})
17921 @cindex @option{-I-} (@command{gnatstub})
17924 @item /NOCURRENT_DIRECTORY
17925 @cindex @option{/NOCURRENT_DIRECTORY} (@command{gnatstub})
17927 ^These switches have ^This switch has^ the same meaning as in calls to
17929 ^They define ^It defines ^ the source search path in the call to
17930 @command{gcc} issued
17931 by @command{gnatstub} to compile an argument source file.
17933 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE=^@var{PATH}
17934 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatstub})
17935 This switch has the same meaning as in calls to @command{gcc}.
17936 It defines the additional configuration file to be passed to the call to
17937 @command{gcc} issued
17938 by @command{gnatstub} to compile an argument source file.
17940 @item ^-gnatyM^/MAX_LINE_LENGTH=^@var{n}
17941 @cindex @option{^-gnatyM^/MAX_LINE_LENGTH^} (@command{gnatstub})
17942 (@var{n} is a non-negative integer). Set the maximum line length in the
17943 body stub to @var{n}; the default is 79. The maximum value that can be
17944 specified is 32767. Note that in the special case of configuration
17945 pragma files, the maximum is always 32767 regardless of whether or
17946 not this switch appears.
17948 @item ^-gnaty^/STYLE_CHECKS=^@var{n}
17949 @cindex @option{^-gnaty^/STYLE_CHECKS=^} (@command{gnatstub})
17950 (@var{n} is a non-negative integer from 1 to 9). Set the indentation level in
17951 the generated body sample to @var{n}.
17952 The default indentation is 3.
17954 @item ^-gnatyo^/ORDERED_SUBPROGRAMS^
17955 @cindex @option{^-gnato^/ORDERED_SUBPROGRAMS^} (@command{gnatstub})
17956 Order local bodies alphabetically. (By default local bodies are ordered
17957 in the same way as the corresponding local specs in the argument spec file.)
17959 @item ^-i^/INDENTATION=^@var{n}
17960 @cindex @option{^-i^/INDENTATION^} (@command{gnatstub})
17961 Same as @option{^-gnaty^/STYLE_CHECKS=^@var{n}}
17963 @item ^-k^/TREE_FILE=SAVE^
17964 @cindex @option{^-k^/TREE_FILE=SAVE^} (@command{gnatstub})
17965 Do not remove the tree file (i.e., the snapshot of the compiler internal
17966 structures used by @command{gnatstub}) after creating the body stub.
17968 @item ^-l^/LINE_LENGTH=^@var{n}
17969 @cindex @option{^-l^/LINE_LENGTH^} (@command{gnatstub})
17970 Same as @option{^-gnatyM^/MAX_LINE_LENGTH=^@var{n}}
17972 @item ^--no-exception^/NO_EXCEPTION^
17973 @cindex @option{^--no-exception^/NO_EXCEPTION^} (@command{gnatstub})
17974 Avoind raising PROGRAM_ERROR in the generated bodies of program unit stubs.
17975 This is not always possible for function stubs.
17977 @item ^--no-local-header^/NO_LOCAL_HEADER^
17978 @cindex @option{^--no-local-header^/NO_LOCAL_HEADER^} (@command{gnatstub})
17979 Do not place local comment header with unit name before body stub for a
17982 @item ^-o ^/BODY=^@var{body-name}
17983 @cindex @option{^-o^/BODY^} (@command{gnatstub})
17984 Body file name. This should be set if the argument file name does not
17986 the GNAT file naming
17987 conventions. If this switch is omitted the default name for the body will be
17989 from the argument file name according to the GNAT file naming conventions.
17992 @cindex @option{^-q^/QUIET^} (@command{gnatstub})
17993 Quiet mode: do not generate a confirmation when a body is
17994 successfully created, and do not generate a message when a body is not
17998 @item ^-r^/TREE_FILE=REUSE^
17999 @cindex @option{^-r^/TREE_FILE=REUSE^} (@command{gnatstub})
18000 Reuse the tree file (if it exists) instead of creating it. Instead of
18001 creating the tree file for the library unit declaration, @command{gnatstub}
18002 tries to find it in the current directory and use it for creating
18003 a body. If the tree file is not found, no body is created. This option
18004 also implies @option{^-k^/SAVE^}, whether or not
18005 the latter is set explicitly.
18007 @item ^-t^/TREE_FILE=OVERWRITE^
18008 @cindex @option{^-t^/TREE_FILE=OVERWRITE^} (@command{gnatstub})
18009 Overwrite the existing tree file. If the current directory already
18010 contains the file which, according to the GNAT file naming rules should
18011 be considered as a tree file for the argument source file,
18013 will refuse to create the tree file needed to create a sample body
18014 unless this option is set.
18016 @item ^-v^/VERBOSE^
18017 @cindex @option{^-v^/VERBOSE^} (@command{gnatstub})
18018 Verbose mode: generate version information.
18022 @c *********************************
18023 @node Generating Ada Bindings for C and C++ headers
18024 @chapter Generating Ada Bindings for C and C++ headers
18028 GNAT now comes with a binding generator for C and C++ headers which is
18029 intended to do 95% of the tedious work of generating Ada specs from C
18030 or C++ header files.
18032 Note that this capability is not intended to generate 100% correct Ada specs,
18033 and will is some cases require manual adjustments, although it can often
18034 be used out of the box in practice.
18036 Some of the known limitations include:
18039 @item only very simple character constant macros are translated into Ada
18040 constants. Function macros (macros with arguments) are partially translated
18041 as comments, to be completed manually if needed.
18042 @item some extensions (e.g. vector types) are not supported
18043 @item pointers to pointers or complex structures are mapped to System.Address
18046 The code generated is using the Ada 2005 syntax, which makes it
18047 easier to interface with other languages than previous versions of Ada.
18050 * Running the binding generator::
18051 * Generating bindings for C++ headers::
18055 @node Running the binding generator
18056 @section Running the binding generator
18059 The binding generator is part of the @command{gcc} compiler and can be
18060 invoked via the @option{-fdump-ada-spec} switch, which will generate Ada
18061 spec files for the header files specified on the command line, and all
18062 header files needed by these files transitivitely. For example:
18065 $ g++ -c -fdump-ada-spec -C /usr/include/time.h
18066 $ gcc -c -gnat05 *.ads
18069 will generate, under GNU/Linux, the following files: @file{time_h.ads},
18070 @file{bits_time_h.ads}, @file{stddef_h.ads}, @file{bits_types_h.ads} which
18071 correspond to the files @file{/usr/include/time.h},
18072 @file{/usr/include/bits/time.h}, etc@dots{}, and will then compile in Ada 2005
18073 mode these Ada specs.
18075 The @code{-C} switch tells @command{gcc} to extract comments from headers,
18076 and will attempt to generate corresponding Ada comments.
18078 If you want to generate a single Ada file and not the transitive closure, you
18079 can use instead the @option{-fdump-ada-spec-slim} switch.
18081 Note that we recommend when possible to use the @command{g++} driver to
18082 generate bindings, even for most C headers, since this will in general
18083 generate better Ada specs. For generating bindings for C++ headers, it is
18084 mandatory to use the @command{g++} command, or @command{gcc -x c++} which
18085 is equivalent in this case. If @command{g++} cannot work on your C headers
18086 because of incompatibilities between C and C++, then you can fallback to
18087 @command{gcc} instead.
18089 For an example of better bindings generated from the C++ front-end,
18090 the name of the parameters (when available) are actually ignored by the C
18091 front-end. Consider the following C header:
18094 extern void foo (int variable);
18097 with the C front-end, @code{variable} is ignored, and the above is handled as:
18100 extern void foo (int);
18103 generating a generic:
18106 procedure foo (param1 : int);
18109 with the C++ front-end, the name is available, and we generate:
18112 procedure foo (variable : int);
18115 In some cases, the generated bindings will be more complete or more meaningful
18116 when defining some macros, which you can do via the @option{-D} switch. This
18117 is for example the case with @file{Xlib.h} under GNU/Linux:
18120 g++ -c -fdump-ada-spec -DXLIB_ILLEGAL_ACCESS -C /usr/include/X11/Xlib.h
18123 The above will generate more complete bindings than a straight call without
18124 the @option{-DXLIB_ILLEGAL_ACCESS} switch.
18126 In other cases, it is not possible to parse a header file in a stand alone
18127 manner, because other include files need to be included first. In this
18128 case, the solution is to create a small header file including the needed
18129 @code{#include} and possible @code{#define} directives. For example, to
18130 generate Ada bindings for @file{readline/readline.h}, you need to first
18131 include @file{stdio.h}, so you can create a file with the following two
18132 lines in e.g. @file{readline1.h}:
18136 #include <readline/readline.h>
18139 and then generate Ada bindings from this file:
18142 $ g++ -c -fdump-ada-spec readline1.h
18145 @node Generating bindings for C++ headers
18146 @section Generating bindings for C++ headers
18149 Generating bindings for C++ headers is done using the same options, always
18150 with the @command{g++} compiler.
18152 In this mode, C++ classes will be mapped to Ada tagged types, constructors
18153 will be mapped using the @code{CPP_Constructor} pragma, and when possible,
18154 multiple inheritance of abstract classes will be mapped to Ada interfaces
18155 (@xref{Interfacing to C++,,,gnat_rm, GNAT Reference Manual}, for additional
18156 information on interfacing to C++).
18158 For example, given the following C++ header file:
18165 virtual int Number_Of_Teeth () = 0;
18170 virtual void Set_Owner (char* Name) = 0;
18176 virtual void Set_Age (int New_Age);
18179 class Dog : Animal, Carnivore, Domestic @{
18184 virtual int Number_Of_Teeth ();
18185 virtual void Set_Owner (char* Name);
18193 The corresponding Ada code is generated:
18195 @smallexample @c ada
18198 package Class_Carnivore is
18199 type Carnivore is limited interface;
18200 pragma Import (CPP, Carnivore);
18202 function Number_Of_Teeth (this : access Carnivore) return int is abstract;
18204 use Class_Carnivore;
18206 package Class_Domestic is
18207 type Domestic is limited interface;
18208 pragma Import (CPP, Domestic);
18210 procedure Set_Owner
18211 (this : access Domestic;
18212 Name : Interfaces.C.Strings.chars_ptr) is abstract;
18214 use Class_Domestic;
18216 package Class_Animal is
18217 type Animal is tagged limited record
18218 Age_Count : aliased int;
18220 pragma Import (CPP, Animal);
18222 procedure Set_Age (this : access Animal; New_Age : int);
18223 pragma Import (CPP, Set_Age, "_ZN6Animal7Set_AgeEi");
18227 package Class_Dog is
18228 type Dog is new Animal and Carnivore and Domestic with record
18229 Tooth_Count : aliased int;
18230 Owner : Interfaces.C.Strings.chars_ptr;
18232 pragma Import (CPP, Dog);
18234 function Number_Of_Teeth (this : access Dog) return int;
18235 pragma Import (CPP, Number_Of_Teeth, "_ZN3Dog15Number_Of_TeethEv");
18237 procedure Set_Owner
18238 (this : access Dog; Name : Interfaces.C.Strings.chars_ptr);
18239 pragma Import (CPP, Set_Owner, "_ZN3Dog9Set_OwnerEPc");
18241 function New_Dog return Dog;
18242 pragma CPP_Constructor (New_Dog);
18243 pragma Import (CPP, New_Dog, "_ZN3DogC1Ev");
18254 @item -fdump-ada-spec
18255 @cindex @option{-fdump-ada-spec} (@command{gcc})
18256 Generate Ada spec files for the given header files transitively (including
18257 all header files that these headers depend upon).
18259 @item -fdump-ada-spec-slim
18260 @cindex @option{-fdump-ada-spec-slim} (@command{gcc})
18261 Generate Ada spec files for the header files specified on the command line
18265 @cindex @option{-C} (@command{gcc})
18266 Extract comments from headers and generate Ada comments in the Ada spec files.
18269 @node Other Utility Programs
18270 @chapter Other Utility Programs
18273 This chapter discusses some other utility programs available in the Ada
18277 * Using Other Utility Programs with GNAT::
18278 * The External Symbol Naming Scheme of GNAT::
18279 * Converting Ada Files to html with gnathtml::
18280 * Installing gnathtml::
18287 @node Using Other Utility Programs with GNAT
18288 @section Using Other Utility Programs with GNAT
18291 The object files generated by GNAT are in standard system format and in
18292 particular the debugging information uses this format. This means
18293 programs generated by GNAT can be used with existing utilities that
18294 depend on these formats.
18297 In general, any utility program that works with C will also often work with
18298 Ada programs generated by GNAT. This includes software utilities such as
18299 gprof (a profiling program), @code{gdb} (the FSF debugger), and utilities such
18303 @node The External Symbol Naming Scheme of GNAT
18304 @section The External Symbol Naming Scheme of GNAT
18307 In order to interpret the output from GNAT, when using tools that are
18308 originally intended for use with other languages, it is useful to
18309 understand the conventions used to generate link names from the Ada
18312 All link names are in all lowercase letters. With the exception of library
18313 procedure names, the mechanism used is simply to use the full expanded
18314 Ada name with dots replaced by double underscores. For example, suppose
18315 we have the following package spec:
18317 @smallexample @c ada
18328 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
18329 the corresponding link name is @code{qrs__mn}.
18331 Of course if a @code{pragma Export} is used this may be overridden:
18333 @smallexample @c ada
18338 pragma Export (Var1, C, External_Name => "var1_name");
18340 pragma Export (Var2, C, Link_Name => "var2_link_name");
18347 In this case, the link name for @var{Var1} is whatever link name the
18348 C compiler would assign for the C function @var{var1_name}. This typically
18349 would be either @var{var1_name} or @var{_var1_name}, depending on operating
18350 system conventions, but other possibilities exist. The link name for
18351 @var{Var2} is @var{var2_link_name}, and this is not operating system
18355 One exception occurs for library level procedures. A potential ambiguity
18356 arises between the required name @code{_main} for the C main program,
18357 and the name we would otherwise assign to an Ada library level procedure
18358 called @code{Main} (which might well not be the main program).
18360 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
18361 names. So if we have a library level procedure such as
18363 @smallexample @c ada
18366 procedure Hello (S : String);
18372 the external name of this procedure will be @var{_ada_hello}.
18375 @node Converting Ada Files to html with gnathtml
18376 @section Converting Ada Files to HTML with @code{gnathtml}
18379 This @code{Perl} script allows Ada source files to be browsed using
18380 standard Web browsers. For installation procedure, see the section
18381 @xref{Installing gnathtml}.
18383 Ada reserved keywords are highlighted in a bold font and Ada comments in
18384 a blue font. Unless your program was compiled with the gcc @option{-gnatx}
18385 switch to suppress the generation of cross-referencing information, user
18386 defined variables and types will appear in a different color; you will
18387 be able to click on any identifier and go to its declaration.
18389 The command line is as follow:
18391 @c $ perl gnathtml.pl @ovar{^switches^options^} @var{ada-files}
18392 @c Expanding @ovar macro inline (explanation in macro def comments)
18393 $ perl gnathtml.pl @r{[}@var{^switches^options^}@r{]} @var{ada-files}
18397 You can pass it as many Ada files as you want. @code{gnathtml} will generate
18398 an html file for every ada file, and a global file called @file{index.htm}.
18399 This file is an index of every identifier defined in the files.
18401 The available ^switches^options^ are the following ones:
18405 @cindex @option{-83} (@code{gnathtml})
18406 Only the Ada 83 subset of keywords will be highlighted.
18408 @item -cc @var{color}
18409 @cindex @option{-cc} (@code{gnathtml})
18410 This option allows you to change the color used for comments. The default
18411 value is green. The color argument can be any name accepted by html.
18414 @cindex @option{-d} (@code{gnathtml})
18415 If the Ada files depend on some other files (for instance through
18416 @code{with} clauses, the latter files will also be converted to html.
18417 Only the files in the user project will be converted to html, not the files
18418 in the run-time library itself.
18421 @cindex @option{-D} (@code{gnathtml})
18422 This command is the same as @option{-d} above, but @command{gnathtml} will
18423 also look for files in the run-time library, and generate html files for them.
18425 @item -ext @var{extension}
18426 @cindex @option{-ext} (@code{gnathtml})
18427 This option allows you to change the extension of the generated HTML files.
18428 If you do not specify an extension, it will default to @file{htm}.
18431 @cindex @option{-f} (@code{gnathtml})
18432 By default, gnathtml will generate html links only for global entities
18433 ('with'ed units, global variables and types,@dots{}). If you specify
18434 @option{-f} on the command line, then links will be generated for local
18437 @item -l @var{number}
18438 @cindex @option{-l} (@code{gnathtml})
18439 If this ^switch^option^ is provided and @var{number} is not 0, then
18440 @code{gnathtml} will number the html files every @var{number} line.
18443 @cindex @option{-I} (@code{gnathtml})
18444 Specify a directory to search for library files (@file{.ALI} files) and
18445 source files. You can provide several -I switches on the command line,
18446 and the directories will be parsed in the order of the command line.
18449 @cindex @option{-o} (@code{gnathtml})
18450 Specify the output directory for html files. By default, gnathtml will
18451 saved the generated html files in a subdirectory named @file{html/}.
18453 @item -p @var{file}
18454 @cindex @option{-p} (@code{gnathtml})
18455 If you are using Emacs and the most recent Emacs Ada mode, which provides
18456 a full Integrated Development Environment for compiling, checking,
18457 running and debugging applications, you may use @file{.gpr} files
18458 to give the directories where Emacs can find sources and object files.
18460 Using this ^switch^option^, you can tell gnathtml to use these files.
18461 This allows you to get an html version of your application, even if it
18462 is spread over multiple directories.
18464 @item -sc @var{color}
18465 @cindex @option{-sc} (@code{gnathtml})
18466 This ^switch^option^ allows you to change the color used for symbol
18468 The default value is red. The color argument can be any name accepted by html.
18470 @item -t @var{file}
18471 @cindex @option{-t} (@code{gnathtml})
18472 This ^switch^option^ provides the name of a file. This file contains a list of
18473 file names to be converted, and the effect is exactly as though they had
18474 appeared explicitly on the command line. This
18475 is the recommended way to work around the command line length limit on some
18480 @node Installing gnathtml
18481 @section Installing @code{gnathtml}
18484 @code{Perl} needs to be installed on your machine to run this script.
18485 @code{Perl} is freely available for almost every architecture and
18486 Operating System via the Internet.
18488 On Unix systems, you may want to modify the first line of the script
18489 @code{gnathtml}, to explicitly tell the Operating system where Perl
18490 is. The syntax of this line is:
18492 #!full_path_name_to_perl
18496 Alternatively, you may run the script using the following command line:
18499 @c $ perl gnathtml.pl @ovar{switches} @var{files}
18500 @c Expanding @ovar macro inline (explanation in macro def comments)
18501 $ perl gnathtml.pl @r{[}@var{switches}@r{]} @var{files}
18510 The GNAT distribution provides an Ada 95 template for the HP Language
18511 Sensitive Editor (LSE), a component of DECset. In order to
18512 access it, invoke LSE with the qualifier /ENVIRONMENT=GNU:[LIB]ADA95.ENV.
18519 GNAT supports The HP Performance Coverage Analyzer (PCA), a component
18520 of DECset. To use it proceed as outlined under ``HELP PCA'', except for running
18521 the collection phase with the /DEBUG qualifier.
18524 $ GNAT MAKE /DEBUG <PROGRAM_NAME>
18525 $ DEFINE LIB$DEBUG PCA$COLLECTOR
18526 $ RUN/DEBUG <PROGRAM_NAME>
18532 @c ******************************
18533 @node Code Coverage and Profiling
18534 @chapter Code Coverage and Profiling
18535 @cindex Code Coverage
18539 This chapter describes how to use @code{gcov} - coverage testing tool - and
18540 @code{gprof} - profiler tool - on your Ada programs.
18543 * Code Coverage of Ada Programs using gcov::
18544 * Profiling an Ada Program using gprof::
18547 @node Code Coverage of Ada Programs using gcov
18548 @section Code Coverage of Ada Programs using gcov
18550 @cindex -fprofile-arcs
18551 @cindex -ftest-coverage
18553 @cindex Code Coverage
18556 @code{gcov} is a test coverage program: it analyzes the execution of a given
18557 program on selected tests, to help you determine the portions of the program
18558 that are still untested.
18560 @code{gcov} is part of the GCC suite, and is described in detail in the GCC
18561 User's Guide. You can refer to this documentation for a more complete
18564 This chapter provides a quick startup guide, and
18565 details some Gnat-specific features.
18568 * Quick startup guide::
18572 @node Quick startup guide
18573 @subsection Quick startup guide
18575 In order to perform coverage analysis of a program using @code{gcov}, 3
18580 Code instrumentation during the compilation process
18582 Execution of the instrumented program
18584 Execution of the @code{gcov} tool to generate the result.
18587 The code instrumentation needed by gcov is created at the object level:
18588 The source code is not modified in any way, because the instrumentation code is
18589 inserted by gcc during the compilation process. To compile your code with code
18590 coverage activated, you need to recompile your whole project using the
18592 @code{-fprofile-arcs} and @code{-ftest-coverage}, and link it using
18593 @code{-fprofile-arcs}.
18596 $ gnatmake -P my_project.gpr -f -cargs -fprofile-arcs -ftest-coverage \
18597 -largs -fprofile-arcs
18600 This compilation process will create @file{.gcno} files together with
18601 the usual object files.
18603 Once the program is compiled with coverage instrumentation, you can
18604 run it as many times as needed - on portions of a test suite for
18605 example. The first execution will produce @file{.gcda} files at the
18606 same location as the @file{.gcno} files. The following executions
18607 will update those files, so that a cumulative result of the covered
18608 portions of the program is generated.
18610 Finally, you need to call the @code{gcov} tool. The different options of
18611 @code{gcov} are available in the GCC User's Guide, section 'Invoking gcov'.
18613 This will create annotated source files with a @file{.gcov} extension:
18614 @file{my_main.adb} file will be analysed in @file{my_main.adb.gcov}.
18616 @node Gnat specifics
18617 @subsection Gnat specifics
18619 Because Ada semantics, portions of the source code may be shared among
18620 several object files. This is the case for example when generics are
18621 involved, when inlining is active or when declarations generate initialisation
18622 calls. In order to take
18623 into account this shared code, you need to call @code{gcov} on all
18624 source files of the tested program at once.
18626 The list of source files might exceed the system's maximum command line
18627 length. In order to bypass this limitation, a new mechanism has been
18628 implemented in @code{gcov}: you can now list all your project's files into a
18629 text file, and provide this file to gcov as a parameter, preceded by a @@
18630 (e.g. @samp{gcov @@mysrclist.txt}).
18632 Note that on AIX compiling a static library with @code{-fprofile-arcs} is
18633 not supported as there can be unresolved symbols during the final link.
18635 @node Profiling an Ada Program using gprof
18636 @section Profiling an Ada Program using gprof
18642 This section is not meant to be an exhaustive documentation of @code{gprof}.
18643 Full documentation for it can be found in the GNU Profiler User's Guide
18644 documentation that is part of this GNAT distribution.
18646 Profiling a program helps determine the parts of a program that are executed
18647 most often, and are therefore the most time-consuming.
18649 @code{gprof} is the standard GNU profiling tool; it has been enhanced to
18650 better handle Ada programs and multitasking.
18651 It is currently supported on the following platforms
18656 solaris sparc/sparc64/x86
18662 In order to profile a program using @code{gprof}, 3 steps are needed:
18666 Code instrumentation, requiring a full recompilation of the project with the
18669 Execution of the program under the analysis conditions, i.e. with the desired
18672 Analysis of the results using the @code{gprof} tool.
18676 The following sections detail the different steps, and indicate how
18677 to interpret the results:
18679 * Compilation for profiling::
18680 * Program execution::
18682 * Interpretation of profiling results::
18685 @node Compilation for profiling
18686 @subsection Compilation for profiling
18690 In order to profile a program the first step is to tell the compiler
18691 to generate the necessary profiling information. The compiler switch to be used
18692 is @code{-pg}, which must be added to other compilation switches. This
18693 switch needs to be specified both during compilation and link stages, and can
18694 be specified once when using gnatmake:
18697 gnatmake -f -pg -P my_project
18701 Note that only the objects that were compiled with the @samp{-pg} switch will be
18702 profiled; if you need to profile your whole project, use the
18703 @samp{-f} gnatmake switch to force full recompilation.
18705 @node Program execution
18706 @subsection Program execution
18709 Once the program has been compiled for profiling, you can run it as usual.
18711 The only constraint imposed by profiling is that the program must terminate
18712 normally. An interrupted program (via a Ctrl-C, kill, etc.) will not be
18715 Once the program completes execution, a data file called @file{gmon.out} is
18716 generated in the directory where the program was launched from. If this file
18717 already exists, it will be overwritten.
18719 @node Running gprof
18720 @subsection Running gprof
18723 The @code{gprof} tool is called as follow:
18726 gprof my_prog gmon.out
18737 The complete form of the gprof command line is the following:
18740 gprof [^switches^options^] [executable [data-file]]
18744 @code{gprof} supports numerous ^switch^options^. The order of these
18745 ^switch^options^ does not matter. The full list of options can be found in
18746 the GNU Profiler User's Guide documentation that comes with this documentation.
18748 The following is the subset of those switches that is most relevant:
18752 @item --demangle[=@var{style}]
18753 @itemx --no-demangle
18754 @cindex @option{--demangle} (@code{gprof})
18755 These options control whether symbol names should be demangled when
18756 printing output. The default is to demangle C++ symbols. The
18757 @code{--no-demangle} option may be used to turn off demangling. Different
18758 compilers have different mangling styles. The optional demangling style
18759 argument can be used to choose an appropriate demangling style for your
18760 compiler, in particular Ada symbols generated by GNAT can be demangled using
18761 @code{--demangle=gnat}.
18763 @item -e @var{function_name}
18764 @cindex @option{-e} (@code{gprof})
18765 The @samp{-e @var{function}} option tells @code{gprof} not to print
18766 information about the function @var{function_name} (and its
18767 children@dots{}) in the call graph. The function will still be listed
18768 as a child of any functions that call it, but its index number will be
18769 shown as @samp{[not printed]}. More than one @samp{-e} option may be
18770 given; only one @var{function_name} may be indicated with each @samp{-e}
18773 @item -E @var{function_name}
18774 @cindex @option{-E} (@code{gprof})
18775 The @code{-E @var{function}} option works like the @code{-e} option, but
18776 execution time spent in the function (and children who were not called from
18777 anywhere else), will not be used to compute the percentages-of-time for
18778 the call graph. More than one @samp{-E} option may be given; only one
18779 @var{function_name} may be indicated with each @samp{-E} option.
18781 @item -f @var{function_name}
18782 @cindex @option{-f} (@code{gprof})
18783 The @samp{-f @var{function}} option causes @code{gprof} to limit the
18784 call graph to the function @var{function_name} and its children (and
18785 their children@dots{}). More than one @samp{-f} option may be given;
18786 only one @var{function_name} may be indicated with each @samp{-f}
18789 @item -F @var{function_name}
18790 @cindex @option{-F} (@code{gprof})
18791 The @samp{-F @var{function}} option works like the @code{-f} option, but
18792 only time spent in the function and its children (and their
18793 children@dots{}) will be used to determine total-time and
18794 percentages-of-time for the call graph. More than one @samp{-F} option
18795 may be given; only one @var{function_name} may be indicated with each
18796 @samp{-F} option. The @samp{-F} option overrides the @samp{-E} option.
18800 @node Interpretation of profiling results
18801 @subsection Interpretation of profiling results
18805 The results of the profiling analysis are represented by two arrays: the
18806 'flat profile' and the 'call graph'. Full documentation of those outputs
18807 can be found in the GNU Profiler User's Guide.
18809 The flat profile shows the time spent in each function of the program, and how
18810 many time it has been called. This allows you to locate easily the most
18811 time-consuming functions.
18813 The call graph shows, for each subprogram, the subprograms that call it,
18814 and the subprograms that it calls. It also provides an estimate of the time
18815 spent in each of those callers/called subprograms.
18818 @c ******************************
18819 @node Running and Debugging Ada Programs
18820 @chapter Running and Debugging Ada Programs
18824 This chapter discusses how to debug Ada programs.
18826 It applies to GNAT on the Alpha OpenVMS platform;
18827 for I64 OpenVMS please refer to the @cite{OpenVMS Debugger Manual},
18828 since HP has implemented Ada support in the OpenVMS debugger on I64.
18831 An incorrect Ada program may be handled in three ways by the GNAT compiler:
18835 The illegality may be a violation of the static semantics of Ada. In
18836 that case GNAT diagnoses the constructs in the program that are illegal.
18837 It is then a straightforward matter for the user to modify those parts of
18841 The illegality may be a violation of the dynamic semantics of Ada. In
18842 that case the program compiles and executes, but may generate incorrect
18843 results, or may terminate abnormally with some exception.
18846 When presented with a program that contains convoluted errors, GNAT
18847 itself may terminate abnormally without providing full diagnostics on
18848 the incorrect user program.
18852 * The GNAT Debugger GDB::
18854 * Introduction to GDB Commands::
18855 * Using Ada Expressions::
18856 * Calling User-Defined Subprograms::
18857 * Using the Next Command in a Function::
18860 * Debugging Generic Units::
18861 * Remote Debugging using gdbserver::
18862 * GNAT Abnormal Termination or Failure to Terminate::
18863 * Naming Conventions for GNAT Source Files::
18864 * Getting Internal Debugging Information::
18865 * Stack Traceback::
18871 @node The GNAT Debugger GDB
18872 @section The GNAT Debugger GDB
18875 @code{GDB} is a general purpose, platform-independent debugger that
18876 can be used to debug mixed-language programs compiled with @command{gcc},
18877 and in particular is capable of debugging Ada programs compiled with
18878 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
18879 complex Ada data structures.
18881 @xref{Top,, Debugging with GDB, gdb, Debugging with GDB},
18883 located in the GNU:[DOCS] directory,
18885 for full details on the usage of @code{GDB}, including a section on
18886 its usage on programs. This manual should be consulted for full
18887 details. The section that follows is a brief introduction to the
18888 philosophy and use of @code{GDB}.
18890 When GNAT programs are compiled, the compiler optionally writes debugging
18891 information into the generated object file, including information on
18892 line numbers, and on declared types and variables. This information is
18893 separate from the generated code. It makes the object files considerably
18894 larger, but it does not add to the size of the actual executable that
18895 will be loaded into memory, and has no impact on run-time performance. The
18896 generation of debug information is triggered by the use of the
18897 ^-g^/DEBUG^ switch in the @command{gcc} or @command{gnatmake} command
18898 used to carry out the compilations. It is important to emphasize that
18899 the use of these options does not change the generated code.
18901 The debugging information is written in standard system formats that
18902 are used by many tools, including debuggers and profilers. The format
18903 of the information is typically designed to describe C types and
18904 semantics, but GNAT implements a translation scheme which allows full
18905 details about Ada types and variables to be encoded into these
18906 standard C formats. Details of this encoding scheme may be found in
18907 the file exp_dbug.ads in the GNAT source distribution. However, the
18908 details of this encoding are, in general, of no interest to a user,
18909 since @code{GDB} automatically performs the necessary decoding.
18911 When a program is bound and linked, the debugging information is
18912 collected from the object files, and stored in the executable image of
18913 the program. Again, this process significantly increases the size of
18914 the generated executable file, but it does not increase the size of
18915 the executable program itself. Furthermore, if this program is run in
18916 the normal manner, it runs exactly as if the debug information were
18917 not present, and takes no more actual memory.
18919 However, if the program is run under control of @code{GDB}, the
18920 debugger is activated. The image of the program is loaded, at which
18921 point it is ready to run. If a run command is given, then the program
18922 will run exactly as it would have if @code{GDB} were not present. This
18923 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
18924 entirely non-intrusive until a breakpoint is encountered. If no
18925 breakpoint is ever hit, the program will run exactly as it would if no
18926 debugger were present. When a breakpoint is hit, @code{GDB} accesses
18927 the debugging information and can respond to user commands to inspect
18928 variables, and more generally to report on the state of execution.
18932 @section Running GDB
18935 This section describes how to initiate the debugger.
18936 @c The above sentence is really just filler, but it was otherwise
18937 @c clumsy to get the first paragraph nonindented given the conditional
18938 @c nature of the description
18941 The debugger can be launched from a @code{GPS} menu or
18942 directly from the command line. The description below covers the latter use.
18943 All the commands shown can be used in the @code{GPS} debug console window,
18944 but there are usually more GUI-based ways to achieve the same effect.
18947 The command to run @code{GDB} is
18950 $ ^gdb program^GDB PROGRAM^
18954 where @code{^program^PROGRAM^} is the name of the executable file. This
18955 activates the debugger and results in a prompt for debugger commands.
18956 The simplest command is simply @code{run}, which causes the program to run
18957 exactly as if the debugger were not present. The following section
18958 describes some of the additional commands that can be given to @code{GDB}.
18960 @c *******************************
18961 @node Introduction to GDB Commands
18962 @section Introduction to GDB Commands
18965 @code{GDB} contains a large repertoire of commands. @xref{Top,,
18966 Debugging with GDB, gdb, Debugging with GDB},
18968 located in the GNU:[DOCS] directory,
18970 for extensive documentation on the use
18971 of these commands, together with examples of their use. Furthermore,
18972 the command @command{help} invoked from within GDB activates a simple help
18973 facility which summarizes the available commands and their options.
18974 In this section we summarize a few of the most commonly
18975 used commands to give an idea of what @code{GDB} is about. You should create
18976 a simple program with debugging information and experiment with the use of
18977 these @code{GDB} commands on the program as you read through the
18981 @item set args @var{arguments}
18982 The @var{arguments} list above is a list of arguments to be passed to
18983 the program on a subsequent run command, just as though the arguments
18984 had been entered on a normal invocation of the program. The @code{set args}
18985 command is not needed if the program does not require arguments.
18988 The @code{run} command causes execution of the program to start from
18989 the beginning. If the program is already running, that is to say if
18990 you are currently positioned at a breakpoint, then a prompt will ask
18991 for confirmation that you want to abandon the current execution and
18994 @item breakpoint @var{location}
18995 The breakpoint command sets a breakpoint, that is to say a point at which
18996 execution will halt and @code{GDB} will await further
18997 commands. @var{location} is
18998 either a line number within a file, given in the format @code{file:linenumber},
18999 or it is the name of a subprogram. If you request that a breakpoint be set on
19000 a subprogram that is overloaded, a prompt will ask you to specify on which of
19001 those subprograms you want to breakpoint. You can also
19002 specify that all of them should be breakpointed. If the program is run
19003 and execution encounters the breakpoint, then the program
19004 stops and @code{GDB} signals that the breakpoint was encountered by
19005 printing the line of code before which the program is halted.
19007 @item catch exception @var{name}
19008 This command causes the program execution to stop whenever exception
19009 @var{name} is raised. If @var{name} is omitted, then the execution is
19010 suspended when any exception is raised.
19012 @item print @var{expression}
19013 This will print the value of the given expression. Most simple
19014 Ada expression formats are properly handled by @code{GDB}, so the expression
19015 can contain function calls, variables, operators, and attribute references.
19018 Continues execution following a breakpoint, until the next breakpoint or the
19019 termination of the program.
19022 Executes a single line after a breakpoint. If the next statement
19023 is a subprogram call, execution continues into (the first statement of)
19024 the called subprogram.
19027 Executes a single line. If this line is a subprogram call, executes and
19028 returns from the call.
19031 Lists a few lines around the current source location. In practice, it
19032 is usually more convenient to have a separate edit window open with the
19033 relevant source file displayed. Successive applications of this command
19034 print subsequent lines. The command can be given an argument which is a
19035 line number, in which case it displays a few lines around the specified one.
19038 Displays a backtrace of the call chain. This command is typically
19039 used after a breakpoint has occurred, to examine the sequence of calls that
19040 leads to the current breakpoint. The display includes one line for each
19041 activation record (frame) corresponding to an active subprogram.
19044 At a breakpoint, @code{GDB} can display the values of variables local
19045 to the current frame. The command @code{up} can be used to
19046 examine the contents of other active frames, by moving the focus up
19047 the stack, that is to say from callee to caller, one frame at a time.
19050 Moves the focus of @code{GDB} down from the frame currently being
19051 examined to the frame of its callee (the reverse of the previous command),
19053 @item frame @var{n}
19054 Inspect the frame with the given number. The value 0 denotes the frame
19055 of the current breakpoint, that is to say the top of the call stack.
19060 The above list is a very short introduction to the commands that
19061 @code{GDB} provides. Important additional capabilities, including conditional
19062 breakpoints, the ability to execute command sequences on a breakpoint,
19063 the ability to debug at the machine instruction level and many other
19064 features are described in detail in @ref{Top,, Debugging with GDB, gdb,
19065 Debugging with GDB}. Note that most commands can be abbreviated
19066 (for example, c for continue, bt for backtrace).
19068 @node Using Ada Expressions
19069 @section Using Ada Expressions
19070 @cindex Ada expressions
19073 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
19074 extensions. The philosophy behind the design of this subset is
19078 That @code{GDB} should provide basic literals and access to operations for
19079 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
19080 leaving more sophisticated computations to subprograms written into the
19081 program (which therefore may be called from @code{GDB}).
19084 That type safety and strict adherence to Ada language restrictions
19085 are not particularly important to the @code{GDB} user.
19088 That brevity is important to the @code{GDB} user.
19092 Thus, for brevity, the debugger acts as if there were
19093 implicit @code{with} and @code{use} clauses in effect for all user-written
19094 packages, thus making it unnecessary to fully qualify most names with
19095 their packages, regardless of context. Where this causes ambiguity,
19096 @code{GDB} asks the user's intent.
19098 For details on the supported Ada syntax, see @ref{Top,, Debugging with
19099 GDB, gdb, Debugging with GDB}.
19101 @node Calling User-Defined Subprograms
19102 @section Calling User-Defined Subprograms
19105 An important capability of @code{GDB} is the ability to call user-defined
19106 subprograms while debugging. This is achieved simply by entering
19107 a subprogram call statement in the form:
19110 call subprogram-name (parameters)
19114 The keyword @code{call} can be omitted in the normal case where the
19115 @code{subprogram-name} does not coincide with any of the predefined
19116 @code{GDB} commands.
19118 The effect is to invoke the given subprogram, passing it the
19119 list of parameters that is supplied. The parameters can be expressions and
19120 can include variables from the program being debugged. The
19121 subprogram must be defined
19122 at the library level within your program, and @code{GDB} will call the
19123 subprogram within the environment of your program execution (which
19124 means that the subprogram is free to access or even modify variables
19125 within your program).
19127 The most important use of this facility is in allowing the inclusion of
19128 debugging routines that are tailored to particular data structures
19129 in your program. Such debugging routines can be written to provide a suitably
19130 high-level description of an abstract type, rather than a low-level dump
19131 of its physical layout. After all, the standard
19132 @code{GDB print} command only knows the physical layout of your
19133 types, not their abstract meaning. Debugging routines can provide information
19134 at the desired semantic level and are thus enormously useful.
19136 For example, when debugging GNAT itself, it is crucial to have access to
19137 the contents of the tree nodes used to represent the program internally.
19138 But tree nodes are represented simply by an integer value (which in turn
19139 is an index into a table of nodes).
19140 Using the @code{print} command on a tree node would simply print this integer
19141 value, which is not very useful. But the PN routine (defined in file
19142 treepr.adb in the GNAT sources) takes a tree node as input, and displays
19143 a useful high level representation of the tree node, which includes the
19144 syntactic category of the node, its position in the source, the integers
19145 that denote descendant nodes and parent node, as well as varied
19146 semantic information. To study this example in more detail, you might want to
19147 look at the body of the PN procedure in the stated file.
19149 @node Using the Next Command in a Function
19150 @section Using the Next Command in a Function
19153 When you use the @code{next} command in a function, the current source
19154 location will advance to the next statement as usual. A special case
19155 arises in the case of a @code{return} statement.
19157 Part of the code for a return statement is the ``epilog'' of the function.
19158 This is the code that returns to the caller. There is only one copy of
19159 this epilog code, and it is typically associated with the last return
19160 statement in the function if there is more than one return. In some
19161 implementations, this epilog is associated with the first statement
19164 The result is that if you use the @code{next} command from a return
19165 statement that is not the last return statement of the function you
19166 may see a strange apparent jump to the last return statement or to
19167 the start of the function. You should simply ignore this odd jump.
19168 The value returned is always that from the first return statement
19169 that was stepped through.
19171 @node Ada Exceptions
19172 @section Stopping when Ada Exceptions are Raised
19176 You can set catchpoints that stop the program execution when your program
19177 raises selected exceptions.
19180 @item catch exception
19181 Set a catchpoint that stops execution whenever (any task in the) program
19182 raises any exception.
19184 @item catch exception @var{name}
19185 Set a catchpoint that stops execution whenever (any task in the) program
19186 raises the exception @var{name}.
19188 @item catch exception unhandled
19189 Set a catchpoint that stops executino whenever (any task in the) program
19190 raises an exception for which there is no handler.
19192 @item info exceptions
19193 @itemx info exceptions @var{regexp}
19194 The @code{info exceptions} command permits the user to examine all defined
19195 exceptions within Ada programs. With a regular expression, @var{regexp}, as
19196 argument, prints out only those exceptions whose name matches @var{regexp}.
19204 @code{GDB} allows the following task-related commands:
19208 This command shows a list of current Ada tasks, as in the following example:
19215 ID TID P-ID Thread Pri State Name
19216 1 8088000 0 807e000 15 Child Activation Wait main_task
19217 2 80a4000 1 80ae000 15 Accept/Select Wait b
19218 3 809a800 1 80a4800 15 Child Activation Wait a
19219 * 4 80ae800 3 80b8000 15 Running c
19223 In this listing, the asterisk before the first task indicates it to be the
19224 currently running task. The first column lists the task ID that is used
19225 to refer to tasks in the following commands.
