1 \input texinfo @c -*-texinfo-*-
3 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
5 @c GNAT DOCUMENTATION o
9 @c Copyright (C) 1992-2007, AdaCore o
11 @c GNAT is free software; you can redistribute it and/or modify it under o
12 @c terms of the GNU General Public License as published by the Free Soft- o
13 @c ware Foundation; either version 2, or (at your option) any later ver- o
14 @c sion. GNAT is distributed in the hope that it will be useful, but WITH- o
15 @c OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY o
16 @c or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License o
17 @c for more details. You should have received a copy of the GNU General o
18 @c Public License distributed with GNAT; see file COPYING. If not, write o
19 @c to the Free Software Foundation, 51 Franklin Street, Fifth Floor, o
20 @c Boston, MA 02110-1301, USA. o
22 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
24 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
26 @c GNAT_UGN Style Guide
28 @c 1. Always put a @noindent on the line before the first paragraph
29 @c after any of these commands:
41 @c 2. DO NOT use @example. Use @smallexample instead.
42 @c a) DO NOT use highlighting commands (@b{}, @i{}) inside an @smallexample
43 @c context. These can interfere with the readability of the texi
44 @c source file. Instead, use one of the following annotated
45 @c @smallexample commands, and preprocess the texi file with the
46 @c ada2texi tool (which generates appropriate highlighting):
47 @c @smallexample @c ada
48 @c @smallexample @c adanocomment
49 @c @smallexample @c projectfile
50 @c b) The "@c ada" markup will result in boldface for reserved words
51 @c and italics for comments
52 @c c) The "@c adanocomment" markup will result only in boldface for
53 @c reserved words (comments are left alone)
54 @c d) The "@c projectfile" markup is like "@c ada" except that the set
55 @c of reserved words include the new reserved words for project files
57 @c 3. Each @chapter, @section, @subsection, @subsubsection, etc.
58 @c command must be preceded by two empty lines
60 @c 4. The @item command should be on a line of its own if it is in an
61 @c @itemize or @enumerate command.
63 @c 5. When talking about ALI files use "ALI" (all uppercase), not "Ali"
66 @c 6. DO NOT put trailing spaces at the end of a line. Such spaces will
67 @c cause the document build to fail.
69 @c 7. DO NOT use @cartouche for examples that are longer than around 10 lines.
70 @c This command inhibits page breaks, so long examples in a @cartouche can
71 @c lead to large, ugly patches of empty space on a page.
73 @c NOTE: This file should be submitted to xgnatugn with either the vms flag
74 @c or the unw flag set. The unw flag covers topics for both Unix and
77 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
79 @setfilename gnat_ugn.info
82 @c This flag is used where the text refers to conditions that exist when the
83 @c text was entered into the document but which may change over time.
84 @c Update the setting for the flag, and (if necessary) the text surrounding,
85 @c the references to the flag, on future doc revisions:
86 @c search for @value{NOW}.
90 @set DEFAULTLANGUAGEVERSION Ada 2005
91 @set NONDEFAULTLANGUAGEVERSION Ada 95
94 @setfilename gnat_ugn_unw.info
99 @set FILE gnat_ugn_unw
103 @set PLATFORM OpenVMS
104 @set FILE gnat_ugn_vms
107 @settitle @value{EDITION} User's Guide @value{PLATFORM}
108 @dircategory GNU Ada tools
110 * @value{EDITION} User's Guide (@value{FILE}) @value{PLATFORM}
113 @include gcc-common.texi
115 @setchapternewpage odd
120 Copyright @copyright{} 1995-2005, Free Software Foundation
122 Permission is granted to copy, distribute and/or modify this document
123 under the terms of the GNU Free Documentation License, Version 1.2
124 or any later version published by the Free Software Foundation;
125 with the Invariant Sections being ``GNU Free Documentation License'', with the
126 Front-Cover Texts being
127 ``@value{EDITION} User's Guide'',
128 and with no Back-Cover Texts.
129 A copy of the license is included in the section entitled
130 ``GNU Free Documentation License''.
134 @title @value{EDITION} User's Guide
138 @titlefont{@i{@value{PLATFORM}}}
144 @subtitle GNAT, The GNU Ada Compiler
149 @vskip 0pt plus 1filll
156 @node Top, About This Guide, (dir), (dir)
157 @top @value{EDITION} User's Guide
160 @value{EDITION} User's Guide @value{PLATFORM}
163 GNAT, The GNU Ada Compiler@*
164 GCC version @value{version-GCC}@*
171 * Getting Started with GNAT::
172 * The GNAT Compilation Model::
173 * Compiling Using gcc::
174 * Binding Using gnatbind::
175 * Linking Using gnatlink::
176 * The GNAT Make Program gnatmake::
177 * Improving Performance::
178 * Renaming Files Using gnatchop::
179 * Configuration Pragmas::
180 * Handling Arbitrary File Naming Conventions Using gnatname::
181 * GNAT Project Manager::
182 * The Cross-Referencing Tools gnatxref and gnatfind::
183 * The GNAT Pretty-Printer gnatpp::
184 * The GNAT Metric Tool gnatmetric::
185 * File Name Krunching Using gnatkr::
186 * Preprocessing Using gnatprep::
188 * The GNAT Run-Time Library Builder gnatlbr::
190 * The GNAT Library Browser gnatls::
191 * Cleaning Up Using gnatclean::
193 * GNAT and Libraries::
194 * Using the GNU make Utility::
196 * Memory Management Issues::
197 * Stack Related Facilities::
198 * Verifying Properties Using gnatcheck::
199 * Creating Sample Bodies Using gnatstub::
200 * Other Utility Programs::
201 * Running and Debugging Ada Programs::
203 * Compatibility with HP Ada::
205 * Platform-Specific Information for the Run-Time Libraries::
206 * Example of Binder Output File::
207 * Elaboration Order Handling in GNAT::
209 * Compatibility and Porting Guide::
211 * Microsoft Windows Topics::
213 * GNU Free Documentation License::
216 --- The Detailed Node Listing ---
220 * What This Guide Contains::
221 * What You Should Know before Reading This Guide::
222 * Related Information::
225 Getting Started with GNAT
228 * Running a Simple Ada Program::
229 * Running a Program with Multiple Units::
230 * Using the gnatmake Utility::
232 * Editing with Emacs::
235 * Introduction to GPS::
238 The GNAT Compilation Model
240 * Source Representation::
241 * Foreign Language Representation::
242 * File Naming Rules::
243 * Using Other File Names::
244 * Alternative File Naming Schemes::
245 * Generating Object Files::
246 * Source Dependencies::
247 * The Ada Library Information Files::
248 * Binding an Ada Program::
249 * Mixed Language Programming::
251 * Building Mixed Ada & C++ Programs::
252 * Comparison between GNAT and C/C++ Compilation Models::
254 * Comparison between GNAT and Conventional Ada Library Models::
256 * Placement of temporary files::
259 Foreign Language Representation
262 * Other 8-Bit Codes::
263 * Wide Character Encodings::
265 Compiling Ada Programs With gcc
267 * Compiling Programs::
269 * Search Paths and the Run-Time Library (RTL)::
270 * Order of Compilation Issues::
275 * Output and Error Message Control::
276 * Warning Message Control::
277 * Debugging and Assertion Control::
278 * Validity Checking::
281 * Using gcc for Syntax Checking::
282 * Using gcc for Semantic Checking::
283 * Compiling Different Versions of Ada::
284 * Character Set Control::
285 * File Naming Control::
286 * Subprogram Inlining Control::
287 * Auxiliary Output Control::
288 * Debugging Control::
289 * Exception Handling Control::
290 * Units to Sources Mapping Files::
291 * Integrated Preprocessing::
296 Binding Ada Programs With gnatbind
299 * Switches for gnatbind::
300 * Command-Line Access::
301 * Search Paths for gnatbind::
302 * Examples of gnatbind Usage::
304 Switches for gnatbind
306 * Consistency-Checking Modes::
307 * Binder Error Message Control::
308 * Elaboration Control::
310 * Binding with Non-Ada Main Programs::
311 * Binding Programs with No Main Subprogram::
313 Linking Using gnatlink
316 * Switches for gnatlink::
318 The GNAT Make Program gnatmake
321 * Switches for gnatmake::
322 * Mode Switches for gnatmake::
323 * Notes on the Command Line::
324 * How gnatmake Works::
325 * Examples of gnatmake Usage::
327 Improving Performance
328 * Performance Considerations::
329 * Reducing Size of Ada Executables with gnatelim::
330 * Reducing Size of Executables with unused subprogram/data elimination::
332 Performance Considerations
333 * Controlling Run-Time Checks::
334 * Use of Restrictions::
335 * Optimization Levels::
336 * Debugging Optimized Code::
337 * Inlining of Subprograms::
338 * Other Optimization Switches::
339 * Optimization and Strict Aliasing::
341 * Coverage Analysis::
344 Reducing Size of Ada Executables with gnatelim
347 * Correcting the List of Eliminate Pragmas::
348 * Making Your Executables Smaller::
349 * Summary of the gnatelim Usage Cycle::
351 Reducing Size of Executables with unused subprogram/data elimination
352 * About unused subprogram/data elimination::
353 * Compilation options::
355 Renaming Files Using gnatchop
357 * Handling Files with Multiple Units::
358 * Operating gnatchop in Compilation Mode::
359 * Command Line for gnatchop::
360 * Switches for gnatchop::
361 * Examples of gnatchop Usage::
363 Configuration Pragmas
365 * Handling of Configuration Pragmas::
366 * The Configuration Pragmas Files::
368 Handling Arbitrary File Naming Conventions Using gnatname
370 * Arbitrary File Naming Conventions::
372 * Switches for gnatname::
373 * Examples of gnatname Usage::
378 * Examples of Project Files::
379 * Project File Syntax::
380 * Objects and Sources in Project Files::
381 * Importing Projects::
382 * Project Extension::
383 * Project Hierarchy Extension::
384 * External References in Project Files::
385 * Packages in Project Files::
386 * Variables from Imported Projects::
389 * Stand-alone Library Projects::
390 * Switches Related to Project Files::
391 * Tools Supporting Project Files::
392 * An Extended Example::
393 * Project File Complete Syntax::
395 The Cross-Referencing Tools gnatxref and gnatfind
397 * gnatxref Switches::
398 * gnatfind Switches::
399 * Project Files for gnatxref and gnatfind::
400 * Regular Expressions in gnatfind and gnatxref::
401 * Examples of gnatxref Usage::
402 * Examples of gnatfind Usage::
404 The GNAT Pretty-Printer gnatpp
406 * Switches for gnatpp::
409 The GNAT Metrics Tool gnatmetric
411 * Switches for gnatmetric::
413 File Name Krunching Using gnatkr
418 * Examples of gnatkr Usage::
420 Preprocessing Using gnatprep
423 * Switches for gnatprep::
424 * Form of Definitions File::
425 * Form of Input Text for gnatprep::
428 The GNAT Run-Time Library Builder gnatlbr
431 * Switches for gnatlbr::
432 * Examples of gnatlbr Usage::
435 The GNAT Library Browser gnatls
438 * Switches for gnatls::
439 * Examples of gnatls Usage::
441 Cleaning Up Using gnatclean
443 * Running gnatclean::
444 * Switches for gnatclean::
445 @c * Examples of gnatclean Usage::
451 * Introduction to Libraries in GNAT::
452 * General Ada Libraries::
453 * Stand-alone Ada Libraries::
454 * Rebuilding the GNAT Run-Time Library::
456 Using the GNU make Utility
458 * Using gnatmake in a Makefile::
459 * Automatically Creating a List of Directories::
460 * Generating the Command Line Switches::
461 * Overcoming Command Line Length Limits::
464 Memory Management Issues
466 * Some Useful Memory Pools::
467 * The GNAT Debug Pool Facility::
472 Stack Related Facilities
474 * Stack Overflow Checking::
475 * Static Stack Usage Analysis::
476 * Dynamic Stack Usage Analysis::
478 Some Useful Memory Pools
480 The GNAT Debug Pool Facility
486 * Switches for gnatmem::
487 * Example of gnatmem Usage::
490 Verifying Properties Using gnatcheck
492 * Format of the Report File::
493 * General gnatcheck Switches::
494 * gnatcheck Rule Options::
495 * Adding the Results of Compiler Checks to gnatcheck Output::
496 * Project-Wide Checks::
499 Sample Bodies Using gnatstub
502 * Switches for gnatstub::
504 Other Utility Programs
506 * Using Other Utility Programs with GNAT::
507 * The External Symbol Naming Scheme of GNAT::
508 * Converting Ada Files to html with gnathtml::
510 Running and Debugging Ada Programs
512 * The GNAT Debugger GDB::
514 * Introduction to GDB Commands::
515 * Using Ada Expressions::
516 * Calling User-Defined Subprograms::
517 * Using the Next Command in a Function::
520 * Debugging Generic Units::
521 * GNAT Abnormal Termination or Failure to Terminate::
522 * Naming Conventions for GNAT Source Files::
523 * Getting Internal Debugging Information::
531 Compatibility with HP Ada
533 * Ada Language Compatibility::
534 * Differences in the Definition of Package System::
535 * Language-Related Features::
536 * The Package STANDARD::
537 * The Package SYSTEM::
538 * Tasking and Task-Related Features::
539 * Pragmas and Pragma-Related Features::
540 * Library of Predefined Units::
542 * Main Program Definition::
543 * Implementation-Defined Attributes::
544 * Compiler and Run-Time Interfacing::
545 * Program Compilation and Library Management::
547 * Implementation Limits::
548 * Tools and Utilities::
550 Language-Related Features
552 * Integer Types and Representations::
553 * Floating-Point Types and Representations::
554 * Pragmas Float_Representation and Long_Float::
555 * Fixed-Point Types and Representations::
556 * Record and Array Component Alignment::
558 * Other Representation Clauses::
560 Tasking and Task-Related Features
562 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
563 * Assigning Task IDs::
564 * Task IDs and Delays::
565 * Task-Related Pragmas::
566 * Scheduling and Task Priority::
568 * External Interrupts::
570 Pragmas and Pragma-Related Features
572 * Restrictions on the Pragma INLINE::
573 * Restrictions on the Pragma INTERFACE::
574 * Restrictions on the Pragma SYSTEM_NAME::
576 Library of Predefined Units
578 * Changes to DECLIB::
582 * Shared Libraries and Options Files::
586 Platform-Specific Information for the Run-Time Libraries
588 * Summary of Run-Time Configurations::
589 * Specifying a Run-Time Library::
590 * Choosing the Scheduling Policy::
591 * Solaris-Specific Considerations::
592 * Linux-Specific Considerations::
593 * AIX-Specific Considerations::
595 Example of Binder Output File
597 Elaboration Order Handling in GNAT
600 * Checking the Elaboration Order::
601 * Controlling the Elaboration Order::
602 * Controlling Elaboration in GNAT - Internal Calls::
603 * Controlling Elaboration in GNAT - External Calls::
604 * Default Behavior in GNAT - Ensuring Safety::
605 * Treatment of Pragma Elaborate::
606 * Elaboration Issues for Library Tasks::
607 * Mixing Elaboration Models::
608 * What to Do If the Default Elaboration Behavior Fails::
609 * Elaboration for Access-to-Subprogram Values::
610 * Summary of Procedures for Elaboration Control::
611 * Other Elaboration Order Considerations::
615 * Basic Assembler Syntax::
616 * A Simple Example of Inline Assembler::
617 * Output Variables in Inline Assembler::
618 * Input Variables in Inline Assembler::
619 * Inlining Inline Assembler Code::
620 * Other Asm Functionality::
622 Compatibility and Porting Guide
624 * Compatibility with Ada 83::
625 * Compatibility between Ada 95 and Ada 2005::
626 * Implementation-dependent characteristics::
628 @c This brief section is only in the non-VMS version
629 @c The complete chapter on HP Ada issues is in the VMS version
630 * Compatibility with HP Ada 83::
632 * Compatibility with Other Ada Systems::
633 * Representation Clauses::
635 * Transitioning to 64-Bit GNAT for OpenVMS::
639 Microsoft Windows Topics
641 * Using GNAT on Windows::
642 * CONSOLE and WINDOWS subsystems::
644 * Mixed-Language Programming on Windows::
645 * Windows Calling Conventions::
646 * Introduction to Dynamic Link Libraries (DLLs)::
647 * Using DLLs with GNAT::
648 * Building DLLs with GNAT::
649 * GNAT and Windows Resources::
651 * Setting Stack Size from gnatlink::
652 * Setting Heap Size from gnatlink::
659 @node About This Guide
660 @unnumbered About This Guide
664 This guide describes the use of @value{EDITION},
665 a compiler and software development toolset for the full Ada
666 programming language, implemented on OpenVMS for HP's Alpha and
667 Integrity server (I64) platforms.
670 This guide describes the use of @value{EDITION},
671 a compiler and software development
672 toolset for the full Ada programming language.
674 It documents the features of the compiler and tools, and explains
675 how to use them to build Ada applications.
677 @value{EDITION} implements Ada 95 and Ada 2005, and it may also be invoked in
678 Ada 83 compatibility mode.
679 By default, @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
680 but you can override with a compiler switch
681 (@pxref{Compiling Different Versions of Ada})
682 to explicitly specify the language version.
683 Throughout this manual, references to ``Ada'' without a year suffix
684 apply to both the Ada 95 and Ada 2005 versions of the language.
688 For ease of exposition, ``@value{EDITION}'' will be referred to simply as
689 ``GNAT'' in the remainder of this document.
696 * What This Guide Contains::
697 * What You Should Know before Reading This Guide::
698 * Related Information::
702 @node What This Guide Contains
703 @unnumberedsec What This Guide Contains
706 This guide contains the following chapters:
710 @ref{Getting Started with GNAT}, describes how to get started compiling
711 and running Ada programs with the GNAT Ada programming environment.
713 @ref{The GNAT Compilation Model}, describes the compilation model used
717 @ref{Compiling Using gcc}, describes how to compile
718 Ada programs with @command{gcc}, the Ada compiler.
721 @ref{Binding Using gnatbind}, describes how to
722 perform binding of Ada programs with @code{gnatbind}, the GNAT binding
726 @ref{Linking Using gnatlink},
727 describes @command{gnatlink}, a
728 program that provides for linking using the GNAT run-time library to
729 construct a program. @command{gnatlink} can also incorporate foreign language
730 object units into the executable.
733 @ref{The GNAT Make Program gnatmake}, describes @command{gnatmake}, a
734 utility that automatically determines the set of sources
735 needed by an Ada compilation unit, and executes the necessary compilations
739 @ref{Improving Performance}, shows various techniques for making your
740 Ada program run faster or take less space.
741 It discusses the effect of the compiler's optimization switch and
742 also describes the @command{gnatelim} tool and unused subprogram/data
746 @ref{Renaming Files Using gnatchop}, describes
747 @code{gnatchop}, a utility that allows you to preprocess a file that
748 contains Ada source code, and split it into one or more new files, one
749 for each compilation unit.
752 @ref{Configuration Pragmas}, describes the configuration pragmas
756 @ref{Handling Arbitrary File Naming Conventions Using gnatname},
757 shows how to override the default GNAT file naming conventions,
758 either for an individual unit or globally.
761 @ref{GNAT Project Manager}, describes how to use project files
762 to organize large projects.
765 @ref{The Cross-Referencing Tools gnatxref and gnatfind}, discusses
766 @code{gnatxref} and @code{gnatfind}, two tools that provide an easy
767 way to navigate through sources.
770 @ref{The GNAT Pretty-Printer gnatpp}, shows how to produce a reformatted
771 version of an Ada source file with control over casing, indentation,
772 comment placement, and other elements of program presentation style.
775 @ref{The GNAT Metric Tool gnatmetric}, shows how to compute various
776 metrics for an Ada source file, such as the number of types and subprograms,
777 and assorted complexity measures.
780 @ref{File Name Krunching Using gnatkr}, describes the @code{gnatkr}
781 file name krunching utility, used to handle shortened
782 file names on operating systems with a limit on the length of names.
785 @ref{Preprocessing Using gnatprep}, describes @code{gnatprep}, a
786 preprocessor utility that allows a single source file to be used to
787 generate multiple or parameterized source files, by means of macro
792 @ref{The GNAT Run-Time Library Builder gnatlbr}, describes @command{gnatlbr},
793 a tool for rebuilding the GNAT run time with user-supplied
794 configuration pragmas.
798 @ref{The GNAT Library Browser gnatls}, describes @code{gnatls}, a
799 utility that displays information about compiled units, including dependences
800 on the corresponding sources files, and consistency of compilations.
803 @ref{Cleaning Up Using gnatclean}, describes @code{gnatclean}, a utility
804 to delete files that are produced by the compiler, binder and linker.
808 @ref{GNAT and Libraries}, describes the process of creating and using
809 Libraries with GNAT. It also describes how to recompile the GNAT run-time
813 @ref{Using the GNU make Utility}, describes some techniques for using
814 the GNAT toolset in Makefiles.
818 @ref{Memory Management Issues}, describes some useful predefined storage pools
819 and in particular the GNAT Debug Pool facility, which helps detect incorrect
822 It also describes @command{gnatmem}, a utility that monitors dynamic
823 allocation and deallocation and helps detect ``memory leaks''.
827 @ref{Stack Related Facilities}, describes some useful tools associated with
828 stack checking and analysis.
831 @ref{Verifying Properties Using gnatcheck}, discusses @code{gnatcheck},
832 a utility that checks Ada code against a set of rules.
835 @ref{Creating Sample Bodies Using gnatstub}, discusses @code{gnatstub},
836 a utility that generates empty but compilable bodies for library units.
839 @ref{Other Utility Programs}, discusses several other GNAT utilities,
840 including @code{gnathtml}.
843 @ref{Running and Debugging Ada Programs}, describes how to run and debug
848 @ref{Compatibility with HP Ada}, details the compatibility of GNAT with
849 HP Ada 83 @footnote{``HP Ada'' refers to the legacy product originally
850 developed by Digital Equipment Corporation and currently supported by HP.}
851 for OpenVMS Alpha. This product was formerly known as DEC Ada,
854 historical compatibility reasons, the relevant libraries still use the
859 @ref{Platform-Specific Information for the Run-Time Libraries},
860 describes the various run-time
861 libraries supported by GNAT on various platforms and explains how to
862 choose a particular library.
865 @ref{Example of Binder Output File}, shows the source code for the binder
866 output file for a sample program.
869 @ref{Elaboration Order Handling in GNAT}, describes how GNAT helps
870 you deal with elaboration order issues.
873 @ref{Inline Assembler}, shows how to use the inline assembly facility
877 @ref{Compatibility and Porting Guide}, contains sections on compatibility
878 of GNAT with other Ada development environments (including Ada 83 systems),
879 to assist in porting code from those environments.
883 @ref{Microsoft Windows Topics}, presents information relevant to the
884 Microsoft Windows platform.
888 @c *************************************************
889 @node What You Should Know before Reading This Guide
890 @c *************************************************
891 @unnumberedsec What You Should Know before Reading This Guide
893 @cindex Ada 95 Language Reference Manual
894 @cindex Ada 2005 Language Reference Manual
896 This guide assumes a basic familiarity with the Ada 95 language, as
897 described in the International Standard ANSI/ISO/IEC-8652:1995, January
899 It does not require knowledge of the new features introduced by Ada 2005,
900 (officially known as ISO/IEC 8652:1995 with Technical Corrigendum 1
902 Both reference manuals are included in the GNAT documentation
905 @node Related Information
906 @unnumberedsec Related Information
909 For further information about related tools, refer to the following
914 @cite{GNAT Reference Manual}, which contains all reference
915 material for the GNAT implementation of Ada.
919 @cite{Using the GNAT Programming Studio}, which describes the GPS
920 Integrated Development Environment.
923 @cite{GNAT Programming Studio Tutorial}, which introduces the
924 main GPS features through examples.
928 @cite{Ada 95 Reference Manual}, which contains reference
929 material for the Ada 95 programming language.
932 @cite{Ada 2005 Reference Manual}, which contains reference
933 material for the Ada 2005 programming language.
936 @cite{Debugging with GDB}
938 , located in the GNU:[DOCS] directory,
940 contains all details on the use of the GNU source-level debugger.
943 @cite{GNU Emacs Manual}
945 , located in the GNU:[DOCS] directory if the EMACS kit is installed,
947 contains full information on the extensible editor and programming
954 @unnumberedsec Conventions
956 @cindex Typographical conventions
959 Following are examples of the typographical and graphic conventions used
964 @code{Functions}, @code{utility program names}, @code{standard names},
971 @file{File Names}, @file{button names}, and @file{field names}.
980 [optional information or parameters]
983 Examples are described by text
985 and then shown this way.
990 Commands that are entered by the user are preceded in this manual by the
991 characters @w{``@code{$ }''} (dollar sign followed by space). If your system
992 uses this sequence as a prompt, then the commands will appear exactly as
993 you see them in the manual. If your system uses some other prompt, then
994 the command will appear with the @code{$} replaced by whatever prompt
995 character you are using.
998 Full file names are shown with the ``@code{/}'' character
999 as the directory separator; e.g., @file{parent-dir/subdir/myfile.adb}.
1000 If you are using GNAT on a Windows platform, please note that
1001 the ``@code{\}'' character should be used instead.
1004 @c ****************************
1005 @node Getting Started with GNAT
1006 @chapter Getting Started with GNAT
1009 This chapter describes some simple ways of using GNAT to build
1010 executable Ada programs.
1012 @ref{Running GNAT}, through @ref{Using the gnatmake Utility},
1013 show how to use the command line environment.
1014 @ref{Introduction to GPS}, provides a brief
1015 introduction to the GNAT Programming Studio, a visually-oriented
1016 Integrated Development Environment for GNAT.
1017 GPS offers a graphical ``look and feel'', support for development in
1018 other programming languages, comprehensive browsing features, and
1019 many other capabilities.
1020 For information on GPS please refer to
1021 @cite{Using the GNAT Programming Studio}.
1026 * Running a Simple Ada Program::
1027 * Running a Program with Multiple Units::
1028 * Using the gnatmake Utility::
1030 * Editing with Emacs::
1033 * Introduction to GPS::
1038 @section Running GNAT
1041 Three steps are needed to create an executable file from an Ada source
1046 The source file(s) must be compiled.
1048 The file(s) must be bound using the GNAT binder.
1050 All appropriate object files must be linked to produce an executable.
1054 All three steps are most commonly handled by using the @command{gnatmake}
1055 utility program that, given the name of the main program, automatically
1056 performs the necessary compilation, binding and linking steps.
1058 @node Running a Simple Ada Program
1059 @section Running a Simple Ada Program
1062 Any text editor may be used to prepare an Ada program.
1064 used, the optional Ada mode may be helpful in laying out the program.)
1066 program text is a normal text file. We will assume in our initial
1067 example that you have used your editor to prepare the following
1068 standard format text file:
1070 @smallexample @c ada
1072 with Ada.Text_IO; use Ada.Text_IO;
1075 Put_Line ("Hello WORLD!");
1081 This file should be named @file{hello.adb}.
1082 With the normal default file naming conventions, GNAT requires
1084 contain a single compilation unit whose file name is the
1086 with periods replaced by hyphens; the
1087 extension is @file{ads} for a
1088 spec and @file{adb} for a body.
1089 You can override this default file naming convention by use of the
1090 special pragma @code{Source_File_Name} (@pxref{Using Other File Names}).
1091 Alternatively, if you want to rename your files according to this default
1092 convention, which is probably more convenient if you will be using GNAT
1093 for all your compilations, then the @code{gnatchop} utility
1094 can be used to generate correctly-named source files
1095 (@pxref{Renaming Files Using gnatchop}).
1097 You can compile the program using the following command (@code{$} is used
1098 as the command prompt in the examples in this document):
1105 @command{gcc} is the command used to run the compiler. This compiler is
1106 capable of compiling programs in several languages, including Ada and
1107 C. It assumes that you have given it an Ada program if the file extension is
1108 either @file{.ads} or @file{.adb}, and it will then call
1109 the GNAT compiler to compile the specified file.
1112 The @option{-c} switch is required. It tells @command{gcc} to only do a
1113 compilation. (For C programs, @command{gcc} can also do linking, but this
1114 capability is not used directly for Ada programs, so the @option{-c}
1115 switch must always be present.)
1118 This compile command generates a file
1119 @file{hello.o}, which is the object
1120 file corresponding to your Ada program. It also generates
1121 an ``Ada Library Information'' file @file{hello.ali},
1122 which contains additional information used to check
1123 that an Ada program is consistent.
1124 To build an executable file,
1125 use @code{gnatbind} to bind the program
1126 and @command{gnatlink} to link it. The
1127 argument to both @code{gnatbind} and @command{gnatlink} is the name of the
1128 @file{ALI} file, but the default extension of @file{.ali} can
1129 be omitted. This means that in the most common case, the argument
1130 is simply the name of the main program:
1138 A simpler method of carrying out these steps is to use
1140 a master program that invokes all the required
1141 compilation, binding and linking tools in the correct order. In particular,
1142 @command{gnatmake} automatically recompiles any sources that have been
1143 modified since they were last compiled, or sources that depend
1144 on such modified sources, so that ``version skew'' is avoided.
1145 @cindex Version skew (avoided by @command{gnatmake})
1148 $ gnatmake hello.adb
1152 The result is an executable program called @file{hello}, which can be
1160 assuming that the current directory is on the search path
1161 for executable programs.
1164 and, if all has gone well, you will see
1171 appear in response to this command.
1173 @c ****************************************
1174 @node Running a Program with Multiple Units
1175 @section Running a Program with Multiple Units
1178 Consider a slightly more complicated example that has three files: a
1179 main program, and the spec and body of a package:
1181 @smallexample @c ada
1184 package Greetings is
1189 with Ada.Text_IO; use Ada.Text_IO;
1190 package body Greetings is
1193 Put_Line ("Hello WORLD!");
1196 procedure Goodbye is
1198 Put_Line ("Goodbye WORLD!");
1215 Following the one-unit-per-file rule, place this program in the
1216 following three separate files:
1220 spec of package @code{Greetings}
1223 body of package @code{Greetings}
1226 body of main program
1230 To build an executable version of
1231 this program, we could use four separate steps to compile, bind, and link
1232 the program, as follows:
1236 $ gcc -c greetings.adb
1242 Note that there is no required order of compilation when using GNAT.
1243 In particular it is perfectly fine to compile the main program first.
1244 Also, it is not necessary to compile package specs in the case where
1245 there is an accompanying body; you only need to compile the body. If you want
1246 to submit these files to the compiler for semantic checking and not code
1247 generation, then use the
1248 @option{-gnatc} switch:
1251 $ gcc -c greetings.ads -gnatc
1255 Although the compilation can be done in separate steps as in the
1256 above example, in practice it is almost always more convenient
1257 to use the @command{gnatmake} tool. All you need to know in this case
1258 is the name of the main program's source file. The effect of the above four
1259 commands can be achieved with a single one:
1262 $ gnatmake gmain.adb
1266 In the next section we discuss the advantages of using @command{gnatmake} in
1269 @c *****************************
1270 @node Using the gnatmake Utility
1271 @section Using the @command{gnatmake} Utility
1274 If you work on a program by compiling single components at a time using
1275 @command{gcc}, you typically keep track of the units you modify. In order to
1276 build a consistent system, you compile not only these units, but also any
1277 units that depend on the units you have modified.
1278 For example, in the preceding case,
1279 if you edit @file{gmain.adb}, you only need to recompile that file. But if
1280 you edit @file{greetings.ads}, you must recompile both
1281 @file{greetings.adb} and @file{gmain.adb}, because both files contain
1282 units that depend on @file{greetings.ads}.
1284 @code{gnatbind} will warn you if you forget one of these compilation
1285 steps, so that it is impossible to generate an inconsistent program as a
1286 result of forgetting to do a compilation. Nevertheless it is tedious and
1287 error-prone to keep track of dependencies among units.
1288 One approach to handle the dependency-bookkeeping is to use a
1289 makefile. However, makefiles present maintenance problems of their own:
1290 if the dependencies change as you change the program, you must make
1291 sure that the makefile is kept up-to-date manually, which is also an
1292 error-prone process.
1294 The @command{gnatmake} utility takes care of these details automatically.
1295 Invoke it using either one of the following forms:
1298 $ gnatmake gmain.adb
1299 $ gnatmake ^gmain^GMAIN^
1303 The argument is the name of the file containing the main program;
1304 you may omit the extension. @command{gnatmake}
1305 examines the environment, automatically recompiles any files that need
1306 recompiling, and binds and links the resulting set of object files,
1307 generating the executable file, @file{^gmain^GMAIN.EXE^}.
1308 In a large program, it
1309 can be extremely helpful to use @command{gnatmake}, because working out by hand
1310 what needs to be recompiled can be difficult.
1312 Note that @command{gnatmake}
1313 takes into account all the Ada rules that
1314 establish dependencies among units. These include dependencies that result
1315 from inlining subprogram bodies, and from
1316 generic instantiation. Unlike some other
1317 Ada make tools, @command{gnatmake} does not rely on the dependencies that were
1318 found by the compiler on a previous compilation, which may possibly
1319 be wrong when sources change. @command{gnatmake} determines the exact set of
1320 dependencies from scratch each time it is run.
1323 @node Editing with Emacs
1324 @section Editing with Emacs
1328 Emacs is an extensible self-documenting text editor that is available in a
1329 separate VMSINSTAL kit.
1331 Invoke Emacs by typing @kbd{Emacs} at the command prompt. To get started,
1332 click on the Emacs Help menu and run the Emacs Tutorial.
1333 In a character cell terminal, Emacs help is invoked with @kbd{Ctrl-h} (also
1334 written as @kbd{C-h}), and the tutorial by @kbd{C-h t}.
1336 Documentation on Emacs and other tools is available in Emacs under the
1337 pull-down menu button: @code{Help - Info}. After selecting @code{Info},
1338 use the middle mouse button to select a topic (e.g. Emacs).
1340 In a character cell terminal, do @kbd{C-h i} to invoke info, and then @kbd{m}
1341 (stands for menu) followed by the menu item desired, as in @kbd{m Emacs}, to
1342 get to the Emacs manual.
1343 Help on Emacs is also available by typing @kbd{HELP EMACS} at the DCL command
1346 The tutorial is highly recommended in order to learn the intricacies of Emacs,
1347 which is sufficiently extensible to provide for a complete programming
1348 environment and shell for the sophisticated user.
1352 @node Introduction to GPS
1353 @section Introduction to GPS
1354 @cindex GPS (GNAT Programming Studio)
1355 @cindex GNAT Programming Studio (GPS)
1357 Although the command line interface (@command{gnatmake}, etc.) alone
1358 is sufficient, a graphical Interactive Development
1359 Environment can make it easier for you to compose, navigate, and debug
1360 programs. This section describes the main features of GPS
1361 (``GNAT Programming Studio''), the GNAT graphical IDE.
1362 You will see how to use GPS to build and debug an executable, and
1363 you will also learn some of the basics of the GNAT ``project'' facility.
1365 GPS enables you to do much more than is presented here;
1366 e.g., you can produce a call graph, interface to a third-party
1367 Version Control System, and inspect the generated assembly language
1369 Indeed, GPS also supports languages other than Ada.
1370 Such additional information, and an explanation of all of the GPS menu
1371 items. may be found in the on-line help, which includes
1372 a user's guide and a tutorial (these are also accessible from the GNAT
1376 * Building a New Program with GPS::
1377 * Simple Debugging with GPS::
1380 @node Building a New Program with GPS
1381 @subsection Building a New Program with GPS
1383 GPS invokes the GNAT compilation tools using information
1384 contained in a @emph{project} (also known as a @emph{project file}):
1385 a collection of properties such
1386 as source directories, identities of main subprograms, tool switches, etc.,
1387 and their associated values.
1388 See @ref{GNAT Project Manager} for details.
1389 In order to run GPS, you will need to either create a new project
1390 or else open an existing one.
1392 This section will explain how you can use GPS to create a project,
1393 to associate Ada source files with a project, and to build and run
1397 @item @emph{Creating a project}
1399 Invoke GPS, either from the command line or the platform's IDE.
1400 After it starts, GPS will display a ``Welcome'' screen with three
1405 @code{Start with default project in directory}
1408 @code{Create new project with wizard}
1411 @code{Open existing project}
1415 Select @code{Create new project with wizard} and press @code{OK}.
1416 A new window will appear. In the text box labeled with
1417 @code{Enter the name of the project to create}, type @file{sample}
1418 as the project name.
1419 In the next box, browse to choose the directory in which you
1420 would like to create the project file.
1421 After selecting an appropriate directory, press @code{Forward}.
1423 A window will appear with the title
1424 @code{Version Control System Configuration}.
1425 Simply press @code{Forward}.
1427 A window will appear with the title
1428 @code{Please select the source directories for this project}.
1429 The directory that you specified for the project file will be selected
1430 by default as the one to use for sources; simply press @code{Forward}.
1432 A window will appear with the title
1433 @code{Please select the build directory for this project}.
1434 The directory that you specified for the project file will be selected
1435 by default for object files and executables;
1436 simply press @code{Forward}.
1438 A window will appear with the title
1439 @code{Please select the main units for this project}.
1440 You will supply this information later, after creating the source file.
1441 Simply press @code{Forward} for now.
1443 A window will appear with the title
1444 @code{Please select the switches to build the project}.
1445 Press @code{Apply}. This will create a project file named
1446 @file{sample.prj} in the directory that you had specified.
1448 @item @emph{Creating and saving the source file}
1450 After you create the new project, a GPS window will appear, which is
1451 partitioned into two main sections:
1455 A @emph{Workspace area}, initially greyed out, which you will use for
1456 creating and editing source files
1459 Directly below, a @emph{Messages area}, which initially displays a
1460 ``Welcome'' message.
1461 (If the Messages area is not visible, drag its border upward to expand it.)
1465 Select @code{File} on the menu bar, and then the @code{New} command.
1466 The Workspace area will become white, and you can now
1467 enter the source program explicitly.
1468 Type the following text
1470 @smallexample @c ada
1472 with Ada.Text_IO; use Ada.Text_IO;
1475 Put_Line("Hello from GPS!");
1481 Select @code{File}, then @code{Save As}, and enter the source file name
1483 The file will be saved in the same directory you specified as the
1484 location of the default project file.
1486 @item @emph{Updating the project file}
1488 You need to add the new source file to the project.
1490 the @code{Project} menu and then @code{Edit project properties}.
1491 Click the @code{Main files} tab on the left, and then the
1493 Choose @file{hello.adb} from the list, and press @code{Open}.
1494 The project settings window will reflect this action.
1497 @item @emph{Building and running the program}
1499 In the main GPS window, now choose the @code{Build} menu, then @code{Make},
1500 and select @file{hello.adb}.
1501 The Messages window will display the resulting invocations of @command{gcc},
1502 @command{gnatbind}, and @command{gnatlink}
1503 (reflecting the default switch settings from the
1504 project file that you created) and then a ``successful compilation/build''
1507 To run the program, choose the @code{Build} menu, then @code{Run}, and
1508 select @command{hello}.
1509 An @emph{Arguments Selection} window will appear.
1510 There are no command line arguments, so just click @code{OK}.
1512 The Messages window will now display the program's output (the string
1513 @code{Hello from GPS}), and at the bottom of the GPS window a status
1514 update is displayed (@code{Run: hello}).
1515 Close the GPS window (or select @code{File}, then @code{Exit}) to
1516 terminate this GPS session.
1519 @node Simple Debugging with GPS
1520 @subsection Simple Debugging with GPS
1522 This section illustrates basic debugging techniques (setting breakpoints,
1523 examining/modifying variables, single stepping).
1526 @item @emph{Opening a project}
1528 Start GPS and select @code{Open existing project}; browse to
1529 specify the project file @file{sample.prj} that you had created in the
1532 @item @emph{Creating a source file}
1534 Select @code{File}, then @code{New}, and type in the following program:
1536 @smallexample @c ada
1538 with Ada.Text_IO; use Ada.Text_IO;
1539 procedure Example is
1540 Line : String (1..80);
1543 Put_Line("Type a line of text at each prompt; an empty line to exit");
1547 Put_Line (Line (1..N) );
1555 Select @code{File}, then @code{Save as}, and enter the file name
1558 @item @emph{Updating the project file}
1560 Add @code{Example} as a new main unit for the project:
1563 Select @code{Project}, then @code{Edit Project Properties}.
1566 Select the @code{Main files} tab, click @code{Add}, then
1567 select the file @file{example.adb} from the list, and
1569 You will see the file name appear in the list of main units
1575 @item @emph{Building/running the executable}
1577 To build the executable
1578 select @code{Build}, then @code{Make}, and then choose @file{example.adb}.
1580 Run the program to see its effect (in the Messages area).
1581 Each line that you enter is displayed; an empty line will
1582 cause the loop to exit and the program to terminate.
1584 @item @emph{Debugging the program}
1586 Note that the @option{-g} switches to @command{gcc} and @command{gnatlink},
1587 which are required for debugging, are on by default when you create
1589 Thus unless you intentionally remove these settings, you will be able
1590 to debug any program that you develop using GPS.
1593 @item @emph{Initializing}
1595 Select @code{Debug}, then @code{Initialize}, then @file{example}
1597 @item @emph{Setting a breakpoint}
1599 After performing the initialization step, you will observe a small
1600 icon to the right of each line number.
1601 This serves as a toggle for breakpoints; clicking the icon will
1602 set a breakpoint at the corresponding line (the icon will change to
1603 a red circle with an ``x''), and clicking it again
1604 will remove the breakpoint / reset the icon.
1606 For purposes of this example, set a breakpoint at line 10 (the
1607 statement @code{Put_Line@ (Line@ (1..N));}
1609 @item @emph{Starting program execution}
1611 Select @code{Debug}, then @code{Run}. When the
1612 @code{Program Arguments} window appears, click @code{OK}.
1613 A console window will appear; enter some line of text,
1614 e.g. @code{abcde}, at the prompt.
1615 The program will pause execution when it gets to the
1616 breakpoint, and the corresponding line is highlighted.
1618 @item @emph{Examining a variable}
1620 Move the mouse over one of the occurrences of the variable @code{N}.
1621 You will see the value (5) displayed, in ``tool tip'' fashion.
1622 Right click on @code{N}, select @code{Debug}, then select @code{Display N}.
1623 You will see information about @code{N} appear in the @code{Debugger Data}
1624 pane, showing the value as 5.
1626 @item @emph{Assigning a new value to a variable}
1628 Right click on the @code{N} in the @code{Debugger Data} pane, and
1629 select @code{Set value of N}.
1630 When the input window appears, enter the value @code{4} and click
1632 This value does not automatically appear in the @code{Debugger Data}
1633 pane; to see it, right click again on the @code{N} in the
1634 @code{Debugger Data} pane and select @code{Update value}.
1635 The new value, 4, will appear in red.
1637 @item @emph{Single stepping}
1639 Select @code{Debug}, then @code{Next}.
1640 This will cause the next statement to be executed, in this case the
1641 call of @code{Put_Line} with the string slice.
1642 Notice in the console window that the displayed string is simply
1643 @code{abcd} and not @code{abcde} which you had entered.
1644 This is because the upper bound of the slice is now 4 rather than 5.
1646 @item @emph{Removing a breakpoint}
1648 Toggle the breakpoint icon at line 10.
1650 @item @emph{Resuming execution from a breakpoint}
1652 Select @code{Debug}, then @code{Continue}.
1653 The program will reach the next iteration of the loop, and
1654 wait for input after displaying the prompt.
1655 This time, just hit the @kbd{Enter} key.
1656 The value of @code{N} will be 0, and the program will terminate.
1657 The console window will disappear.
1662 @node The GNAT Compilation Model
1663 @chapter The GNAT Compilation Model
1664 @cindex GNAT compilation model
1665 @cindex Compilation model
1668 * Source Representation::
1669 * Foreign Language Representation::
1670 * File Naming Rules::
1671 * Using Other File Names::
1672 * Alternative File Naming Schemes::
1673 * Generating Object Files::
1674 * Source Dependencies::
1675 * The Ada Library Information Files::
1676 * Binding an Ada Program::
1677 * Mixed Language Programming::
1679 * Building Mixed Ada & C++ Programs::
1680 * Comparison between GNAT and C/C++ Compilation Models::
1682 * Comparison between GNAT and Conventional Ada Library Models::
1684 * Placement of temporary files::
1689 This chapter describes the compilation model used by GNAT. Although
1690 similar to that used by other languages, such as C and C++, this model
1691 is substantially different from the traditional Ada compilation models,
1692 which are based on a library. The model is initially described without
1693 reference to the library-based model. If you have not previously used an
1694 Ada compiler, you need only read the first part of this chapter. The
1695 last section describes and discusses the differences between the GNAT
1696 model and the traditional Ada compiler models. If you have used other
1697 Ada compilers, this section will help you to understand those
1698 differences, and the advantages of the GNAT model.