19227 @item break @var{linespec} task @var{taskid}
19228 @itemx break @var{linespec} task @var{taskid} if @dots{}
19229 @cindex Breakpoints and tasks
19230 These commands are like the @code{break @dots{} thread @dots{}}.
19231 @var{linespec} specifies source lines.
19233 Use the qualifier @samp{task @var{taskid}} with a breakpoint command
19234 to specify that you only want @code{GDB} to stop the program when a
19235 particular Ada task reaches this breakpoint. @var{taskid} is one of the
19236 numeric task identifiers assigned by @code{GDB}, shown in the first
19237 column of the @samp{info tasks} display.
19239 If you do not specify @samp{task @var{taskid}} when you set a
19240 breakpoint, the breakpoint applies to @emph{all} tasks of your
19243 You can use the @code{task} qualifier on conditional breakpoints as
19244 well; in this case, place @samp{task @var{taskid}} before the
19245 breakpoint condition (before the @code{if}).
19247 @item task @var{taskno}
19248 @cindex Task switching
19250 This command allows to switch to the task referred by @var{taskno}. In
19251 particular, This allows to browse the backtrace of the specified
19252 task. It is advised to switch back to the original task before
19253 continuing execution otherwise the scheduling of the program may be
19258 For more detailed information on the tasking support,
19259 see @ref{Top,, Debugging with GDB, gdb, Debugging with GDB}.
19261 @node Debugging Generic Units
19262 @section Debugging Generic Units
19263 @cindex Debugging Generic Units
19267 GNAT always uses code expansion for generic instantiation. This means that
19268 each time an instantiation occurs, a complete copy of the original code is
19269 made, with appropriate substitutions of formals by actuals.
19271 It is not possible to refer to the original generic entities in
19272 @code{GDB}, but it is always possible to debug a particular instance of
19273 a generic, by using the appropriate expanded names. For example, if we have
19275 @smallexample @c ada
19280 generic package k is
19281 procedure kp (v1 : in out integer);
19285 procedure kp (v1 : in out integer) is
19291 package k1 is new k;
19292 package k2 is new k;
19294 var : integer := 1;
19307 Then to break on a call to procedure kp in the k2 instance, simply
19311 (gdb) break g.k2.kp
19315 When the breakpoint occurs, you can step through the code of the
19316 instance in the normal manner and examine the values of local variables, as for
19319 @node Remote Debugging using gdbserver
19320 @section Remote Debugging using gdbserver
19321 @cindex Remote Debugging using gdbserver
19324 On platforms where gdbserver is supported, it is possible to use this tool
19325 to debug your application remotely. This can be useful in situations
19326 where the program needs to be run on a target host that is different
19327 from the host used for development, particularly when the target has
19328 a limited amount of resources (either CPU and/or memory).
19330 To do so, start your program using gdbserver on the target machine.
19331 gdbserver then automatically suspends the execution of your program
19332 at its entry point, waiting for a debugger to connect to it. The
19333 following commands starts an application and tells gdbserver to
19334 wait for a connection with the debugger on localhost port 4444.
19337 $ gdbserver localhost:4444 program
19338 Process program created; pid = 5685
19339 Listening on port 4444
19342 Once gdbserver has started listening, we can tell the debugger to establish
19343 a connection with this gdbserver, and then start the same debugging session
19344 as if the program was being debugged on the same host, directly under
19345 the control of GDB.
19349 (gdb) target remote targethost:4444
19350 Remote debugging using targethost:4444
19351 0x00007f29936d0af0 in ?? () from /lib64/ld-linux-x86-64.so.
19353 Breakpoint 1 at 0x401f0c: file foo.adb, line 3.
19357 Breakpoint 1, foo () at foo.adb:4
19361 It is also possible to use gdbserver to attach to an already running
19362 program, in which case the execution of that program is simply suspended
19363 until the connection between the debugger and gdbserver is established.
19365 For more information on how to use gdbserver, @ref{Top, Server, Using
19366 the gdbserver Program, gdb, Debugging with GDB}. GNAT Pro provides support
19367 for gdbserver on x86-linux, x86-windows and x86_64-linux.
19369 @node GNAT Abnormal Termination or Failure to Terminate
19370 @section GNAT Abnormal Termination or Failure to Terminate
19371 @cindex GNAT Abnormal Termination or Failure to Terminate
19374 When presented with programs that contain serious errors in syntax
19376 GNAT may on rare occasions experience problems in operation, such
19378 segmentation fault or illegal memory access, raising an internal
19379 exception, terminating abnormally, or failing to terminate at all.
19380 In such cases, you can activate
19381 various features of GNAT that can help you pinpoint the construct in your
19382 program that is the likely source of the problem.
19384 The following strategies are presented in increasing order of
19385 difficulty, corresponding to your experience in using GNAT and your
19386 familiarity with compiler internals.
19390 Run @command{gcc} with the @option{-gnatf}. This first
19391 switch causes all errors on a given line to be reported. In its absence,
19392 only the first error on a line is displayed.
19394 The @option{-gnatdO} switch causes errors to be displayed as soon as they
19395 are encountered, rather than after compilation is terminated. If GNAT
19396 terminates prematurely or goes into an infinite loop, the last error
19397 message displayed may help to pinpoint the culprit.
19400 Run @command{gcc} with the @option{^-v (verbose)^/VERBOSE^} switch. In this
19401 mode, @command{gcc} produces ongoing information about the progress of the
19402 compilation and provides the name of each procedure as code is
19403 generated. This switch allows you to find which Ada procedure was being
19404 compiled when it encountered a code generation problem.
19407 @cindex @option{-gnatdc} switch
19408 Run @command{gcc} with the @option{-gnatdc} switch. This is a GNAT specific
19409 switch that does for the front-end what @option{^-v^VERBOSE^} does
19410 for the back end. The system prints the name of each unit,
19411 either a compilation unit or nested unit, as it is being analyzed.
19413 Finally, you can start
19414 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
19415 front-end of GNAT, and can be run independently (normally it is just
19416 called from @command{gcc}). You can use @code{gdb} on @code{gnat1} as you
19417 would on a C program (but @pxref{The GNAT Debugger GDB} for caveats). The
19418 @code{where} command is the first line of attack; the variable
19419 @code{lineno} (seen by @code{print lineno}), used by the second phase of
19420 @code{gnat1} and by the @command{gcc} backend, indicates the source line at
19421 which the execution stopped, and @code{input_file name} indicates the name of
19425 @node Naming Conventions for GNAT Source Files
19426 @section Naming Conventions for GNAT Source Files
19429 In order to examine the workings of the GNAT system, the following
19430 brief description of its organization may be helpful:
19434 Files with prefix @file{^sc^SC^} contain the lexical scanner.
19437 All files prefixed with @file{^par^PAR^} are components of the parser. The
19438 numbers correspond to chapters of the Ada Reference Manual. For example,
19439 parsing of select statements can be found in @file{par-ch9.adb}.
19442 All files prefixed with @file{^sem^SEM^} perform semantic analysis. The
19443 numbers correspond to chapters of the Ada standard. For example, all
19444 issues involving context clauses can be found in @file{sem_ch10.adb}. In
19445 addition, some features of the language require sufficient special processing
19446 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
19447 dynamic dispatching, etc.
19450 All files prefixed with @file{^exp^EXP^} perform normalization and
19451 expansion of the intermediate representation (abstract syntax tree, or AST).
19452 these files use the same numbering scheme as the parser and semantics files.
19453 For example, the construction of record initialization procedures is done in
19454 @file{exp_ch3.adb}.
19457 The files prefixed with @file{^bind^BIND^} implement the binder, which
19458 verifies the consistency of the compilation, determines an order of
19459 elaboration, and generates the bind file.
19462 The files @file{atree.ads} and @file{atree.adb} detail the low-level
19463 data structures used by the front-end.
19466 The files @file{sinfo.ads} and @file{sinfo.adb} detail the structure of
19467 the abstract syntax tree as produced by the parser.
19470 The files @file{einfo.ads} and @file{einfo.adb} detail the attributes of
19471 all entities, computed during semantic analysis.
19474 Library management issues are dealt with in files with prefix
19480 Ada files with the prefix @file{^a-^A-^} are children of @code{Ada}, as
19481 defined in Annex A.
19486 Files with prefix @file{^i-^I-^} are children of @code{Interfaces}, as
19487 defined in Annex B.
19491 Files with prefix @file{^s-^S-^} are children of @code{System}. This includes
19492 both language-defined children and GNAT run-time routines.
19496 Files with prefix @file{^g-^G-^} are children of @code{GNAT}. These are useful
19497 general-purpose packages, fully documented in their specs. All
19498 the other @file{.c} files are modifications of common @command{gcc} files.
19501 @node Getting Internal Debugging Information
19502 @section Getting Internal Debugging Information
19505 Most compilers have internal debugging switches and modes. GNAT
19506 does also, except GNAT internal debugging switches and modes are not
19507 secret. A summary and full description of all the compiler and binder
19508 debug flags are in the file @file{debug.adb}. You must obtain the
19509 sources of the compiler to see the full detailed effects of these flags.
19511 The switches that print the source of the program (reconstructed from
19512 the internal tree) are of general interest for user programs, as are the
19514 the full internal tree, and the entity table (the symbol table
19515 information). The reconstructed source provides a readable version of the
19516 program after the front-end has completed analysis and expansion,
19517 and is useful when studying the performance of specific constructs.
19518 For example, constraint checks are indicated, complex aggregates
19519 are replaced with loops and assignments, and tasking primitives
19520 are replaced with run-time calls.
19522 @node Stack Traceback
19523 @section Stack Traceback
19525 @cindex stack traceback
19526 @cindex stack unwinding
19529 Traceback is a mechanism to display the sequence of subprogram calls that
19530 leads to a specified execution point in a program. Often (but not always)
19531 the execution point is an instruction at which an exception has been raised.
19532 This mechanism is also known as @i{stack unwinding} because it obtains
19533 its information by scanning the run-time stack and recovering the activation
19534 records of all active subprograms. Stack unwinding is one of the most
19535 important tools for program debugging.
19537 The first entry stored in traceback corresponds to the deepest calling level,
19538 that is to say the subprogram currently executing the instruction
19539 from which we want to obtain the traceback.
19541 Note that there is no runtime performance penalty when stack traceback
19542 is enabled, and no exception is raised during program execution.
19545 * Non-Symbolic Traceback::
19546 * Symbolic Traceback::
19549 @node Non-Symbolic Traceback
19550 @subsection Non-Symbolic Traceback
19551 @cindex traceback, non-symbolic
19554 Note: this feature is not supported on all platforms. See
19555 @file{GNAT.Traceback spec in g-traceb.ads} for a complete list of supported
19559 * Tracebacks From an Unhandled Exception::
19560 * Tracebacks From Exception Occurrences (non-symbolic)::
19561 * Tracebacks From Anywhere in a Program (non-symbolic)::
19564 @node Tracebacks From an Unhandled Exception
19565 @subsubsection Tracebacks From an Unhandled Exception
19568 A runtime non-symbolic traceback is a list of addresses of call instructions.
19569 To enable this feature you must use the @option{-E}
19570 @code{gnatbind}'s option. With this option a stack traceback is stored as part
19571 of exception information. You can retrieve this information using the
19572 @code{addr2line} tool.
19574 Here is a simple example:
19576 @smallexample @c ada
19582 raise Constraint_Error;
19597 $ gnatmake stb -bargs -E
19600 Execution terminated by unhandled exception
19601 Exception name: CONSTRAINT_ERROR
19603 Call stack traceback locations:
19604 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
19608 As we see the traceback lists a sequence of addresses for the unhandled
19609 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
19610 guess that this exception come from procedure P1. To translate these
19611 addresses into the source lines where the calls appear, the
19612 @code{addr2line} tool, described below, is invaluable. The use of this tool
19613 requires the program to be compiled with debug information.
19616 $ gnatmake -g stb -bargs -E
19619 Execution terminated by unhandled exception
19620 Exception name: CONSTRAINT_ERROR
19622 Call stack traceback locations:
19623 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
19625 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
19626 0x4011f1 0x77e892a4
19628 00401373 at d:/stb/stb.adb:5
19629 0040138B at d:/stb/stb.adb:10
19630 0040139C at d:/stb/stb.adb:14
19631 00401335 at d:/stb/b~stb.adb:104
19632 004011C4 at /build/@dots{}/crt1.c:200
19633 004011F1 at /build/@dots{}/crt1.c:222
19634 77E892A4 in ?? at ??:0
19638 The @code{addr2line} tool has several other useful options:
19642 to get the function name corresponding to any location
19644 @item --demangle=gnat
19645 to use the gnat decoding mode for the function names. Note that
19646 for binutils version 2.9.x the option is simply @option{--demangle}.
19650 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
19651 0x40139c 0x401335 0x4011c4 0x4011f1
19653 00401373 in stb.p1 at d:/stb/stb.adb:5
19654 0040138B in stb.p2 at d:/stb/stb.adb:10
19655 0040139C in stb at d:/stb/stb.adb:14
19656 00401335 in main at d:/stb/b~stb.adb:104
19657 004011C4 in <__mingw_CRTStartup> at /build/@dots{}/crt1.c:200
19658 004011F1 in <mainCRTStartup> at /build/@dots{}/crt1.c:222
19662 From this traceback we can see that the exception was raised in
19663 @file{stb.adb} at line 5, which was reached from a procedure call in
19664 @file{stb.adb} at line 10, and so on. The @file{b~std.adb} is the binder file,
19665 which contains the call to the main program.
19666 @xref{Running gnatbind}. The remaining entries are assorted runtime routines,
19667 and the output will vary from platform to platform.
19669 It is also possible to use @code{GDB} with these traceback addresses to debug
19670 the program. For example, we can break at a given code location, as reported
19671 in the stack traceback:
19677 Furthermore, this feature is not implemented inside Windows DLL. Only
19678 the non-symbolic traceback is reported in this case.
19681 (gdb) break *0x401373
19682 Breakpoint 1 at 0x401373: file stb.adb, line 5.
19686 It is important to note that the stack traceback addresses
19687 do not change when debug information is included. This is particularly useful
19688 because it makes it possible to release software without debug information (to
19689 minimize object size), get a field report that includes a stack traceback
19690 whenever an internal bug occurs, and then be able to retrieve the sequence
19691 of calls with the same program compiled with debug information.
19693 @node Tracebacks From Exception Occurrences (non-symbolic)
19694 @subsubsection Tracebacks From Exception Occurrences
19697 Non-symbolic tracebacks are obtained by using the @option{-E} binder argument.
19698 The stack traceback is attached to the exception information string, and can
19699 be retrieved in an exception handler within the Ada program, by means of the
19700 Ada facilities defined in @code{Ada.Exceptions}. Here is a simple example:
19702 @smallexample @c ada
19704 with Ada.Exceptions;
19709 use Ada.Exceptions;
19717 Text_IO.Put_Line (Exception_Information (E));
19731 This program will output:
19736 Exception name: CONSTRAINT_ERROR
19737 Message: stb.adb:12
19738 Call stack traceback locations:
19739 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
19742 @node Tracebacks From Anywhere in a Program (non-symbolic)
19743 @subsubsection Tracebacks From Anywhere in a Program
19746 It is also possible to retrieve a stack traceback from anywhere in a
19747 program. For this you need to
19748 use the @code{GNAT.Traceback} API. This package includes a procedure called
19749 @code{Call_Chain} that computes a complete stack traceback, as well as useful
19750 display procedures described below. It is not necessary to use the
19751 @option{-E gnatbind} option in this case, because the stack traceback mechanism
19752 is invoked explicitly.
19755 In the following example we compute a traceback at a specific location in
19756 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
19757 convert addresses to strings:
19759 @smallexample @c ada
19761 with GNAT.Traceback;
19762 with GNAT.Debug_Utilities;
19768 use GNAT.Traceback;
19771 TB : Tracebacks_Array (1 .. 10);
19772 -- We are asking for a maximum of 10 stack frames.
19774 -- Len will receive the actual number of stack frames returned.
19776 Call_Chain (TB, Len);
19778 Text_IO.Put ("In STB.P1 : ");
19780 for K in 1 .. Len loop
19781 Text_IO.Put (Debug_Utilities.Image (TB (K)));
19802 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
19803 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
19807 You can then get further information by invoking the @code{addr2line}
19808 tool as described earlier (note that the hexadecimal addresses
19809 need to be specified in C format, with a leading ``0x'').
19811 @node Symbolic Traceback
19812 @subsection Symbolic Traceback
19813 @cindex traceback, symbolic
19816 A symbolic traceback is a stack traceback in which procedure names are
19817 associated with each code location.
19820 Note that this feature is not supported on all platforms. See
19821 @file{GNAT.Traceback.Symbolic spec in g-trasym.ads} for a complete
19822 list of currently supported platforms.
19825 Note that the symbolic traceback requires that the program be compiled
19826 with debug information. If it is not compiled with debug information
19827 only the non-symbolic information will be valid.
19830 * Tracebacks From Exception Occurrences (symbolic)::
19831 * Tracebacks From Anywhere in a Program (symbolic)::
19834 @node Tracebacks From Exception Occurrences (symbolic)
19835 @subsubsection Tracebacks From Exception Occurrences
19837 @smallexample @c ada
19839 with GNAT.Traceback.Symbolic;
19845 raise Constraint_Error;
19862 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
19867 $ gnatmake -g .\stb -bargs -E -largs -lgnat -laddr2line -lintl
19870 0040149F in stb.p1 at stb.adb:8
19871 004014B7 in stb.p2 at stb.adb:13
19872 004014CF in stb.p3 at stb.adb:18
19873 004015DD in ada.stb at stb.adb:22
19874 00401461 in main at b~stb.adb:168
19875 004011C4 in __mingw_CRTStartup at crt1.c:200
19876 004011F1 in mainCRTStartup at crt1.c:222
19877 77E892A4 in ?? at ??:0
19881 In the above example the ``.\'' syntax in the @command{gnatmake} command
19882 is currently required by @command{addr2line} for files that are in
19883 the current working directory.
19884 Moreover, the exact sequence of linker options may vary from platform
19886 The above @option{-largs} section is for Windows platforms. By contrast,
19887 under Unix there is no need for the @option{-largs} section.
19888 Differences across platforms are due to details of linker implementation.
19890 @node Tracebacks From Anywhere in a Program (symbolic)
19891 @subsubsection Tracebacks From Anywhere in a Program
19894 It is possible to get a symbolic stack traceback
19895 from anywhere in a program, just as for non-symbolic tracebacks.
19896 The first step is to obtain a non-symbolic
19897 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
19898 information. Here is an example:
19900 @smallexample @c ada
19902 with GNAT.Traceback;
19903 with GNAT.Traceback.Symbolic;
19908 use GNAT.Traceback;
19909 use GNAT.Traceback.Symbolic;
19912 TB : Tracebacks_Array (1 .. 10);
19913 -- We are asking for a maximum of 10 stack frames.
19915 -- Len will receive the actual number of stack frames returned.
19917 Call_Chain (TB, Len);
19918 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
19931 @c ******************************
19933 @node Compatibility with HP Ada
19934 @chapter Compatibility with HP Ada
19935 @cindex Compatibility
19940 @cindex Compatibility between GNAT and HP Ada
19941 This chapter compares HP Ada (formerly known as ``DEC Ada'')
19942 for OpenVMS Alpha and GNAT for OpenVMS for Alpha and for I64.
19943 GNAT is highly compatible
19944 with HP Ada, and it should generally be straightforward to port code
19945 from the HP Ada environment to GNAT. However, there are a few language
19946 and implementation differences of which the user must be aware. These
19947 differences are discussed in this chapter. In
19948 addition, the operating environment and command structure for the
19949 compiler are different, and these differences are also discussed.
19951 For further details on these and other compatibility issues,
19952 see Appendix E of the HP publication
19953 @cite{HP Ada, Technical Overview and Comparison on HP Platforms}.
19955 Except where otherwise indicated, the description of GNAT for OpenVMS
19956 applies to both the Alpha and I64 platforms.
19958 For information on porting Ada code from GNAT on Alpha OpenVMS to GNAT on
19959 I64 OpenVMS, see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
19961 The discussion in this chapter addresses specifically the implementation
19962 of Ada 83 for HP OpenVMS Alpha Systems. In cases where the implementation
19963 of HP Ada differs between OpenVMS Alpha Systems and OpenVMS VAX Systems,
19964 GNAT always follows the Alpha implementation.
19966 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
19967 attributes are recognized, although only a subset of them can sensibly
19968 be implemented. The description of pragmas in
19969 @xref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference Manual},
19970 indicates whether or not they are applicable to non-VMS systems.
19973 * Ada Language Compatibility::
19974 * Differences in the Definition of Package System::
19975 * Language-Related Features::
19976 * The Package STANDARD::
19977 * The Package SYSTEM::
19978 * Tasking and Task-Related Features::
19979 * Pragmas and Pragma-Related Features::
19980 * Library of Predefined Units::
19982 * Main Program Definition::
19983 * Implementation-Defined Attributes::
19984 * Compiler and Run-Time Interfacing::
19985 * Program Compilation and Library Management::
19987 * Implementation Limits::
19988 * Tools and Utilities::
19991 @node Ada Language Compatibility
19992 @section Ada Language Compatibility
19995 GNAT handles Ada 95 and Ada 2005 as well as Ada 83, whereas HP Ada is only
19996 for Ada 83. Ada 95 and Ada 2005 are almost completely upwards compatible
19997 with Ada 83, and therefore Ada 83 programs will compile
19998 and run under GNAT with
19999 no changes or only minor changes. The @cite{Annotated Ada Reference Manual}
20000 provides details on specific incompatibilities.
20002 GNAT provides the switch @option{/83} on the @command{GNAT COMPILE} command,
20003 as well as the pragma @code{ADA_83}, to force the compiler to
20004 operate in Ada 83 mode. This mode does not guarantee complete
20005 conformance to Ada 83, but in practice is sufficient to
20006 eliminate most sources of incompatibilities.
20007 In particular, it eliminates the recognition of the
20008 additional Ada 95 and Ada 2005 keywords, so that their use as identifiers
20009 in Ada 83 programs is legal, and handles the cases of packages
20010 with optional bodies, and generics that instantiate unconstrained
20011 types without the use of @code{(<>)}.
20013 @node Differences in the Definition of Package System
20014 @section Differences in the Definition of Package @code{System}
20017 An Ada compiler is allowed to add
20018 implementation-dependent declarations to package @code{System}.
20020 GNAT does not take advantage of this permission, and the version of
20021 @code{System} provided by GNAT exactly matches that defined in the Ada
20024 However, HP Ada adds an extensive set of declarations to package
20026 as fully documented in the HP Ada manuals. To minimize changes required
20027 for programs that make use of these extensions, GNAT provides the pragma
20028 @code{Extend_System} for extending the definition of package System. By using:
20029 @cindex pragma @code{Extend_System}
20030 @cindex @code{Extend_System} pragma
20032 @smallexample @c ada
20035 pragma Extend_System (Aux_DEC);
20041 the set of definitions in @code{System} is extended to include those in
20042 package @code{System.Aux_DEC}.
20043 @cindex @code{System.Aux_DEC} package
20044 @cindex @code{Aux_DEC} package (child of @code{System})
20045 These definitions are incorporated directly into package @code{System},
20046 as though they had been declared there. For a
20047 list of the declarations added, see the spec of this package,
20048 which can be found in the file @file{s-auxdec.ads} in the GNAT library.
20049 @cindex @file{s-auxdec.ads} file
20050 The pragma @code{Extend_System} is a configuration pragma, which means that
20051 it can be placed in the file @file{gnat.adc}, so that it will automatically
20052 apply to all subsequent compilations. See @ref{Configuration Pragmas},
20053 for further details.
20055 An alternative approach that avoids the use of the non-standard
20056 @code{Extend_System} pragma is to add a context clause to the unit that
20057 references these facilities:
20059 @smallexample @c ada
20061 with System.Aux_DEC;
20062 use System.Aux_DEC;
20067 The effect is not quite semantically identical to incorporating
20068 the declarations directly into package @code{System},
20069 but most programs will not notice a difference
20070 unless they use prefix notation (e.g.@: @code{System.Integer_8})
20071 to reference the entities directly in package @code{System}.
20072 For units containing such references,
20073 the prefixes must either be removed, or the pragma @code{Extend_System}
20076 @node Language-Related Features
20077 @section Language-Related Features
20080 The following sections highlight differences in types,
20081 representations of types, operations, alignment, and
20085 * Integer Types and Representations::
20086 * Floating-Point Types and Representations::
20087 * Pragmas Float_Representation and Long_Float::
20088 * Fixed-Point Types and Representations::
20089 * Record and Array Component Alignment::
20090 * Address Clauses::
20091 * Other Representation Clauses::
20094 @node Integer Types and Representations
20095 @subsection Integer Types and Representations
20098 The set of predefined integer types is identical in HP Ada and GNAT.
20099 Furthermore the representation of these integer types is also identical,
20100 including the capability of size clauses forcing biased representation.
20103 HP Ada for OpenVMS Alpha systems has defined the
20104 following additional integer types in package @code{System}:
20121 @code{LARGEST_INTEGER}
20125 In GNAT, the first four of these types may be obtained from the
20126 standard Ada package @code{Interfaces}.
20127 Alternatively, by use of the pragma @code{Extend_System}, identical
20128 declarations can be referenced directly in package @code{System}.
20129 On both GNAT and HP Ada, the maximum integer size is 64 bits.
20131 @node Floating-Point Types and Representations
20132 @subsection Floating-Point Types and Representations
20133 @cindex Floating-Point types
20136 The set of predefined floating-point types is identical in HP Ada and GNAT.
20137 Furthermore the representation of these floating-point
20138 types is also identical. One important difference is that the default
20139 representation for HP Ada is @code{VAX_Float}, but the default representation
20142 Specific types may be declared to be @code{VAX_Float} or IEEE, using the
20143 pragma @code{Float_Representation} as described in the HP Ada
20145 For example, the declarations:
20147 @smallexample @c ada
20149 type F_Float is digits 6;
20150 pragma Float_Representation (VAX_Float, F_Float);
20155 declares a type @code{F_Float} that will be represented in @code{VAX_Float}
20157 This set of declarations actually appears in @code{System.Aux_DEC},
20159 the full set of additional floating-point declarations provided in
20160 the HP Ada version of package @code{System}.
20161 This and similar declarations may be accessed in a user program
20162 by using pragma @code{Extend_System}. The use of this
20163 pragma, and the related pragma @code{Long_Float} is described in further
20164 detail in the following section.
20166 @node Pragmas Float_Representation and Long_Float
20167 @subsection Pragmas @code{Float_Representation} and @code{Long_Float}
20170 HP Ada provides the pragma @code{Float_Representation}, which
20171 acts as a program library switch to allow control over
20172 the internal representation chosen for the predefined
20173 floating-point types declared in the package @code{Standard}.
20174 The format of this pragma is as follows:
20176 @smallexample @c ada
20178 pragma Float_Representation(VAX_Float | IEEE_Float);
20183 This pragma controls the representation of floating-point
20188 @code{VAX_Float} specifies that floating-point
20189 types are represented by default with the VAX system hardware types
20190 @code{F-floating}, @code{D-floating}, @code{G-floating}.
20191 Note that the @code{H-floating}
20192 type was available only on VAX systems, and is not available
20193 in either HP Ada or GNAT.
20196 @code{IEEE_Float} specifies that floating-point
20197 types are represented by default with the IEEE single and
20198 double floating-point types.
20202 GNAT provides an identical implementation of the pragma
20203 @code{Float_Representation}, except that it functions as a
20204 configuration pragma. Note that the
20205 notion of configuration pragma corresponds closely to the
20206 HP Ada notion of a program library switch.
20208 When no pragma is used in GNAT, the default is @code{IEEE_Float},
20210 from HP Ada 83, where the default is @code{VAX_Float}. In addition, the
20211 predefined libraries in GNAT are built using @code{IEEE_Float}, so it is not
20212 advisable to change the format of numbers passed to standard library
20213 routines, and if necessary explicit type conversions may be needed.
20215 The use of @code{IEEE_Float} is recommended in GNAT since it is more
20216 efficient, and (given that it conforms to an international standard)
20217 potentially more portable.
20218 The situation in which @code{VAX_Float} may be useful is in interfacing
20219 to existing code and data that expect the use of @code{VAX_Float}.
20220 In such a situation use the predefined @code{VAX_Float}
20221 types in package @code{System}, as extended by
20222 @code{Extend_System}. For example, use @code{System.F_Float}
20223 to specify the 32-bit @code{F-Float} format.
20226 On OpenVMS systems, HP Ada provides the pragma @code{Long_Float}
20227 to allow control over the internal representation chosen
20228 for the predefined type @code{Long_Float} and for floating-point
20229 type declarations with digits specified in the range 7 .. 15.
20230 The format of this pragma is as follows:
20232 @smallexample @c ada
20234 pragma Long_Float (D_FLOAT | G_FLOAT);
20238 @node Fixed-Point Types and Representations
20239 @subsection Fixed-Point Types and Representations
20242 On HP Ada for OpenVMS Alpha systems, rounding is
20243 away from zero for both positive and negative numbers.
20244 Therefore, @code{+0.5} rounds to @code{1},
20245 and @code{-0.5} rounds to @code{-1}.
20247 On GNAT the results of operations
20248 on fixed-point types are in accordance with the Ada
20249 rules. In particular, results of operations on decimal
20250 fixed-point types are truncated.
20252 @node Record and Array Component Alignment
20253 @subsection Record and Array Component Alignment
20256 On HP Ada for OpenVMS Alpha, all non-composite components
20257 are aligned on natural boundaries. For example, 1-byte
20258 components are aligned on byte boundaries, 2-byte
20259 components on 2-byte boundaries, 4-byte components on 4-byte
20260 byte boundaries, and so on. The OpenVMS Alpha hardware
20261 runs more efficiently with naturally aligned data.
20263 On GNAT, alignment rules are compatible
20264 with HP Ada for OpenVMS Alpha.
20266 @node Address Clauses
20267 @subsection Address Clauses
20270 In HP Ada and GNAT, address clauses are supported for
20271 objects and imported subprograms.
20272 The predefined type @code{System.Address} is a private type
20273 in both compilers on Alpha OpenVMS, with the same representation
20274 (it is simply a machine pointer). Addition, subtraction, and comparison
20275 operations are available in the standard Ada package
20276 @code{System.Storage_Elements}, or in package @code{System}
20277 if it is extended to include @code{System.Aux_DEC} using a
20278 pragma @code{Extend_System} as previously described.
20280 Note that code that @code{with}'s both this extended package @code{System}
20281 and the package @code{System.Storage_Elements} should not @code{use}
20282 both packages, or ambiguities will result. In general it is better
20283 not to mix these two sets of facilities. The Ada package was
20284 designed specifically to provide the kind of features that HP Ada
20285 adds directly to package @code{System}.
20287 The type @code{System.Address} is a 64-bit integer type in GNAT for
20288 I64 OpenVMS. For more information,
20289 see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
20291 GNAT is compatible with HP Ada in its handling of address
20292 clauses, except for some limitations in
20293 the form of address clauses for composite objects with
20294 initialization. Such address clauses are easily replaced
20295 by the use of an explicitly-defined constant as described
20296 in the Ada Reference Manual (13.1(22)). For example, the sequence
20299 @smallexample @c ada
20301 X, Y : Integer := Init_Func;
20302 Q : String (X .. Y) := "abc";
20304 for Q'Address use Compute_Address;
20309 will be rejected by GNAT, since the address cannot be computed at the time
20310 that @code{Q} is declared. To achieve the intended effect, write instead:
20312 @smallexample @c ada
20315 X, Y : Integer := Init_Func;
20316 Q_Address : constant Address := Compute_Address;
20317 Q : String (X .. Y) := "abc";
20319 for Q'Address use Q_Address;
20325 which will be accepted by GNAT (and other Ada compilers), and is also
20326 compatible with Ada 83. A fuller description of the restrictions
20327 on address specifications is found in @ref{Top, GNAT Reference Manual,
20328 About This Guide, gnat_rm, GNAT Reference Manual}.
20330 @node Other Representation Clauses
20331 @subsection Other Representation Clauses
20334 GNAT implements in a compatible manner all the representation
20335 clauses supported by HP Ada. In addition, GNAT
20336 implements the representation clause forms that were introduced in Ada 95,
20337 including @code{COMPONENT_SIZE} and @code{SIZE} clauses for objects.
20339 @node The Package STANDARD
20340 @section The Package @code{STANDARD}
20343 The package @code{STANDARD}, as implemented by HP Ada, is fully
20344 described in the @cite{Ada Reference Manual} and in the
20345 @cite{HP Ada Language Reference Manual}. As implemented by GNAT, the
20346 package @code{STANDARD} is described in the @cite{Ada Reference Manual}.
20348 In addition, HP Ada supports the Latin-1 character set in
20349 the type @code{CHARACTER}. GNAT supports the Latin-1 character set
20350 in the type @code{CHARACTER} and also Unicode (ISO 10646 BMP) in
20351 the type @code{WIDE_CHARACTER}.
20353 The floating-point types supported by GNAT are those
20354 supported by HP Ada, but the defaults are different, and are controlled by
20355 pragmas. See @ref{Floating-Point Types and Representations}, for details.
20357 @node The Package SYSTEM
20358 @section The Package @code{SYSTEM}
20361 HP Ada provides a specific version of the package
20362 @code{SYSTEM} for each platform on which the language is implemented.
20363 For the complete spec of the package @code{SYSTEM}, see
20364 Appendix F of the @cite{HP Ada Language Reference Manual}.
20366 On HP Ada, the package @code{SYSTEM} includes the following conversion
20369 @item @code{TO_ADDRESS(INTEGER)}
20371 @item @code{TO_ADDRESS(UNSIGNED_LONGWORD)}
20373 @item @code{TO_ADDRESS(}@i{universal_integer}@code{)}
20375 @item @code{TO_INTEGER(ADDRESS)}
20377 @item @code{TO_UNSIGNED_LONGWORD(ADDRESS)}
20379 @item Function @code{IMPORT_VALUE return UNSIGNED_LONGWORD} and the
20380 functions @code{IMPORT_ADDRESS} and @code{IMPORT_LARGEST_VALUE}
20384 By default, GNAT supplies a version of @code{SYSTEM} that matches
20385 the definition given in the @cite{Ada Reference Manual}.
20387 is a subset of the HP system definitions, which is as
20388 close as possible to the original definitions. The only difference
20389 is that the definition of @code{SYSTEM_NAME} is different:
20391 @smallexample @c ada
20393 type Name is (SYSTEM_NAME_GNAT);
20394 System_Name : constant Name := SYSTEM_NAME_GNAT;
20399 Also, GNAT adds the Ada declarations for
20400 @code{BIT_ORDER} and @code{DEFAULT_BIT_ORDER}.
20402 However, the use of the following pragma causes GNAT
20403 to extend the definition of package @code{SYSTEM} so that it
20404 encompasses the full set of HP-specific extensions,
20405 including the functions listed above:
20407 @smallexample @c ada
20409 pragma Extend_System (Aux_DEC);
20414 The pragma @code{Extend_System} is a configuration pragma that
20415 is most conveniently placed in the @file{gnat.adc} file. @xref{Pragma
20416 Extend_System,,, gnat_rm, GNAT Reference Manual}, for further details.
20418 HP Ada does not allow the recompilation of the package
20419 @code{SYSTEM}. Instead HP Ada provides several pragmas
20420 (@code{SYSTEM_NAME}, @code{STORAGE_UNIT}, and @code{MEMORY_SIZE})
20421 to modify values in the package @code{SYSTEM}.
20422 On OpenVMS Alpha systems, the pragma
20423 @code{SYSTEM_NAME} takes the enumeration literal @code{OPENVMS_AXP} as
20424 its single argument.
20426 GNAT does permit the recompilation of package @code{SYSTEM} using
20427 the special switch @option{-gnatg}, and this switch can be used if
20428 it is necessary to modify the definitions in @code{SYSTEM}. GNAT does
20429 not permit the specification of @code{SYSTEM_NAME}, @code{STORAGE_UNIT}
20430 or @code{MEMORY_SIZE} by any other means.
20432 On GNAT systems, the pragma @code{SYSTEM_NAME} takes the
20433 enumeration literal @code{SYSTEM_NAME_GNAT}.
20435 The definitions provided by the use of
20437 @smallexample @c ada
20438 pragma Extend_System (AUX_Dec);
20442 are virtually identical to those provided by the HP Ada 83 package
20443 @code{SYSTEM}. One important difference is that the name of the
20445 function for type @code{UNSIGNED_LONGWORD} is changed to
20446 @code{TO_ADDRESS_LONG}.
20447 @xref{Address Clauses,,, gnat_rm, GNAT Reference Manual}, for a
20448 discussion of why this change was necessary.
20451 The version of @code{TO_ADDRESS} taking a @i{universal_integer} argument
20453 an extension to Ada 83 not strictly compatible with the reference manual.
20454 GNAT, in order to be exactly compatible with the standard,
20455 does not provide this capability. In HP Ada 83, the
20456 point of this definition is to deal with a call like:
20458 @smallexample @c ada
20459 TO_ADDRESS (16#12777#);
20463 Normally, according to Ada 83 semantics, one would expect this to be
20464 ambiguous, since it matches both the @code{INTEGER} and
20465 @code{UNSIGNED_LONGWORD} forms of @code{TO_ADDRESS}.
20466 However, in HP Ada 83, there is no ambiguity, since the
20467 definition using @i{universal_integer} takes precedence.
20469 In GNAT, since the version with @i{universal_integer} cannot be supplied,
20471 not possible to be 100% compatible. Since there are many programs using
20472 numeric constants for the argument to @code{TO_ADDRESS}, the decision in
20474 to change the name of the function in the @code{UNSIGNED_LONGWORD} case,
20475 so the declarations provided in the GNAT version of @code{AUX_Dec} are:
20477 @smallexample @c ada
20478 function To_Address (X : Integer) return Address;
20479 pragma Pure_Function (To_Address);
20481 function To_Address_Long (X : Unsigned_Longword) return Address;
20482 pragma Pure_Function (To_Address_Long);
20486 This means that programs using @code{TO_ADDRESS} for
20487 @code{UNSIGNED_LONGWORD} must change the name to @code{TO_ADDRESS_LONG}.
20489 @node Tasking and Task-Related Features
20490 @section Tasking and Task-Related Features
20493 This section compares the treatment of tasking in GNAT
20494 and in HP Ada for OpenVMS Alpha.
20495 The GNAT description applies to both Alpha and I64 OpenVMS.
20496 For detailed information on tasking in
20497 HP Ada, see the @cite{HP Ada Language Reference Manual} and the
20498 relevant run-time reference manual.
20501 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
20502 * Assigning Task IDs::
20503 * Task IDs and Delays::
20504 * Task-Related Pragmas::
20505 * Scheduling and Task Priority::
20507 * External Interrupts::
20510 @node Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
20511 @subsection Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
20514 On OpenVMS Alpha systems, each Ada task (except a passive
20515 task) is implemented as a single stream of execution
20516 that is created and managed by the kernel. On these
20517 systems, HP Ada tasking support is based on DECthreads,
20518 an implementation of the POSIX standard for threads.
20520 Also, on OpenVMS Alpha systems, HP Ada tasks and foreign
20521 code that calls DECthreads routines can be used together.
20522 The interaction between Ada tasks and DECthreads routines
20523 can have some benefits. For example when on OpenVMS Alpha,
20524 HP Ada can call C code that is already threaded.
20526 GNAT uses the facilities of DECthreads,
20527 and Ada tasks are mapped to threads.
20529 @node Assigning Task IDs
20530 @subsection Assigning Task IDs
20533 The HP Ada Run-Time Library always assigns @code{%TASK 1} to
20534 the environment task that executes the main program. On
20535 OpenVMS Alpha systems, @code{%TASK 0} is often used for tasks
20536 that have been created but are not yet activated.
20538 On OpenVMS Alpha systems, task IDs are assigned at
20539 activation. On GNAT systems, task IDs are also assigned at
20540 task creation but do not have the same form or values as
20541 task ID values in HP Ada. There is no null task, and the
20542 environment task does not have a specific task ID value.
20544 @node Task IDs and Delays
20545 @subsection Task IDs and Delays
20548 On OpenVMS Alpha systems, tasking delays are implemented
20549 using Timer System Services. The Task ID is used for the
20550 identification of the timer request (the @code{REQIDT} parameter).
20551 If Timers are used in the application take care not to use
20552 @code{0} for the identification, because cancelling such a timer
20553 will cancel all timers and may lead to unpredictable results.
20555 @node Task-Related Pragmas
20556 @subsection Task-Related Pragmas
20559 Ada supplies the pragma @code{TASK_STORAGE}, which allows
20560 specification of the size of the guard area for a task
20561 stack. (The guard area forms an area of memory that has no
20562 read or write access and thus helps in the detection of
20563 stack overflow.) On OpenVMS Alpha systems, if the pragma
20564 @code{TASK_STORAGE} specifies a value of zero, a minimal guard
20565 area is created. In the absence of a pragma @code{TASK_STORAGE},
20566 a default guard area is created.
20568 GNAT supplies the following task-related pragmas:
20571 @item @code{TASK_INFO}
20573 This pragma appears within a task definition and
20574 applies to the task in which it appears. The argument
20575 must be of type @code{SYSTEM.TASK_INFO.TASK_INFO_TYPE}.