1700 @node Source Representation
1701 @section Source Representation
1705 Ada source programs are represented in standard text files, using
1706 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1707 7-bit ASCII set, plus additional characters used for
1708 representing foreign languages (@pxref{Foreign Language Representation}
1709 for support of non-USA character sets). The format effector characters
1710 are represented using their standard ASCII encodings, as follows:
1715 Vertical tab, @code{16#0B#}
1719 Horizontal tab, @code{16#09#}
1723 Carriage return, @code{16#0D#}
1727 Line feed, @code{16#0A#}
1731 Form feed, @code{16#0C#}
1735 Source files are in standard text file format. In addition, GNAT will
1736 recognize a wide variety of stream formats, in which the end of
1737 physical lines is marked by any of the following sequences:
1738 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1739 in accommodating files that are imported from other operating systems.
1741 @cindex End of source file
1742 @cindex Source file, end
1744 The end of a source file is normally represented by the physical end of
1745 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1746 recognized as signalling the end of the source file. Again, this is
1747 provided for compatibility with other operating systems where this
1748 code is used to represent the end of file.
1750 Each file contains a single Ada compilation unit, including any pragmas
1751 associated with the unit. For example, this means you must place a
1752 package declaration (a package @dfn{spec}) and the corresponding body in
1753 separate files. An Ada @dfn{compilation} (which is a sequence of
1754 compilation units) is represented using a sequence of files. Similarly,
1755 you will place each subunit or child unit in a separate file.
1757 @node Foreign Language Representation
1758 @section Foreign Language Representation
1761 GNAT supports the standard character sets defined in Ada as well as
1762 several other non-standard character sets for use in localized versions
1763 of the compiler (@pxref{Character Set Control}).
1766 * Other 8-Bit Codes::
1767 * Wide Character Encodings::
1775 The basic character set is Latin-1. This character set is defined by ISO
1776 standard 8859, part 1. The lower half (character codes @code{16#00#}
1777 ... @code{16#7F#)} is identical to standard ASCII coding, but the upper half
1778 is used to represent additional characters. These include extended letters
1779 used by European languages, such as French accents, the vowels with umlauts
1780 used in German, and the extra letter A-ring used in Swedish.
1782 @findex Ada.Characters.Latin_1
1783 For a complete list of Latin-1 codes and their encodings, see the source
1784 file of library unit @code{Ada.Characters.Latin_1} in file
1785 @file{a-chlat1.ads}.
1786 You may use any of these extended characters freely in character or
1787 string literals. In addition, the extended characters that represent
1788 letters can be used in identifiers.
1790 @node Other 8-Bit Codes
1791 @subsection Other 8-Bit Codes
1794 GNAT also supports several other 8-bit coding schemes:
1797 @item ISO 8859-2 (Latin-2)
1800 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1803 @item ISO 8859-3 (Latin-3)
1806 Latin-3 letters allowed in identifiers, with uppercase and lowercase
1809 @item ISO 8859-4 (Latin-4)
1812 Latin-4 letters allowed in identifiers, with uppercase and lowercase
1815 @item ISO 8859-5 (Cyrillic)
1818 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
1819 lowercase equivalence.
1821 @item ISO 8859-15 (Latin-9)
1824 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
1825 lowercase equivalence
1827 @item IBM PC (code page 437)
1828 @cindex code page 437
1829 This code page is the normal default for PCs in the U.S. It corresponds
1830 to the original IBM PC character set. This set has some, but not all, of
1831 the extended Latin-1 letters, but these letters do not have the same
1832 encoding as Latin-1. In this mode, these letters are allowed in
1833 identifiers with uppercase and lowercase equivalence.
1835 @item IBM PC (code page 850)
1836 @cindex code page 850
1837 This code page is a modification of 437 extended to include all the
1838 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
1839 mode, all these letters are allowed in identifiers with uppercase and
1840 lowercase equivalence.
1842 @item Full Upper 8-bit
1843 Any character in the range 80-FF allowed in identifiers, and all are
1844 considered distinct. In other words, there are no uppercase and lowercase
1845 equivalences in this range. This is useful in conjunction with
1846 certain encoding schemes used for some foreign character sets (e.g.
1847 the typical method of representing Chinese characters on the PC).
1850 No upper-half characters in the range 80-FF are allowed in identifiers.
1851 This gives Ada 83 compatibility for identifier names.
1855 For precise data on the encodings permitted, and the uppercase and lowercase
1856 equivalences that are recognized, see the file @file{csets.adb} in
1857 the GNAT compiler sources. You will need to obtain a full source release
1858 of GNAT to obtain this file.
1860 @node Wide Character Encodings
1861 @subsection Wide Character Encodings
1864 GNAT allows wide character codes to appear in character and string
1865 literals, and also optionally in identifiers, by means of the following
1866 possible encoding schemes:
1871 In this encoding, a wide character is represented by the following five
1879 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1880 characters (using uppercase letters) of the wide character code. For
1881 example, ESC A345 is used to represent the wide character with code
1883 This scheme is compatible with use of the full Wide_Character set.
1885 @item Upper-Half Coding
1886 @cindex Upper-Half Coding
1887 The wide character with encoding @code{16#abcd#} where the upper bit is on
1888 (in other words, ``a'' is in the range 8-F) is represented as two bytes,
1889 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
1890 character, but is not required to be in the upper half. This method can
1891 be also used for shift-JIS or EUC, where the internal coding matches the
1894 @item Shift JIS Coding
1895 @cindex Shift JIS Coding
1896 A wide character is represented by a two-character sequence,
1898 @code{16#cd#}, with the restrictions described for upper-half encoding as
1899 described above. The internal character code is the corresponding JIS
1900 character according to the standard algorithm for Shift-JIS
1901 conversion. Only characters defined in the JIS code set table can be
1902 used with this encoding method.
1906 A wide character is represented by a two-character sequence
1908 @code{16#cd#}, with both characters being in the upper half. The internal
1909 character code is the corresponding JIS character according to the EUC
1910 encoding algorithm. Only characters defined in the JIS code set table
1911 can be used with this encoding method.
1914 A wide character is represented using
1915 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
1916 10646-1/Am.2. Depending on the character value, the representation
1917 is a one, two, or three byte sequence:
1922 16#0000#-16#007f#: 2#0xxxxxxx#
1923 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
1924 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
1929 where the xxx bits correspond to the left-padded bits of the
1930 16-bit character value. Note that all lower half ASCII characters
1931 are represented as ASCII bytes and all upper half characters and
1932 other wide characters are represented as sequences of upper-half
1933 (The full UTF-8 scheme allows for encoding 31-bit characters as
1934 6-byte sequences, but in this implementation, all UTF-8 sequences
1935 of four or more bytes length will be treated as illegal).
1936 @item Brackets Coding
1937 In this encoding, a wide character is represented by the following eight
1945 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
1946 characters (using uppercase letters) of the wide character code. For
1947 example, [``A345''] is used to represent the wide character with code
1948 @code{16#A345#}. It is also possible (though not required) to use the
1949 Brackets coding for upper half characters. For example, the code
1950 @code{16#A3#} can be represented as @code{[``A3'']}.
1952 This scheme is compatible with use of the full Wide_Character set,
1953 and is also the method used for wide character encoding in the standard
1954 ACVC (Ada Compiler Validation Capability) test suite distributions.
1959 Note: Some of these coding schemes do not permit the full use of the
1960 Ada character set. For example, neither Shift JIS, nor EUC allow the
1961 use of the upper half of the Latin-1 set.
1963 @node File Naming Rules
1964 @section File Naming Rules
1967 The default file name is determined by the name of the unit that the
1968 file contains. The name is formed by taking the full expanded name of
1969 the unit and replacing the separating dots with hyphens and using
1970 ^lowercase^uppercase^ for all letters.
1972 An exception arises if the file name generated by the above rules starts
1973 with one of the characters
1980 and the second character is a
1981 minus. In this case, the character ^tilde^dollar sign^ is used in place
1982 of the minus. The reason for this special rule is to avoid clashes with
1983 the standard names for child units of the packages System, Ada,
1984 Interfaces, and GNAT, which use the prefixes
1993 The file extension is @file{.ads} for a spec and
1994 @file{.adb} for a body. The following list shows some
1995 examples of these rules.
2002 @item arith_functions.ads
2003 Arith_Functions (package spec)
2004 @item arith_functions.adb
2005 Arith_Functions (package body)
2007 Func.Spec (child package spec)
2009 Func.Spec (child package body)
2011 Sub (subunit of Main)
2012 @item ^a~bad.adb^A$BAD.ADB^
2013 A.Bad (child package body)
2017 Following these rules can result in excessively long
2018 file names if corresponding
2019 unit names are long (for example, if child units or subunits are
2020 heavily nested). An option is available to shorten such long file names
2021 (called file name ``krunching''). This may be particularly useful when
2022 programs being developed with GNAT are to be used on operating systems
2023 with limited file name lengths. @xref{Using gnatkr}.
2025 Of course, no file shortening algorithm can guarantee uniqueness over
2026 all possible unit names; if file name krunching is used, it is your
2027 responsibility to ensure no name clashes occur. Alternatively you
2028 can specify the exact file names that you want used, as described
2029 in the next section. Finally, if your Ada programs are migrating from a
2030 compiler with a different naming convention, you can use the gnatchop
2031 utility to produce source files that follow the GNAT naming conventions.
2032 (For details @pxref{Renaming Files Using gnatchop}.)
2034 Note: in the case of @code{Windows NT/XP} or @code{OpenVMS} operating
2035 systems, case is not significant. So for example on @code{Windows XP}
2036 if the canonical name is @code{main-sub.adb}, you can use the file name
2037 @code{Main-Sub.adb} instead. However, case is significant for other
2038 operating systems, so for example, if you want to use other than
2039 canonically cased file names on a Unix system, you need to follow
2040 the procedures described in the next section.
2042 @node Using Other File Names
2043 @section Using Other File Names
2047 In the previous section, we have described the default rules used by
2048 GNAT to determine the file name in which a given unit resides. It is
2049 often convenient to follow these default rules, and if you follow them,
2050 the compiler knows without being explicitly told where to find all
2053 However, in some cases, particularly when a program is imported from
2054 another Ada compiler environment, it may be more convenient for the
2055 programmer to specify which file names contain which units. GNAT allows
2056 arbitrary file names to be used by means of the Source_File_Name pragma.
2057 The form of this pragma is as shown in the following examples:
2058 @cindex Source_File_Name pragma
2060 @smallexample @c ada
2062 pragma Source_File_Name (My_Utilities.Stacks,
2063 Spec_File_Name => "myutilst_a.ada");
2064 pragma Source_File_name (My_Utilities.Stacks,
2065 Body_File_Name => "myutilst.ada");
2070 As shown in this example, the first argument for the pragma is the unit
2071 name (in this example a child unit). The second argument has the form
2072 of a named association. The identifier
2073 indicates whether the file name is for a spec or a body;
2074 the file name itself is given by a string literal.
2076 The source file name pragma is a configuration pragma, which means that
2077 normally it will be placed in the @file{gnat.adc}
2078 file used to hold configuration
2079 pragmas that apply to a complete compilation environment.
2080 For more details on how the @file{gnat.adc} file is created and used
2081 see @ref{Handling of Configuration Pragmas}.
2082 @cindex @file{gnat.adc}
2085 GNAT allows completely arbitrary file names to be specified using the
2086 source file name pragma. However, if the file name specified has an
2087 extension other than @file{.ads} or @file{.adb} it is necessary to use
2088 a special syntax when compiling the file. The name in this case must be
2089 preceded by the special sequence @code{-x} followed by a space and the name
2090 of the language, here @code{ada}, as in:
2093 $ gcc -c -x ada peculiar_file_name.sim
2098 @command{gnatmake} handles non-standard file names in the usual manner (the
2099 non-standard file name for the main program is simply used as the
2100 argument to gnatmake). Note that if the extension is also non-standard,
2101 then it must be included in the gnatmake command, it may not be omitted.
2103 @node Alternative File Naming Schemes
2104 @section Alternative File Naming Schemes
2105 @cindex File naming schemes, alternative
2108 In the previous section, we described the use of the @code{Source_File_Name}
2109 pragma to allow arbitrary names to be assigned to individual source files.
2110 However, this approach requires one pragma for each file, and especially in
2111 large systems can result in very long @file{gnat.adc} files, and also create
2112 a maintenance problem.
2114 GNAT also provides a facility for specifying systematic file naming schemes
2115 other than the standard default naming scheme previously described. An
2116 alternative scheme for naming is specified by the use of
2117 @code{Source_File_Name} pragmas having the following format:
2118 @cindex Source_File_Name pragma
2120 @smallexample @c ada
2121 pragma Source_File_Name (
2122 Spec_File_Name => FILE_NAME_PATTERN
2123 [,Casing => CASING_SPEC]
2124 [,Dot_Replacement => STRING_LITERAL]);
2126 pragma Source_File_Name (
2127 Body_File_Name => FILE_NAME_PATTERN
2128 [,Casing => CASING_SPEC]
2129 [,Dot_Replacement => STRING_LITERAL]);
2131 pragma Source_File_Name (
2132 Subunit_File_Name => FILE_NAME_PATTERN
2133 [,Casing => CASING_SPEC]
2134 [,Dot_Replacement => STRING_LITERAL]);
2136 FILE_NAME_PATTERN ::= STRING_LITERAL
2137 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
2141 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
2142 It contains a single asterisk character, and the unit name is substituted
2143 systematically for this asterisk. The optional parameter
2144 @code{Casing} indicates
2145 whether the unit name is to be all upper-case letters, all lower-case letters,
2146 or mixed-case. If no
2147 @code{Casing} parameter is used, then the default is all
2148 ^lower-case^upper-case^.
2150 The optional @code{Dot_Replacement} string is used to replace any periods
2151 that occur in subunit or child unit names. If no @code{Dot_Replacement}
2152 argument is used then separating dots appear unchanged in the resulting
2154 Although the above syntax indicates that the
2155 @code{Casing} argument must appear
2156 before the @code{Dot_Replacement} argument, but it
2157 is also permissible to write these arguments in the opposite order.
2159 As indicated, it is possible to specify different naming schemes for
2160 bodies, specs, and subunits. Quite often the rule for subunits is the
2161 same as the rule for bodies, in which case, there is no need to give
2162 a separate @code{Subunit_File_Name} rule, and in this case the
2163 @code{Body_File_name} rule is used for subunits as well.
2165 The separate rule for subunits can also be used to implement the rather
2166 unusual case of a compilation environment (e.g. a single directory) which
2167 contains a subunit and a child unit with the same unit name. Although
2168 both units cannot appear in the same partition, the Ada Reference Manual
2169 allows (but does not require) the possibility of the two units coexisting
2170 in the same environment.
2172 The file name translation works in the following steps:
2177 If there is a specific @code{Source_File_Name} pragma for the given unit,
2178 then this is always used, and any general pattern rules are ignored.
2181 If there is a pattern type @code{Source_File_Name} pragma that applies to
2182 the unit, then the resulting file name will be used if the file exists. If
2183 more than one pattern matches, the latest one will be tried first, and the
2184 first attempt resulting in a reference to a file that exists will be used.
2187 If no pattern type @code{Source_File_Name} pragma that applies to the unit
2188 for which the corresponding file exists, then the standard GNAT default
2189 naming rules are used.
2194 As an example of the use of this mechanism, consider a commonly used scheme
2195 in which file names are all lower case, with separating periods copied
2196 unchanged to the resulting file name, and specs end with @file{.1.ada}, and
2197 bodies end with @file{.2.ada}. GNAT will follow this scheme if the following
2200 @smallexample @c ada
2201 pragma Source_File_Name
2202 (Spec_File_Name => "*.1.ada");
2203 pragma Source_File_Name
2204 (Body_File_Name => "*.2.ada");
2208 The default GNAT scheme is actually implemented by providing the following
2209 default pragmas internally:
2211 @smallexample @c ada
2212 pragma Source_File_Name
2213 (Spec_File_Name => "*.ads", Dot_Replacement => "-");
2214 pragma Source_File_Name
2215 (Body_File_Name => "*.adb", Dot_Replacement => "-");
2219 Our final example implements a scheme typically used with one of the
2220 Ada 83 compilers, where the separator character for subunits was ``__''
2221 (two underscores), specs were identified by adding @file{_.ADA}, bodies
2222 by adding @file{.ADA}, and subunits by
2223 adding @file{.SEP}. All file names were
2224 upper case. Child units were not present of course since this was an
2225 Ada 83 compiler, but it seems reasonable to extend this scheme to use
2226 the same double underscore separator for child units.
2228 @smallexample @c ada
2229 pragma Source_File_Name
2230 (Spec_File_Name => "*_.ADA",
2231 Dot_Replacement => "__",
2232 Casing = Uppercase);
2233 pragma Source_File_Name
2234 (Body_File_Name => "*.ADA",
2235 Dot_Replacement => "__",
2236 Casing = Uppercase);
2237 pragma Source_File_Name
2238 (Subunit_File_Name => "*.SEP",
2239 Dot_Replacement => "__",
2240 Casing = Uppercase);
2243 @node Generating Object Files
2244 @section Generating Object Files
2247 An Ada program consists of a set of source files, and the first step in
2248 compiling the program is to generate the corresponding object files.
2249 These are generated by compiling a subset of these source files.
2250 The files you need to compile are the following:
2254 If a package spec has no body, compile the package spec to produce the
2255 object file for the package.
2258 If a package has both a spec and a body, compile the body to produce the
2259 object file for the package. The source file for the package spec need
2260 not be compiled in this case because there is only one object file, which
2261 contains the code for both the spec and body of the package.
2264 For a subprogram, compile the subprogram body to produce the object file
2265 for the subprogram. The spec, if one is present, is as usual in a
2266 separate file, and need not be compiled.
2270 In the case of subunits, only compile the parent unit. A single object
2271 file is generated for the entire subunit tree, which includes all the
2275 Compile child units independently of their parent units
2276 (though, of course, the spec of all the ancestor unit must be present in order
2277 to compile a child unit).
2281 Compile generic units in the same manner as any other units. The object
2282 files in this case are small dummy files that contain at most the
2283 flag used for elaboration checking. This is because GNAT always handles generic
2284 instantiation by means of macro expansion. However, it is still necessary to
2285 compile generic units, for dependency checking and elaboration purposes.
2289 The preceding rules describe the set of files that must be compiled to
2290 generate the object files for a program. Each object file has the same
2291 name as the corresponding source file, except that the extension is
2294 You may wish to compile other files for the purpose of checking their
2295 syntactic and semantic correctness. For example, in the case where a
2296 package has a separate spec and body, you would not normally compile the
2297 spec. However, it is convenient in practice to compile the spec to make
2298 sure it is error-free before compiling clients of this spec, because such
2299 compilations will fail if there is an error in the spec.
2301 GNAT provides an option for compiling such files purely for the
2302 purposes of checking correctness; such compilations are not required as
2303 part of the process of building a program. To compile a file in this
2304 checking mode, use the @option{-gnatc} switch.
2306 @node Source Dependencies
2307 @section Source Dependencies
2310 A given object file clearly depends on the source file which is compiled
2311 to produce it. Here we are using @dfn{depends} in the sense of a typical
2312 @code{make} utility; in other words, an object file depends on a source
2313 file if changes to the source file require the object file to be
2315 In addition to this basic dependency, a given object may depend on
2316 additional source files as follows:
2320 If a file being compiled @code{with}'s a unit @var{X}, the object file
2321 depends on the file containing the spec of unit @var{X}. This includes
2322 files that are @code{with}'ed implicitly either because they are parents
2323 of @code{with}'ed child units or they are run-time units required by the
2324 language constructs used in a particular unit.
2327 If a file being compiled instantiates a library level generic unit, the
2328 object file depends on both the spec and body files for this generic
2332 If a file being compiled instantiates a generic unit defined within a
2333 package, the object file depends on the body file for the package as
2334 well as the spec file.
2338 @cindex @option{-gnatn} switch
2339 If a file being compiled contains a call to a subprogram for which
2340 pragma @code{Inline} applies and inlining is activated with the
2341 @option{-gnatn} switch, the object file depends on the file containing the
2342 body of this subprogram as well as on the file containing the spec. Note
2343 that for inlining to actually occur as a result of the use of this switch,
2344 it is necessary to compile in optimizing mode.
2346 @cindex @option{-gnatN} switch
2347 The use of @option{-gnatN} activates a more extensive inlining optimization
2348 that is performed by the front end of the compiler. This inlining does
2349 not require that the code generation be optimized. Like @option{-gnatn},
2350 the use of this switch generates additional dependencies.
2352 @option{-gnatN} automatically implies @option{-gnatn} so it is not necessary
2353 to specify both options.
2356 If an object file @file{O} depends on the proper body of a subunit through
2357 inlining or instantiation, it depends on the parent unit of the subunit.
2358 This means that any modification of the parent unit or one of its subunits
2359 affects the compilation of @file{O}.
2362 The object file for a parent unit depends on all its subunit body files.
2365 The previous two rules meant that for purposes of computing dependencies and
2366 recompilation, a body and all its subunits are treated as an indivisible whole.
2369 These rules are applied transitively: if unit @code{A} @code{with}'s
2370 unit @code{B}, whose elaboration calls an inlined procedure in package
2371 @code{C}, the object file for unit @code{A} will depend on the body of
2372 @code{C}, in file @file{c.adb}.
2374 The set of dependent files described by these rules includes all the
2375 files on which the unit is semantically dependent, as dictated by the
2376 Ada language standard. However, it is a superset of what the
2377 standard describes, because it includes generic, inline, and subunit
2380 An object file must be recreated by recompiling the corresponding source
2381 file if any of the source files on which it depends are modified. For
2382 example, if the @code{make} utility is used to control compilation,
2383 the rule for an Ada object file must mention all the source files on
2384 which the object file depends, according to the above definition.
2385 The determination of the necessary
2386 recompilations is done automatically when one uses @command{gnatmake}.
2389 @node The Ada Library Information Files
2390 @section The Ada Library Information Files
2391 @cindex Ada Library Information files
2392 @cindex @file{ALI} files
2395 Each compilation actually generates two output files. The first of these
2396 is the normal object file that has a @file{.o} extension. The second is a
2397 text file containing full dependency information. It has the same
2398 name as the source file, but an @file{.ali} extension.
2399 This file is known as the Ada Library Information (@file{ALI}) file.
2400 The following information is contained in the @file{ALI} file.
2404 Version information (indicates which version of GNAT was used to compile
2405 the unit(s) in question)
2408 Main program information (including priority and time slice settings,
2409 as well as the wide character encoding used during compilation).
2412 List of arguments used in the @command{gcc} command for the compilation
2415 Attributes of the unit, including configuration pragmas used, an indication
2416 of whether the compilation was successful, exception model used etc.
2419 A list of relevant restrictions applying to the unit (used for consistency)
2423 Categorization information (e.g. use of pragma @code{Pure}).
2426 Information on all @code{with}'ed units, including presence of
2427 @code{Elaborate} or @code{Elaborate_All} pragmas.
2430 Information from any @code{Linker_Options} pragmas used in the unit
2433 Information on the use of @code{Body_Version} or @code{Version}
2434 attributes in the unit.
2437 Dependency information. This is a list of files, together with
2438 time stamp and checksum information. These are files on which
2439 the unit depends in the sense that recompilation is required
2440 if any of these units are modified.
2443 Cross-reference data. Contains information on all entities referenced
2444 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
2445 provide cross-reference information.
2450 For a full detailed description of the format of the @file{ALI} file,
2451 see the source of the body of unit @code{Lib.Writ}, contained in file
2452 @file{lib-writ.adb} in the GNAT compiler sources.
2454 @node Binding an Ada Program
2455 @section Binding an Ada Program
2458 When using languages such as C and C++, once the source files have been
2459 compiled the only remaining step in building an executable program
2460 is linking the object modules together. This means that it is possible to
2461 link an inconsistent version of a program, in which two units have
2462 included different versions of the same header.
2464 The rules of Ada do not permit such an inconsistent program to be built.
2465 For example, if two clients have different versions of the same package,
2466 it is illegal to build a program containing these two clients.
2467 These rules are enforced by the GNAT binder, which also determines an
2468 elaboration order consistent with the Ada rules.
2470 The GNAT binder is run after all the object files for a program have
2471 been created. It is given the name of the main program unit, and from
2472 this it determines the set of units required by the program, by reading the
2473 corresponding ALI files. It generates error messages if the program is
2474 inconsistent or if no valid order of elaboration exists.
2476 If no errors are detected, the binder produces a main program, in Ada by
2477 default, that contains calls to the elaboration procedures of those
2478 compilation unit that require them, followed by
2479 a call to the main program. This Ada program is compiled to generate the
2480 object file for the main program. The name of
2481 the Ada file is @file{b~@var{xxx}.adb} (with the corresponding spec
2482 @file{b~@var{xxx}.ads}) where @var{xxx} is the name of the
2485 Finally, the linker is used to build the resulting executable program,
2486 using the object from the main program from the bind step as well as the
2487 object files for the Ada units of the program.
2489 @node Mixed Language Programming
2490 @section Mixed Language Programming
2491 @cindex Mixed Language Programming
2494 This section describes how to develop a mixed-language program,
2495 specifically one that comprises units in both Ada and C.
2498 * Interfacing to C::
2499 * Calling Conventions::
2502 @node Interfacing to C
2503 @subsection Interfacing to C
2505 Interfacing Ada with a foreign language such as C involves using
2506 compiler directives to import and/or export entity definitions in each
2507 language---using @code{extern} statements in C, for instance, and the
2508 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada.
2509 A full treatment of these topics is provided in Appendix B, section 1
2510 of the Ada Reference Manual.
2512 There are two ways to build a program using GNAT that contains some Ada
2513 sources and some foreign language sources, depending on whether or not
2514 the main subprogram is written in Ada. Here is a source example with
2515 the main subprogram in Ada:
2521 void print_num (int num)
2523 printf ("num is %d.\n", num);
2529 /* num_from_Ada is declared in my_main.adb */
2530 extern int num_from_Ada;
2534 return num_from_Ada;
2538 @smallexample @c ada
2540 procedure My_Main is
2542 -- Declare then export an Integer entity called num_from_Ada
2543 My_Num : Integer := 10;
2544 pragma Export (C, My_Num, "num_from_Ada");
2546 -- Declare an Ada function spec for Get_Num, then use
2547 -- C function get_num for the implementation.
2548 function Get_Num return Integer;
2549 pragma Import (C, Get_Num, "get_num");
2551 -- Declare an Ada procedure spec for Print_Num, then use
2552 -- C function print_num for the implementation.
2553 procedure Print_Num (Num : Integer);
2554 pragma Import (C, Print_Num, "print_num");
2557 Print_Num (Get_Num);
2563 To build this example, first compile the foreign language files to
2564 generate object files:
2566 ^gcc -c file1.c^gcc -c FILE1.C^
2567 ^gcc -c file2.c^gcc -c FILE2.C^
2571 Then, compile the Ada units to produce a set of object files and ALI
2574 gnatmake ^-c^/ACTIONS=COMPILE^ my_main.adb
2578 Run the Ada binder on the Ada main program:
2580 gnatbind my_main.ali
2584 Link the Ada main program, the Ada objects and the other language
2587 gnatlink my_main.ali file1.o file2.o
2591 The last three steps can be grouped in a single command:
2593 gnatmake my_main.adb -largs file1.o file2.o
2596 @cindex Binder output file
2598 If the main program is in a language other than Ada, then you may have
2599 more than one entry point into the Ada subsystem. You must use a special
2600 binder option to generate callable routines that initialize and
2601 finalize the Ada units (@pxref{Binding with Non-Ada Main Programs}).
2602 Calls to the initialization and finalization routines must be inserted
2603 in the main program, or some other appropriate point in the code. The
2604 call to initialize the Ada units must occur before the first Ada
2605 subprogram is called, and the call to finalize the Ada units must occur
2606 after the last Ada subprogram returns. The binder will place the
2607 initialization and finalization subprograms into the
2608 @file{b~@var{xxx}.adb} file where they can be accessed by your C
2609 sources. To illustrate, we have the following example:
2613 extern void adainit (void);
2614 extern void adafinal (void);
2615 extern int add (int, int);
2616 extern int sub (int, int);
2618 int main (int argc, char *argv[])
2624 /* Should print "21 + 7 = 28" */
2625 printf ("%d + %d = %d\n", a, b, add (a, b));
2626 /* Should print "21 - 7 = 14" */
2627 printf ("%d - %d = %d\n", a, b, sub (a, b));
2633 @smallexample @c ada
2636 function Add (A, B : Integer) return Integer;
2637 pragma Export (C, Add, "add");
2641 package body Unit1 is
2642 function Add (A, B : Integer) return Integer is
2650 function Sub (A, B : Integer) return Integer;
2651 pragma Export (C, Sub, "sub");
2655 package body Unit2 is
2656 function Sub (A, B : Integer) return Integer is
2665 The build procedure for this application is similar to the last
2666 example's. First, compile the foreign language files to generate object
2669 ^gcc -c main.c^gcc -c main.c^
2673 Next, compile the Ada units to produce a set of object files and ALI
2676 gnatmake ^-c^/ACTIONS=COMPILE^ unit1.adb
2677 gnatmake ^-c^/ACTIONS=COMPILE^ unit2.adb
2681 Run the Ada binder on every generated ALI file. Make sure to use the
2682 @option{-n} option to specify a foreign main program:
2684 gnatbind ^-n^/NOMAIN^ unit1.ali unit2.ali
2688 Link the Ada main program, the Ada objects and the foreign language
2689 objects. You need only list the last ALI file here:
2691 gnatlink unit2.ali main.o -o exec_file
2694 This procedure yields a binary executable called @file{exec_file}.
2698 Depending on the circumstances (for example when your non-Ada main object
2699 does not provide symbol @code{main}), you may also need to instruct the
2700 GNAT linker not to include the standard startup objects by passing the
2701 @option{^-nostartfiles^/NOSTART_FILES^} switch to @command{gnatlink}.
2703 @node Calling Conventions
2704 @subsection Calling Conventions
2705 @cindex Foreign Languages
2706 @cindex Calling Conventions
2707 GNAT follows standard calling sequence conventions and will thus interface
2708 to any other language that also follows these conventions. The following
2709 Convention identifiers are recognized by GNAT:
2712 @cindex Interfacing to Ada
2713 @cindex Other Ada compilers
2714 @cindex Convention Ada
2716 This indicates that the standard Ada calling sequence will be
2717 used and all Ada data items may be passed without any limitations in the
2718 case where GNAT is used to generate both the caller and callee. It is also
2719 possible to mix GNAT generated code and code generated by another Ada
2720 compiler. In this case, the data types should be restricted to simple
2721 cases, including primitive types. Whether complex data types can be passed
2722 depends on the situation. Probably it is safe to pass simple arrays, such
2723 as arrays of integers or floats. Records may or may not work, depending
2724 on whether both compilers lay them out identically. Complex structures
2725 involving variant records, access parameters, tasks, or protected types,
2726 are unlikely to be able to be passed.
2728 Note that in the case of GNAT running
2729 on a platform that supports HP Ada 83, a higher degree of compatibility
2730 can be guaranteed, and in particular records are layed out in an identical
2731 manner in the two compilers. Note also that if output from two different
2732 compilers is mixed, the program is responsible for dealing with elaboration
2733 issues. Probably the safest approach is to write the main program in the
2734 version of Ada other than GNAT, so that it takes care of its own elaboration
2735 requirements, and then call the GNAT-generated adainit procedure to ensure
2736 elaboration of the GNAT components. Consult the documentation of the other
2737 Ada compiler for further details on elaboration.
2739 However, it is not possible to mix the tasking run time of GNAT and
2740 HP Ada 83, All the tasking operations must either be entirely within
2741 GNAT compiled sections of the program, or entirely within HP Ada 83
2742 compiled sections of the program.
2744 @cindex Interfacing to Assembly
2745 @cindex Convention Assembler
2747 Specifies assembler as the convention. In practice this has the
2748 same effect as convention Ada (but is not equivalent in the sense of being
2749 considered the same convention).
2751 @cindex Convention Asm
2754 Equivalent to Assembler.
2756 @cindex Interfacing to COBOL
2757 @cindex Convention COBOL
2760 Data will be passed according to the conventions described
2761 in section B.4 of the Ada Reference Manual.
2764 @cindex Interfacing to C
2765 @cindex Convention C
2767 Data will be passed according to the conventions described
2768 in section B.3 of the Ada Reference Manual.
2770 A note on interfacing to a C ``varargs'' function:
2771 @findex C varargs function
2772 @cindex Interfacing to C varargs function
2773 @cindex varargs function interfaces
2777 In C, @code{varargs} allows a function to take a variable number of
2778 arguments. There is no direct equivalent in this to Ada. One
2779 approach that can be used is to create a C wrapper for each
2780 different profile and then interface to this C wrapper. For
2781 example, to print an @code{int} value using @code{printf},
2782 create a C function @code{printfi} that takes two arguments, a
2783 pointer to a string and an int, and calls @code{printf}.
2784 Then in the Ada program, use pragma @code{Import} to
2785 interface to @code{printfi}.
2788 It may work on some platforms to directly interface to
2789 a @code{varargs} function by providing a specific Ada profile
2790 for a particular call. However, this does not work on
2791 all platforms, since there is no guarantee that the
2792 calling sequence for a two argument normal C function
2793 is the same as for calling a @code{varargs} C function with
2794 the same two arguments.
2797 @cindex Convention Default
2802 @cindex Convention External
2809 @cindex Interfacing to C++
2810 @cindex Convention C++
2811 @item C_Plus_Plus (or CPP)
2812 This stands for C++. For most purposes this is identical to C.
2813 See the separate description of the specialized GNAT pragmas relating to
2814 C++ interfacing for further details.
2818 @cindex Interfacing to Fortran
2819 @cindex Convention Fortran
2821 Data will be passed according to the conventions described
2822 in section B.5 of the Ada Reference Manual.
2825 This applies to an intrinsic operation, as defined in the Ada
2826 Reference Manual. If a pragma Import (Intrinsic) applies to a subprogram,
2827 this means that the body of the subprogram is provided by the compiler itself,
2828 usually by means of an efficient code sequence, and that the user does not
2829 supply an explicit body for it. In an application program, the pragma can
2830 only be applied to the following two sets of names, which the GNAT compiler
2835 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right,
2836 Shift_Right_Arithmetic. The corresponding subprogram declaration must have
2837 two formal parameters. The
2838 first one must be a signed integer type or a modular type with a binary
2839 modulus, and the second parameter must be of type Natural.
2840 The return type must be the same as the type of the first argument. The size
2841 of this type can only be 8, 16, 32, or 64.
2842 @item binary arithmetic operators: ``+'', ``-'', ``*'', ``/''
2843 The corresponding operator declaration must have parameters and result type
2844 that have the same root numeric type (for example, all three are long_float
2845 types). This simplifies the definition of operations that use type checking
2846 to perform dimensional checks:
2848 @smallexample @c ada
2849 type Distance is new Long_Float;
2850 type Time is new Long_Float;
2851 type Velocity is new Long_Float;
2852 function "/" (D : Distance; T : Time)
2854 pragma Import (Intrinsic, "/");
2858 This common idiom is often programmed with a generic definition and an
2859 explicit body. The pragma makes it simpler to introduce such declarations.
2860 It incurs no overhead in compilation time or code size, because it is
2861 implemented as a single machine instruction.
2867 @cindex Convention Stdcall
2869 This is relevant only to Windows XP/2000/NT implementations of GNAT,
2870 and specifies that the @code{Stdcall} calling sequence will be used,
2871 as defined by the NT API. Nevertheless, to ease building
2872 cross-platform bindings this convention will be handled as a @code{C} calling
2873 convention on non Windows platforms.
2876 @cindex Convention DLL
2878 This is equivalent to @code{Stdcall}.
2881 @cindex Convention Win32
2883 This is equivalent to @code{Stdcall}.
2887 @cindex Convention Stubbed
2889 This is a special convention that indicates that the compiler
2890 should provide a stub body that raises @code{Program_Error}.
2894 GNAT additionally provides a useful pragma @code{Convention_Identifier}
2895 that can be used to parametrize conventions and allow additional synonyms
2896 to be specified. For example if you have legacy code in which the convention
2897 identifier Fortran77 was used for Fortran, you can use the configuration
2900 @smallexample @c ada
2901 pragma Convention_Identifier (Fortran77, Fortran);
2905 And from now on the identifier Fortran77 may be used as a convention
2906 identifier (for example in an @code{Import} pragma) with the same
2910 @node Building Mixed Ada & C++ Programs
2911 @section Building Mixed Ada and C++ Programs
2914 A programmer inexperienced with mixed-language development may find that
2915 building an application containing both Ada and C++ code can be a
2916 challenge. This section gives a few
2917 hints that should make this task easier. The first section addresses
2918 the differences between interfacing with C and interfacing with C++.
2920 looks into the delicate problem of linking the complete application from
2921 its Ada and C++ parts. The last section gives some hints on how the GNAT
2922 run-time library can be adapted in order to allow inter-language dispatching
2923 with a new C++ compiler.
2926 * Interfacing to C++::
2927 * Linking a Mixed C++ & Ada Program::
2928 * A Simple Example::
2929 * Interfacing with C++ at the Class Level::
2932 @node Interfacing to C++
2933 @subsection Interfacing to C++
2936 GNAT supports interfacing with the G++ compiler (or any C++ compiler
2937 generating code that is compatible with the G++ Application Binary
2938 Interface ---see http://www.codesourcery.com/archives/cxx-abi).
2941 Interfacing can be done at 3 levels: simple data, subprograms, and
2942 classes. In the first two cases, GNAT offers a specific @var{Convention
2943 C_Plus_Plus} (or @var{CPP}) that behaves exactly like @var{Convention C}.
2944 Usually, C++ mangles the names of subprograms, and currently, GNAT does
2945 not provide any help to solve the demangling problem. This problem can be
2946 addressed in two ways:
2949 by modifying the C++ code in order to force a C convention using
2950 the @code{extern "C"} syntax.
2953 by figuring out the mangled name and use it as the Link_Name argument of
2958 Interfacing at the class level can be achieved by using the GNAT specific
2959 pragmas such as @code{CPP_Constructor}. See the GNAT Reference Manual for
2960 additional information.
2962 @node Linking a Mixed C++ & Ada Program
2963 @subsection Linking a Mixed C++ & Ada Program
2966 Usually the linker of the C++ development system must be used to link
2967 mixed applications because most C++ systems will resolve elaboration
2968 issues (such as calling constructors on global class instances)
2969 transparently during the link phase. GNAT has been adapted to ease the
2970 use of a foreign linker for the last phase. Three cases can be
2975 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
2976 The C++ linker can simply be called by using the C++ specific driver
2977 called @code{c++}. Note that this setup is not very common because it
2978 may involve recompiling the whole GCC tree from sources, which makes it
2979 harder to upgrade the compilation system for one language without
2980 destabilizing the other.
2985 $ gnatmake ada_unit -largs file1.o file2.o --LINK=c++
2989 Using GNAT and G++ from two different GCC installations: If both
2990 compilers are on the PATH, the previous method may be used. It is
2991 important to note that environment variables such as C_INCLUDE_PATH,
2992 GCC_EXEC_PREFIX, BINUTILS_ROOT, and GCC_ROOT will affect both compilers
2993 at the same time and may make one of the two compilers operate
2994 improperly if set during invocation of the wrong compiler. It is also
2995 very important that the linker uses the proper @file{libgcc.a} GCC
2996 library -- that is, the one from the C++ compiler installation. The
2997 implicit link command as suggested in the gnatmake command from the
2998 former example can be replaced by an explicit link command with the
2999 full-verbosity option in order to verify which library is used:
3002 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
3004 If there is a problem due to interfering environment variables, it can
3005 be worked around by using an intermediate script. The following example
3006 shows the proper script to use when GNAT has not been installed at its
3007 default location and g++ has been installed at its default location:
3015 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
3019 Using a non-GNU C++ compiler: The commands previously described can be
3020 used to insure that the C++ linker is used. Nonetheless, you need to add
3021 a few more parameters to the link command line, depending on the exception
3024 If the @code{setjmp/longjmp} exception mechanism is used, only the paths
3025 to the libgcc libraries are required:
3030 CC $* `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a`
3031 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3034 Where CC is the name of the non-GNU C++ compiler.
3036 If the @code{zero cost} exception mechanism is used, and the platform
3037 supports automatic registration of exception tables (e.g. Solaris or IRIX),
3038 paths to more objects are required:
3043 CC `gcc -print-file-name=crtbegin.o` $* \
3044 `gcc -print-file-name=libgcc.a` `gcc -print-file-name=libgcc_eh.a` \
3045 `gcc -print-file-name=crtend.o`
3046 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3049 If the @code{zero cost} exception mechanism is used, and the platform
3050 doesn't support automatic registration of exception tables (e.g. HP-UX,
3051 Tru64 or AIX), the simple approach described above will not work and
3052 a pre-linking phase using GNAT will be necessary.
3056 @node A Simple Example
3057 @subsection A Simple Example
3059 The following example, provided as part of the GNAT examples, shows how
3060 to achieve procedural interfacing between Ada and C++ in both
3061 directions. The C++ class A has two methods. The first method is exported
3062 to Ada by the means of an extern C wrapper function. The second method
3063 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
3064 a limited record with a layout comparable to the C++ class. The Ada
3065 subprogram, in turn, calls the C++ method. So, starting from the C++
3066 main program, the process passes back and forth between the two
3070 Here are the compilation commands:
3072 $ gnatmake -c simple_cpp_interface
3075 $ gnatbind -n simple_cpp_interface
3076 $ gnatlink simple_cpp_interface -o cpp_main --LINK=$(CPLUSPLUS)
3077 -lstdc++ ex7.o cpp_main.o
3081 Here are the corresponding sources:
3089 void adainit (void);
3090 void adafinal (void);
3091 void method1 (A *t);
3113 class A : public Origin @{
3115 void method1 (void);
3116 void method2 (int v);
3126 extern "C" @{ void ada_method2 (A *t, int v);@}
3128 void A::method1 (void)
3131 printf ("in A::method1, a_value = %d \n",a_value);
3135 void A::method2 (int v)
3137 ada_method2 (this, v);
3138 printf ("in A::method2, a_value = %d \n",a_value);
3145 printf ("in A::A, a_value = %d \n",a_value);
3149 @smallexample @c ada
3151 package body Simple_Cpp_Interface is
3153 procedure Ada_Method2 (This : in out A; V : Integer) is
3159 end Simple_Cpp_Interface;
3162 package Simple_Cpp_Interface is
3165 Vptr : System.Address;
3169 pragma Convention (C, A);
3171 procedure Method1 (This : in out A);
3172 pragma Import (C, Method1);
3174 procedure Ada_Method2 (This : in out A; V : Integer);
3175 pragma Export (C, Ada_Method2);
3177 end Simple_Cpp_Interface;
3180 @node Interfacing with C++ at the Class Level
3181 @subsection Interfacing with C++ at the Class Level
3183 In this section we demonstrate the GNAT features for interfacing with
3184 C++ by means of an example making use of Ada 2005 abstract interface
3185 types. This example consists of a classification of animals; classes
3186 have been used to model our main classification of animals, and
3187 interfaces provide support for the management of secondary
3188 classifications. We first demonstrate a case in which the types and
3189 constructors are defined on the C++ side and imported from the Ada
3190 side, and latter the reverse case.
3192 The root of our derivation will be the @code{Animal} class, with a
3193 single private attribute (the @code{Age} of the animal) and two public
3194 primitives to set and get the value of this attribute.
3199 @b{virtual} void Set_Age (int New_Age);
3200 @b{virtual} int Age ();
3206 Abstract interface types are defined in C++ by means of classes with pure
3207 virtual functions and no data members. In our example we will use two
3208 interfaces that provide support for the common management of @code{Carnivore}
3209 and @code{Domestic} animals:
3212 @b{class} Carnivore @{
3214 @b{virtual} int Number_Of_Teeth () = 0;
3217 @b{class} Domestic @{
3219 @b{virtual void} Set_Owner (char* Name) = 0;
3223 Using these declarations, we can now say that a @code{Dog} is an animal that is
3224 both Carnivore and Domestic, that is:
3227 @b{class} Dog : Animal, Carnivore, Domestic @{
3229 @b{virtual} int Number_Of_Teeth ();
3230 @b{virtual} void Set_Owner (char* Name);
3232 Dog(); // Constructor
3239 In the following examples we will assume that the previous declarations are
3240 located in a file named @code{animals.h}. The following package demonstrates
3241 how to import these C++ declarations from the Ada side:
3243 @smallexample @c ada
3244 with Interfaces.C.Strings; use Interfaces.C.Strings;
3246 type Carnivore is interface;
3247 pragma Convention (C_Plus_Plus, Carnivore);
3248 function Number_Of_Teeth (X : Carnivore)
3249 return Natural is abstract;
3251 type Domestic is interface;
3252 pragma Convention (C_Plus_Plus, Set_Owner);
3254 (X : in out Domestic;
3255 Name : Chars_Ptr) is abstract;
3257 type Animal is tagged record
3260 pragma Import (C_Plus_Plus, Animal);
3262 procedure Set_Age (X : in out Animal; Age : Integer);
3263 pragma Import (C_Plus_Plus, Set_Age);
3265 function Age (X : Animal) return Integer;
3266 pragma Import (C_Plus_Plus, Age);
3268 type Dog is new Animal and Carnivore and Domestic with record
3269 Tooth_Count : Natural;
3270 Owner : String (1 .. 30);
3272 pragma Import (C_Plus_Plus, Dog);
3274 function Number_Of_Teeth (A : Dog) return Integer;
3275 pragma Import (C_Plus_Plus, Number_Of_Teeth);
3277 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3278 pragma Import (C_Plus_Plus, Set_Owner);
3280 function New_Dog return Dog'Class;
3281 pragma CPP_Constructor (New_Dog);
3282 pragma Import (CPP, New_Dog, "_ZN3DogC2Ev");
3286 Thanks to the compatibility between GNAT run-time structures and the C++ ABI,
3287 interfacing with these C++ classes is easy. The only requirement is that all
3288 the primitives and components must be declared exactly in the same order in
3291 Regarding the abstract interfaces, we must indicate to the GNAT compiler by
3292 means of a @code{pragma Convention (C_Plus_Plus)}, the convention used to pass
3293 the arguments to the called primitives will be the same as for C++. For the
3294 imported classes we use @code{pragma Import} with convention @code{C_Plus_Plus}
3295 to indicate that they have been defined on the C++ side; this is required
3296 because the dispatch table associated with these tagged types will be built
3297 in the C++ side and therefore will not contain the predefined Ada primitives
3298 which Ada would otherwise expect.
3300 As the reader can see there is no need to indicate the C++ mangled names
3301 associated with each subprogram because it is assumed that all the calls to
3302 these primitives will be dispatching calls. The only exception is the
3303 constructor, which must be registered with the compiler by means of
3304 @code{pragma CPP_Constructor} and needs to provide its associated C++
3305 mangled name because the Ada compiler generates direct calls to it.
3307 With the above packages we can now declare objects of type Dog on the Ada side
3308 and dispatch calls to the corresponding subprograms on the C++ side. We can
3309 also extend the tagged type Dog with further fields and primitives, and
3310 override some of its C++ primitives on the Ada side. For example, here we have
3311 a type derivation defined on the Ada side that inherits all the dispatching
3312 primitives of the ancestor from the C++ side.
3315 @b{with} Animals; @b{use} Animals;
3316 @b{package} Vaccinated_Animals @b{is}
3317 @b{type} Vaccinated_Dog @b{is new} Dog @b{with null record};
3318 @b{function} Vaccination_Expired (A : Vaccinated_Dog) @b{return} Boolean;
3319 @b{end} Vaccinated_Animals;
3322 It is important to note that, because of the ABI compatibility, the programmer
3323 does not need to add any further information to indicate either the object
3324 layout or the dispatch table entry associated with each dispatching operation.
3326 Now let us define all the types and constructors on the Ada side and export
3327 them to C++, using the same hierarchy of our previous example:
3329 @smallexample @c ada
3330 with Interfaces.C.Strings;
3331 use Interfaces.C.Strings;
3333 type Carnivore is interface;
3334 pragma Convention (C_Plus_Plus, Carnivore);
3335 function Number_Of_Teeth (X : Carnivore)
3336 return Natural is abstract;
3338 type Domestic is interface;
3339 pragma Convention (C_Plus_Plus, Set_Owner);
3341 (X : in out Domestic;
3342 Name : Chars_Ptr) is abstract;
3344 type Animal is tagged record
3347 pragma Convention (C_Plus_Plus, Animal);
3349 procedure Set_Age (X : in out Animal; Age : Integer);
3350 pragma Export (C_Plus_Plus, Set_Age);
3352 function Age (X : Animal) return Integer;
3353 pragma Export (C_Plus_Plus, Age);
3355 type Dog is new Animal and Carnivore and Domestic with record
3356 Tooth_Count : Natural;
3357 Owner : String (1 .. 30);
3359 pragma Convention (C_Plus_Plus, Dog);
3361 function Number_Of_Teeth (A : Dog) return Integer;
3362 pragma Export (C_Plus_Plus, Number_Of_Teeth);
3364 procedure Set_Owner (A : in out Dog; Name : Chars_Ptr);
3365 pragma Export (C_Plus_Plus, Set_Owner);
3367 function New_Dog return Dog'Class;
3368 pragma Export (C_Plus_Plus, New_Dog);
3372 Compared with our previous example the only difference is the use of
3373 @code{pragma Export} to indicate to the GNAT compiler that the primitives will
3374 be available to C++. Thanks to the ABI compatibility, on the C++ side there is
3375 nothing else to be done; as explained above, the only requirement is that all
3376 the primitives and components are declared in exactly the same order.
3378 For completeness, let us see a brief C++ main program that uses the
3379 declarations available in @code{animals.h} (presented in our first example) to
3380 import and use the declarations from the Ada side, properly initializing and
3381 finalizing the Ada run-time system along the way:
3384 @b{#include} "animals.h"
3385 @b{#include} <iostream>
3386 @b{using namespace} std;
3388 void Check_Carnivore (Carnivore *obj) @{ ... @}
3389 void Check_Domestic (Domestic *obj) @{ ... @}
3390 void Check_Animal (Animal *obj) @{ ... @}
3391 void Check_Dog (Dog *obj) @{ ... @}
3394 void adainit (void);
3395 void adafinal (void);
3401 Dog *obj = new_dog(); // Ada constructor
3402 Check_Carnivore (obj); // Check secondary DT
3403 Check_Domestic (obj); // Check secondary DT
3404 Check_Animal (obj); // Check primary DT
3405 Check_Dog (obj); // Check primary DT
3410 adainit (); test(); adafinal ();
3415 @node Comparison between GNAT and C/C++ Compilation Models
3416 @section Comparison between GNAT and C/C++ Compilation Models
3419 The GNAT model of compilation is close to the C and C++ models. You can
3420 think of Ada specs as corresponding to header files in C. As in C, you
3421 don't need to compile specs; they are compiled when they are used. The
3422 Ada @code{with} is similar in effect to the @code{#include} of a C
3425 One notable difference is that, in Ada, you may compile specs separately
3426 to check them for semantic and syntactic accuracy. This is not always
3427 possible with C headers because they are fragments of programs that have
3428 less specific syntactic or semantic rules.
3430 The other major difference is the requirement for running the binder,
3431 which performs two important functions. First, it checks for
3432 consistency. In C or C++, the only defense against assembling
3433 inconsistent programs lies outside the compiler, in a makefile, for
3434 example. The binder satisfies the Ada requirement that it be impossible
3435 to construct an inconsistent program when the compiler is used in normal
3438 @cindex Elaboration order control
3439 The other important function of the binder is to deal with elaboration
3440 issues. There are also elaboration issues in C++ that are handled
3441 automatically. This automatic handling has the advantage of being
3442 simpler to use, but the C++ programmer has no control over elaboration.
3443 Where @code{gnatbind} might complain there was no valid order of
3444 elaboration, a C++ compiler would simply construct a program that
3445 malfunctioned at run time.
3448 @node Comparison between GNAT and Conventional Ada Library Models
3449 @section Comparison between GNAT and Conventional Ada Library Models
3452 This section is intended for Ada programmers who have
3453 used an Ada compiler implementing the traditional Ada library
3454 model, as described in the Ada Reference Manual.
3456 @cindex GNAT library
3457 In GNAT, there is no ``library'' in the normal sense. Instead, the set of
3458 source files themselves acts as the library. Compiling Ada programs does
3459 not generate any centralized information, but rather an object file and
3460 a ALI file, which are of interest only to the binder and linker.
3461 In a traditional system, the compiler reads information not only from
3462 the source file being compiled, but also from the centralized library.
3463 This means that the effect of a compilation depends on what has been
3464 previously compiled. In particular:
3468 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3469 to the version of the unit most recently compiled into the library.
3472 Inlining is effective only if the necessary body has already been
3473 compiled into the library.
3476 Compiling a unit may obsolete other units in the library.
3480 In GNAT, compiling one unit never affects the compilation of any other
3481 units because the compiler reads only source files. Only changes to source
3482 files can affect the results of a compilation. In particular:
3486 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3487 to the source version of the unit that is currently accessible to the
3492 Inlining requires the appropriate source files for the package or
3493 subprogram bodies to be available to the compiler. Inlining is always
3494 effective, independent of the order in which units are complied.
3497 Compiling a unit never affects any other compilations. The editing of
3498 sources may cause previous compilations to be out of date if they
3499 depended on the source file being modified.
3503 The most important result of these differences is that order of compilation
3504 is never significant in GNAT. There is no situation in which one is
3505 required to do one compilation before another. What shows up as order of
3506 compilation requirements in the traditional Ada library becomes, in
3507 GNAT, simple source dependencies; in other words, there is only a set
3508 of rules saying what source files must be present when a file is
3512 @node Placement of temporary files
3513 @section Placement of temporary files
3514 @cindex Temporary files (user control over placement)
3517 GNAT creates temporary files in the directory designated by the environment
3518 variable @env{TMPDIR}.
3519 (See the HP @emph{C RTL Reference Manual} on the function @code{getenv()}
3520 for detailed information on how environment variables are resolved.
3521 For most users the easiest way to make use of this feature is to simply
3522 define @env{TMPDIR} as a job level logical name).
3523 For example, if you wish to use a Ramdisk (assuming DECRAM is installed)
3524 for compiler temporary files, then you can include something like the
3525 following command in your @file{LOGIN.COM} file:
3528 $ define/job TMPDIR "/disk$scratchram/000000/temp/"
3532 If @env{TMPDIR} is not defined, then GNAT uses the directory designated by
3533 @env{TMP}; if @env{TMP} is not defined, then GNAT uses the directory
3534 designated by @env{TEMP}.
3535 If none of these environment variables are defined then GNAT uses the
3536 directory designated by the logical name @code{SYS$SCRATCH:}
3537 (by default the user's home directory). If all else fails
3538 GNAT uses the current directory for temporary files.
3541 @c *************************
3542 @node Compiling Using gcc
3543 @chapter Compiling Using @command{gcc}
3546 This chapter discusses how to compile Ada programs using the @command{gcc}
3547 command. It also describes the set of switches
3548 that can be used to control the behavior of the compiler.
3550 * Compiling Programs::
3551 * Switches for gcc::
3552 * Search Paths and the Run-Time Library (RTL)::
3553 * Order of Compilation Issues::
3557 @node Compiling Programs
3558 @section Compiling Programs
3561 The first step in creating an executable program is to compile the units
3562 of the program using the @command{gcc} command. You must compile the
3567 the body file (@file{.adb}) for a library level subprogram or generic
3571 the spec file (@file{.ads}) for a library level package or generic
3572 package that has no body
3575 the body file (@file{.adb}) for a library level package
3576 or generic package that has a body
3581 You need @emph{not} compile the following files
3586 the spec of a library unit which has a body
3593 because they are compiled as part of compiling related units. GNAT
3595 when the corresponding body is compiled, and subunits when the parent is
3598 @cindex cannot generate code
3599 If you attempt to compile any of these files, you will get one of the
3600 following error messages (where fff is the name of the file you compiled):
3603 cannot generate code for file @var{fff} (package spec)
3604 to check package spec, use -gnatc
3606 cannot generate code for file @var{fff} (missing subunits)
3607 to check parent unit, use -gnatc
3609 cannot generate code for file @var{fff} (subprogram spec)
3610 to check subprogram spec, use -gnatc
3612 cannot generate code for file @var{fff} (subunit)
3613 to check subunit, use -gnatc
3617 As indicated by the above error messages, if you want to submit
3618 one of these files to the compiler to check for correct semantics
3619 without generating code, then use the @option{-gnatc} switch.
3621 The basic command for compiling a file containing an Ada unit is
3624 $ gcc -c [@var{switches}] @file{file name}
3628 where @var{file name} is the name of the Ada file (usually
3630 @file{.ads} for a spec or @file{.adb} for a body).
3633 @option{-c} switch to tell @command{gcc} to compile, but not link, the file.
3635 The result of a successful compilation is an object file, which has the
3636 same name as the source file but an extension of @file{.o} and an Ada
3637 Library Information (ALI) file, which also has the same name as the
3638 source file, but with @file{.ali} as the extension. GNAT creates these
3639 two output files in the current directory, but you may specify a source
3640 file in any directory using an absolute or relative path specification
3641 containing the directory information.
3644 @command{gcc} is actually a driver program that looks at the extensions of
3645 the file arguments and loads the appropriate compiler. For example, the
3646 GNU C compiler is @file{cc1}, and the Ada compiler is @file{gnat1}.
3647 These programs are in directories known to the driver program (in some
3648 configurations via environment variables you set), but need not be in
3649 your path. The @command{gcc} driver also calls the assembler and any other
3650 utilities needed to complete the generation of the required object
3653 It is possible to supply several file names on the same @command{gcc}
3654 command. This causes @command{gcc} to call the appropriate compiler for
3655 each file. For example, the following command lists three separate
3656 files to be compiled:
3659 $ gcc -c x.adb y.adb z.c
3663 calls @code{gnat1} (the Ada compiler) twice to compile @file{x.adb} and
3664 @file{y.adb}, and @code{cc1} (the C compiler) once to compile @file{z.c}.
3665 The compiler generates three object files @file{x.o}, @file{y.o} and
3666 @file{z.o} and the two ALI files @file{x.ali} and @file{y.ali} from the
3667 Ada compilations. Any switches apply to all the files ^listed,^listed.^
3670 @option{-gnat@var{x}} switches, which apply only to Ada compilations.
3673 @node Switches for gcc
3674 @section Switches for @command{gcc}
3677 The @command{gcc} command accepts switches that control the
3678 compilation process. These switches are fully described in this section.
3679 First we briefly list all the switches, in alphabetical order, then we
3680 describe the switches in more detail in functionally grouped sections.
3682 More switches exist for GCC than those documented here, especially
3683 for specific targets. However, their use is not recommended as
3684 they may change code generation in ways that are incompatible with
3685 the Ada run-time library, or can cause inconsistencies between
3689 * Output and Error Message Control::
3690 * Warning Message Control::
3691 * Debugging and Assertion Control::
3692 * Validity Checking::
3695 * Using gcc for Syntax Checking::
3696 * Using gcc for Semantic Checking::
3697 * Compiling Different Versions of Ada::
3698 * Character Set Control::
3699 * File Naming Control::
3700 * Subprogram Inlining Control::
3701 * Auxiliary Output Control::
3702 * Debugging Control::
3703 * Exception Handling Control::
3704 * Units to Sources Mapping Files::
3705 * Integrated Preprocessing::
3706 * Code Generation Control::
3715 @cindex @option{-b} (@command{gcc})
3716 @item -b @var{target}
3717 Compile your program to run on @var{target}, which is the name of a
3718 system configuration. You must have a GNAT cross-compiler built if
3719 @var{target} is not the same as your host system.
3722 @cindex @option{-B} (@command{gcc})
3723 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
3724 from @var{dir} instead of the default location. Only use this switch
3725 when multiple versions of the GNAT compiler are available. See the
3726 @command{gcc} manual page for further details. You would normally use the
3727 @option{-b} or @option{-V} switch instead.
3730 @cindex @option{-c} (@command{gcc})
3731 Compile. Always use this switch when compiling Ada programs.
3733 Note: for some other languages when using @command{gcc}, notably in
3734 the case of C and C++, it is possible to use
3735 use @command{gcc} without a @option{-c} switch to
3736 compile and link in one step. In the case of GNAT, you
3737 cannot use this approach, because the binder must be run
3738 and @command{gcc} cannot be used to run the GNAT binder.
3742 @cindex @option{-fno-inline} (@command{gcc})
3743 Suppresses all back-end inlining, even if other optimization or inlining
3745 This includes suppression of inlining that results
3746 from the use of the pragma @code{Inline_Always}.
3747 See also @option{-gnatn} and @option{-gnatN}.
3749 @item -fno-strict-aliasing
3750 @cindex @option{-fno-strict-aliasing} (@command{gcc})
3751 Causes the compiler to avoid assumptions regarding non-aliasing
3752 of objects of different types. See
3753 @ref{Optimization and Strict Aliasing} for details.
3756 @cindex @option{-fstack-check} (@command{gcc})
3757 Activates stack checking.
3758 See @ref{Stack Overflow Checking} for details.
3761 @cindex @option{-fstack-usage} (@command{gcc})
3762 Makes the compiler output stack usage information for the program, on a
3763 per-function basis. See @ref{Static Stack Usage Analysis} for details.
3765 @item -fcallgraph-info[=su]
3766 @cindex @option{-fcallgraph-info} (@command{gcc})
3767 Makes the compiler output callgraph information for the program, on a
3768 per-file basis. The information is generated in the VCG format. It can
3769 be decorated with stack-usage per-node information.
3772 @cindex @option{^-g^/DEBUG^} (@command{gcc})
3773 Generate debugging information. This information is stored in the object
3774 file and copied from there to the final executable file by the linker,
3775 where it can be read by the debugger. You must use the
3776 @option{^-g^/DEBUG^} switch if you plan on using the debugger.
3779 @cindex @option{-gnat83} (@command{gcc})
3780 Enforce Ada 83 restrictions.
3783 @cindex @option{-gnat95} (@command{gcc})
3784 Enforce Ada 95 restrictions.
3787 @cindex @option{-gnat05} (@command{gcc})
3788 Allow full Ada 2005 features.
3791 @cindex @option{-gnata} (@command{gcc})
3792 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
3793 activated. Note that these pragmas can also be controlled using the
3794 configuration pragmas @code{Assertion_Policy} and @code{Debug_Policy}.
3797 @cindex @option{-gnatA} (@command{gcc})
3798 Avoid processing @file{gnat.adc}. If a gnat.adc file is present,
3802 @cindex @option{-gnatb} (@command{gcc})
3803 Generate brief messages to @file{stderr} even if verbose mode set.
3806 @cindex @option{-gnatc} (@command{gcc})
3807 Check syntax and semantics only (no code generation attempted).
3810 @cindex @option{-gnatd} (@command{gcc})
3811 Specify debug options for the compiler. The string of characters after
3812 the @option{-gnatd} specify the specific debug options. The possible
3813 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
3814 compiler source file @file{debug.adb} for details of the implemented
3815 debug options. Certain debug options are relevant to applications
3816 programmers, and these are documented at appropriate points in this
3820 @cindex @option{-gnatD} (@command{gcc})
3821 Create expanded source files for source level debugging. This switch
3822 also suppress generation of cross-reference information
3823 (see @option{-gnatx}).
3825 @item -gnatec=@var{path}
3826 @cindex @option{-gnatec} (@command{gcc})
3827 Specify a configuration pragma file
3829 (the equal sign is optional)
3831 (@pxref{The Configuration Pragmas Files}).
3833 @item ^-gnateD^/DATA_PREPROCESSING=^symbol[=value]
3834 @cindex @option{-gnateD} (@command{gcc})
3835 Defines a symbol, associated with value, for preprocessing.
3836 (@pxref{Integrated Preprocessing}).
3839 @cindex @option{-gnatef} (@command{gcc})
3840 Display full source path name in brief error messages.
3842 @item -gnatem=@var{path}
3843 @cindex @option{-gnatem} (@command{gcc})
3844 Specify a mapping file
3846 (the equal sign is optional)
3848 (@pxref{Units to Sources Mapping Files}).
3850 @item -gnatep=@var{file}
3851 @cindex @option{-gnatep} (@command{gcc})
3852 Specify a preprocessing data file
3854 (the equal sign is optional)
3856 (@pxref{Integrated Preprocessing}).
3859 @cindex @option{-gnatE} (@command{gcc})
3860 Full dynamic elaboration checks.
3863 @cindex @option{-gnatf} (@command{gcc})
3864 Full errors. Multiple errors per line, all undefined references, do not
3865 attempt to suppress cascaded errors.
3868 @cindex @option{-gnatF} (@command{gcc})
3869 Externals names are folded to all uppercase.
3871 @item ^-gnatg^/GNAT_INTERNAL^
3872 @cindex @option{^-gnatg^/GNAT_INTERNAL^} (@command{gcc})
3873 Internal GNAT implementation mode. This should not be used for
3874 applications programs, it is intended only for use by the compiler
3875 and its run-time library. For documentation, see the GNAT sources.
3876 Note that @option{^-gnatg^/GNAT_INTERNAL^} implies
3877 @option{^-gnatwae^/WARNINGS=ALL,ERRORS^} and
3878 @option{^-gnatyg^/STYLE_CHECKS=GNAT^}
3879 so that all standard warnings and all standard style options are turned on.
3880 All warnings and style error messages are treated as errors.
3883 @cindex @option{-gnatG} (@command{gcc})
3884 List generated expanded code in source form.
3886 @item ^-gnath^/HELP^
3887 @cindex @option{^-gnath^/HELP^} (@command{gcc})
3888 Output usage information. The output is written to @file{stdout}.
3890 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
3891 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
3892 Identifier character set
3894 (@var{c}=1/2/3/4/8/9/p/f/n/w).
3897 For details of the possible selections for @var{c},
3898 see @ref{Character Set Control}.
3902 @cindex @option{-gnatjnn} (@command{gcc})
3903 Reformat error messages to fit on nn character lines
3905 @item -gnatk=@var{n}
3906 @cindex @option{-gnatk} (@command{gcc})
3907 Limit file names to @var{n} (1-999) characters ^(@code{k} = krunch)^^.
3910 @cindex @option{-gnatl} (@command{gcc})
3911 Output full source listing with embedded error messages.
3914 @cindex @option{-gnatL} (@command{gcc})
3915 Used in conjunction with -gnatG or -gnatD to intersperse original
3916 source lines (as comment lines with line numbers) in the expanded
3919 @item -gnatm=@var{n}
3920 @cindex @option{-gnatm} (@command{gcc})
3921 Limit number of detected error or warning messages to @var{n}
3922 where @var{n} is in the range 1..999_999. The default setting if
3923 no switch is given is 9999. Compilation is terminated if this
3924 limit is exceeded. The equal sign here is optional.
3927 @cindex @option{-gnatn} (@command{gcc})
3928 Activate inlining for subprograms for which
3929 pragma @code{inline} is specified. This inlining is performed
3930 by the GCC back-end.
3933 @cindex @option{-gnatN} (@command{gcc})
3934 Activate front end inlining for subprograms for which
3935 pragma @code{Inline} is specified. This inlining is performed
3936 by the front end and will be visible in the
3937 @option{-gnatG} output.
3938 In some cases, this has proved more effective than the back end
3939 inlining resulting from the use of
3942 @option{-gnatN} automatically implies
3943 @option{-gnatn} so it is not necessary
3944 to specify both options. There are a few cases that the back-end inlining
3945 catches that cannot be dealt with in the front-end.
3948 @cindex @option{-gnato} (@command{gcc})
3949 Enable numeric overflow checking (which is not normally enabled by
3950 default). Not that division by zero is a separate check that is not
3951 controlled by this switch (division by zero checking is on by default).
3954 @cindex @option{-gnatp} (@command{gcc})
3955 Suppress all checks.
3958 @cindex @option{-gnatP} (@command{gcc})
3959 Enable polling. This is required on some systems (notably Windows NT) to
3960 obtain asynchronous abort and asynchronous transfer of control capability.
3961 See the description of pragma Polling in the GNAT Reference Manual for
3965 @cindex @option{-gnatq} (@command{gcc})
3966 Don't quit; try semantics, even if parse errors.
3969 @cindex @option{-gnatQ} (@command{gcc})
3970 Don't quit; generate @file{ALI} and tree files even if illegalities.
3972 @item ^-gnatR[0/1/2/3[s]]^/REPRESENTATION_INFO^
3973 @cindex @option{-gnatR} (@command{gcc})
3974 Output representation information for declared types and objects.
3977 @cindex @option{-gnats} (@command{gcc})
3981 @cindex @option{-gnatS} (@command{gcc})
3982 Print package Standard.
3985 @cindex @option{-gnatt} (@command{gcc})
3986 Generate tree output file.
3988 @item ^-gnatT^/TABLE_MULTIPLIER=^@var{nnn}
3989 @cindex @option{^-gnatT^/TABLE_MULTIPLIER^} (@command{gcc})
3990 All compiler tables start at @var{nnn} times usual starting size.
3993 @cindex @option{-gnatu} (@command{gcc})
3994 List units for this compilation.
3997 @cindex @option{-gnatU} (@command{gcc})
3998 Tag all error messages with the unique string ``error:''
4001 @cindex @option{-gnatv} (@command{gcc})
4002 Verbose mode. Full error output with source lines to @file{stdout}.
4005 @cindex @option{-gnatV} (@command{gcc})
4006 Control level of validity checking. See separate section describing
4009 @item ^-gnatw@var{xxx}^/WARNINGS=(@var{option}[,...])^
4010 @cindex @option{^-gnatw^/WARNINGS^} (@command{gcc})
4012 ^@var{xxx} is a string of option letters that^the list of options^ denotes
4013 the exact warnings that
4014 are enabled or disabled (@pxref{Warning Message Control}).
4016 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
4017 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
4018 Wide character encoding method
4020 (@var{e}=n/h/u/s/e/8).
4023 (@var{e}=@code{BRACKETS, NONE, HEX, UPPER, SHIFT_JIS, EUC, UTF8})
4027 @cindex @option{-gnatx} (@command{gcc})
4028 Suppress generation of cross-reference information.
4030 @item ^-gnaty^/STYLE_CHECKS=(option,option..)^
4031 @cindex @option{^-gnaty^/STYLE_CHECKS^} (@command{gcc})
4032 Enable built-in style checks (@pxref{Style Checking}).
4034 @item ^-gnatz^/DISTRIBUTION_STUBS=^@var{m}
4035 @cindex @option{^-gnatz^/DISTRIBUTION_STUBS^} (@command{gcc})
4036 Distribution stub generation and compilation
4038 (@var{m}=r/c for receiver/caller stubs).
4041 (@var{m}=@code{RECEIVER} or @code{CALLER} to specify the type of stubs
4042 to be generated and compiled).
4045 @item ^-I^/SEARCH=^@var{dir}
4046 @cindex @option{^-I^/SEARCH^} (@command{gcc})
4048 Direct GNAT to search the @var{dir} directory for source files needed by
4049 the current compilation
4050 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4052 @item ^-I-^/NOCURRENT_DIRECTORY^
4053 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gcc})
4055 Except for the source file named in the command line, do not look for source
4056 files in the directory containing the source file named in the command line
4057 (@pxref{Search Paths and the Run-Time Library (RTL)}).
4061 @cindex @option{-mbig-switch} (@command{gcc})
4062 @cindex @code{case} statement (effect of @option{-mbig-switch} option)
4063 This standard gcc switch causes the compiler to use larger offsets in its
4064 jump table representation for @code{case} statements.
4065 This may result in less efficient code, but is sometimes necessary
4066 (for example on HP-UX targets)
4067 @cindex HP-UX and @option{-mbig-switch} option
4068 in order to compile large and/or nested @code{case} statements.
4071 @cindex @option{-o} (@command{gcc})
4072 This switch is used in @command{gcc} to redirect the generated object file
4073 and its associated ALI file. Beware of this switch with GNAT, because it may
4074 cause the object file and ALI file to have different names which in turn
4075 may confuse the binder and the linker.
4079 @cindex @option{-nostdinc} (@command{gcc})
4080 Inhibit the search of the default location for the GNAT Run Time
4081 Library (RTL) source files.
4084 @cindex @option{-nostdlib} (@command{gcc})
4085 Inhibit the search of the default location for the GNAT Run Time
4086 Library (RTL) ALI files.
4090 @cindex @option{-O} (@command{gcc})
4091 @var{n} controls the optimization level.
4095 No optimization, the default setting if no @option{-O} appears
4098 Normal optimization, the default if you specify @option{-O} without
4099 an operand. A good compromise between code quality and compilation
4103 Extensive optimization, may improve execution time, possibly at the cost of
4104 substantially increased compilation time.
4107 Same as @option{-O2}, and also includes inline expansion for small subprograms
4111 Optimize space usage
4115 See also @ref{Optimization Levels}.
4120 @cindex @option{/NOOPTIMIZE} (@code{GNAT COMPILE})
4121 Equivalent to @option{/OPTIMIZE=NONE}.
4122 This is the default behavior in the absence of an @option{/OPTIMIZE}
4125 @item /OPTIMIZE[=(keyword[,...])]
4126 @cindex @option{/OPTIMIZE} (@code{GNAT COMPILE})
4127 Selects the level of optimization for your program. The supported
4128 keywords are as follows:
4131 Perform most optimizations, including those that
4133 This is the default if the @option{/OPTIMIZE} qualifier is supplied
4134 without keyword options.
4137 Do not do any optimizations. Same as @code{/NOOPTIMIZE}.
4140 Perform some optimizations, but omit ones that are costly.
4143 Same as @code{SOME}.
4146 Full optimization as in @option{/OPTIMIZE=ALL}, and also attempts
4147 automatic inlining of small subprograms within a unit
4150 Try to unroll loops. This keyword may be specified together with
4151 any keyword above other than @code{NONE}. Loop unrolling
4152 usually, but not always, improves the performance of programs.
4155 Optimize space usage
4159 See also @ref{Optimization Levels}.
4163 @item -pass-exit-codes
4164 @cindex @option{-pass-exit-codes} (@command{gcc})
4165 Catch exit codes from the compiler and use the most meaningful as
4169 @item --RTS=@var{rts-path}
4170 @cindex @option{--RTS} (@command{gcc})
4171 Specifies the default location of the runtime library. Same meaning as the
4172 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
4175 @cindex @option{^-S^/ASM^} (@command{gcc})
4176 ^Used in place of @option{-c} to^Used to^
4177 cause the assembler source file to be
4178 generated, using @file{^.s^.S^} as the extension,
4179 instead of the object file.
4180 This may be useful if you need to examine the generated assembly code.
4182 @item ^-fverbose-asm^/VERBOSE_ASM^
4183 @cindex @option{^-fverbose-asm^/VERBOSE_ASM^} (@command{gcc})
4184 ^Used in conjunction with @option{-S}^Used in place of @option{/ASM}^
4185 to cause the generated assembly code file to be annotated with variable
4186 names, making it significantly easier to follow.
4189 @cindex @option{^-v^/VERBOSE^} (@command{gcc})
4190 Show commands generated by the @command{gcc} driver. Normally used only for
4191 debugging purposes or if you need to be sure what version of the
4192 compiler you are executing.
4196 @cindex @option{-V} (@command{gcc})
4197 Execute @var{ver} version of the compiler. This is the @command{gcc}
4198 version, not the GNAT version.
4201 @item ^-w^NO_BACK_END_WARNINGS^
4202 @cindex @option{-w} (@command{gcc})
4203 Turn off warnings generated by the back end of the compiler. Use of
4204 this switch also causes the default for front end warnings to be set
4205 to suppress (as though @option{-gnatws} had appeared at the start of
4211 @c Combining qualifiers does not work on VMS
4212 You may combine a sequence of GNAT switches into a single switch. For
4213 example, the combined switch
4215 @cindex Combining GNAT switches
4221 is equivalent to specifying the following sequence of switches:
4224 -gnato -gnatf -gnati3
4229 The following restrictions apply to the combination of switches
4234 The switch @option{-gnatc} if combined with other switches must come
4235 first in the string.
4238 The switch @option{-gnats} if combined with other switches must come
4239 first in the string.
4243 @option{^-gnatz^/DISTRIBUTION_STUBS^}, @option{-gnatzc}, and @option{-gnatzr}
4244 may not be combined with any other switches.
4248 Once a ``y'' appears in the string (that is a use of the @option{-gnaty}
4249 switch), then all further characters in the switch are interpreted
4250 as style modifiers (see description of @option{-gnaty}).
4253 Once a ``d'' appears in the string (that is a use of the @option{-gnatd}
4254 switch), then all further characters in the switch are interpreted
4255 as debug flags (see description of @option{-gnatd}).
4258 Once a ``w'' appears in the string (that is a use of the @option{-gnatw}
4259 switch), then all further characters in the switch are interpreted
4260 as warning mode modifiers (see description of @option{-gnatw}).
4263 Once a ``V'' appears in the string (that is a use of the @option{-gnatV}
4264 switch), then all further characters in the switch are interpreted
4265 as validity checking options (see description of @option{-gnatV}).
4269 @node Output and Error Message Control
4270 @subsection Output and Error Message Control
4274 The standard default format for error messages is called ``brief format''.
4275 Brief format messages are written to @file{stderr} (the standard error
4276 file) and have the following form:
4279 e.adb:3:04: Incorrect spelling of keyword "function"
4280 e.adb:4:20: ";" should be "is"
4284 The first integer after the file name is the line number in the file,
4285 and the second integer is the column number within the line.
4287 @code{GPS} can parse the error messages
4288 and point to the referenced character.
4290 The following switches provide control over the error message
4296 @cindex @option{-gnatv} (@command{gcc})
4299 The v stands for verbose.
4301 The effect of this setting is to write long-format error
4302 messages to @file{stdout} (the standard output file.
4303 The same program compiled with the
4304 @option{-gnatv} switch would generate:
4308 3. funcion X (Q : Integer)
4310 >>> Incorrect spelling of keyword "function"
4313 >>> ";" should be "is"
4318 The vertical bar indicates the location of the error, and the @samp{>>>}
4319 prefix can be used to search for error messages. When this switch is
4320 used the only source lines output are those with errors.
4323 @cindex @option{-gnatl} (@command{gcc})
4325 The @code{l} stands for list.
4327 This switch causes a full listing of
4328 the file to be generated. In the case where a body is
4329 compiled, the corresponding spec is also listed, along
4330 with any subunits. Typical output from compiling a package
4331 body @file{p.adb} might look like:
4333 @smallexample @c ada
4337 1. package body p is
4339 3. procedure a is separate;
4350 2. pragma Elaborate_Body
4374 When you specify the @option{-gnatv} or @option{-gnatl} switches and
4375 standard output is redirected, a brief summary is written to
4376 @file{stderr} (standard error) giving the number of error messages and
4377 warning messages generated.
4379 @item -^gnatl^OUTPUT_FILE^=file
4380 @cindex @option{^-gnatl^OUTPUT_FILE^=fname} (@command{gcc})
4381 This has the same effect as @code{-gnatl} except that the output is
4382 written to a file instead of to standard output. If the given name
4383 @file{fname} does not start with a period, then it is the full name
4384 of the file to be written. If @file{fname} is an extension, it is
4385 appended to the name of the file being compiled. For example, if
4386 file @file{xyz.adb} is compiled with @option{^-gnatl^OUTPUT_FILE^=.lst},
4387 then the output is written to file ^xyz.adb.lst^xyz.adb_lst^.
4390 @cindex @option{-gnatU} (@command{gcc})
4391 This switch forces all error messages to be preceded by the unique
4392 string ``error:''. This means that error messages take a few more
4393 characters in space, but allows easy searching for and identification
4397 @cindex @option{-gnatb} (@command{gcc})
4399 The @code{b} stands for brief.
4401 This switch causes GNAT to generate the
4402 brief format error messages to @file{stderr} (the standard error
4403 file) as well as the verbose
4404 format message or full listing (which as usual is written to
4405 @file{stdout} (the standard output file).
4407 @item -gnatm=@var{n}
4408 @cindex @option{-gnatm} (@command{gcc})
4410 The @code{m} stands for maximum.
4412 @var{n} is a decimal integer in the
4413 range of 1 to 999 and limits the number of error messages to be
4414 generated. For example, using @option{-gnatm2} might yield
4417 e.adb:3:04: Incorrect spelling of keyword "function"
4418 e.adb:5:35: missing ".."
4419 fatal error: maximum errors reached
4420 compilation abandoned
4424 Note that the equal sign is optional, so the switches
4425 @option{-gnatm2} and @option{-gnatm=2} are equivalent.
4428 @cindex @option{-gnatf} (@command{gcc})
4429 @cindex Error messages, suppressing
4431 The @code{f} stands for full.
4433 Normally, the compiler suppresses error messages that are likely to be
4434 redundant. This switch causes all error
4435 messages to be generated. In particular, in the case of
4436 references to undefined variables. If a given variable is referenced
4437 several times, the normal format of messages is
4439 e.adb:7:07: "V" is undefined (more references follow)
4443 where the parenthetical comment warns that there are additional
4444 references to the variable @code{V}. Compiling the same program with the
4445 @option{-gnatf} switch yields
4448 e.adb:7:07: "V" is undefined
4449 e.adb:8:07: "V" is undefined
4450 e.adb:8:12: "V" is undefined
4451 e.adb:8:16: "V" is undefined
4452 e.adb:9:07: "V" is undefined
4453 e.adb:9:12: "V" is undefined
4457 The @option{-gnatf} switch also generates additional information for
4458 some error messages. Some examples are:
4462 Full details on entities not available in high integrity mode
4464 Details on possibly non-portable unchecked conversion
4466 List possible interpretations for ambiguous calls
4468 Additional details on incorrect parameters
4472 @cindex @option{-gnatjnn} (@command{gcc})
4473 In normal operation mode (or if @option{-gnatj0} is used, then error messages
4474 with continuation lines are treated as though the continuation lines were
4475 separate messages (and so a warning with two continuation lines counts as
4476 three warnings, and is listed as three separate messages).
4478 If the @option{-gnatjnn} switch is used with a positive value for nn, then
4479 messages are output in a different manner. A message and all its continuation
4480 lines are treated as a unit, and count as only one warning or message in the
4481 statistics totals. Furthermore, the message is reformatted so that no line
4482 is longer than nn characters.
4485 @cindex @option{-gnatq} (@command{gcc})
4487 The @code{q} stands for quit (really ``don't quit'').
4489 In normal operation mode, the compiler first parses the program and
4490 determines if there are any syntax errors. If there are, appropriate
4491 error messages are generated and compilation is immediately terminated.
4493 GNAT to continue with semantic analysis even if syntax errors have been
4494 found. This may enable the detection of more errors in a single run. On
4495 the other hand, the semantic analyzer is more likely to encounter some
4496 internal fatal error when given a syntactically invalid tree.
4499 @cindex @option{-gnatQ} (@command{gcc})
4500 In normal operation mode, the @file{ALI} file is not generated if any
4501 illegalities are detected in the program. The use of @option{-gnatQ} forces
4502 generation of the @file{ALI} file. This file is marked as being in
4503 error, so it cannot be used for binding purposes, but it does contain
4504 reasonably complete cross-reference information, and thus may be useful
4505 for use by tools (e.g. semantic browsing tools or integrated development
4506 environments) that are driven from the @file{ALI} file. This switch
4507 implies @option{-gnatq}, since the semantic phase must be run to get a
4508 meaningful ALI file.
4510 In addition, if @option{-gnatt} is also specified, then the tree file is
4511 generated even if there are illegalities. It may be useful in this case
4512 to also specify @option{-gnatq} to ensure that full semantic processing
4513 occurs. The resulting tree file can be processed by ASIS, for the purpose
4514 of providing partial information about illegal units, but if the error
4515 causes the tree to be badly malformed, then ASIS may crash during the
4518 When @option{-gnatQ} is used and the generated @file{ALI} file is marked as
4519 being in error, @command{gnatmake} will attempt to recompile the source when it
4520 finds such an @file{ALI} file, including with switch @option{-gnatc}.
4522 Note that @option{-gnatQ} has no effect if @option{-gnats} is specified,
4523 since ALI files are never generated if @option{-gnats} is set.
4527 @node Warning Message Control
4528 @subsection Warning Message Control
4529 @cindex Warning messages
4531 In addition to error messages, which correspond to illegalities as defined
4532 in the Ada Reference Manual, the compiler detects two kinds of warning
4535 First, the compiler considers some constructs suspicious and generates a
4536 warning message to alert you to a possible error. Second, if the
4537 compiler detects a situation that is sure to raise an exception at
4538 run time, it generates a warning message. The following shows an example
4539 of warning messages:
4541 e.adb:4:24: warning: creation of object may raise Storage_Error
4542 e.adb:10:17: warning: static value out of range
4543 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
4547 GNAT considers a large number of situations as appropriate
4548 for the generation of warning messages. As always, warnings are not
4549 definite indications of errors. For example, if you do an out-of-range
4550 assignment with the deliberate intention of raising a
4551 @code{Constraint_Error} exception, then the warning that may be
4552 issued does not indicate an error. Some of the situations for which GNAT
4553 issues warnings (at least some of the time) are given in the following
4554 list. This list is not complete, and new warnings are often added to
4555 subsequent versions of GNAT. The list is intended to give a general idea
4556 of the kinds of warnings that are generated.
4560 Possible infinitely recursive calls
4563 Out-of-range values being assigned
4566 Possible order of elaboration problems
4572 Address clauses with possibly unaligned values, or where an attempt is
4573 made to overlay a smaller variable with a larger one.
4576 Fixed-point type declarations with a null range
4579 Direct_IO or Sequential_IO instantiated with a type that has access values
4582 Variables that are never assigned a value
4585 Variables that are referenced before being initialized
4588 Task entries with no corresponding @code{accept} statement
4591 Duplicate accepts for the same task entry in a @code{select}
4594 Objects that take too much storage
4597 Unchecked conversion between types of differing sizes
4600 Missing @code{return} statement along some execution path in a function
4603 Incorrect (unrecognized) pragmas
4606 Incorrect external names
4609 Allocation from empty storage pool
4612 Potentially blocking operation in protected type
4615 Suspicious parenthesization of expressions
4618 Mismatching bounds in an aggregate
4621 Attempt to return local value by reference
4624 Premature instantiation of a generic body
4627 Attempt to pack aliased components
4630 Out of bounds array subscripts
4633 Wrong length on string assignment
4636 Violations of style rules if style checking is enabled
4639 Unused @code{with} clauses
4642 @code{Bit_Order} usage that does not have any effect
4645 @code{Standard.Duration} used to resolve universal fixed expression
4648 Dereference of possibly null value
4651 Declaration that is likely to cause storage error
4654 Internal GNAT unit @code{with}'ed by application unit
4657 Values known to be out of range at compile time
4660 Unreferenced labels and variables
4663 Address overlays that could clobber memory
4666 Unexpected initialization when address clause present
4669 Bad alignment for address clause
4672 Useless type conversions
4675 Redundant assignment statements and other redundant constructs
4678 Useless exception handlers
4681 Accidental hiding of name by child unit
4684 Access before elaboration detected at compile time
4687 A range in a @code{for} loop that is known to be null or might be null
4692 The following section lists compiler switches that are available
4693 to control the handling of warning messages. It is also possible
4694 to exercise much finer control over what warnings are issued and
4695 suppressed using the GNAT pragma Warnings, which is documented
4696 in the GNAT Reference manual.
4701 @emph{Activate all optional errors.}
4702 @cindex @option{-gnatwa} (@command{gcc})
4703 This switch activates most optional warning messages, see remaining list
4704 in this section for details on optional warning messages that can be
4705 individually controlled. The warnings that are not turned on by this
4707 @option{-gnatwd} (implicit dereferencing),
4708 @option{-gnatwh} (hiding),
4709 @option{-gnatwl} (elaboration warnings),
4710 and @option{-gnatwt} (tracking of deleted conditional code).
4711 All other optional warnings are turned on.
4714 @emph{Suppress all optional errors.}
4715 @cindex @option{-gnatwA} (@command{gcc})
4716 This switch suppresses all optional warning messages, see remaining list
4717 in this section for details on optional warning messages that can be
4718 individually controlled.
4721 @emph{Activate warnings on bad fixed values.}
4722 @cindex @option{-gnatwb} (@command{gcc})
4723 @cindex Bad fixed values
4724 @cindex Fixed-point Small value
4726 This switch activates warnings for static fixed-point expressions whose
4727 value is not an exact multiple of Small. Such values are implementation
4728 dependent, since an implementation is free to choose either of the multiples
4729 that surround the value. GNAT always chooses the closer one, but this is not
4730 required behavior, and it is better to specify a value that is an exact
4731 multiple, ensuring predictable execution. The default is that such warnings
4735 @emph{Suppress warnings on bad fixed values.}
4736 @cindex @option{-gnatwB} (@command{gcc})
4737 This switch suppresses warnings for static fixed-point expressions whose
4738 value is not an exact multiple of Small.
4741 @emph{Activate warnings on conditionals.}
4742 @cindex @option{-gnatwc} (@command{gcc})
4743 @cindex Conditionals, constant
4744 This switch activates warnings for conditional expressions used in
4745 tests that are known to be True or False at compile time. The default
4746 is that such warnings are not generated.