20577 @item @code{TASK_STORAGE}
20579 GNAT implements pragma @code{TASK_STORAGE} in the same way as HP Ada.
20580 Both HP Ada and GNAT supply the pragmas @code{PASSIVE},
20581 @code{SUPPRESS}, and @code{VOLATILE}.
20583 @node Scheduling and Task Priority
20584 @subsection Scheduling and Task Priority
20587 HP Ada implements the Ada language requirement that
20588 when two tasks are eligible for execution and they have
20589 different priorities, the lower priority task does not
20590 execute while the higher priority task is waiting. The HP
20591 Ada Run-Time Library keeps a task running until either the
20592 task is suspended or a higher priority task becomes ready.
20594 On OpenVMS Alpha systems, the default strategy is round-
20595 robin with preemption. Tasks of equal priority take turns
20596 at the processor. A task is run for a certain period of
20597 time and then placed at the tail of the ready queue for
20598 its priority level.
20600 HP Ada provides the implementation-defined pragma @code{TIME_SLICE},
20601 which can be used to enable or disable round-robin
20602 scheduling of tasks with the same priority.
20603 See the relevant HP Ada run-time reference manual for
20604 information on using the pragmas to control HP Ada task
20607 GNAT follows the scheduling rules of Annex D (Real-Time
20608 Annex) of the @cite{Ada Reference Manual}. In general, this
20609 scheduling strategy is fully compatible with HP Ada
20610 although it provides some additional constraints (as
20611 fully documented in Annex D).
20612 GNAT implements time slicing control in a manner compatible with
20613 HP Ada 83, by means of the pragma @code{Time_Slice}, whose semantics
20614 are identical to the HP Ada 83 pragma of the same name.
20615 Note that it is not possible to mix GNAT tasking and
20616 HP Ada 83 tasking in the same program, since the two run-time
20617 libraries are not compatible.
20619 @node The Task Stack
20620 @subsection The Task Stack
20623 In HP Ada, a task stack is allocated each time a
20624 non-passive task is activated. As soon as the task is
20625 terminated, the storage for the task stack is deallocated.
20626 If you specify a size of zero (bytes) with @code{T'STORAGE_SIZE},
20627 a default stack size is used. Also, regardless of the size
20628 specified, some additional space is allocated for task
20629 management purposes. On OpenVMS Alpha systems, at least
20630 one page is allocated.
20632 GNAT handles task stacks in a similar manner. In accordance with
20633 the Ada rules, it provides the pragma @code{STORAGE_SIZE} as
20634 an alternative method for controlling the task stack size.
20635 The specification of the attribute @code{T'STORAGE_SIZE} is also
20636 supported in a manner compatible with HP Ada.
20638 @node External Interrupts
20639 @subsection External Interrupts
20642 On HP Ada, external interrupts can be associated with task entries.
20643 GNAT is compatible with HP Ada in its handling of external interrupts.
20645 @node Pragmas and Pragma-Related Features
20646 @section Pragmas and Pragma-Related Features
20649 Both HP Ada and GNAT supply all language-defined pragmas
20650 as specified by the Ada 83 standard. GNAT also supplies all
20651 language-defined pragmas introduced by Ada 95 and Ada 2005.
20652 In addition, GNAT implements the implementation-defined pragmas
20656 @item @code{AST_ENTRY}
20658 @item @code{COMMON_OBJECT}
20660 @item @code{COMPONENT_ALIGNMENT}
20662 @item @code{EXPORT_EXCEPTION}
20664 @item @code{EXPORT_FUNCTION}
20666 @item @code{EXPORT_OBJECT}
20668 @item @code{EXPORT_PROCEDURE}
20670 @item @code{EXPORT_VALUED_PROCEDURE}
20672 @item @code{FLOAT_REPRESENTATION}
20676 @item @code{IMPORT_EXCEPTION}
20678 @item @code{IMPORT_FUNCTION}
20680 @item @code{IMPORT_OBJECT}
20682 @item @code{IMPORT_PROCEDURE}
20684 @item @code{IMPORT_VALUED_PROCEDURE}
20686 @item @code{INLINE_GENERIC}
20688 @item @code{INTERFACE_NAME}
20690 @item @code{LONG_FLOAT}
20692 @item @code{MAIN_STORAGE}
20694 @item @code{PASSIVE}
20696 @item @code{PSECT_OBJECT}
20698 @item @code{SHARE_GENERIC}
20700 @item @code{SUPPRESS_ALL}
20702 @item @code{TASK_STORAGE}
20704 @item @code{TIME_SLICE}
20710 These pragmas are all fully implemented, with the exception of @code{TITLE},
20711 @code{PASSIVE}, and @code{SHARE_GENERIC}, which are
20712 recognized, but which have no
20713 effect in GNAT. The effect of @code{PASSIVE} may be obtained by the
20714 use of Ada protected objects. In GNAT, all generics are inlined.
20716 Unlike HP Ada, the GNAT ``@code{EXPORT_}@i{subprogram}'' pragmas require
20717 a separate subprogram specification which must appear before the
20720 GNAT also supplies a number of implementation-defined pragmas including the
20724 @item @code{ABORT_DEFER}
20726 @item @code{ADA_83}
20728 @item @code{ADA_95}
20730 @item @code{ADA_05}
20732 @item @code{Ada_2005}
20734 @item @code{Ada_12}
20736 @item @code{Ada_2012}
20738 @item @code{ANNOTATE}
20740 @item @code{ASSERT}
20742 @item @code{C_PASS_BY_COPY}
20744 @item @code{CPP_CLASS}
20746 @item @code{CPP_CONSTRUCTOR}
20748 @item @code{CPP_DESTRUCTOR}
20752 @item @code{EXTEND_SYSTEM}
20754 @item @code{LINKER_ALIAS}
20756 @item @code{LINKER_SECTION}
20758 @item @code{MACHINE_ATTRIBUTE}
20760 @item @code{NO_RETURN}
20762 @item @code{PURE_FUNCTION}
20764 @item @code{SOURCE_FILE_NAME}
20766 @item @code{SOURCE_REFERENCE}
20768 @item @code{TASK_INFO}
20770 @item @code{UNCHECKED_UNION}
20772 @item @code{UNIMPLEMENTED_UNIT}
20774 @item @code{UNIVERSAL_DATA}
20776 @item @code{UNSUPPRESS}
20778 @item @code{WARNINGS}
20780 @item @code{WEAK_EXTERNAL}
20784 For full details on these and other GNAT implementation-defined pragmas,
20785 see @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT Reference
20789 * Restrictions on the Pragma INLINE::
20790 * Restrictions on the Pragma INTERFACE::
20791 * Restrictions on the Pragma SYSTEM_NAME::
20794 @node Restrictions on the Pragma INLINE
20795 @subsection Restrictions on Pragma @code{INLINE}
20798 HP Ada enforces the following restrictions on the pragma @code{INLINE}:
20800 @item Parameters cannot have a task type.
20802 @item Function results cannot be task types, unconstrained
20803 array types, or unconstrained types with discriminants.
20805 @item Bodies cannot declare the following:
20807 @item Subprogram body or stub (imported subprogram is allowed)
20811 @item Generic declarations
20813 @item Instantiations
20817 @item Access types (types derived from access types allowed)
20819 @item Array or record types
20821 @item Dependent tasks
20823 @item Direct recursive calls of subprogram or containing
20824 subprogram, directly or via a renaming
20830 In GNAT, the only restriction on pragma @code{INLINE} is that the
20831 body must occur before the call if both are in the same
20832 unit, and the size must be appropriately small. There are
20833 no other specific restrictions which cause subprograms to
20834 be incapable of being inlined.
20836 @node Restrictions on the Pragma INTERFACE
20837 @subsection Restrictions on Pragma @code{INTERFACE}
20840 The following restrictions on pragma @code{INTERFACE}
20841 are enforced by both HP Ada and GNAT:
20843 @item Languages accepted: Ada, Bliss, C, Fortran, Default.
20844 Default is the default on OpenVMS Alpha systems.
20846 @item Parameter passing: Language specifies default
20847 mechanisms but can be overridden with an @code{EXPORT} pragma.
20850 @item Ada: Use internal Ada rules.
20852 @item Bliss, C: Parameters must be mode @code{in}; cannot be
20853 record or task type. Result cannot be a string, an
20854 array, or a record.
20856 @item Fortran: Parameters cannot have a task type. Result cannot
20857 be a string, an array, or a record.
20862 GNAT is entirely upwards compatible with HP Ada, and in addition allows
20863 record parameters for all languages.
20865 @node Restrictions on the Pragma SYSTEM_NAME
20866 @subsection Restrictions on Pragma @code{SYSTEM_NAME}
20869 For HP Ada for OpenVMS Alpha, the enumeration literal
20870 for the type @code{NAME} is @code{OPENVMS_AXP}.
20871 In GNAT, the enumeration
20872 literal for the type @code{NAME} is @code{SYSTEM_NAME_GNAT}.
20874 @node Library of Predefined Units
20875 @section Library of Predefined Units
20878 A library of predefined units is provided as part of the
20879 HP Ada and GNAT implementations. HP Ada does not provide
20880 the package @code{MACHINE_CODE} but instead recommends importing
20883 The GNAT versions of the HP Ada Run-Time Library (@code{ADA$PREDEFINED:})
20884 units are taken from the OpenVMS Alpha version, not the OpenVMS VAX
20886 The HP Ada Predefined Library units are modified to remove post-Ada 83
20887 incompatibilities and to make them interoperable with GNAT
20888 (@pxref{Changes to DECLIB}, for details).
20889 The units are located in the @file{DECLIB} directory.
20891 The GNAT RTL is contained in
20892 the @file{ADALIB} directory, and
20893 the default search path is set up to find @code{DECLIB} units in preference
20894 to @code{ADALIB} units with the same name (@code{TEXT_IO},
20895 @code{SEQUENTIAL_IO}, and @code{DIRECT_IO}, for example).
20898 * Changes to DECLIB::
20901 @node Changes to DECLIB
20902 @subsection Changes to @code{DECLIB}
20905 The changes made to the HP Ada predefined library for GNAT and post-Ada 83
20906 compatibility are minor and include the following:
20909 @item Adjusting the location of pragmas and record representation
20910 clauses to obey Ada 95 (and thus Ada 2005) rules
20912 @item Adding the proper notation to generic formal parameters
20913 that take unconstrained types in instantiation
20915 @item Adding pragma @code{ELABORATE_BODY} to package specs
20916 that have package bodies not otherwise allowed
20918 @item Replacing occurrences of the identifier ``@code{PROTECTED}'' by
20919 ``@code{PROTECTD}''.
20920 Currently these are found only in the @code{STARLET} package spec.
20922 @item Changing @code{SYSTEM.ADDRESS} to @code{SYSTEM.SHORT_ADDRESS}
20923 where the address size is constrained to 32 bits.
20927 None of the above changes is visible to users.
20933 On OpenVMS Alpha, HP Ada provides the following strongly-typed bindings:
20936 @item Command Language Interpreter (CLI interface)
20938 @item DECtalk Run-Time Library (DTK interface)
20940 @item Librarian utility routines (LBR interface)
20942 @item General Purpose Run-Time Library (LIB interface)
20944 @item Math Run-Time Library (MTH interface)
20946 @item National Character Set Run-Time Library (NCS interface)
20948 @item Compiled Code Support Run-Time Library (OTS interface)
20950 @item Parallel Processing Run-Time Library (PPL interface)
20952 @item Screen Management Run-Time Library (SMG interface)
20954 @item Sort Run-Time Library (SOR interface)
20956 @item String Run-Time Library (STR interface)
20958 @item STARLET System Library
20961 @item X Window System Version 11R4 and 11R5 (X, XLIB interface)
20963 @item X Windows Toolkit (XT interface)
20965 @item X/Motif Version 1.1.3 and 1.2 (XM interface)
20969 GNAT provides implementations of these HP bindings in the @code{DECLIB}
20970 directory, on both the Alpha and I64 OpenVMS platforms.
20972 The X/Motif bindings used to build @code{DECLIB} are whatever versions are
20974 HP Ada @file{ADA$PREDEFINED} directory with extension @file{.ADC}.
20975 A pragma @code{Linker_Options} has been added to packages @code{Xm},
20976 @code{Xt}, and @code{X_Lib}
20977 causing the default X/Motif sharable image libraries to be linked in. This
20978 is done via options files named @file{xm.opt}, @file{xt.opt}, and
20979 @file{x_lib.opt} (also located in the @file{DECLIB} directory).
20981 It may be necessary to edit these options files to update or correct the
20982 library names if, for example, the newer X/Motif bindings from
20983 @file{ADA$EXAMPLES}
20984 had been (previous to installing GNAT) copied and renamed to supersede the
20985 default @file{ADA$PREDEFINED} versions.
20988 * Shared Libraries and Options Files::
20989 * Interfaces to C::
20992 @node Shared Libraries and Options Files
20993 @subsection Shared Libraries and Options Files
20996 When using the HP Ada
20997 predefined X and Motif bindings, the linking with their sharable images is
20998 done automatically by @command{GNAT LINK}.
20999 When using other X and Motif bindings, you need
21000 to add the corresponding sharable images to the command line for
21001 @code{GNAT LINK}. When linking with shared libraries, or with
21002 @file{.OPT} files, you must
21003 also add them to the command line for @command{GNAT LINK}.
21005 A shared library to be used with GNAT is built in the same way as other
21006 libraries under VMS. The VMS Link command can be used in standard fashion.
21008 @node Interfaces to C
21009 @subsection Interfaces to C
21013 provides the following Ada types and operations:
21016 @item C types package (@code{C_TYPES})
21018 @item C strings (@code{C_TYPES.NULL_TERMINATED})
21020 @item Other_types (@code{SHORT_INT})
21024 Interfacing to C with GNAT, you can use the above approach
21025 described for HP Ada or the facilities of Annex B of
21026 the @cite{Ada Reference Manual} (packages @code{INTERFACES.C},
21027 @code{INTERFACES.C.STRINGS} and @code{INTERFACES.C.POINTERS}). For more
21028 information, see @ref{Interfacing to C,,, gnat_rm, GNAT Reference Manual}.
21030 The @option{-gnatF} qualifier forces default and explicit
21031 @code{External_Name} parameters in pragmas @code{Import} and @code{Export}
21032 to be uppercased for compatibility with the default behavior
21033 of HP C. The qualifier has no effect on @code{Link_Name} parameters.
21035 @node Main Program Definition
21036 @section Main Program Definition
21039 The following section discusses differences in the
21040 definition of main programs on HP Ada and GNAT.
21041 On HP Ada, main programs are defined to meet the
21042 following conditions:
21044 @item Procedure with no formal parameters (returns @code{0} upon
21047 @item Procedure with no formal parameters (returns @code{42} when
21048 an unhandled exception is raised)
21050 @item Function with no formal parameters whose returned value
21051 is of a discrete type
21053 @item Procedure with one @code{out} formal of a discrete type for
21054 which a specification of pragma @code{EXPORT_VALUED_PROCEDURE} is given.
21059 When declared with the pragma @code{EXPORT_VALUED_PROCEDURE},
21060 a main function or main procedure returns a discrete
21061 value whose size is less than 64 bits (32 on VAX systems),
21062 the value is zero- or sign-extended as appropriate.
21063 On GNAT, main programs are defined as follows:
21065 @item Must be a non-generic, parameterless subprogram that
21066 is either a procedure or function returning an Ada
21067 @code{STANDARD.INTEGER} (the predefined type)
21069 @item Cannot be a generic subprogram or an instantiation of a
21073 @node Implementation-Defined Attributes
21074 @section Implementation-Defined Attributes
21077 GNAT provides all HP Ada implementation-defined
21080 @node Compiler and Run-Time Interfacing
21081 @section Compiler and Run-Time Interfacing
21084 HP Ada provides the following qualifiers to pass options to the linker
21087 @item @option{/WAIT} and @option{/SUBMIT}
21089 @item @option{/COMMAND}
21091 @item @option{/@r{[}NO@r{]}MAP}
21093 @item @option{/OUTPUT=@var{file-spec}}
21095 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
21099 To pass options to the linker, GNAT provides the following
21103 @item @option{/EXECUTABLE=@var{exec-name}}
21105 @item @option{/VERBOSE}
21107 @item @option{/@r{[}NO@r{]}DEBUG} and @option{/@r{[}NO@r{]}TRACEBACK}
21111 For more information on these switches, see
21112 @ref{Switches for gnatlink}.
21113 In HP Ada, the command-line switch @option{/OPTIMIZE} is available
21114 to control optimization. HP Ada also supplies the
21117 @item @code{OPTIMIZE}
21119 @item @code{INLINE}
21121 @item @code{INLINE_GENERIC}
21123 @item @code{SUPPRESS_ALL}
21125 @item @code{PASSIVE}
21129 In GNAT, optimization is controlled strictly by command
21130 line parameters, as described in the corresponding section of this guide.
21131 The HP pragmas for control of optimization are
21132 recognized but ignored.
21134 Note that in GNAT, the default is optimization off, whereas in HP Ada
21135 the default is that optimization is turned on.
21137 @node Program Compilation and Library Management
21138 @section Program Compilation and Library Management
21141 HP Ada and GNAT provide a comparable set of commands to
21142 build programs. HP Ada also provides a program library,
21143 which is a concept that does not exist on GNAT. Instead,
21144 GNAT provides directories of sources that are compiled as
21147 The following table summarizes
21148 the HP Ada commands and provides
21149 equivalent GNAT commands. In this table, some GNAT
21150 equivalents reflect the fact that GNAT does not use the
21151 concept of a program library. Instead, it uses a model
21152 in which collections of source and object files are used
21153 in a manner consistent with other languages like C and
21154 Fortran. Therefore, standard system file commands are used
21155 to manipulate these elements. Those GNAT commands are marked with
21157 Note that, unlike HP Ada, none of the GNAT commands accepts wild cards.
21160 @multitable @columnfractions .35 .65
21162 @item @emph{HP Ada Command}
21163 @tab @emph{GNAT Equivalent / Description}
21165 @item @command{ADA}
21166 @tab @command{GNAT COMPILE}@*
21167 Invokes the compiler to compile one or more Ada source files.
21169 @item @command{ACS ATTACH}@*
21170 @tab [No equivalent]@*
21171 Switches control of terminal from current process running the program
21174 @item @command{ACS CHECK}
21175 @tab @command{GNAT MAKE /DEPENDENCY_LIST}@*
21176 Forms the execution closure of one
21177 or more compiled units and checks completeness and currency.
21179 @item @command{ACS COMPILE}
21180 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
21181 Forms the execution closure of one or
21182 more specified units, checks completeness and currency,
21183 identifies units that have revised source files, compiles same,
21184 and recompiles units that are or will become obsolete.
21185 Also completes incomplete generic instantiations.
21187 @item @command{ACS COPY FOREIGN}
21189 Copies a foreign object file into the program library as a
21192 @item @command{ACS COPY UNIT}
21194 Copies a compiled unit from one program library to another.
21196 @item @command{ACS CREATE LIBRARY}
21197 @tab Create /directory (*)@*
21198 Creates a program library.
21200 @item @command{ACS CREATE SUBLIBRARY}
21201 @tab Create /directory (*)@*
21202 Creates a program sublibrary.
21204 @item @command{ACS DELETE LIBRARY}
21206 Deletes a program library and its contents.
21208 @item @command{ACS DELETE SUBLIBRARY}
21210 Deletes a program sublibrary and its contents.
21212 @item @command{ACS DELETE UNIT}
21213 @tab Delete file (*)@*
21214 On OpenVMS systems, deletes one or more compiled units from
21215 the current program library.
21217 @item @command{ACS DIRECTORY}
21218 @tab Directory (*)@*
21219 On OpenVMS systems, lists units contained in the current
21222 @item @command{ACS ENTER FOREIGN}
21224 Allows the import of a foreign body as an Ada library
21225 spec and enters a reference to a pointer.
21227 @item @command{ACS ENTER UNIT}
21229 Enters a reference (pointer) from the current program library to
21230 a unit compiled into another program library.
21232 @item @command{ACS EXIT}
21233 @tab [No equivalent]@*
21234 Exits from the program library manager.
21236 @item @command{ACS EXPORT}
21238 Creates an object file that contains system-specific object code
21239 for one or more units. With GNAT, object files can simply be copied
21240 into the desired directory.
21242 @item @command{ACS EXTRACT SOURCE}
21244 Allows access to the copied source file for each Ada compilation unit
21246 @item @command{ACS HELP}
21247 @tab @command{HELP GNAT}@*
21248 Provides online help.
21250 @item @command{ACS LINK}
21251 @tab @command{GNAT LINK}@*
21252 Links an object file containing Ada units into an executable file.
21254 @item @command{ACS LOAD}
21256 Loads (partially compiles) Ada units into the program library.
21257 Allows loading a program from a collection of files into a library
21258 without knowing the relationship among units.
21260 @item @command{ACS MERGE}
21262 Merges into the current program library, one or more units from
21263 another library where they were modified.
21265 @item @command{ACS RECOMPILE}
21266 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
21267 Recompiles from external or copied source files any obsolete
21268 unit in the closure. Also, completes any incomplete generic
21271 @item @command{ACS REENTER}
21272 @tab @command{GNAT MAKE}@*
21273 Reenters current references to units compiled after last entered
21274 with the @command{ACS ENTER UNIT} command.
21276 @item @command{ACS SET LIBRARY}
21277 @tab Set default (*)@*
21278 Defines a program library to be the compilation context as well
21279 as the target library for compiler output and commands in general.
21281 @item @command{ACS SET PRAGMA}
21282 @tab Edit @file{gnat.adc} (*)@*
21283 Redefines specified values of the library characteristics
21284 @code{LONG_ FLOAT}, @code{MEMORY_SIZE}, @code{SYSTEM_NAME},
21285 and @code{Float_Representation}.
21287 @item @command{ACS SET SOURCE}
21288 @tab Define @code{ADA_INCLUDE_PATH} path (*)@*
21289 Defines the source file search list for the @command{ACS COMPILE} command.
21291 @item @command{ACS SHOW LIBRARY}
21292 @tab Directory (*)@*
21293 Lists information about one or more program libraries.
21295 @item @command{ACS SHOW PROGRAM}
21296 @tab [No equivalent]@*
21297 Lists information about the execution closure of one or
21298 more units in the program library.
21300 @item @command{ACS SHOW SOURCE}
21301 @tab Show logical @code{ADA_INCLUDE_PATH}@*
21302 Shows the source file search used when compiling units.
21304 @item @command{ACS SHOW VERSION}
21305 @tab Compile with @option{VERBOSE} option
21306 Displays the version number of the compiler and program library
21309 @item @command{ACS SPAWN}
21310 @tab [No equivalent]@*
21311 Creates a subprocess of the current process (same as @command{DCL SPAWN}
21314 @item @command{ACS VERIFY}
21315 @tab [No equivalent]@*
21316 Performs a series of consistency checks on a program library to
21317 determine whether the library structure and library files are in
21324 @section Input-Output
21327 On OpenVMS Alpha systems, HP Ada uses OpenVMS Record
21328 Management Services (RMS) to perform operations on
21332 HP Ada and GNAT predefine an identical set of input-
21333 output packages. To make the use of the
21334 generic @code{TEXT_IO} operations more convenient, HP Ada
21335 provides predefined library packages that instantiate the
21336 integer and floating-point operations for the predefined
21337 integer and floating-point types as shown in the following table.
21339 @multitable @columnfractions .45 .55
21340 @item @emph{Package Name} @tab Instantiation
21342 @item @code{INTEGER_TEXT_IO}
21343 @tab @code{INTEGER_IO(INTEGER)}
21345 @item @code{SHORT_INTEGER_TEXT_IO}
21346 @tab @code{INTEGER_IO(SHORT_INTEGER)}
21348 @item @code{SHORT_SHORT_INTEGER_TEXT_IO}
21349 @tab @code{INTEGER_IO(SHORT_SHORT_INTEGER)}
21351 @item @code{FLOAT_TEXT_IO}
21352 @tab @code{FLOAT_IO(FLOAT)}
21354 @item @code{LONG_FLOAT_TEXT_IO}
21355 @tab @code{FLOAT_IO(LONG_FLOAT)}
21359 The HP Ada predefined packages and their operations
21360 are implemented using OpenVMS Alpha files and input-output
21361 facilities. HP Ada supports asynchronous input-output on OpenVMS Alpha.
21362 Familiarity with the following is recommended:
21364 @item RMS file organizations and access methods
21366 @item OpenVMS file specifications and directories
21368 @item OpenVMS File Definition Language (FDL)
21372 GNAT provides I/O facilities that are completely
21373 compatible with HP Ada. The distribution includes the
21374 standard HP Ada versions of all I/O packages, operating
21375 in a manner compatible with HP Ada. In particular, the
21376 following packages are by default the HP Ada (Ada 83)
21377 versions of these packages rather than the renamings
21378 suggested in Annex J of the Ada Reference Manual:
21380 @item @code{TEXT_IO}
21382 @item @code{SEQUENTIAL_IO}
21384 @item @code{DIRECT_IO}
21388 The use of the standard child package syntax (for
21389 example, @code{ADA.TEXT_IO}) retrieves the post-Ada 83 versions of these
21391 GNAT provides HP-compatible predefined instantiations
21392 of the @code{TEXT_IO} packages, and also
21393 provides the standard predefined instantiations required
21394 by the @cite{Ada Reference Manual}.
21396 For further information on how GNAT interfaces to the file
21397 system or how I/O is implemented in programs written in
21398 mixed languages, see @ref{Implementation of the Standard I/O,,,
21399 gnat_rm, GNAT Reference Manual}.
21400 This chapter covers the following:
21402 @item Standard I/O packages
21404 @item @code{FORM} strings
21406 @item @code{ADA.DIRECT_IO}
21408 @item @code{ADA.SEQUENTIAL_IO}
21410 @item @code{ADA.TEXT_IO}
21412 @item Stream pointer positioning
21414 @item Reading and writing non-regular files
21416 @item @code{GET_IMMEDIATE}
21418 @item Treating @code{TEXT_IO} files as streams
21425 @node Implementation Limits
21426 @section Implementation Limits
21429 The following table lists implementation limits for HP Ada
21431 @multitable @columnfractions .60 .20 .20
21433 @item @emph{Compilation Parameter}
21438 @item In a subprogram or entry declaration, maximum number of
21439 formal parameters that are of an unconstrained record type
21444 @item Maximum identifier length (number of characters)
21449 @item Maximum number of characters in a source line
21454 @item Maximum collection size (number of bytes)
21459 @item Maximum number of discriminants for a record type
21464 @item Maximum number of formal parameters in an entry or
21465 subprogram declaration
21470 @item Maximum number of dimensions in an array type
21475 @item Maximum number of library units and subunits in a compilation.
21480 @item Maximum number of library units and subunits in an execution.
21485 @item Maximum number of objects declared with the pragma @code{COMMON_OBJECT}
21486 or @code{PSECT_OBJECT}
21491 @item Maximum number of enumeration literals in an enumeration type
21497 @item Maximum number of lines in a source file
21502 @item Maximum number of bits in any object
21507 @item Maximum size of the static portion of a stack frame (approximate)
21512 @node Tools and Utilities
21513 @section Tools and Utilities
21516 The following table lists some of the OpenVMS development tools
21517 available for HP Ada, and the corresponding tools for
21518 use with @value{EDITION} on Alpha and I64 platforms.
21519 Aside from the debugger, all the OpenVMS tools identified are part
21520 of the DECset package.
21523 @c Specify table in TeX since Texinfo does a poor job
21527 \settabs\+Language-Sensitive Editor\quad
21528 &Product with HP Ada\quad
21531 &\it Product with HP Ada
21532 & \it Product with GNAT Pro\cr
21534 \+Code Management System
21538 \+Language-Sensitive Editor
21540 & emacs or HP LSE (Alpha)\cr
21550 & OpenVMS Debug (I64)\cr
21552 \+Source Code Analyzer /
21569 \+Coverage Analyzer
21573 \+Module Management
21575 & Not applicable\cr
21585 @c This is the Texinfo version of the table. It renders poorly in pdf, hence
21586 @c the TeX version above for the printed version
21588 @c @multitable @columnfractions .3 .4 .4
21589 @multitable {Source Code Analyzer /}{Tool with HP Ada}{Tool with GNAT Pro}
21591 @tab @i{Tool with HP Ada}
21592 @tab @i{Tool with @value{EDITION}}
21593 @item Code Management@*System
21596 @item Language-Sensitive@*Editor
21598 @tab emacs or HP LSE (Alpha)
21607 @tab OpenVMS Debug (I64)
21608 @item Source Code Analyzer /@*Cross Referencer
21612 @tab HP Digital Test@*Manager (DTM)
21614 @item Performance and@*Coverage Analyzer
21617 @item Module Management@*System
21619 @tab Not applicable
21626 @c **************************************
21627 @node Platform-Specific Information for the Run-Time Libraries
21628 @appendix Platform-Specific Information for the Run-Time Libraries
21629 @cindex Tasking and threads libraries
21630 @cindex Threads libraries and tasking
21631 @cindex Run-time libraries (platform-specific information)
21634 The GNAT run-time implementation may vary with respect to both the
21635 underlying threads library and the exception handling scheme.
21636 For threads support, one or more of the following are supplied:
21638 @item @b{native threads library}, a binding to the thread package from
21639 the underlying operating system
21641 @item @b{pthreads library} (Sparc Solaris only), a binding to the Solaris
21642 POSIX thread package
21646 For exception handling, either or both of two models are supplied:
21648 @item @b{Zero-Cost Exceptions} (``ZCX''),@footnote{
21649 Most programs should experience a substantial speed improvement by
21650 being compiled with a ZCX run-time.
21651 This is especially true for
21652 tasking applications or applications with many exception handlers.}
21653 @cindex Zero-Cost Exceptions
21654 @cindex ZCX (Zero-Cost Exceptions)
21655 which uses binder-generated tables that
21656 are interrogated at run time to locate a handler
21658 @item @b{setjmp / longjmp} (``SJLJ''),
21659 @cindex setjmp/longjmp Exception Model
21660 @cindex SJLJ (setjmp/longjmp Exception Model)
21661 which uses dynamically-set data to establish
21662 the set of handlers
21666 This appendix summarizes which combinations of threads and exception support
21667 are supplied on various GNAT platforms.
21668 It then shows how to select a particular library either
21669 permanently or temporarily,
21670 explains the properties of (and tradeoffs among) the various threads
21671 libraries, and provides some additional
21672 information about several specific platforms.
21675 * Summary of Run-Time Configurations::
21676 * Specifying a Run-Time Library::
21677 * Choosing the Scheduling Policy::
21678 * Solaris-Specific Considerations::
21679 * Linux-Specific Considerations::
21680 * AIX-Specific Considerations::
21681 * Irix-Specific Considerations::
21682 * RTX-Specific Considerations::
21683 * HP-UX-Specific Considerations::
21686 @node Summary of Run-Time Configurations
21687 @section Summary of Run-Time Configurations
21689 @multitable @columnfractions .30 .70
21690 @item @b{alpha-openvms}
21691 @item @code{@ @ }@i{rts-native (default)}
21692 @item @code{@ @ @ @ }Tasking @tab native VMS threads
21693 @item @code{@ @ @ @ }Exceptions @tab ZCX
21695 @item @b{alpha-tru64}
21696 @item @code{@ @ }@i{rts-native (default)}
21697 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
21698 @item @code{@ @ @ @ }Exceptions @tab ZCX
21700 @item @code{@ @ }@i{rts-sjlj}
21701 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
21702 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21704 @item @b{ia64-hp_linux}
21705 @item @code{@ @ }@i{rts-native (default)}
21706 @item @code{@ @ @ @ }Tasking @tab pthread library
21707 @item @code{@ @ @ @ }Exceptions @tab ZCX
21709 @item @b{ia64-hpux}
21710 @item @code{@ @ }@i{rts-native (default)}
21711 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
21712 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21714 @item @b{ia64-openvms}
21715 @item @code{@ @ }@i{rts-native (default)}
21716 @item @code{@ @ @ @ }Tasking @tab native VMS threads
21717 @item @code{@ @ @ @ }Exceptions @tab ZCX
21719 @item @b{ia64-sgi_linux}
21720 @item @code{@ @ }@i{rts-native (default)}
21721 @item @code{@ @ @ @ }Tasking @tab pthread library
21722 @item @code{@ @ @ @ }Exceptions @tab ZCX
21724 @item @b{mips-irix}
21725 @item @code{@ @ }@i{rts-native (default)}
21726 @item @code{@ @ @ @ }Tasking @tab native IRIX threads
21727 @item @code{@ @ @ @ }Exceptions @tab ZCX
21730 @item @code{@ @ }@i{rts-native (default)}
21731 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
21732 @item @code{@ @ @ @ }Exceptions @tab ZCX
21734 @item @code{@ @ }@i{rts-sjlj}
21735 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
21736 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21739 @item @code{@ @ }@i{rts-native (default)}
21740 @item @code{@ @ @ @ }Tasking @tab native AIX threads
21741 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21743 @item @b{ppc-darwin}
21744 @item @code{@ @ }@i{rts-native (default)}
21745 @item @code{@ @ @ @ }Tasking @tab native MacOS threads
21746 @item @code{@ @ @ @ }Exceptions @tab ZCX
21748 @item @b{sparc-solaris} @tab
21749 @item @code{@ @ }@i{rts-native (default)}
21750 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
21751 @item @code{@ @ @ @ }Exceptions @tab ZCX
21753 @item @code{@ @ }@i{rts-pthread}
21754 @item @code{@ @ @ @ }Tasking @tab pthread library
21755 @item @code{@ @ @ @ }Exceptions @tab ZCX
21757 @item @code{@ @ }@i{rts-sjlj}
21758 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
21759 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21761 @item @b{sparc64-solaris} @tab
21762 @item @code{@ @ }@i{rts-native (default)}
21763 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
21764 @item @code{@ @ @ @ }Exceptions @tab ZCX
21766 @item @b{x86-linux}
21767 @item @code{@ @ }@i{rts-native (default)}
21768 @item @code{@ @ @ @ }Tasking @tab pthread library
21769 @item @code{@ @ @ @ }Exceptions @tab ZCX
21771 @item @code{@ @ }@i{rts-sjlj}
21772 @item @code{@ @ @ @ }Tasking @tab pthread library
21773 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21776 @item @code{@ @ }@i{rts-native (default)}
21777 @item @code{@ @ @ @ }Tasking @tab native LynxOS threads
21778 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21780 @item @b{x86-solaris}
21781 @item @code{@ @ }@i{rts-native (default)}
21782 @item @code{@ @ @ @ }Tasking @tab native Solaris threads
21783 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21785 @item @b{x86-windows}
21786 @item @code{@ @ }@i{rts-native (default)}
21787 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
21788 @item @code{@ @ @ @ }Exceptions @tab ZCX
21790 @item @code{@ @ }@i{rts-sjlj (default)}
21791 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
21792 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21794 @item @b{x86-windows-rtx}
21795 @item @code{@ @ }@i{rts-rtx-rtss (default)}
21796 @item @code{@ @ @ @ }Tasking @tab RTX real-time subsystem RTSS threads (kernel mode)
21797 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21799 @item @code{@ @ }@i{rts-rtx-w32}
21800 @item @code{@ @ @ @ }Tasking @tab RTX Win32 threads (user mode)
21801 @item @code{@ @ @ @ }Exceptions @tab ZCX
21803 @item @b{x86_64-linux}
21804 @item @code{@ @ }@i{rts-native (default)}
21805 @item @code{@ @ @ @ }Tasking @tab pthread library
21806 @item @code{@ @ @ @ }Exceptions @tab ZCX
21808 @item @code{@ @ }@i{rts-sjlj}
21809 @item @code{@ @ @ @ }Tasking @tab pthread library
21810 @item @code{@ @ @ @ }Exceptions @tab SJLJ
21814 @node Specifying a Run-Time Library
21815 @section Specifying a Run-Time Library
21818 The @file{adainclude} subdirectory containing the sources of the GNAT
21819 run-time library, and the @file{adalib} subdirectory containing the
21820 @file{ALI} files and the static and/or shared GNAT library, are located
21821 in the gcc target-dependent area:
21824 target=$prefix/lib/gcc/gcc-@i{dumpmachine}/gcc-@i{dumpversion}/
21828 As indicated above, on some platforms several run-time libraries are supplied.
21829 These libraries are installed in the target dependent area and
21830 contain a complete source and binary subdirectory. The detailed description
21831 below explains the differences between the different libraries in terms of
21832 their thread support.
21834 The default run-time library (when GNAT is installed) is @emph{rts-native}.
21835 This default run time is selected by the means of soft links.
21836 For example on x86-linux:
21842 +--- adainclude----------+
21844 +--- adalib-----------+ |
21846 +--- rts-native | |
21848 | +--- adainclude <---+
21850 | +--- adalib <----+
21861 If the @i{rts-sjlj} library is to be selected on a permanent basis,
21862 these soft links can be modified with the following commands:
21866 $ rm -f adainclude adalib
21867 $ ln -s rts-sjlj/adainclude adainclude
21868 $ ln -s rts-sjlj/adalib adalib
21872 Alternatively, you can specify @file{rts-sjlj/adainclude} in the file
21873 @file{$target/ada_source_path} and @file{rts-sjlj/adalib} in
21874 @file{$target/ada_object_path}.
21876 Selecting another run-time library temporarily can be
21877 achieved by using the @option{--RTS} switch, e.g., @option{--RTS=sjlj}
21878 @cindex @option{--RTS} option
21880 @node Choosing the Scheduling Policy
21881 @section Choosing the Scheduling Policy
21884 When using a POSIX threads implementation, you have a choice of several
21885 scheduling policies: @code{SCHED_FIFO},
21886 @cindex @code{SCHED_FIFO} scheduling policy
21888 @cindex @code{SCHED_RR} scheduling policy
21889 and @code{SCHED_OTHER}.
21890 @cindex @code{SCHED_OTHER} scheduling policy
21891 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
21892 or @code{SCHED_RR} requires special (e.g., root) privileges.
21894 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
21896 @cindex @code{SCHED_FIFO} scheduling policy
21897 you can use one of the following:
21901 @code{pragma Time_Slice (0.0)}
21902 @cindex pragma Time_Slice
21904 the corresponding binder option @option{-T0}
21905 @cindex @option{-T0} option
21907 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
21908 @cindex pragma Task_Dispatching_Policy
21912 To specify @code{SCHED_RR},
21913 @cindex @code{SCHED_RR} scheduling policy
21914 you should use @code{pragma Time_Slice} with a
21915 value greater than @code{0.0}, or else use the corresponding @option{-T}
21918 @node Solaris-Specific Considerations
21919 @section Solaris-Specific Considerations
21920 @cindex Solaris Sparc threads libraries
21923 This section addresses some topics related to the various threads libraries
21927 * Solaris Threads Issues::
21930 @node Solaris Threads Issues
21931 @subsection Solaris Threads Issues
21934 GNAT under Solaris/Sparc 32 bits comes with an alternate tasking run-time
21935 library based on POSIX threads --- @emph{rts-pthread}.
21936 @cindex rts-pthread threads library
21937 This run-time library has the advantage of being mostly shared across all
21938 POSIX-compliant thread implementations, and it also provides under
21939 @w{Solaris 8} the @code{PTHREAD_PRIO_INHERIT}
21940 @cindex @code{PTHREAD_PRIO_INHERIT} policy (under rts-pthread)
21941 and @code{PTHREAD_PRIO_PROTECT}
21942 @cindex @code{PTHREAD_PRIO_PROTECT} policy (under rts-pthread)
21943 semantics that can be selected using the predefined pragma
21944 @code{Locking_Policy}
21945 @cindex pragma Locking_Policy (under rts-pthread)
21947 @code{Inheritance_Locking} and @code{Ceiling_Locking} as the policy.
21948 @cindex @code{Inheritance_Locking} (under rts-pthread)
21949 @cindex @code{Ceiling_Locking} (under rts-pthread)
21951 As explained above, the native run-time library is based on the Solaris thread
21952 library (@code{libthread}) and is the default library.
21954 When the Solaris threads library is used (this is the default), programs
21955 compiled with GNAT can automatically take advantage of
21956 and can thus execute on multiple processors.
21957 The user can alternatively specify a processor on which the program should run
21958 to emulate a single-processor system. The multiprocessor / uniprocessor choice
21960 setting the environment variable @env{GNAT_PROCESSOR}
21961 @cindex @env{GNAT_PROCESSOR} environment variable (on Sparc Solaris)
21962 to one of the following:
21966 Use the default configuration (run the program on all
21967 available processors) - this is the same as having @code{GNAT_PROCESSOR}
21971 Let the run-time implementation choose one processor and run the program on
21974 @item 0 .. Last_Proc
21975 Run the program on the specified processor.
21976 @code{Last_Proc} is equal to @code{_SC_NPROCESSORS_CONF - 1}
21977 (where @code{_SC_NPROCESSORS_CONF} is a system variable).
21980 @node Linux-Specific Considerations
21981 @section Linux-Specific Considerations
21982 @cindex Linux threads libraries
21985 On GNU/Linux without NPTL support (usually system with GNU C Library
21986 older than 2.3), the signal model is not POSIX compliant, which means
21987 that to send a signal to the process, you need to send the signal to all
21988 threads, e.g.@: by using @code{killpg()}.
21990 @node AIX-Specific Considerations
21991 @section AIX-Specific Considerations
21992 @cindex AIX resolver library
21995 On AIX, the resolver library initializes some internal structure on
21996 the first call to @code{get*by*} functions, which are used to implement
21997 @code{GNAT.Sockets.Get_Host_By_Name} and
21998 @code{GNAT.Sockets.Get_Host_By_Address}.