4747 Note that this warning does
4748 not get issued for the use of boolean variables or constants whose
4749 values are known at compile time, since this is a standard technique
4750 for conditional compilation in Ada, and this would generate too many
4751 false positive warnings.
4753 This warning option also activates a special test for comparisons using
4754 the operators ``>='' and`` <=''.
4755 If the compiler can tell that only the equality condition is possible,
4756 then it will warn that the ``>'' or ``<'' part of the test
4757 is useless and that the operator could be replaced by ``=''.
4758 An example would be comparing a @code{Natural} variable <= 0.
4760 This warning can also be turned on using @option{-gnatwa}.
4763 @emph{Suppress warnings on conditionals.}
4764 @cindex @option{-gnatwC} (@command{gcc})
4765 This switch suppresses warnings for conditional expressions used in
4766 tests that are known to be True or False at compile time.
4769 @emph{Activate warnings on missing component clauses.}
4770 @cindex @option{-gnatw.c} (@command{gcc})
4771 @cindex Component clause, missing
4772 This switch activates warnings for record components where a record
4773 representation clause is present and has component clauses for the
4774 majority, but not all, of the components. A warning is given for each
4775 component for which no component clause is present.
4777 This warning can also be turned on using @option{-gnatwa}.
4780 @emph{Suppress warnings on missing component clauses.}
4781 @cindex @option{-gnatwC} (@command{gcc})
4782 This switch suppresses warnings for record components that are
4783 missing a component clause in the situation described above.
4786 @emph{Activate warnings on implicit dereferencing.}
4787 @cindex @option{-gnatwd} (@command{gcc})
4788 If this switch is set, then the use of a prefix of an access type
4789 in an indexed component, slice, or selected component without an
4790 explicit @code{.all} will generate a warning. With this warning
4791 enabled, access checks occur only at points where an explicit
4792 @code{.all} appears in the source code (assuming no warnings are
4793 generated as a result of this switch). The default is that such
4794 warnings are not generated.
4795 Note that @option{-gnatwa} does not affect the setting of
4796 this warning option.
4799 @emph{Suppress warnings on implicit dereferencing.}
4800 @cindex @option{-gnatwD} (@command{gcc})
4801 @cindex Implicit dereferencing
4802 @cindex Dereferencing, implicit
4803 This switch suppresses warnings for implicit dereferences in
4804 indexed components, slices, and selected components.
4807 @emph{Treat warnings as errors.}
4808 @cindex @option{-gnatwe} (@command{gcc})
4809 @cindex Warnings, treat as error
4810 This switch causes warning messages to be treated as errors.
4811 The warning string still appears, but the warning messages are counted
4812 as errors, and prevent the generation of an object file.
4815 @emph{Activate warnings on unreferenced formals.}
4816 @cindex @option{-gnatwf} (@command{gcc})
4817 @cindex Formals, unreferenced
4818 This switch causes a warning to be generated if a formal parameter
4819 is not referenced in the body of the subprogram. This warning can
4820 also be turned on using @option{-gnatwa} or @option{-gnatwu}. The
4821 default is that these warnings are not generated.
4824 @emph{Suppress warnings on unreferenced formals.}
4825 @cindex @option{-gnatwF} (@command{gcc})
4826 This switch suppresses warnings for unreferenced formal
4827 parameters. Note that the
4828 combination @option{-gnatwu} followed by @option{-gnatwF} has the
4829 effect of warning on unreferenced entities other than subprogram
4833 @emph{Activate warnings on unrecognized pragmas.}
4834 @cindex @option{-gnatwg} (@command{gcc})
4835 @cindex Pragmas, unrecognized
4836 This switch causes a warning to be generated if an unrecognized
4837 pragma is encountered. Apart from issuing this warning, the
4838 pragma is ignored and has no effect. This warning can
4839 also be turned on using @option{-gnatwa}. The default
4840 is that such warnings are issued (satisfying the Ada Reference
4841 Manual requirement that such warnings appear).
4844 @emph{Suppress warnings on unrecognized pragmas.}
4845 @cindex @option{-gnatwG} (@command{gcc})
4846 This switch suppresses warnings for unrecognized pragmas.
4849 @emph{Activate warnings on hiding.}
4850 @cindex @option{-gnatwh} (@command{gcc})
4851 @cindex Hiding of Declarations
4852 This switch activates warnings on hiding declarations.
4853 A declaration is considered hiding
4854 if it is for a non-overloadable entity, and it declares an entity with the
4855 same name as some other entity that is directly or use-visible. The default
4856 is that such warnings are not generated.
4857 Note that @option{-gnatwa} does not affect the setting of this warning option.
4860 @emph{Suppress warnings on hiding.}
4861 @cindex @option{-gnatwH} (@command{gcc})
4862 This switch suppresses warnings on hiding declarations.
4865 @emph{Activate warnings on implementation units.}
4866 @cindex @option{-gnatwi} (@command{gcc})
4867 This switch activates warnings for a @code{with} of an internal GNAT
4868 implementation unit, defined as any unit from the @code{Ada},
4869 @code{Interfaces}, @code{GNAT},
4870 ^^@code{DEC},^ or @code{System}
4871 hierarchies that is not
4872 documented in either the Ada Reference Manual or the GNAT
4873 Programmer's Reference Manual. Such units are intended only
4874 for internal implementation purposes and should not be @code{with}'ed
4875 by user programs. The default is that such warnings are generated
4876 This warning can also be turned on using @option{-gnatwa}.
4879 @emph{Disable warnings on implementation units.}
4880 @cindex @option{-gnatwI} (@command{gcc})
4881 This switch disables warnings for a @code{with} of an internal GNAT
4882 implementation unit.
4885 @emph{Activate warnings on obsolescent features (Annex J).}
4886 @cindex @option{-gnatwj} (@command{gcc})
4887 @cindex Features, obsolescent
4888 @cindex Obsolescent features
4889 If this warning option is activated, then warnings are generated for
4890 calls to subprograms marked with @code{pragma Obsolescent} and
4891 for use of features in Annex J of the Ada Reference Manual. In the
4892 case of Annex J, not all features are flagged. In particular use
4893 of the renamed packages (like @code{Text_IO}) and use of package
4894 @code{ASCII} are not flagged, since these are very common and
4895 would generate many annoying positive warnings. The default is that
4896 such warnings are not generated. This warning is also turned on by
4897 the use of @option{-gnatwa}.
4899 In addition to the above cases, warnings are also generated for
4900 GNAT features that have been provided in past versions but which
4901 have been superseded (typically by features in the new Ada standard).
4902 For example, @code{pragma Ravenscar} will be flagged since its
4903 function is replaced by @code{pragma Profile(Ravenscar)}.
4905 Note that this warning option functions differently from the
4906 restriction @code{No_Obsolescent_Features} in two respects.
4907 First, the restriction applies only to annex J features.
4908 Second, the restriction does flag uses of package @code{ASCII}.
4911 @emph{Suppress warnings on obsolescent features (Annex J).}
4912 @cindex @option{-gnatwJ} (@command{gcc})
4913 This switch disables warnings on use of obsolescent features.
4916 @emph{Activate warnings on variables that could be constants.}
4917 @cindex @option{-gnatwk} (@command{gcc})
4918 This switch activates warnings for variables that are initialized but
4919 never modified, and then could be declared constants. The default is that
4920 such warnings are not given.
4921 This warning can also be turned on using @option{-gnatwa}.
4924 @emph{Suppress warnings on variables that could be constants.}
4925 @cindex @option{-gnatwK} (@command{gcc})
4926 This switch disables warnings on variables that could be declared constants.
4929 @emph{Activate warnings for missing elaboration pragmas.}
4930 @cindex @option{-gnatwl} (@command{gcc})
4931 @cindex Elaboration, warnings
4932 This switch activates warnings on missing
4933 @code{Elaborate_All} and @code{Elaborate} pragmas.
4934 See the section in this guide on elaboration checking for details on
4935 when such pragmas should be used. Warnings are also generated if you
4936 are using the static mode of elaboration, and a @code{pragma Elaborate}
4937 is encountered. The default is that such warnings
4939 This warning is not automatically turned on by the use of @option{-gnatwa}.
4942 @emph{Suppress warnings for missing elaboration pragmas.}
4943 @cindex @option{-gnatwL} (@command{gcc})
4944 This switch suppresses warnings on missing Elaborate and Elaborate_All pragmas.
4945 See the section in this guide on elaboration checking for details on
4946 when such pragmas should be used.
4949 @emph{Activate warnings on modified but unreferenced variables.}
4950 @cindex @option{-gnatwm} (@command{gcc})
4951 This switch activates warnings for variables that are assigned (using
4952 an initialization value or with one or more assignment statements) but
4953 whose value is never read. The warning is suppressed for volatile
4954 variables and also for variables that are renamings of other variables
4955 or for which an address clause is given.
4956 This warning can also be turned on using @option{-gnatwa}.
4957 The default is that these warnings are not given.
4960 @emph{Disable warnings on modified but unreferenced variables.}
4961 @cindex @option{-gnatwM} (@command{gcc})
4962 This switch disables warnings for variables that are assigned or
4963 initialized, but never read.
4966 @emph{Set normal warnings mode.}
4967 @cindex @option{-gnatwn} (@command{gcc})
4968 This switch sets normal warning mode, in which enabled warnings are
4969 issued and treated as warnings rather than errors. This is the default
4970 mode. the switch @option{-gnatwn} can be used to cancel the effect of
4971 an explicit @option{-gnatws} or
4972 @option{-gnatwe}. It also cancels the effect of the
4973 implicit @option{-gnatwe} that is activated by the
4974 use of @option{-gnatg}.
4977 @emph{Activate warnings on address clause overlays.}
4978 @cindex @option{-gnatwo} (@command{gcc})
4979 @cindex Address Clauses, warnings
4980 This switch activates warnings for possibly unintended initialization
4981 effects of defining address clauses that cause one variable to overlap
4982 another. The default is that such warnings are generated.
4983 This warning can also be turned on using @option{-gnatwa}.
4986 @emph{Suppress warnings on address clause overlays.}
4987 @cindex @option{-gnatwO} (@command{gcc})
4988 This switch suppresses warnings on possibly unintended initialization
4989 effects of defining address clauses that cause one variable to overlap
4993 @emph{Activate warnings on ineffective pragma Inlines.}
4994 @cindex @option{-gnatwp} (@command{gcc})
4995 @cindex Inlining, warnings
4996 This switch activates warnings for failure of front end inlining
4997 (activated by @option{-gnatN}) to inline a particular call. There are
4998 many reasons for not being able to inline a call, including most
4999 commonly that the call is too complex to inline. The default is
5000 that such warnings are not given.
5001 This warning can also be turned on using @option{-gnatwa}.
5002 Warnings on ineffective inlining by the gcc back-end can be activated
5003 separately, using the gcc switch -Winline.
5006 @emph{Suppress warnings on ineffective pragma Inlines.}
5007 @cindex @option{-gnatwP} (@command{gcc})
5008 This switch suppresses warnings on ineffective pragma Inlines. If the
5009 inlining mechanism cannot inline a call, it will simply ignore the
5013 @emph{Activate warnings on questionable missing parentheses.}
5014 @cindex @option{-gnatwq} (@command{gcc})
5015 @cindex Parentheses, warnings
5016 This switch activates warnings for cases where parentheses are not used and
5017 the result is potential ambiguity from a readers point of view. For example
5018 (not a > b) when a and b are modular means ((not a) > b) and very likely the
5019 programmer intended (not (a > b)). Similarly (-x mod 5) means (-(x mod 5)) and
5020 quite likely ((-x) mod 5) was intended. In such situations it seems best to
5021 follow the rule of always parenthesizing to make the association clear, and
5022 this warning switch warns if such parentheses are not present. The default
5023 is that these warnings are given.
5024 This warning can also be turned on using @option{-gnatwa}.
5027 @emph{Suppress warnings on questionable missing parentheses.}
5028 @cindex @option{-gnatwQ} (@command{gcc})
5029 This switch suppresses warnings for cases where the association is not
5030 clear and the use of parentheses is preferred.
5033 @emph{Activate warnings on redundant constructs.}
5034 @cindex @option{-gnatwr} (@command{gcc})
5035 This switch activates warnings for redundant constructs. The following
5036 is the current list of constructs regarded as redundant:
5040 Assignment of an item to itself.
5042 Type conversion that converts an expression to its own type.
5044 Use of the attribute @code{Base} where @code{typ'Base} is the same
5047 Use of pragma @code{Pack} when all components are placed by a record
5048 representation clause.
5050 Exception handler containing only a reraise statement (raise with no
5051 operand) which has no effect.
5053 Use of the operator abs on an operand that is known at compile time
5056 Comparison of boolean expressions to an explicit True value.
5059 This warning can also be turned on using @option{-gnatwa}.
5060 The default is that warnings for redundant constructs are not given.
5063 @emph{Suppress warnings on redundant constructs.}
5064 @cindex @option{-gnatwR} (@command{gcc})
5065 This switch suppresses warnings for redundant constructs.
5068 @emph{Suppress all warnings.}
5069 @cindex @option{-gnatws} (@command{gcc})
5070 This switch completely suppresses the
5071 output of all warning messages from the GNAT front end.
5072 Note that it does not suppress warnings from the @command{gcc} back end.
5073 To suppress these back end warnings as well, use the switch @option{-w}
5074 in addition to @option{-gnatws}.
5077 @emph{Activate warnings for tracking of deleted conditional code.}
5078 @cindex @option{-gnatwt} (@command{gcc})
5079 @cindex Deactivated code, warnings
5080 @cindex Deleted code, warnings
5081 This switch activates warnings for tracking of code in conditionals (IF and
5082 CASE statements) that is detected to be dead code which cannot be executed, and
5083 which is removed by the front end. This warning is off by default, and is not
5084 turned on by @option{-gnatwa}, it has to be turned on explicitly. This may be
5085 useful for detecting deactivated code in certified applications.
5088 @emph{Suppress warnings for tracking of deleted conditional code.}
5089 @cindex @option{-gnatwT} (@command{gcc})
5090 This switch suppresses warnings for tracking of deleted conditional code.
5093 @emph{Activate warnings on unused entities.}
5094 @cindex @option{-gnatwu} (@command{gcc})
5095 This switch activates warnings to be generated for entities that
5096 are declared but not referenced, and for units that are @code{with}'ed
5098 referenced. In the case of packages, a warning is also generated if
5099 no entities in the package are referenced. This means that if the package
5100 is referenced but the only references are in @code{use}
5101 clauses or @code{renames}
5102 declarations, a warning is still generated. A warning is also generated
5103 for a generic package that is @code{with}'ed but never instantiated.
5104 In the case where a package or subprogram body is compiled, and there
5105 is a @code{with} on the corresponding spec
5106 that is only referenced in the body,
5107 a warning is also generated, noting that the
5108 @code{with} can be moved to the body. The default is that
5109 such warnings are not generated.
5110 This switch also activates warnings on unreferenced formals
5111 (it includes the effect of @option{-gnatwf}).
5112 This warning can also be turned on using @option{-gnatwa}.
5115 @emph{Suppress warnings on unused entities.}
5116 @cindex @option{-gnatwU} (@command{gcc})
5117 This switch suppresses warnings for unused entities and packages.
5118 It also turns off warnings on unreferenced formals (and thus includes
5119 the effect of @option{-gnatwF}).
5122 @emph{Activate warnings on unassigned variables.}
5123 @cindex @option{-gnatwv} (@command{gcc})
5124 @cindex Unassigned variable warnings
5125 This switch activates warnings for access to variables which
5126 may not be properly initialized. The default is that
5127 such warnings are generated.
5128 This warning can also be turned on using @option{-gnatwa}.
5131 @emph{Suppress warnings on unassigned variables.}
5132 @cindex @option{-gnatwV} (@command{gcc})
5133 This switch suppresses warnings for access to variables which
5134 may not be properly initialized.
5135 For variables of a composite type, the warning can also be suppressed in
5136 Ada 2005 by using a default initialization with a box. For example, if
5137 Table is an array of records whose components are only partially uninitialized,
5138 then the following code:
5140 @smallexample @c ada
5141 Tab : Table := (others => <>);
5144 will suppress warnings on subsequent statements that access components
5148 @emph{Activate warnings on wrong low bound assumption.}
5149 @cindex @option{-gnatww} (@command{gcc})
5150 @cindex String indexing warnings
5151 This switch activates warnings for indexing an unconstrained string parameter
5152 with a literal or S'Length. This is a case where the code is assuming that the
5153 low bound is one, which is in general not true (for example when a slice is
5154 passed). The default is that such warnings are generated.
5155 This warning can also be turned on using @option{-gnatwa}.
5158 @emph{Suppress warnings on wrong low bound assumption.}
5159 @cindex @option{-gnatwW} (@command{gcc})
5160 This switch activates warnings for indexing an unconstrained string parameter
5161 with a literal or S'Length. This warning can also be suppressed by providing
5162 an Assert pragma that checks the low bound, for example:
5164 @smallexample @c ada
5165 procedure K (S : String) is
5166 pragma Assert (S'First = 1);
5171 @emph{Activate warnings on Export/Import pragmas.}
5172 @cindex @option{-gnatwx} (@command{gcc})
5173 @cindex Export/Import pragma warnings
5174 This switch activates warnings on Export/Import pragmas when
5175 the compiler detects a possible conflict between the Ada and
5176 foreign language calling sequences. For example, the use of
5177 default parameters in a convention C procedure is dubious
5178 because the C compiler cannot supply the proper default, so
5179 a warning is issued. The default is that such warnings are
5181 This warning can also be turned on using @option{-gnatwa}.
5184 @emph{Suppress warnings on Export/Import pragmas.}
5185 @cindex @option{-gnatwX} (@command{gcc})
5186 This switch suppresses warnings on Export/Import pragmas.
5187 The sense of this is that you are telling the compiler that
5188 you know what you are doing in writing the pragma, and it
5189 should not complain at you.
5192 @emph{Activate warnings for No_Exception_Propagation mode.}
5193 @cindex @option{-gnatwm} (@command{gcc})
5194 This switch activates warnings for exception usage when pragma Restrictions
5195 (No_Exception_Propagation) is in effect. Warnings are given for implicit or
5196 explicit exception raises which are not covered by a local handler, and for
5197 exception handlers which do not cover a local raise. The default is that these
5201 @emph{Disable warnings for No_Exception_Propagation mode.}
5202 This switch disables warnings for exception usage when pragma Restrictions
5203 (No_Exception_Propagation) is in effect.
5206 @emph{Activate warnings for Ada 2005 compatibility issues.}
5207 @cindex @option{-gnatwy} (@command{gcc})
5208 @cindex Ada 2005 compatibility issues warnings
5209 For the most part Ada 2005 is upwards compatible with Ada 95,
5210 but there are some exceptions (for example the fact that
5211 @code{interface} is now a reserved word in Ada 2005). This
5212 switch activates several warnings to help in identifying
5213 and correcting such incompatibilities. The default is that
5214 these warnings are generated. Note that at one point Ada 2005
5215 was called Ada 0Y, hence the choice of character.
5216 This warning can also be turned on using @option{-gnatwa}.
5219 @emph{Disable warnings for Ada 2005 compatibility issues.}
5220 @cindex @option{-gnatwY} (@command{gcc})
5221 @cindex Ada 2005 compatibility issues warnings
5222 This switch suppresses several warnings intended to help in identifying
5223 incompatibilities between Ada 95 and Ada 2005.
5226 @emph{Activate warnings on unchecked conversions.}
5227 @cindex @option{-gnatwz} (@command{gcc})
5228 @cindex Unchecked_Conversion warnings
5229 This switch activates warnings for unchecked conversions
5230 where the types are known at compile time to have different
5232 is that such warnings are generated.
5233 This warning can also be turned on using @option{-gnatwa}.
5236 @emph{Suppress warnings on unchecked conversions.}
5237 @cindex @option{-gnatwZ} (@command{gcc})
5238 This switch suppresses warnings for unchecked conversions
5239 where the types are known at compile time to have different
5242 @item ^-Wuninitialized^WARNINGS=UNINITIALIZED^
5243 @cindex @option{-Wuninitialized}
5244 The warnings controlled by the @option{-gnatw} switch are generated by the
5245 front end of the compiler. In some cases, the @option{^gcc^GCC^} back end
5246 can provide additional warnings. One such useful warning is provided by
5247 @option{^-Wuninitialized^WARNINGS=UNINITIALIZED^}. This must be used in
5248 conjunction with turning on optimization mode. This causes the flow
5249 analysis circuits of the back end optimizer to output additional
5250 warnings about uninitialized variables.
5252 @item ^-w^/NO_BACK_END_WARNINGS^
5254 This switch suppresses warnings from the @option{^gcc^GCC^} back end. The
5255 code generator detects a number of warning situations that are missed
5256 by the @option{GNAT} front end, and this switch can be used to suppress them.
5257 The use of this switch also sets the default front end warning mode to
5258 @option{-gnatws}, that is, front end warnings suppressed as well.
5264 A string of warning parameters can be used in the same parameter. For example:
5271 will turn on all optional warnings except for elaboration pragma warnings,
5272 and also specify that warnings should be treated as errors.
5274 When no switch @option{^-gnatw^/WARNINGS^} is used, this is equivalent to:
5299 @node Debugging and Assertion Control
5300 @subsection Debugging and Assertion Control
5304 @cindex @option{-gnata} (@command{gcc})
5310 The pragmas @code{Assert} and @code{Debug} normally have no effect and
5311 are ignored. This switch, where @samp{a} stands for assert, causes
5312 @code{Assert} and @code{Debug} pragmas to be activated.
5314 The pragmas have the form:
5318 @b{pragma} Assert (@var{Boolean-expression} [,
5319 @var{static-string-expression}])
5320 @b{pragma} Debug (@var{procedure call})
5325 The @code{Assert} pragma causes @var{Boolean-expression} to be tested.
5326 If the result is @code{True}, the pragma has no effect (other than
5327 possible side effects from evaluating the expression). If the result is
5328 @code{False}, the exception @code{Assert_Failure} declared in the package
5329 @code{System.Assertions} is
5330 raised (passing @var{static-string-expression}, if present, as the
5331 message associated with the exception). If no string expression is
5332 given the default is a string giving the file name and line number
5335 The @code{Debug} pragma causes @var{procedure} to be called. Note that
5336 @code{pragma Debug} may appear within a declaration sequence, allowing
5337 debugging procedures to be called between declarations.
5340 @item /DEBUG[=debug-level]
5342 Specifies how much debugging information is to be included in
5343 the resulting object file where 'debug-level' is one of the following:
5346 Include both debugger symbol records and traceback
5348 This is the default setting.
5350 Include both debugger symbol records and traceback in
5353 Excludes both debugger symbol records and traceback
5354 the object file. Same as /NODEBUG.
5356 Includes only debugger symbol records in the object
5357 file. Note that this doesn't include traceback information.
5362 @node Validity Checking
5363 @subsection Validity Checking
5364 @findex Validity Checking
5367 The Ada Reference Manual has specific requirements for checking
5368 for invalid values. In particular, RM 13.9.1 requires that the
5369 evaluation of invalid values (for example from unchecked conversions),
5370 not result in erroneous execution. In GNAT, the result of such an
5371 evaluation in normal default mode is to either use the value
5372 unmodified, or to raise Constraint_Error in those cases where use
5373 of the unmodified value would cause erroneous execution. The cases
5374 where unmodified values might lead to erroneous execution are case
5375 statements (where a wild jump might result from an invalid value),
5376 and subscripts on the left hand side (where memory corruption could
5377 occur as a result of an invalid value).
5379 The @option{-gnatV^@var{x}^^} switch allows more control over the validity
5382 The @code{x} argument is a string of letters that
5383 indicate validity checks that are performed or not performed in addition
5384 to the default checks described above.
5387 The options allowed for this qualifier
5388 indicate validity checks that are performed or not performed in addition
5389 to the default checks described above.
5395 @emph{All validity checks.}
5396 @cindex @option{-gnatVa} (@command{gcc})
5397 All validity checks are turned on.
5399 That is, @option{-gnatVa} is
5400 equivalent to @option{gnatVcdfimorst}.
5404 @emph{Validity checks for copies.}
5405 @cindex @option{-gnatVc} (@command{gcc})
5406 The right hand side of assignments, and the initializing values of
5407 object declarations are validity checked.
5410 @emph{Default (RM) validity checks.}
5411 @cindex @option{-gnatVd} (@command{gcc})
5412 Some validity checks are done by default following normal Ada semantics
5414 A check is done in case statements that the expression is within the range
5415 of the subtype. If it is not, Constraint_Error is raised.
5416 For assignments to array components, a check is done that the expression used
5417 as index is within the range. If it is not, Constraint_Error is raised.
5418 Both these validity checks may be turned off using switch @option{-gnatVD}.
5419 They are turned on by default. If @option{-gnatVD} is specified, a subsequent
5420 switch @option{-gnatVd} will leave the checks turned on.
5421 Switch @option{-gnatVD} should be used only if you are sure that all such
5422 expressions have valid values. If you use this switch and invalid values
5423 are present, then the program is erroneous, and wild jumps or memory
5424 overwriting may occur.
5427 @emph{Validity checks for elementary components.}
5428 @cindex @option{-gnatVe} (@command{gcc})
5429 In the absence of this switch, assignments to record or array components are
5430 not validity checked, even if validity checks for assignments generally
5431 (@option{-gnatVc}) are turned on. In Ada, assignment of composite values do not
5432 require valid data, but assignment of individual components does. So for
5433 example, there is a difference between copying the elements of an array with a
5434 slice assignment, compared to assigning element by element in a loop. This
5435 switch allows you to turn off validity checking for components, even when they
5436 are assigned component by component.
5439 @emph{Validity checks for floating-point values.}
5440 @cindex @option{-gnatVf} (@command{gcc})
5441 In the absence of this switch, validity checking occurs only for discrete
5442 values. If @option{-gnatVf} is specified, then validity checking also applies
5443 for floating-point values, and NaN's and infinities are considered invalid,
5444 as well as out of range values for constrained types. Note that this means
5445 that standard @code{IEEE} infinity mode is not allowed. The exact contexts
5446 in which floating-point values are checked depends on the setting of other
5447 options. For example,
5448 @option{^-gnatVif^VALIDITY_CHECKING=(IN_PARAMS,FLOATS)^} or
5449 @option{^-gnatVfi^VALIDITY_CHECKING=(FLOATS,IN_PARAMS)^}
5450 (the order does not matter) specifies that floating-point parameters of mode
5451 @code{in} should be validity checked.
5454 @emph{Validity checks for @code{in} mode parameters}
5455 @cindex @option{-gnatVi} (@command{gcc})
5456 Arguments for parameters of mode @code{in} are validity checked in function
5457 and procedure calls at the point of call.
5460 @emph{Validity checks for @code{in out} mode parameters.}
5461 @cindex @option{-gnatVm} (@command{gcc})
5462 Arguments for parameters of mode @code{in out} are validity checked in
5463 procedure calls at the point of call. The @code{'m'} here stands for
5464 modify, since this concerns parameters that can be modified by the call.
5465 Note that there is no specific option to test @code{out} parameters,
5466 but any reference within the subprogram will be tested in the usual
5467 manner, and if an invalid value is copied back, any reference to it
5468 will be subject to validity checking.
5471 @emph{No validity checks.}
5472 @cindex @option{-gnatVn} (@command{gcc})
5473 This switch turns off all validity checking, including the default checking
5474 for case statements and left hand side subscripts. Note that the use of
5475 the switch @option{-gnatp} suppresses all run-time checks, including
5476 validity checks, and thus implies @option{-gnatVn}. When this switch
5477 is used, it cancels any other @option{-gnatV} previously issued.
5480 @emph{Validity checks for operator and attribute operands.}
5481 @cindex @option{-gnatVo} (@command{gcc})
5482 Arguments for predefined operators and attributes are validity checked.
5483 This includes all operators in package @code{Standard},
5484 the shift operators defined as intrinsic in package @code{Interfaces}
5485 and operands for attributes such as @code{Pos}. Checks are also made
5486 on individual component values for composite comparisons, and on the
5487 expressions in type conversions and qualified expressions. Checks are
5488 also made on explicit ranges using .. (e.g. slices, loops etc).
5491 @emph{Validity checks for parameters.}
5492 @cindex @option{-gnatVp} (@command{gcc})
5493 This controls the treatment of parameters within a subprogram (as opposed
5494 to @option{-gnatVi} and @option{-gnatVm} which control validity testing
5495 of parameters on a call. If either of these call options is used, then
5496 normally an assumption is made within a subprogram that the input arguments
5497 have been validity checking at the point of call, and do not need checking
5498 again within a subprogram). If @option{-gnatVp} is set, then this assumption
5499 is not made, and parameters are not assumed to be valid, so their validity
5500 will be checked (or rechecked) within the subprogram.
5503 @emph{Validity checks for function returns.}
5504 @cindex @option{-gnatVr} (@command{gcc})
5505 The expression in @code{return} statements in functions is validity
5509 @emph{Validity checks for subscripts.}
5510 @cindex @option{-gnatVs} (@command{gcc})
5511 All subscripts expressions are checked for validity, whether they appear
5512 on the right side or left side (in default mode only left side subscripts
5513 are validity checked).
5516 @emph{Validity checks for tests.}
5517 @cindex @option{-gnatVt} (@command{gcc})
5518 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
5519 statements are checked, as well as guard expressions in entry calls.
5524 The @option{-gnatV} switch may be followed by
5525 ^a string of letters^a list of options^
5526 to turn on a series of validity checking options.
5528 @option{^-gnatVcr^/VALIDITY_CHECKING=(COPIES, RETURNS)^}
5529 specifies that in addition to the default validity checking, copies and
5530 function return expressions are to be validity checked.
5531 In order to make it easier
5532 to specify the desired combination of effects,
5534 the upper case letters @code{CDFIMORST} may
5535 be used to turn off the corresponding lower case option.
5538 the prefix @code{NO} on an option turns off the corresponding validity
5541 @item @code{NOCOPIES}
5542 @item @code{NODEFAULT}
5543 @item @code{NOFLOATS}
5544 @item @code{NOIN_PARAMS}
5545 @item @code{NOMOD_PARAMS}
5546 @item @code{NOOPERANDS}
5547 @item @code{NORETURNS}
5548 @item @code{NOSUBSCRIPTS}
5549 @item @code{NOTESTS}
5553 @option{^-gnatVaM^/VALIDITY_CHECKING=(ALL, NOMOD_PARAMS)^}
5554 turns on all validity checking options except for
5555 checking of @code{@b{in out}} procedure arguments.
5557 The specification of additional validity checking generates extra code (and
5558 in the case of @option{-gnatVa} the code expansion can be substantial.
5559 However, these additional checks can be very useful in detecting
5560 uninitialized variables, incorrect use of unchecked conversion, and other
5561 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
5562 is useful in conjunction with the extra validity checking, since this
5563 ensures that wherever possible uninitialized variables have invalid values.
5565 See also the pragma @code{Validity_Checks} which allows modification of
5566 the validity checking mode at the program source level, and also allows for
5567 temporary disabling of validity checks.
5569 @node Style Checking
5570 @subsection Style Checking
5571 @findex Style checking
5574 The @option{-gnaty^x^(option,option,...)^} switch
5575 @cindex @option{-gnaty} (@command{gcc})
5576 causes the compiler to
5577 enforce specified style rules. A limited set of style rules has been used
5578 in writing the GNAT sources themselves. This switch allows user programs
5579 to activate all or some of these checks. If the source program fails a
5580 specified style check, an appropriate warning message is given, preceded by
5581 the character sequence ``(style)''.
5583 @code{(option,option,...)} is a sequence of keywords
5586 The string @var{x} is a sequence of letters or digits
5588 indicating the particular style
5589 checks to be performed. The following checks are defined:
5594 @emph{Specify indentation level.}
5595 If a digit from 1-9 appears
5596 ^in the string after @option{-gnaty}^as an option for /STYLE_CHECKS^
5597 then proper indentation is checked, with the digit indicating the
5598 indentation level required.
5599 The general style of required indentation is as specified by
5600 the examples in the Ada Reference Manual. Full line comments must be
5601 aligned with the @code{--} starting on a column that is a multiple of
5602 the alignment level.
5605 @emph{Check attribute casing.}
5606 If the ^letter a^word ATTRIBUTE^ appears in the string after @option{-gnaty}
5607 then attribute names, including the case of keywords such as @code{digits}
5608 used as attributes names, must be written in mixed case, that is, the
5609 initial letter and any letter following an underscore must be uppercase.
5610 All other letters must be lowercase.
5612 @item ^A^ARRAY_INDEXES^
5613 @emph{Use of array index numbers in array attributes.}
5614 If the ^letter A^word ARRAY_INDEXES^ appears in the string after
5615 @option{-gnaty} then when using the array attributes First, Last, Range,
5616 or Length, the index number must be omitted for one-dimensional arrays
5617 and is required for multi-dimensional arrays.
5620 @emph{Blanks not allowed at statement end.}
5621 If the ^letter b^word BLANKS^ appears in the string after @option{-gnaty} then
5622 trailing blanks are not allowed at the end of statements. The purpose of this
5623 rule, together with h (no horizontal tabs), is to enforce a canonical format
5624 for the use of blanks to separate source tokens.
5627 @emph{Check comments.}
5628 If the ^letter c^word COMMENTS^ appears in the string after @option{-gnaty}
5629 then comments must meet the following set of rules:
5634 The ``@code{--}'' that starts the column must either start in column one,
5635 or else at least one blank must precede this sequence.
5638 Comments that follow other tokens on a line must have at least one blank
5639 following the ``@code{--}'' at the start of the comment.
5642 Full line comments must have two blanks following the ``@code{--}'' that
5643 starts the comment, with the following exceptions.
5646 A line consisting only of the ``@code{--}'' characters, possibly preceded
5647 by blanks is permitted.
5650 A comment starting with ``@code{--x}'' where @code{x} is a special character
5652 This allows proper processing of the output generated by specialized tools
5653 including @command{gnatprep} (where ``@code{--!}'' is used) and the SPARK
5655 language (where ``@code{--#}'' is used). For the purposes of this rule, a
5656 special character is defined as being in one of the ASCII ranges
5657 @code{16#21#..16#2F#} or @code{16#3A#..16#3F#}.
5658 Note that this usage is not permitted
5659 in GNAT implementation units (i.e. when @option{-gnatg} is used).
5662 A line consisting entirely of minus signs, possibly preceded by blanks, is
5663 permitted. This allows the construction of box comments where lines of minus
5664 signs are used to form the top and bottom of the box.
5667 A comment that starts and ends with ``@code{--}'' is permitted as long as at
5668 least one blank follows the initial ``@code{--}''. Together with the preceding
5669 rule, this allows the construction of box comments, as shown in the following
5672 ---------------------------
5673 -- This is a box comment --
5674 -- with two text lines. --
5675 ---------------------------
5679 @item ^d^DOS_LINE_ENDINGS^
5680 @emph{Check no DOS line terminators present.}
5681 If the ^letter d^word DOS_LINE_ENDINGS^ appears in the string after
5682 @option{-gnaty} then all lines must be terminated by a single ASCII.LF
5683 character (in particular the DOS line terminator sequence CR/LF is not
5687 @emph{Check end/exit labels.}
5688 If the ^letter e^word END^ appears in the string after @option{-gnaty} then
5689 optional labels on @code{end} statements ending subprograms and on
5690 @code{exit} statements exiting named loops, are required to be present.
5693 @emph{No form feeds or vertical tabs.}
5694 If the ^letter f^word VTABS^ appears in the string after @option{-gnaty} then
5695 neither form feeds nor vertical tab characters are permitted
5699 @emph{GNAT style mode}
5700 If the ^letter g^word GNAT^ appears in the string after @option{-gnaty} then
5701 the set of style check switches is set to match that used by the GNAT sources.
5702 This may be useful when developing code that is eventually intended to be
5703 incorporated into GNAT. For further details, see GNAT sources.
5706 @emph{No horizontal tabs.}
5707 If the ^letter h^word HTABS^ appears in the string after @option{-gnaty} then
5708 horizontal tab characters are not permitted in the source text.
5709 Together with the b (no blanks at end of line) check, this
5710 enforces a canonical form for the use of blanks to separate
5714 @emph{Check if-then layout.}
5715 If the ^letter i^word IF_THEN^ appears in the string after @option{-gnaty},
5716 then the keyword @code{then} must appear either on the same
5717 line as corresponding @code{if}, or on a line on its own, lined
5718 up under the @code{if} with at least one non-blank line in between
5719 containing all or part of the condition to be tested.
5722 @emph{check mode IN keywords}
5723 If the ^letter I (upper case)^word IN_MODE^ appears in the string
5724 after @option{-gnaty} then mode @code{in} (the default mode) is not
5725 allowed to be given explicitly. @code{in out} is fine,
5726 but not @code{in} on its own.
5729 @emph{Check keyword casing.}
5730 If the ^letter k^word KEYWORD^ appears in the string after @option{-gnaty} then
5731 all keywords must be in lower case (with the exception of keywords
5732 such as @code{digits} used as attribute names to which this check
5736 @emph{Check layout.}
5737 If the ^letter l^word LAYOUT^ appears in the string after @option{-gnaty} then
5738 layout of statement and declaration constructs must follow the
5739 recommendations in the Ada Reference Manual, as indicated by the
5740 form of the syntax rules. For example an @code{else} keyword must
5741 be lined up with the corresponding @code{if} keyword.
5743 There are two respects in which the style rule enforced by this check
5744 option are more liberal than those in the Ada Reference Manual. First
5745 in the case of record declarations, it is permissible to put the
5746 @code{record} keyword on the same line as the @code{type} keyword, and
5747 then the @code{end} in @code{end record} must line up under @code{type}.
5748 This is also permitted when the type declaration is split on two lines.
5749 For example, any of the following three layouts is acceptable:
5751 @smallexample @c ada
5774 Second, in the case of a block statement, a permitted alternative
5775 is to put the block label on the same line as the @code{declare} or
5776 @code{begin} keyword, and then line the @code{end} keyword up under
5777 the block label. For example both the following are permitted:
5779 @smallexample @c ada
5797 The same alternative format is allowed for loops. For example, both of
5798 the following are permitted:
5800 @smallexample @c ada
5802 Clear : while J < 10 loop
5813 @item ^Lnnn^MAX_NESTING=nnn^
5814 @emph{Set maximum nesting level}
5815 If the sequence ^Lnnn^MAX_NESTING=nnn^, where nnn is a decimal number in
5816 the range 0-999, appears in the string after @option{-gnaty} then the
5817 maximum level of nesting of constructs (including subprograms, loops,
5818 blocks, packages, and conditionals) may not exceed the given value. A
5819 value of zero disconnects this style check.
5821 @item ^m^LINE_LENGTH^
5822 @emph{Check maximum line length.}
5823 If the ^letter m^word LINE_LENGTH^ appears in the string after @option{-gnaty}
5824 then the length of source lines must not exceed 79 characters, including
5825 any trailing blanks. The value of 79 allows convenient display on an
5826 80 character wide device or window, allowing for possible special
5827 treatment of 80 character lines. Note that this count is of
5828 characters in the source text. This means that a tab character counts
5829 as one character in this count but a wide character sequence counts as
5830 a single character (however many bytes are needed in the encoding).
5832 @item ^Mnnn^MAX_LENGTH=nnn^
5833 @emph{Set maximum line length.}
5834 If the sequence ^M^MAX_LENGTH=^nnn, where nnn is a decimal number, appears in
5835 the string after @option{-gnaty} then the length of lines must not exceed the
5836 given value. The maximum value that can be specified is 32767.
5838 @item ^n^STANDARD_CASING^
5839 @emph{Check casing of entities in Standard.}
5840 If the ^letter n^word STANDARD_CASING^ appears in the string
5841 after @option{-gnaty} then any identifier from Standard must be cased
5842 to match the presentation in the Ada Reference Manual (for example,
5843 @code{Integer} and @code{ASCII.NUL}).
5845 @item ^o^ORDERED_SUBPROGRAMS^
5846 @emph{Check order of subprogram bodies.}
5847 If the ^letter o^word ORDERED_SUBPROGRAMS^ appears in the string
5848 after @option{-gnaty} then all subprogram bodies in a given scope
5849 (e.g. a package body) must be in alphabetical order. The ordering
5850 rule uses normal Ada rules for comparing strings, ignoring casing
5851 of letters, except that if there is a trailing numeric suffix, then
5852 the value of this suffix is used in the ordering (e.g. Junk2 comes
5856 @emph{Check pragma casing.}
5857 If the ^letter p^word PRAGMA^ appears in the string after @option{-gnaty} then
5858 pragma names must be written in mixed case, that is, the
5859 initial letter and any letter following an underscore must be uppercase.
5860 All other letters must be lowercase.
5862 @item ^r^REFERENCES^
5863 @emph{Check references.}
5864 If the ^letter r^word REFERENCES^ appears in the string after @option{-gnaty}
5865 then all identifier references must be cased in the same way as the
5866 corresponding declaration. No specific casing style is imposed on
5867 identifiers. The only requirement is for consistency of references
5871 @emph{Check separate specs.}
5872 If the ^letter s^word SPECS^ appears in the string after @option{-gnaty} then
5873 separate declarations (``specs'') are required for subprograms (a
5874 body is not allowed to serve as its own declaration). The only
5875 exception is that parameterless library level procedures are
5876 not required to have a separate declaration. This exception covers
5877 the most frequent form of main program procedures.
5880 @emph{Check token spacing.}
5881 If the ^letter t^word TOKEN^ appears in the string after @option{-gnaty} then
5882 the following token spacing rules are enforced:
5887 The keywords @code{@b{abs}} and @code{@b{not}} must be followed by a space.
5890 The token @code{=>} must be surrounded by spaces.
5893 The token @code{<>} must be preceded by a space or a left parenthesis.
5896 Binary operators other than @code{**} must be surrounded by spaces.
5897 There is no restriction on the layout of the @code{**} binary operator.
5900 Colon must be surrounded by spaces.
5903 Colon-equal (assignment, initialization) must be surrounded by spaces.
5906 Comma must be the first non-blank character on the line, or be
5907 immediately preceded by a non-blank character, and must be followed
5911 If the token preceding a left parenthesis ends with a letter or digit, then
5912 a space must separate the two tokens.
5915 A right parenthesis must either be the first non-blank character on
5916 a line, or it must be preceded by a non-blank character.
5919 A semicolon must not be preceded by a space, and must not be followed by
5920 a non-blank character.
5923 A unary plus or minus may not be followed by a space.
5926 A vertical bar must be surrounded by spaces.
5929 @item ^u^UNNECESSARY_BLANK_LINES^
5930 @emph{Check unnecessary blank lines.}
5931 Check for unnecessary blank lines. A blank line is considered
5932 unnecessary if it appears at the end of the file, or if more than
5933 one blank line occurs in sequence.
5935 @item ^x^XTRA_PARENS^
5936 @emph{Check extra parentheses.}
5937 Check for the use of an unnecessary extra level of parentheses (C-style)
5938 around conditions in @code{if} statements, @code{while} statements and
5939 @code{exit} statements.
5944 In the above rules, appearing in column one is always permitted, that is,
5945 counts as meeting either a requirement for a required preceding space,
5946 or as meeting a requirement for no preceding space.
5948 Appearing at the end of a line is also always permitted, that is, counts
5949 as meeting either a requirement for a following space, or as meeting
5950 a requirement for no following space.
5953 If any of these style rules is violated, a message is generated giving
5954 details on the violation. The initial characters of such messages are
5955 always ``@code{(style)}''. Note that these messages are treated as warning
5956 messages, so they normally do not prevent the generation of an object
5957 file. The @option{-gnatwe} switch can be used to treat warning messages,
5958 including style messages, as fatal errors.
5962 @option{-gnaty} on its own (that is not
5963 followed by any letters or digits),
5964 is equivalent to @code{gnaty3aAbcefhiklmnprst}, that is all checking
5965 options enabled with the exception of @option{-gnatyo},
5966 @option{-gnatyd}, @option{-gnatyu}, and @option{-gnatyx}.
5969 /STYLE_CHECKS=ALL_BUILTIN enables all checking options with
5970 the exception of ORDERED_SUBPROGRAMS, UNNECESSARY_BLANK_LINES,
5971 XTRA_PARENS, and DOS_LINE_ENDINGS. In addition
5973 an indentation level of 3 is set. This is similar to the standard
5974 checking option that is used for the GNAT sources.
5983 clears any previously set style checks.