21999 If such initialization occurs within an Ada task, and the stack size for
22000 the task is the default size, a stack overflow may occur.
22002 To avoid this overflow, the user should either ensure that the first call
22003 to @code{GNAT.Sockets.Get_Host_By_Name} or
22004 @code{GNAT.Sockets.Get_Host_By_Addrss}
22005 occurs in the environment task, or use @code{pragma Storage_Size} to
22006 specify a sufficiently large size for the stack of the task that contains
22009 @node Irix-Specific Considerations
22010 @section Irix-Specific Considerations
22011 @cindex Irix libraries
22014 The GCC support libraries coming with the Irix compiler have moved to
22015 their canonical place with respect to the general Irix ABI related
22016 conventions. Running applications built with the default shared GNAT
22017 run-time now requires the LD_LIBRARY_PATH environment variable to
22018 include this location. A possible way to achieve this is to issue the
22019 following command line on a bash prompt:
22023 $ LD_LIBRARY_PATH=$LD_LIBRARY_PATH:`dirname \`gcc --print-file-name=libgcc_s.so\``
22027 @node RTX-Specific Considerations
22028 @section RTX-Specific Considerations
22029 @cindex RTX libraries
22032 The Real-time Extension (RTX) to Windows is based on the Windows Win32
22033 API. Applications can be built to work in two different modes:
22037 Windows executables that run in Ring 3 to utilize memory protection
22038 (@emph{rts-rtx-w32}).
22041 Real-time subsystem (RTSS) executables that run in Ring 0, where
22042 performance can be optimized with RTSS applications taking precedent
22043 over all Windows applications (@emph{rts-rtx-rtss}). This mode requires
22044 the Microsoft linker to handle RTSS libraries.
22048 @node HP-UX-Specific Considerations
22049 @section HP-UX-Specific Considerations
22050 @cindex HP-UX Scheduling
22053 On HP-UX, appropriate privileges are required to change the scheduling
22054 parameters of a task. The calling process must have appropriate
22055 privileges or be a member of a group having @code{PRIV_RTSCHED} access to
22056 successfully change the scheduling parameters.
22058 By default, GNAT uses the @code{SCHED_HPUX} policy. To have access to the
22059 priority range 0-31 either the @code{FIFO_Within_Priorities} or the
22060 @code{Round_Robin_Within_Priorities} scheduling policies need to be set.
22062 To specify the @code{FIFO_Within_Priorities} scheduling policy you can use
22063 one of the following:
22067 @code{pragma Time_Slice (0.0)}
22068 @cindex pragma Time_Slice
22070 the corresponding binder option @option{-T0}
22071 @cindex @option{-T0} option
22073 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
22074 @cindex pragma Task_Dispatching_Policy
22078 To specify the @code{Round_Robin_Within_Priorities}, scheduling policy
22079 you should use @code{pragma Time_Slice} with a
22080 value greater than @code{0.0}, or use the corresponding @option{-T}
22081 binder option, or set the @code{pragma Task_Dispatching_Policy
22082 (Round_Robin_Within_Priorities)}.
22084 @c *******************************
22085 @node Example of Binder Output File
22086 @appendix Example of Binder Output File
22089 This Appendix displays the source code for @command{gnatbind}'s output
22090 file generated for a simple ``Hello World'' program.
22091 Comments have been added for clarification purposes.
22093 @smallexample @c adanocomment
22097 -- The package is called Ada_Main unless this name is actually used
22098 -- as a unit name in the partition, in which case some other unique
22102 package ada_main is
22104 Elab_Final_Code : Integer;
22105 pragma Import (C, Elab_Final_Code, "__gnat_inside_elab_final_code");
22107 -- The main program saves the parameters (argument count,
22108 -- argument values, environment pointer) in global variables
22109 -- for later access by other units including
22110 -- Ada.Command_Line.
22112 gnat_argc : Integer;
22113 gnat_argv : System.Address;
22114 gnat_envp : System.Address;
22116 -- The actual variables are stored in a library routine. This
22117 -- is useful for some shared library situations, where there
22118 -- are problems if variables are not in the library.
22120 pragma Import (C, gnat_argc);
22121 pragma Import (C, gnat_argv);
22122 pragma Import (C, gnat_envp);
22124 -- The exit status is similarly an external location
22126 gnat_exit_status : Integer;
22127 pragma Import (C, gnat_exit_status);
22129 GNAT_Version : constant String :=
22130 "GNAT Version: 6.0.0w (20061115)";
22131 pragma Export (C, GNAT_Version, "__gnat_version");
22133 -- This is the generated adafinal routine that performs
22134 -- finalization at the end of execution. In the case where
22135 -- Ada is the main program, this main program makes a call
22136 -- to adafinal at program termination.
22138 procedure adafinal;
22139 pragma Export (C, adafinal, "adafinal");
22141 -- This is the generated adainit routine that performs
22142 -- initialization at the start of execution. In the case
22143 -- where Ada is the main program, this main program makes
22144 -- a call to adainit at program startup.
22147 pragma Export (C, adainit, "adainit");
22149 -- This routine is called at the start of execution. It is
22150 -- a dummy routine that is used by the debugger to breakpoint
22151 -- at the start of execution.
22153 procedure Break_Start;
22154 pragma Import (C, Break_Start, "__gnat_break_start");
22156 -- This is the actual generated main program (it would be
22157 -- suppressed if the no main program switch were used). As
22158 -- required by standard system conventions, this program has
22159 -- the external name main.
22163 argv : System.Address;
22164 envp : System.Address)
22166 pragma Export (C, main, "main");
22168 -- The following set of constants give the version
22169 -- identification values for every unit in the bound
22170 -- partition. This identification is computed from all
22171 -- dependent semantic units, and corresponds to the
22172 -- string that would be returned by use of the
22173 -- Body_Version or Version attributes.
22175 type Version_32 is mod 2 ** 32;
22176 u00001 : constant Version_32 := 16#7880BEB3#;
22177 u00002 : constant Version_32 := 16#0D24CBD0#;
22178 u00003 : constant Version_32 := 16#3283DBEB#;
22179 u00004 : constant Version_32 := 16#2359F9ED#;
22180 u00005 : constant Version_32 := 16#664FB847#;
22181 u00006 : constant Version_32 := 16#68E803DF#;
22182 u00007 : constant Version_32 := 16#5572E604#;
22183 u00008 : constant Version_32 := 16#46B173D8#;
22184 u00009 : constant Version_32 := 16#156A40CF#;
22185 u00010 : constant Version_32 := 16#033DABE0#;
22186 u00011 : constant Version_32 := 16#6AB38FEA#;
22187 u00012 : constant Version_32 := 16#22B6217D#;
22188 u00013 : constant Version_32 := 16#68A22947#;
22189 u00014 : constant Version_32 := 16#18CC4A56#;
22190 u00015 : constant Version_32 := 16#08258E1B#;
22191 u00016 : constant Version_32 := 16#367D5222#;
22192 u00017 : constant Version_32 := 16#20C9ECA4#;
22193 u00018 : constant Version_32 := 16#50D32CB6#;
22194 u00019 : constant Version_32 := 16#39A8BB77#;
22195 u00020 : constant Version_32 := 16#5CF8FA2B#;
22196 u00021 : constant Version_32 := 16#2F1EB794#;
22197 u00022 : constant Version_32 := 16#31AB6444#;
22198 u00023 : constant Version_32 := 16#1574B6E9#;
22199 u00024 : constant Version_32 := 16#5109C189#;
22200 u00025 : constant Version_32 := 16#56D770CD#;
22201 u00026 : constant Version_32 := 16#02F9DE3D#;
22202 u00027 : constant Version_32 := 16#08AB6B2C#;
22203 u00028 : constant Version_32 := 16#3FA37670#;
22204 u00029 : constant Version_32 := 16#476457A0#;
22205 u00030 : constant Version_32 := 16#731E1B6E#;
22206 u00031 : constant Version_32 := 16#23C2E789#;
22207 u00032 : constant Version_32 := 16#0F1BD6A1#;
22208 u00033 : constant Version_32 := 16#7C25DE96#;
22209 u00034 : constant Version_32 := 16#39ADFFA2#;
22210 u00035 : constant Version_32 := 16#571DE3E7#;
22211 u00036 : constant Version_32 := 16#5EB646AB#;
22212 u00037 : constant Version_32 := 16#4249379B#;
22213 u00038 : constant Version_32 := 16#0357E00A#;
22214 u00039 : constant Version_32 := 16#3784FB72#;
22215 u00040 : constant Version_32 := 16#2E723019#;
22216 u00041 : constant Version_32 := 16#623358EA#;
22217 u00042 : constant Version_32 := 16#107F9465#;
22218 u00043 : constant Version_32 := 16#6843F68A#;
22219 u00044 : constant Version_32 := 16#63305874#;
22220 u00045 : constant Version_32 := 16#31E56CE1#;
22221 u00046 : constant Version_32 := 16#02917970#;
22222 u00047 : constant Version_32 := 16#6CCBA70E#;
22223 u00048 : constant Version_32 := 16#41CD4204#;
22224 u00049 : constant Version_32 := 16#572E3F58#;
22225 u00050 : constant Version_32 := 16#20729FF5#;
22226 u00051 : constant Version_32 := 16#1D4F93E8#;
22227 u00052 : constant Version_32 := 16#30B2EC3D#;
22228 u00053 : constant Version_32 := 16#34054F96#;
22229 u00054 : constant Version_32 := 16#5A199860#;
22230 u00055 : constant Version_32 := 16#0E7F912B#;
22231 u00056 : constant Version_32 := 16#5760634A#;
22232 u00057 : constant Version_32 := 16#5D851835#;
22234 -- The following Export pragmas export the version numbers
22235 -- with symbolic names ending in B (for body) or S
22236 -- (for spec) so that they can be located in a link. The
22237 -- information provided here is sufficient to track down
22238 -- the exact versions of units used in a given build.
22240 pragma Export (C, u00001, "helloB");
22241 pragma Export (C, u00002, "system__standard_libraryB");
22242 pragma Export (C, u00003, "system__standard_libraryS");
22243 pragma Export (C, u00004, "adaS");
22244 pragma Export (C, u00005, "ada__text_ioB");
22245 pragma Export (C, u00006, "ada__text_ioS");
22246 pragma Export (C, u00007, "ada__exceptionsB");
22247 pragma Export (C, u00008, "ada__exceptionsS");
22248 pragma Export (C, u00009, "gnatS");
22249 pragma Export (C, u00010, "gnat__heap_sort_aB");
22250 pragma Export (C, u00011, "gnat__heap_sort_aS");
22251 pragma Export (C, u00012, "systemS");
22252 pragma Export (C, u00013, "system__exception_tableB");
22253 pragma Export (C, u00014, "system__exception_tableS");
22254 pragma Export (C, u00015, "gnat__htableB");
22255 pragma Export (C, u00016, "gnat__htableS");
22256 pragma Export (C, u00017, "system__exceptionsS");
22257 pragma Export (C, u00018, "system__machine_state_operationsB");
22258 pragma Export (C, u00019, "system__machine_state_operationsS");
22259 pragma Export (C, u00020, "system__machine_codeS");
22260 pragma Export (C, u00021, "system__storage_elementsB");
22261 pragma Export (C, u00022, "system__storage_elementsS");
22262 pragma Export (C, u00023, "system__secondary_stackB");
22263 pragma Export (C, u00024, "system__secondary_stackS");
22264 pragma Export (C, u00025, "system__parametersB");
22265 pragma Export (C, u00026, "system__parametersS");
22266 pragma Export (C, u00027, "system__soft_linksB");
22267 pragma Export (C, u00028, "system__soft_linksS");
22268 pragma Export (C, u00029, "system__stack_checkingB");
22269 pragma Export (C, u00030, "system__stack_checkingS");
22270 pragma Export (C, u00031, "system__tracebackB");
22271 pragma Export (C, u00032, "system__tracebackS");
22272 pragma Export (C, u00033, "ada__streamsS");
22273 pragma Export (C, u00034, "ada__tagsB");
22274 pragma Export (C, u00035, "ada__tagsS");
22275 pragma Export (C, u00036, "system__string_opsB");
22276 pragma Export (C, u00037, "system__string_opsS");
22277 pragma Export (C, u00038, "interfacesS");
22278 pragma Export (C, u00039, "interfaces__c_streamsB");
22279 pragma Export (C, u00040, "interfaces__c_streamsS");
22280 pragma Export (C, u00041, "system__file_ioB");
22281 pragma Export (C, u00042, "system__file_ioS");
22282 pragma Export (C, u00043, "ada__finalizationB");
22283 pragma Export (C, u00044, "ada__finalizationS");
22284 pragma Export (C, u00045, "system__finalization_rootB");
22285 pragma Export (C, u00046, "system__finalization_rootS");
22286 pragma Export (C, u00047, "system__finalization_implementationB");
22287 pragma Export (C, u00048, "system__finalization_implementationS");
22288 pragma Export (C, u00049, "system__string_ops_concat_3B");
22289 pragma Export (C, u00050, "system__string_ops_concat_3S");
22290 pragma Export (C, u00051, "system__stream_attributesB");
22291 pragma Export (C, u00052, "system__stream_attributesS");
22292 pragma Export (C, u00053, "ada__io_exceptionsS");
22293 pragma Export (C, u00054, "system__unsigned_typesS");
22294 pragma Export (C, u00055, "system__file_control_blockS");
22295 pragma Export (C, u00056, "ada__finalization__list_controllerB");
22296 pragma Export (C, u00057, "ada__finalization__list_controllerS");
22298 -- BEGIN ELABORATION ORDER
22301 -- gnat.heap_sort_a (spec)
22302 -- gnat.heap_sort_a (body)
22303 -- gnat.htable (spec)
22304 -- gnat.htable (body)
22305 -- interfaces (spec)
22307 -- system.machine_code (spec)
22308 -- system.parameters (spec)
22309 -- system.parameters (body)
22310 -- interfaces.c_streams (spec)
22311 -- interfaces.c_streams (body)
22312 -- system.standard_library (spec)
22313 -- ada.exceptions (spec)
22314 -- system.exception_table (spec)
22315 -- system.exception_table (body)
22316 -- ada.io_exceptions (spec)
22317 -- system.exceptions (spec)
22318 -- system.storage_elements (spec)
22319 -- system.storage_elements (body)
22320 -- system.machine_state_operations (spec)
22321 -- system.machine_state_operations (body)
22322 -- system.secondary_stack (spec)
22323 -- system.stack_checking (spec)
22324 -- system.soft_links (spec)
22325 -- system.soft_links (body)
22326 -- system.stack_checking (body)
22327 -- system.secondary_stack (body)
22328 -- system.standard_library (body)
22329 -- system.string_ops (spec)
22330 -- system.string_ops (body)
22333 -- ada.streams (spec)
22334 -- system.finalization_root (spec)
22335 -- system.finalization_root (body)
22336 -- system.string_ops_concat_3 (spec)
22337 -- system.string_ops_concat_3 (body)
22338 -- system.traceback (spec)
22339 -- system.traceback (body)
22340 -- ada.exceptions (body)
22341 -- system.unsigned_types (spec)
22342 -- system.stream_attributes (spec)
22343 -- system.stream_attributes (body)
22344 -- system.finalization_implementation (spec)
22345 -- system.finalization_implementation (body)
22346 -- ada.finalization (spec)
22347 -- ada.finalization (body)
22348 -- ada.finalization.list_controller (spec)
22349 -- ada.finalization.list_controller (body)
22350 -- system.file_control_block (spec)
22351 -- system.file_io (spec)
22352 -- system.file_io (body)
22353 -- ada.text_io (spec)
22354 -- ada.text_io (body)
22356 -- END ELABORATION ORDER
22360 -- The following source file name pragmas allow the generated file
22361 -- names to be unique for different main programs. They are needed
22362 -- since the package name will always be Ada_Main.
22364 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
22365 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
22367 -- Generated package body for Ada_Main starts here
22369 package body ada_main is
22371 -- The actual finalization is performed by calling the
22372 -- library routine in System.Standard_Library.Adafinal
22374 procedure Do_Finalize;
22375 pragma Import (C, Do_Finalize, "system__standard_library__adafinal");
22382 procedure adainit is
22384 -- These booleans are set to True once the associated unit has
22385 -- been elaborated. It is also used to avoid elaborating the
22386 -- same unit twice.
22389 pragma Import (Ada, E040, "interfaces__c_streams_E");
22392 pragma Import (Ada, E008, "ada__exceptions_E");
22395 pragma Import (Ada, E014, "system__exception_table_E");
22398 pragma Import (Ada, E053, "ada__io_exceptions_E");
22401 pragma Import (Ada, E017, "system__exceptions_E");
22404 pragma Import (Ada, E024, "system__secondary_stack_E");
22407 pragma Import (Ada, E030, "system__stack_checking_E");
22410 pragma Import (Ada, E028, "system__soft_links_E");
22413 pragma Import (Ada, E035, "ada__tags_E");
22416 pragma Import (Ada, E033, "ada__streams_E");
22419 pragma Import (Ada, E046, "system__finalization_root_E");
22422 pragma Import (Ada, E048, "system__finalization_implementation_E");
22425 pragma Import (Ada, E044, "ada__finalization_E");
22428 pragma Import (Ada, E057, "ada__finalization__list_controller_E");
22431 pragma Import (Ada, E055, "system__file_control_block_E");
22434 pragma Import (Ada, E042, "system__file_io_E");
22437 pragma Import (Ada, E006, "ada__text_io_E");
22439 -- Set_Globals is a library routine that stores away the
22440 -- value of the indicated set of global values in global
22441 -- variables within the library.
22443 procedure Set_Globals
22444 (Main_Priority : Integer;
22445 Time_Slice_Value : Integer;
22446 WC_Encoding : Character;
22447 Locking_Policy : Character;
22448 Queuing_Policy : Character;
22449 Task_Dispatching_Policy : Character;
22450 Adafinal : System.Address;
22451 Unreserve_All_Interrupts : Integer;
22452 Exception_Tracebacks : Integer);
22453 @findex __gnat_set_globals
22454 pragma Import (C, Set_Globals, "__gnat_set_globals");
22456 -- SDP_Table_Build is a library routine used to build the
22457 -- exception tables. See unit Ada.Exceptions in files
22458 -- a-except.ads/adb for full details of how zero cost
22459 -- exception handling works. This procedure, the call to
22460 -- it, and the two following tables are all omitted if the
22461 -- build is in longjmp/setjmp exception mode.
22463 @findex SDP_Table_Build
22464 @findex Zero Cost Exceptions
22465 procedure SDP_Table_Build
22466 (SDP_Addresses : System.Address;
22467 SDP_Count : Natural;
22468 Elab_Addresses : System.Address;
22469 Elab_Addr_Count : Natural);
22470 pragma Import (C, SDP_Table_Build, "__gnat_SDP_Table_Build");
22472 -- Table of Unit_Exception_Table addresses. Used for zero
22473 -- cost exception handling to build the top level table.
22475 ST : aliased constant array (1 .. 23) of System.Address := (
22477 Ada.Text_Io'UET_Address,
22478 Ada.Exceptions'UET_Address,
22479 Gnat.Heap_Sort_A'UET_Address,
22480 System.Exception_Table'UET_Address,
22481 System.Machine_State_Operations'UET_Address,
22482 System.Secondary_Stack'UET_Address,
22483 System.Parameters'UET_Address,
22484 System.Soft_Links'UET_Address,
22485 System.Stack_Checking'UET_Address,
22486 System.Traceback'UET_Address,
22487 Ada.Streams'UET_Address,
22488 Ada.Tags'UET_Address,
22489 System.String_Ops'UET_Address,
22490 Interfaces.C_Streams'UET_Address,
22491 System.File_Io'UET_Address,
22492 Ada.Finalization'UET_Address,
22493 System.Finalization_Root'UET_Address,
22494 System.Finalization_Implementation'UET_Address,
22495 System.String_Ops_Concat_3'UET_Address,
22496 System.Stream_Attributes'UET_Address,
22497 System.File_Control_Block'UET_Address,
22498 Ada.Finalization.List_Controller'UET_Address);
22500 -- Table of addresses of elaboration routines. Used for
22501 -- zero cost exception handling to make sure these
22502 -- addresses are included in the top level procedure
22505 EA : aliased constant array (1 .. 23) of System.Address := (
22506 adainit'Code_Address,
22507 Do_Finalize'Code_Address,
22508 Ada.Exceptions'Elab_Spec'Address,
22509 System.Exceptions'Elab_Spec'Address,
22510 Interfaces.C_Streams'Elab_Spec'Address,
22511 System.Exception_Table'Elab_Body'Address,
22512 Ada.Io_Exceptions'Elab_Spec'Address,
22513 System.Stack_Checking'Elab_Spec'Address,
22514 System.Soft_Links'Elab_Body'Address,
22515 System.Secondary_Stack'Elab_Body'Address,
22516 Ada.Tags'Elab_Spec'Address,
22517 Ada.Tags'Elab_Body'Address,
22518 Ada.Streams'Elab_Spec'Address,
22519 System.Finalization_Root'Elab_Spec'Address,
22520 Ada.Exceptions'Elab_Body'Address,
22521 System.Finalization_Implementation'Elab_Spec'Address,
22522 System.Finalization_Implementation'Elab_Body'Address,
22523 Ada.Finalization'Elab_Spec'Address,
22524 Ada.Finalization.List_Controller'Elab_Spec'Address,
22525 System.File_Control_Block'Elab_Spec'Address,
22526 System.File_Io'Elab_Body'Address,
22527 Ada.Text_Io'Elab_Spec'Address,
22528 Ada.Text_Io'Elab_Body'Address);
22530 -- Start of processing for adainit
22534 -- Call SDP_Table_Build to build the top level procedure
22535 -- table for zero cost exception handling (omitted in
22536 -- longjmp/setjmp mode).
22538 SDP_Table_Build (ST'Address, 23, EA'Address, 23);
22540 -- Call Set_Globals to record various information for
22541 -- this partition. The values are derived by the binder
22542 -- from information stored in the ali files by the compiler.
22544 @findex __gnat_set_globals
22546 (Main_Priority => -1,
22547 -- Priority of main program, -1 if no pragma Priority used
22549 Time_Slice_Value => -1,
22550 -- Time slice from Time_Slice pragma, -1 if none used
22552 WC_Encoding => 'b',
22553 -- Wide_Character encoding used, default is brackets
22555 Locking_Policy => ' ',
22556 -- Locking_Policy used, default of space means not
22557 -- specified, otherwise it is the first character of
22558 -- the policy name.
22560 Queuing_Policy => ' ',
22561 -- Queuing_Policy used, default of space means not
22562 -- specified, otherwise it is the first character of
22563 -- the policy name.
22565 Task_Dispatching_Policy => ' ',
22566 -- Task_Dispatching_Policy used, default of space means
22567 -- not specified, otherwise first character of the
22570 Adafinal => System.Null_Address,
22571 -- Address of Adafinal routine, not used anymore
22573 Unreserve_All_Interrupts => 0,
22574 -- Set true if pragma Unreserve_All_Interrupts was used
22576 Exception_Tracebacks => 0);
22577 -- Indicates if exception tracebacks are enabled
22579 Elab_Final_Code := 1;
22581 -- Now we have the elaboration calls for all units in the partition.
22582 -- The Elab_Spec and Elab_Body attributes generate references to the
22583 -- implicit elaboration procedures generated by the compiler for
22584 -- each unit that requires elaboration.
22587 Interfaces.C_Streams'Elab_Spec;
22591 Ada.Exceptions'Elab_Spec;
22594 System.Exception_Table'Elab_Body;
22598 Ada.Io_Exceptions'Elab_Spec;
22602 System.Exceptions'Elab_Spec;
22606 System.Stack_Checking'Elab_Spec;
22609 System.Soft_Links'Elab_Body;
22614 System.Secondary_Stack'Elab_Body;
22618 Ada.Tags'Elab_Spec;
22621 Ada.Tags'Elab_Body;
22625 Ada.Streams'Elab_Spec;
22629 System.Finalization_Root'Elab_Spec;
22633 Ada.Exceptions'Elab_Body;
22637 System.Finalization_Implementation'Elab_Spec;
22640 System.Finalization_Implementation'Elab_Body;
22644 Ada.Finalization'Elab_Spec;
22648 Ada.Finalization.List_Controller'Elab_Spec;
22652 System.File_Control_Block'Elab_Spec;
22656 System.File_Io'Elab_Body;
22660 Ada.Text_Io'Elab_Spec;
22663 Ada.Text_Io'Elab_Body;
22667 Elab_Final_Code := 0;
22675 procedure adafinal is
22684 -- main is actually a function, as in the ANSI C standard,
22685 -- defined to return the exit status. The three parameters
22686 -- are the argument count, argument values and environment
22689 @findex Main Program
22692 argv : System.Address;
22693 envp : System.Address)
22696 -- The initialize routine performs low level system
22697 -- initialization using a standard library routine which
22698 -- sets up signal handling and performs any other
22699 -- required setup. The routine can be found in file
22702 @findex __gnat_initialize
22703 procedure initialize;
22704 pragma Import (C, initialize, "__gnat_initialize");
22706 -- The finalize routine performs low level system
22707 -- finalization using a standard library routine. The
22708 -- routine is found in file a-final.c and in the standard
22709 -- distribution is a dummy routine that does nothing, so
22710 -- really this is a hook for special user finalization.
22712 @findex __gnat_finalize
22713 procedure finalize;
22714 pragma Import (C, finalize, "__gnat_finalize");
22716 -- We get to the main program of the partition by using
22717 -- pragma Import because if we try to with the unit and
22718 -- call it Ada style, then not only do we waste time
22719 -- recompiling it, but also, we don't really know the right
22720 -- switches (e.g.@: identifier character set) to be used
22723 procedure Ada_Main_Program;
22724 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
22726 -- Start of processing for main
22729 -- Save global variables
22735 -- Call low level system initialization
22739 -- Call our generated Ada initialization routine
22743 -- This is the point at which we want the debugger to get
22748 -- Now we call the main program of the partition
22752 -- Perform Ada finalization
22756 -- Perform low level system finalization
22760 -- Return the proper exit status
22761 return (gnat_exit_status);
22764 -- This section is entirely comments, so it has no effect on the
22765 -- compilation of the Ada_Main package. It provides the list of
22766 -- object files and linker options, as well as some standard
22767 -- libraries needed for the link. The gnatlink utility parses
22768 -- this b~hello.adb file to read these comment lines to generate
22769 -- the appropriate command line arguments for the call to the
22770 -- system linker. The BEGIN/END lines are used for sentinels for
22771 -- this parsing operation.
22773 -- The exact file names will of course depend on the environment,
22774 -- host/target and location of files on the host system.
22776 @findex Object file list
22777 -- BEGIN Object file/option list
22780 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
22781 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
22782 -- END Object file/option list
22788 The Ada code in the above example is exactly what is generated by the
22789 binder. We have added comments to more clearly indicate the function
22790 of each part of the generated @code{Ada_Main} package.
22792 The code is standard Ada in all respects, and can be processed by any
22793 tools that handle Ada. In particular, it is possible to use the debugger
22794 in Ada mode to debug the generated @code{Ada_Main} package. For example,
22795 suppose that for reasons that you do not understand, your program is crashing
22796 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
22797 you can place a breakpoint on the call:
22799 @smallexample @c ada
22800 Ada.Text_Io'Elab_Body;
22804 and trace the elaboration routine for this package to find out where
22805 the problem might be (more usually of course you would be debugging
22806 elaboration code in your own application).
22808 @node Elaboration Order Handling in GNAT
22809 @appendix Elaboration Order Handling in GNAT
22810 @cindex Order of elaboration
22811 @cindex Elaboration control
22814 * Elaboration Code::
22815 * Checking the Elaboration Order::
22816 * Controlling the Elaboration Order::
22817 * Controlling Elaboration in GNAT - Internal Calls::
22818 * Controlling Elaboration in GNAT - External Calls::
22819 * Default Behavior in GNAT - Ensuring Safety::
22820 * Treatment of Pragma Elaborate::
22821 * Elaboration Issues for Library Tasks::
22822 * Mixing Elaboration Models::
22823 * What to Do If the Default Elaboration Behavior Fails::
22824 * Elaboration for Access-to-Subprogram Values::
22825 * Summary of Procedures for Elaboration Control::
22826 * Other Elaboration Order Considerations::
22830 This chapter describes the handling of elaboration code in Ada and
22831 in GNAT, and discusses how the order of elaboration of program units can
22832 be controlled in GNAT, either automatically or with explicit programming
22835 @node Elaboration Code
22836 @section Elaboration Code
22839 Ada provides rather general mechanisms for executing code at elaboration
22840 time, that is to say before the main program starts executing. Such code arises
22844 @item Initializers for variables.
22845 Variables declared at the library level, in package specs or bodies, can
22846 require initialization that is performed at elaboration time, as in:
22847 @smallexample @c ada
22849 Sqrt_Half : Float := Sqrt (0.5);
22853 @item Package initialization code
22854 Code in a @code{BEGIN-END} section at the outer level of a package body is
22855 executed as part of the package body elaboration code.
22857 @item Library level task allocators
22858 Tasks that are declared using task allocators at the library level
22859 start executing immediately and hence can execute at elaboration time.
22863 Subprogram calls are possible in any of these contexts, which means that
22864 any arbitrary part of the program may be executed as part of the elaboration
22865 code. It is even possible to write a program which does all its work at
22866 elaboration time, with a null main program, although stylistically this
22867 would usually be considered an inappropriate way to structure
22870 An important concern arises in the context of elaboration code:
22871 we have to be sure that it is executed in an appropriate order. What we
22872 have is a series of elaboration code sections, potentially one section
22873 for each unit in the program. It is important that these execute
22874 in the correct order. Correctness here means that, taking the above
22875 example of the declaration of @code{Sqrt_Half},
22876 if some other piece of
22877 elaboration code references @code{Sqrt_Half},
22878 then it must run after the
22879 section of elaboration code that contains the declaration of
22882 There would never be any order of elaboration problem if we made a rule
22883 that whenever you @code{with} a unit, you must elaborate both the spec and body
22884 of that unit before elaborating the unit doing the @code{with}'ing:
22886 @smallexample @c ada
22890 package Unit_2 is @dots{}
22896 would require that both the body and spec of @code{Unit_1} be elaborated
22897 before the spec of @code{Unit_2}. However, a rule like that would be far too
22898 restrictive. In particular, it would make it impossible to have routines
22899 in separate packages that were mutually recursive.
22901 You might think that a clever enough compiler could look at the actual
22902 elaboration code and determine an appropriate correct order of elaboration,
22903 but in the general case, this is not possible. Consider the following
22906 In the body of @code{Unit_1}, we have a procedure @code{Func_1}
22908 the variable @code{Sqrt_1}, which is declared in the elaboration code
22909 of the body of @code{Unit_1}:
22911 @smallexample @c ada
22913 Sqrt_1 : Float := Sqrt (0.1);
22918 The elaboration code of the body of @code{Unit_1} also contains:
22920 @smallexample @c ada
22923 if expression_1 = 1 then
22924 Q := Unit_2.Func_2;
22931 @code{Unit_2} is exactly parallel,
22932 it has a procedure @code{Func_2} that references
22933 the variable @code{Sqrt_2}, which is declared in the elaboration code of
22934 the body @code{Unit_2}:
22936 @smallexample @c ada
22938 Sqrt_2 : Float := Sqrt (0.1);
22943 The elaboration code of the body of @code{Unit_2} also contains:
22945 @smallexample @c ada
22948 if expression_2 = 2 then
22949 Q := Unit_1.Func_1;
22956 Now the question is, which of the following orders of elaboration is
22981 If you carefully analyze the flow here, you will see that you cannot tell
22982 at compile time the answer to this question.
22983 If @code{expression_1} is not equal to 1,
22984 and @code{expression_2} is not equal to 2,
22985 then either order is acceptable, because neither of the function calls is
22986 executed. If both tests evaluate to true, then neither order is acceptable
22987 and in fact there is no correct order.
22989 If one of the two expressions is true, and the other is false, then one
22990 of the above orders is correct, and the other is incorrect. For example,
22991 if @code{expression_1} /= 1 and @code{expression_2} = 2,
22992 then the call to @code{Func_1}
22993 will occur, but not the call to @code{Func_2.}
22994 This means that it is essential
22995 to elaborate the body of @code{Unit_1} before
22996 the body of @code{Unit_2}, so the first
22997 order of elaboration is correct and the second is wrong.
22999 By making @code{expression_1} and @code{expression_2}
23000 depend on input data, or perhaps
23001 the time of day, we can make it impossible for the compiler or binder
23002 to figure out which of these expressions will be true, and hence it
23003 is impossible to guarantee a safe order of elaboration at run time.
23005 @node Checking the Elaboration Order
23006 @section Checking the Elaboration Order
23009 In some languages that involve the same kind of elaboration problems,
23010 e.g.@: Java and C++, the programmer is expected to worry about these
23011 ordering problems himself, and it is common to
23012 write a program in which an incorrect elaboration order gives
23013 surprising results, because it references variables before they
23015 Ada is designed to be a safe language, and a programmer-beware approach is
23016 clearly not sufficient. Consequently, the language provides three lines
23020 @item Standard rules
23021 Some standard rules restrict the possible choice of elaboration
23022 order. In particular, if you @code{with} a unit, then its spec is always
23023 elaborated before the unit doing the @code{with}. Similarly, a parent
23024 spec is always elaborated before the child spec, and finally
23025 a spec is always elaborated before its corresponding body.
23027 @item Dynamic elaboration checks
23028 @cindex Elaboration checks
23029 @cindex Checks, elaboration
23030 Dynamic checks are made at run time, so that if some entity is accessed
23031 before it is elaborated (typically by means of a subprogram call)
23032 then the exception (@code{Program_Error}) is raised.
23034 @item Elaboration control
23035 Facilities are provided for the programmer to specify the desired order
23039 Let's look at these facilities in more detail. First, the rules for
23040 dynamic checking. One possible rule would be simply to say that the
23041 exception is raised if you access a variable which has not yet been
23042 elaborated. The trouble with this approach is that it could require
23043 expensive checks on every variable reference. Instead Ada has two
23044 rules which are a little more restrictive, but easier to check, and
23048 @item Restrictions on calls
23049 A subprogram can only be called at elaboration time if its body
23050 has been elaborated. The rules for elaboration given above guarantee
23051 that the spec of the subprogram has been elaborated before the
23052 call, but not the body. If this rule is violated, then the
23053 exception @code{Program_Error} is raised.
23055 @item Restrictions on instantiations
23056 A generic unit can only be instantiated if the body of the generic
23057 unit has been elaborated. Again, the rules for elaboration given above
23058 guarantee that the spec of the generic unit has been elaborated
23059 before the instantiation, but not the body. If this rule is
23060 violated, then the exception @code{Program_Error} is raised.
23064 The idea is that if the body has been elaborated, then any variables
23065 it references must have been elaborated; by checking for the body being
23066 elaborated we guarantee that none of its references causes any
23067 trouble. As we noted above, this is a little too restrictive, because a
23068 subprogram that has no non-local references in its body may in fact be safe
23069 to call. However, it really would be unsafe to rely on this, because
23070 it would mean that the caller was aware of details of the implementation
23071 in the body. This goes against the basic tenets of Ada.
23073 A plausible implementation can be described as follows.
23074 A Boolean variable is associated with each subprogram
23075 and each generic unit. This variable is initialized to False, and is set to
23076 True at the point body is elaborated. Every call or instantiation checks the
23077 variable, and raises @code{Program_Error} if the variable is False.
23079 Note that one might think that it would be good enough to have one Boolean
23080 variable for each package, but that would not deal with cases of trying
23081 to call a body in the same package as the call
23082 that has not been elaborated yet.
23083 Of course a compiler may be able to do enough analysis to optimize away
23084 some of the Boolean variables as unnecessary, and @code{GNAT} indeed
23085 does such optimizations, but still the easiest conceptual model is to
23086 think of there being one variable per subprogram.
23088 @node Controlling the Elaboration Order
23089 @section Controlling the Elaboration Order
23092 In the previous section we discussed the rules in Ada which ensure
23093 that @code{Program_Error} is raised if an incorrect elaboration order is
23094 chosen. This prevents erroneous executions, but we need mechanisms to
23095 specify a correct execution and avoid the exception altogether.
23096 To achieve this, Ada provides a number of features for controlling
23097 the order of elaboration. We discuss these features in this section.
23099 First, there are several ways of indicating to the compiler that a given
23100 unit has no elaboration problems:
23103 @item packages that do not require a body
23104 A library package that does not require a body does not permit
23105 a body (this rule was introduced in Ada 95).
23106 Thus if we have a such a package, as in:
23108 @smallexample @c ada
23111 package Definitions is
23113 type m is new integer;
23115 type a is array (1 .. 10) of m;
23116 type b is array (1 .. 20) of m;
23124 A package that @code{with}'s @code{Definitions} may safely instantiate
23125 @code{Definitions.Subp} because the compiler can determine that there
23126 definitely is no package body to worry about in this case
23129 @cindex pragma Pure
23131 Places sufficient restrictions on a unit to guarantee that
23132 no call to any subprogram in the unit can result in an
23133 elaboration problem. This means that the compiler does not need
23134 to worry about the point of elaboration of such units, and in
23135 particular, does not need to check any calls to any subprograms
23138 @item pragma Preelaborate
23139 @findex Preelaborate
23140 @cindex pragma Preelaborate
23141 This pragma places slightly less stringent restrictions on a unit than
23143 but these restrictions are still sufficient to ensure that there
23144 are no elaboration problems with any calls to the unit.
23146 @item pragma Elaborate_Body
23147 @findex Elaborate_Body
23148 @cindex pragma Elaborate_Body
23149 This pragma requires that the body of a unit be elaborated immediately
23150 after its spec. Suppose a unit @code{A} has such a pragma,
23151 and unit @code{B} does
23152 a @code{with} of unit @code{A}. Recall that the standard rules require
23153 the spec of unit @code{A}
23154 to be elaborated before the @code{with}'ing unit; given the pragma in
23155 @code{A}, we also know that the body of @code{A}
23156 will be elaborated before @code{B}, so
23157 that calls to @code{A} are safe and do not need a check.
23162 unlike pragma @code{Pure} and pragma @code{Preelaborate},
23164 @code{Elaborate_Body} does not guarantee that the program is
23165 free of elaboration problems, because it may not be possible
23166 to satisfy the requested elaboration order.
23167 Let's go back to the example with @code{Unit_1} and @code{Unit_2}.
23169 marks @code{Unit_1} as @code{Elaborate_Body},
23170 and not @code{Unit_2,} then the order of
23171 elaboration will be:
23183 Now that means that the call to @code{Func_1} in @code{Unit_2}
23184 need not be checked,
23185 it must be safe. But the call to @code{Func_2} in
23186 @code{Unit_1} may still fail if
23187 @code{Expression_1} is equal to 1,
23188 and the programmer must still take
23189 responsibility for this not being the case.
23191 If all units carry a pragma @code{Elaborate_Body}, then all problems are
23192 eliminated, except for calls entirely within a body, which are
23193 in any case fully under programmer control. However, using the pragma
23194 everywhere is not always possible.
23195 In particular, for our @code{Unit_1}/@code{Unit_2} example, if
23196 we marked both of them as having pragma @code{Elaborate_Body}, then
23197 clearly there would be no possible elaboration order.
23199 The above pragmas allow a server to guarantee safe use by clients, and
23200 clearly this is the preferable approach. Consequently a good rule
23201 is to mark units as @code{Pure} or @code{Preelaborate} if possible,
23202 and if this is not possible,
23203 mark them as @code{Elaborate_Body} if possible.
23204 As we have seen, there are situations where neither of these
23205 three pragmas can be used.
23206 So we also provide methods for clients to control the
23207 order of elaboration of the servers on which they depend:
23210 @item pragma Elaborate (unit)
23212 @cindex pragma Elaborate
23213 This pragma is placed in the context clause, after a @code{with} clause,
23214 and it requires that the body of the named unit be elaborated before
23215 the unit in which the pragma occurs. The idea is to use this pragma
23216 if the current unit calls at elaboration time, directly or indirectly,
23217 some subprogram in the named unit.
23219 @item pragma Elaborate_All (unit)
23220 @findex Elaborate_All
23221 @cindex pragma Elaborate_All
23222 This is a stronger version of the Elaborate pragma. Consider the
23226 Unit A @code{with}'s unit B and calls B.Func in elab code
23227 Unit B @code{with}'s unit C, and B.Func calls C.Func
23231 Now if we put a pragma @code{Elaborate (B)}
23232 in unit @code{A}, this ensures that the
23233 body of @code{B} is elaborated before the call, but not the
23234 body of @code{C}, so
23235 the call to @code{C.Func} could still cause @code{Program_Error} to
23238 The effect of a pragma @code{Elaborate_All} is stronger, it requires
23239 not only that the body of the named unit be elaborated before the
23240 unit doing the @code{with}, but also the bodies of all units that the
23241 named unit uses, following @code{with} links transitively. For example,
23242 if we put a pragma @code{Elaborate_All (B)} in unit @code{A},
23244 not only that the body of @code{B} be elaborated before @code{A},
23246 body of @code{C}, because @code{B} @code{with}'s @code{C}.
23250 We are now in a position to give a usage rule in Ada for avoiding
23251 elaboration problems, at least if dynamic dispatching and access to
23252 subprogram values are not used. We will handle these cases separately
23255 The rule is simple. If a unit has elaboration code that can directly or
23256 indirectly make a call to a subprogram in a @code{with}'ed unit, or instantiate
23257 a generic package in a @code{with}'ed unit,
23258 then if the @code{with}'ed unit does not have
23259 pragma @code{Pure} or @code{Preelaborate}, then the client should have
23260 a pragma @code{Elaborate_All}
23261 for the @code{with}'ed unit. By following this rule a client is
23262 assured that calls can be made without risk of an exception.
23264 For generic subprogram instantiations, the rule can be relaxed to
23265 require only a pragma @code{Elaborate} since elaborating the body
23266 of a subprogram cannot cause any transitive elaboration (we are
23267 not calling the subprogram in this case, just elaborating its
23270 If this rule is not followed, then a program may be in one of four
23274 @item No order exists
23275 No order of elaboration exists which follows the rules, taking into
23276 account any @code{Elaborate}, @code{Elaborate_All},
23277 or @code{Elaborate_Body} pragmas. In
23278 this case, an Ada compiler must diagnose the situation at bind
23279 time, and refuse to build an executable program.
23281 @item One or more orders exist, all incorrect
23282 One or more acceptable elaboration orders exist, and all of them
23283 generate an elaboration order problem. In this case, the binder
23284 can build an executable program, but @code{Program_Error} will be raised
23285 when the program is run.
23287 @item Several orders exist, some right, some incorrect
23288 One or more acceptable elaboration orders exists, and some of them
23289 work, and some do not. The programmer has not controlled
23290 the order of elaboration, so the binder may or may not pick one of
23291 the correct orders, and the program may or may not raise an
23292 exception when it is run. This is the worst case, because it means
23293 that the program may fail when moved to another compiler, or even
23294 another version of the same compiler.
23296 @item One or more orders exists, all correct
23297 One ore more acceptable elaboration orders exist, and all of them
23298 work. In this case the program runs successfully. This state of
23299 affairs can be guaranteed by following the rule we gave above, but
23300 may be true even if the rule is not followed.
23304 Note that one additional advantage of following our rules on the use
23305 of @code{Elaborate} and @code{Elaborate_All}
23306 is that the program continues to stay in the ideal (all orders OK) state
23307 even if maintenance
23308 changes some bodies of some units. Conversely, if a program that does
23309 not follow this rule happens to be safe at some point, this state of affairs
23310 may deteriorate silently as a result of maintenance changes.
23312 You may have noticed that the above discussion did not mention
23313 the use of @code{Elaborate_Body}. This was a deliberate omission. If you
23314 @code{with} an @code{Elaborate_Body} unit, it still may be the case that
23315 code in the body makes calls to some other unit, so it is still necessary
23316 to use @code{Elaborate_All} on such units.
23318 @node Controlling Elaboration in GNAT - Internal Calls
23319 @section Controlling Elaboration in GNAT - Internal Calls
23322 In the case of internal calls, i.e., calls within a single package, the
23323 programmer has full control over the order of elaboration, and it is up
23324 to the programmer to elaborate declarations in an appropriate order. For
23327 @smallexample @c ada
23330 function One return Float;
23334 function One return Float is
23343 will obviously raise @code{Program_Error} at run time, because function
23344 One will be called before its body is elaborated. In this case GNAT will
23345 generate a warning that the call will raise @code{Program_Error}:
23351 2. function One return Float;
23353 4. Q : Float := One;
23355 >>> warning: cannot call "One" before body is elaborated
23356 >>> warning: Program_Error will be raised at run time
23359 6. function One return Float is
23372 Note that in this particular case, it is likely that the call is safe, because
23373 the function @code{One} does not access any global variables.
23374 Nevertheless in Ada, we do not want the validity of the check to depend on
23375 the contents of the body (think about the separate compilation case), so this
23376 is still wrong, as we discussed in the previous sections.
23378 The error is easily corrected by rearranging the declarations so that the
23379 body of @code{One} appears before the declaration containing the call
23380 (note that in Ada 95 and Ada 2005,
23381 declarations can appear in any order, so there is no restriction that
23382 would prevent this reordering, and if we write:
23384 @smallexample @c ada
23387 function One return Float;
23389 function One return Float is
23400 then all is well, no warning is generated, and no
23401 @code{Program_Error} exception
23403 Things are more complicated when a chain of subprograms is executed:
23405 @smallexample @c ada
23408 function A return Integer;
23409 function B return Integer;
23410 function C return Integer;
23412 function B return Integer is begin return A; end;
23413 function C return Integer is begin return B; end;
23417 function A return Integer is begin return 1; end;
23423 Now the call to @code{C}
23424 at elaboration time in the declaration of @code{X} is correct, because
23425 the body of @code{C} is already elaborated,
23426 and the call to @code{B} within the body of
23427 @code{C} is correct, but the call
23428 to @code{A} within the body of @code{B} is incorrect, because the body
23429 of @code{A} has not been elaborated, so @code{Program_Error}
23430 will be raised on the call to @code{A}.
23431 In this case GNAT will generate a
23432 warning that @code{Program_Error} may be
23433 raised at the point of the call. Let's look at the warning:
23439 2. function A return Integer;
23440 3. function B return Integer;
23441 4. function C return Integer;
23443 6. function B return Integer is begin return A; end;
23445 >>> warning: call to "A" before body is elaborated may
23446 raise Program_Error
23447 >>> warning: "B" called at line 7
23448 >>> warning: "C" called at line 9
23450 7. function C return Integer is begin return B; end;
23452 9. X : Integer := C;
23454 11. function A return Integer is begin return 1; end;
23464 Note that the message here says ``may raise'', instead of the direct case,
23465 where the message says ``will be raised''. That's because whether
23467 actually called depends in general on run-time flow of control.
23468 For example, if the body of @code{B} said
23470 @smallexample @c ada
23473 function B return Integer is
23475 if some-condition-depending-on-input-data then
23486 then we could not know until run time whether the incorrect call to A would
23487 actually occur, so @code{Program_Error} might
23488 or might not be raised. It is possible for a compiler to
23489 do a better job of analyzing bodies, to
23490 determine whether or not @code{Program_Error}
23491 might be raised, but it certainly
23492 couldn't do a perfect job (that would require solving the halting problem
23493 and is provably impossible), and because this is a warning anyway, it does
23494 not seem worth the effort to do the analysis. Cases in which it
23495 would be relevant are rare.
23497 In practice, warnings of either of the forms given
23498 above will usually correspond to
23499 real errors, and should be examined carefully and eliminated.
23500 In the rare case where a warning is bogus, it can be suppressed by any of
23501 the following methods:
23505 Compile with the @option{-gnatws} switch set
23508 Suppress @code{Elaboration_Check} for the called subprogram
23511 Use pragma @code{Warnings_Off} to turn warnings off for the call
23515 For the internal elaboration check case,
23516 GNAT by default generates the
23517 necessary run-time checks to ensure
23518 that @code{Program_Error} is raised if any
23519 call fails an elaboration check. Of course this can only happen if a
23520 warning has been issued as described above. The use of pragma
23521 @code{Suppress (Elaboration_Check)} may (but is not guaranteed to) suppress
23522 some of these checks, meaning that it may be possible (but is not
23523 guaranteed) for a program to be able to call a subprogram whose body
23524 is not yet elaborated, without raising a @code{Program_Error} exception.
23526 @node Controlling Elaboration in GNAT - External Calls
23527 @section Controlling Elaboration in GNAT - External Calls
23530 The previous section discussed the case in which the execution of a
23531 particular thread of elaboration code occurred entirely within a
23532 single unit. This is the easy case to handle, because a programmer
23533 has direct and total control over the order of elaboration, and
23534 furthermore, checks need only be generated in cases which are rare
23535 and which the compiler can easily detect.
23536 The situation is more complex when separate compilation is taken into account.
23537 Consider the following:
23539 @smallexample @c ada
23543 function Sqrt (Arg : Float) return Float;
23546 package body Math is
23547 function Sqrt (Arg : Float) return Float is
23556 X : Float := Math.Sqrt (0.5);
23569 where @code{Main} is the main program. When this program is executed, the
23570 elaboration code must first be executed, and one of the jobs of the
23571 binder is to determine the order in which the units of a program are
23572 to be elaborated. In this case we have four units: the spec and body
23574 the spec of @code{Stuff} and the body of @code{Main}).
23575 In what order should the four separate sections of elaboration code
23578 There are some restrictions in the order of elaboration that the binder
23579 can choose. In particular, if unit U has a @code{with}
23580 for a package @code{X}, then you
23581 are assured that the spec of @code{X}
23582 is elaborated before U , but you are
23583 not assured that the body of @code{X}
23584 is elaborated before U.
23585 This means that in the above case, the binder is allowed to choose the
23596 but that's not good, because now the call to @code{Math.Sqrt}
23597 that happens during
23598 the elaboration of the @code{Stuff}
23599 spec happens before the body of @code{Math.Sqrt} is
23600 elaborated, and hence causes @code{Program_Error} exception to be raised.
23601 At first glance, one might say that the binder is misbehaving, because
23602 obviously you want to elaborate the body of something you @code{with}
23604 that is not a general rule that can be followed in all cases. Consider
23606 @smallexample @c ada
23609 package X is @dots{}
23611 package Y is @dots{}
23614 package body Y is @dots{}
23617 package body X is @dots{}
23623 This is a common arrangement, and, apart from the order of elaboration
23624 problems that might arise in connection with elaboration code, this works fine.
23625 A rule that says that you must first elaborate the body of anything you
23626 @code{with} cannot work in this case:
23627 the body of @code{X} @code{with}'s @code{Y},
23628 which means you would have to
23629 elaborate the body of @code{Y} first, but that @code{with}'s @code{X},
23631 you have to elaborate the body of @code{X} first, but @dots{} and we have a
23632 loop that cannot be broken.
23634 It is true that the binder can in many cases guess an order of elaboration
23635 that is unlikely to cause a @code{Program_Error}
23636 exception to be raised, and it tries to do so (in the
23637 above example of @code{Math/Stuff/Spec}, the GNAT binder will
23639 elaborate the body of @code{Math} right after its spec, so all will be well).
23641 However, a program that blindly relies on the binder to be helpful can
23642 get into trouble, as we discussed in the previous sections, so
23644 provides a number of facilities for assisting the programmer in
23645 developing programs that are robust with respect to elaboration order.
23647 @node Default Behavior in GNAT - Ensuring Safety
23648 @section Default Behavior in GNAT - Ensuring Safety
23651 The default behavior in GNAT ensures elaboration safety. In its
23652 default mode GNAT implements the
23653 rule we previously described as the right approach. Let's restate it:
23657 @emph{If a unit has elaboration code that can directly or indirectly make a
23658 call to a subprogram in a @code{with}'ed unit, or instantiate a generic
23659 package in a @code{with}'ed unit, then if the @code{with}'ed unit
23660 does not have pragma @code{Pure} or
23661 @code{Preelaborate}, then the client should have an
23662 @code{Elaborate_All} pragma for the @code{with}'ed unit.}
23664 @emph{In the case of instantiating a generic subprogram, it is always
23665 sufficient to have only an @code{Elaborate} pragma for the
23666 @code{with}'ed unit.}
23670 By following this rule a client is assured that calls and instantiations
23671 can be made without risk of an exception.
23673 In this mode GNAT traces all calls that are potentially made from
23674 elaboration code, and puts in any missing implicit @code{Elaborate}
23675 and @code{Elaborate_All} pragmas.
23676 The advantage of this approach is that no elaboration problems
23677 are possible if the binder can find an elaboration order that is
23678 consistent with these implicit @code{Elaborate} and
23679 @code{Elaborate_All} pragmas. The
23680 disadvantage of this approach is that no such order may exist.
23682 If the binder does not generate any diagnostics, then it means that it has
23683 found an elaboration order that is guaranteed to be safe. However, the binder
23684 may still be relying on implicitly generated @code{Elaborate} and
23685 @code{Elaborate_All} pragmas so portability to other compilers than GNAT is not
23688 If it is important to guarantee portability, then the compilations should
23691 (warn on elaboration problems) switch. This will cause warning messages
23692 to be generated indicating the missing @code{Elaborate} and
23693 @code{Elaborate_All} pragmas.
23694 Consider the following source program:
23696 @smallexample @c ada
23701 m : integer := k.r;
23708 where it is clear that there
23709 should be a pragma @code{Elaborate_All}
23710 for unit @code{k}. An implicit pragma will be generated, and it is
23711 likely that the binder will be able to honor it. However, if you want
23712 to port this program to some other Ada compiler than GNAT.
23713 it is safer to include the pragma explicitly in the source. If this
23714 unit is compiled with the
23716 switch, then the compiler outputs a warning:
23723 3. m : integer := k.r;
23725 >>> warning: call to "r" may raise Program_Error
23726 >>> warning: missing pragma Elaborate_All for "k"
23734 and these warnings can be used as a guide for supplying manually
23735 the missing pragmas. It is usually a bad idea to use this warning
23736 option during development. That's because it will warn you when
23737 you need to put in a pragma, but cannot warn you when it is time
23738 to take it out. So the use of pragma @code{Elaborate_All} may lead to
23739 unnecessary dependencies and even false circularities.
23741 This default mode is more restrictive than the Ada Reference
23742 Manual, and it is possible to construct programs which will compile
23743 using the dynamic model described there, but will run into a
23744 circularity using the safer static model we have described.
23746 Of course any Ada compiler must be able to operate in a mode
23747 consistent with the requirements of the Ada Reference Manual,
23748 and in particular must have the capability of implementing the
23749 standard dynamic model of elaboration with run-time checks.
23751 In GNAT, this standard mode can be achieved either by the use of
23752 the @option{-gnatE} switch on the compiler (@command{gcc} or
23753 @command{gnatmake}) command, or by the use of the configuration pragma:
23755 @smallexample @c ada
23756 pragma Elaboration_Checks (DYNAMIC);
23760 Either approach will cause the unit affected to be compiled using the
23761 standard dynamic run-time elaboration checks described in the Ada
23762 Reference Manual. The static model is generally preferable, since it
23763 is clearly safer to rely on compile and link time checks rather than
23764 run-time checks. However, in the case of legacy code, it may be
23765 difficult to meet the requirements of the static model. This
23766 issue is further discussed in
23767 @ref{What to Do If the Default Elaboration Behavior Fails}.
23769 Note that the static model provides a strict subset of the allowed
23770 behavior and programs of the Ada Reference Manual, so if you do
23771 adhere to the static model and no circularities exist,
23772 then you are assured that your program will
23773 work using the dynamic model, providing that you remove any
23774 pragma Elaborate statements from the source.
23776 @node Treatment of Pragma Elaborate
23777 @section Treatment of Pragma Elaborate
23778 @cindex Pragma Elaborate
23781 The use of @code{pragma Elaborate}
23782 should generally be avoided in Ada 95 and Ada 2005 programs,
23783 since there is no guarantee that transitive calls
23784 will be properly handled. Indeed at one point, this pragma was placed
23785 in Annex J (Obsolescent Features), on the grounds that it is never useful.
23787 Now that's a bit restrictive. In practice, the case in which
23788 @code{pragma Elaborate} is useful is when the caller knows that there
23789 are no transitive calls, or that the called unit contains all necessary
23790 transitive @code{pragma Elaborate} statements, and legacy code often
23791 contains such uses.
23793 Strictly speaking the static mode in GNAT should ignore such pragmas,
23794 since there is no assurance at compile time that the necessary safety
23795 conditions are met. In practice, this would cause GNAT to be incompatible
23796 with correctly written Ada 83 code that had all necessary
23797 @code{pragma Elaborate} statements in place. Consequently, we made the
23798 decision that GNAT in its default mode will believe that if it encounters
23799 a @code{pragma Elaborate} then the programmer knows what they are doing,
23800 and it will trust that no elaboration errors can occur.
23802 The result of this decision is two-fold. First to be safe using the
23803 static mode, you should remove all @code{pragma Elaborate} statements.
23804 Second, when fixing circularities in existing code, you can selectively
23805 use @code{pragma Elaborate} statements to convince the static mode of
23806 GNAT that it need not generate an implicit @code{pragma Elaborate_All}
23809 When using the static mode with @option{-gnatwl}, any use of
23810 @code{pragma Elaborate} will generate a warning about possible
23813 @node Elaboration Issues for Library Tasks
23814 @section Elaboration Issues for Library Tasks
23815 @cindex Library tasks, elaboration issues
23816 @cindex Elaboration of library tasks
23819 In this section we examine special elaboration issues that arise for
23820 programs that declare library level tasks.
23822 Generally the model of execution of an Ada program is that all units are
23823 elaborated, and then execution of the program starts. However, the
23824 declaration of library tasks definitely does not fit this model. The
23825 reason for this is that library tasks start as soon as they are declared
23826 (more precisely, as soon as the statement part of the enclosing package
23827 body is reached), that is to say before elaboration
23828 of the program is complete. This means that if such a task calls a
23829 subprogram, or an entry in another task, the callee may or may not be
23830 elaborated yet, and in the standard
23831 Reference Manual model of dynamic elaboration checks, you can even
23832 get timing dependent Program_Error exceptions, since there can be
23833 a race between the elaboration code and the task code.
23835 The static model of elaboration in GNAT seeks to avoid all such
23836 dynamic behavior, by being conservative, and the conservative
23837 approach in this particular case is to assume that all the code
23838 in a task body is potentially executed at elaboration time if
23839 a task is declared at the library level.
23841 This can definitely result in unexpected circularities. Consider
23842 the following example
23844 @smallexample @c ada
23850 type My_Int is new Integer;
23852 function Ident (M : My_Int) return My_Int;
23856 package body Decls is
23857 task body Lib_Task is
23863 function Ident (M : My_Int) return My_Int is
23871 procedure Put_Val (Arg : Decls.My_Int);
23875 package body Utils is
23876 procedure Put_Val (Arg : Decls.My_Int) is
23878 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
23885 Decls.Lib_Task.Start;
23890 If the above example is compiled in the default static elaboration
23891 mode, then a circularity occurs. The circularity comes from the call
23892 @code{Utils.Put_Val} in the task body of @code{Decls.Lib_Task}. Since
23893 this call occurs in elaboration code, we need an implicit pragma
23894 @code{Elaborate_All} for @code{Utils}. This means that not only must
23895 the spec and body of @code{Utils} be elaborated before the body
23896 of @code{Decls}, but also the spec and body of any unit that is
23897 @code{with'ed} by the body of @code{Utils} must also be elaborated before
23898 the body of @code{Decls}. This is the transitive implication of
23899 pragma @code{Elaborate_All} and it makes sense, because in general
23900 the body of @code{Put_Val} might have a call to something in a
23901 @code{with'ed} unit.
23903 In this case, the body of Utils (actually its spec) @code{with's}
23904 @code{Decls}. Unfortunately this means that the body of @code{Decls}
23905 must be elaborated before itself, in case there is a call from the
23906 body of @code{Utils}.
23908 Here is the exact chain of events we are worrying about:
23912 In the body of @code{Decls} a call is made from within the body of a library
23913 task to a subprogram in the package @code{Utils}. Since this call may
23914 occur at elaboration time (given that the task is activated at elaboration
23915 time), we have to assume the worst, i.e., that the
23916 call does happen at elaboration time.
23919 This means that the body and spec of @code{Util} must be elaborated before
23920 the body of @code{Decls} so that this call does not cause an access before
23924 Within the body of @code{Util}, specifically within the body of
23925 @code{Util.Put_Val} there may be calls to any unit @code{with}'ed
23929 One such @code{with}'ed package is package @code{Decls}, so there
23930 might be a call to a subprogram in @code{Decls} in @code{Put_Val}.
23931 In fact there is such a call in this example, but we would have to
23932 assume that there was such a call even if it were not there, since
23933 we are not supposed to write the body of @code{Decls} knowing what
23934 is in the body of @code{Utils}; certainly in the case of the
23935 static elaboration model, the compiler does not know what is in
23936 other bodies and must assume the worst.
23939 This means that the spec and body of @code{Decls} must also be
23940 elaborated before we elaborate the unit containing the call, but
23941 that unit is @code{Decls}! This means that the body of @code{Decls}
23942 must be elaborated before itself, and that's a circularity.
23946 Indeed, if you add an explicit pragma @code{Elaborate_All} for @code{Utils} in
23947 the body of @code{Decls} you will get a true Ada Reference Manual
23948 circularity that makes the program illegal.
23950 In practice, we have found that problems with the static model of
23951 elaboration in existing code often arise from library tasks, so
23952 we must address this particular situation.
23954 Note that if we compile and run the program above, using the dynamic model of
23955 elaboration (that is to say use the @option{-gnatE} switch),
23956 then it compiles, binds,
23957 links, and runs, printing the expected result of 2. Therefore in some sense
23958 the circularity here is only apparent, and we need to capture
23959 the properties of this program that distinguish it from other library-level
23960 tasks that have real elaboration problems.
23962 We have four possible answers to this question:
23967 Use the dynamic model of elaboration.
23969 If we use the @option{-gnatE} switch, then as noted above, the program works.
23970 Why is this? If we examine the task body, it is apparent that the task cannot
23972 @code{accept} statement until after elaboration has been completed, because
23973 the corresponding entry call comes from the main program, not earlier.
23974 This is why the dynamic model works here. But that's really giving
23975 up on a precise analysis, and we prefer to take this approach only if we cannot
23977 problem in any other manner. So let us examine two ways to reorganize
23978 the program to avoid the potential elaboration problem.
23981 Split library tasks into separate packages.
23983 Write separate packages, so that library tasks are isolated from
23984 other declarations as much as possible. Let us look at a variation on
23987 @smallexample @c ada
23995 package body Decls1 is
23996 task body Lib_Task is
24004 type My_Int is new Integer;
24005 function Ident (M : My_Int) return My_Int;
24009 package body Decls2 is
24010 function Ident (M : My_Int) return My_Int is
24018 procedure Put_Val (Arg : Decls2.My_Int);
24022 package body Utils is
24023 procedure Put_Val (Arg : Decls2.My_Int) is
24025 Text_IO.Put_Line (Decls2.My_Int'Image (Decls2.Ident (Arg)));
24032 Decls1.Lib_Task.Start;
24037 All we have done is to split @code{Decls} into two packages, one
24038 containing the library task, and one containing everything else. Now
24039 there is no cycle, and the program compiles, binds, links and executes
24040 using the default static model of elaboration.
24043 Declare separate task types.
24045 A significant part of the problem arises because of the use of the
24046 single task declaration form. This means that the elaboration of
24047 the task type, and the elaboration of the task itself (i.e.@: the
24048 creation of the task) happen at the same time. A good rule
24049 of style in Ada is to always create explicit task types. By
24050 following the additional step of placing task objects in separate
24051 packages from the task type declaration, many elaboration problems
24052 are avoided. Here is another modified example of the example program:
24054 @smallexample @c ada
24056 task type Lib_Task_Type is
24060 type My_Int is new Integer;
24062 function Ident (M : My_Int) return My_Int;
24066 package body Decls is
24067 task body Lib_Task_Type is
24073 function Ident (M : My_Int) return My_Int is
24081 procedure Put_Val (Arg : Decls.My_Int);
24085 package body Utils is
24086 procedure Put_Val (Arg : Decls.My_Int) is
24088 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
24094 Lib_Task : Decls.Lib_Task_Type;
24100 Declst.Lib_Task.Start;
24105 What we have done here is to replace the @code{task} declaration in
24106 package @code{Decls} with a @code{task type} declaration. Then we
24107 introduce a separate package @code{Declst} to contain the actual
24108 task object. This separates the elaboration issues for
24109 the @code{task type}
24110 declaration, which causes no trouble, from the elaboration issues
24111 of the task object, which is also unproblematic, since it is now independent
24112 of the elaboration of @code{Utils}.
24113 This separation of concerns also corresponds to
24114 a generally sound engineering principle of separating declarations
24115 from instances. This version of the program also compiles, binds, links,
24116 and executes, generating the expected output.
24119 Use No_Entry_Calls_In_Elaboration_Code restriction.
24120 @cindex No_Entry_Calls_In_Elaboration_Code
24122 The previous two approaches described how a program can be restructured
24123 to avoid the special problems caused by library task bodies. in practice,
24124 however, such restructuring may be difficult to apply to existing legacy code,
24125 so we must consider solutions that do not require massive rewriting.
24127 Let us consider more carefully why our original sample program works
24128 under the dynamic model of elaboration. The reason is that the code
24129 in the task body blocks immediately on the @code{accept}
24130 statement. Now of course there is nothing to prohibit elaboration
24131 code from making entry calls (for example from another library level task),
24132 so we cannot tell in isolation that
24133 the task will not execute the accept statement during elaboration.
24135 However, in practice it is very unusual to see elaboration code
24136 make any entry calls, and the pattern of tasks starting
24137 at elaboration time and then immediately blocking on @code{accept} or
24138 @code{select} statements is very common. What this means is that
24139 the compiler is being too pessimistic when it analyzes the
24140 whole package body as though it might be executed at elaboration
24143 If we know that the elaboration code contains no entry calls, (a very safe
24144 assumption most of the time, that could almost be made the default
24145 behavior), then we can compile all units of the program under control
24146 of the following configuration pragma:
24149 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
24153 This pragma can be placed in the @file{gnat.adc} file in the usual
24154 manner. If we take our original unmodified program and compile it
24155 in the presence of a @file{gnat.adc} containing the above pragma,
24156 then once again, we can compile, bind, link, and execute, obtaining
24157 the expected result. In the presence of this pragma, the compiler does
24158 not trace calls in a task body, that appear after the first @code{accept}
24159 or @code{select} statement, and therefore does not report a potential
24160 circularity in the original program.
24162 The compiler will check to the extent it can that the above
24163 restriction is not violated, but it is not always possible to do a
24164 complete check at compile time, so it is important to use this
24165 pragma only if the stated restriction is in fact met, that is to say
24166 no task receives an entry call before elaboration of all units is completed.
24170 @node Mixing Elaboration Models
24171 @section Mixing Elaboration Models
24173 So far, we have assumed that the entire program is either compiled
24174 using the dynamic model or static model, ensuring consistency. It
24175 is possible to mix the two models, but rules have to be followed
24176 if this mixing is done to ensure that elaboration checks are not
24179 The basic rule is that @emph{a unit compiled with the static model cannot
24180 be @code{with'ed} by a unit compiled with the dynamic model}. The
24181 reason for this is that in the static model, a unit assumes that
24182 its clients guarantee to use (the equivalent of) pragma
24183 @code{Elaborate_All} so that no elaboration checks are required
24184 in inner subprograms, and this assumption is violated if the
24185 client is compiled with dynamic checks.
24187 The precise rule is as follows. A unit that is compiled with dynamic
24188 checks can only @code{with} a unit that meets at least one of the
24189 following criteria:
24194 The @code{with'ed} unit is itself compiled with dynamic elaboration
24195 checks (that is with the @option{-gnatE} switch.
24198 The @code{with'ed} unit is an internal GNAT implementation unit from
24199 the System, Interfaces, Ada, or GNAT hierarchies.
24202 The @code{with'ed} unit has pragma Preelaborate or pragma Pure.
24205 The @code{with'ing} unit (that is the client) has an explicit pragma
24206 @code{Elaborate_All} for the @code{with'ed} unit.
24211 If this rule is violated, that is if a unit with dynamic elaboration
24212 checks @code{with's} a unit that does not meet one of the above four
24213 criteria, then the binder (@code{gnatbind}) will issue a warning
24214 similar to that in the following example:
24217 warning: "x.ads" has dynamic elaboration checks and with's
24218 warning: "y.ads" which has static elaboration checks
24222 These warnings indicate that the rule has been violated, and that as a result
24223 elaboration checks may be missed in the resulting executable file.
24224 This warning may be suppressed using the @option{-ws} binder switch
24225 in the usual manner.
24227 One useful application of this mixing rule is in the case of a subsystem
24228 which does not itself @code{with} units from the remainder of the
24229 application. In this case, the entire subsystem can be compiled with
24230 dynamic checks to resolve a circularity in the subsystem, while
24231 allowing the main application that uses this subsystem to be compiled
24232 using the more reliable default static model.
24234 @node What to Do If the Default Elaboration Behavior Fails
24235 @section What to Do If the Default Elaboration Behavior Fails
24238 If the binder cannot find an acceptable order, it outputs detailed
24239 diagnostics. For example:
24245 error: elaboration circularity detected
24246 info: "proc (body)" must be elaborated before "pack (body)"
24247 info: reason: Elaborate_All probably needed in unit "pack (body)"
24248 info: recompile "pack (body)" with -gnatwl
24249 info: for full details
24250 info: "proc (body)"
24251 info: is needed by its spec:
24252 info: "proc (spec)"
24253 info: which is withed by:
24254 info: "pack (body)"
24255 info: "pack (body)" must be elaborated before "proc (body)"
24256 info: reason: pragma Elaborate in unit "proc (body)"
24262 In this case we have a cycle that the binder cannot break. On the one
24263 hand, there is an explicit pragma Elaborate in @code{proc} for
24264 @code{pack}. This means that the body of @code{pack} must be elaborated
24265 before the body of @code{proc}. On the other hand, there is elaboration
24266 code in @code{pack} that calls a subprogram in @code{proc}. This means
24267 that for maximum safety, there should really be a pragma
24268 Elaborate_All in @code{pack} for @code{proc} which would require that
24269 the body of @code{proc} be elaborated before the body of
24270 @code{pack}. Clearly both requirements cannot be satisfied.
24271 Faced with a circularity of this kind, you have three different options.
24274 @item Fix the program
24275 The most desirable option from the point of view of long-term maintenance
24276 is to rearrange the program so that the elaboration problems are avoided.
24277 One useful technique is to place the elaboration code into separate
24278 child packages. Another is to move some of the initialization code to
24279 explicitly called subprograms, where the program controls the order
24280 of initialization explicitly. Although this is the most desirable option,
24281 it may be impractical and involve too much modification, especially in
24282 the case of complex legacy code.
24284 @item Perform dynamic checks
24285 If the compilations are done using the
24287 (dynamic elaboration check) switch, then GNAT behaves in a quite different
24288 manner. Dynamic checks are generated for all calls that could possibly result
24289 in raising an exception. With this switch, the compiler does not generate
24290 implicit @code{Elaborate} or @code{Elaborate_All} pragmas. The behavior then is
24291 exactly as specified in the @cite{Ada Reference Manual}.
24292 The binder will generate
24293 an executable program that may or may not raise @code{Program_Error}, and then
24294 it is the programmer's job to ensure that it does not raise an exception. Note
24295 that it is important to compile all units with the switch, it cannot be used
24298 @item Suppress checks
24299 The drawback of dynamic checks is that they generate a
24300 significant overhead at run time, both in space and time. If you
24301 are absolutely sure that your program cannot raise any elaboration
24302 exceptions, and you still want to use the dynamic elaboration model,
24303 then you can use the configuration pragma
24304 @code{Suppress (Elaboration_Check)} to suppress all such checks. For
24305 example this pragma could be placed in the @file{gnat.adc} file.
24307 @item Suppress checks selectively
24308 When you know that certain calls or instantiations in elaboration code cannot
24309 possibly lead to an elaboration error, and the binder nevertheless complains
24310 about implicit @code{Elaborate} and @code{Elaborate_All} pragmas that lead to
24311 elaboration circularities, it is possible to remove those warnings locally and
24312 obtain a program that will bind. Clearly this can be unsafe, and it is the
24313 responsibility of the programmer to make sure that the resulting program has no
24314 elaboration anomalies. The pragma @code{Suppress (Elaboration_Check)} can be
24315 used with different granularity to suppress warnings and break elaboration
24320 Place the pragma that names the called subprogram in the declarative part
24321 that contains the call.
24324 Place the pragma in the declarative part, without naming an entity. This
24325 disables warnings on all calls in the corresponding declarative region.
24328 Place the pragma in the package spec that declares the called subprogram,
24329 and name the subprogram. This disables warnings on all elaboration calls to
24333 Place the pragma in the package spec that declares the called subprogram,
24334 without naming any entity. This disables warnings on all elaboration calls to
24335 all subprograms declared in this spec.
24337 @item Use Pragma Elaborate
24338 As previously described in section @xref{Treatment of Pragma Elaborate},
24339 GNAT in static mode assumes that a @code{pragma} Elaborate indicates correctly
24340 that no elaboration checks are required on calls to the designated unit.
24341 There may be cases in which the caller knows that no transitive calls
24342 can occur, so that a @code{pragma Elaborate} will be sufficient in a
24343 case where @code{pragma Elaborate_All} would cause a circularity.
24347 These five cases are listed in order of decreasing safety, and therefore
24348 require increasing programmer care in their application. Consider the
24351 @smallexample @c adanocomment
24353 function F1 return Integer;
24358 function F2 return Integer;
24359 function Pure (x : integer) return integer;
24360 -- pragma Suppress (Elaboration_Check, On => Pure); -- (3)
24361 -- pragma Suppress (Elaboration_Check); -- (4)
24365 package body Pack1 is
24366 function F1 return Integer is
24370 Val : integer := Pack2.Pure (11); -- Elab. call (1)
24373 -- pragma Suppress(Elaboration_Check, Pack2.F2); -- (1)
24374 -- pragma Suppress(Elaboration_Check); -- (2)
24376 X1 := Pack2.F2 + 1; -- Elab. call (2)
24381 package body Pack2 is
24382 function F2 return Integer is
24386 function Pure (x : integer) return integer is
24388 return x ** 3 - 3 * x;
24392 with Pack1, Ada.Text_IO;
24395 Ada.Text_IO.Put_Line(Pack1.X1'Img); -- 101
24398 In the absence of any pragmas, an attempt to bind this program produces
24399 the following diagnostics:
24405 error: elaboration circularity detected
24406 info: "pack1 (body)" must be elaborated before "pack1 (body)"
24407 info: reason: Elaborate_All probably needed in unit "pack1 (body)"
24408 info: recompile "pack1 (body)" with -gnatwl for full details
24409 info: "pack1 (body)"
24410 info: must be elaborated along with its spec:
24411 info: "pack1 (spec)"
24412 info: which is withed by:
24413 info: "pack2 (body)"
24414 info: which must be elaborated along with its spec:
24415 info: "pack2 (spec)"
24416 info: which is withed by:
24417 info: "pack1 (body)"
24420 The sources of the circularity are the two calls to @code{Pack2.Pure} and
24421 @code{Pack2.F2} in the body of @code{Pack1}. We can see that the call to
24422 F2 is safe, even though F2 calls F1, because the call appears after the
24423 elaboration of the body of F1. Therefore the pragma (1) is safe, and will
24424 remove the warning on the call. It is also possible to use pragma (2)
24425 because there are no other potentially unsafe calls in the block.
24428 The call to @code{Pure} is safe because this function does not depend on the
24429 state of @code{Pack2}. Therefore any call to this function is safe, and it
24430 is correct to place pragma (3) in the corresponding package spec.
24433 Finally, we could place pragma (4) in the spec of @code{Pack2} to disable
24434 warnings on all calls to functions declared therein. Note that this is not
24435 necessarily safe, and requires more detailed examination of the subprogram
24436 bodies involved. In particular, a call to @code{F2} requires that @code{F1}
24437 be already elaborated.
24441 It is hard to generalize on which of these four approaches should be
24442 taken. Obviously if it is possible to fix the program so that the default
24443 treatment works, this is preferable, but this may not always be practical.
24444 It is certainly simple enough to use
24446 but the danger in this case is that, even if the GNAT binder
24447 finds a correct elaboration order, it may not always do so,
24448 and certainly a binder from another Ada compiler might not. A
24449 combination of testing and analysis (for which the warnings generated
24452 switch can be useful) must be used to ensure that the program is free
24453 of errors. One switch that is useful in this testing is the
24454 @option{^-p (pessimistic elaboration order)^/PESSIMISTIC_ELABORATION_ORDER^}
24457 Normally the binder tries to find an order that has the best chance
24458 of avoiding elaboration problems. However, if this switch is used, the binder
24459 plays a devil's advocate role, and tries to choose the order that
24460 has the best chance of failing. If your program works even with this
24461 switch, then it has a better chance of being error free, but this is still
24464 For an example of this approach in action, consider the C-tests (executable
24465 tests) from the ACVC suite. If these are compiled and run with the default
24466 treatment, then all but one of them succeed without generating any error
24467 diagnostics from the binder. However, there is one test that fails, and
24468 this is not surprising, because the whole point of this test is to ensure
24469 that the compiler can handle cases where it is impossible to determine
24470 a correct order statically, and it checks that an exception is indeed
24471 raised at run time.
24473 This one test must be compiled and run using the
24475 switch, and then it passes. Alternatively, the entire suite can
24476 be run using this switch. It is never wrong to run with the dynamic
24477 elaboration switch if your code is correct, and we assume that the
24478 C-tests are indeed correct (it is less efficient, but efficiency is
24479 not a factor in running the ACVC tests.)
24481 @node Elaboration for Access-to-Subprogram Values
24482 @section Elaboration for Access-to-Subprogram Values
24483 @cindex Access-to-subprogram
24486 Access-to-subprogram types (introduced in Ada 95) complicate
24487 the handling of elaboration. The trouble is that it becomes
24488 impossible to tell at compile time which procedure
24489 is being called. This means that it is not possible for the binder
24490 to analyze the elaboration requirements in this case.
24492 If at the point at which the access value is created
24493 (i.e., the evaluation of @code{P'Access} for a subprogram @code{P}),
24494 the body of the subprogram is
24495 known to have been elaborated, then the access value is safe, and its use
24496 does not require a check. This may be achieved by appropriate arrangement
24497 of the order of declarations if the subprogram is in the current unit,
24498 or, if the subprogram is in another unit, by using pragma
24499 @code{Pure}, @code{Preelaborate}, or @code{Elaborate_Body}
24500 on the referenced unit.
24502 If the referenced body is not known to have been elaborated at the point
24503 the access value is created, then any use of the access value must do a
24504 dynamic check, and this dynamic check will fail and raise a
24505 @code{Program_Error} exception if the body has not been elaborated yet.
24506 GNAT will generate the necessary checks, and in addition, if the
24508 switch is set, will generate warnings that such checks are required.
24510 The use of dynamic dispatching for tagged types similarly generates
24511 a requirement for dynamic checks, and premature calls to any primitive
24512 operation of a tagged type before the body of the operation has been
24513 elaborated, will result in the raising of @code{Program_Error}.
24515 @node Summary of Procedures for Elaboration Control
24516 @section Summary of Procedures for Elaboration Control
24517 @cindex Elaboration control
24520 First, compile your program with the default options, using none of
24521 the special elaboration control switches. If the binder successfully
24522 binds your program, then you can be confident that, apart from issues
24523 raised by the use of access-to-subprogram types and dynamic dispatching,
24524 the program is free of elaboration errors. If it is important that the
24525 program be portable, then use the
24527 switch to generate warnings about missing @code{Elaborate} or
24528 @code{Elaborate_All} pragmas, and supply the missing pragmas.
24530 If the program fails to bind using the default static elaboration
24531 handling, then you can fix the program to eliminate the binder
24532 message, or recompile the entire program with the
24533 @option{-gnatE} switch to generate dynamic elaboration checks,
24534 and, if you are sure there really are no elaboration problems,
24535 use a global pragma @code{Suppress (Elaboration_Check)}.
24537 @node Other Elaboration Order Considerations
24538 @section Other Elaboration Order Considerations
24540 This section has been entirely concerned with the issue of finding a valid
24541 elaboration order, as defined by the Ada Reference Manual. In a case
24542 where several elaboration orders are valid, the task is to find one
24543 of the possible valid elaboration orders (and the static model in GNAT
24544 will ensure that this is achieved).
24546 The purpose of the elaboration rules in the Ada Reference Manual is to
24547 make sure that no entity is accessed before it has been elaborated. For
24548 a subprogram, this means that the spec and body must have been elaborated
24549 before the subprogram is called. For an object, this means that the object
24550 must have been elaborated before its value is read or written. A violation
24551 of either of these two requirements is an access before elaboration order,
24552 and this section has been all about avoiding such errors.
24554 In the case where more than one order of elaboration is possible, in the
24555 sense that access before elaboration errors are avoided, then any one of
24556 the orders is ``correct'' in the sense that it meets the requirements of
24557 the Ada Reference Manual, and no such error occurs.
24559 However, it may be the case for a given program, that there are
24560 constraints on the order of elaboration that come not from consideration
24561 of avoiding elaboration errors, but rather from extra-lingual logic
24562 requirements. Consider this example:
24564 @smallexample @c ada
24565 with Init_Constants;
24566 package Constants is
24571 package Init_Constants is
24572 procedure P; -- require a body
24573 end Init_Constants;
24576 package body Init_Constants is
24577 procedure P is begin null; end;
24581 end Init_Constants;
24585 Z : Integer := Constants.X + Constants.Y;
24589 with Text_IO; use Text_IO;
24592 Put_Line (Calc.Z'Img);
24597 In this example, there is more than one valid order of elaboration. For
24598 example both the following are correct orders:
24601 Init_Constants spec
24604 Init_Constants body
24609 Init_Constants spec
24610 Init_Constants body
24617 There is no language rule to prefer one or the other, both are correct
24618 from an order of elaboration point of view. But the programmatic effects
24619 of the two orders are very different. In the first, the elaboration routine
24620 of @code{Calc} initializes @code{Z} to zero, and then the main program
24621 runs with this value of zero. But in the second order, the elaboration
24622 routine of @code{Calc} runs after the body of Init_Constants has set
24623 @code{X} and @code{Y} and thus @code{Z} is set to 7 before @code{Main}
24626 One could perhaps by applying pretty clever non-artificial intelligence
24627 to the situation guess that it is more likely that the second order of
24628 elaboration is the one desired, but there is no formal linguistic reason
24629 to prefer one over the other. In fact in this particular case, GNAT will
24630 prefer the second order, because of the rule that bodies are elaborated
24631 as soon as possible, but it's just luck that this is what was wanted
24632 (if indeed the second order was preferred).