5985 @node Run-Time Checks
5986 @subsection Run-Time Checks
5987 @cindex Division by zero
5988 @cindex Access before elaboration
5989 @cindex Checks, division by zero
5990 @cindex Checks, access before elaboration
5991 @cindex Checks, stack overflow checking
5994 If you compile with the default options, GNAT will insert many run-time
5995 checks into the compiled code, including code that performs range
5996 checking against constraints, but not arithmetic overflow checking for
5997 integer operations (including division by zero), checks for access
5998 before elaboration on subprogram calls, or stack overflow checking. All
5999 other run-time checks, as required by the Ada Reference Manual, are
6000 generated by default. The following @command{gcc} switches refine this
6006 @cindex @option{-gnatp} (@command{gcc})
6007 @cindex Suppressing checks
6008 @cindex Checks, suppressing
6010 Suppress all run-time checks as though @code{pragma Suppress (all_checks})
6011 had been present in the source. Validity checks are also suppressed (in
6012 other words @option{-gnatp} also implies @option{-gnatVn}.
6013 Use this switch to improve the performance
6014 of the code at the expense of safety in the presence of invalid data or
6018 @cindex @option{-gnato} (@command{gcc})
6019 @cindex Overflow checks
6020 @cindex Check, overflow
6021 Enables overflow checking for integer operations.
6022 This causes GNAT to generate slower and larger executable
6023 programs by adding code to check for overflow (resulting in raising
6024 @code{Constraint_Error} as required by standard Ada
6025 semantics). These overflow checks correspond to situations in which
6026 the true value of the result of an operation may be outside the base
6027 range of the result type. The following example shows the distinction:
6029 @smallexample @c ada
6030 X1 : Integer := Integer'Last;
6031 X2 : Integer range 1 .. 5 := 5;
6032 X3 : Integer := Integer'Last;
6033 X4 : Integer range 1 .. 5 := 5;
6034 F : Float := 2.0E+20;
6043 Here the first addition results in a value that is outside the base range
6044 of Integer, and hence requires an overflow check for detection of the
6045 constraint error. Thus the first assignment to @code{X1} raises a
6046 @code{Constraint_Error} exception only if @option{-gnato} is set.
6048 The second increment operation results in a violation
6049 of the explicit range constraint, and such range checks are always
6050 performed (unless specifically suppressed with a pragma @code{suppress}
6051 or the use of @option{-gnatp}).
6053 The two conversions of @code{F} both result in values that are outside
6054 the base range of type @code{Integer} and thus will raise
6055 @code{Constraint_Error} exceptions only if @option{-gnato} is used.
6056 The fact that the result of the second conversion is assigned to
6057 variable @code{X4} with a restricted range is irrelevant, since the problem
6058 is in the conversion, not the assignment.
6060 Basically the rule is that in the default mode (@option{-gnato} not
6061 used), the generated code assures that all integer variables stay
6062 within their declared ranges, or within the base range if there is
6063 no declared range. This prevents any serious problems like indexes
6064 out of range for array operations.
6066 What is not checked in default mode is an overflow that results in
6067 an in-range, but incorrect value. In the above example, the assignments
6068 to @code{X1}, @code{X2}, @code{X3} all give results that are within the
6069 range of the target variable, but the result is wrong in the sense that
6070 it is too large to be represented correctly. Typically the assignment
6071 to @code{X1} will result in wrap around to the largest negative number.
6072 The conversions of @code{F} will result in some @code{Integer} value
6073 and if that integer value is out of the @code{X4} range then the
6074 subsequent assignment would generate an exception.
6076 @findex Machine_Overflows
6077 Note that the @option{-gnato} switch does not affect the code generated
6078 for any floating-point operations; it applies only to integer
6080 For floating-point, GNAT has the @code{Machine_Overflows}
6081 attribute set to @code{False} and the normal mode of operation is to
6082 generate IEEE NaN and infinite values on overflow or invalid operations
6083 (such as dividing 0.0 by 0.0).
6085 The reason that we distinguish overflow checking from other kinds of
6086 range constraint checking is that a failure of an overflow check can
6087 generate an incorrect value, but cannot cause erroneous behavior. This
6088 is unlike the situation with a constraint check on an array subscript,
6089 where failure to perform the check can result in random memory description,
6090 or the range check on a case statement, where failure to perform the check
6091 can cause a wild jump.
6093 Note again that @option{-gnato} is off by default, so overflow checking is
6094 not performed in default mode. This means that out of the box, with the
6095 default settings, GNAT does not do all the checks expected from the
6096 language description in the Ada Reference Manual. If you want all constraint
6097 checks to be performed, as described in this Manual, then you must
6098 explicitly use the -gnato switch either on the @command{gnatmake} or
6099 @command{gcc} command.
6102 @cindex @option{-gnatE} (@command{gcc})
6103 @cindex Elaboration checks
6104 @cindex Check, elaboration
6105 Enables dynamic checks for access-before-elaboration
6106 on subprogram calls and generic instantiations.
6107 For full details of the effect and use of this switch,
6108 @xref{Compiling Using gcc}.
6111 @cindex @option{-fstack-check} (@command{gcc})
6112 @cindex Stack Overflow Checking
6113 @cindex Checks, stack overflow checking
6114 Activates stack overflow checking. For full details of the effect and use of
6115 this switch see @ref{Stack Overflow Checking}.
6120 The setting of these switches only controls the default setting of the
6121 checks. You may modify them using either @code{Suppress} (to remove
6122 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
6125 @node Using gcc for Syntax Checking
6126 @subsection Using @command{gcc} for Syntax Checking
6129 @cindex @option{-gnats} (@command{gcc})
6133 The @code{s} stands for ``syntax''.
6136 Run GNAT in syntax checking only mode. For
6137 example, the command
6140 $ gcc -c -gnats x.adb
6144 compiles file @file{x.adb} in syntax-check-only mode. You can check a
6145 series of files in a single command
6147 , and can use wild cards to specify such a group of files.
6148 Note that you must specify the @option{-c} (compile
6149 only) flag in addition to the @option{-gnats} flag.
6152 You may use other switches in conjunction with @option{-gnats}. In
6153 particular, @option{-gnatl} and @option{-gnatv} are useful to control the
6154 format of any generated error messages.
6156 When the source file is empty or contains only empty lines and/or comments,
6157 the output is a warning:
6160 $ gcc -c -gnats -x ada toto.txt
6161 toto.txt:1:01: warning: empty file, contains no compilation units
6165 Otherwise, the output is simply the error messages, if any. No object file or
6166 ALI file is generated by a syntax-only compilation. Also, no units other
6167 than the one specified are accessed. For example, if a unit @code{X}
6168 @code{with}'s a unit @code{Y}, compiling unit @code{X} in syntax
6169 check only mode does not access the source file containing unit
6172 @cindex Multiple units, syntax checking
6173 Normally, GNAT allows only a single unit in a source file. However, this
6174 restriction does not apply in syntax-check-only mode, and it is possible
6175 to check a file containing multiple compilation units concatenated
6176 together. This is primarily used by the @code{gnatchop} utility
6177 (@pxref{Renaming Files Using gnatchop}).
6180 @node Using gcc for Semantic Checking
6181 @subsection Using @command{gcc} for Semantic Checking
6184 @cindex @option{-gnatc} (@command{gcc})
6188 The @code{c} stands for ``check''.
6190 Causes the compiler to operate in semantic check mode,
6191 with full checking for all illegalities specified in the
6192 Ada Reference Manual, but without generation of any object code
6193 (no object file is generated).
6195 Because dependent files must be accessed, you must follow the GNAT
6196 semantic restrictions on file structuring to operate in this mode:
6200 The needed source files must be accessible
6201 (@pxref{Search Paths and the Run-Time Library (RTL)}).
6204 Each file must contain only one compilation unit.
6207 The file name and unit name must match (@pxref{File Naming Rules}).
6210 The output consists of error messages as appropriate. No object file is
6211 generated. An @file{ALI} file is generated for use in the context of
6212 cross-reference tools, but this file is marked as not being suitable
6213 for binding (since no object file is generated).
6214 The checking corresponds exactly to the notion of
6215 legality in the Ada Reference Manual.
6217 Any unit can be compiled in semantics-checking-only mode, including
6218 units that would not normally be compiled (subunits,
6219 and specifications where a separate body is present).
6222 @node Compiling Different Versions of Ada
6223 @subsection Compiling Different Versions of Ada
6226 The switches described in this section allow you to explicitly specify
6227 the version of the Ada language that your programs are written in.
6228 By default @value{EDITION} assumes @value{DEFAULTLANGUAGEVERSION},
6229 but you can also specify @value{NONDEFAULTLANGUAGEVERSION} or
6230 indicate Ada 83 compatibility mode.
6233 @cindex Compatibility with Ada 83
6235 @item -gnat83 (Ada 83 Compatibility Mode)
6236 @cindex @option{-gnat83} (@command{gcc})
6237 @cindex ACVC, Ada 83 tests
6241 Although GNAT is primarily an Ada 95 / Ada 2005 compiler, this switch
6242 specifies that the program is to be compiled in Ada 83 mode. With
6243 @option{-gnat83}, GNAT rejects most post-Ada 83 extensions and applies Ada 83
6244 semantics where this can be done easily.
6245 It is not possible to guarantee this switch does a perfect
6246 job; some subtle tests, such as are
6247 found in earlier ACVC tests (and that have been removed from the ACATS suite
6248 for Ada 95), might not compile correctly.
6249 Nevertheless, this switch may be useful in some circumstances, for example
6250 where, due to contractual reasons, existing code needs to be maintained
6251 using only Ada 83 features.
6253 With few exceptions (most notably the need to use @code{<>} on
6254 @cindex Generic formal parameters
6255 unconstrained generic formal parameters, the use of the new Ada 95 / Ada 2005
6256 reserved words, and the use of packages
6257 with optional bodies), it is not necessary to specify the
6258 @option{-gnat83} switch when compiling Ada 83 programs, because, with rare
6259 exceptions, Ada 95 and Ada 2005 are upwardly compatible with Ada 83. Thus
6260 a correct Ada 83 program is usually also a correct program
6261 in these later versions of the language standard.
6262 For further information, please refer to @ref{Compatibility and Porting Guide}.
6264 @item -gnat95 (Ada 95 mode)
6265 @cindex @option{-gnat95} (@command{gcc})
6269 This switch directs the compiler to implement the Ada 95 version of the
6271 Since Ada 95 is almost completely upwards
6272 compatible with Ada 83, Ada 83 programs may generally be compiled using
6273 this switch (see the description of the @option{-gnat83} switch for further
6274 information about Ada 83 mode).
6275 If an Ada 2005 program is compiled in Ada 95 mode,
6276 uses of the new Ada 2005 features will cause error
6277 messages or warnings.
6279 This switch also can be used to cancel the effect of a previous
6280 @option{-gnat83} or @option{-gnat05} switch earlier in the command line.
6282 @item -gnat05 (Ada 2005 mode)
6283 @cindex @option{-gnat05} (@command{gcc})
6284 @cindex Ada 2005 mode
6287 This switch directs the compiler to implement the Ada 2005 version of the
6289 Since Ada 2005 is almost completely upwards
6290 compatible with Ada 95 (and thus also with Ada 83), Ada 83 and Ada 95 programs
6291 may generally be compiled using this switch (see the description of the
6292 @option{-gnat83} and @option{-gnat95} switches for further
6295 For information about the approved ``Ada Issues'' that have been incorporated
6296 into Ada 2005, see @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs}.
6297 Included with GNAT releases is a file @file{features-ada0y} that describes
6298 the set of implemented Ada 2005 features.
6302 @node Character Set Control
6303 @subsection Character Set Control
6305 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
6306 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@command{gcc})
6309 Normally GNAT recognizes the Latin-1 character set in source program
6310 identifiers, as described in the Ada Reference Manual.
6312 GNAT to recognize alternate character sets in identifiers. @var{c} is a
6313 single character ^^or word^ indicating the character set, as follows:
6317 ISO 8859-1 (Latin-1) identifiers
6320 ISO 8859-2 (Latin-2) letters allowed in identifiers
6323 ISO 8859-3 (Latin-3) letters allowed in identifiers
6326 ISO 8859-4 (Latin-4) letters allowed in identifiers
6329 ISO 8859-5 (Cyrillic) letters allowed in identifiers
6332 ISO 8859-15 (Latin-9) letters allowed in identifiers
6335 IBM PC letters (code page 437) allowed in identifiers
6338 IBM PC letters (code page 850) allowed in identifiers
6340 @item ^f^FULL_UPPER^
6341 Full upper-half codes allowed in identifiers
6344 No upper-half codes allowed in identifiers
6347 Wide-character codes (that is, codes greater than 255)
6348 allowed in identifiers
6351 @xref{Foreign Language Representation}, for full details on the
6352 implementation of these character sets.
6354 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
6355 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@command{gcc})
6356 Specify the method of encoding for wide characters.
6357 @var{e} is one of the following:
6362 Hex encoding (brackets coding also recognized)
6365 Upper half encoding (brackets encoding also recognized)
6368 Shift/JIS encoding (brackets encoding also recognized)
6371 EUC encoding (brackets encoding also recognized)
6374 UTF-8 encoding (brackets encoding also recognized)
6377 Brackets encoding only (default value)
6379 For full details on these encoding
6380 methods see @ref{Wide Character Encodings}.
6381 Note that brackets coding is always accepted, even if one of the other
6382 options is specified, so for example @option{-gnatW8} specifies that both
6383 brackets and @code{UTF-8} encodings will be recognized. The units that are
6384 with'ed directly or indirectly will be scanned using the specified
6385 representation scheme, and so if one of the non-brackets scheme is
6386 used, it must be used consistently throughout the program. However,
6387 since brackets encoding is always recognized, it may be conveniently
6388 used in standard libraries, allowing these libraries to be used with
6389 any of the available coding schemes.
6390 scheme. If no @option{-gnatW?} parameter is present, then the default
6391 representation is Brackets encoding only.
6393 Note that the wide character representation that is specified (explicitly
6394 or by default) for the main program also acts as the default encoding used
6395 for Wide_Text_IO files if not specifically overridden by a WCEM form
6399 @node File Naming Control
6400 @subsection File Naming Control
6403 @item ^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{n}
6404 @cindex @option{-gnatk} (@command{gcc})
6405 Activates file name ``krunching''. @var{n}, a decimal integer in the range
6406 1-999, indicates the maximum allowable length of a file name (not
6407 including the @file{.ads} or @file{.adb} extension). The default is not
6408 to enable file name krunching.
6410 For the source file naming rules, @xref{File Naming Rules}.
6413 @node Subprogram Inlining Control
6414 @subsection Subprogram Inlining Control
6419 @cindex @option{-gnatn} (@command{gcc})
6421 The @code{n} here is intended to suggest the first syllable of the
6424 GNAT recognizes and processes @code{Inline} pragmas. However, for the
6425 inlining to actually occur, optimization must be enabled. To enable
6426 inlining of subprograms specified by pragma @code{Inline},
6427 you must also specify this switch.
6428 In the absence of this switch, GNAT does not attempt
6429 inlining and does not need to access the bodies of
6430 subprograms for which @code{pragma Inline} is specified if they are not
6431 in the current unit.
6433 If you specify this switch the compiler will access these bodies,
6434 creating an extra source dependency for the resulting object file, and
6435 where possible, the call will be inlined.
6436 For further details on when inlining is possible
6437 see @ref{Inlining of Subprograms}.
6440 @cindex @option{-gnatN} (@command{gcc})
6441 The front end inlining activated by this switch is generally more extensive,
6442 and quite often more effective than the standard @option{-gnatn} inlining mode.
6443 It will also generate additional dependencies.
6445 @option{-gnatN} automatically implies @option{-gnatn} so it is not necessary
6446 to specify both options.
6449 @node Auxiliary Output Control
6450 @subsection Auxiliary Output Control
6454 @cindex @option{-gnatt} (@command{gcc})
6455 @cindex Writing internal trees
6456 @cindex Internal trees, writing to file
6457 Causes GNAT to write the internal tree for a unit to a file (with the
6458 extension @file{.adt}.
6459 This not normally required, but is used by separate analysis tools.
6461 these tools do the necessary compilations automatically, so you should
6462 not have to specify this switch in normal operation.
6465 @cindex @option{-gnatu} (@command{gcc})
6466 Print a list of units required by this compilation on @file{stdout}.
6467 The listing includes all units on which the unit being compiled depends
6468 either directly or indirectly.
6471 @item -pass-exit-codes
6472 @cindex @option{-pass-exit-codes} (@command{gcc})
6473 If this switch is not used, the exit code returned by @command{gcc} when
6474 compiling multiple files indicates whether all source files have
6475 been successfully used to generate object files or not.
6477 When @option{-pass-exit-codes} is used, @command{gcc} exits with an extended
6478 exit status and allows an integrated development environment to better
6479 react to a compilation failure. Those exit status are:
6483 There was an error in at least one source file.
6485 At least one source file did not generate an object file.
6487 The compiler died unexpectedly (internal error for example).
6489 An object file has been generated for every source file.
6494 @node Debugging Control
6495 @subsection Debugging Control
6499 @cindex Debugging options
6502 @cindex @option{-gnatd} (@command{gcc})
6503 Activate internal debugging switches. @var{x} is a letter or digit, or
6504 string of letters or digits, which specifies the type of debugging
6505 outputs desired. Normally these are used only for internal development
6506 or system debugging purposes. You can find full documentation for these
6507 switches in the body of the @code{Debug} unit in the compiler source
6508 file @file{debug.adb}.
6512 @cindex @option{-gnatG} (@command{gcc})
6513 This switch causes the compiler to generate auxiliary output containing
6514 a pseudo-source listing of the generated expanded code. Like most Ada
6515 compilers, GNAT works by first transforming the high level Ada code into
6516 lower level constructs. For example, tasking operations are transformed
6517 into calls to the tasking run-time routines. A unique capability of GNAT
6518 is to list this expanded code in a form very close to normal Ada source.
6519 This is very useful in understanding the implications of various Ada
6520 usage on the efficiency of the generated code. There are many cases in
6521 Ada (e.g. the use of controlled types), where simple Ada statements can
6522 generate a lot of run-time code. By using @option{-gnatG} you can identify
6523 these cases, and consider whether it may be desirable to modify the coding
6524 approach to improve efficiency.
6526 The format of the output is very similar to standard Ada source, and is
6527 easily understood by an Ada programmer. The following special syntactic
6528 additions correspond to low level features used in the generated code that
6529 do not have any exact analogies in pure Ada source form. The following
6530 is a partial list of these special constructions. See the specification
6531 of package @code{Sprint} in file @file{sprint.ads} for a full list.
6533 If the switch @option{-gnatL} is used in conjunction with
6534 @cindex @option{-gnatL} (@command{gcc})
6535 @option{-gnatG}, then the original source lines are interspersed
6536 in the expanded source (as comment lines with the original line number).
6539 @item new @var{xxx} [storage_pool = @var{yyy}]
6540 Shows the storage pool being used for an allocator.
6542 @item at end @var{procedure-name};
6543 Shows the finalization (cleanup) procedure for a scope.
6545 @item (if @var{expr} then @var{expr} else @var{expr})
6546 Conditional expression equivalent to the @code{x?y:z} construction in C.
6548 @item @var{target}^^^(@var{source})
6549 A conversion with floating-point truncation instead of rounding.
6551 @item @var{target}?(@var{source})
6552 A conversion that bypasses normal Ada semantic checking. In particular
6553 enumeration types and fixed-point types are treated simply as integers.
6555 @item @var{target}?^^^(@var{source})
6556 Combines the above two cases.
6558 @item @var{x} #/ @var{y}
6559 @itemx @var{x} #mod @var{y}
6560 @itemx @var{x} #* @var{y}
6561 @itemx @var{x} #rem @var{y}
6562 A division or multiplication of fixed-point values which are treated as
6563 integers without any kind of scaling.
6565 @item free @var{expr} [storage_pool = @var{xxx}]
6566 Shows the storage pool associated with a @code{free} statement.
6568 @item [subtype or type declaration]
6569 Used to list an equivalent declaration for an internally generated
6570 type that is referenced elsewhere in the listing.
6572 @item freeze @var{type-name} [@var{actions}]
6573 Shows the point at which @var{type-name} is frozen, with possible
6574 associated actions to be performed at the freeze point.
6576 @item reference @var{itype}
6577 Reference (and hence definition) to internal type @var{itype}.
6579 @item @var{function-name}! (@var{arg}, @var{arg}, @var{arg})
6580 Intrinsic function call.
6582 @item @var{label-name} : label
6583 Declaration of label @var{labelname}.
6585 @item #$ @var{subprogram-name}
6586 An implicit call to a run-time support routine
6587 (to meet the requirement of H.3.1(9) in a
6590 @item @var{expr} && @var{expr} && @var{expr} ... && @var{expr}
6591 A multiple concatenation (same effect as @var{expr} & @var{expr} &
6592 @var{expr}, but handled more efficiently).
6594 @item [constraint_error]
6595 Raise the @code{Constraint_Error} exception.
6597 @item @var{expression}'reference
6598 A pointer to the result of evaluating @var{expression}.
6600 @item @var{target-type}!(@var{source-expression})
6601 An unchecked conversion of @var{source-expression} to @var{target-type}.
6603 @item [@var{numerator}/@var{denominator}]
6604 Used to represent internal real literals (that) have no exact
6605 representation in base 2-16 (for example, the result of compile time
6606 evaluation of the expression 1.0/27.0).
6610 @cindex @option{-gnatD} (@command{gcc})
6611 When used in conjunction with @option{-gnatG}, this switch causes
6612 the expanded source, as described above for
6613 @option{-gnatG} to be written to files with names
6614 @file{^xxx.dg^XXX_DG^}, where @file{xxx} is the normal file name,
6615 instead of to the standard output file. For
6616 example, if the source file name is @file{hello.adb}, then a file
6617 @file{^hello.adb.dg^HELLO.ADB_DG^} will be written. The debugging
6618 information generated by the @command{gcc} @option{^-g^/DEBUG^} switch
6619 will refer to the generated @file{^xxx.dg^XXX_DG^} file. This allows
6620 you to do source level debugging using the generated code which is
6621 sometimes useful for complex code, for example to find out exactly
6622 which part of a complex construction raised an exception. This switch
6623 also suppress generation of cross-reference information (see
6624 @option{-gnatx}) since otherwise the cross-reference information
6625 would refer to the @file{^.dg^.DG^} file, which would cause
6626 confusion since this is not the original source file.
6628 Note that @option{-gnatD} actually implies @option{-gnatG}
6629 automatically, so it is not necessary to give both options.
6630 In other words @option{-gnatD} is equivalent to @option{-gnatDG}).
6632 If the switch @option{-gnatL} is used in conjunction with
6633 @cindex @option{-gnatL} (@command{gcc})
6634 @option{-gnatDG}, then the original source lines are interspersed
6635 in the expanded source (as comment lines with the original line number).
6638 @item -gnatR[0|1|2|3[s]]
6639 @cindex @option{-gnatR} (@command{gcc})
6640 This switch controls output from the compiler of a listing showing
6641 representation information for declared types and objects. For
6642 @option{-gnatR0}, no information is output (equivalent to omitting
6643 the @option{-gnatR} switch). For @option{-gnatR1} (which is the default,
6644 so @option{-gnatR} with no parameter has the same effect), size and alignment
6645 information is listed for declared array and record types. For
6646 @option{-gnatR2}, size and alignment information is listed for all
6647 expressions for values that are computed at run time for
6648 variant records. These symbolic expressions have a mostly obvious
6649 format with #n being used to represent the value of the n'th
6650 discriminant. See source files @file{repinfo.ads/adb} in the
6651 @code{GNAT} sources for full details on the format of @option{-gnatR3}
6652 output. If the switch is followed by an s (e.g. @option{-gnatR2s}), then
6653 the output is to a file with the name @file{^file.rep^file_REP^} where
6654 file is the name of the corresponding source file.
6657 @item /REPRESENTATION_INFO
6658 @cindex @option{/REPRESENTATION_INFO} (@command{gcc})
6659 This qualifier controls output from the compiler of a listing showing
6660 representation information for declared types and objects. For
6661 @option{/REPRESENTATION_INFO=NONE}, no information is output
6662 (equivalent to omitting the @option{/REPRESENTATION_INFO} qualifier).
6663 @option{/REPRESENTATION_INFO} without option is equivalent to
6664 @option{/REPRESENTATION_INFO=ARRAYS}.
6665 For @option{/REPRESENTATION_INFO=ARRAYS}, size and alignment
6666 information is listed for declared array and record types. For
6667 @option{/REPRESENTATION_INFO=OBJECTS}, size and alignment information
6668 is listed for all expression information for values that are computed
6669 at run time for variant records. These symbolic expressions have a mostly
6670 obvious format with #n being used to represent the value of the n'th
6671 discriminant. See source files @file{REPINFO.ADS/ADB} in the
6672 @code{GNAT} sources for full details on the format of
6673 @option{/REPRESENTATION_INFO=SYMBOLIC} output.
6674 If _FILE is added at the end of an option
6675 (e.g. @option{/REPRESENTATION_INFO=ARRAYS_FILE}),
6676 then the output is to a file with the name @file{file_REP} where
6677 file is the name of the corresponding source file.
6679 Note that it is possible for record components to have zero size. In
6680 this case, the component clause uses an obvious extension of permitted
6681 Ada syntax, for example @code{at 0 range 0 .. -1}.
6683 Representation information requires that code be generated (since it is the
6684 code generator that lays out complex data structures). If an attempt is made
6685 to output representation information when no code is generated, for example
6686 when a subunit is compiled on its own, then no information can be generated
6687 and the compiler outputs a message to this effect.
6690 @cindex @option{-gnatS} (@command{gcc})
6691 The use of the switch @option{-gnatS} for an
6692 Ada compilation will cause the compiler to output a
6693 representation of package Standard in a form very
6694 close to standard Ada. It is not quite possible to
6695 do this entirely in standard Ada (since new
6696 numeric base types cannot be created in standard
6697 Ada), but the output is easily
6698 readable to any Ada programmer, and is useful to
6699 determine the characteristics of target dependent
6700 types in package Standard.
6703 @cindex @option{-gnatx} (@command{gcc})
6704 Normally the compiler generates full cross-referencing information in
6705 the @file{ALI} file. This information is used by a number of tools,
6706 including @code{gnatfind} and @code{gnatxref}. The @option{-gnatx} switch
6707 suppresses this information. This saves some space and may slightly
6708 speed up compilation, but means that these tools cannot be used.
6711 @node Exception Handling Control
6712 @subsection Exception Handling Control
6715 GNAT uses two methods for handling exceptions at run-time. The
6716 @code{setjmp/longjmp} method saves the context when entering
6717 a frame with an exception handler. Then when an exception is
6718 raised, the context can be restored immediately, without the
6719 need for tracing stack frames. This method provides very fast
6720 exception propagation, but introduces significant overhead for
6721 the use of exception handlers, even if no exception is raised.
6723 The other approach is called ``zero cost'' exception handling.
6724 With this method, the compiler builds static tables to describe
6725 the exception ranges. No dynamic code is required when entering
6726 a frame containing an exception handler. When an exception is
6727 raised, the tables are used to control a back trace of the
6728 subprogram invocation stack to locate the required exception
6729 handler. This method has considerably poorer performance for
6730 the propagation of exceptions, but there is no overhead for
6731 exception handlers if no exception is raised. Note that in this
6732 mode and in the context of mixed Ada and C/C++ programming,
6733 to propagate an exception through a C/C++ code, the C/C++ code
6734 must be compiled with the @option{-funwind-tables} GCC's
6737 The following switches may be used to control which of the
6738 two exception handling methods is used.
6744 @cindex @option{--RTS=sjlj} (@command{gnatmake})
6745 This switch causes the setjmp/longjmp run-time (when available) to be used
6746 for exception handling. If the default
6747 mechanism for the target is zero cost exceptions, then
6748 this switch can be used to modify this default, and must be
6749 used for all units in the partition.
6750 This option is rarely used. One case in which it may be
6751 advantageous is if you have an application where exception
6752 raising is common and the overall performance of the
6753 application is improved by favoring exception propagation.
6756 @cindex @option{--RTS=zcx} (@command{gnatmake})
6757 @cindex Zero Cost Exceptions
6758 This switch causes the zero cost approach to be used
6759 for exception handling. If this is the default mechanism for the
6760 target (see below), then this switch is unneeded. If the default
6761 mechanism for the target is setjmp/longjmp exceptions, then
6762 this switch can be used to modify this default, and must be
6763 used for all units in the partition.
6764 This option can only be used if the zero cost approach
6765 is available for the target in use, otherwise it will generate an error.
6769 The same option @option{--RTS} must be used both for @command{gcc}
6770 and @command{gnatbind}. Passing this option to @command{gnatmake}
6771 (@pxref{Switches for gnatmake}) will ensure the required consistency
6772 through the compilation and binding steps.
6774 @node Units to Sources Mapping Files
6775 @subsection Units to Sources Mapping Files
6779 @item -gnatem^^=^@var{path}
6780 @cindex @option{-gnatem} (@command{gcc})
6781 A mapping file is a way to communicate to the compiler two mappings:
6782 from unit names to file names (without any directory information) and from
6783 file names to path names (with full directory information). These mappings
6784 are used by the compiler to short-circuit the path search.
6786 The use of mapping files is not required for correct operation of the
6787 compiler, but mapping files can improve efficiency, particularly when
6788 sources are read over a slow network connection. In normal operation,
6789 you need not be concerned with the format or use of mapping files,
6790 and the @option{-gnatem} switch is not a switch that you would use
6791 explicitly. it is intended only for use by automatic tools such as
6792 @command{gnatmake} running under the project file facility. The
6793 description here of the format of mapping files is provided
6794 for completeness and for possible use by other tools.
6796 A mapping file is a sequence of sets of three lines. In each set,
6797 the first line is the unit name, in lower case, with ``@code{%s}''
6799 specifications and ``@code{%b}'' appended for bodies; the second line is the
6800 file name; and the third line is the path name.
6806 /gnat/project1/sources/main.2.ada
6809 When the switch @option{-gnatem} is specified, the compiler will create
6810 in memory the two mappings from the specified file. If there is any problem
6811 (non existent file, truncated file or duplicate entries), no mapping
6814 Several @option{-gnatem} switches may be specified; however, only the last
6815 one on the command line will be taken into account.
6817 When using a project file, @command{gnatmake} create a temporary mapping file
6818 and communicates it to the compiler using this switch.
6822 @node Integrated Preprocessing
6823 @subsection Integrated Preprocessing
6826 GNAT sources may be preprocessed immediately before compilation; the actual
6827 text of the source is not the text of the source file, but is derived from it
6828 through a process called preprocessing. Integrated preprocessing is specified
6829 through switches @option{-gnatep} and/or @option{-gnateD}. @option{-gnatep}
6830 indicates, through a text file, the preprocessing data to be used.
6831 @option{-gnateD} specifies or modifies the values of preprocessing symbol.
6834 It is recommended that @command{gnatmake} switch ^-s^/SWITCH_CHECK^ should be
6835 used when Integrated Preprocessing is used. The reason is that preprocessing
6836 with another Preprocessing Data file without changing the sources will
6837 not trigger recompilation without this switch.
6840 Note that @command{gnatmake} switch ^-m^/MINIMAL_RECOMPILATION^ will almost
6841 always trigger recompilation for sources that are preprocessed,
6842 because @command{gnatmake} cannot compute the checksum of the source after
6846 The actual preprocessing function is described in details in section
6847 @ref{Preprocessing Using gnatprep}. This section only describes how integrated
6848 preprocessing is triggered and parameterized.
6852 @item -gnatep=@var{file}
6853 @cindex @option{-gnatep} (@command{gcc})
6854 This switch indicates to the compiler the file name (without directory
6855 information) of the preprocessor data file to use. The preprocessor data file
6856 should be found in the source directories.
6859 A preprocessing data file is a text file with significant lines indicating
6860 how should be preprocessed either a specific source or all sources not
6861 mentioned in other lines. A significant line is a non empty, non comment line.
6862 Comments are similar to Ada comments.
6865 Each significant line starts with either a literal string or the character '*'.
6866 A literal string is the file name (without directory information) of the source
6867 to preprocess. A character '*' indicates the preprocessing for all the sources
6868 that are not specified explicitly on other lines (order of the lines is not
6869 significant). It is an error to have two lines with the same file name or two
6870 lines starting with the character '*'.
6873 After the file name or the character '*', another optional literal string
6874 indicating the file name of the definition file to be used for preprocessing
6875 (@pxref{Form of Definitions File}). The definition files are found by the
6876 compiler in one of the source directories. In some cases, when compiling
6877 a source in a directory other than the current directory, if the definition
6878 file is in the current directory, it may be necessary to add the current
6879 directory as a source directory through switch ^-I.^/SEARCH=[]^, otherwise
6880 the compiler would not find the definition file.
6883 Then, optionally, ^switches^switches^ similar to those of @code{gnatprep} may
6884 be found. Those ^switches^switches^ are:
6889 Causes both preprocessor lines and the lines deleted by
6890 preprocessing to be replaced by blank lines, preserving the line number.
6891 This ^switch^switch^ is always implied; however, if specified after @option{-c}
6892 it cancels the effect of @option{-c}.
6895 Causes both preprocessor lines and the lines deleted
6896 by preprocessing to be retained as comments marked
6897 with the special string ``@code{--! }''.
6899 @item -Dsymbol=value
6900 Define or redefine a symbol, associated with value. A symbol is an Ada
6901 identifier, or an Ada reserved word, with the exception of @code{if},
6902 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
6903 @code{value} is either a literal string, an Ada identifier or any Ada reserved
6904 word. A symbol declared with this ^switch^switch^ replaces a symbol with the
6905 same name defined in a definition file.
6908 Causes a sorted list of symbol names and values to be
6909 listed on the standard output file.
6912 Causes undefined symbols to be treated as having the value @code{FALSE}
6914 of a preprocessor test. In the absence of this option, an undefined symbol in
6915 a @code{#if} or @code{#elsif} test will be treated as an error.
6920 Examples of valid lines in a preprocessor data file:
6923 "toto.adb" "prep.def" -u
6924 -- preprocess "toto.adb", using definition file "prep.def",
6925 -- undefined symbol are False.
6928 -- preprocess all other sources without a definition file;
6929 -- suppressed lined are commented; symbol VERSION has the value V101.
6931 "titi.adb" "prep2.def" -s
6932 -- preprocess "titi.adb", using definition file "prep2.def";
6933 -- list all symbols with their values.
6936 @item ^-gnateD^/DATA_PREPROCESSING=^symbol[=value]
6937 @cindex @option{-gnateD} (@command{gcc})
6938 Define or redefine a preprocessing symbol, associated with value. If no value
6939 is given on the command line, then the value of the symbol is @code{True}.
6940 A symbol is an identifier, following normal Ada (case-insensitive)
6941 rules for its syntax, and value is any sequence (including an empty sequence)
6942 of characters from the set (letters, digits, period, underline).
6943 Ada reserved words may be used as symbols, with the exceptions of @code{if},
6944 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
6947 A symbol declared with this ^switch^switch^ on the command line replaces a
6948 symbol with the same name either in a definition file or specified with a
6949 ^switch^switch^ -D in the preprocessor data file.
6952 This switch is similar to switch @option{^-D^/ASSOCIATE^} of @code{gnatprep}.
6956 @node Code Generation Control
6957 @subsection Code Generation Control
6961 The GCC technology provides a wide range of target dependent
6962 @option{-m} switches for controlling
6963 details of code generation with respect to different versions of
6964 architectures. This includes variations in instruction sets (e.g.
6965 different members of the power pc family), and different requirements
6966 for optimal arrangement of instructions (e.g. different members of
6967 the x86 family). The list of available @option{-m} switches may be
6968 found in the GCC documentation.
6970 Use of these @option{-m} switches may in some cases result in improved
6973 The GNAT Pro technology is tested and qualified without any
6974 @option{-m} switches,
6975 so generally the most reliable approach is to avoid the use of these
6976 switches. However, we generally expect most of these switches to work
6977 successfully with GNAT Pro, and many customers have reported successful
6978 use of these options.
6980 Our general advice is to avoid the use of @option{-m} switches unless
6981 special needs lead to requirements in this area. In particular,
6982 there is no point in using @option{-m} switches to improve performance
6983 unless you actually see a performance improvement.
6987 @subsection Return Codes
6988 @cindex Return Codes
6989 @cindex @option{/RETURN_CODES=VMS}
6992 On VMS, GNAT compiled programs return POSIX-style codes by default,
6993 e.g. @option{/RETURN_CODES=POSIX}.
6995 To enable VMS style return codes, use GNAT BIND and LINK with the option
6996 @option{/RETURN_CODES=VMS}. For example:
6999 GNAT BIND MYMAIN.ALI /RETURN_CODES=VMS
7000 GNAT LINK MYMAIN.ALI /RETURN_CODES=VMS
7004 Programs built with /RETURN_CODES=VMS are suitable to be called in
7005 VMS DCL scripts. Programs compiled with the default /RETURN_CODES=POSIX
7006 are suitable for spawning with appropriate GNAT RTL routines.
7010 @node Search Paths and the Run-Time Library (RTL)
7011 @section Search Paths and the Run-Time Library (RTL)
7014 With the GNAT source-based library system, the compiler must be able to
7015 find source files for units that are needed by the unit being compiled.
7016 Search paths are used to guide this process.
7018 The compiler compiles one source file whose name must be given
7019 explicitly on the command line. In other words, no searching is done
7020 for this file. To find all other source files that are needed (the most
7021 common being the specs of units), the compiler examines the following
7022 directories, in the following order:
7026 The directory containing the source file of the main unit being compiled
7027 (the file name on the command line).
7030 Each directory named by an @option{^-I^/SOURCE_SEARCH^} switch given on the
7031 @command{gcc} command line, in the order given.
7034 @findex ADA_PRJ_INCLUDE_FILE
7035 Each of the directories listed in the text file whose name is given
7036 by the @code{ADA_PRJ_INCLUDE_FILE} ^environment variable^logical name^.
7039 @code{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
7040 driver when project files are used. It should not normally be set
7044 @findex ADA_INCLUDE_PATH
7045 Each of the directories listed in the value of the
7046 @code{ADA_INCLUDE_PATH} ^environment variable^logical name^.
7048 Construct this value
7049 exactly as the @code{PATH} environment variable: a list of directory
7050 names separated by colons (semicolons when working with the NT version).
7053 Normally, define this value as a logical name containing a comma separated
7054 list of directory names.
7056 This variable can also be defined by means of an environment string
7057 (an argument to the HP C exec* set of functions).
7061 DEFINE ANOTHER_PATH FOO:[BAG]
7062 DEFINE ADA_INCLUDE_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
7065 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
7066 first, followed by the standard Ada
7067 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADAINCLUDE].
7068 If this is not redefined, the user will obtain the HP Ada 83 IO packages
7069 (Text_IO, Sequential_IO, etc)
7070 instead of the standard Ada packages. Thus, in order to get the standard Ada
7071 packages by default, ADA_INCLUDE_PATH must be redefined.
7075 The content of the @file{ada_source_path} file which is part of the GNAT
7076 installation tree and is used to store standard libraries such as the
7077 GNAT Run Time Library (RTL) source files.
7079 @ref{Installing a library}
7084 Specifying the switch @option{^-I-^/NOCURRENT_DIRECTORY^}
7085 inhibits the use of the directory
7086 containing the source file named in the command line. You can still
7087 have this directory on your search path, but in this case it must be
7088 explicitly requested with a @option{^-I^/SOURCE_SEARCH^} switch.
7090 Specifying the switch @option{-nostdinc}
7091 inhibits the search of the default location for the GNAT Run Time
7092 Library (RTL) source files.
7094 The compiler outputs its object files and ALI files in the current
7097 Caution: The object file can be redirected with the @option{-o} switch;
7098 however, @command{gcc} and @code{gnat1} have not been coordinated on this
7099 so the @file{ALI} file will not go to the right place. Therefore, you should
7100 avoid using the @option{-o} switch.
7104 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
7105 children make up the GNAT RTL, together with the simple @code{System.IO}
7106 package used in the @code{"Hello World"} example. The sources for these units
7107 are needed by the compiler and are kept together in one directory. Not
7108 all of the bodies are needed, but all of the sources are kept together
7109 anyway. In a normal installation, you need not specify these directory
7110 names when compiling or binding. Either the environment variables or
7111 the built-in defaults cause these files to be found.
7113 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
7114 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
7115 consisting of child units of @code{GNAT}. This is a collection of generally
7116 useful types, subprograms, etc. See the @cite{GNAT Reference Manual} for
7119 Besides simplifying access to the RTL, a major use of search paths is
7120 in compiling sources from multiple directories. This can make
7121 development environments much more flexible.
7123 @node Order of Compilation Issues
7124 @section Order of Compilation Issues
7127 If, in our earlier example, there was a spec for the @code{hello}
7128 procedure, it would be contained in the file @file{hello.ads}; yet this
7129 file would not have to be explicitly compiled. This is the result of the
7130 model we chose to implement library management. Some of the consequences
7131 of this model are as follows:
7135 There is no point in compiling specs (except for package
7136 specs with no bodies) because these are compiled as needed by clients. If
7137 you attempt a useless compilation, you will receive an error message.
7138 It is also useless to compile subunits because they are compiled as needed
7142 There are no order of compilation requirements: performing a
7143 compilation never obsoletes anything. The only way you can obsolete
7144 something and require recompilations is to modify one of the
7145 source files on which it depends.
7148 There is no library as such, apart from the ALI files
7149 (@pxref{The Ada Library Information Files}, for information on the format
7150 of these files). For now we find it convenient to create separate ALI files,
7151 but eventually the information therein may be incorporated into the object
7155 When you compile a unit, the source files for the specs of all units
7156 that it @code{with}'s, all its subunits, and the bodies of any generics it
7157 instantiates must be available (reachable by the search-paths mechanism
7158 described above), or you will receive a fatal error message.
7165 The following are some typical Ada compilation command line examples:
7168 @item $ gcc -c xyz.adb
7169 Compile body in file @file{xyz.adb} with all default options.
7172 @item $ gcc -c -O2 -gnata xyz-def.adb
7175 @item $ GNAT COMPILE /OPTIMIZE=ALL -gnata xyz-def.adb
7178 Compile the child unit package in file @file{xyz-def.adb} with extensive
7179 optimizations, and pragma @code{Assert}/@code{Debug} statements
7182 @item $ gcc -c -gnatc abc-def.adb
7183 Compile the subunit in file @file{abc-def.adb} in semantic-checking-only
7187 @node Binding Using gnatbind
7188 @chapter Binding Using @code{gnatbind}
7192 * Running gnatbind::
7193 * Switches for gnatbind::
7194 * Command-Line Access::
7195 * Search Paths for gnatbind::
7196 * Examples of gnatbind Usage::
7200 This chapter describes the GNAT binder, @code{gnatbind}, which is used
7201 to bind compiled GNAT objects.
7203 Note: to invoke @code{gnatbind} with a project file, use the @code{gnat}
7204 driver (see @ref{The GNAT Driver and Project Files}).
7206 The @code{gnatbind} program performs four separate functions:
7210 Checks that a program is consistent, in accordance with the rules in
7211 Chapter 10 of the Ada Reference Manual. In particular, error
7212 messages are generated if a program uses inconsistent versions of a
7216 Checks that an acceptable order of elaboration exists for the program
7217 and issues an error message if it cannot find an order of elaboration
7218 that satisfies the rules in Chapter 10 of the Ada Language Manual.
7221 Generates a main program incorporating the given elaboration order.
7222 This program is a small Ada package (body and spec) that
7223 must be subsequently compiled
7224 using the GNAT compiler. The necessary compilation step is usually
7225 performed automatically by @command{gnatlink}. The two most important
7226 functions of this program
7227 are to call the elaboration routines of units in an appropriate order
7228 and to call the main program.
7231 Determines the set of object files required by the given main program.
7232 This information is output in the forms of comments in the generated program,
7233 to be read by the @command{gnatlink} utility used to link the Ada application.
7236 @node Running gnatbind
7237 @section Running @code{gnatbind}
7240 The form of the @code{gnatbind} command is
7243 $ gnatbind [@i{switches}] @i{mainprog}[.ali] [@i{switches}]
7247 where @file{@i{mainprog}.adb} is the Ada file containing the main program
7248 unit body. If no switches are specified, @code{gnatbind} constructs an Ada
7249 package in two files whose names are
7250 @file{b~@i{mainprog}.ads}, and @file{b~@i{mainprog}.adb}.
7251 For example, if given the
7252 parameter @file{hello.ali}, for a main program contained in file
7253 @file{hello.adb}, the binder output files would be @file{b~hello.ads}
7254 and @file{b~hello.adb}.