24634 If the program cares about the order of elaboration routines in a case like
24635 this, it is important to specify the order required. In this particular
24636 case, that could have been achieved by adding to the spec of Calc:
24638 @smallexample @c ada
24639 pragma Elaborate_All (Constants);
24643 which requires that the body (if any) and spec of @code{Constants},
24644 as well as the body and spec of any unit @code{with}'ed by
24645 @code{Constants} be elaborated before @code{Calc} is elaborated.
24647 Clearly no automatic method can always guess which alternative you require,
24648 and if you are working with legacy code that had constraints of this kind
24649 which were not properly specified by adding @code{Elaborate} or
24650 @code{Elaborate_All} pragmas, then indeed it is possible that two different
24651 compilers can choose different orders.
24653 However, GNAT does attempt to diagnose the common situation where there
24654 are uninitialized variables in the visible part of a package spec, and the
24655 corresponding package body has an elaboration block that directly or
24656 indirectly initialized one or more of these variables. This is the situation
24657 in which a pragma Elaborate_Body is usually desirable, and GNAT will generate
24658 a warning that suggests this addition if it detects this situation.
24660 The @code{gnatbind}
24661 @option{^-p^/PESSIMISTIC_ELABORATION^} switch may be useful in smoking
24662 out problems. This switch causes bodies to be elaborated as late as possible
24663 instead of as early as possible. In the example above, it would have forced
24664 the choice of the first elaboration order. If you get different results
24665 when using this switch, and particularly if one set of results is right,
24666 and one is wrong as far as you are concerned, it shows that you have some
24667 missing @code{Elaborate} pragmas. For the example above, we have the
24671 gnatmake -f -q main
24674 gnatmake -f -q main -bargs -p
24680 It is of course quite unlikely that both these results are correct, so
24681 it is up to you in a case like this to investigate the source of the
24682 difference, by looking at the two elaboration orders that are chosen,
24683 and figuring out which is correct, and then adding the necessary
24684 @code{Elaborate} or @code{Elaborate_All} pragmas to ensure the desired order.
24688 @c *******************************
24689 @node Conditional Compilation
24690 @appendix Conditional Compilation
24691 @c *******************************
24692 @cindex Conditional compilation
24695 It is often necessary to arrange for a single source program
24696 to serve multiple purposes, where it is compiled in different
24697 ways to achieve these different goals. Some examples of the
24698 need for this feature are
24701 @item Adapting a program to a different hardware environment
24702 @item Adapting a program to a different target architecture
24703 @item Turning debugging features on and off
24704 @item Arranging for a program to compile with different compilers
24708 In C, or C++, the typical approach would be to use the preprocessor
24709 that is defined as part of the language. The Ada language does not
24710 contain such a feature. This is not an oversight, but rather a very
24711 deliberate design decision, based on the experience that overuse of
24712 the preprocessing features in C and C++ can result in programs that
24713 are extremely difficult to maintain. For example, if we have ten
24714 switches that can be on or off, this means that there are a thousand
24715 separate programs, any one of which might not even be syntactically
24716 correct, and even if syntactically correct, the resulting program
24717 might not work correctly. Testing all combinations can quickly become
24720 Nevertheless, the need to tailor programs certainly exists, and in
24721 this Appendix we will discuss how this can
24722 be achieved using Ada in general, and GNAT in particular.
24725 * Use of Boolean Constants::
24726 * Debugging - A Special Case::
24727 * Conditionalizing Declarations::
24728 * Use of Alternative Implementations::
24732 @node Use of Boolean Constants
24733 @section Use of Boolean Constants
24736 In the case where the difference is simply which code
24737 sequence is executed, the cleanest solution is to use Boolean
24738 constants to control which code is executed.
24740 @smallexample @c ada
24742 FP_Initialize_Required : constant Boolean := True;
24744 if FP_Initialize_Required then
24751 Not only will the code inside the @code{if} statement not be executed if
24752 the constant Boolean is @code{False}, but it will also be completely
24753 deleted from the program.
24754 However, the code is only deleted after the @code{if} statement
24755 has been checked for syntactic and semantic correctness.
24756 (In contrast, with preprocessors the code is deleted before the
24757 compiler ever gets to see it, so it is not checked until the switch
24759 @cindex Preprocessors (contrasted with conditional compilation)
24761 Typically the Boolean constants will be in a separate package,
24764 @smallexample @c ada
24767 FP_Initialize_Required : constant Boolean := True;
24768 Reset_Available : constant Boolean := False;
24775 The @code{Config} package exists in multiple forms for the various targets,
24776 with an appropriate script selecting the version of @code{Config} needed.
24777 Then any other unit requiring conditional compilation can do a @code{with}
24778 of @code{Config} to make the constants visible.
24781 @node Debugging - A Special Case
24782 @section Debugging - A Special Case
24785 A common use of conditional code is to execute statements (for example
24786 dynamic checks, or output of intermediate results) under control of a
24787 debug switch, so that the debugging behavior can be turned on and off.
24788 This can be done using a Boolean constant to control whether the code
24791 @smallexample @c ada
24794 Put_Line ("got to the first stage!");
24802 @smallexample @c ada
24804 if Debugging and then Temperature > 999.0 then
24805 raise Temperature_Crazy;
24811 Since this is a common case, there are special features to deal with
24812 this in a convenient manner. For the case of tests, Ada 2005 has added
24813 a pragma @code{Assert} that can be used for such tests. This pragma is modeled
24814 @cindex pragma @code{Assert}
24815 on the @code{Assert} pragma that has always been available in GNAT, so this
24816 feature may be used with GNAT even if you are not using Ada 2005 features.
24817 The use of pragma @code{Assert} is described in
24818 @ref{Pragma Assert,,, gnat_rm, GNAT Reference Manual}, but as an
24819 example, the last test could be written:
24821 @smallexample @c ada
24822 pragma Assert (Temperature <= 999.0, "Temperature Crazy");
24828 @smallexample @c ada
24829 pragma Assert (Temperature <= 999.0);
24833 In both cases, if assertions are active and the temperature is excessive,
24834 the exception @code{Assert_Failure} will be raised, with the given string in
24835 the first case or a string indicating the location of the pragma in the second
24836 case used as the exception message.
24838 You can turn assertions on and off by using the @code{Assertion_Policy}
24840 @cindex pragma @code{Assertion_Policy}
24841 This is an Ada 2005 pragma which is implemented in all modes by
24842 GNAT, but only in the latest versions of GNAT which include Ada 2005
24843 capability. Alternatively, you can use the @option{-gnata} switch
24844 @cindex @option{-gnata} switch
24845 to enable assertions from the command line (this is recognized by all versions
24848 For the example above with the @code{Put_Line}, the GNAT-specific pragma
24849 @code{Debug} can be used:
24850 @cindex pragma @code{Debug}
24852 @smallexample @c ada
24853 pragma Debug (Put_Line ("got to the first stage!"));
24857 If debug pragmas are enabled, the argument, which must be of the form of
24858 a procedure call, is executed (in this case, @code{Put_Line} will be called).
24859 Only one call can be present, but of course a special debugging procedure
24860 containing any code you like can be included in the program and then
24861 called in a pragma @code{Debug} argument as needed.
24863 One advantage of pragma @code{Debug} over the @code{if Debugging then}
24864 construct is that pragma @code{Debug} can appear in declarative contexts,
24865 such as at the very beginning of a procedure, before local declarations have
24868 Debug pragmas are enabled using either the @option{-gnata} switch that also
24869 controls assertions, or with a separate Debug_Policy pragma.
24870 @cindex pragma @code{Debug_Policy}
24871 The latter pragma is new in the Ada 2005 versions of GNAT (but it can be used
24872 in Ada 95 and Ada 83 programs as well), and is analogous to
24873 pragma @code{Assertion_Policy} to control assertions.
24875 @code{Assertion_Policy} and @code{Debug_Policy} are configuration pragmas,
24876 and thus they can appear in @file{gnat.adc} if you are not using a
24877 project file, or in the file designated to contain configuration pragmas
24879 They then apply to all subsequent compilations. In practice the use of
24880 the @option{-gnata} switch is often the most convenient method of controlling
24881 the status of these pragmas.
24883 Note that a pragma is not a statement, so in contexts where a statement
24884 sequence is required, you can't just write a pragma on its own. You have
24885 to add a @code{null} statement.
24887 @smallexample @c ada
24890 @dots{} -- some statements
24892 pragma Assert (Num_Cases < 10);
24899 @node Conditionalizing Declarations
24900 @section Conditionalizing Declarations
24903 In some cases, it may be necessary to conditionalize declarations to meet
24904 different requirements. For example we might want a bit string whose length
24905 is set to meet some hardware message requirement.
24907 In some cases, it may be possible to do this using declare blocks controlled
24908 by conditional constants:
24910 @smallexample @c ada
24912 if Small_Machine then
24914 X : Bit_String (1 .. 10);
24920 X : Large_Bit_String (1 .. 1000);
24929 Note that in this approach, both declarations are analyzed by the
24930 compiler so this can only be used where both declarations are legal,
24931 even though one of them will not be used.
24933 Another approach is to define integer constants, e.g.@: @code{Bits_Per_Word}, or
24934 Boolean constants, e.g.@: @code{Little_Endian}, and then write declarations
24935 that are parameterized by these constants. For example
24937 @smallexample @c ada
24940 Field1 at 0 range Boolean'Pos (Little_Endian) * 10 .. Bits_Per_Word;
24946 If @code{Bits_Per_Word} is set to 32, this generates either
24948 @smallexample @c ada
24951 Field1 at 0 range 0 .. 32;
24957 for the big endian case, or
24959 @smallexample @c ada
24962 Field1 at 0 range 10 .. 32;
24968 for the little endian case. Since a powerful subset of Ada expression
24969 notation is usable for creating static constants, clever use of this
24970 feature can often solve quite difficult problems in conditionalizing
24971 compilation (note incidentally that in Ada 95, the little endian
24972 constant was introduced as @code{System.Default_Bit_Order}, so you do not
24973 need to define this one yourself).
24976 @node Use of Alternative Implementations
24977 @section Use of Alternative Implementations
24980 In some cases, none of the approaches described above are adequate. This
24981 can occur for example if the set of declarations required is radically
24982 different for two different configurations.
24984 In this situation, the official Ada way of dealing with conditionalizing
24985 such code is to write separate units for the different cases. As long as
24986 this does not result in excessive duplication of code, this can be done
24987 without creating maintenance problems. The approach is to share common
24988 code as far as possible, and then isolate the code and declarations
24989 that are different. Subunits are often a convenient method for breaking
24990 out a piece of a unit that is to be conditionalized, with separate files
24991 for different versions of the subunit for different targets, where the
24992 build script selects the right one to give to the compiler.
24993 @cindex Subunits (and conditional compilation)
24995 As an example, consider a situation where a new feature in Ada 2005
24996 allows something to be done in a really nice way. But your code must be able
24997 to compile with an Ada 95 compiler. Conceptually you want to say:
24999 @smallexample @c ada
25002 @dots{} neat Ada 2005 code
25004 @dots{} not quite as neat Ada 95 code
25010 where @code{Ada_2005} is a Boolean constant.
25012 But this won't work when @code{Ada_2005} is set to @code{False},
25013 since the @code{then} clause will be illegal for an Ada 95 compiler.
25014 (Recall that although such unreachable code would eventually be deleted
25015 by the compiler, it still needs to be legal. If it uses features
25016 introduced in Ada 2005, it will be illegal in Ada 95.)
25018 So instead we write
25020 @smallexample @c ada
25021 procedure Insert is separate;
25025 Then we have two files for the subunit @code{Insert}, with the two sets of
25027 If the package containing this is called @code{File_Queries}, then we might
25031 @item @file{file_queries-insert-2005.adb}
25032 @item @file{file_queries-insert-95.adb}
25036 and the build script renames the appropriate file to
25039 file_queries-insert.adb
25043 and then carries out the compilation.
25045 This can also be done with project files' naming schemes. For example:
25047 @smallexample @c project
25048 For Body ("File_Queries.Insert") use "file_queries-insert-2005.ada";
25052 Note also that with project files it is desirable to use a different extension
25053 than @file{ads} / @file{adb} for alternative versions. Otherwise a naming
25054 conflict may arise through another commonly used feature: to declare as part
25055 of the project a set of directories containing all the sources obeying the
25056 default naming scheme.
25058 The use of alternative units is certainly feasible in all situations,
25059 and for example the Ada part of the GNAT run-time is conditionalized
25060 based on the target architecture using this approach. As a specific example,
25061 consider the implementation of the AST feature in VMS. There is one
25069 which is the same for all architectures, and three bodies:
25073 used for all non-VMS operating systems
25074 @item s-asthan-vms-alpha.adb
25075 used for VMS on the Alpha
25076 @item s-asthan-vms-ia64.adb
25077 used for VMS on the ia64
25081 The dummy version @file{s-asthan.adb} simply raises exceptions noting that
25082 this operating system feature is not available, and the two remaining
25083 versions interface with the corresponding versions of VMS to provide
25084 VMS-compatible AST handling. The GNAT build script knows the architecture
25085 and operating system, and automatically selects the right version,
25086 renaming it if necessary to @file{s-asthan.adb} before the run-time build.
25088 Another style for arranging alternative implementations is through Ada's
25089 access-to-subprogram facility.
25090 In case some functionality is to be conditionally included,
25091 you can declare an access-to-procedure variable @code{Ref} that is initialized
25092 to designate a ``do nothing'' procedure, and then invoke @code{Ref.all}
25094 In some library package, set @code{Ref} to @code{Proc'Access} for some
25095 procedure @code{Proc} that performs the relevant processing.
25096 The initialization only occurs if the library package is included in the
25098 The same idea can also be implemented using tagged types and dispatching
25102 @node Preprocessing
25103 @section Preprocessing
25104 @cindex Preprocessing
25107 Although it is quite possible to conditionalize code without the use of
25108 C-style preprocessing, as described earlier in this section, it is
25109 nevertheless convenient in some cases to use the C approach. Moreover,
25110 older Ada compilers have often provided some preprocessing capability,
25111 so legacy code may depend on this approach, even though it is not
25114 To accommodate such use, GNAT provides a preprocessor (modeled to a large
25115 extent on the various preprocessors that have been used
25116 with legacy code on other compilers, to enable easier transition).
25118 The preprocessor may be used in two separate modes. It can be used quite
25119 separately from the compiler, to generate a separate output source file
25120 that is then fed to the compiler as a separate step. This is the
25121 @code{gnatprep} utility, whose use is fully described in
25122 @ref{Preprocessing Using gnatprep}.
25123 @cindex @code{gnatprep}
25125 The preprocessing language allows such constructs as
25129 #if DEBUG or PRIORITY > 4 then
25130 bunch of declarations
25132 completely different bunch of declarations
25138 The values of the symbols @code{DEBUG} and @code{PRIORITY} can be
25139 defined either on the command line or in a separate file.
25141 The other way of running the preprocessor is even closer to the C style and
25142 often more convenient. In this approach the preprocessing is integrated into
25143 the compilation process. The compiler is fed the preprocessor input which
25144 includes @code{#if} lines etc, and then the compiler carries out the
25145 preprocessing internally and processes the resulting output.
25146 For more details on this approach, see @ref{Integrated Preprocessing}.
25149 @c *******************************
25150 @node Inline Assembler
25151 @appendix Inline Assembler
25152 @c *******************************
25155 If you need to write low-level software that interacts directly
25156 with the hardware, Ada provides two ways to incorporate assembly
25157 language code into your program. First, you can import and invoke
25158 external routines written in assembly language, an Ada feature fully
25159 supported by GNAT@. However, for small sections of code it may be simpler
25160 or more efficient to include assembly language statements directly
25161 in your Ada source program, using the facilities of the implementation-defined
25162 package @code{System.Machine_Code}, which incorporates the gcc
25163 Inline Assembler. The Inline Assembler approach offers a number of advantages,
25164 including the following:
25167 @item No need to use non-Ada tools
25168 @item Consistent interface over different targets
25169 @item Automatic usage of the proper calling conventions
25170 @item Access to Ada constants and variables
25171 @item Definition of intrinsic routines
25172 @item Possibility of inlining a subprogram comprising assembler code
25173 @item Code optimizer can take Inline Assembler code into account
25176 This chapter presents a series of examples to show you how to use
25177 the Inline Assembler. Although it focuses on the Intel x86,
25178 the general approach applies also to other processors.
25179 It is assumed that you are familiar with Ada
25180 and with assembly language programming.
25183 * Basic Assembler Syntax::
25184 * A Simple Example of Inline Assembler::
25185 * Output Variables in Inline Assembler::
25186 * Input Variables in Inline Assembler::
25187 * Inlining Inline Assembler Code::
25188 * Other Asm Functionality::
25191 @c ---------------------------------------------------------------------------
25192 @node Basic Assembler Syntax
25193 @section Basic Assembler Syntax
25196 The assembler used by GNAT and gcc is based not on the Intel assembly
25197 language, but rather on a language that descends from the AT&T Unix
25198 assembler @emph{as} (and which is often referred to as ``AT&T syntax'').
25199 The following table summarizes the main features of @emph{as} syntax
25200 and points out the differences from the Intel conventions.
25201 See the gcc @emph{as} and @emph{gas} (an @emph{as} macro
25202 pre-processor) documentation for further information.
25205 @item Register names
25206 gcc / @emph{as}: Prefix with ``%''; for example @code{%eax}
25208 Intel: No extra punctuation; for example @code{eax}
25210 @item Immediate operand
25211 gcc / @emph{as}: Prefix with ``$''; for example @code{$4}
25213 Intel: No extra punctuation; for example @code{4}
25216 gcc / @emph{as}: Prefix with ``$''; for example @code{$loc}
25218 Intel: No extra punctuation; for example @code{loc}
25220 @item Memory contents
25221 gcc / @emph{as}: No extra punctuation; for example @code{loc}
25223 Intel: Square brackets; for example @code{[loc]}
25225 @item Register contents
25226 gcc / @emph{as}: Parentheses; for example @code{(%eax)}
25228 Intel: Square brackets; for example @code{[eax]}
25230 @item Hexadecimal numbers
25231 gcc / @emph{as}: Leading ``0x'' (C language syntax); for example @code{0xA0}
25233 Intel: Trailing ``h''; for example @code{A0h}
25236 gcc / @emph{as}: Explicit in op code; for example @code{movw} to move
25239 Intel: Implicit, deduced by assembler; for example @code{mov}
25241 @item Instruction repetition
25242 gcc / @emph{as}: Split into two lines; for example
25248 Intel: Keep on one line; for example @code{rep stosl}
25250 @item Order of operands
25251 gcc / @emph{as}: Source first; for example @code{movw $4, %eax}
25253 Intel: Destination first; for example @code{mov eax, 4}
25256 @c ---------------------------------------------------------------------------
25257 @node A Simple Example of Inline Assembler
25258 @section A Simple Example of Inline Assembler
25261 The following example will generate a single assembly language statement,
25262 @code{nop}, which does nothing. Despite its lack of run-time effect,
25263 the example will be useful in illustrating the basics of
25264 the Inline Assembler facility.
25266 @smallexample @c ada
25268 with System.Machine_Code; use System.Machine_Code;
25269 procedure Nothing is
25276 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
25277 here it takes one parameter, a @emph{template string} that must be a static
25278 expression and that will form the generated instruction.
25279 @code{Asm} may be regarded as a compile-time procedure that parses
25280 the template string and additional parameters (none here),
25281 from which it generates a sequence of assembly language instructions.
25283 The examples in this chapter will illustrate several of the forms
25284 for invoking @code{Asm}; a complete specification of the syntax
25285 is found in @ref{Machine Code Insertions,,, gnat_rm, GNAT Reference
25288 Under the standard GNAT conventions, the @code{Nothing} procedure
25289 should be in a file named @file{nothing.adb}.
25290 You can build the executable in the usual way:
25294 However, the interesting aspect of this example is not its run-time behavior
25295 but rather the generated assembly code.
25296 To see this output, invoke the compiler as follows:
25298 gcc -c -S -fomit-frame-pointer -gnatp @file{nothing.adb}
25300 where the options are:
25304 compile only (no bind or link)
25306 generate assembler listing
25307 @item -fomit-frame-pointer
25308 do not set up separate stack frames
25310 do not add runtime checks
25313 This gives a human-readable assembler version of the code. The resulting
25314 file will have the same name as the Ada source file, but with a @code{.s}
25315 extension. In our example, the file @file{nothing.s} has the following
25320 .file "nothing.adb"
25322 ___gnu_compiled_ada:
25325 .globl __ada_nothing
25337 The assembly code you included is clearly indicated by
25338 the compiler, between the @code{#APP} and @code{#NO_APP}
25339 delimiters. The character before the 'APP' and 'NOAPP'
25340 can differ on different targets. For example, GNU/Linux uses '#APP' while
25341 on NT you will see '/APP'.
25343 If you make a mistake in your assembler code (such as using the
25344 wrong size modifier, or using a wrong operand for the instruction) GNAT
25345 will report this error in a temporary file, which will be deleted when
25346 the compilation is finished. Generating an assembler file will help
25347 in such cases, since you can assemble this file separately using the
25348 @emph{as} assembler that comes with gcc.
25350 Assembling the file using the command
25353 as @file{nothing.s}
25356 will give you error messages whose lines correspond to the assembler
25357 input file, so you can easily find and correct any mistakes you made.
25358 If there are no errors, @emph{as} will generate an object file
25359 @file{nothing.out}.
25361 @c ---------------------------------------------------------------------------
25362 @node Output Variables in Inline Assembler
25363 @section Output Variables in Inline Assembler
25366 The examples in this section, showing how to access the processor flags,
25367 illustrate how to specify the destination operands for assembly language
25370 @smallexample @c ada
25372 with Interfaces; use Interfaces;
25373 with Ada.Text_IO; use Ada.Text_IO;
25374 with System.Machine_Code; use System.Machine_Code;
25375 procedure Get_Flags is
25376 Flags : Unsigned_32;
25379 Asm ("pushfl" & LF & HT & -- push flags on stack
25380 "popl %%eax" & LF & HT & -- load eax with flags
25381 "movl %%eax, %0", -- store flags in variable
25382 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
25383 Put_Line ("Flags register:" & Flags'Img);
25388 In order to have a nicely aligned assembly listing, we have separated
25389 multiple assembler statements in the Asm template string with linefeed
25390 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
25391 The resulting section of the assembly output file is:
25398 movl %eax, -40(%ebp)
25403 It would have been legal to write the Asm invocation as:
25406 Asm ("pushfl popl %%eax movl %%eax, %0")
25409 but in the generated assembler file, this would come out as:
25413 pushfl popl %eax movl %eax, -40(%ebp)
25417 which is not so convenient for the human reader.
25419 We use Ada comments
25420 at the end of each line to explain what the assembler instructions
25421 actually do. This is a useful convention.
25423 When writing Inline Assembler instructions, you need to precede each register
25424 and variable name with a percent sign. Since the assembler already requires
25425 a percent sign at the beginning of a register name, you need two consecutive
25426 percent signs for such names in the Asm template string, thus @code{%%eax}.
25427 In the generated assembly code, one of the percent signs will be stripped off.
25429 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
25430 variables: operands you later define using @code{Input} or @code{Output}
25431 parameters to @code{Asm}.
25432 An output variable is illustrated in
25433 the third statement in the Asm template string:
25437 The intent is to store the contents of the eax register in a variable that can
25438 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
25439 necessarily work, since the compiler might optimize by using a register
25440 to hold Flags, and the expansion of the @code{movl} instruction would not be
25441 aware of this optimization. The solution is not to store the result directly
25442 but rather to advise the compiler to choose the correct operand form;
25443 that is the purpose of the @code{%0} output variable.
25445 Information about the output variable is supplied in the @code{Outputs}
25446 parameter to @code{Asm}:
25448 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
25451 The output is defined by the @code{Asm_Output} attribute of the target type;
25452 the general format is
25454 Type'Asm_Output (constraint_string, variable_name)
25457 The constraint string directs the compiler how
25458 to store/access the associated variable. In the example
25460 Unsigned_32'Asm_Output ("=m", Flags);
25462 the @code{"m"} (memory) constraint tells the compiler that the variable
25463 @code{Flags} should be stored in a memory variable, thus preventing
25464 the optimizer from keeping it in a register. In contrast,
25466 Unsigned_32'Asm_Output ("=r", Flags);
25468 uses the @code{"r"} (register) constraint, telling the compiler to
25469 store the variable in a register.
25471 If the constraint is preceded by the equal character (@strong{=}), it tells
25472 the compiler that the variable will be used to store data into it.
25474 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
25475 allowing the optimizer to choose whatever it deems best.
25477 There are a fairly large number of constraints, but the ones that are
25478 most useful (for the Intel x86 processor) are the following:
25484 global (i.e.@: can be stored anywhere)
25502 use one of eax, ebx, ecx or edx
25504 use one of eax, ebx, ecx, edx, esi or edi
25507 The full set of constraints is described in the gcc and @emph{as}
25508 documentation; note that it is possible to combine certain constraints
25509 in one constraint string.
25511 You specify the association of an output variable with an assembler operand
25512 through the @code{%}@emph{n} notation, where @emph{n} is a non-negative
25514 @smallexample @c ada
25516 Asm ("pushfl" & LF & HT & -- push flags on stack
25517 "popl %%eax" & LF & HT & -- load eax with flags
25518 "movl %%eax, %0", -- store flags in variable
25519 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
25523 @code{%0} will be replaced in the expanded code by the appropriate operand,
25525 the compiler decided for the @code{Flags} variable.
25527 In general, you may have any number of output variables:
25530 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
25532 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
25533 of @code{Asm_Output} attributes
25537 @smallexample @c ada
25539 Asm ("movl %%eax, %0" & LF & HT &
25540 "movl %%ebx, %1" & LF & HT &
25542 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
25543 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
25544 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
25548 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
25549 in the Ada program.
25551 As a variation on the @code{Get_Flags} example, we can use the constraints
25552 string to direct the compiler to store the eax register into the @code{Flags}
25553 variable, instead of including the store instruction explicitly in the
25554 @code{Asm} template string:
25556 @smallexample @c ada
25558 with Interfaces; use Interfaces;
25559 with Ada.Text_IO; use Ada.Text_IO;
25560 with System.Machine_Code; use System.Machine_Code;
25561 procedure Get_Flags_2 is
25562 Flags : Unsigned_32;
25565 Asm ("pushfl" & LF & HT & -- push flags on stack
25566 "popl %%eax", -- save flags in eax
25567 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
25568 Put_Line ("Flags register:" & Flags'Img);
25574 The @code{"a"} constraint tells the compiler that the @code{Flags}
25575 variable will come from the eax register. Here is the resulting code:
25583 movl %eax,-40(%ebp)
25588 The compiler generated the store of eax into Flags after
25589 expanding the assembler code.
25591 Actually, there was no need to pop the flags into the eax register;
25592 more simply, we could just pop the flags directly into the program variable:
25594 @smallexample @c ada
25596 with Interfaces; use Interfaces;
25597 with Ada.Text_IO; use Ada.Text_IO;
25598 with System.Machine_Code; use System.Machine_Code;
25599 procedure Get_Flags_3 is
25600 Flags : Unsigned_32;
25603 Asm ("pushfl" & LF & HT & -- push flags on stack
25604 "pop %0", -- save flags in Flags
25605 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
25606 Put_Line ("Flags register:" & Flags'Img);
25611 @c ---------------------------------------------------------------------------
25612 @node Input Variables in Inline Assembler
25613 @section Input Variables in Inline Assembler
25616 The example in this section illustrates how to specify the source operands
25617 for assembly language statements.
25618 The program simply increments its input value by 1:
25620 @smallexample @c ada
25622 with Interfaces; use Interfaces;
25623 with Ada.Text_IO; use Ada.Text_IO;
25624 with System.Machine_Code; use System.Machine_Code;
25625 procedure Increment is
25627 function Incr (Value : Unsigned_32) return Unsigned_32 is
25628 Result : Unsigned_32;
25631 Inputs => Unsigned_32'Asm_Input ("a", Value),
25632 Outputs => Unsigned_32'Asm_Output ("=a", Result));
25636 Value : Unsigned_32;
25640 Put_Line ("Value before is" & Value'Img);
25641 Value := Incr (Value);
25642 Put_Line ("Value after is" & Value'Img);
25647 The @code{Outputs} parameter to @code{Asm} specifies
25648 that the result will be in the eax register and that it is to be stored
25649 in the @code{Result} variable.
25651 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
25652 but with an @code{Asm_Input} attribute.
25653 The @code{"="} constraint, indicating an output value, is not present.
25655 You can have multiple input variables, in the same way that you can have more
25656 than one output variable.
25658 The parameter count (%0, %1) etc, now starts at the first input
25659 statement, and continues with the output statements.
25660 When both parameters use the same variable, the
25661 compiler will treat them as the same %n operand, which is the case here.
25663 Just as the @code{Outputs} parameter causes the register to be stored into the
25664 target variable after execution of the assembler statements, so does the
25665 @code{Inputs} parameter cause its variable to be loaded into the register
25666 before execution of the assembler statements.
25668 Thus the effect of the @code{Asm} invocation is:
25670 @item load the 32-bit value of @code{Value} into eax
25671 @item execute the @code{incl %eax} instruction
25672 @item store the contents of eax into the @code{Result} variable
25675 The resulting assembler file (with @option{-O2} optimization) contains:
25678 _increment__incr.1:
25691 @c ---------------------------------------------------------------------------
25692 @node Inlining Inline Assembler Code
25693 @section Inlining Inline Assembler Code
25696 For a short subprogram such as the @code{Incr} function in the previous
25697 section, the overhead of the call and return (creating / deleting the stack
25698 frame) can be significant, compared to the amount of code in the subprogram
25699 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
25700 which directs the compiler to expand invocations of the subprogram at the
25701 point(s) of call, instead of setting up a stack frame for out-of-line calls.
25702 Here is the resulting program:
25704 @smallexample @c ada
25706 with Interfaces; use Interfaces;
25707 with Ada.Text_IO; use Ada.Text_IO;
25708 with System.Machine_Code; use System.Machine_Code;
25709 procedure Increment_2 is
25711 function Incr (Value : Unsigned_32) return Unsigned_32 is
25712 Result : Unsigned_32;
25715 Inputs => Unsigned_32'Asm_Input ("a", Value),
25716 Outputs => Unsigned_32'Asm_Output ("=a", Result));
25719 pragma Inline (Increment);
25721 Value : Unsigned_32;
25725 Put_Line ("Value before is" & Value'Img);
25726 Value := Increment (Value);
25727 Put_Line ("Value after is" & Value'Img);
25732 Compile the program with both optimization (@option{-O2}) and inlining
25733 (@option{-gnatn}) enabled.
25735 The @code{Incr} function is still compiled as usual, but at the
25736 point in @code{Increment} where our function used to be called:
25741 call _increment__incr.1
25746 the code for the function body directly appears:
25759 thus saving the overhead of stack frame setup and an out-of-line call.
25761 @c ---------------------------------------------------------------------------
25762 @node Other Asm Functionality
25763 @section Other @code{Asm} Functionality
25766 This section describes two important parameters to the @code{Asm}
25767 procedure: @code{Clobber}, which identifies register usage;
25768 and @code{Volatile}, which inhibits unwanted optimizations.
25771 * The Clobber Parameter::
25772 * The Volatile Parameter::
25775 @c ---------------------------------------------------------------------------
25776 @node The Clobber Parameter
25777 @subsection The @code{Clobber} Parameter
25780 One of the dangers of intermixing assembly language and a compiled language
25781 such as Ada is that the compiler needs to be aware of which registers are
25782 being used by the assembly code. In some cases, such as the earlier examples,
25783 the constraint string is sufficient to indicate register usage (e.g.,
25785 the eax register). But more generally, the compiler needs an explicit
25786 identification of the registers that are used by the Inline Assembly
25789 Using a register that the compiler doesn't know about
25790 could be a side effect of an instruction (like @code{mull}
25791 storing its result in both eax and edx).
25792 It can also arise from explicit register usage in your
25793 assembly code; for example:
25796 Asm ("movl %0, %%ebx" & LF & HT &
25798 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
25799 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out));
25803 where the compiler (since it does not analyze the @code{Asm} template string)
25804 does not know you are using the ebx register.
25806 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
25807 to identify the registers that will be used by your assembly code:
25811 Asm ("movl %0, %%ebx" & LF & HT &
25813 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
25814 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
25819 The Clobber parameter is a static string expression specifying the
25820 register(s) you are using. Note that register names are @emph{not} prefixed
25821 by a percent sign. Also, if more than one register is used then their names
25822 are separated by commas; e.g., @code{"eax, ebx"}
25824 The @code{Clobber} parameter has several additional uses:
25826 @item Use ``register'' name @code{cc} to indicate that flags might have changed
25827 @item Use ``register'' name @code{memory} if you changed a memory location
25830 @c ---------------------------------------------------------------------------
25831 @node The Volatile Parameter
25832 @subsection The @code{Volatile} Parameter
25833 @cindex Volatile parameter
25836 Compiler optimizations in the presence of Inline Assembler may sometimes have
25837 unwanted effects. For example, when an @code{Asm} invocation with an input
25838 variable is inside a loop, the compiler might move the loading of the input
25839 variable outside the loop, regarding it as a one-time initialization.
25841 If this effect is not desired, you can disable such optimizations by setting
25842 the @code{Volatile} parameter to @code{True}; for example:
25844 @smallexample @c ada
25846 Asm ("movl %0, %%ebx" & LF & HT &
25848 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
25849 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
25855 By default, @code{Volatile} is set to @code{False} unless there is no
25856 @code{Outputs} parameter.
25858 Although setting @code{Volatile} to @code{True} prevents unwanted
25859 optimizations, it will also disable other optimizations that might be
25860 important for efficiency. In general, you should set @code{Volatile}
25861 to @code{True} only if the compiler's optimizations have created
25863 @c END OF INLINE ASSEMBLER CHAPTER
25864 @c ===============================
25866 @c ***********************************
25867 @c * Compatibility and Porting Guide *
25868 @c ***********************************
25869 @node Compatibility and Porting Guide
25870 @appendix Compatibility and Porting Guide
25873 This chapter describes the compatibility issues that may arise between
25874 GNAT and other Ada compilation systems (including those for Ada 83),
25875 and shows how GNAT can expedite porting
25876 applications developed in other Ada environments.
25879 * Compatibility with Ada 83::
25880 * Compatibility between Ada 95 and Ada 2005::
25881 * Implementation-dependent characteristics::
25882 * Compatibility with Other Ada Systems::
25883 * Representation Clauses::
25885 @c Brief section is only in non-VMS version
25886 @c Full chapter is in VMS version
25887 * Compatibility with HP Ada 83::
25890 * Transitioning to 64-Bit GNAT for OpenVMS::
25894 @node Compatibility with Ada 83
25895 @section Compatibility with Ada 83
25896 @cindex Compatibility (between Ada 83 and Ada 95 / Ada 2005)
25899 Ada 95 and Ada 2005 are highly upwards compatible with Ada 83. In
25900 particular, the design intention was that the difficulties associated
25901 with moving from Ada 83 to Ada 95 or Ada 2005 should be no greater than those
25902 that occur when moving from one Ada 83 system to another.
25904 However, there are a number of points at which there are minor
25905 incompatibilities. The @cite{Ada 95 Annotated Reference Manual} contains
25906 full details of these issues,
25907 and should be consulted for a complete treatment.
25909 following subsections treat the most likely issues to be encountered.
25912 * Legal Ada 83 programs that are illegal in Ada 95::
25913 * More deterministic semantics::
25914 * Changed semantics::
25915 * Other language compatibility issues::
25918 @node Legal Ada 83 programs that are illegal in Ada 95
25919 @subsection Legal Ada 83 programs that are illegal in Ada 95
25921 Some legal Ada 83 programs are illegal (i.e., they will fail to compile) in
25922 Ada 95 and thus also in Ada 2005:
25925 @item Character literals
25926 Some uses of character literals are ambiguous. Since Ada 95 has introduced
25927 @code{Wide_Character} as a new predefined character type, some uses of
25928 character literals that were legal in Ada 83 are illegal in Ada 95.
25930 @smallexample @c ada
25931 for Char in 'A' .. 'Z' loop @dots{} end loop;
25935 The problem is that @code{'A'} and @code{'Z'} could be from either
25936 @code{Character} or @code{Wide_Character}. The simplest correction
25937 is to make the type explicit; e.g.:
25938 @smallexample @c ada
25939 for Char in Character range 'A' .. 'Z' loop @dots{} end loop;
25942 @item New reserved words
25943 The identifiers @code{abstract}, @code{aliased}, @code{protected},
25944 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
25945 Existing Ada 83 code using any of these identifiers must be edited to
25946 use some alternative name.
25948 @item Freezing rules
25949 The rules in Ada 95 are slightly different with regard to the point at
25950 which entities are frozen, and representation pragmas and clauses are
25951 not permitted past the freeze point. This shows up most typically in
25952 the form of an error message complaining that a representation item
25953 appears too late, and the appropriate corrective action is to move
25954 the item nearer to the declaration of the entity to which it refers.
25956 A particular case is that representation pragmas
25959 extended HP Ada 83 compatibility pragmas such as @code{Export_Procedure})
25961 cannot be applied to a subprogram body. If necessary, a separate subprogram
25962 declaration must be introduced to which the pragma can be applied.
25964 @item Optional bodies for library packages
25965 In Ada 83, a package that did not require a package body was nevertheless
25966 allowed to have one. This lead to certain surprises in compiling large
25967 systems (situations in which the body could be unexpectedly ignored by the
25968 binder). In Ada 95, if a package does not require a body then it is not
25969 permitted to have a body. To fix this problem, simply remove a redundant
25970 body if it is empty, or, if it is non-empty, introduce a dummy declaration
25971 into the spec that makes the body required. One approach is to add a private
25972 part to the package declaration (if necessary), and define a parameterless
25973 procedure called @code{Requires_Body}, which must then be given a dummy
25974 procedure body in the package body, which then becomes required.
25975 Another approach (assuming that this does not introduce elaboration
25976 circularities) is to add an @code{Elaborate_Body} pragma to the package spec,
25977 since one effect of this pragma is to require the presence of a package body.
25979 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
25980 In Ada 95, the exception @code{Numeric_Error} is a renaming of
25981 @code{Constraint_Error}.
25982 This means that it is illegal to have separate exception handlers for
25983 the two exceptions. The fix is simply to remove the handler for the
25984 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
25985 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
25987 @item Indefinite subtypes in generics
25988 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String})
25989 as the actual for a generic formal private type, but then the instantiation
25990 would be illegal if there were any instances of declarations of variables
25991 of this type in the generic body. In Ada 95, to avoid this clear violation
25992 of the methodological principle known as the ``contract model'',
25993 the generic declaration explicitly indicates whether
25994 or not such instantiations are permitted. If a generic formal parameter
25995 has explicit unknown discriminants, indicated by using @code{(<>)} after the
25996 type name, then it can be instantiated with indefinite types, but no
25997 stand-alone variables can be declared of this type. Any attempt to declare
25998 such a variable will result in an illegality at the time the generic is
25999 declared. If the @code{(<>)} notation is not used, then it is illegal
26000 to instantiate the generic with an indefinite type.
26001 This is the potential incompatibility issue when porting Ada 83 code to Ada 95.
26002 It will show up as a compile time error, and
26003 the fix is usually simply to add the @code{(<>)} to the generic declaration.
26006 @node More deterministic semantics
26007 @subsection More deterministic semantics
26011 Conversions from real types to integer types round away from 0. In Ada 83
26012 the conversion Integer(2.5) could deliver either 2 or 3 as its value. This
26013 implementation freedom was intended to support unbiased rounding in
26014 statistical applications, but in practice it interfered with portability.
26015 In Ada 95 the conversion semantics are unambiguous, and rounding away from 0
26016 is required. Numeric code may be affected by this change in semantics.
26017 Note, though, that this issue is no worse than already existed in Ada 83
26018 when porting code from one vendor to another.
26021 The Real-Time Annex introduces a set of policies that define the behavior of
26022 features that were implementation dependent in Ada 83, such as the order in
26023 which open select branches are executed.
26026 @node Changed semantics
26027 @subsection Changed semantics
26030 The worst kind of incompatibility is one where a program that is legal in
26031 Ada 83 is also legal in Ada 95 but can have an effect in Ada 95 that was not
26032 possible in Ada 83. Fortunately this is extremely rare, but the one
26033 situation that you should be alert to is the change in the predefined type
26034 @code{Character} from 7-bit ASCII to 8-bit Latin-1.
26037 @item Range of type @code{Character}
26038 The range of @code{Standard.Character} is now the full 256 characters
26039 of Latin-1, whereas in most Ada 83 implementations it was restricted
26040 to 128 characters. Although some of the effects of
26041 this change will be manifest in compile-time rejection of legal
26042 Ada 83 programs it is possible for a working Ada 83 program to have
26043 a different effect in Ada 95, one that was not permitted in Ada 83.
26044 As an example, the expression
26045 @code{Character'Pos(Character'Last)} returned @code{127} in Ada 83 and now
26046 delivers @code{255} as its value.
26047 In general, you should look at the logic of any
26048 character-processing Ada 83 program and see whether it needs to be adapted
26049 to work correctly with Latin-1. Note that the predefined Ada 95 API has a
26050 character handling package that may be relevant if code needs to be adapted
26051 to account for the additional Latin-1 elements.
26052 The desirable fix is to
26053 modify the program to accommodate the full character set, but in some cases
26054 it may be convenient to define a subtype or derived type of Character that
26055 covers only the restricted range.