7256 When doing consistency checking, the binder takes into consideration
7257 any source files it can locate. For example, if the binder determines
7258 that the given main program requires the package @code{Pack}, whose
7260 file is @file{pack.ali} and whose corresponding source spec file is
7261 @file{pack.ads}, it attempts to locate the source file @file{pack.ads}
7262 (using the same search path conventions as previously described for the
7263 @command{gcc} command). If it can locate this source file, it checks that
7265 or source checksums of the source and its references to in @file{ALI} files
7266 match. In other words, any @file{ALI} files that mentions this spec must have
7267 resulted from compiling this version of the source file (or in the case
7268 where the source checksums match, a version close enough that the
7269 difference does not matter).
7271 @cindex Source files, use by binder
7272 The effect of this consistency checking, which includes source files, is
7273 that the binder ensures that the program is consistent with the latest
7274 version of the source files that can be located at bind time. Editing a
7275 source file without compiling files that depend on the source file cause
7276 error messages to be generated by the binder.
7278 For example, suppose you have a main program @file{hello.adb} and a
7279 package @code{P}, from file @file{p.ads} and you perform the following
7284 Enter @code{gcc -c hello.adb} to compile the main program.
7287 Enter @code{gcc -c p.ads} to compile package @code{P}.
7290 Edit file @file{p.ads}.
7293 Enter @code{gnatbind hello}.
7297 At this point, the file @file{p.ali} contains an out-of-date time stamp
7298 because the file @file{p.ads} has been edited. The attempt at binding
7299 fails, and the binder generates the following error messages:
7302 error: "hello.adb" must be recompiled ("p.ads" has been modified)
7303 error: "p.ads" has been modified and must be recompiled
7307 Now both files must be recompiled as indicated, and then the bind can
7308 succeed, generating a main program. You need not normally be concerned
7309 with the contents of this file, but for reference purposes a sample
7310 binder output file is given in @ref{Example of Binder Output File}.
7312 In most normal usage, the default mode of @command{gnatbind} which is to
7313 generate the main package in Ada, as described in the previous section.
7314 In particular, this means that any Ada programmer can read and understand
7315 the generated main program. It can also be debugged just like any other
7316 Ada code provided the @option{^-g^/DEBUG^} switch is used for
7317 @command{gnatbind} and @command{gnatlink}.
7319 However for some purposes it may be convenient to generate the main
7320 program in C rather than Ada. This may for example be helpful when you
7321 are generating a mixed language program with the main program in C. The
7322 GNAT compiler itself is an example.
7323 The use of the @option{^-C^/BIND_FILE=C^} switch
7324 for both @code{gnatbind} and @command{gnatlink} will cause the program to
7325 be generated in C (and compiled using the gnu C compiler).
7327 @node Switches for gnatbind
7328 @section Switches for @command{gnatbind}
7331 The following switches are available with @code{gnatbind}; details will
7332 be presented in subsequent sections.
7335 * Consistency-Checking Modes::
7336 * Binder Error Message Control::
7337 * Elaboration Control::
7339 * Binding with Non-Ada Main Programs::
7340 * Binding Programs with No Main Subprogram::
7347 @cindex @option{-a} @command{gnatbind}
7348 Indicates that, if supported by the platform, the adainit procedure should
7349 be treated as an initialisation routine by the linker (a constructor). This
7350 is intended to be used by the Project Manager to automatically initialize
7351 shared Stand-Alone Libraries.
7353 @item ^-aO^/OBJECT_SEARCH^
7354 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatbind})
7355 Specify directory to be searched for ALI files.
7357 @item ^-aI^/SOURCE_SEARCH^
7358 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
7359 Specify directory to be searched for source file.
7361 @item ^-A^/BIND_FILE=ADA^
7362 @cindex @option{^-A^/BIND_FILE=ADA^} (@command{gnatbind})
7363 Generate binder program in Ada (default)
7365 @item ^-b^/REPORT_ERRORS=BRIEF^
7366 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@command{gnatbind})
7367 Generate brief messages to @file{stderr} even if verbose mode set.
7369 @item ^-c^/NOOUTPUT^
7370 @cindex @option{^-c^/NOOUTPUT^} (@command{gnatbind})
7371 Check only, no generation of binder output file.
7373 @item ^-C^/BIND_FILE=C^
7374 @cindex @option{^-C^/BIND_FILE=C^} (@command{gnatbind})
7375 Generate binder program in C
7377 @item ^-d^/DEFAULT_STACK_SIZE=^@var{nn}[k|m]
7378 @cindex @option{^-d^/DEFAULT_STACK_SIZE=^@var{nn}[k|m]} (@command{gnatbind})
7379 This switch can be used to change the default task stack size value
7380 to a specified size @var{nn}, which is expressed in bytes by default, or
7381 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
7383 In the absence of a [k|m] suffix, this switch is equivalent, in effect,
7384 to completing all task specs with
7385 @smallexample @c ada
7386 pragma Storage_Size (nn);
7388 When they do not already have such a pragma.
7390 @item ^-D^/DEFAULT_SECONDARY_STACK_SIZE=^@var{nn}[k|m]
7391 @cindex @option{^-D^/DEFAULT_SECONDARY_STACK_SIZE=nnnnn^} (@command{gnatbind})
7392 This switch can be used to change the default secondary stack size value
7393 to a specified size @var{nn}, which is expressed in bytes by default, or
7394 in kilobytes when suffixed with @var{k} or in megabytes when suffixed
7397 The secondary stack is used to deal with functions that return a variable
7398 sized result, for example a function returning an unconstrained
7399 String. There are two ways in which this secondary stack is allocated.
7401 For most targets, the secondary stack is growing on demand and is allocated
7402 as a chain of blocks in the heap. The -D option is not very
7403 relevant. It only give some control over the size of the allocated
7404 blocks (whose size is the minimum of the default secondary stack size value,
7405 and the actual size needed for the current allocation request).
7407 For certain targets, notably VxWorks 653,
7408 the secondary stack is allocated by carving off a fixed ratio chunk of the
7409 primary task stack. The -D option is used to defined the
7410 size of the environment task's secondary stack.
7412 @item ^-e^/ELABORATION_DEPENDENCIES^
7413 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@command{gnatbind})
7414 Output complete list of elaboration-order dependencies.
7416 @item ^-E^/STORE_TRACEBACKS^
7417 @cindex @option{^-E^/STORE_TRACEBACKS^} (@command{gnatbind})
7418 Store tracebacks in exception occurrences when the target supports it.
7419 This is the default with the zero cost exception mechanism.
7421 @c The following may get moved to an appendix
7422 This option is currently supported on the following targets:
7423 all x86 ports, Solaris, Windows, HP-UX, AIX, PowerPC VxWorks and Alpha VxWorks.
7425 See also the packages @code{GNAT.Traceback} and
7426 @code{GNAT.Traceback.Symbolic} for more information.
7428 Note that on x86 ports, you must not use @option{-fomit-frame-pointer}
7429 @command{gcc} option.
7432 @item ^-F^/FORCE_ELABS_FLAGS^
7433 @cindex @option{^-F^/FORCE_ELABS_FLAGS^} (@command{gnatbind})
7434 Force the checks of elaboration flags. @command{gnatbind} does not normally
7435 generate checks of elaboration flags for the main executable, except when
7436 a Stand-Alone Library is used. However, there are cases when this cannot be
7437 detected by gnatbind. An example is importing an interface of a Stand-Alone
7438 Library through a pragma Import and only specifying through a linker switch
7439 this Stand-Alone Library. This switch is used to guarantee that elaboration
7440 flag checks are generated.
7443 @cindex @option{^-h^/HELP^} (@command{gnatbind})
7444 Output usage (help) information
7447 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
7448 Specify directory to be searched for source and ALI files.
7450 @item ^-I-^/NOCURRENT_DIRECTORY^
7451 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatbind})
7452 Do not look for sources in the current directory where @code{gnatbind} was
7453 invoked, and do not look for ALI files in the directory containing the
7454 ALI file named in the @code{gnatbind} command line.
7456 @item ^-l^/ORDER_OF_ELABORATION^
7457 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@command{gnatbind})
7458 Output chosen elaboration order.
7460 @item ^-Lxxx^/BUILD_LIBRARY=xxx^
7461 @cindex @option{^-L^/BUILD_LIBRARY^} (@command{gnatbind})
7462 Bind the units for library building. In this case the adainit and
7463 adafinal procedures (@pxref{Binding with Non-Ada Main Programs})
7464 are renamed to ^xxxinit^XXXINIT^ and
7465 ^xxxfinal^XXXFINAL^.
7466 Implies ^-n^/NOCOMPILE^.
7468 (@xref{GNAT and Libraries}, for more details.)
7471 On OpenVMS, these init and final procedures are exported in uppercase
7472 letters. For example if /BUILD_LIBRARY=toto is used, the exported name of
7473 the init procedure will be "TOTOINIT" and the exported name of the final
7474 procedure will be "TOTOFINAL".
7477 @item ^-Mxyz^/RENAME_MAIN=xyz^
7478 @cindex @option{^-M^/RENAME_MAIN^} (@command{gnatbind})
7479 Rename generated main program from main to xyz. This option is
7480 supported on cross environments only.
7482 @item ^-m^/ERROR_LIMIT=^@var{n}
7483 @cindex @option{^-m^/ERROR_LIMIT^} (@command{gnatbind})
7484 Limit number of detected errors to @var{n}, where @var{n} is
7485 in the range 1..999_999. The default value if no switch is
7486 given is 9999. Binding is terminated if the limit is exceeded.
7488 Furthermore, under Windows, the sources pointed to by the libraries path
7489 set in the registry are not searched for.
7493 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
7497 @cindex @option{-nostdinc} (@command{gnatbind})
7498 Do not look for sources in the system default directory.
7501 @cindex @option{-nostdlib} (@command{gnatbind})
7502 Do not look for library files in the system default directory.
7504 @item --RTS=@var{rts-path}
7505 @cindex @option{--RTS} (@code{gnatbind})
7506 Specifies the default location of the runtime library. Same meaning as the
7507 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
7509 @item ^-o ^/OUTPUT=^@var{file}
7510 @cindex @option{^-o ^/OUTPUT^} (@command{gnatbind})
7511 Name the output file @var{file} (default is @file{b~@var{xxx}.adb}).
7512 Note that if this option is used, then linking must be done manually,
7513 gnatlink cannot be used.
7515 @item ^-O^/OBJECT_LIST^
7516 @cindex @option{^-O^/OBJECT_LIST^} (@command{gnatbind})
7519 @item ^-p^/PESSIMISTIC_ELABORATION^
7520 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@command{gnatbind})
7521 Pessimistic (worst-case) elaboration order
7524 @cindex @option{^-R^-R^} (@command{gnatbind})
7525 Output closure source list.
7527 @item ^-s^/READ_SOURCES=ALL^
7528 @cindex @option{^-s^/READ_SOURCES=ALL^} (@command{gnatbind})
7529 Require all source files to be present.
7531 @item ^-S@var{xxx}^/INITIALIZE_SCALARS=@var{xxx}^
7532 @cindex @option{^-S^/INITIALIZE_SCALARS^} (@command{gnatbind})
7533 Specifies the value to be used when detecting uninitialized scalar
7534 objects with pragma Initialize_Scalars.
7535 The @var{xxx} ^string specified with the switch^option^ may be either
7537 @item ``@option{^in^INVALID^}'' requesting an invalid value where possible
7538 @item ``@option{^lo^LOW^}'' for the lowest possible value
7539 @item ``@option{^hi^HIGH^}'' for the highest possible value
7540 @item ``@option{xx}'' for a value consisting of repeated bytes with the
7541 value 16#xx# (i.e. xx is a string of two hexadecimal digits).
7544 In addition, you can specify @option{-Sev} to indicate that the value is
7545 to be set at run time. In this case, the program will look for an environment
7546 @cindex GNAT_INIT_SCALARS
7547 variable of the form @code{GNAT_INIT_SCALARS=xx}, where xx is one
7548 of @option{in/lo/hi/xx} with the same meanings as above.
7549 If no environment variable is found, or if it does not have a valid value,
7550 then the default is @option{in} (invalid values).
7554 @cindex @option{-static} (@code{gnatbind})
7555 Link against a static GNAT run time.
7558 @cindex @option{-shared} (@code{gnatbind})
7559 Link against a shared GNAT run time when available.
7562 @item ^-t^/NOTIME_STAMP_CHECK^
7563 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
7564 Tolerate time stamp and other consistency errors
7566 @item ^-T@var{n}^/TIME_SLICE=@var{n}^
7567 @cindex @option{^-T^/TIME_SLICE^} (@code{gnatbind})
7568 Set the time slice value to @var{n} milliseconds. If the system supports
7569 the specification of a specific time slice value, then the indicated value
7570 is used. If the system does not support specific time slice values, but
7571 does support some general notion of round-robin scheduling, then any
7572 nonzero value will activate round-robin scheduling.
7574 A value of zero is treated specially. It turns off time
7575 slicing, and in addition, indicates to the tasking run time that the
7576 semantics should match as closely as possible the Annex D
7577 requirements of the Ada RM, and in particular sets the default
7578 scheduling policy to @code{FIFO_Within_Priorities}.
7580 @item ^-u@var{n}^/DYNAMIC_STACK_USAGE=@var{n}^
7581 @cindex @option{^-u^/DYNAMIC_STACK_USAGE^} (@code{gnatbind})
7582 Enable dynamic stack usage, with n result stored and displayed at program
7583 termination. Results that can't be stored are displayed on the fly, at task
7584 termination. This option is currently not supported on OpenVMS I64 platforms.
7586 @item ^-v^/REPORT_ERRORS=VERBOSE^
7587 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
7588 Verbose mode. Write error messages, header, summary output to
7593 @cindex @option{-w} (@code{gnatbind})
7594 Warning mode (@var{x}=s/e for suppress/treat as error)
7598 @item /WARNINGS=NORMAL
7599 @cindex @option{/WARNINGS} (@code{gnatbind})
7600 Normal warnings mode. Warnings are issued but ignored
7602 @item /WARNINGS=SUPPRESS
7603 @cindex @option{/WARNINGS} (@code{gnatbind})
7604 All warning messages are suppressed
7606 @item /WARNINGS=ERROR
7607 @cindex @option{/WARNINGS} (@code{gnatbind})
7608 Warning messages are treated as fatal errors
7611 @item ^-x^/READ_SOURCES=NONE^
7612 @cindex @option{^-x^/READ_SOURCES^} (@code{gnatbind})
7613 Exclude source files (check object consistency only).
7616 @item /READ_SOURCES=AVAILABLE
7617 @cindex @option{/READ_SOURCES} (@code{gnatbind})
7618 Default mode, in which sources are checked for consistency only if
7622 @item ^-z^/ZERO_MAIN^
7623 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
7629 You may obtain this listing of switches by running @code{gnatbind} with
7633 @node Consistency-Checking Modes
7634 @subsection Consistency-Checking Modes
7637 As described earlier, by default @code{gnatbind} checks
7638 that object files are consistent with one another and are consistent
7639 with any source files it can locate. The following switches control binder
7644 @item ^-s^/READ_SOURCES=ALL^
7645 @cindex @option{^-s^/READ_SOURCES=ALL^} (@code{gnatbind})
7646 Require source files to be present. In this mode, the binder must be
7647 able to locate all source files that are referenced, in order to check
7648 their consistency. In normal mode, if a source file cannot be located it
7649 is simply ignored. If you specify this switch, a missing source
7652 @item ^-x^/READ_SOURCES=NONE^
7653 @cindex @option{^-x^/READ_SOURCES=NONE^} (@code{gnatbind})
7654 Exclude source files. In this mode, the binder only checks that ALI
7655 files are consistent with one another. Source files are not accessed.
7656 The binder runs faster in this mode, and there is still a guarantee that
7657 the resulting program is self-consistent.
7658 If a source file has been edited since it was last compiled, and you
7659 specify this switch, the binder will not detect that the object
7660 file is out of date with respect to the source file. Note that this is the
7661 mode that is automatically used by @command{gnatmake} because in this
7662 case the checking against sources has already been performed by
7663 @command{gnatmake} in the course of compilation (i.e. before binding).
7666 @item /READ_SOURCES=AVAILABLE
7667 @cindex @code{/READ_SOURCES=AVAILABLE} (@code{gnatbind})
7668 This is the default mode in which source files are checked if they are
7669 available, and ignored if they are not available.
7673 @node Binder Error Message Control
7674 @subsection Binder Error Message Control
7677 The following switches provide control over the generation of error
7678 messages from the binder:
7682 @item ^-v^/REPORT_ERRORS=VERBOSE^
7683 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
7684 Verbose mode. In the normal mode, brief error messages are generated to
7685 @file{stderr}. If this switch is present, a header is written
7686 to @file{stdout} and any error messages are directed to @file{stdout}.
7687 All that is written to @file{stderr} is a brief summary message.
7689 @item ^-b^/REPORT_ERRORS=BRIEF^
7690 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@code{gnatbind})
7691 Generate brief error messages to @file{stderr} even if verbose mode is
7692 specified. This is relevant only when used with the
7693 @option{^-v^/REPORT_ERRORS=VERBOSE^} switch.
7697 @cindex @option{-m} (@code{gnatbind})
7698 Limits the number of error messages to @var{n}, a decimal integer in the
7699 range 1-999. The binder terminates immediately if this limit is reached.
7702 @cindex @option{-M} (@code{gnatbind})
7703 Renames the generated main program from @code{main} to @code{xxx}.
7704 This is useful in the case of some cross-building environments, where
7705 the actual main program is separate from the one generated
7709 @item ^-ws^/WARNINGS=SUPPRESS^
7710 @cindex @option{^-ws^/WARNINGS=SUPPRESS^} (@code{gnatbind})
7712 Suppress all warning messages.
7714 @item ^-we^/WARNINGS=ERROR^
7715 @cindex @option{^-we^/WARNINGS=ERROR^} (@code{gnatbind})
7716 Treat any warning messages as fatal errors.
7719 @item /WARNINGS=NORMAL
7720 Standard mode with warnings generated, but warnings do not get treated
7724 @item ^-t^/NOTIME_STAMP_CHECK^
7725 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
7726 @cindex Time stamp checks, in binder
7727 @cindex Binder consistency checks
7728 @cindex Consistency checks, in binder
7729 The binder performs a number of consistency checks including:
7733 Check that time stamps of a given source unit are consistent
7735 Check that checksums of a given source unit are consistent
7737 Check that consistent versions of @code{GNAT} were used for compilation
7739 Check consistency of configuration pragmas as required
7743 Normally failure of such checks, in accordance with the consistency
7744 requirements of the Ada Reference Manual, causes error messages to be
7745 generated which abort the binder and prevent the output of a binder
7746 file and subsequent link to obtain an executable.
7748 The @option{^-t^/NOTIME_STAMP_CHECK^} switch converts these error messages
7749 into warnings, so that
7750 binding and linking can continue to completion even in the presence of such
7751 errors. The result may be a failed link (due to missing symbols), or a
7752 non-functional executable which has undefined semantics.
7753 @emph{This means that
7754 @option{^-t^/NOTIME_STAMP_CHECK^} should be used only in unusual situations,
7758 @node Elaboration Control
7759 @subsection Elaboration Control
7762 The following switches provide additional control over the elaboration
7763 order. For full details see @ref{Elaboration Order Handling in GNAT}.
7766 @item ^-p^/PESSIMISTIC_ELABORATION^
7767 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@code{gnatbind})
7768 Normally the binder attempts to choose an elaboration order that is
7769 likely to minimize the likelihood of an elaboration order error resulting
7770 in raising a @code{Program_Error} exception. This switch reverses the
7771 action of the binder, and requests that it deliberately choose an order
7772 that is likely to maximize the likelihood of an elaboration error.
7773 This is useful in ensuring portability and avoiding dependence on
7774 accidental fortuitous elaboration ordering.
7776 Normally it only makes sense to use the @option{^-p^/PESSIMISTIC_ELABORATION^}
7778 elaboration checking is used (@option{-gnatE} switch used for compilation).
7779 This is because in the default static elaboration mode, all necessary
7780 @code{Elaborate} and @code{Elaborate_All} pragmas are implicitly inserted.
7781 These implicit pragmas are still respected by the binder in
7782 @option{^-p^/PESSIMISTIC_ELABORATION^} mode, so a
7783 safe elaboration order is assured.
7786 @node Output Control
7787 @subsection Output Control
7790 The following switches allow additional control over the output
7791 generated by the binder.
7796 @item ^-A^/BIND_FILE=ADA^
7797 @cindex @option{^-A^/BIND_FILE=ADA^} (@code{gnatbind})
7798 Generate binder program in Ada (default). The binder program is named
7799 @file{b~@var{mainprog}.adb} by default. This can be changed with
7800 @option{^-o^/OUTPUT^} @code{gnatbind} option.
7802 @item ^-c^/NOOUTPUT^
7803 @cindex @option{^-c^/NOOUTPUT^} (@code{gnatbind})
7804 Check only. Do not generate the binder output file. In this mode the
7805 binder performs all error checks but does not generate an output file.
7807 @item ^-C^/BIND_FILE=C^
7808 @cindex @option{^-C^/BIND_FILE=C^} (@code{gnatbind})
7809 Generate binder program in C. The binder program is named
7810 @file{b_@var{mainprog}.c}.
7811 This can be changed with @option{^-o^/OUTPUT^} @code{gnatbind}
7814 @item ^-e^/ELABORATION_DEPENDENCIES^
7815 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@code{gnatbind})
7816 Output complete list of elaboration-order dependencies, showing the
7817 reason for each dependency. This output can be rather extensive but may
7818 be useful in diagnosing problems with elaboration order. The output is
7819 written to @file{stdout}.
7822 @cindex @option{^-h^/HELP^} (@code{gnatbind})
7823 Output usage information. The output is written to @file{stdout}.
7825 @item ^-K^/LINKER_OPTION_LIST^
7826 @cindex @option{^-K^/LINKER_OPTION_LIST^} (@code{gnatbind})
7827 Output linker options to @file{stdout}. Includes library search paths,
7828 contents of pragmas Ident and Linker_Options, and libraries added
7831 @item ^-l^/ORDER_OF_ELABORATION^
7832 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@code{gnatbind})
7833 Output chosen elaboration order. The output is written to @file{stdout}.
7835 @item ^-O^/OBJECT_LIST^
7836 @cindex @option{^-O^/OBJECT_LIST^} (@code{gnatbind})
7837 Output full names of all the object files that must be linked to provide
7838 the Ada component of the program. The output is written to @file{stdout}.
7839 This list includes the files explicitly supplied and referenced by the user
7840 as well as implicitly referenced run-time unit files. The latter are
7841 omitted if the corresponding units reside in shared libraries. The
7842 directory names for the run-time units depend on the system configuration.
7844 @item ^-o ^/OUTPUT=^@var{file}
7845 @cindex @option{^-o^/OUTPUT^} (@code{gnatbind})
7846 Set name of output file to @var{file} instead of the normal
7847 @file{b~@var{mainprog}.adb} default. Note that @var{file} denote the Ada
7848 binder generated body filename. In C mode you would normally give
7849 @var{file} an extension of @file{.c} because it will be a C source program.
7850 Note that if this option is used, then linking must be done manually.
7851 It is not possible to use gnatlink in this case, since it cannot locate
7854 @item ^-r^/RESTRICTION_LIST^
7855 @cindex @option{^-r^/RESTRICTION_LIST^} (@code{gnatbind})
7856 Generate list of @code{pragma Restrictions} that could be applied to
7857 the current unit. This is useful for code audit purposes, and also may
7858 be used to improve code generation in some cases.
7862 @node Binding with Non-Ada Main Programs
7863 @subsection Binding with Non-Ada Main Programs
7866 In our description so far we have assumed that the main
7867 program is in Ada, and that the task of the binder is to generate a
7868 corresponding function @code{main} that invokes this Ada main
7869 program. GNAT also supports the building of executable programs where
7870 the main program is not in Ada, but some of the called routines are
7871 written in Ada and compiled using GNAT (@pxref{Mixed Language Programming}).
7872 The following switch is used in this situation:
7876 @cindex @option{^-n^/NOMAIN^} (@code{gnatbind})
7877 No main program. The main program is not in Ada.
7881 In this case, most of the functions of the binder are still required,
7882 but instead of generating a main program, the binder generates a file
7883 containing the following callable routines:
7888 You must call this routine to initialize the Ada part of the program by
7889 calling the necessary elaboration routines. A call to @code{adainit} is
7890 required before the first call to an Ada subprogram.
7892 Note that it is assumed that the basic execution environment must be setup
7893 to be appropriate for Ada execution at the point where the first Ada
7894 subprogram is called. In particular, if the Ada code will do any
7895 floating-point operations, then the FPU must be setup in an appropriate
7896 manner. For the case of the x86, for example, full precision mode is
7897 required. The procedure GNAT.Float_Control.Reset may be used to ensure
7898 that the FPU is in the right state.
7902 You must call this routine to perform any library-level finalization
7903 required by the Ada subprograms. A call to @code{adafinal} is required
7904 after the last call to an Ada subprogram, and before the program
7909 If the @option{^-n^/NOMAIN^} switch
7910 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
7911 @cindex Binder, multiple input files
7912 is given, more than one ALI file may appear on
7913 the command line for @code{gnatbind}. The normal @dfn{closure}
7914 calculation is performed for each of the specified units. Calculating
7915 the closure means finding out the set of units involved by tracing
7916 @code{with} references. The reason it is necessary to be able to
7917 specify more than one ALI file is that a given program may invoke two or
7918 more quite separate groups of Ada units.
7920 The binder takes the name of its output file from the last specified ALI
7921 file, unless overridden by the use of the @option{^-o file^/OUTPUT=file^}.
7922 @cindex @option{^-o^/OUTPUT^} (@command{gnatbind})
7923 The output is an Ada unit in source form that can
7924 be compiled with GNAT unless the -C switch is used in which case the
7925 output is a C source file, which must be compiled using the C compiler.
7926 This compilation occurs automatically as part of the @command{gnatlink}
7929 Currently the GNAT run time requires a FPU using 80 bits mode
7930 precision. Under targets where this is not the default it is required to
7931 call GNAT.Float_Control.Reset before using floating point numbers (this
7932 include float computation, float input and output) in the Ada code. A
7933 side effect is that this could be the wrong mode for the foreign code
7934 where floating point computation could be broken after this call.
7936 @node Binding Programs with No Main Subprogram
7937 @subsection Binding Programs with No Main Subprogram
7940 It is possible to have an Ada program which does not have a main
7941 subprogram. This program will call the elaboration routines of all the
7942 packages, then the finalization routines.
7944 The following switch is used to bind programs organized in this manner:
7947 @item ^-z^/ZERO_MAIN^
7948 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
7949 Normally the binder checks that the unit name given on the command line
7950 corresponds to a suitable main subprogram. When this switch is used,
7951 a list of ALI files can be given, and the execution of the program
7952 consists of elaboration of these units in an appropriate order.
7955 @node Command-Line Access
7956 @section Command-Line Access
7959 The package @code{Ada.Command_Line} provides access to the command-line
7960 arguments and program name. In order for this interface to operate
7961 correctly, the two variables
7973 are declared in one of the GNAT library routines. These variables must
7974 be set from the actual @code{argc} and @code{argv} values passed to the
7975 main program. With no @option{^n^/NOMAIN^} present, @code{gnatbind}
7976 generates the C main program to automatically set these variables.
7977 If the @option{^n^/NOMAIN^} switch is used, there is no automatic way to
7978 set these variables. If they are not set, the procedures in
7979 @code{Ada.Command_Line} will not be available, and any attempt to use
7980 them will raise @code{Constraint_Error}. If command line access is
7981 required, your main program must set @code{gnat_argc} and
7982 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
7985 @node Search Paths for gnatbind
7986 @section Search Paths for @code{gnatbind}
7989 The binder takes the name of an ALI file as its argument and needs to
7990 locate source files as well as other ALI files to verify object consistency.
7992 For source files, it follows exactly the same search rules as @command{gcc}
7993 (@pxref{Search Paths and the Run-Time Library (RTL)}). For ALI files the
7994 directories searched are:
7998 The directory containing the ALI file named in the command line, unless
7999 the switch @option{^-I-^/NOCURRENT_DIRECTORY^} is specified.
8002 All directories specified by @option{^-I^/SEARCH^}
8003 switches on the @code{gnatbind}
8004 command line, in the order given.
8007 @findex ADA_PRJ_OBJECTS_FILE
8008 Each of the directories listed in the text file whose name is given
8009 by the @code{ADA_PRJ_OBJECTS_FILE} ^environment variable^logical name^.
8012 @code{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
8013 driver when project files are used. It should not normally be set
8017 @findex ADA_OBJECTS_PATH
8018 Each of the directories listed in the value of the
8019 @code{ADA_OBJECTS_PATH} ^environment variable^logical name^.
8021 Construct this value
8022 exactly as the @code{PATH} environment variable: a list of directory
8023 names separated by colons (semicolons when working with the NT version
8027 Normally, define this value as a logical name containing a comma separated
8028 list of directory names.
8030 This variable can also be defined by means of an environment string
8031 (an argument to the HP C exec* set of functions).
8035 DEFINE ANOTHER_PATH FOO:[BAG]
8036 DEFINE ADA_OBJECTS_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
8039 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
8040 first, followed by the standard Ada
8041 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB].
8042 If this is not redefined, the user will obtain the HP Ada 83 IO packages
8043 (Text_IO, Sequential_IO, etc)
8044 instead of the standard Ada packages. Thus, in order to get the standard Ada
8045 packages by default, ADA_OBJECTS_PATH must be redefined.
8049 The content of the @file{ada_object_path} file which is part of the GNAT
8050 installation tree and is used to store standard libraries such as the
8051 GNAT Run Time Library (RTL) unless the switch @option{-nostdlib} is
8054 @ref{Installing a library}
8059 In the binder the switch @option{^-I^/SEARCH^}
8060 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
8061 is used to specify both source and
8062 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
8063 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
8064 instead if you want to specify
8065 source paths only, and @option{^-aO^/LIBRARY_SEARCH^}
8066 @cindex @option{^-aO^/LIBRARY_SEARCH^} (@command{gnatbind})
8067 if you want to specify library paths
8068 only. This means that for the binder
8069 @option{^-I^/SEARCH=^}@var{dir} is equivalent to
8070 @option{^-aI^/SOURCE_SEARCH=^}@var{dir}
8071 @option{^-aO^/OBJECT_SEARCH=^}@var{dir}.
8072 The binder generates the bind file (a C language source file) in the
8073 current working directory.
8079 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
8080 children make up the GNAT Run-Time Library, together with the package
8081 GNAT and its children, which contain a set of useful additional
8082 library functions provided by GNAT. The sources for these units are
8083 needed by the compiler and are kept together in one directory. The ALI
8084 files and object files generated by compiling the RTL are needed by the
8085 binder and the linker and are kept together in one directory, typically
8086 different from the directory containing the sources. In a normal
8087 installation, you need not specify these directory names when compiling
8088 or binding. Either the environment variables or the built-in defaults
8089 cause these files to be found.
8091 Besides simplifying access to the RTL, a major use of search paths is
8092 in compiling sources from multiple directories. This can make
8093 development environments much more flexible.
8095 @node Examples of gnatbind Usage
8096 @section Examples of @code{gnatbind} Usage
8099 This section contains a number of examples of using the GNAT binding
8100 utility @code{gnatbind}.
8103 @item gnatbind hello
8104 The main program @code{Hello} (source program in @file{hello.adb}) is
8105 bound using the standard switch settings. The generated main program is
8106 @file{b~hello.adb}. This is the normal, default use of the binder.
8109 @item gnatbind hello -o mainprog.adb
8112 @item gnatbind HELLO.ALI /OUTPUT=Mainprog.ADB
8114 The main program @code{Hello} (source program in @file{hello.adb}) is
8115 bound using the standard switch settings. The generated main program is
8116 @file{mainprog.adb} with the associated spec in
8117 @file{mainprog.ads}. Note that you must specify the body here not the
8118 spec, in the case where the output is in Ada. Note that if this option
8119 is used, then linking must be done manually, since gnatlink will not
8120 be able to find the generated file.
8123 @item gnatbind main -C -o mainprog.c -x
8126 @item gnatbind MAIN.ALI /BIND_FILE=C /OUTPUT=Mainprog.C /READ_SOURCES=NONE
8128 The main program @code{Main} (source program in
8129 @file{main.adb}) is bound, excluding source files from the
8130 consistency checking, generating
8131 the file @file{mainprog.c}.
8134 @item gnatbind -x main_program -C -o mainprog.c
8135 This command is exactly the same as the previous example. Switches may
8136 appear anywhere in the command line, and single letter switches may be
8137 combined into a single switch.
8141 @item gnatbind -n math dbase -C -o ada-control.c
8144 @item gnatbind /NOMAIN math dbase /BIND_FILE=C /OUTPUT=ada-control.c
8146 The main program is in a language other than Ada, but calls to
8147 subprograms in packages @code{Math} and @code{Dbase} appear. This call
8148 to @code{gnatbind} generates the file @file{ada-control.c} containing
8149 the @code{adainit} and @code{adafinal} routines to be called before and
8150 after accessing the Ada units.
8153 @c ------------------------------------
8154 @node Linking Using gnatlink
8155 @chapter Linking Using @command{gnatlink}
8156 @c ------------------------------------
8160 This chapter discusses @command{gnatlink}, a tool that links
8161 an Ada program and builds an executable file. This utility
8162 invokes the system linker ^(via the @command{gcc} command)^^
8163 with a correct list of object files and library references.
8164 @command{gnatlink} automatically determines the list of files and
8165 references for the Ada part of a program. It uses the binder file
8166 generated by the @command{gnatbind} to determine this list.
8168 Note: to invoke @code{gnatlink} with a project file, use the @code{gnat}
8169 driver (see @ref{The GNAT Driver and Project Files}).
8172 * Running gnatlink::
8173 * Switches for gnatlink::
8176 @node Running gnatlink
8177 @section Running @command{gnatlink}
8180 The form of the @command{gnatlink} command is
8183 $ gnatlink [@var{switches}] @var{mainprog}[.ali]
8184 [@var{non-Ada objects}] [@var{linker options}]
8188 The arguments of @command{gnatlink} (switches, main @file{ALI} file,
8190 or linker options) may be in any order, provided that no non-Ada object may
8191 be mistaken for a main @file{ALI} file.
8192 Any file name @file{F} without the @file{.ali}
8193 extension will be taken as the main @file{ALI} file if a file exists
8194 whose name is the concatenation of @file{F} and @file{.ali}.
8197 @file{@var{mainprog}.ali} references the ALI file of the main program.
8198 The @file{.ali} extension of this file can be omitted. From this
8199 reference, @command{gnatlink} locates the corresponding binder file
8200 @file{b~@var{mainprog}.adb} and, using the information in this file along
8201 with the list of non-Ada objects and linker options, constructs a
8202 linker command file to create the executable.
8204 The arguments other than the @command{gnatlink} switches and the main
8205 @file{ALI} file are passed to the linker uninterpreted.
8206 They typically include the names of
8207 object files for units written in other languages than Ada and any library
8208 references required to resolve references in any of these foreign language
8209 units, or in @code{Import} pragmas in any Ada units.
8211 @var{linker options} is an optional list of linker specific
8213 The default linker called by gnatlink is @var{gcc} which in
8214 turn calls the appropriate system linker.
8215 Standard options for the linker such as @option{-lmy_lib} or
8216 @option{-Ldir} can be added as is.
8217 For options that are not recognized by
8218 @var{gcc} as linker options, use the @var{gcc} switches @option{-Xlinker} or
8220 Refer to the GCC documentation for
8221 details. Here is an example showing how to generate a linker map:
8224 $ ^gnatlink my_prog -Wl,-Map,MAPFILE^GNAT LINK my_prog.ali /MAP^
8227 Using @var{linker options} it is possible to set the program stack and
8230 See @ref{Setting Stack Size from gnatlink} and
8231 @ref{Setting Heap Size from gnatlink}.
8234 @command{gnatlink} determines the list of objects required by the Ada
8235 program and prepends them to the list of objects passed to the linker.
8236 @command{gnatlink} also gathers any arguments set by the use of
8237 @code{pragma Linker_Options} and adds them to the list of arguments
8238 presented to the linker.
8241 @command{gnatlink} accepts the following types of extra files on the command
8242 line: objects (.OBJ), libraries (.OLB), sharable images (.EXE), and
8243 options files (.OPT). These are recognized and handled according to their
8247 @node Switches for gnatlink
8248 @section Switches for @command{gnatlink}
8251 The following switches are available with the @command{gnatlink} utility:
8256 @item ^-A^/BIND_FILE=ADA^
8257 @cindex @option{^-A^/BIND_FILE=ADA^} (@command{gnatlink})
8258 The binder has generated code in Ada. This is the default.
8260 @item ^-C^/BIND_FILE=C^
8261 @cindex @option{^-C^/BIND_FILE=C^} (@command{gnatlink})
8262 If instead of generating a file in Ada, the binder has generated one in
8263 C, then the linker needs to know about it. Use this switch to signal
8264 to @command{gnatlink} that the binder has generated C code rather than
8267 @item ^-f^/FORCE_OBJECT_FILE_LIST^
8268 @cindex Command line length
8269 @cindex @option{^-f^/FORCE_OBJECT_FILE_LIST^} (@command{gnatlink})
8270 On some targets, the command line length is limited, and @command{gnatlink}
8271 will generate a separate file for the linker if the list of object files
8273 The @option{^-f^/FORCE_OBJECT_FILE_LIST^} switch forces this file
8274 to be generated even if
8275 the limit is not exceeded. This is useful in some cases to deal with
8276 special situations where the command line length is exceeded.
8279 @cindex Debugging information, including
8280 @cindex @option{^-g^/DEBUG^} (@command{gnatlink})
8281 The option to include debugging information causes the Ada bind file (in
8282 other words, @file{b~@var{mainprog}.adb}) to be compiled with
8283 @option{^-g^/DEBUG^}.
8284 In addition, the binder does not delete the @file{b~@var{mainprog}.adb},
8285 @file{b~@var{mainprog}.o} and @file{b~@var{mainprog}.ali} files.
8286 Without @option{^-g^/DEBUG^}, the binder removes these files by
8287 default. The same procedure apply if a C bind file was generated using
8288 @option{^-C^/BIND_FILE=C^} @code{gnatbind} option, in this case the filenames
8289 are @file{b_@var{mainprog}.c} and @file{b_@var{mainprog}.o}.
8291 @item ^-n^/NOCOMPILE^
8292 @cindex @option{^-n^/NOCOMPILE^} (@command{gnatlink})
8293 Do not compile the file generated by the binder. This may be used when
8294 a link is rerun with different options, but there is no need to recompile
8298 @cindex @option{^-v^/VERBOSE^} (@command{gnatlink})
8299 Causes additional information to be output, including a full list of the
8300 included object files. This switch option is most useful when you want
8301 to see what set of object files are being used in the link step.
8303 @item ^-v -v^/VERBOSE/VERBOSE^
8304 @cindex @option{^-v -v^/VERBOSE/VERBOSE^} (@command{gnatlink})
8305 Very verbose mode. Requests that the compiler operate in verbose mode when
8306 it compiles the binder file, and that the system linker run in verbose mode.
8308 @item ^-o ^/EXECUTABLE=^@var{exec-name}
8309 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatlink})
8310 @var{exec-name} specifies an alternate name for the generated
8311 executable program. If this switch is omitted, the executable has the same
8312 name as the main unit. For example, @code{gnatlink try.ali} creates
8313 an executable called @file{^try^TRY.EXE^}.
8316 @item -b @var{target}
8317 @cindex @option{-b} (@command{gnatlink})
8318 Compile your program to run on @var{target}, which is the name of a
8319 system configuration. You must have a GNAT cross-compiler built if
8320 @var{target} is not the same as your host system.
8323 @cindex @option{-B} (@command{gnatlink})
8324 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
8325 from @var{dir} instead of the default location. Only use this switch
8326 when multiple versions of the GNAT compiler are available. See the
8327 @command{gcc} manual page for further details. You would normally use the
8328 @option{-b} or @option{-V} switch instead.
8330 @item --GCC=@var{compiler_name}
8331 @cindex @option{--GCC=compiler_name} (@command{gnatlink})
8332 Program used for compiling the binder file. The default is
8333 @command{gcc}. You need to use quotes around @var{compiler_name} if
8334 @code{compiler_name} contains spaces or other separator characters.
8335 As an example @option{--GCC="foo -x -y"} will instruct @command{gnatlink} to
8336 use @code{foo -x -y} as your compiler. Note that switch @option{-c} is always
8337 inserted after your command name. Thus in the above example the compiler
8338 command that will be used by @command{gnatlink} will be @code{foo -c -x -y}.
8339 A limitation of this syntax is that the name and path name of the executable
8340 itself must not include any embedded spaces. If several
8341 @option{--GCC=compiler_name} are used, only the last @var{compiler_name}
8342 is taken into account. However, all the additional switches are also taken
8344 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
8345 @option{--GCC="bar -x -y -z -t"}.
8347 @item --LINK=@var{name}
8348 @cindex @option{--LINK=} (@command{gnatlink})
8349 @var{name} is the name of the linker to be invoked. This is especially
8350 useful in mixed language programs since languages such as C++ require
8351 their own linker to be used. When this switch is omitted, the default
8352 name for the linker is @command{gcc}. When this switch is used, the
8353 specified linker is called instead of @command{gcc} with exactly the same
8354 parameters that would have been passed to @command{gcc} so if the desired
8355 linker requires different parameters it is necessary to use a wrapper
8356 script that massages the parameters before invoking the real linker. It
8357 may be useful to control the exact invocation by using the verbose
8363 @item /DEBUG=TRACEBACK
8364 @cindex @code{/DEBUG=TRACEBACK} (@command{gnatlink})
8365 This qualifier causes sufficient information to be included in the
8366 executable file to allow a traceback, but does not include the full
8367 symbol information needed by the debugger.
8369 @item /IDENTIFICATION="<string>"
8370 @code{"<string>"} specifies the string to be stored in the image file
8371 identification field in the image header.
8372 It overrides any pragma @code{Ident} specified string.
8374 @item /NOINHIBIT-EXEC
8375 Generate the executable file even if there are linker warnings.
8377 @item /NOSTART_FILES
8378 Don't link in the object file containing the ``main'' transfer address.
8379 Used when linking with a foreign language main program compiled with an
8383 Prefer linking with object libraries over sharable images, even without
8389 @node The GNAT Make Program gnatmake
8390 @chapter The GNAT Make Program @command{gnatmake}
8394 * Running gnatmake::
8395 * Switches for gnatmake::
8396 * Mode Switches for gnatmake::
8397 * Notes on the Command Line::
8398 * How gnatmake Works::
8399 * Examples of gnatmake Usage::
8402 A typical development cycle when working on an Ada program consists of
8403 the following steps:
8407 Edit some sources to fix bugs.
8413 Compile all sources affected.
8423 The third step can be tricky, because not only do the modified files
8424 @cindex Dependency rules
8425 have to be compiled, but any files depending on these files must also be
8426 recompiled. The dependency rules in Ada can be quite complex, especially
8427 in the presence of overloading, @code{use} clauses, generics and inlined
8430 @command{gnatmake} automatically takes care of the third and fourth steps
8431 of this process. It determines which sources need to be compiled,
8432 compiles them, and binds and links the resulting object files.
8434 Unlike some other Ada make programs, the dependencies are always
8435 accurately recomputed from the new sources. The source based approach of
8436 the GNAT compilation model makes this possible. This means that if
8437 changes to the source program cause corresponding changes in
8438 dependencies, they will always be tracked exactly correctly by
8441 @node Running gnatmake
8442 @section Running @command{gnatmake}
8445 The usual form of the @command{gnatmake} command is
8448 $ gnatmake [@var{switches}] @var{file_name}
8449 [@var{file_names}] [@var{mode_switches}]
8453 The only required argument is one @var{file_name}, which specifies
8454 a compilation unit that is a main program. Several @var{file_names} can be
8455 specified: this will result in several executables being built.
8456 If @code{switches} are present, they can be placed before the first
8457 @var{file_name}, between @var{file_names} or after the last @var{file_name}.
8458 If @var{mode_switches} are present, they must always be placed after
8459 the last @var{file_name} and all @code{switches}.
8461 If you are using standard file extensions (.adb and .ads), then the
8462 extension may be omitted from the @var{file_name} arguments. However, if
8463 you are using non-standard extensions, then it is required that the
8464 extension be given. A relative or absolute directory path can be
8465 specified in a @var{file_name}, in which case, the input source file will
8466 be searched for in the specified directory only. Otherwise, the input
8467 source file will first be searched in the directory where
8468 @command{gnatmake} was invoked and if it is not found, it will be search on
8469 the source path of the compiler as described in
8470 @ref{Search Paths and the Run-Time Library (RTL)}.