26059 @node Other language compatibility issues
26060 @subsection Other language compatibility issues
26063 @item @option{-gnat83} switch
26064 All implementations of GNAT provide a switch that causes GNAT to operate
26065 in Ada 83 mode. In this mode, some but not all compatibility problems
26066 of the type described above are handled automatically. For example, the
26067 new reserved words introduced in Ada 95 and Ada 2005 are treated simply
26068 as identifiers as in Ada 83.
26070 in practice, it is usually advisable to make the necessary modifications
26071 to the program to remove the need for using this switch.
26072 See @ref{Compiling Different Versions of Ada}.
26074 @item Support for removed Ada 83 pragmas and attributes
26075 A number of pragmas and attributes from Ada 83 were removed from Ada 95,
26076 generally because they were replaced by other mechanisms. Ada 95 and Ada 2005
26077 compilers are allowed, but not required, to implement these missing
26078 elements. In contrast with some other compilers, GNAT implements all
26079 such pragmas and attributes, eliminating this compatibility concern. These
26080 include @code{pragma Interface} and the floating point type attributes
26081 (@code{Emax}, @code{Mantissa}, etc.), among other items.
26085 @node Compatibility between Ada 95 and Ada 2005
26086 @section Compatibility between Ada 95 and Ada 2005
26087 @cindex Compatibility between Ada 95 and Ada 2005
26090 Although Ada 2005 was designed to be upwards compatible with Ada 95, there are
26091 a number of incompatibilities. Several are enumerated below;
26092 for a complete description please see the
26093 Annotated Ada 2005 Reference Manual, or section 9.1.1 in
26094 @cite{Rationale for Ada 2005}.
26097 @item New reserved words.
26098 The words @code{interface}, @code{overriding} and @code{synchronized} are
26099 reserved in Ada 2005.
26100 A pre-Ada 2005 program that uses any of these as an identifier will be
26103 @item New declarations in predefined packages.
26104 A number of packages in the predefined environment contain new declarations:
26105 @code{Ada.Exceptions}, @code{Ada.Real_Time}, @code{Ada.Strings},
26106 @code{Ada.Strings.Fixed}, @code{Ada.Strings.Bounded},
26107 @code{Ada.Strings.Unbounded}, @code{Ada.Strings.Wide_Fixed},
26108 @code{Ada.Strings.Wide_Bounded}, @code{Ada.Strings.Wide_Unbounded},
26109 @code{Ada.Tags}, @code{Ada.Text_IO}, and @code{Interfaces.C}.
26110 If an Ada 95 program does a @code{with} and @code{use} of any of these
26111 packages, the new declarations may cause name clashes.
26113 @item Access parameters.
26114 A nondispatching subprogram with an access parameter cannot be renamed
26115 as a dispatching operation. This was permitted in Ada 95.
26117 @item Access types, discriminants, and constraints.
26118 Rule changes in this area have led to some incompatibilities; for example,
26119 constrained subtypes of some access types are not permitted in Ada 2005.
26121 @item Aggregates for limited types.
26122 The allowance of aggregates for limited types in Ada 2005 raises the
26123 possibility of ambiguities in legal Ada 95 programs, since additional types
26124 now need to be considered in expression resolution.
26126 @item Fixed-point multiplication and division.
26127 Certain expressions involving ``*'' or ``/'' for a fixed-point type, which
26128 were legal in Ada 95 and invoked the predefined versions of these operations,
26130 The ambiguity may be resolved either by applying a type conversion to the
26131 expression, or by explicitly invoking the operation from package
26134 @item Return-by-reference types.
26135 The Ada 95 return-by-reference mechanism has been removed. Instead, the user
26136 can declare a function returning a value from an anonymous access type.
26140 @node Implementation-dependent characteristics
26141 @section Implementation-dependent characteristics
26143 Although the Ada language defines the semantics of each construct as
26144 precisely as practical, in some situations (for example for reasons of
26145 efficiency, or where the effect is heavily dependent on the host or target
26146 platform) the implementation is allowed some freedom. In porting Ada 83
26147 code to GNAT, you need to be aware of whether / how the existing code
26148 exercised such implementation dependencies. Such characteristics fall into
26149 several categories, and GNAT offers specific support in assisting the
26150 transition from certain Ada 83 compilers.
26153 * Implementation-defined pragmas::
26154 * Implementation-defined attributes::
26156 * Elaboration order::
26157 * Target-specific aspects::
26160 @node Implementation-defined pragmas
26161 @subsection Implementation-defined pragmas
26164 Ada compilers are allowed to supplement the language-defined pragmas, and
26165 these are a potential source of non-portability. All GNAT-defined pragmas
26166 are described in @ref{Implementation Defined Pragmas,,, gnat_rm, GNAT
26167 Reference Manual}, and these include several that are specifically
26168 intended to correspond to other vendors' Ada 83 pragmas.
26169 For migrating from VADS, the pragma @code{Use_VADS_Size} may be useful.
26170 For compatibility with HP Ada 83, GNAT supplies the pragmas
26171 @code{Extend_System}, @code{Ident}, @code{Inline_Generic},
26172 @code{Interface_Name}, @code{Passive}, @code{Suppress_All},
26173 and @code{Volatile}.
26174 Other relevant pragmas include @code{External} and @code{Link_With}.
26175 Some vendor-specific
26176 Ada 83 pragmas (@code{Share_Generic}, @code{Subtitle}, and @code{Title}) are
26178 avoiding compiler rejection of units that contain such pragmas; they are not
26179 relevant in a GNAT context and hence are not otherwise implemented.
26181 @node Implementation-defined attributes
26182 @subsection Implementation-defined attributes
26184 Analogous to pragmas, the set of attributes may be extended by an
26185 implementation. All GNAT-defined attributes are described in
26186 @ref{Implementation Defined Attributes,,, gnat_rm, GNAT Reference
26187 Manual}, and these include several that are specifically intended
26188 to correspond to other vendors' Ada 83 attributes. For migrating from VADS,
26189 the attribute @code{VADS_Size} may be useful. For compatibility with HP
26190 Ada 83, GNAT supplies the attributes @code{Bit}, @code{Machine_Size} and
26194 @subsection Libraries
26196 Vendors may supply libraries to supplement the standard Ada API. If Ada 83
26197 code uses vendor-specific libraries then there are several ways to manage
26198 this in Ada 95 or Ada 2005:
26201 If the source code for the libraries (specs and bodies) are
26202 available, then the libraries can be migrated in the same way as the
26205 If the source code for the specs but not the bodies are
26206 available, then you can reimplement the bodies.
26208 Some features introduced by Ada 95 obviate the need for library support. For
26209 example most Ada 83 vendors supplied a package for unsigned integers. The
26210 Ada 95 modular type feature is the preferred way to handle this need, so
26211 instead of migrating or reimplementing the unsigned integer package it may
26212 be preferable to retrofit the application using modular types.
26215 @node Elaboration order
26216 @subsection Elaboration order
26218 The implementation can choose any elaboration order consistent with the unit
26219 dependency relationship. This freedom means that some orders can result in
26220 Program_Error being raised due to an ``Access Before Elaboration'': an attempt
26221 to invoke a subprogram its body has been elaborated, or to instantiate a
26222 generic before the generic body has been elaborated. By default GNAT
26223 attempts to choose a safe order (one that will not encounter access before
26224 elaboration problems) by implicitly inserting @code{Elaborate} or
26225 @code{Elaborate_All} pragmas where
26226 needed. However, this can lead to the creation of elaboration circularities
26227 and a resulting rejection of the program by gnatbind. This issue is
26228 thoroughly described in @ref{Elaboration Order Handling in GNAT}.
26229 In brief, there are several
26230 ways to deal with this situation:
26234 Modify the program to eliminate the circularities, e.g.@: by moving
26235 elaboration-time code into explicitly-invoked procedures
26237 Constrain the elaboration order by including explicit @code{Elaborate_Body} or
26238 @code{Elaborate} pragmas, and then inhibit the generation of implicit
26239 @code{Elaborate_All}
26240 pragmas either globally (as an effect of the @option{-gnatE} switch) or locally
26241 (by selectively suppressing elaboration checks via pragma
26242 @code{Suppress(Elaboration_Check)} when it is safe to do so).
26245 @node Target-specific aspects
26246 @subsection Target-specific aspects
26248 Low-level applications need to deal with machine addresses, data
26249 representations, interfacing with assembler code, and similar issues. If
26250 such an Ada 83 application is being ported to different target hardware (for
26251 example where the byte endianness has changed) then you will need to
26252 carefully examine the program logic; the porting effort will heavily depend
26253 on the robustness of the original design. Moreover, Ada 95 (and thus
26254 Ada 2005) are sometimes
26255 incompatible with typical Ada 83 compiler practices regarding implicit
26256 packing, the meaning of the Size attribute, and the size of access values.
26257 GNAT's approach to these issues is described in @ref{Representation Clauses}.
26259 @node Compatibility with Other Ada Systems
26260 @section Compatibility with Other Ada Systems
26263 If programs avoid the use of implementation dependent and
26264 implementation defined features, as documented in the @cite{Ada
26265 Reference Manual}, there should be a high degree of portability between
26266 GNAT and other Ada systems. The following are specific items which
26267 have proved troublesome in moving Ada 95 programs from GNAT to other Ada 95
26268 compilers, but do not affect porting code to GNAT@.
26269 (As of @value{NOW}, GNAT is the only compiler available for Ada 2005;
26270 the following issues may or may not arise for Ada 2005 programs
26271 when other compilers appear.)
26274 @item Ada 83 Pragmas and Attributes
26275 Ada 95 compilers are allowed, but not required, to implement the missing
26276 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
26277 GNAT implements all such pragmas and attributes, eliminating this as
26278 a compatibility concern, but some other Ada 95 compilers reject these
26279 pragmas and attributes.
26281 @item Specialized Needs Annexes
26282 GNAT implements the full set of special needs annexes. At the
26283 current time, it is the only Ada 95 compiler to do so. This means that
26284 programs making use of these features may not be portable to other Ada
26285 95 compilation systems.
26287 @item Representation Clauses
26288 Some other Ada 95 compilers implement only the minimal set of
26289 representation clauses required by the Ada 95 reference manual. GNAT goes
26290 far beyond this minimal set, as described in the next section.
26293 @node Representation Clauses
26294 @section Representation Clauses
26297 The Ada 83 reference manual was quite vague in describing both the minimal
26298 required implementation of representation clauses, and also their precise
26299 effects. Ada 95 (and thus also Ada 2005) are much more explicit, but the
26300 minimal set of capabilities required is still quite limited.
26302 GNAT implements the full required set of capabilities in
26303 Ada 95 and Ada 2005, but also goes much further, and in particular
26304 an effort has been made to be compatible with existing Ada 83 usage to the
26305 greatest extent possible.
26307 A few cases exist in which Ada 83 compiler behavior is incompatible with
26308 the requirements in Ada 95 (and thus also Ada 2005). These are instances of
26309 intentional or accidental dependence on specific implementation dependent
26310 characteristics of these Ada 83 compilers. The following is a list of
26311 the cases most likely to arise in existing Ada 83 code.
26314 @item Implicit Packing
26315 Some Ada 83 compilers allowed a Size specification to cause implicit
26316 packing of an array or record. This could cause expensive implicit
26317 conversions for change of representation in the presence of derived
26318 types, and the Ada design intends to avoid this possibility.
26319 Subsequent AI's were issued to make it clear that such implicit
26320 change of representation in response to a Size clause is inadvisable,
26321 and this recommendation is represented explicitly in the Ada 95 (and Ada 2005)
26322 Reference Manuals as implementation advice that is followed by GNAT@.
26323 The problem will show up as an error
26324 message rejecting the size clause. The fix is simply to provide
26325 the explicit pragma @code{Pack}, or for more fine tuned control, provide
26326 a Component_Size clause.
26328 @item Meaning of Size Attribute
26329 The Size attribute in Ada 95 (and Ada 2005) for discrete types is defined as
26330 the minimal number of bits required to hold values of the type. For example,
26331 on a 32-bit machine, the size of @code{Natural} will typically be 31 and not
26332 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
26333 some 32 in this situation. This problem will usually show up as a compile
26334 time error, but not always. It is a good idea to check all uses of the
26335 'Size attribute when porting Ada 83 code. The GNAT specific attribute
26336 Object_Size can provide a useful way of duplicating the behavior of
26337 some Ada 83 compiler systems.
26339 @item Size of Access Types
26340 A common assumption in Ada 83 code is that an access type is in fact a pointer,
26341 and that therefore it will be the same size as a System.Address value. This
26342 assumption is true for GNAT in most cases with one exception. For the case of
26343 a pointer to an unconstrained array type (where the bounds may vary from one
26344 value of the access type to another), the default is to use a ``fat pointer'',
26345 which is represented as two separate pointers, one to the bounds, and one to
26346 the array. This representation has a number of advantages, including improved
26347 efficiency. However, it may cause some difficulties in porting existing Ada 83
26348 code which makes the assumption that, for example, pointers fit in 32 bits on
26349 a machine with 32-bit addressing.
26351 To get around this problem, GNAT also permits the use of ``thin pointers'' for
26352 access types in this case (where the designated type is an unconstrained array
26353 type). These thin pointers are indeed the same size as a System.Address value.
26354 To specify a thin pointer, use a size clause for the type, for example:
26356 @smallexample @c ada
26357 type X is access all String;
26358 for X'Size use Standard'Address_Size;
26362 which will cause the type X to be represented using a single pointer.
26363 When using this representation, the bounds are right behind the array.
26364 This representation is slightly less efficient, and does not allow quite
26365 such flexibility in the use of foreign pointers or in using the
26366 Unrestricted_Access attribute to create pointers to non-aliased objects.
26367 But for any standard portable use of the access type it will work in
26368 a functionally correct manner and allow porting of existing code.
26369 Note that another way of forcing a thin pointer representation
26370 is to use a component size clause for the element size in an array,
26371 or a record representation clause for an access field in a record.
26375 @c This brief section is only in the non-VMS version
26376 @c The complete chapter on HP Ada is in the VMS version
26377 @node Compatibility with HP Ada 83
26378 @section Compatibility with HP Ada 83
26381 The VMS version of GNAT fully implements all the pragmas and attributes
26382 provided by HP Ada 83, as well as providing the standard HP Ada 83
26383 libraries, including Starlet. In addition, data layouts and parameter
26384 passing conventions are highly compatible. This means that porting
26385 existing HP Ada 83 code to GNAT in VMS systems should be easier than
26386 most other porting efforts. The following are some of the most
26387 significant differences between GNAT and HP Ada 83.
26390 @item Default floating-point representation
26391 In GNAT, the default floating-point format is IEEE, whereas in HP Ada 83,
26392 it is VMS format. GNAT does implement the necessary pragmas
26393 (Long_Float, Float_Representation) for changing this default.
26396 The package System in GNAT exactly corresponds to the definition in the
26397 Ada 95 reference manual, which means that it excludes many of the
26398 HP Ada 83 extensions. However, a separate package Aux_DEC is provided
26399 that contains the additional definitions, and a special pragma,
26400 Extend_System allows this package to be treated transparently as an
26401 extension of package System.
26404 The definitions provided by Aux_DEC are exactly compatible with those
26405 in the HP Ada 83 version of System, with one exception.
26406 HP Ada provides the following declarations:
26408 @smallexample @c ada
26409 TO_ADDRESS (INTEGER)
26410 TO_ADDRESS (UNSIGNED_LONGWORD)
26411 TO_ADDRESS (@i{universal_integer})
26415 The version of TO_ADDRESS taking a @i{universal integer} argument is in fact
26416 an extension to Ada 83 not strictly compatible with the reference manual.
26417 In GNAT, we are constrained to be exactly compatible with the standard,
26418 and this means we cannot provide this capability. In HP Ada 83, the
26419 point of this definition is to deal with a call like:
26421 @smallexample @c ada
26422 TO_ADDRESS (16#12777#);
26426 Normally, according to the Ada 83 standard, one would expect this to be
26427 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
26428 of TO_ADDRESS@. However, in HP Ada 83, there is no ambiguity, since the
26429 definition using @i{universal_integer} takes precedence.
26431 In GNAT, since the version with @i{universal_integer} cannot be supplied, it
26432 is not possible to be 100% compatible. Since there are many programs using
26433 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
26434 to change the name of the function in the UNSIGNED_LONGWORD case, so the
26435 declarations provided in the GNAT version of AUX_Dec are:
26437 @smallexample @c ada
26438 function To_Address (X : Integer) return Address;
26439 pragma Pure_Function (To_Address);
26441 function To_Address_Long (X : Unsigned_Longword)
26443 pragma Pure_Function (To_Address_Long);
26447 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
26448 change the name to TO_ADDRESS_LONG@.
26450 @item Task_Id values
26451 The Task_Id values assigned will be different in the two systems, and GNAT
26452 does not provide a specified value for the Task_Id of the environment task,
26453 which in GNAT is treated like any other declared task.
26457 For full details on these and other less significant compatibility issues,
26458 see appendix E of the HP publication entitled @cite{HP Ada, Technical
26459 Overview and Comparison on HP Platforms}.
26461 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
26462 attributes are recognized, although only a subset of them can sensibly
26463 be implemented. The description of pragmas in @ref{Implementation
26464 Defined Pragmas,,, gnat_rm, GNAT Reference Manual}
26465 indicates whether or not they are applicable to non-VMS systems.
26469 @node Transitioning to 64-Bit GNAT for OpenVMS
26470 @section Transitioning to 64-Bit @value{EDITION} for OpenVMS
26473 This section is meant to assist users of pre-2006 @value{EDITION}
26474 for Alpha OpenVMS who are transitioning to 64-bit @value{EDITION},
26475 the version of the GNAT technology supplied in 2006 and later for
26476 OpenVMS on both Alpha and I64.
26479 * Introduction to transitioning::
26480 * Migration of 32 bit code::
26481 * Taking advantage of 64 bit addressing::
26482 * Technical details::
26485 @node Introduction to transitioning
26486 @subsection Introduction
26489 64-bit @value{EDITION} for Open VMS has been designed to meet
26494 Providing a full conforming implementation of Ada 95 and Ada 2005
26497 Allowing maximum backward compatibility, thus easing migration of existing
26501 Supplying a path for exploiting the full 64-bit address range
26505 Ada's strong typing semantics has made it
26506 impractical to have different 32-bit and 64-bit modes. As soon as
26507 one object could possibly be outside the 32-bit address space, this
26508 would make it necessary for the @code{System.Address} type to be 64 bits.
26509 In particular, this would cause inconsistencies if 32-bit code is
26510 called from 64-bit code that raises an exception.
26512 This issue has been resolved by always using 64-bit addressing
26513 at the system level, but allowing for automatic conversions between
26514 32-bit and 64-bit addresses where required. Thus users who
26515 do not currently require 64-bit addressing capabilities, can
26516 recompile their code with only minimal changes (and indeed
26517 if the code is written in portable Ada, with no assumptions about
26518 the size of the @code{Address} type, then no changes at all are necessary).
26520 this approach provides a simple, gradual upgrade path to future
26521 use of larger memories than available for 32-bit systems.
26522 Also, newly written applications or libraries will by default
26523 be fully compatible with future systems exploiting 64-bit
26524 addressing capabilities.
26526 @ref{Migration of 32 bit code}, will focus on porting applications
26527 that do not require more than 2 GB of
26528 addressable memory. This code will be referred to as
26529 @emph{32-bit code}.
26530 For applications intending to exploit the full 64-bit address space,
26531 @ref{Taking advantage of 64 bit addressing},
26532 will consider further changes that may be required.
26533 Such code will be referred to below as @emph{64-bit code}.
26535 @node Migration of 32 bit code
26536 @subsection Migration of 32-bit code
26541 * Unchecked conversions::
26542 * Predefined constants::
26543 * Interfacing with C::
26544 * Experience with source compatibility::
26547 @node Address types
26548 @subsubsection Address types
26551 To solve the problem of mixing 64-bit and 32-bit addressing,
26552 while maintaining maximum backward compatibility, the following
26553 approach has been taken:
26557 @code{System.Address} always has a size of 64 bits
26560 @code{System.Short_Address} is a 32-bit subtype of @code{System.Address}
26564 Since @code{System.Short_Address} is a subtype of @code{System.Address},
26565 a @code{Short_Address}
26566 may be used where an @code{Address} is required, and vice versa, without
26567 needing explicit type conversions.
26568 By virtue of the Open VMS parameter passing conventions,
26570 and exported subprograms that have 32-bit address parameters are
26571 compatible with those that have 64-bit address parameters.
26572 (See @ref{Making code 64 bit clean} for details.)
26574 The areas that may need attention are those where record types have
26575 been defined that contain components of the type @code{System.Address}, and
26576 where objects of this type are passed to code expecting a record layout with
26579 Different compilers on different platforms cannot be
26580 expected to represent the same type in the same way,
26581 since alignment constraints
26582 and other system-dependent properties affect the compiler's decision.
26583 For that reason, Ada code
26584 generally uses representation clauses to specify the expected
26585 layout where required.
26587 If such a representation clause uses 32 bits for a component having
26588 the type @code{System.Address}, 64-bit @value{EDITION} for OpenVMS
26589 will detect that error and produce a specific diagnostic message.
26590 The developer should then determine whether the representation
26591 should be 64 bits or not and make either of two changes:
26592 change the size to 64 bits and leave the type as @code{System.Address}, or
26593 leave the size as 32 bits and change the type to @code{System.Short_Address}.
26594 Since @code{Short_Address} is a subtype of @code{Address}, no changes are
26595 required in any code setting or accessing the field; the compiler will
26596 automatically perform any needed conversions between address
26600 @subsubsection Access types
26603 By default, objects designated by access values are always
26604 allocated in the 32-bit
26605 address space. Thus legacy code will never contain
26606 any objects that are not addressable with 32-bit addresses, and
26607 the compiler will never raise exceptions as result of mixing
26608 32-bit and 64-bit addresses.
26610 However, the access values themselves are represented in 64 bits, for optimum
26611 performance and future compatibility with 64-bit code. As was
26612 the case with @code{System.Address}, the compiler will give an error message
26613 if an object or record component has a representation clause that
26614 requires the access value to fit in 32 bits. In such a situation,
26615 an explicit size clause for the access type, specifying 32 bits,
26616 will have the desired effect.
26618 General access types (declared with @code{access all}) can never be
26619 32 bits, as values of such types must be able to refer to any object
26620 of the designated type,
26621 including objects residing outside the 32-bit address range.
26622 Existing Ada 83 code will not contain such type definitions,
26623 however, since general access types were introduced in Ada 95.
26625 @node Unchecked conversions
26626 @subsubsection Unchecked conversions
26629 In the case of an @code{Unchecked_Conversion} where the source type is a
26630 64-bit access type or the type @code{System.Address}, and the target
26631 type is a 32-bit type, the compiler will generate a warning.
26632 Even though the generated code will still perform the required
26633 conversions, it is highly recommended in these cases to use
26634 respectively a 32-bit access type or @code{System.Short_Address}
26635 as the source type.
26637 @node Predefined constants
26638 @subsubsection Predefined constants
26641 The following table shows the correspondence between pre-2006 versions of
26642 @value{EDITION} on Alpha OpenVMS (``Old'') and 64-bit @value{EDITION}
26645 @multitable {@code{System.Short_Memory_Size}} {2**32} {2**64}
26646 @item @b{Constant} @tab @b{Old} @tab @b{New}
26647 @item @code{System.Word_Size} @tab 32 @tab 64
26648 @item @code{System.Memory_Size} @tab 2**32 @tab 2**64
26649 @item @code{System.Short_Memory_Size} @tab 2**32 @tab 2**32
26650 @item @code{System.Address_Size} @tab 32 @tab 64
26654 If you need to refer to the specific
26655 memory size of a 32-bit implementation, instead of the
26656 actual memory size, use @code{System.Short_Memory_Size}
26657 rather than @code{System.Memory_Size}.
26658 Similarly, references to @code{System.Address_Size} may need
26659 to be replaced by @code{System.Short_Address'Size}.
26660 The program @command{gnatfind} may be useful for locating
26661 references to the above constants, so that you can verify that they
26664 @node Interfacing with C
26665 @subsubsection Interfacing with C
26668 In order to minimize the impact of the transition to 64-bit addresses on
26669 legacy programs, some fundamental types in the @code{Interfaces.C}
26670 package hierarchy continue to be represented in 32 bits.
26671 These types are: @code{ptrdiff_t}, @code{size_t}, and @code{chars_ptr}.
26672 This eases integration with the default HP C layout choices, for example
26673 as found in the system routines in @code{DECC$SHR.EXE}.
26674 Because of this implementation choice, the type fully compatible with
26675 @code{chars_ptr} is now @code{Short_Address} and not @code{Address}.
26676 Depending on the context the compiler will issue a
26677 warning or an error when type @code{Address} is used, alerting the user to a
26678 potential problem. Otherwise 32-bit programs that use
26679 @code{Interfaces.C} should normally not require code modifications
26681 The other issue arising with C interfacing concerns pragma @code{Convention}.
26682 For VMS 64-bit systems, there is an issue of the appropriate default size
26683 of C convention pointers in the absence of an explicit size clause. The HP
26684 C compiler can choose either 32 or 64 bits depending on compiler options.
26685 GNAT chooses 32-bits rather than 64-bits in the default case where no size
26686 clause is given. This proves a better choice for porting 32-bit legacy
26687 applications. In order to have a 64-bit representation, it is necessary to
26688 specify a size representation clause. For example:
26690 @smallexample @c ada
26691 type int_star is access Interfaces.C.int;
26692 pragma Convention(C, int_star);
26693 for int_star'Size use 64; -- Necessary to get 64 and not 32 bits
26696 @node Experience with source compatibility
26697 @subsubsection Experience with source compatibility
26700 The Security Server and STARLET on I64 provide an interesting ``test case''
26701 for source compatibility issues, since it is in such system code
26702 where assumptions about @code{Address} size might be expected to occur.
26703 Indeed, there were a small number of occasions in the Security Server
26704 file @file{jibdef.ads}
26705 where a representation clause for a record type specified
26706 32 bits for a component of type @code{Address}.
26707 All of these errors were detected by the compiler.
26708 The repair was obvious and immediate; to simply replace @code{Address} by
26709 @code{Short_Address}.
26711 In the case of STARLET, there were several record types that should
26712 have had representation clauses but did not. In these record types
26713 there was an implicit assumption that an @code{Address} value occupied
26715 These compiled without error, but their usage resulted in run-time error
26716 returns from STARLET system calls.
26717 Future GNAT technology enhancements may include a tool that detects and flags
26718 these sorts of potential source code porting problems.
26720 @c ****************************************
26721 @node Taking advantage of 64 bit addressing
26722 @subsection Taking advantage of 64-bit addressing
26725 * Making code 64 bit clean::
26726 * Allocating memory from the 64 bit storage pool::
26727 * Restrictions on use of 64 bit objects::
26728 * Using 64 bit storage pools by default::
26729 * General access types::
26730 * STARLET and other predefined libraries::
26733 @node Making code 64 bit clean
26734 @subsubsection Making code 64-bit clean
26737 In order to prevent problems that may occur when (parts of) a
26738 system start using memory outside the 32-bit address range,
26739 we recommend some additional guidelines:
26743 For imported subprograms that take parameters of the
26744 type @code{System.Address}, ensure that these subprograms can
26745 indeed handle 64-bit addresses. If not, or when in doubt,
26746 change the subprogram declaration to specify
26747 @code{System.Short_Address} instead.
26750 Resolve all warnings related to size mismatches in
26751 unchecked conversions. Failing to do so causes
26752 erroneous execution if the source object is outside
26753 the 32-bit address space.
26756 (optional) Explicitly use the 32-bit storage pool
26757 for access types used in a 32-bit context, or use
26758 generic access types where possible
26759 (@pxref{Restrictions on use of 64 bit objects}).
26763 If these rules are followed, the compiler will automatically insert
26764 any necessary checks to ensure that no addresses or access values
26765 passed to 32-bit code ever refer to objects outside the 32-bit
26767 Any attempt to do this will raise @code{Constraint_Error}.
26769 @node Allocating memory from the 64 bit storage pool
26770 @subsubsection Allocating memory from the 64-bit storage pool
26773 For any access type @code{T} that potentially requires memory allocations
26774 beyond the 32-bit address space,
26775 use the following representation clause:
26777 @smallexample @c ada
26778 for T'Storage_Pool use System.Pool_64;
26781 @node Restrictions on use of 64 bit objects
26782 @subsubsection Restrictions on use of 64-bit objects
26785 Taking the address of an object allocated from a 64-bit storage pool,
26786 and then passing this address to a subprogram expecting
26787 @code{System.Short_Address},
26788 or assigning it to a variable of type @code{Short_Address}, will cause
26789 @code{Constraint_Error} to be raised. In case the code is not 64-bit clean
26790 (@pxref{Making code 64 bit clean}), or checks are suppressed,
26791 no exception is raised and execution
26792 will become erroneous.
26794 @node Using 64 bit storage pools by default
26795 @subsubsection Using 64-bit storage pools by default
26798 In some cases it may be desirable to have the compiler allocate
26799 from 64-bit storage pools by default. This may be the case for
26800 libraries that are 64-bit clean, but may be used in both 32-bit
26801 and 64-bit contexts. For these cases the following configuration
26802 pragma may be specified:
26804 @smallexample @c ada
26805 pragma Pool_64_Default;
26809 Any code compiled in the context of this pragma will by default
26810 use the @code{System.Pool_64} storage pool. This default may be overridden
26811 for a specific access type @code{T} by the representation clause:
26813 @smallexample @c ada
26814 for T'Storage_Pool use System.Pool_32;
26818 Any object whose address may be passed to a subprogram with a
26819 @code{Short_Address} argument, or assigned to a variable of type
26820 @code{Short_Address}, needs to be allocated from this pool.
26822 @node General access types
26823 @subsubsection General access types
26826 Objects designated by access values from a
26827 general access type (declared with @code{access all}) are never allocated
26828 from a 64-bit storage pool. Code that uses general access types will
26829 accept objects allocated in either 32-bit or 64-bit address spaces,
26830 but never allocate objects outside the 32-bit address space.
26831 Using general access types ensures maximum compatibility with both
26832 32-bit and 64-bit code.
26834 @node STARLET and other predefined libraries
26835 @subsubsection STARLET and other predefined libraries
26838 All code that comes as part of GNAT is 64-bit clean, but the
26839 restrictions given in @ref{Restrictions on use of 64 bit objects},
26840 still apply. Look at the package
26841 specs to see in which contexts objects allocated
26842 in 64-bit address space are acceptable.
26844 @node Technical details
26845 @subsection Technical details
26848 64-bit @value{EDITION} for Open VMS takes advantage of the freedom given in the
26849 Ada standard with respect to the type of @code{System.Address}. Previous
26850 versions of GNAT Pro have defined this type as private and implemented it as a
26853 In order to allow defining @code{System.Short_Address} as a proper subtype,
26854 and to match the implicit sign extension in parameter passing,
26855 in 64-bit @value{EDITION} for Open VMS, @code{System.Address} is defined as a
26856 visible (i.e., non-private) integer type.
26857 Standard operations on the type, such as the binary operators ``+'', ``-'',
26858 etc., that take @code{Address} operands and return an @code{Address} result,
26859 have been hidden by declaring these
26860 @code{abstract}, a feature introduced in Ada 95 that helps avoid the potential
26861 ambiguities that would otherwise result from overloading.
26862 (Note that, although @code{Address} is a visible integer type,
26863 good programming practice dictates against exploiting the type's
26864 integer properties such as literals, since this will compromise
26867 Defining @code{Address} as a visible integer type helps achieve
26868 maximum compatibility for existing Ada code,
26869 without sacrificing the capabilities of the 64-bit architecture.
26872 @c ************************************************
26874 @node Microsoft Windows Topics
26875 @appendix Microsoft Windows Topics
26881 This chapter describes topics that are specific to the Microsoft Windows
26882 platforms (NT, 2000, and XP Professional).
26885 * Using GNAT on Windows::
26886 * Using a network installation of GNAT::
26887 * CONSOLE and WINDOWS subsystems::
26888 * Temporary Files::
26889 * Mixed-Language Programming on Windows::
26890 * Windows Calling Conventions::
26891 * Introduction to Dynamic Link Libraries (DLLs)::
26892 * Using DLLs with GNAT::
26893 * Building DLLs with GNAT Project files::
26894 * Building DLLs with GNAT::
26895 * Building DLLs with gnatdll::
26896 * GNAT and Windows Resources::
26897 * Debugging a DLL::
26898 * Setting Stack Size from gnatlink::
26899 * Setting Heap Size from gnatlink::
26902 @node Using GNAT on Windows
26903 @section Using GNAT on Windows
26906 One of the strengths of the GNAT technology is that its tool set
26907 (@command{gcc}, @command{gnatbind}, @command{gnatlink}, @command{gnatmake}, the
26908 @code{gdb} debugger, etc.) is used in the same way regardless of the
26911 On Windows this tool set is complemented by a number of Microsoft-specific
26912 tools that have been provided to facilitate interoperability with Windows
26913 when this is required. With these tools:
26918 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
26922 You can use any Dynamically Linked Library (DLL) in your Ada code (both
26923 relocatable and non-relocatable DLLs are supported).
26926 You can build Ada DLLs for use in other applications. These applications
26927 can be written in a language other than Ada (e.g., C, C++, etc). Again both
26928 relocatable and non-relocatable Ada DLLs are supported.
26931 You can include Windows resources in your Ada application.
26934 You can use or create COM/DCOM objects.
26938 Immediately below are listed all known general GNAT-for-Windows restrictions.
26939 Other restrictions about specific features like Windows Resources and DLLs
26940 are listed in separate sections below.
26945 It is not possible to use @code{GetLastError} and @code{SetLastError}
26946 when tasking, protected records, or exceptions are used. In these
26947 cases, in order to implement Ada semantics, the GNAT run-time system
26948 calls certain Win32 routines that set the last error variable to 0 upon
26949 success. It should be possible to use @code{GetLastError} and
26950 @code{SetLastError} when tasking, protected record, and exception
26951 features are not used, but it is not guaranteed to work.
26954 It is not possible to link against Microsoft libraries except for
26955 import libraries. Interfacing must be done by the mean of DLLs.
26958 When the compilation environment is located on FAT32 drives, users may
26959 experience recompilations of the source files that have not changed if
26960 Daylight Saving Time (DST) state has changed since the last time files
26961 were compiled. NTFS drives do not have this problem.
26964 No components of the GNAT toolset use any entries in the Windows
26965 registry. The only entries that can be created are file associations and
26966 PATH settings, provided the user has chosen to create them at installation
26967 time, as well as some minimal book-keeping information needed to correctly
26968 uninstall or integrate different GNAT products.
26971 @node Using a network installation of GNAT
26972 @section Using a network installation of GNAT
26975 Make sure the system on which GNAT is installed is accessible from the
26976 current machine, i.e., the install location is shared over the network.
26977 Shared resources are accessed on Windows by means of UNC paths, which
26978 have the format @code{\\server\sharename\path}
26980 In order to use such a network installation, simply add the UNC path of the
26981 @file{bin} directory of your GNAT installation in front of your PATH. For
26982 example, if GNAT is installed in @file{\GNAT} directory of a share location
26983 called @file{c-drive} on a machine @file{LOKI}, the following command will
26986 @code{@ @ @ path \\loki\c-drive\gnat\bin;%path%}
26988 Be aware that every compilation using the network installation results in the
26989 transfer of large amounts of data across the network and will likely cause
26990 serious performance penalty.
26992 @node CONSOLE and WINDOWS subsystems
26993 @section CONSOLE and WINDOWS subsystems
26994 @cindex CONSOLE Subsystem
26995 @cindex WINDOWS Subsystem
26999 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
27000 (which is the default subsystem) will always create a console when
27001 launching the application. This is not something desirable when the
27002 application has a Windows GUI. To get rid of this console the
27003 application must be using the @code{WINDOWS} subsystem. To do so
27004 the @option{-mwindows} linker option must be specified.
27007 $ gnatmake winprog -largs -mwindows
27010 @node Temporary Files
27011 @section Temporary Files
27012 @cindex Temporary files
27015 It is possible to control where temporary files gets created by setting
27016 the @env{TMP} environment variable. The file will be created:
27019 @item Under the directory pointed to by the @env{TMP} environment variable if
27020 this directory exists.
27022 @item Under @file{c:\temp}, if the @env{TMP} environment variable is not
27023 set (or not pointing to a directory) and if this directory exists.
27025 @item Under the current working directory otherwise.
27029 This allows you to determine exactly where the temporary
27030 file will be created. This is particularly useful in networked
27031 environments where you may not have write access to some
27034 @node Mixed-Language Programming on Windows
27035 @section Mixed-Language Programming on Windows
27038 Developing pure Ada applications on Windows is no different than on
27039 other GNAT-supported platforms. However, when developing or porting an
27040 application that contains a mix of Ada and C/C++, the choice of your
27041 Windows C/C++ development environment conditions your overall
27042 interoperability strategy.
27044 If you use @command{gcc} to compile the non-Ada part of your application,
27045 there are no Windows-specific restrictions that affect the overall
27046 interoperability with your Ada code. If you do want to use the
27047 Microsoft tools for your non-Ada code, you have two choices:
27051 Encapsulate your non-Ada code in a DLL to be linked with your Ada
27052 application. In this case, use the Microsoft or whatever environment to
27053 build the DLL and use GNAT to build your executable
27054 (@pxref{Using DLLs with GNAT}).
27057 Or you can encapsulate your Ada code in a DLL to be linked with the
27058 other part of your application. In this case, use GNAT to build the DLL
27059 (@pxref{Building DLLs with GNAT Project files}) and use the Microsoft
27060 or whatever environment to build your executable.
27063 @node Windows Calling Conventions
27064 @section Windows Calling Conventions
27068 This section pertain only to Win32. On Win64 there is a single native
27069 calling convention. All convention specifiers are ignored on this
27073 * C Calling Convention::
27074 * Stdcall Calling Convention::
27075 * Win32 Calling Convention::
27076 * DLL Calling Convention::
27080 When a subprogram @code{F} (caller) calls a subprogram @code{G}
27081 (callee), there are several ways to push @code{G}'s parameters on the
27082 stack and there are several possible scenarios to clean up the stack
27083 upon @code{G}'s return. A calling convention is an agreed upon software
27084 protocol whereby the responsibilities between the caller (@code{F}) and
27085 the callee (@code{G}) are clearly defined. Several calling conventions
27086 are available for Windows:
27090 @code{C} (Microsoft defined)
27093 @code{Stdcall} (Microsoft defined)
27096 @code{Win32} (GNAT specific)
27099 @code{DLL} (GNAT specific)
27102 @node C Calling Convention
27103 @subsection @code{C} Calling Convention
27106 This is the default calling convention used when interfacing to C/C++
27107 routines compiled with either @command{gcc} or Microsoft Visual C++.
27109 In the @code{C} calling convention subprogram parameters are pushed on the
27110 stack by the caller from right to left. The caller itself is in charge of
27111 cleaning up the stack after the call. In addition, the name of a routine
27112 with @code{C} calling convention is mangled by adding a leading underscore.
27114 The name to use on the Ada side when importing (or exporting) a routine
27115 with @code{C} calling convention is the name of the routine. For
27116 instance the C function:
27119 int get_val (long);
27123 should be imported from Ada as follows:
27125 @smallexample @c ada
27127 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
27128 pragma Import (C, Get_Val, External_Name => "get_val");
27133 Note that in this particular case the @code{External_Name} parameter could
27134 have been omitted since, when missing, this parameter is taken to be the
27135 name of the Ada entity in lower case. When the @code{Link_Name} parameter
27136 is missing, as in the above example, this parameter is set to be the
27137 @code{External_Name} with a leading underscore.
27139 When importing a variable defined in C, you should always use the @code{C}
27140 calling convention unless the object containing the variable is part of a
27141 DLL (in which case you should use the @code{Stdcall} calling
27142 convention, @pxref{Stdcall Calling Convention}).
27144 @node Stdcall Calling Convention
27145 @subsection @code{Stdcall} Calling Convention
27148 This convention, which was the calling convention used for Pascal
27149 programs, is used by Microsoft for all the routines in the Win32 API for
27150 efficiency reasons. It must be used to import any routine for which this
27151 convention was specified.
27153 In the @code{Stdcall} calling convention subprogram parameters are pushed
27154 on the stack by the caller from right to left. The callee (and not the
27155 caller) is in charge of cleaning the stack on routine exit. In addition,
27156 the name of a routine with @code{Stdcall} calling convention is mangled by
27157 adding a leading underscore (as for the @code{C} calling convention) and a
27158 trailing @code{@@}@code{@var{nn}}, where @var{nn} is the overall size (in
27159 bytes) of the parameters passed to the routine.
27161 The name to use on the Ada side when importing a C routine with a
27162 @code{Stdcall} calling convention is the name of the C routine. The leading
27163 underscore and trailing @code{@@}@code{@var{nn}} are added automatically by
27164 the compiler. For instance the Win32 function:
27167 @b{APIENTRY} int get_val (long);
27171 should be imported from Ada as follows:
27173 @smallexample @c ada
27175 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
27176 pragma Import (Stdcall, Get_Val);
27177 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
27182 As for the @code{C} calling convention, when the @code{External_Name}
27183 parameter is missing, it is taken to be the name of the Ada entity in lower
27184 case. If instead of writing the above import pragma you write:
27186 @smallexample @c ada
27188 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
27189 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
27194 then the imported routine is @code{_retrieve_val@@4}. However, if instead
27195 of specifying the @code{External_Name} parameter you specify the
27196 @code{Link_Name} as in the following example:
27198 @smallexample @c ada
27200 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
27201 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
27206 then the imported routine is @code{retrieve_val}, that is, there is no
27207 decoration at all. No leading underscore and no Stdcall suffix
27208 @code{@@}@code{@var{nn}}.