8472 All @command{gnatmake} output (except when you specify
8473 @option{^-M^/DEPENDENCIES_LIST^}) is to
8474 @file{stderr}. The output produced by the
8475 @option{^-M^/DEPENDENCIES_LIST^} switch is send to
8478 @node Switches for gnatmake
8479 @section Switches for @command{gnatmake}
8482 You may specify any of the following switches to @command{gnatmake}:
8487 @item --GCC=@var{compiler_name}
8488 @cindex @option{--GCC=compiler_name} (@command{gnatmake})
8489 Program used for compiling. The default is `@command{gcc}'. You need to use
8490 quotes around @var{compiler_name} if @code{compiler_name} contains
8491 spaces or other separator characters. As an example @option{--GCC="foo -x
8492 -y"} will instruct @command{gnatmake} to use @code{foo -x -y} as your
8493 compiler. A limitation of this syntax is that the name and path name of
8494 the executable itself must not include any embedded spaces. Note that
8495 switch @option{-c} is always inserted after your command name. Thus in the
8496 above example the compiler command that will be used by @command{gnatmake}
8497 will be @code{foo -c -x -y}. If several @option{--GCC=compiler_name} are
8498 used, only the last @var{compiler_name} is taken into account. However,
8499 all the additional switches are also taken into account. Thus,
8500 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
8501 @option{--GCC="bar -x -y -z -t"}.
8503 @item --GNATBIND=@var{binder_name}
8504 @cindex @option{--GNATBIND=binder_name} (@command{gnatmake})
8505 Program used for binding. The default is `@code{gnatbind}'. You need to
8506 use quotes around @var{binder_name} if @var{binder_name} contains spaces
8507 or other separator characters. As an example @option{--GNATBIND="bar -x
8508 -y"} will instruct @command{gnatmake} to use @code{bar -x -y} as your
8509 binder. Binder switches that are normally appended by @command{gnatmake}
8510 to `@code{gnatbind}' are now appended to the end of @code{bar -x -y}.
8511 A limitation of this syntax is that the name and path name of the executable
8512 itself must not include any embedded spaces.
8514 @item --GNATLINK=@var{linker_name}
8515 @cindex @option{--GNATLINK=linker_name} (@command{gnatmake})
8516 Program used for linking. The default is `@command{gnatlink}'. You need to
8517 use quotes around @var{linker_name} if @var{linker_name} contains spaces
8518 or other separator characters. As an example @option{--GNATLINK="lan -x
8519 -y"} will instruct @command{gnatmake} to use @code{lan -x -y} as your
8520 linker. Linker switches that are normally appended by @command{gnatmake} to
8521 `@command{gnatlink}' are now appended to the end of @code{lan -x -y}.
8522 A limitation of this syntax is that the name and path name of the executable
8523 itself must not include any embedded spaces.
8527 @item ^-a^/ALL_FILES^
8528 @cindex @option{^-a^/ALL_FILES^} (@command{gnatmake})
8529 Consider all files in the make process, even the GNAT internal system
8530 files (for example, the predefined Ada library files), as well as any
8531 locked files. Locked files are files whose ALI file is write-protected.
8533 @command{gnatmake} does not check these files,
8534 because the assumption is that the GNAT internal files are properly up
8535 to date, and also that any write protected ALI files have been properly
8536 installed. Note that if there is an installation problem, such that one
8537 of these files is not up to date, it will be properly caught by the
8539 You may have to specify this switch if you are working on GNAT
8540 itself. The switch @option{^-a^/ALL_FILES^} is also useful
8541 in conjunction with @option{^-f^/FORCE_COMPILE^}
8542 if you need to recompile an entire application,
8543 including run-time files, using special configuration pragmas,
8544 such as a @code{Normalize_Scalars} pragma.
8547 @code{gnatmake ^-a^/ALL_FILES^} compiles all GNAT
8550 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
8553 the @code{/CHECKS=SUPPRESS_ALL /STYLE_CHECKS=GNAT} switch.
8556 @item ^-b^/ACTIONS=BIND^
8557 @cindex @option{^-b^/ACTIONS=BIND^} (@command{gnatmake})
8558 Bind only. Can be combined with @option{^-c^/ACTIONS=COMPILE^} to do
8559 compilation and binding, but no link.
8560 Can be combined with @option{^-l^/ACTIONS=LINK^}
8561 to do binding and linking. When not combined with
8562 @option{^-c^/ACTIONS=COMPILE^}
8563 all the units in the closure of the main program must have been previously
8564 compiled and must be up to date. The root unit specified by @var{file_name}
8565 may be given without extension, with the source extension or, if no GNAT
8566 Project File is specified, with the ALI file extension.
8568 @item ^-c^/ACTIONS=COMPILE^
8569 @cindex @option{^-c^/ACTIONS=COMPILE^} (@command{gnatmake})
8570 Compile only. Do not perform binding, except when @option{^-b^/ACTIONS=BIND^}
8571 is also specified. Do not perform linking, except if both
8572 @option{^-b^/ACTIONS=BIND^} and
8573 @option{^-l^/ACTIONS=LINK^} are also specified.
8574 If the root unit specified by @var{file_name} is not a main unit, this is the
8575 default. Otherwise @command{gnatmake} will attempt binding and linking
8576 unless all objects are up to date and the executable is more recent than
8580 @cindex @option{^-C^/MAPPING^} (@command{gnatmake})
8581 Use a temporary mapping file. A mapping file is a way to communicate to the
8582 compiler two mappings: from unit names to file names (without any directory
8583 information) and from file names to path names (with full directory
8584 information). These mappings are used by the compiler to short-circuit the path
8585 search. When @command{gnatmake} is invoked with this switch, it will create
8586 a temporary mapping file, initially populated by the project manager,
8587 if @option{^-P^/PROJECT_FILE^} is used, otherwise initially empty.
8588 Each invocation of the compiler will add the newly accessed sources to the
8589 mapping file. This will improve the source search during the next invocation
8592 @item ^-C=^/USE_MAPPING_FILE=^@var{file}
8593 @cindex @option{^-C=^/USE_MAPPING^} (@command{gnatmake})
8594 Use a specific mapping file. The file, specified as a path name (absolute or
8595 relative) by this switch, should already exist, otherwise the switch is
8596 ineffective. The specified mapping file will be communicated to the compiler.
8597 This switch is not compatible with a project file
8598 (^-P^/PROJECT_FILE=^@var{file}) or with multiple compiling processes
8599 (^-j^/PROCESSES=^nnn, when nnn is greater than 1).
8601 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
8602 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@command{gnatmake})
8603 Put all object files and ALI file in directory @var{dir}.
8604 If the @option{^-D^/DIRECTORY_OBJECTS^} switch is not used, all object files
8605 and ALI files go in the current working directory.
8607 This switch cannot be used when using a project file.
8611 @cindex @option{-eL} (@command{gnatmake})
8612 Follow all symbolic links when processing project files.
8615 @item ^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^
8616 @cindex @option{^-eS^/STANDARD_OUTPUT_FOR_COMMANDS^} (@command{gnatmake})
8617 Output the commands for the compiler, the binder and the linker
8618 on ^standard output^SYS$OUTPUT^,
8619 instead of ^standard error^SYS$ERROR^.
8621 @item ^-f^/FORCE_COMPILE^
8622 @cindex @option{^-f^/FORCE_COMPILE^} (@command{gnatmake})
8623 Force recompilations. Recompile all sources, even though some object
8624 files may be up to date, but don't recompile predefined or GNAT internal
8625 files or locked files (files with a write-protected ALI file),
8626 unless the @option{^-a^/ALL_FILES^} switch is also specified.
8628 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
8629 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@command{gnatmake})
8630 When using project files, if some errors or warnings are detected during
8631 parsing and verbose mode is not in effect (no use of switch
8632 ^-v^/VERBOSE^), then error lines start with the full path name of the project
8633 file, rather than its simple file name.
8636 @cindex @option{^-g^/DEBUG^} (@command{gnatmake})
8637 Enable debugging. This switch is simply passed to the compiler and to the
8640 @item ^-i^/IN_PLACE^
8641 @cindex @option{^-i^/IN_PLACE^} (@command{gnatmake})
8642 In normal mode, @command{gnatmake} compiles all object files and ALI files
8643 into the current directory. If the @option{^-i^/IN_PLACE^} switch is used,
8644 then instead object files and ALI files that already exist are overwritten
8645 in place. This means that once a large project is organized into separate
8646 directories in the desired manner, then @command{gnatmake} will automatically
8647 maintain and update this organization. If no ALI files are found on the
8648 Ada object path (@ref{Search Paths and the Run-Time Library (RTL)}),
8649 the new object and ALI files are created in the
8650 directory containing the source being compiled. If another organization
8651 is desired, where objects and sources are kept in different directories,
8652 a useful technique is to create dummy ALI files in the desired directories.
8653 When detecting such a dummy file, @command{gnatmake} will be forced to
8654 recompile the corresponding source file, and it will be put the resulting
8655 object and ALI files in the directory where it found the dummy file.
8657 @item ^-j^/PROCESSES=^@var{n}
8658 @cindex @option{^-j^/PROCESSES^} (@command{gnatmake})
8659 @cindex Parallel make
8660 Use @var{n} processes to carry out the (re)compilations. On a
8661 multiprocessor machine compilations will occur in parallel. In the
8662 event of compilation errors, messages from various compilations might
8663 get interspersed (but @command{gnatmake} will give you the full ordered
8664 list of failing compiles at the end). If this is problematic, rerun
8665 the make process with n set to 1 to get a clean list of messages.
8667 @item ^-k^/CONTINUE_ON_ERROR^
8668 @cindex @option{^-k^/CONTINUE_ON_ERROR^} (@command{gnatmake})
8669 Keep going. Continue as much as possible after a compilation error. To
8670 ease the programmer's task in case of compilation errors, the list of
8671 sources for which the compile fails is given when @command{gnatmake}
8674 If @command{gnatmake} is invoked with several @file{file_names} and with this
8675 switch, if there are compilation errors when building an executable,
8676 @command{gnatmake} will not attempt to build the following executables.
8678 @item ^-l^/ACTIONS=LINK^
8679 @cindex @option{^-l^/ACTIONS=LINK^} (@command{gnatmake})
8680 Link only. Can be combined with @option{^-b^/ACTIONS=BIND^} to binding
8681 and linking. Linking will not be performed if combined with
8682 @option{^-c^/ACTIONS=COMPILE^}
8683 but not with @option{^-b^/ACTIONS=BIND^}.
8684 When not combined with @option{^-b^/ACTIONS=BIND^}
8685 all the units in the closure of the main program must have been previously
8686 compiled and must be up to date, and the main program needs to have been bound.
8687 The root unit specified by @var{file_name}
8688 may be given without extension, with the source extension or, if no GNAT
8689 Project File is specified, with the ALI file extension.
8691 @item ^-m^/MINIMAL_RECOMPILATION^
8692 @cindex @option{^-m^/MINIMAL_RECOMPILATION^} (@command{gnatmake})
8693 Specify that the minimum necessary amount of recompilations
8694 be performed. In this mode @command{gnatmake} ignores time
8695 stamp differences when the only
8696 modifications to a source file consist in adding/removing comments,
8697 empty lines, spaces or tabs. This means that if you have changed the
8698 comments in a source file or have simply reformatted it, using this
8699 switch will tell gnatmake not to recompile files that depend on it
8700 (provided other sources on which these files depend have undergone no
8701 semantic modifications). Note that the debugging information may be
8702 out of date with respect to the sources if the @option{-m} switch causes
8703 a compilation to be switched, so the use of this switch represents a
8704 trade-off between compilation time and accurate debugging information.
8706 @item ^-M^/DEPENDENCIES_LIST^
8707 @cindex Dependencies, producing list
8708 @cindex @option{^-M^/DEPENDENCIES_LIST^} (@command{gnatmake})
8709 Check if all objects are up to date. If they are, output the object
8710 dependences to @file{stdout} in a form that can be directly exploited in
8711 a @file{Makefile}. By default, each source file is prefixed with its
8712 (relative or absolute) directory name. This name is whatever you
8713 specified in the various @option{^-aI^/SOURCE_SEARCH^}
8714 and @option{^-I^/SEARCH^} switches. If you use
8715 @code{gnatmake ^-M^/DEPENDENCIES_LIST^}
8716 @option{^-q^/QUIET^}
8717 (see below), only the source file names,
8718 without relative paths, are output. If you just specify the
8719 @option{^-M^/DEPENDENCIES_LIST^}
8720 switch, dependencies of the GNAT internal system files are omitted. This
8721 is typically what you want. If you also specify
8722 the @option{^-a^/ALL_FILES^} switch,
8723 dependencies of the GNAT internal files are also listed. Note that
8724 dependencies of the objects in external Ada libraries (see switch
8725 @option{^-aL^/SKIP_MISSING=^}@var{dir} in the following list)
8728 @item ^-n^/DO_OBJECT_CHECK^
8729 @cindex @option{^-n^/DO_OBJECT_CHECK^} (@command{gnatmake})
8730 Don't compile, bind, or link. Checks if all objects are up to date.
8731 If they are not, the full name of the first file that needs to be
8732 recompiled is printed.
8733 Repeated use of this option, followed by compiling the indicated source
8734 file, will eventually result in recompiling all required units.
8736 @item ^-o ^/EXECUTABLE=^@var{exec_name}
8737 @cindex @option{^-o^/EXECUTABLE^} (@command{gnatmake})
8738 Output executable name. The name of the final executable program will be
8739 @var{exec_name}. If the @option{^-o^/EXECUTABLE^} switch is omitted the default
8740 name for the executable will be the name of the input file in appropriate form
8741 for an executable file on the host system.
8743 This switch cannot be used when invoking @command{gnatmake} with several
8746 @item ^-p or --create-missing-dirs^/CREATE_MISSING_DIRS^
8747 @cindex @option{^-p^/CREATE_MISSING_DIRS^} (@command{gnatmake})
8748 When using project files (^-P^/PROJECT_FILE=^@var{project}), create
8749 automatically missing object directories, library directories and exec
8752 @item ^-P^/PROJECT_FILE=^@var{project}
8753 @cindex @option{^-P^/PROJECT_FILE^} (@command{gnatmake})
8754 Use project file @var{project}. Only one such switch can be used.
8755 @xref{gnatmake and Project Files}.
8758 @cindex @option{^-q^/QUIET^} (@command{gnatmake})
8759 Quiet. When this flag is not set, the commands carried out by
8760 @command{gnatmake} are displayed.
8762 @item ^-s^/SWITCH_CHECK/^
8763 @cindex @option{^-s^/SWITCH_CHECK^} (@command{gnatmake})
8764 Recompile if compiler switches have changed since last compilation.
8765 All compiler switches but -I and -o are taken into account in the
8767 orders between different ``first letter'' switches are ignored, but
8768 orders between same switches are taken into account. For example,
8769 @option{-O -O2} is different than @option{-O2 -O}, but @option{-g -O}
8770 is equivalent to @option{-O -g}.
8772 This switch is recommended when Integrated Preprocessing is used.
8775 @cindex @option{^-u^/UNIQUE^} (@command{gnatmake})
8776 Unique. Recompile at most the main files. It implies -c. Combined with
8777 -f, it is equivalent to calling the compiler directly. Note that using
8778 ^-u^/UNIQUE^ with a project file and no main has a special meaning
8779 (@pxref{Project Files and Main Subprograms}).
8781 @item ^-U^/ALL_PROJECTS^
8782 @cindex @option{^-U^/ALL_PROJECTS^} (@command{gnatmake})
8783 When used without a project file or with one or several mains on the command
8784 line, is equivalent to ^-u^/UNIQUE^. When used with a project file and no main
8785 on the command line, all sources of all project files are checked and compiled
8786 if not up to date, and libraries are rebuilt, if necessary.
8789 @cindex @option{^-v^/REASONS^} (@command{gnatmake})
8790 Verbose. Display the reason for all recompilations @command{gnatmake}
8791 decides are necessary, with the highest verbosity level.
8793 @item ^-vl^/LOW_VERBOSITY^
8794 @cindex @option{^-vl^/LOW_VERBOSITY^} (@command{gnatmake})
8795 Verbosity level Low. Display fewer lines than in verbosity Medium.
8797 @item ^-vm^/MEDIUM_VERBOSITY^
8798 @cindex @option{^-vm^/MEDIUM_VERBOSITY^} (@command{gnatmake})
8799 Verbosity level Medium. Potentially display fewer lines than in verbosity High.
8801 @item ^-vh^/HIGH_VERBOSITY^
8802 @cindex @option{^-vm^/HIGH_VERBOSITY^} (@command{gnatmake})
8803 Verbosity level High. Equivalent to ^-v^/REASONS^.
8805 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
8806 Indicate the verbosity of the parsing of GNAT project files.
8807 @xref{Switches Related to Project Files}.
8809 @item ^-x^/NON_PROJECT_UNIT_COMPILATION^
8810 @cindex @option{^-x^/NON_PROJECT_UNIT_COMPILATION^} (@command{gnatmake})
8811 Indicate that sources that are not part of any Project File may be compiled.
8812 Normally, when using Project Files, only sources that are part of a Project
8813 File may be compile. When this switch is used, a source outside of all Project
8814 Files may be compiled. The ALI file and the object file will be put in the
8815 object directory of the main Project. The compilation switches used will only
8816 be those specified on the command line.
8818 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
8819 Indicate that external variable @var{name} has the value @var{value}.
8820 The Project Manager will use this value for occurrences of
8821 @code{external(name)} when parsing the project file.
8822 @xref{Switches Related to Project Files}.
8825 @cindex @option{^-z^/NOMAIN^} (@command{gnatmake})
8826 No main subprogram. Bind and link the program even if the unit name
8827 given on the command line is a package name. The resulting executable
8828 will execute the elaboration routines of the package and its closure,
8829 then the finalization routines.
8834 @item @command{gcc} @asis{switches}
8836 Any uppercase or multi-character switch that is not a @command{gnatmake} switch
8837 is passed to @command{gcc} (e.g. @option{-O}, @option{-gnato,} etc.)
8840 Any qualifier that cannot be recognized as a qualifier for @code{GNAT MAKE}
8841 but is recognizable as a valid qualifier for @code{GNAT COMPILE} is
8842 automatically treated as a compiler switch, and passed on to all
8843 compilations that are carried out.
8848 Source and library search path switches:
8852 @item ^-aI^/SOURCE_SEARCH=^@var{dir}
8853 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatmake})
8854 When looking for source files also look in directory @var{dir}.
8855 The order in which source files search is undertaken is
8856 described in @ref{Search Paths and the Run-Time Library (RTL)}.
8858 @item ^-aL^/SKIP_MISSING=^@var{dir}
8859 @cindex @option{^-aL^/SKIP_MISSING^} (@command{gnatmake})
8860 Consider @var{dir} as being an externally provided Ada library.
8861 Instructs @command{gnatmake} to skip compilation units whose @file{.ALI}
8862 files have been located in directory @var{dir}. This allows you to have
8863 missing bodies for the units in @var{dir} and to ignore out of date bodies
8864 for the same units. You still need to specify
8865 the location of the specs for these units by using the switches
8866 @option{^-aI^/SOURCE_SEARCH=^@var{dir}}
8867 or @option{^-I^/SEARCH=^@var{dir}}.
8868 Note: this switch is provided for compatibility with previous versions
8869 of @command{gnatmake}. The easier method of causing standard libraries
8870 to be excluded from consideration is to write-protect the corresponding
8873 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
8874 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatmake})
8875 When searching for library and object files, look in directory
8876 @var{dir}. The order in which library files are searched is described in
8877 @ref{Search Paths for gnatbind}.
8879 @item ^-A^/CONDITIONAL_SOURCE_SEARCH=^@var{dir}
8880 @cindex Search paths, for @command{gnatmake}
8881 @cindex @option{^-A^/CONDITIONAL_SOURCE_SEARCH^} (@command{gnatmake})
8882 Equivalent to @option{^-aL^/SKIP_MISSING=^@var{dir}
8883 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
8885 @item ^-I^/SEARCH=^@var{dir}
8886 @cindex @option{^-I^/SEARCH^} (@command{gnatmake})
8887 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}
8888 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
8890 @item ^-I-^/NOCURRENT_DIRECTORY^
8891 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatmake})
8892 @cindex Source files, suppressing search
8893 Do not look for source files in the directory containing the source
8894 file named in the command line.
8895 Do not look for ALI or object files in the directory
8896 where @command{gnatmake} was invoked.
8898 @item ^-L^/LIBRARY_SEARCH=^@var{dir}
8899 @cindex @option{^-L^/LIBRARY_SEARCH^} (@command{gnatmake})
8900 @cindex Linker libraries
8901 Add directory @var{dir} to the list of directories in which the linker
8902 will search for libraries. This is equivalent to
8903 @option{-largs ^-L^/LIBRARY_SEARCH=^}@var{dir}.
8905 Furthermore, under Windows, the sources pointed to by the libraries path
8906 set in the registry are not searched for.
8910 @cindex @option{-nostdinc} (@command{gnatmake})
8911 Do not look for source files in the system default directory.
8914 @cindex @option{-nostdlib} (@command{gnatmake})
8915 Do not look for library files in the system default directory.
8917 @item --RTS=@var{rts-path}
8918 @cindex @option{--RTS} (@command{gnatmake})
8919 Specifies the default location of the runtime library. GNAT looks for the
8921 in the following directories, and stops as soon as a valid runtime is found
8922 (@file{adainclude} or @file{ada_source_path}, and @file{adalib} or
8923 @file{ada_object_path} present):
8926 @item <current directory>/$rts_path
8928 @item <default-search-dir>/$rts_path
8930 @item <default-search-dir>/rts-$rts_path
8934 The selected path is handled like a normal RTS path.
8938 @node Mode Switches for gnatmake
8939 @section Mode Switches for @command{gnatmake}
8942 The mode switches (referred to as @code{mode_switches}) allow the
8943 inclusion of switches that are to be passed to the compiler itself, the
8944 binder or the linker. The effect of a mode switch is to cause all
8945 subsequent switches up to the end of the switch list, or up to the next
8946 mode switch, to be interpreted as switches to be passed on to the
8947 designated component of GNAT.
8951 @item -cargs @var{switches}
8952 @cindex @option{-cargs} (@command{gnatmake})
8953 Compiler switches. Here @var{switches} is a list of switches
8954 that are valid switches for @command{gcc}. They will be passed on to
8955 all compile steps performed by @command{gnatmake}.
8957 @item -bargs @var{switches}
8958 @cindex @option{-bargs} (@command{gnatmake})
8959 Binder switches. Here @var{switches} is a list of switches
8960 that are valid switches for @code{gnatbind}. They will be passed on to
8961 all bind steps performed by @command{gnatmake}.
8963 @item -largs @var{switches}
8964 @cindex @option{-largs} (@command{gnatmake})
8965 Linker switches. Here @var{switches} is a list of switches
8966 that are valid switches for @command{gnatlink}. They will be passed on to
8967 all link steps performed by @command{gnatmake}.
8969 @item -margs @var{switches}
8970 @cindex @option{-margs} (@command{gnatmake})
8971 Make switches. The switches are directly interpreted by @command{gnatmake},
8972 regardless of any previous occurrence of @option{-cargs}, @option{-bargs}
8976 @node Notes on the Command Line
8977 @section Notes on the Command Line
8980 This section contains some additional useful notes on the operation
8981 of the @command{gnatmake} command.
8985 @cindex Recompilation, by @command{gnatmake}
8986 If @command{gnatmake} finds no ALI files, it recompiles the main program
8987 and all other units required by the main program.
8988 This means that @command{gnatmake}
8989 can be used for the initial compile, as well as during subsequent steps of
8990 the development cycle.
8993 If you enter @code{gnatmake @var{file}.adb}, where @file{@var{file}.adb}
8994 is a subunit or body of a generic unit, @command{gnatmake} recompiles
8995 @file{@var{file}.adb} (because it finds no ALI) and stops, issuing a
8999 In @command{gnatmake} the switch @option{^-I^/SEARCH^}
9000 is used to specify both source and
9001 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
9002 instead if you just want to specify
9003 source paths only and @option{^-aO^/OBJECT_SEARCH^}
9004 if you want to specify library paths
9008 @command{gnatmake} will ignore any files whose ALI file is write-protected.
9009 This may conveniently be used to exclude standard libraries from
9010 consideration and in particular it means that the use of the
9011 @option{^-f^/FORCE_COMPILE^} switch will not recompile these files
9012 unless @option{^-a^/ALL_FILES^} is also specified.
9015 @command{gnatmake} has been designed to make the use of Ada libraries
9016 particularly convenient. Assume you have an Ada library organized
9017 as follows: @i{^obj-dir^[OBJ_DIR]^} contains the objects and ALI files for
9018 of your Ada compilation units,
9019 whereas @i{^include-dir^[INCLUDE_DIR]^} contains the
9020 specs of these units, but no bodies. Then to compile a unit
9021 stored in @code{main.adb}, which uses this Ada library you would just type
9025 $ gnatmake -aI@var{include-dir} -aL@var{obj-dir} main
9028 $ gnatmake /SOURCE_SEARCH=@i{[INCLUDE_DIR]}
9029 /SKIP_MISSING=@i{[OBJ_DIR]} main
9034 Using @command{gnatmake} along with the
9035 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}
9036 switch provides a mechanism for avoiding unnecessary rcompilations. Using
9038 you can update the comments/format of your
9039 source files without having to recompile everything. Note, however, that
9040 adding or deleting lines in a source files may render its debugging
9041 info obsolete. If the file in question is a spec, the impact is rather
9042 limited, as that debugging info will only be useful during the
9043 elaboration phase of your program. For bodies the impact can be more
9044 significant. In all events, your debugger will warn you if a source file
9045 is more recent than the corresponding object, and alert you to the fact
9046 that the debugging information may be out of date.
9049 @node How gnatmake Works
9050 @section How @command{gnatmake} Works
9053 Generally @command{gnatmake} automatically performs all necessary
9054 recompilations and you don't need to worry about how it works. However,
9055 it may be useful to have some basic understanding of the @command{gnatmake}
9056 approach and in particular to understand how it uses the results of
9057 previous compilations without incorrectly depending on them.
9059 First a definition: an object file is considered @dfn{up to date} if the
9060 corresponding ALI file exists and if all the source files listed in the
9061 dependency section of this ALI file have time stamps matching those in
9062 the ALI file. This means that neither the source file itself nor any
9063 files that it depends on have been modified, and hence there is no need
9064 to recompile this file.
9066 @command{gnatmake} works by first checking if the specified main unit is up
9067 to date. If so, no compilations are required for the main unit. If not,
9068 @command{gnatmake} compiles the main program to build a new ALI file that
9069 reflects the latest sources. Then the ALI file of the main unit is
9070 examined to find all the source files on which the main program depends,
9071 and @command{gnatmake} recursively applies the above procedure on all these
9074 This process ensures that @command{gnatmake} only trusts the dependencies
9075 in an existing ALI file if they are known to be correct. Otherwise it
9076 always recompiles to determine a new, guaranteed accurate set of
9077 dependencies. As a result the program is compiled ``upside down'' from what may
9078 be more familiar as the required order of compilation in some other Ada
9079 systems. In particular, clients are compiled before the units on which
9080 they depend. The ability of GNAT to compile in any order is critical in
9081 allowing an order of compilation to be chosen that guarantees that
9082 @command{gnatmake} will recompute a correct set of new dependencies if
9085 When invoking @command{gnatmake} with several @var{file_names}, if a unit is
9086 imported by several of the executables, it will be recompiled at most once.
9088 Note: when using non-standard naming conventions
9089 (@pxref{Using Other File Names}), changing through a configuration pragmas
9090 file the version of a source and invoking @command{gnatmake} to recompile may
9091 have no effect, if the previous version of the source is still accessible
9092 by @command{gnatmake}. It may be necessary to use the switch
9093 ^-f^/FORCE_COMPILE^.
9095 @node Examples of gnatmake Usage
9096 @section Examples of @command{gnatmake} Usage
9099 @item gnatmake hello.adb
9100 Compile all files necessary to bind and link the main program
9101 @file{hello.adb} (containing unit @code{Hello}) and bind and link the
9102 resulting object files to generate an executable file @file{^hello^HELLO.EXE^}.
9104 @item gnatmake main1 main2 main3
9105 Compile all files necessary to bind and link the main programs
9106 @file{main1.adb} (containing unit @code{Main1}), @file{main2.adb}
9107 (containing unit @code{Main2}) and @file{main3.adb}
9108 (containing unit @code{Main3}) and bind and link the resulting object files
9109 to generate three executable files @file{^main1^MAIN1.EXE^},
9110 @file{^main2^MAIN2.EXE^}
9111 and @file{^main3^MAIN3.EXE^}.
9114 @item gnatmake -q Main_Unit -cargs -O2 -bargs -l
9118 @item gnatmake Main_Unit /QUIET
9119 /COMPILER_QUALIFIERS /OPTIMIZE=ALL
9120 /BINDER_QUALIFIERS /ORDER_OF_ELABORATION
9122 Compile all files necessary to bind and link the main program unit
9123 @code{Main_Unit} (from file @file{main_unit.adb}). All compilations will
9124 be done with optimization level 2 and the order of elaboration will be
9125 listed by the binder. @command{gnatmake} will operate in quiet mode, not
9126 displaying commands it is executing.
9129 @c *************************
9130 @node Improving Performance
9131 @chapter Improving Performance
9132 @cindex Improving performance
9135 This chapter presents several topics related to program performance.
9136 It first describes some of the tradeoffs that need to be considered
9137 and some of the techniques for making your program run faster.
9138 It then documents the @command{gnatelim} tool and unused subprogram/data
9139 elimination feature, which can reduce the size of program executables.
9141 Note: to invoke @command{gnatelim} with a project file, use the @code{gnat}
9142 driver (see @ref{The GNAT Driver and Project Files}).
9146 * Performance Considerations::
9147 * Reducing Size of Ada Executables with gnatelim::
9148 * Reducing Size of Executables with unused subprogram/data elimination::
9152 @c *****************************
9153 @node Performance Considerations
9154 @section Performance Considerations
9157 The GNAT system provides a number of options that allow a trade-off
9162 performance of the generated code
9165 speed of compilation
9168 minimization of dependences and recompilation
9171 the degree of run-time checking.
9175 The defaults (if no options are selected) aim at improving the speed
9176 of compilation and minimizing dependences, at the expense of performance
9177 of the generated code:
9184 no inlining of subprogram calls
9187 all run-time checks enabled except overflow and elaboration checks
9191 These options are suitable for most program development purposes. This
9192 chapter describes how you can modify these choices, and also provides
9193 some guidelines on debugging optimized code.
9196 * Controlling Run-Time Checks::
9197 * Use of Restrictions::
9198 * Optimization Levels::
9199 * Debugging Optimized Code::
9200 * Inlining of Subprograms::
9201 * Other Optimization Switches::
9202 * Optimization and Strict Aliasing::
9205 * Coverage Analysis::
9209 @node Controlling Run-Time Checks
9210 @subsection Controlling Run-Time Checks
9213 By default, GNAT generates all run-time checks, except arithmetic overflow
9214 checking for integer operations and checks for access before elaboration on
9215 subprogram calls. The latter are not required in default mode, because all
9216 necessary checking is done at compile time.
9217 @cindex @option{-gnatp} (@command{gcc})
9218 @cindex @option{-gnato} (@command{gcc})
9219 Two gnat switches, @option{-gnatp} and @option{-gnato} allow this default to
9220 be modified. @xref{Run-Time Checks}.
9222 Our experience is that the default is suitable for most development
9225 We treat integer overflow specially because these
9226 are quite expensive and in our experience are not as important as other
9227 run-time checks in the development process. Note that division by zero
9228 is not considered an overflow check, and divide by zero checks are
9229 generated where required by default.
9231 Elaboration checks are off by default, and also not needed by default, since
9232 GNAT uses a static elaboration analysis approach that avoids the need for
9233 run-time checking. This manual contains a full chapter discussing the issue
9234 of elaboration checks, and if the default is not satisfactory for your use,
9235 you should read this chapter.
9237 For validity checks, the minimal checks required by the Ada Reference
9238 Manual (for case statements and assignments to array elements) are on
9239 by default. These can be suppressed by use of the @option{-gnatVn} switch.
9240 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
9241 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
9242 it may be reasonable to routinely use @option{-gnatVn}. Validity checks
9243 are also suppressed entirely if @option{-gnatp} is used.
9245 @cindex Overflow checks
9246 @cindex Checks, overflow
9249 @cindex pragma Suppress
9250 @cindex pragma Unsuppress
9251 Note that the setting of the switches controls the default setting of
9252 the checks. They may be modified using either @code{pragma Suppress} (to
9253 remove checks) or @code{pragma Unsuppress} (to add back suppressed
9254 checks) in the program source.
9256 @node Use of Restrictions
9257 @subsection Use of Restrictions
9260 The use of pragma Restrictions allows you to control which features are
9261 permitted in your program. Apart from the obvious point that if you avoid
9262 relatively expensive features like finalization (enforceable by the use
9263 of pragma Restrictions (No_Finalization), the use of this pragma does not
9264 affect the generated code in most cases.
9266 One notable exception to this rule is that the possibility of task abort
9267 results in some distributed overhead, particularly if finalization or
9268 exception handlers are used. The reason is that certain sections of code
9269 have to be marked as non-abortable.
9271 If you use neither the @code{abort} statement, nor asynchronous transfer
9272 of control (@code{select .. then abort}), then this distributed overhead
9273 is removed, which may have a general positive effect in improving
9274 overall performance. Especially code involving frequent use of tasking
9275 constructs and controlled types will show much improved performance.
9276 The relevant restrictions pragmas are
9278 @smallexample @c ada
9279 pragma Restrictions (No_Abort_Statements);
9280 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
9284 It is recommended that these restriction pragmas be used if possible. Note
9285 that this also means that you can write code without worrying about the
9286 possibility of an immediate abort at any point.
9288 @node Optimization Levels
9289 @subsection Optimization Levels
9290 @cindex @option{^-O^/OPTIMIZE^} (@command{gcc})
9293 The default is optimization off. This results in the fastest compile
9294 times, but GNAT makes absolutely no attempt to optimize, and the
9295 generated programs are considerably larger and slower than when
9296 optimization is enabled. You can use the
9298 @option{-O} switch (the permitted forms are @option{-O0}, @option{-O1}
9299 @option{-O2}, @option{-O3}, and @option{-Os})
9302 @code{OPTIMIZE} qualifier
9304 to @command{gcc} to control the optimization level:
9307 @item ^-O0^/OPTIMIZE=NONE^
9308 No optimization (the default);
9309 generates unoptimized code but has
9310 the fastest compilation time.
9312 Note that many other compilers do fairly extensive optimization
9313 even if "no optimization" is specified. When using gcc, it is
9314 very unusual to use ^-O0^/OPTIMIZE=NONE^ for production if
9315 execution time is of any concern, since ^-O0^/OPTIMIZE=NONE^
9316 really does mean no optimization at all. This difference between
9317 gcc and other compilers should be kept in mind when doing
9318 performance comparisons.
9320 @item ^-O1^/OPTIMIZE=SOME^
9321 Moderate optimization;
9322 optimizes reasonably well but does not
9323 degrade compilation time significantly.
9325 @item ^-O2^/OPTIMIZE=ALL^
9327 @itemx /OPTIMIZE=DEVELOPMENT
9330 generates highly optimized code and has
9331 the slowest compilation time.
9333 @item ^-O3^/OPTIMIZE=INLINING^
9334 Full optimization as in @option{-O2},
9335 and also attempts automatic inlining of small
9336 subprograms within a unit (@pxref{Inlining of Subprograms}).
9338 @item ^-Os^/OPTIMIZE=SPACE^
9339 Optimize space usage of resulting program.
9343 Higher optimization levels perform more global transformations on the
9344 program and apply more expensive analysis algorithms in order to generate
9345 faster and more compact code. The price in compilation time, and the
9346 resulting improvement in execution time,
9347 both depend on the particular application and the hardware environment.
9348 You should experiment to find the best level for your application.
9350 The @option{^-Os^/OPTIMIZE=SPACE^} switch is independent of the time
9351 optimizations, so you can specify both @option{^-Os^/OPTIMIZE=SPACE^}
9352 and a time optimization on the same compile command.
9354 Since the precise set of optimizations done at each level will vary from
9355 release to release (and sometime from target to target), it is best to think
9356 of the optimization settings in general terms.
9357 The @cite{Using GNU GCC} manual contains details about
9358 ^the @option{-O} settings and a number of @option{-f} options that^how to^
9359 individually enable or disable specific optimizations.
9361 Unlike some other compilation systems, ^@command{gcc}^GNAT^ has
9362 been tested extensively at all optimization levels. There are some bugs
9363 which appear only with optimization turned on, but there have also been
9364 bugs which show up only in @emph{unoptimized} code. Selecting a lower
9365 level of optimization does not improve the reliability of the code
9366 generator, which in practice is highly reliable at all optimization
9369 Note regarding the use of @option{-O3}: The use of this optimization level
9370 is generally discouraged with GNAT, since it often results in larger
9371 executables which run more slowly. See further discussion of this point
9372 in @ref{Inlining of Subprograms}.
9374 @node Debugging Optimized Code
9375 @subsection Debugging Optimized Code
9376 @cindex Debugging optimized code
9377 @cindex Optimization and debugging
9380 Although it is possible to do a reasonable amount of debugging at
9382 nonzero optimization levels,
9383 the higher the level the more likely that
9386 @option{/OPTIMIZE} settings other than @code{NONE},
9387 such settings will make it more likely that
9389 source-level constructs will have been eliminated by optimization.
9390 For example, if a loop is strength-reduced, the loop
9391 control variable may be completely eliminated and thus cannot be
9392 displayed in the debugger.
9393 This can only happen at @option{-O2} or @option{-O3}.
9394 Explicit temporary variables that you code might be eliminated at
9395 ^level^setting^ @option{-O1} or higher.
9397 The use of the @option{^-g^/DEBUG^} switch,
9398 @cindex @option{^-g^/DEBUG^} (@command{gcc})
9399 which is needed for source-level debugging,
9400 affects the size of the program executable on disk,
9401 and indeed the debugging information can be quite large.
9402 However, it has no effect on the generated code (and thus does not
9403 degrade performance)
9405 Since the compiler generates debugging tables for a compilation unit before
9406 it performs optimizations, the optimizing transformations may invalidate some
9407 of the debugging data. You therefore need to anticipate certain
9408 anomalous situations that may arise while debugging optimized code.
9409 These are the most common cases:
9413 @i{The ``hopping Program Counter'':} Repeated @code{step} or @code{next}
9415 the PC bouncing back and forth in the code. This may result from any of
9416 the following optimizations:
9420 @i{Common subexpression elimination:} using a single instance of code for a
9421 quantity that the source computes several times. As a result you
9422 may not be able to stop on what looks like a statement.
9425 @i{Invariant code motion:} moving an expression that does not change within a
9426 loop, to the beginning of the loop.
9429 @i{Instruction scheduling:} moving instructions so as to
9430 overlap loads and stores (typically) with other code, or in
9431 general to move computations of values closer to their uses. Often
9432 this causes you to pass an assignment statement without the assignment
9433 happening and then later bounce back to the statement when the
9434 value is actually needed. Placing a breakpoint on a line of code
9435 and then stepping over it may, therefore, not always cause all the
9436 expected side-effects.
9440 @i{The ``big leap'':} More commonly known as @emph{cross-jumping}, in which
9441 two identical pieces of code are merged and the program counter suddenly
9442 jumps to a statement that is not supposed to be executed, simply because
9443 it (and the code following) translates to the same thing as the code
9444 that @emph{was} supposed to be executed. This effect is typically seen in
9445 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
9446 a @code{break} in a C @code{^switch^switch^} statement.
9449 @i{The ``roving variable'':} The symptom is an unexpected value in a variable.
9450 There are various reasons for this effect:
9454 In a subprogram prologue, a parameter may not yet have been moved to its
9458 A variable may be dead, and its register re-used. This is
9459 probably the most common cause.
9462 As mentioned above, the assignment of a value to a variable may
9466 A variable may be eliminated entirely by value propagation or
9467 other means. In this case, GCC may incorrectly generate debugging
9468 information for the variable
9472 In general, when an unexpected value appears for a local variable or parameter
9473 you should first ascertain if that value was actually computed by
9474 your program, as opposed to being incorrectly reported by the debugger.
9476 array elements in an object designated by an access value
9477 are generally less of a problem, once you have ascertained that the access
9479 Typically, this means checking variables in the preceding code and in the
9480 calling subprogram to verify that the value observed is explainable from other
9481 values (one must apply the procedure recursively to those
9482 other values); or re-running the code and stopping a little earlier
9483 (perhaps before the call) and stepping to better see how the variable obtained
9484 the value in question; or continuing to step @emph{from} the point of the
9485 strange value to see if code motion had simply moved the variable's
9490 In light of such anomalies, a recommended technique is to use @option{-O0}
9491 early in the software development cycle, when extensive debugging capabilities
9492 are most needed, and then move to @option{-O1} and later @option{-O2} as
9493 the debugger becomes less critical.
9494 Whether to use the @option{^-g^/DEBUG^} switch in the release version is
9495 a release management issue.
9497 Note that if you use @option{-g} you can then use the @command{strip} program
9498 on the resulting executable,
9499 which removes both debugging information and global symbols.
9502 @node Inlining of Subprograms
9503 @subsection Inlining of Subprograms
9506 A call to a subprogram in the current unit is inlined if all the
9507 following conditions are met:
9511 The optimization level is at least @option{-O1}.
9514 The called subprogram is suitable for inlining: It must be small enough
9515 and not contain nested subprograms or anything else that @command{gcc}
9516 cannot support in inlined subprograms.
9519 The call occurs after the definition of the body of the subprogram.
9522 @cindex pragma Inline
9524 Either @code{pragma Inline} applies to the subprogram or it is
9525 small and automatic inlining (optimization level @option{-O3}) is
9530 Calls to subprograms in @code{with}'ed units are normally not inlined.
9531 To achieve actual inlining (that is, replacement of the call by the code
9532 in the body of the subprogram), the following conditions must all be true.
9536 The optimization level is at least @option{-O1}.
9539 The called subprogram is suitable for inlining: It must be small enough
9540 and not contain nested subprograms or anything else @command{gcc} cannot
9541 support in inlined subprograms.
9544 The call appears in a body (not in a package spec).
9547 There is a @code{pragma Inline} for the subprogram.
9550 @cindex @option{-gnatn} (@command{gcc})
9551 The @option{^-gnatn^/INLINE^} switch
9552 is used in the @command{gcc} command line
9555 Even if all these conditions are met, it may not be possible for
9556 the compiler to inline the call, due to the length of the body,
9557 or features in the body that make it impossible for the compiler
9560 Note that specifying the @option{-gnatn} switch causes additional
9561 compilation dependencies. Consider the following:
9563 @smallexample @c ada
9583 With the default behavior (no @option{-gnatn} switch specified), the
9584 compilation of the @code{Main} procedure depends only on its own source,
9585 @file{main.adb}, and the spec of the package in file @file{r.ads}. This
9586 means that editing the body of @code{R} does not require recompiling
9589 On the other hand, the call @code{R.Q} is not inlined under these
9590 circumstances. If the @option{-gnatn} switch is present when @code{Main}
9591 is compiled, the call will be inlined if the body of @code{Q} is small
9592 enough, but now @code{Main} depends on the body of @code{R} in
9593 @file{r.adb} as well as on the spec. This means that if this body is edited,
9594 the main program must be recompiled. Note that this extra dependency
9595 occurs whether or not the call is in fact inlined by @command{gcc}.
9597 The use of front end inlining with @option{-gnatN} generates similar
9598 additional dependencies.
9600 @cindex @option{^-fno-inline^/INLINE=SUPPRESS^} (@command{gcc})
9601 Note: The @option{^-fno-inline^/INLINE=SUPPRESS^} switch
9602 can be used to prevent
9603 all inlining. This switch overrides all other conditions and ensures
9604 that no inlining occurs. The extra dependences resulting from
9605 @option{-gnatn} will still be active, even if
9606 this switch is used to suppress the resulting inlining actions.
9608 Note regarding the use of @option{-O3}: There is no difference in inlining
9609 behavior between @option{-O2} and @option{-O3} for subprograms with an explicit
9610 pragma @code{Inline} assuming the use of @option{-gnatn}
9611 or @option{-gnatN} (the switches that activate inlining). If you have used
9612 pragma @code{Inline} in appropriate cases, then it is usually much better
9613 to use @option{-O2} and @option{-gnatn} and avoid the use of @option{-O3} which
9614 in this case only has the effect of inlining subprograms you did not
9615 think should be inlined. We often find that the use of @option{-O3} slows
9616 down code by performing excessive inlining, leading to increased instruction
9617 cache pressure from the increased code size. So the bottom line here is
9618 that you should not automatically assume that @option{-O3} is better than
9619 @option{-O2}, and indeed you should use @option{-O3} only if tests show that
9620 it actually improves performance.