27211 This is especially important as in some special cases a DLL's entry
27212 point name lacks a trailing @code{@@}@code{@var{nn}} while the exported
27213 name generated for a call has it.
27216 It is also possible to import variables defined in a DLL by using an
27217 import pragma for a variable. As an example, if a DLL contains a
27218 variable defined as:
27225 then, to access this variable from Ada you should write:
27227 @smallexample @c ada
27229 My_Var : Interfaces.C.int;
27230 pragma Import (Stdcall, My_Var);
27235 Note that to ease building cross-platform bindings this convention
27236 will be handled as a @code{C} calling convention on non-Windows platforms.
27238 @node Win32 Calling Convention
27239 @subsection @code{Win32} Calling Convention
27242 This convention, which is GNAT-specific is fully equivalent to the
27243 @code{Stdcall} calling convention described above.
27245 @node DLL Calling Convention
27246 @subsection @code{DLL} Calling Convention
27249 This convention, which is GNAT-specific is fully equivalent to the
27250 @code{Stdcall} calling convention described above.
27252 @node Introduction to Dynamic Link Libraries (DLLs)
27253 @section Introduction to Dynamic Link Libraries (DLLs)
27257 A Dynamically Linked Library (DLL) is a library that can be shared by
27258 several applications running under Windows. A DLL can contain any number of
27259 routines and variables.
27261 One advantage of DLLs is that you can change and enhance them without
27262 forcing all the applications that depend on them to be relinked or
27263 recompiled. However, you should be aware than all calls to DLL routines are
27264 slower since, as you will understand below, such calls are indirect.
27266 To illustrate the remainder of this section, suppose that an application
27267 wants to use the services of a DLL @file{API.dll}. To use the services
27268 provided by @file{API.dll} you must statically link against the DLL or
27269 an import library which contains a jump table with an entry for each
27270 routine and variable exported by the DLL. In the Microsoft world this
27271 import library is called @file{API.lib}. When using GNAT this import
27272 library is called either @file{libAPI.dll.a}, @file{libapi.dll.a},
27273 @file{libAPI.a} or @file{libapi.a} (names are case insensitive).
27275 After you have linked your application with the DLL or the import library
27276 and you run your application, here is what happens:
27280 Your application is loaded into memory.
27283 The DLL @file{API.dll} is mapped into the address space of your
27284 application. This means that:
27288 The DLL will use the stack of the calling thread.
27291 The DLL will use the virtual address space of the calling process.
27294 The DLL will allocate memory from the virtual address space of the calling
27298 Handles (pointers) can be safely exchanged between routines in the DLL
27299 routines and routines in the application using the DLL.
27303 The entries in the jump table (from the import library @file{libAPI.dll.a}
27304 or @file{API.lib} or automatically created when linking against a DLL)
27305 which is part of your application are initialized with the addresses
27306 of the routines and variables in @file{API.dll}.
27309 If present in @file{API.dll}, routines @code{DllMain} or
27310 @code{DllMainCRTStartup} are invoked. These routines typically contain
27311 the initialization code needed for the well-being of the routines and
27312 variables exported by the DLL.
27316 There is an additional point which is worth mentioning. In the Windows
27317 world there are two kind of DLLs: relocatable and non-relocatable
27318 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
27319 in the target application address space. If the addresses of two
27320 non-relocatable DLLs overlap and these happen to be used by the same
27321 application, a conflict will occur and the application will run
27322 incorrectly. Hence, when possible, it is always preferable to use and
27323 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
27324 supported by GNAT. Note that the @option{-s} linker option (see GNU Linker
27325 User's Guide) removes the debugging symbols from the DLL but the DLL can
27326 still be relocated.
27328 As a side note, an interesting difference between Microsoft DLLs and
27329 Unix shared libraries, is the fact that on most Unix systems all public
27330 routines are exported by default in a Unix shared library, while under
27331 Windows it is possible (but not required) to list exported routines in
27332 a definition file (@pxref{The Definition File}).
27334 @node Using DLLs with GNAT
27335 @section Using DLLs with GNAT
27338 * Creating an Ada Spec for the DLL Services::
27339 * Creating an Import Library::
27343 To use the services of a DLL, say @file{API.dll}, in your Ada application
27348 The Ada spec for the routines and/or variables you want to access in
27349 @file{API.dll}. If not available this Ada spec must be built from the C/C++
27350 header files provided with the DLL.
27353 The import library (@file{libAPI.dll.a} or @file{API.lib}). As previously
27354 mentioned an import library is a statically linked library containing the
27355 import table which will be filled at load time to point to the actual
27356 @file{API.dll} routines. Sometimes you don't have an import library for the
27357 DLL you want to use. The following sections will explain how to build
27358 one. Note that this is optional.
27361 The actual DLL, @file{API.dll}.
27365 Once you have all the above, to compile an Ada application that uses the
27366 services of @file{API.dll} and whose main subprogram is @code{My_Ada_App},
27367 you simply issue the command
27370 $ gnatmake my_ada_app -largs -lAPI
27374 The argument @option{-largs -lAPI} at the end of the @command{gnatmake} command
27375 tells the GNAT linker to look for an import library. The linker will
27376 look for a library name in this specific order:
27379 @item @file{libAPI.dll.a}
27380 @item @file{API.dll.a}
27381 @item @file{libAPI.a}
27382 @item @file{API.lib}
27383 @item @file{libAPI.dll}
27384 @item @file{API.dll}
27387 The first three are the GNU style import libraries. The third is the
27388 Microsoft style import libraries. The last two are the DLL themself.
27390 Note that if the Ada package spec for @file{API.dll} contains the
27393 @smallexample @c ada
27394 pragma Linker_Options ("-lAPI");
27398 you do not have to add @option{-largs -lAPI} at the end of the
27399 @command{gnatmake} command.
27401 If any one of the items above is missing you will have to create it
27402 yourself. The following sections explain how to do so using as an
27403 example a fictitious DLL called @file{API.dll}.
27405 @node Creating an Ada Spec for the DLL Services
27406 @subsection Creating an Ada Spec for the DLL Services
27409 A DLL typically comes with a C/C++ header file which provides the
27410 definitions of the routines and variables exported by the DLL. The Ada
27411 equivalent of this header file is a package spec that contains definitions
27412 for the imported entities. If the DLL you intend to use does not come with
27413 an Ada spec you have to generate one such spec yourself. For example if
27414 the header file of @file{API.dll} is a file @file{api.h} containing the
27415 following two definitions:
27427 then the equivalent Ada spec could be:
27429 @smallexample @c ada
27432 with Interfaces.C.Strings;
27437 function Get (Str : C.Strings.Chars_Ptr) return C.int;
27440 pragma Import (C, Get);
27441 pragma Import (DLL, Some_Var);
27448 Note that a variable is
27449 @strong{always imported with a DLL convention}. A function
27450 can have @code{C} or @code{Stdcall} convention.
27451 (@pxref{Windows Calling Conventions}).
27453 @node Creating an Import Library
27454 @subsection Creating an Import Library
27455 @cindex Import library
27458 * The Definition File::
27459 * GNAT-Style Import Library::
27460 * Microsoft-Style Import Library::
27464 If a Microsoft-style import library @file{API.lib} or a GNAT-style
27465 import library @file{libAPI.dll.a} or @file{libAPI.a} is available
27466 with @file{API.dll} you can skip this section. You can also skip this
27467 section if @file{API.dll} or @file{libAPI.dll} is built with GNU tools
27468 as in this case it is possible to link directly against the
27469 DLL. Otherwise read on.
27471 @node The Definition File
27472 @subsubsection The Definition File
27473 @cindex Definition file
27477 As previously mentioned, and unlike Unix systems, the list of symbols
27478 that are exported from a DLL must be provided explicitly in Windows.
27479 The main goal of a definition file is precisely that: list the symbols
27480 exported by a DLL. A definition file (usually a file with a @code{.def}
27481 suffix) has the following structure:
27486 @r{[}LIBRARY @var{name}@r{]}
27487 @r{[}DESCRIPTION @var{string}@r{]}
27497 @item LIBRARY @var{name}
27498 This section, which is optional, gives the name of the DLL.
27500 @item DESCRIPTION @var{string}
27501 This section, which is optional, gives a description string that will be
27502 embedded in the import library.
27505 This section gives the list of exported symbols (procedures, functions or
27506 variables). For instance in the case of @file{API.dll} the @code{EXPORTS}
27507 section of @file{API.def} looks like:
27521 Note that you must specify the correct suffix (@code{@@}@code{@var{nn}})
27522 (@pxref{Windows Calling Conventions}) for a Stdcall
27523 calling convention function in the exported symbols list.
27526 There can actually be other sections in a definition file, but these
27527 sections are not relevant to the discussion at hand.
27529 @node GNAT-Style Import Library
27530 @subsubsection GNAT-Style Import Library
27533 To create a static import library from @file{API.dll} with the GNAT tools
27534 you should proceed as follows:
27538 Create the definition file @file{API.def} (@pxref{The Definition File}).
27539 For that use the @code{dll2def} tool as follows:
27542 $ dll2def API.dll > API.def
27546 @code{dll2def} is a very simple tool: it takes as input a DLL and prints
27547 to standard output the list of entry points in the DLL. Note that if
27548 some routines in the DLL have the @code{Stdcall} convention
27549 (@pxref{Windows Calling Conventions}) with stripped @code{@@}@var{nn}
27550 suffix then you'll have to edit @file{api.def} to add it, and specify
27551 @option{-k} to @command{gnatdll} when creating the import library.
27554 Here are some hints to find the right @code{@@}@var{nn} suffix.
27558 If you have the Microsoft import library (.lib), it is possible to get
27559 the right symbols by using Microsoft @code{dumpbin} tool (see the
27560 corresponding Microsoft documentation for further details).
27563 $ dumpbin /exports api.lib
27567 If you have a message about a missing symbol at link time the compiler
27568 tells you what symbol is expected. You just have to go back to the
27569 definition file and add the right suffix.
27573 Build the import library @code{libAPI.dll.a}, using @code{gnatdll}
27574 (@pxref{Using gnatdll}) as follows:
27577 $ gnatdll -e API.def -d API.dll
27581 @code{gnatdll} takes as input a definition file @file{API.def} and the
27582 name of the DLL containing the services listed in the definition file
27583 @file{API.dll}. The name of the static import library generated is
27584 computed from the name of the definition file as follows: if the
27585 definition file name is @var{xyz}@code{.def}, the import library name will
27586 be @code{lib}@var{xyz}@code{.a}. Note that in the previous example option
27587 @option{-e} could have been removed because the name of the definition
27588 file (before the ``@code{.def}'' suffix) is the same as the name of the
27589 DLL (@pxref{Using gnatdll} for more information about @code{gnatdll}).
27592 @node Microsoft-Style Import Library
27593 @subsubsection Microsoft-Style Import Library
27596 With GNAT you can either use a GNAT-style or Microsoft-style import
27597 library. A Microsoft import library is needed only if you plan to make an
27598 Ada DLL available to applications developed with Microsoft
27599 tools (@pxref{Mixed-Language Programming on Windows}).
27601 To create a Microsoft-style import library for @file{API.dll} you
27602 should proceed as follows:
27606 Create the definition file @file{API.def} from the DLL. For this use either
27607 the @code{dll2def} tool as described above or the Microsoft @code{dumpbin}
27608 tool (see the corresponding Microsoft documentation for further details).
27611 Build the actual import library using Microsoft's @code{lib} utility:
27614 $ lib -machine:IX86 -def:API.def -out:API.lib
27618 If you use the above command the definition file @file{API.def} must
27619 contain a line giving the name of the DLL:
27626 See the Microsoft documentation for further details about the usage of
27630 @node Building DLLs with GNAT Project files
27631 @section Building DLLs with GNAT Project files
27632 @cindex DLLs, building
27635 There is nothing specific to Windows in the build process.
27636 @pxref{Library Projects}.
27639 Due to a system limitation, it is not possible under Windows to create threads
27640 when inside the @code{DllMain} routine which is used for auto-initialization
27641 of shared libraries, so it is not possible to have library level tasks in SALs.
27643 @node Building DLLs with GNAT
27644 @section Building DLLs with GNAT
27645 @cindex DLLs, building
27648 This section explain how to build DLLs using the GNAT built-in DLL
27649 support. With the following procedure it is straight forward to build
27650 and use DLLs with GNAT.
27654 @item building object files
27656 The first step is to build all objects files that are to be included
27657 into the DLL. This is done by using the standard @command{gnatmake} tool.
27659 @item building the DLL
27661 To build the DLL you must use @command{gcc}'s @option{-shared}
27662 option. It is quite simple to use this method:
27665 $ gcc -shared -o api.dll obj1.o obj2.o @dots{}
27668 It is important to note that in this case all symbols found in the
27669 object files are automatically exported. It is possible to restrict
27670 the set of symbols to export by passing to @command{gcc} a definition
27671 file, @pxref{The Definition File}. For example:
27674 $ gcc -shared -o api.dll api.def obj1.o obj2.o @dots{}
27677 If you use a definition file you must export the elaboration procedures
27678 for every package that required one. Elaboration procedures are named
27679 using the package name followed by "_E".
27681 @item preparing DLL to be used
27683 For the DLL to be used by client programs the bodies must be hidden
27684 from it and the .ali set with read-only attribute. This is very important
27685 otherwise GNAT will recompile all packages and will not actually use
27686 the code in the DLL. For example:
27690 $ copy *.ads *.ali api.dll apilib
27691 $ attrib +R apilib\*.ali
27696 At this point it is possible to use the DLL by directly linking
27697 against it. Note that you must use the GNAT shared runtime when using
27698 GNAT shared libraries. This is achieved by using @option{-shared} binder's
27702 $ gnatmake main -Iapilib -bargs -shared -largs -Lapilib -lAPI
27705 @node Building DLLs with gnatdll
27706 @section Building DLLs with gnatdll
27707 @cindex DLLs, building
27710 * Limitations When Using Ada DLLs from Ada::
27711 * Exporting Ada Entities::
27712 * Ada DLLs and Elaboration::
27713 * Ada DLLs and Finalization::
27714 * Creating a Spec for Ada DLLs::
27715 * Creating the Definition File::
27720 Note that it is preferred to use GNAT Project files
27721 (@pxref{Building DLLs with GNAT Project files}) or the built-in GNAT
27722 DLL support (@pxref{Building DLLs with GNAT}) or to build DLLs.
27724 This section explains how to build DLLs containing Ada code using
27725 @code{gnatdll}. These DLLs will be referred to as Ada DLLs in the
27726 remainder of this section.
27728 The steps required to build an Ada DLL that is to be used by Ada as well as
27729 non-Ada applications are as follows:
27733 You need to mark each Ada @i{entity} exported by the DLL with a @code{C} or
27734 @code{Stdcall} calling convention to avoid any Ada name mangling for the
27735 entities exported by the DLL (@pxref{Exporting Ada Entities}). You can
27736 skip this step if you plan to use the Ada DLL only from Ada applications.
27739 Your Ada code must export an initialization routine which calls the routine
27740 @code{adainit} generated by @command{gnatbind} to perform the elaboration of
27741 the Ada code in the DLL (@pxref{Ada DLLs and Elaboration}). The initialization
27742 routine exported by the Ada DLL must be invoked by the clients of the DLL
27743 to initialize the DLL.
27746 When useful, the DLL should also export a finalization routine which calls
27747 routine @code{adafinal} generated by @command{gnatbind} to perform the
27748 finalization of the Ada code in the DLL (@pxref{Ada DLLs and Finalization}).
27749 The finalization routine exported by the Ada DLL must be invoked by the
27750 clients of the DLL when the DLL services are no further needed.
27753 You must provide a spec for the services exported by the Ada DLL in each
27754 of the programming languages to which you plan to make the DLL available.
27757 You must provide a definition file listing the exported entities
27758 (@pxref{The Definition File}).
27761 Finally you must use @code{gnatdll} to produce the DLL and the import
27762 library (@pxref{Using gnatdll}).
27766 Note that a relocatable DLL stripped using the @code{strip}
27767 binutils tool will not be relocatable anymore. To build a DLL without
27768 debug information pass @code{-largs -s} to @code{gnatdll}. This
27769 restriction does not apply to a DLL built using a Library Project.
27770 @pxref{Library Projects}.
27772 @node Limitations When Using Ada DLLs from Ada
27773 @subsection Limitations When Using Ada DLLs from Ada
27776 When using Ada DLLs from Ada applications there is a limitation users
27777 should be aware of. Because on Windows the GNAT run time is not in a DLL of
27778 its own, each Ada DLL includes a part of the GNAT run time. Specifically,
27779 each Ada DLL includes the services of the GNAT run time that are necessary
27780 to the Ada code inside the DLL. As a result, when an Ada program uses an
27781 Ada DLL there are two independent GNAT run times: one in the Ada DLL and
27782 one in the main program.
27784 It is therefore not possible to exchange GNAT run-time objects between the
27785 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
27786 handles (e.g.@: @code{Text_IO.File_Type}), tasks types, protected objects
27789 It is completely safe to exchange plain elementary, array or record types,
27790 Windows object handles, etc.
27792 @node Exporting Ada Entities
27793 @subsection Exporting Ada Entities
27794 @cindex Export table
27797 Building a DLL is a way to encapsulate a set of services usable from any
27798 application. As a result, the Ada entities exported by a DLL should be
27799 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
27800 any Ada name mangling. As an example here is an Ada package
27801 @code{API}, spec and body, exporting two procedures, a function, and a
27804 @smallexample @c ada
27807 with Interfaces.C; use Interfaces;
27809 Count : C.int := 0;
27810 function Factorial (Val : C.int) return C.int;
27812 procedure Initialize_API;
27813 procedure Finalize_API;
27814 -- Initialization & Finalization routines. More in the next section.
27816 pragma Export (C, Initialize_API);
27817 pragma Export (C, Finalize_API);
27818 pragma Export (C, Count);
27819 pragma Export (C, Factorial);
27825 @smallexample @c ada
27828 package body API is
27829 function Factorial (Val : C.int) return C.int is
27832 Count := Count + 1;
27833 for K in 1 .. Val loop
27839 procedure Initialize_API is
27841 pragma Import (C, Adainit);
27844 end Initialize_API;
27846 procedure Finalize_API is
27847 procedure Adafinal;
27848 pragma Import (C, Adafinal);
27858 If the Ada DLL you are building will only be used by Ada applications
27859 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
27860 convention. As an example, the previous package could be written as
27863 @smallexample @c ada
27867 Count : Integer := 0;
27868 function Factorial (Val : Integer) return Integer;
27870 procedure Initialize_API;
27871 procedure Finalize_API;
27872 -- Initialization and Finalization routines.
27878 @smallexample @c ada
27881 package body API is
27882 function Factorial (Val : Integer) return Integer is
27883 Fact : Integer := 1;
27885 Count := Count + 1;
27886 for K in 1 .. Val loop
27893 -- The remainder of this package body is unchanged.
27900 Note that if you do not export the Ada entities with a @code{C} or
27901 @code{Stdcall} convention you will have to provide the mangled Ada names
27902 in the definition file of the Ada DLL
27903 (@pxref{Creating the Definition File}).
27905 @node Ada DLLs and Elaboration
27906 @subsection Ada DLLs and Elaboration
27907 @cindex DLLs and elaboration
27910 The DLL that you are building contains your Ada code as well as all the
27911 routines in the Ada library that are needed by it. The first thing a
27912 user of your DLL must do is elaborate the Ada code
27913 (@pxref{Elaboration Order Handling in GNAT}).
27915 To achieve this you must export an initialization routine
27916 (@code{Initialize_API} in the previous example), which must be invoked
27917 before using any of the DLL services. This elaboration routine must call
27918 the Ada elaboration routine @code{adainit} generated by the GNAT binder
27919 (@pxref{Binding with Non-Ada Main Programs}). See the body of
27920 @code{Initialize_Api} for an example. Note that the GNAT binder is
27921 automatically invoked during the DLL build process by the @code{gnatdll}
27922 tool (@pxref{Using gnatdll}).
27924 When a DLL is loaded, Windows systematically invokes a routine called
27925 @code{DllMain}. It would therefore be possible to call @code{adainit}
27926 directly from @code{DllMain} without having to provide an explicit
27927 initialization routine. Unfortunately, it is not possible to call
27928 @code{adainit} from the @code{DllMain} if your program has library level
27929 tasks because access to the @code{DllMain} entry point is serialized by
27930 the system (that is, only a single thread can execute ``through'' it at a
27931 time), which means that the GNAT run time will deadlock waiting for the
27932 newly created task to complete its initialization.
27934 @node Ada DLLs and Finalization
27935 @subsection Ada DLLs and Finalization
27936 @cindex DLLs and finalization
27939 When the services of an Ada DLL are no longer needed, the client code should
27940 invoke the DLL finalization routine, if available. The DLL finalization
27941 routine is in charge of releasing all resources acquired by the DLL. In the
27942 case of the Ada code contained in the DLL, this is achieved by calling
27943 routine @code{adafinal} generated by the GNAT binder
27944 (@pxref{Binding with Non-Ada Main Programs}).
27945 See the body of @code{Finalize_Api} for an
27946 example. As already pointed out the GNAT binder is automatically invoked
27947 during the DLL build process by the @code{gnatdll} tool
27948 (@pxref{Using gnatdll}).
27950 @node Creating a Spec for Ada DLLs
27951 @subsection Creating a Spec for Ada DLLs
27954 To use the services exported by the Ada DLL from another programming
27955 language (e.g.@: C), you have to translate the specs of the exported Ada
27956 entities in that language. For instance in the case of @code{API.dll},
27957 the corresponding C header file could look like:
27962 extern int *_imp__count;
27963 #define count (*_imp__count)
27964 int factorial (int);
27970 It is important to understand that when building an Ada DLL to be used by
27971 other Ada applications, you need two different specs for the packages
27972 contained in the DLL: one for building the DLL and the other for using
27973 the DLL. This is because the @code{DLL} calling convention is needed to
27974 use a variable defined in a DLL, but when building the DLL, the variable
27975 must have either the @code{Ada} or @code{C} calling convention. As an
27976 example consider a DLL comprising the following package @code{API}:
27978 @smallexample @c ada
27982 Count : Integer := 0;
27984 -- Remainder of the package omitted.
27991 After producing a DLL containing package @code{API}, the spec that
27992 must be used to import @code{API.Count} from Ada code outside of the
27995 @smallexample @c ada
28000 pragma Import (DLL, Count);
28006 @node Creating the Definition File
28007 @subsection Creating the Definition File
28010 The definition file is the last file needed to build the DLL. It lists
28011 the exported symbols. As an example, the definition file for a DLL
28012 containing only package @code{API} (where all the entities are exported
28013 with a @code{C} calling convention) is:
28028 If the @code{C} calling convention is missing from package @code{API},
28029 then the definition file contains the mangled Ada names of the above
28030 entities, which in this case are:
28039 api__initialize_api
28044 @node Using gnatdll
28045 @subsection Using @code{gnatdll}
28049 * gnatdll Example::
28050 * gnatdll behind the Scenes::
28055 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
28056 and non-Ada sources that make up your DLL have been compiled.
28057 @code{gnatdll} is actually in charge of two distinct tasks: build the
28058 static import library for the DLL and the actual DLL. The form of the
28059 @code{gnatdll} command is
28063 @c $ gnatdll @ovar{switches} @var{list-of-files} @r{[}-largs @var{opts}@r{]}
28064 @c Expanding @ovar macro inline (explanation in macro def comments)
28065 $ gnatdll @r{[}@var{switches}@r{]} @var{list-of-files} @r{[}-largs @var{opts}@r{]}
28070 where @var{list-of-files} is a list of ALI and object files. The object
28071 file list must be the exact list of objects corresponding to the non-Ada
28072 sources whose services are to be included in the DLL. The ALI file list
28073 must be the exact list of ALI files for the corresponding Ada sources
28074 whose services are to be included in the DLL. If @var{list-of-files} is
28075 missing, only the static import library is generated.
28078 You may specify any of the following switches to @code{gnatdll}:
28081 @c @item -a@ovar{address}
28082 @c Expanding @ovar macro inline (explanation in macro def comments)
28083 @item -a@r{[}@var{address}@r{]}
28084 @cindex @option{-a} (@code{gnatdll})
28085 Build a non-relocatable DLL at @var{address}. If @var{address} is not
28086 specified the default address @var{0x11000000} will be used. By default,
28087 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
28088 advise the reader to build relocatable DLL.
28090 @item -b @var{address}
28091 @cindex @option{-b} (@code{gnatdll})
28092 Set the relocatable DLL base address. By default the address is
28095 @item -bargs @var{opts}
28096 @cindex @option{-bargs} (@code{gnatdll})
28097 Binder options. Pass @var{opts} to the binder.
28099 @item -d @var{dllfile}
28100 @cindex @option{-d} (@code{gnatdll})
28101 @var{dllfile} is the name of the DLL. This switch must be present for
28102 @code{gnatdll} to do anything. The name of the generated import library is
28103 obtained algorithmically from @var{dllfile} as shown in the following
28104 example: if @var{dllfile} is @code{xyz.dll}, the import library name is
28105 @code{libxyz.dll.a}. The name of the definition file to use (if not specified
28106 by option @option{-e}) is obtained algorithmically from @var{dllfile}
28107 as shown in the following example:
28108 if @var{dllfile} is @code{xyz.dll}, the definition
28109 file used is @code{xyz.def}.
28111 @item -e @var{deffile}
28112 @cindex @option{-e} (@code{gnatdll})
28113 @var{deffile} is the name of the definition file.
28116 @cindex @option{-g} (@code{gnatdll})
28117 Generate debugging information. This information is stored in the object
28118 file and copied from there to the final DLL file by the linker,
28119 where it can be read by the debugger. You must use the
28120 @option{-g} switch if you plan on using the debugger or the symbolic
28124 @cindex @option{-h} (@code{gnatdll})
28125 Help mode. Displays @code{gnatdll} switch usage information.
28128 @cindex @option{-I} (@code{gnatdll})
28129 Direct @code{gnatdll} to search the @var{dir} directory for source and
28130 object files needed to build the DLL.
28131 (@pxref{Search Paths and the Run-Time Library (RTL)}).
28134 @cindex @option{-k} (@code{gnatdll})
28135 Removes the @code{@@}@var{nn} suffix from the import library's exported
28136 names, but keeps them for the link names. You must specify this
28137 option if you want to use a @code{Stdcall} function in a DLL for which
28138 the @code{@@}@var{nn} suffix has been removed. This is the case for most
28139 of the Windows NT DLL for example. This option has no effect when
28140 @option{-n} option is specified.
28142 @item -l @var{file}
28143 @cindex @option{-l} (@code{gnatdll})
28144 The list of ALI and object files used to build the DLL are listed in
28145 @var{file}, instead of being given in the command line. Each line in
28146 @var{file} contains the name of an ALI or object file.
28149 @cindex @option{-n} (@code{gnatdll})
28150 No Import. Do not create the import library.
28153 @cindex @option{-q} (@code{gnatdll})
28154 Quiet mode. Do not display unnecessary messages.
28157 @cindex @option{-v} (@code{gnatdll})
28158 Verbose mode. Display extra information.
28160 @item -largs @var{opts}
28161 @cindex @option{-largs} (@code{gnatdll})
28162 Linker options. Pass @var{opts} to the linker.
28165 @node gnatdll Example
28166 @subsubsection @code{gnatdll} Example
28169 As an example the command to build a relocatable DLL from @file{api.adb}
28170 once @file{api.adb} has been compiled and @file{api.def} created is
28173 $ gnatdll -d api.dll api.ali
28177 The above command creates two files: @file{libapi.dll.a} (the import
28178 library) and @file{api.dll} (the actual DLL). If you want to create
28179 only the DLL, just type:
28182 $ gnatdll -d api.dll -n api.ali
28186 Alternatively if you want to create just the import library, type:
28189 $ gnatdll -d api.dll
28192 @node gnatdll behind the Scenes
28193 @subsubsection @code{gnatdll} behind the Scenes
28196 This section details the steps involved in creating a DLL. @code{gnatdll}
28197 does these steps for you. Unless you are interested in understanding what
28198 goes on behind the scenes, you should skip this section.
28200 We use the previous example of a DLL containing the Ada package @code{API},
28201 to illustrate the steps necessary to build a DLL. The starting point is a
28202 set of objects that will make up the DLL and the corresponding ALI
28203 files. In the case of this example this means that @file{api.o} and
28204 @file{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
28209 @code{gnatdll} builds the base file (@file{api.base}). A base file gives
28210 the information necessary to generate relocation information for the
28216 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
28221 In addition to the base file, the @command{gnatlink} command generates an
28222 output file @file{api.jnk} which can be discarded. The @option{-mdll} switch
28223 asks @command{gnatlink} to generate the routines @code{DllMain} and
28224 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
28225 is loaded into memory.
28228 @code{gnatdll} uses @code{dlltool} (@pxref{Using dlltool}) to build the
28229 export table (@file{api.exp}). The export table contains the relocation
28230 information in a form which can be used during the final link to ensure
28231 that the Windows loader is able to place the DLL anywhere in memory.
28235 $ dlltool --dllname api.dll --def api.def --base-file api.base \
28236 --output-exp api.exp
28241 @code{gnatdll} builds the base file using the new export table. Note that
28242 @command{gnatbind} must be called once again since the binder generated file
28243 has been deleted during the previous call to @command{gnatlink}.
28248 $ gnatlink api -o api.jnk api.exp -mdll
28249 -Wl,--base-file,api.base
28254 @code{gnatdll} builds the new export table using the new base file and
28255 generates the DLL import library @file{libAPI.dll.a}.
28259 $ dlltool --dllname api.dll --def api.def --base-file api.base \
28260 --output-exp api.exp --output-lib libAPI.a
28265 Finally @code{gnatdll} builds the relocatable DLL using the final export
28271 $ gnatlink api api.exp -o api.dll -mdll
28276 @node Using dlltool
28277 @subsubsection Using @code{dlltool}
28280 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
28281 DLLs and static import libraries. This section summarizes the most
28282 common @code{dlltool} switches. The form of the @code{dlltool} command
28286 @c $ dlltool @ovar{switches}
28287 @c Expanding @ovar macro inline (explanation in macro def comments)
28288 $ dlltool @r{[}@var{switches}@r{]}
28292 @code{dlltool} switches include:
28295 @item --base-file @var{basefile}
28296 @cindex @option{--base-file} (@command{dlltool})
28297 Read the base file @var{basefile} generated by the linker. This switch
28298 is used to create a relocatable DLL.
28300 @item --def @var{deffile}
28301 @cindex @option{--def} (@command{dlltool})
28302 Read the definition file.
28304 @item --dllname @var{name}
28305 @cindex @option{--dllname} (@command{dlltool})
28306 Gives the name of the DLL. This switch is used to embed the name of the
28307 DLL in the static import library generated by @code{dlltool} with switch
28308 @option{--output-lib}.
28311 @cindex @option{-k} (@command{dlltool})
28312 Kill @code{@@}@var{nn} from exported names
28313 (@pxref{Windows Calling Conventions}
28314 for a discussion about @code{Stdcall}-style symbols.
28317 @cindex @option{--help} (@command{dlltool})
28318 Prints the @code{dlltool} switches with a concise description.
28320 @item --output-exp @var{exportfile}
28321 @cindex @option{--output-exp} (@command{dlltool})
28322 Generate an export file @var{exportfile}. The export file contains the
28323 export table (list of symbols in the DLL) and is used to create the DLL.
28325 @item --output-lib @var{libfile}
28326 @cindex @option{--output-lib} (@command{dlltool})
28327 Generate a static import library @var{libfile}.
28330 @cindex @option{-v} (@command{dlltool})
28333 @item --as @var{assembler-name}
28334 @cindex @option{--as} (@command{dlltool})
28335 Use @var{assembler-name} as the assembler. The default is @code{as}.
28338 @node GNAT and Windows Resources
28339 @section GNAT and Windows Resources
28340 @cindex Resources, windows
28343 * Building Resources::
28344 * Compiling Resources::
28345 * Using Resources::
28349 Resources are an easy way to add Windows specific objects to your
28350 application. The objects that can be added as resources include:
28379 This section explains how to build, compile and use resources.
28381 @node Building Resources
28382 @subsection Building Resources
28383 @cindex Resources, building
28386 A resource file is an ASCII file. By convention resource files have an
28387 @file{.rc} extension.
28388 The easiest way to build a resource file is to use Microsoft tools
28389 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
28390 @code{dlgedit.exe} to build dialogs.
28391 It is always possible to build an @file{.rc} file yourself by writing a
28394 It is not our objective to explain how to write a resource file. A
28395 complete description of the resource script language can be found in the
28396 Microsoft documentation.
28398 @node Compiling Resources
28399 @subsection Compiling Resources
28402 @cindex Resources, compiling
28405 This section describes how to build a GNAT-compatible (COFF) object file
28406 containing the resources. This is done using the Resource Compiler
28407 @code{windres} as follows:
28410 $ windres -i myres.rc -o myres.o
28414 By default @code{windres} will run @command{gcc} to preprocess the @file{.rc}
28415 file. You can specify an alternate preprocessor (usually named
28416 @file{cpp.exe}) using the @code{windres} @option{--preprocessor}
28417 parameter. A list of all possible options may be obtained by entering
28418 the command @code{windres} @option{--help}.
28420 It is also possible to use the Microsoft resource compiler @code{rc.exe}
28421 to produce a @file{.res} file (binary resource file). See the
28422 corresponding Microsoft documentation for further details. In this case
28423 you need to use @code{windres} to translate the @file{.res} file to a
28424 GNAT-compatible object file as follows:
28427 $ windres -i myres.res -o myres.o
28430 @node Using Resources
28431 @subsection Using Resources
28432 @cindex Resources, using
28435 To include the resource file in your program just add the
28436 GNAT-compatible object file for the resource(s) to the linker
28437 arguments. With @command{gnatmake} this is done by using the @option{-largs}
28441 $ gnatmake myprog -largs myres.o
28444 @node Debugging a DLL
28445 @section Debugging a DLL
28446 @cindex DLL debugging
28449 * Program and DLL Both Built with GCC/GNAT::
28450 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
28454 Debugging a DLL is similar to debugging a standard program. But
28455 we have to deal with two different executable parts: the DLL and the
28456 program that uses it. We have the following four possibilities:
28460 The program and the DLL are built with @code{GCC/GNAT}.
28462 The program is built with foreign tools and the DLL is built with
28465 The program is built with @code{GCC/GNAT} and the DLL is built with
28470 In this section we address only cases one and two above.
28471 There is no point in trying to debug
28472 a DLL with @code{GNU/GDB}, if there is no GDB-compatible debugging
28473 information in it. To do so you must use a debugger compatible with the
28474 tools suite used to build the DLL.
28476 @node Program and DLL Both Built with GCC/GNAT
28477 @subsection Program and DLL Both Built with GCC/GNAT
28480 This is the simplest case. Both the DLL and the program have @code{GDB}
28481 compatible debugging information. It is then possible to break anywhere in
28482 the process. Let's suppose here that the main procedure is named
28483 @code{ada_main} and that in the DLL there is an entry point named
28487 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) and
28488 program must have been built with the debugging information (see GNAT -g
28489 switch). Here are the step-by-step instructions for debugging it:
28492 @item Launch @code{GDB} on the main program.
28498 @item Start the program and stop at the beginning of the main procedure
28505 This step is required to be able to set a breakpoint inside the DLL. As long
28506 as the program is not run, the DLL is not loaded. This has the
28507 consequence that the DLL debugging information is also not loaded, so it is not
28508 possible to set a breakpoint in the DLL.
28510 @item Set a breakpoint inside the DLL
28513 (gdb) break ada_dll
28520 At this stage a breakpoint is set inside the DLL. From there on
28521 you can use the standard approach to debug the whole program
28522 (@pxref{Running and Debugging Ada Programs}).
28525 @c This used to work, probably because the DLLs were non-relocatable
28526 @c keep this section around until the problem is sorted out.
28528 To break on the @code{DllMain} routine it is not possible to follow
28529 the procedure above. At the time the program stop on @code{ada_main}
28530 the @code{DllMain} routine as already been called. Either you can use
28531 the procedure below @pxref{Debugging the DLL Directly} or this procedure:
28534 @item Launch @code{GDB} on the main program.
28540 @item Load DLL symbols
28543 (gdb) add-sym api.dll
28546 @item Set a breakpoint inside the DLL
28549 (gdb) break ada_dll.adb:45
28552 Note that at this point it is not possible to break using the routine symbol
28553 directly as the program is not yet running. The solution is to break
28554 on the proper line (break in @file{ada_dll.adb} line 45).
28556 @item Start the program
28565 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT
28566 @subsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
28569 * Debugging the DLL Directly::
28570 * Attaching to a Running Process::
28574 In this case things are slightly more complex because it is not possible to
28575 start the main program and then break at the beginning to load the DLL and the
28576 associated DLL debugging information. It is not possible to break at the
28577 beginning of the program because there is no @code{GDB} debugging information,
28578 and therefore there is no direct way of getting initial control. This
28579 section addresses this issue by describing some methods that can be used
28580 to break somewhere in the DLL to debug it.
28583 First suppose that the main procedure is named @code{main} (this is for
28584 example some C code built with Microsoft Visual C) and that there is a
28585 DLL named @code{test.dll} containing an Ada entry point named
28589 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) must have
28590 been built with debugging information (see GNAT -g option).
28592 @node Debugging the DLL Directly
28593 @subsubsection Debugging the DLL Directly
28597 Find out the executable starting address
28600 $ objdump --file-header main.exe
28603 The starting address is reported on the last line. For example:
28606 main.exe: file format pei-i386
28607 architecture: i386, flags 0x0000010a:
28608 EXEC_P, HAS_DEBUG, D_PAGED
28609 start address 0x00401010
28613 Launch the debugger on the executable.
28620 Set a breakpoint at the starting address, and launch the program.
28623 $ (gdb) break *0x00401010
28627 The program will stop at the given address.
28630 Set a breakpoint on a DLL subroutine.
28633 (gdb) break ada_dll.adb:45
28636 Or if you want to break using a symbol on the DLL, you need first to
28637 select the Ada language (language used by the DLL).
28640 (gdb) set language ada
28641 (gdb) break ada_dll
28645 Continue the program.
28652 This will run the program until it reaches the breakpoint that has been
28653 set. From that point you can use the standard way to debug a program
28654 as described in (@pxref{Running and Debugging Ada Programs}).
28659 It is also possible to debug the DLL by attaching to a running process.
28661 @node Attaching to a Running Process
28662 @subsubsection Attaching to a Running Process
28663 @cindex DLL debugging, attach to process
28666 With @code{GDB} it is always possible to debug a running process by
28667 attaching to it. It is possible to debug a DLL this way. The limitation
28668 of this approach is that the DLL must run long enough to perform the
28669 attach operation. It may be useful for instance to insert a time wasting
28670 loop in the code of the DLL to meet this criterion.
28674 @item Launch the main program @file{main.exe}.
28680 @item Use the Windows @i{Task Manager} to find the process ID. Let's say
28681 that the process PID for @file{main.exe} is 208.
28689 @item Attach to the running process to be debugged.
28695 @item Load the process debugging information.
28698 (gdb) symbol-file main.exe
28701 @item Break somewhere in the DLL.
28704 (gdb) break ada_dll
28707 @item Continue process execution.
28716 This last step will resume the process execution, and stop at
28717 the breakpoint we have set. From there you can use the standard
28718 approach to debug a program as described in
28719 (@pxref{Running and Debugging Ada Programs}).
28721 @node Setting Stack Size from gnatlink
28722 @section Setting Stack Size from @command{gnatlink}
28725 It is possible to specify the program stack size at link time. On modern
28726 versions of Windows, starting with XP, this is mostly useful to set the size of
28727 the main stack (environment task). The other task stacks are set with pragma
28728 Storage_Size or with the @command{gnatbind -d} command.
28730 Since older versions of Windows (2000, NT4, etc.) do not allow setting the
28731 reserve size of individual tasks, the link-time stack size applies to all
28732 tasks, and pragma Storage_Size has no effect.
28733 In particular, Stack Overflow checks are made against this
28734 link-time specified size.
28736 This setting can be done with
28737 @command{gnatlink} using either:
28741 @item using @option{-Xlinker} linker option
28744 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
28747 This sets the stack reserve size to 0x10000 bytes and the stack commit
28748 size to 0x1000 bytes.
28750 @item using @option{-Wl} linker option
28753 $ gnatlink hello -Wl,--stack=0x1000000
28756 This sets the stack reserve size to 0x1000000 bytes. Note that with
28757 @option{-Wl} option it is not possible to set the stack commit size
28758 because the coma is a separator for this option.
28762 @node Setting Heap Size from gnatlink
28763 @section Setting Heap Size from @command{gnatlink}
28766 Under Windows systems, it is possible to specify the program heap size from
28767 @command{gnatlink} using either:
28771 @item using @option{-Xlinker} linker option
28774 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
28777 This sets the heap reserve size to 0x10000 bytes and the heap commit
28778 size to 0x1000 bytes.
28780 @item using @option{-Wl} linker option
28783 $ gnatlink hello -Wl,--heap=0x1000000
28786 This sets the heap reserve size to 0x1000000 bytes. Note that with
28787 @option{-Wl} option it is not possible to set the heap commit size
28788 because the coma is a separator for this option.
28794 @c **********************************
28795 @c * GNU Free Documentation License *
28796 @c **********************************
28798 @c GNU Free Documentation License
28800 @node Index,,GNU Free Documentation License, Top
28806 @c Put table of contents at end, otherwise it precedes the "title page" in
28807 @c the .txt version
28808 @c Edit the pdf file to move the contents to the beginning, after the title