9622 @node Other Optimization Switches
9623 @subsection Other Optimization Switches
9624 @cindex Optimization Switches
9626 Since @code{GNAT} uses the @code{gcc} back end, all the specialized
9627 @code{gcc} optimization switches are potentially usable. These switches
9628 have not been extensively tested with GNAT but can generally be expected
9629 to work. Examples of switches in this category are
9630 @option{-funroll-loops} and
9631 the various target-specific @option{-m} options (in particular, it has been
9632 observed that @option{-march=pentium4} can significantly improve performance
9633 on appropriate machines). For full details of these switches, see the
9636 @node Optimization and Strict Aliasing
9637 @subsection Optimization and Strict Aliasing
9639 @cindex Strict Aliasing
9640 @cindex No_Strict_Aliasing
9643 The strong typing capabilities of Ada allow an optimizer to generate
9644 efficient code in situations where other languages would be forced to
9645 make worst case assumptions preventing such optimizations. Consider
9646 the following example:
9648 @smallexample @c ada
9651 type Int1 is new Integer;
9652 type Int2 is new Integer;
9653 type Int1A is access Int1;
9654 type Int2A is access Int2;
9661 for J in Data'Range loop
9662 if Data (J) = Int1V.all then
9663 Int2V.all := Int2V.all + 1;
9672 In this example, since the variable @code{Int1V} can only access objects
9673 of type @code{Int1}, and @code{Int2V} can only access objects of type
9674 @code{Int2}, there is no possibility that the assignment to
9675 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
9676 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
9677 for all iterations of the loop and avoid the extra memory reference
9678 required to dereference it each time through the loop.
9680 This kind of optimization, called strict aliasing analysis, is
9681 triggered by specifying an optimization level of @option{-O2} or
9682 higher and allows @code{GNAT} to generate more efficient code
9683 when access values are involved.
9685 However, although this optimization is always correct in terms of
9686 the formal semantics of the Ada Reference Manual, difficulties can
9687 arise if features like @code{Unchecked_Conversion} are used to break
9688 the typing system. Consider the following complete program example:
9690 @smallexample @c ada
9693 type int1 is new integer;
9694 type int2 is new integer;
9695 type a1 is access int1;
9696 type a2 is access int2;
9701 function to_a2 (Input : a1) return a2;
9704 with Unchecked_Conversion;
9706 function to_a2 (Input : a1) return a2 is
9708 new Unchecked_Conversion (a1, a2);
9710 return to_a2u (Input);
9716 with Text_IO; use Text_IO;
9718 v1 : a1 := new int1;
9719 v2 : a2 := to_a2 (v1);
9723 put_line (int1'image (v1.all));
9729 This program prints out 0 in @code{-O0} or @code{-O1}
9730 mode, but it prints out 1 in @code{-O2} mode. That's
9731 because in strict aliasing mode, the compiler can and
9732 does assume that the assignment to @code{v2.all} could not
9733 affect the value of @code{v1.all}, since different types
9736 This behavior is not a case of non-conformance with the standard, since
9737 the Ada RM specifies that an unchecked conversion where the resulting
9738 bit pattern is not a correct value of the target type can result in an
9739 abnormal value and attempting to reference an abnormal value makes the
9740 execution of a program erroneous. That's the case here since the result
9741 does not point to an object of type @code{int2}. This means that the
9742 effect is entirely unpredictable.
9744 However, although that explanation may satisfy a language
9745 lawyer, in practice an applications programmer expects an
9746 unchecked conversion involving pointers to create true
9747 aliases and the behavior of printing 1 seems plain wrong.
9748 In this case, the strict aliasing optimization is unwelcome.
9750 Indeed the compiler recognizes this possibility, and the
9751 unchecked conversion generates a warning:
9754 p2.adb:5:07: warning: possible aliasing problem with type "a2"
9755 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
9756 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
9760 Unfortunately the problem is recognized when compiling the body of
9761 package @code{p2}, but the actual "bad" code is generated while
9762 compiling the body of @code{m} and this latter compilation does not see
9763 the suspicious @code{Unchecked_Conversion}.
9765 As implied by the warning message, there are approaches you can use to
9766 avoid the unwanted strict aliasing optimization in a case like this.
9768 One possibility is to simply avoid the use of @code{-O2}, but
9769 that is a bit drastic, since it throws away a number of useful
9770 optimizations that do not involve strict aliasing assumptions.
9772 A less drastic approach is to compile the program using the
9773 option @code{-fno-strict-aliasing}. Actually it is only the
9774 unit containing the dereferencing of the suspicious pointer
9775 that needs to be compiled. So in this case, if we compile
9776 unit @code{m} with this switch, then we get the expected
9777 value of zero printed. Analyzing which units might need
9778 the switch can be painful, so a more reasonable approach
9779 is to compile the entire program with options @code{-O2}
9780 and @code{-fno-strict-aliasing}. If the performance is
9781 satisfactory with this combination of options, then the
9782 advantage is that the entire issue of possible "wrong"
9783 optimization due to strict aliasing is avoided.
9785 To avoid the use of compiler switches, the configuration
9786 pragma @code{No_Strict_Aliasing} with no parameters may be
9787 used to specify that for all access types, the strict
9788 aliasing optimization should be suppressed.
9790 However, these approaches are still overkill, in that they causes
9791 all manipulations of all access values to be deoptimized. A more
9792 refined approach is to concentrate attention on the specific
9793 access type identified as problematic.
9795 First, if a careful analysis of uses of the pointer shows
9796 that there are no possible problematic references, then
9797 the warning can be suppressed by bracketing the
9798 instantiation of @code{Unchecked_Conversion} to turn
9801 @smallexample @c ada
9802 pragma Warnings (Off);
9804 new Unchecked_Conversion (a1, a2);
9805 pragma Warnings (On);
9809 Of course that approach is not appropriate for this particular
9810 example, since indeed there is a problematic reference. In this
9811 case we can take one of two other approaches.
9813 The first possibility is to move the instantiation of unchecked
9814 conversion to the unit in which the type is declared. In
9815 this example, we would move the instantiation of
9816 @code{Unchecked_Conversion} from the body of package
9817 @code{p2} to the spec of package @code{p1}. Now the
9818 warning disappears. That's because any use of the
9819 access type knows there is a suspicious unchecked
9820 conversion, and the strict aliasing optimization
9821 is automatically suppressed for the type.
9823 If it is not practical to move the unchecked conversion to the same unit
9824 in which the destination access type is declared (perhaps because the
9825 source type is not visible in that unit), you may use pragma
9826 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
9827 same declarative sequence as the declaration of the access type:
9829 @smallexample @c ada
9830 type a2 is access int2;
9831 pragma No_Strict_Aliasing (a2);
9835 Here again, the compiler now knows that the strict aliasing optimization
9836 should be suppressed for any reference to type @code{a2} and the
9837 expected behavior is obtained.
9839 Finally, note that although the compiler can generate warnings for
9840 simple cases of unchecked conversions, there are tricker and more
9841 indirect ways of creating type incorrect aliases which the compiler
9842 cannot detect. Examples are the use of address overlays and unchecked
9843 conversions involving composite types containing access types as
9844 components. In such cases, no warnings are generated, but there can
9845 still be aliasing problems. One safe coding practice is to forbid the
9846 use of address clauses for type overlaying, and to allow unchecked
9847 conversion only for primitive types. This is not really a significant
9848 restriction since any possible desired effect can be achieved by
9849 unchecked conversion of access values.
9852 @node Coverage Analysis
9853 @subsection Coverage Analysis
9856 GNAT supports the HP Performance Coverage Analyzer (PCA), which allows
9857 the user to determine the distribution of execution time across a program,
9858 @pxref{Profiling} for details of usage.
9861 @node Reducing Size of Ada Executables with gnatelim
9862 @section Reducing Size of Ada Executables with @code{gnatelim}
9866 This section describes @command{gnatelim}, a tool which detects unused
9867 subprograms and helps the compiler to create a smaller executable for your
9872 * Running gnatelim::
9873 * Correcting the List of Eliminate Pragmas::
9874 * Making Your Executables Smaller::
9875 * Summary of the gnatelim Usage Cycle::
9878 @node About gnatelim
9879 @subsection About @code{gnatelim}
9882 When a program shares a set of Ada
9883 packages with other programs, it may happen that this program uses
9884 only a fraction of the subprograms defined in these packages. The code
9885 created for these unused subprograms increases the size of the executable.
9887 @code{gnatelim} tracks unused subprograms in an Ada program and
9888 outputs a list of GNAT-specific pragmas @code{Eliminate} marking all the
9889 subprograms that are declared but never called. By placing the list of
9890 @code{Eliminate} pragmas in the GNAT configuration file @file{gnat.adc} and
9891 recompiling your program, you may decrease the size of its executable,
9892 because the compiler will not generate the code for 'eliminated' subprograms.
9893 See GNAT Reference Manual for more information about this pragma.
9895 @code{gnatelim} needs as its input data the name of the main subprogram
9896 and a bind file for a main subprogram.
9898 To create a bind file for @code{gnatelim}, run @code{gnatbind} for
9899 the main subprogram. @code{gnatelim} can work with both Ada and C
9900 bind files; when both are present, it uses the Ada bind file.
9901 The following commands will build the program and create the bind file:
9904 $ gnatmake ^-c Main_Prog^/ACTIONS=COMPILE MAIN_PROG^
9905 $ gnatbind main_prog
9908 Note that @code{gnatelim} needs neither object nor ALI files.
9910 @node Running gnatelim
9911 @subsection Running @code{gnatelim}
9914 @code{gnatelim} has the following command-line interface:
9917 $ gnatelim [options] name
9921 @code{name} should be a name of a source file that contains the main subprogram
9922 of a program (partition).
9924 @code{gnatelim} has the following switches:
9929 @cindex @option{^-q^/QUIET^} (@command{gnatelim})
9930 Quiet mode: by default @code{gnatelim} outputs to the standard error
9931 stream the number of program units left to be processed. This option turns
9935 @cindex @option{^-v^/VERBOSE^} (@command{gnatelim})
9936 Verbose mode: @code{gnatelim} version information is printed as Ada
9937 comments to the standard output stream. Also, in addition to the number of
9938 program units left @code{gnatelim} will output the name of the current unit
9942 @cindex @option{^-a^/ALL^} (@command{gnatelim})
9943 Also look for subprograms from the GNAT run time that can be eliminated. Note
9944 that when @file{gnat.adc} is produced using this switch, the entire program
9945 must be recompiled with switch @option{^-a^/ALL_FILES^} to @command{gnatmake}.
9947 @item ^-I^/INCLUDE_DIRS=^@var{dir}
9948 @cindex @option{^-I^/INCLUDE_DIRS^} (@command{gnatelim})
9949 When looking for source files also look in directory @var{dir}. Specifying
9950 @option{^-I-^/INCLUDE_DIRS=-^} instructs @code{gnatelim} not to look for
9951 sources in the current directory.
9953 @item ^-b^/BIND_FILE=^@var{bind_file}
9954 @cindex @option{^-b^/BIND_FILE^} (@command{gnatelim})
9955 Specifies @var{bind_file} as the bind file to process. If not set, the name
9956 of the bind file is computed from the full expanded Ada name
9957 of a main subprogram.
9959 @item ^-C^/CONFIG_FILE=^@var{config_file}
9960 @cindex @option{^-C^/CONFIG_FILE^} (@command{gnatelim})
9961 Specifies a file @var{config_file} that contains configuration pragmas. The
9962 file must be specified with full path.
9964 @item ^--GCC^/COMPILER^=@var{compiler_name}
9965 @cindex @option{^-GCC^/COMPILER^} (@command{gnatelim})
9966 Instructs @code{gnatelim} to use specific @command{gcc} compiler instead of one
9967 available on the path.
9969 @item ^--GNATMAKE^/GNATMAKE^=@var{gnatmake_name}
9970 @cindex @option{^--GNATMAKE^/GNATMAKE^} (@command{gnatelim})
9971 Instructs @code{gnatelim} to use specific @command{gnatmake} instead of one
9972 available on the path.
9976 @code{gnatelim} sends its output to the standard output stream, and all the
9977 tracing and debug information is sent to the standard error stream.
9978 In order to produce a proper GNAT configuration file
9979 @file{gnat.adc}, redirection must be used:
9983 $ PIPE GNAT ELIM MAIN_PROG.ADB > GNAT.ADC
9986 $ gnatelim main_prog.adb > gnat.adc
9995 $ gnatelim main_prog.adb >> gnat.adc
9999 in order to append the @code{gnatelim} output to the existing contents of
10003 @node Correcting the List of Eliminate Pragmas
10004 @subsection Correcting the List of Eliminate Pragmas
10007 In some rare cases @code{gnatelim} may try to eliminate
10008 subprograms that are actually called in the program. In this case, the
10009 compiler will generate an error message of the form:
10012 file.adb:106:07: cannot call eliminated subprogram "My_Prog"
10016 You will need to manually remove the wrong @code{Eliminate} pragmas from
10017 the @file{gnat.adc} file. You should recompile your program
10018 from scratch after that, because you need a consistent @file{gnat.adc} file
10019 during the entire compilation.
10021 @node Making Your Executables Smaller
10022 @subsection Making Your Executables Smaller
10025 In order to get a smaller executable for your program you now have to
10026 recompile the program completely with the new @file{gnat.adc} file
10027 created by @code{gnatelim} in your current directory:
10030 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10034 (Use the @option{^-f^/FORCE_COMPILE^} option for @command{gnatmake} to
10035 recompile everything
10036 with the set of pragmas @code{Eliminate} that you have obtained with
10037 @command{gnatelim}).
10039 Be aware that the set of @code{Eliminate} pragmas is specific to each
10040 program. It is not recommended to merge sets of @code{Eliminate}
10041 pragmas created for different programs in one @file{gnat.adc} file.
10043 @node Summary of the gnatelim Usage Cycle
10044 @subsection Summary of the gnatelim Usage Cycle
10047 Here is a quick summary of the steps to be taken in order to reduce
10048 the size of your executables with @code{gnatelim}. You may use
10049 other GNAT options to control the optimization level,
10050 to produce the debugging information, to set search path, etc.
10054 Produce a bind file
10057 $ gnatmake ^-c main_prog^/ACTIONS=COMPILE MAIN_PROG^
10058 $ gnatbind main_prog
10062 Generate a list of @code{Eliminate} pragmas
10065 $ PIPE GNAT ELIM MAIN_PROG > GNAT.ADC
10068 $ gnatelim main_prog >[>] gnat.adc
10073 Recompile the application
10076 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
10081 @node Reducing Size of Executables with unused subprogram/data elimination
10082 @section Reducing Size of Executables with Unused Subprogram/Data Elimination
10083 @findex unused subprogram/data elimination
10086 This section describes how you can eliminate unused subprograms and data from
10087 your executable just by setting options at compilation time.
10090 * About unused subprogram/data elimination::
10091 * Compilation options::
10092 * Example of unused subprogram/data elimination::
10095 @node About unused subprogram/data elimination
10096 @subsection About unused subprogram/data elimination
10099 By default, an executable contains all code and data of its composing objects
10100 (directly linked or coming from statically linked libraries), even data or code
10101 never used by this executable.
10103 This feature will allow you to eliminate such unused code from your
10104 executable, making it smaller (in disk and in memory).
10106 This functionality is available on all platforms using elf binary format and
10107 having GNU binutils version 2.16.1.
10109 @node Compilation options
10110 @subsection Compilation options
10113 The operation of eliminating the unused code and data from the final executable
10114 is directly performed by the linker.
10116 In order to do this, it has to work with objects compiled with the
10118 @option{-ffunction-sections} @option{-fdata-sections}.
10119 @cindex @option{-ffunction-sections} (@command{gcc})
10120 @cindex @option{-fdata-sections} (@command{gcc})
10121 These options are usable with C and Ada files.
10122 They will place respectively each
10123 function or data in a separate section in the resulting object file.
10125 Once the objects and static libraries are created with these options, the
10126 linker can perform the dead code elimination. You can do this by setting
10127 the @option{-Wl,--gc-sections} option to gcc command or in the
10128 @option{-largs} section of gnatmake. This will perform a garbage collection of
10129 code and data never referenced.
10131 If the linker performs a partial link (@option{-r} ld linker option), then you
10132 will need to provide one or several entry point using the
10133 @option{-e} / @option{--entry} ld option.
10135 Note that objects compiled without the @option{-ffunction-sections} and
10136 @option{-fdata-sections} options can still be linked with the executable.
10137 However, no dead code elimination will be performed on those objects (they will
10140 The GNAT static library is now compiled with -ffunction-sections and
10141 -fdata-sections on some platforms. This allows you to eliminate the unused code
10142 and data of the GNAT library from your executable.
10144 @node Example of unused subprogram/data elimination
10145 @subsection Example of unused subprogram/data elimination
10148 Here is a simple example:
10150 @smallexample @c ada
10159 Used_Data : Integer;
10160 Unused_Data : Integer;
10162 procedure Used (Data : Integer);
10163 procedure Unused (Data : Integer);
10166 package body Aux is
10167 procedure Used (Data : Integer) is
10172 procedure Unused (Data : Integer) is
10174 Unused_Data := Data;
10180 @code{Unused} and @code{Unused_Data} are never referenced in this code
10181 excerpt, and hence they may be safely removed from the final executable.
10186 $ nm test | grep used
10187 020015f0 T aux__unused
10188 02005d88 B aux__unused_data
10189 020015cc T aux__used
10190 02005d84 B aux__used_data
10192 $ gnatmake test -cargs -fdata-sections -ffunction-sections \
10193 -largs -Wl,--gc-sections
10195 $ nm test | grep used
10196 02005350 T aux__used
10197 0201ffe0 B aux__used_data
10201 It can be observed that the procedure @code{Unused} and the object
10202 @code{Unused_Data} are removed by the linker when using the
10203 appropriate options.
10205 @c ********************************
10206 @node Renaming Files Using gnatchop
10207 @chapter Renaming Files Using @code{gnatchop}
10211 This chapter discusses how to handle files with multiple units by using
10212 the @code{gnatchop} utility. This utility is also useful in renaming
10213 files to meet the standard GNAT default file naming conventions.
10216 * Handling Files with Multiple Units::
10217 * Operating gnatchop in Compilation Mode::
10218 * Command Line for gnatchop::
10219 * Switches for gnatchop::
10220 * Examples of gnatchop Usage::
10223 @node Handling Files with Multiple Units
10224 @section Handling Files with Multiple Units
10227 The basic compilation model of GNAT requires that a file submitted to the
10228 compiler have only one unit and there be a strict correspondence
10229 between the file name and the unit name.
10231 The @code{gnatchop} utility allows both of these rules to be relaxed,
10232 allowing GNAT to process files which contain multiple compilation units
10233 and files with arbitrary file names. @code{gnatchop}
10234 reads the specified file and generates one or more output files,
10235 containing one unit per file. The unit and the file name correspond,
10236 as required by GNAT.
10238 If you want to permanently restructure a set of ``foreign'' files so that
10239 they match the GNAT rules, and do the remaining development using the
10240 GNAT structure, you can simply use @command{gnatchop} once, generate the
10241 new set of files and work with them from that point on.
10243 Alternatively, if you want to keep your files in the ``foreign'' format,
10244 perhaps to maintain compatibility with some other Ada compilation
10245 system, you can set up a procedure where you use @command{gnatchop} each
10246 time you compile, regarding the source files that it writes as temporary
10247 files that you throw away.
10249 @node Operating gnatchop in Compilation Mode
10250 @section Operating gnatchop in Compilation Mode
10253 The basic function of @code{gnatchop} is to take a file with multiple units
10254 and split it into separate files. The boundary between files is reasonably
10255 clear, except for the issue of comments and pragmas. In default mode, the
10256 rule is that any pragmas between units belong to the previous unit, except
10257 that configuration pragmas always belong to the following unit. Any comments
10258 belong to the following unit. These rules
10259 almost always result in the right choice of
10260 the split point without needing to mark it explicitly and most users will
10261 find this default to be what they want. In this default mode it is incorrect to
10262 submit a file containing only configuration pragmas, or one that ends in
10263 configuration pragmas, to @code{gnatchop}.
10265 However, using a special option to activate ``compilation mode'',
10267 can perform another function, which is to provide exactly the semantics
10268 required by the RM for handling of configuration pragmas in a compilation.
10269 In the absence of configuration pragmas (at the main file level), this
10270 option has no effect, but it causes such configuration pragmas to be handled
10271 in a quite different manner.
10273 First, in compilation mode, if @code{gnatchop} is given a file that consists of
10274 only configuration pragmas, then this file is appended to the
10275 @file{gnat.adc} file in the current directory. This behavior provides
10276 the required behavior described in the RM for the actions to be taken
10277 on submitting such a file to the compiler, namely that these pragmas
10278 should apply to all subsequent compilations in the same compilation
10279 environment. Using GNAT, the current directory, possibly containing a
10280 @file{gnat.adc} file is the representation
10281 of a compilation environment. For more information on the
10282 @file{gnat.adc} file, see @ref{Handling of Configuration Pragmas}.
10284 Second, in compilation mode, if @code{gnatchop}
10285 is given a file that starts with
10286 configuration pragmas, and contains one or more units, then these
10287 configuration pragmas are prepended to each of the chopped files. This
10288 behavior provides the required behavior described in the RM for the
10289 actions to be taken on compiling such a file, namely that the pragmas
10290 apply to all units in the compilation, but not to subsequently compiled
10293 Finally, if configuration pragmas appear between units, they are appended
10294 to the previous unit. This results in the previous unit being illegal,
10295 since the compiler does not accept configuration pragmas that follow
10296 a unit. This provides the required RM behavior that forbids configuration
10297 pragmas other than those preceding the first compilation unit of a
10300 For most purposes, @code{gnatchop} will be used in default mode. The
10301 compilation mode described above is used only if you need exactly
10302 accurate behavior with respect to compilations, and you have files
10303 that contain multiple units and configuration pragmas. In this
10304 circumstance the use of @code{gnatchop} with the compilation mode
10305 switch provides the required behavior, and is for example the mode
10306 in which GNAT processes the ACVC tests.
10308 @node Command Line for gnatchop
10309 @section Command Line for @code{gnatchop}
10312 The @code{gnatchop} command has the form:
10315 $ gnatchop switches @var{file name} [@var{file name} @var{file name} ...]
10320 The only required argument is the file name of the file to be chopped.
10321 There are no restrictions on the form of this file name. The file itself
10322 contains one or more Ada units, in normal GNAT format, concatenated
10323 together. As shown, more than one file may be presented to be chopped.
10325 When run in default mode, @code{gnatchop} generates one output file in
10326 the current directory for each unit in each of the files.
10328 @var{directory}, if specified, gives the name of the directory to which
10329 the output files will be written. If it is not specified, all files are
10330 written to the current directory.
10332 For example, given a
10333 file called @file{hellofiles} containing
10335 @smallexample @c ada
10340 with Text_IO; use Text_IO;
10343 Put_Line ("Hello");
10353 $ gnatchop ^hellofiles^HELLOFILES.^
10357 generates two files in the current directory, one called
10358 @file{hello.ads} containing the single line that is the procedure spec,
10359 and the other called @file{hello.adb} containing the remaining text. The
10360 original file is not affected. The generated files can be compiled in
10364 When gnatchop is invoked on a file that is empty or that contains only empty
10365 lines and/or comments, gnatchop will not fail, but will not produce any
10368 For example, given a
10369 file called @file{toto.txt} containing
10371 @smallexample @c ada
10383 $ gnatchop ^toto.txt^TOT.TXT^
10387 will not produce any new file and will result in the following warnings:
10390 toto.txt:1:01: warning: empty file, contains no compilation units
10391 no compilation units found
10392 no source files written
10395 @node Switches for gnatchop
10396 @section Switches for @code{gnatchop}
10399 @command{gnatchop} recognizes the following switches:
10404 @item ^-c^/COMPILATION^
10405 @cindex @option{^-c^/COMPILATION^} (@code{gnatchop})
10406 Causes @code{gnatchop} to operate in compilation mode, in which
10407 configuration pragmas are handled according to strict RM rules. See
10408 previous section for a full description of this mode.
10412 This passes the given @option{-gnatxxx} switch to @code{gnat} which is
10413 used to parse the given file. Not all @code{xxx} options make sense,
10414 but for example, the use of @option{-gnati2} allows @code{gnatchop} to
10415 process a source file that uses Latin-2 coding for identifiers.
10419 Causes @code{gnatchop} to generate a brief help summary to the standard
10420 output file showing usage information.
10422 @item ^-k@var{mm}^/FILE_NAME_MAX_LENGTH=@var{mm}^
10423 @cindex @option{^-k^/FILE_NAME_MAX_LENGTH^} (@code{gnatchop})
10424 Limit generated file names to the specified number @code{mm}
10426 This is useful if the
10427 resulting set of files is required to be interoperable with systems
10428 which limit the length of file names.
10430 If no value is given, or
10431 if no @code{/FILE_NAME_MAX_LENGTH} qualifier is given,
10432 a default of 39, suitable for OpenVMS Alpha
10433 Systems, is assumed
10436 No space is allowed between the @option{-k} and the numeric value. The numeric
10437 value may be omitted in which case a default of @option{-k8},
10439 with DOS-like file systems, is used. If no @option{-k} switch
10441 there is no limit on the length of file names.
10444 @item ^-p^/PRESERVE^
10445 @cindex @option{^-p^/PRESERVE^} (@code{gnatchop})
10446 Causes the file ^modification^creation^ time stamp of the input file to be
10447 preserved and used for the time stamp of the output file(s). This may be
10448 useful for preserving coherency of time stamps in an environment where
10449 @code{gnatchop} is used as part of a standard build process.
10452 @cindex @option{^-q^/QUIET^} (@code{gnatchop})
10453 Causes output of informational messages indicating the set of generated
10454 files to be suppressed. Warnings and error messages are unaffected.
10456 @item ^-r^/REFERENCE^
10457 @cindex @option{^-r^/REFERENCE^} (@code{gnatchop})
10458 @findex Source_Reference
10459 Generate @code{Source_Reference} pragmas. Use this switch if the output
10460 files are regarded as temporary and development is to be done in terms
10461 of the original unchopped file. This switch causes
10462 @code{Source_Reference} pragmas to be inserted into each of the
10463 generated files to refers back to the original file name and line number.
10464 The result is that all error messages refer back to the original
10466 In addition, the debugging information placed into the object file (when
10467 the @option{^-g^/DEBUG^} switch of @command{gcc} or @command{gnatmake} is
10469 also refers back to this original file so that tools like profilers and
10470 debuggers will give information in terms of the original unchopped file.
10472 If the original file to be chopped itself contains
10473 a @code{Source_Reference}
10474 pragma referencing a third file, then gnatchop respects
10475 this pragma, and the generated @code{Source_Reference} pragmas
10476 in the chopped file refer to the original file, with appropriate
10477 line numbers. This is particularly useful when @code{gnatchop}
10478 is used in conjunction with @code{gnatprep} to compile files that
10479 contain preprocessing statements and multiple units.
10481 @item ^-v^/VERBOSE^
10482 @cindex @option{^-v^/VERBOSE^} (@code{gnatchop})
10483 Causes @code{gnatchop} to operate in verbose mode. The version
10484 number and copyright notice are output, as well as exact copies of
10485 the gnat1 commands spawned to obtain the chop control information.
10487 @item ^-w^/OVERWRITE^
10488 @cindex @option{^-w^/OVERWRITE^} (@code{gnatchop})
10489 Overwrite existing file names. Normally @code{gnatchop} regards it as a
10490 fatal error if there is already a file with the same name as a
10491 file it would otherwise output, in other words if the files to be
10492 chopped contain duplicated units. This switch bypasses this
10493 check, and causes all but the last instance of such duplicated
10494 units to be skipped.
10498 @cindex @option{--GCC=} (@code{gnatchop})
10499 Specify the path of the GNAT parser to be used. When this switch is used,
10500 no attempt is made to add the prefix to the GNAT parser executable.
10504 @node Examples of gnatchop Usage
10505 @section Examples of @code{gnatchop} Usage
10509 @item gnatchop /OVERWRITE HELLO_S.ADA [PRERELEASE.FILES]
10512 @item gnatchop -w hello_s.ada prerelease/files
10515 Chops the source file @file{hello_s.ada}. The output files will be
10516 placed in the directory @file{^prerelease/files^[PRERELEASE.FILES]^},
10518 files with matching names in that directory (no files in the current
10519 directory are modified).
10521 @item gnatchop ^archive^ARCHIVE.^
10522 Chops the source file @file{^archive^ARCHIVE.^}
10523 into the current directory. One
10524 useful application of @code{gnatchop} is in sending sets of sources
10525 around, for example in email messages. The required sources are simply
10526 concatenated (for example, using a ^Unix @code{cat}^VMS @code{APPEND/NEW}^
10528 @command{gnatchop} is used at the other end to reconstitute the original
10531 @item gnatchop file1 file2 file3 direc
10532 Chops all units in files @file{file1}, @file{file2}, @file{file3}, placing
10533 the resulting files in the directory @file{direc}. Note that if any units
10534 occur more than once anywhere within this set of files, an error message
10535 is generated, and no files are written. To override this check, use the
10536 @option{^-w^/OVERWRITE^} switch,
10537 in which case the last occurrence in the last file will
10538 be the one that is output, and earlier duplicate occurrences for a given
10539 unit will be skipped.
10542 @node Configuration Pragmas
10543 @chapter Configuration Pragmas
10544 @cindex Configuration pragmas
10545 @cindex Pragmas, configuration
10548 Configuration pragmas include those pragmas described as
10549 such in the Ada Reference Manual, as well as
10550 implementation-dependent pragmas that are configuration pragmas. See the
10551 individual descriptions of pragmas in the @cite{GNAT Reference Manual} for
10552 details on these additional GNAT-specific configuration pragmas. Most
10553 notably, the pragma @code{Source_File_Name}, which allows
10554 specifying non-default names for source files, is a configuration
10555 pragma. The following is a complete list of configuration pragmas
10556 recognized by GNAT:
10563 Component_Alignment
10569 External_Name_Casing
10570 Float_Representation
10581 Propagate_Exceptions
10584 Restricted_Run_Time
10586 Restrictions_Warnings
10591 Task_Dispatching_Policy
10600 * Handling of Configuration Pragmas::
10601 * The Configuration Pragmas Files::
10604 @node Handling of Configuration Pragmas
10605 @section Handling of Configuration Pragmas
10607 Configuration pragmas may either appear at the start of a compilation
10608 unit, in which case they apply only to that unit, or they may apply to
10609 all compilations performed in a given compilation environment.
10611 GNAT also provides the @code{gnatchop} utility to provide an automatic
10612 way to handle configuration pragmas following the semantics for
10613 compilations (that is, files with multiple units), described in the RM.
10614 See @ref{Operating gnatchop in Compilation Mode} for details.
10615 However, for most purposes, it will be more convenient to edit the
10616 @file{gnat.adc} file that contains configuration pragmas directly,
10617 as described in the following section.
10619 @node The Configuration Pragmas Files
10620 @section The Configuration Pragmas Files
10621 @cindex @file{gnat.adc}
10624 In GNAT a compilation environment is defined by the current
10625 directory at the time that a compile command is given. This current
10626 directory is searched for a file whose name is @file{gnat.adc}. If
10627 this file is present, it is expected to contain one or more
10628 configuration pragmas that will be applied to the current compilation.
10629 However, if the switch @option{-gnatA} is used, @file{gnat.adc} is not
10632 Configuration pragmas may be entered into the @file{gnat.adc} file
10633 either by running @code{gnatchop} on a source file that consists only of
10634 configuration pragmas, or more conveniently by
10635 direct editing of the @file{gnat.adc} file, which is a standard format
10638 In addition to @file{gnat.adc}, additional files containing configuration
10639 pragmas may be applied to the current compilation using the switch
10640 @option{-gnatec}@var{path}. @var{path} must designate an existing file that
10641 contains only configuration pragmas. These configuration pragmas are
10642 in addition to those found in @file{gnat.adc} (provided @file{gnat.adc}
10643 is present and switch @option{-gnatA} is not used).
10645 It is allowed to specify several switches @option{-gnatec}, all of which
10646 will be taken into account.
10648 If you are using project file, a separate mechanism is provided using
10649 project attributes, see @ref{Specifying Configuration Pragmas} for more
10653 Of special interest to GNAT OpenVMS Alpha is the following
10654 configuration pragma:
10656 @smallexample @c ada
10658 pragma Extend_System (Aux_DEC);
10663 In the presence of this pragma, GNAT adds to the definition of the
10664 predefined package SYSTEM all the additional types and subprograms that are
10665 defined in HP Ada. See @ref{Compatibility with HP Ada} for details.
10668 @node Handling Arbitrary File Naming Conventions Using gnatname
10669 @chapter Handling Arbitrary File Naming Conventions Using @code{gnatname}
10670 @cindex Arbitrary File Naming Conventions
10673 * Arbitrary File Naming Conventions::
10674 * Running gnatname::
10675 * Switches for gnatname::
10676 * Examples of gnatname Usage::
10679 @node Arbitrary File Naming Conventions
10680 @section Arbitrary File Naming Conventions
10683 The GNAT compiler must be able to know the source file name of a compilation
10684 unit. When using the standard GNAT default file naming conventions
10685 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
10686 does not need additional information.
10689 When the source file names do not follow the standard GNAT default file naming
10690 conventions, the GNAT compiler must be given additional information through
10691 a configuration pragmas file (@pxref{Configuration Pragmas})
10693 When the non standard file naming conventions are well-defined,
10694 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
10695 (@pxref{Alternative File Naming Schemes}) may be sufficient. However,
10696 if the file naming conventions are irregular or arbitrary, a number
10697 of pragma @code{Source_File_Name} for individual compilation units
10699 To help maintain the correspondence between compilation unit names and
10700 source file names within the compiler,
10701 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
10704 @node Running gnatname
10705 @section Running @code{gnatname}
10708 The usual form of the @code{gnatname} command is
10711 $ gnatname [@var{switches}] @var{naming_pattern} [@var{naming_patterns}]
10715 All of the arguments are optional. If invoked without any argument,
10716 @code{gnatname} will display its usage.
10719 When used with at least one naming pattern, @code{gnatname} will attempt to
10720 find all the compilation units in files that follow at least one of the
10721 naming patterns. To find these compilation units,
10722 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
10726 One or several Naming Patterns may be given as arguments to @code{gnatname}.
10727 Each Naming Pattern is enclosed between double quotes.
10728 A Naming Pattern is a regular expression similar to the wildcard patterns
10729 used in file names by the Unix shells or the DOS prompt.
10732 Examples of Naming Patterns are
10741 For a more complete description of the syntax of Naming Patterns,
10742 see the second kind of regular expressions described in @file{g-regexp.ads}
10743 (the ``Glob'' regular expressions).
10746 When invoked with no switches, @code{gnatname} will create a configuration
10747 pragmas file @file{gnat.adc} in the current working directory, with pragmas
10748 @code{Source_File_Name} for each file that contains a valid Ada unit.
10750 @node Switches for gnatname
10751 @section Switches for @code{gnatname}
10754 Switches for @code{gnatname} must precede any specified Naming Pattern.
10757 You may specify any of the following switches to @code{gnatname}:
10762 @item ^-c^/CONFIG_FILE=^@file{file}
10763 @cindex @option{^-c^/CONFIG_FILE^} (@code{gnatname})
10764 Create a configuration pragmas file @file{file} (instead of the default
10767 There may be zero, one or more space between @option{-c} and
10770 @file{file} may include directory information. @file{file} must be
10771 writable. There may be only one switch @option{^-c^/CONFIG_FILE^}.
10772 When a switch @option{^-c^/CONFIG_FILE^} is
10773 specified, no switch @option{^-P^/PROJECT_FILE^} may be specified (see below).
10775 @item ^-d^/SOURCE_DIRS=^@file{dir}
10776 @cindex @option{^-d^/SOURCE_DIRS^} (@code{gnatname})
10777 Look for source files in directory @file{dir}. There may be zero, one or more
10778 spaces between @option{^-d^/SOURCE_DIRS=^} and @file{dir}.
10779 When a switch @option{^-d^/SOURCE_DIRS^}
10780 is specified, the current working directory will not be searched for source
10781 files, unless it is explicitly specified with a @option{^-d^/SOURCE_DIRS^}
10782 or @option{^-D^/DIR_FILES^} switch.
10783 Several switches @option{^-d^/SOURCE_DIRS^} may be specified.
10784 If @file{dir} is a relative path, it is relative to the directory of
10785 the configuration pragmas file specified with switch
10786 @option{^-c^/CONFIG_FILE^},
10787 or to the directory of the project file specified with switch
10788 @option{^-P^/PROJECT_FILE^} or,
10789 if neither switch @option{^-c^/CONFIG_FILE^}
10790 nor switch @option{^-P^/PROJECT_FILE^} are specified, it is relative to the
10791 current working directory. The directory
10792 specified with switch @option{^-d^/SOURCE_DIRS^} must exist and be readable.
10794 @item ^-D^/DIRS_FILE=^@file{file}
10795 @cindex @option{^-D^/DIRS_FILE^} (@code{gnatname})
10796 Look for source files in all directories listed in text file @file{file}.
10797 There may be zero, one or more spaces between @option{^-D^/DIRS_FILE=^}
10799 @file{file} must be an existing, readable text file.
10800 Each non empty line in @file{file} must be a directory.
10801 Specifying switch @option{^-D^/DIRS_FILE^} is equivalent to specifying as many
10802 switches @option{^-d^/SOURCE_DIRS^} as there are non empty lines in
10805 @item ^-f^/FOREIGN_PATTERN=^@file{pattern}
10806 @cindex @option{^-f^/FOREIGN_PATTERN^} (@code{gnatname})
10807 Foreign patterns. Using this switch, it is possible to add sources of languages
10808 other than Ada to the list of sources of a project file.
10809 It is only useful if a ^-P^/PROJECT_FILE^ switch is used.
10812 gnatname ^-Pprj -f"*.c"^/PROJECT_FILE=PRJ /FOREIGN_PATTERN=*.C^ "*.ada"
10815 will look for Ada units in all files with the @file{.ada} extension,
10816 and will add to the list of file for project @file{prj.gpr} the C files
10817 with extension ".^c^C^".
10820 @cindex @option{^-h^/HELP^} (@code{gnatname})
10821 Output usage (help) information. The output is written to @file{stdout}.
10823 @item ^-P^/PROJECT_FILE=^@file{proj}
10824 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatname})
10825 Create or update project file @file{proj}. There may be zero, one or more space
10826 between @option{-P} and @file{proj}. @file{proj} may include directory
10827 information. @file{proj} must be writable.
10828 There may be only one switch @option{^-P^/PROJECT_FILE^}.
10829 When a switch @option{^-P^/PROJECT_FILE^} is specified,
10830 no switch @option{^-c^/CONFIG_FILE^} may be specified.
10832 @item ^-v^/VERBOSE^
10833 @cindex @option{^-v^/VERBOSE^} (@code{gnatname})
10834 Verbose mode. Output detailed explanation of behavior to @file{stdout}.
10835 This includes name of the file written, the name of the directories to search
10836 and, for each file in those directories whose name matches at least one of
10837 the Naming Patterns, an indication of whether the file contains a unit,
10838 and if so the name of the unit.
10840 @item ^-v -v^/VERBOSE /VERBOSE^
10841 @cindex @option{^-v -v^/VERBOSE /VERBOSE^} (@code{gnatname})
10842 Very Verbose mode. In addition to the output produced in verbose mode,
10843 for each file in the searched directories whose name matches none of
10844 the Naming Patterns, an indication is given that there is no match.
10846 @item ^-x^/EXCLUDED_PATTERN=^@file{pattern}
10847 @cindex @option{^-x^/EXCLUDED_PATTERN^} (@code{gnatname})
10848 Excluded patterns. Using this switch, it is possible to exclude some files
10849 that would match the name patterns. For example,
10851 gnatname ^-x "*_nt.ada"^/EXCLUDED_PATTERN=*_nt.ada^ "*.ada"
10854 will look for Ada units in all files with the @file{.ada} extension,
10855 except those whose names end with @file{_nt.ada}.
10859 @node Examples of gnatname Usage
10860 @section Examples of @code{gnatname} Usage
10864 $ gnatname /CONFIG_FILE=[HOME.ME]NAMES.ADC /SOURCE_DIRS=SOURCES "[a-z]*.ada*"
10870 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
10875 In this example, the directory @file{^/home/me^[HOME.ME]^} must already exist
10876 and be writable. In addition, the directory
10877 @file{^/home/me/sources^[HOME.ME.SOURCES]^} (specified by
10878 @option{^-d sources^/SOURCE_DIRS=SOURCES^}) must exist and be readable.
10881 Note the optional spaces after @option{-c} and @option{-d}.
10886 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
10887 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
10890 $ gnatname /PROJECT_FILE=[HOME.ME]PROJ
10891 /EXCLUDED_PATTERN=*_nt_body.ada
10892 /SOURCE_DIRS=(SOURCES,[SOURCES.PLUS])
10893 /DIRS_FILE=COMMON_DIRS.TXT "body_*" "spec_*"
10897 Note that several switches @option{^-d^/SOURCE_DIRS^} may be used,
10898 even in conjunction with one or several switches
10899 @option{^-D^/DIRS_FILE^}. Several Naming Patterns and one excluded pattern
10900 are used in this example.
10902 @c *****************************************
10903 @c * G N A T P r o j e c t M a n a g e r *
10904 @c *****************************************
10905 @node GNAT Project Manager
10906 @chapter GNAT Project Manager
10910 * Examples of Project Files::
10911 * Project File Syntax::
10912 * Objects and Sources in Project Files::
10913 * Importing Projects::
10914 * Project Extension::
10915 * Project Hierarchy Extension::
10916 * External References in Project Files::
10917 * Packages in Project Files::
10918 * Variables from Imported Projects::
10920 * Library Projects::
10921 * Stand-alone Library Projects::
10922 * Switches Related to Project Files::
10923 * Tools Supporting Project Files::
10924 * An Extended Example::
10925 * Project File Complete Syntax::
10928 @c ****************
10929 @c * Introduction *
10930 @c ****************
10933 @section Introduction
10936 This chapter describes GNAT's @emph{Project Manager}, a facility that allows
10937 you to manage complex builds involving a number of source files, directories,
10938 and compilation options for different system configurations. In particular,
10939 project files allow you to specify:
10942 The directory or set of directories containing the source files, and/or the
10943 names of the specific source files themselves
10945 The directory in which the compiler's output
10946 (@file{ALI} files, object files, tree files) is to be placed
10948 The directory in which the executable programs is to be placed
10950 ^Switch^Switch^ settings for any of the project-enabled tools
10951 (@command{gnatmake}, compiler, binder, linker, @code{gnatls}, @code{gnatxref},
10952 @code{gnatfind}); you can apply these settings either globally or to individual
10955 The source files containing the main subprogram(s) to be built
10957 The source programming language(s) (currently Ada and/or C)
10959 Source file naming conventions; you can specify these either globally or for
10960 individual compilation units
10967 @node Project Files
10968 @subsection Project Files
10971 Project files are written in a syntax close to that of Ada, using familiar
10972 notions such as packages, context clauses, declarations, default values,
10973 assignments, and inheritance. Finally, project files can be built
10974 hierarchically from other project files, simplifying complex system
10975 integration and project reuse.
10977 A @dfn{project} is a specific set of values for various compilation properties.
10978 The settings for a given project are described by means of
10979 a @dfn{project file}, which is a text file written in an Ada-like syntax.
10980 Property values in project files are either strings or lists of strings.
10981 Properties that are not explicitly set receive default values. A project
10982 file may interrogate the values of @dfn{external variables} (user-defined
10983 command-line switches or environment variables), and it may specify property
10984 settings conditionally, based on the value of such variables.
10986 In simple cases, a project's source files depend only on other source files
10987 in the same project, or on the predefined libraries. (@emph{Dependence} is
10989 the Ada technical sense; as in one Ada unit @code{with}ing another.) However,
10990 the Project Manager also allows more sophisticated arrangements,
10991 where the source files in one project depend on source files in other
10995 One project can @emph{import} other projects containing needed source files.
10997 You can organize GNAT projects in a hierarchy: a @emph{child} project
10998 can extend a @emph{parent} project, inheriting the parent's source files and
10999 optionally overriding any of them with alternative versions
11003 More generally, the Project Manager lets you structure large development
11004 efforts into hierarchical subsystems, where build decisions are delegated
11005 to the subsystem level, and thus different compilation environments
11006 (^switch^switch^ settings) used for different subsystems.
11008 The Project Manager is invoked through the
11009 @option{^-P^/PROJECT_FILE=^@emph{projectfile}}
11010 switch to @command{gnatmake} or to the @command{^gnat^GNAT^} front driver.
11012 There may be zero, one or more spaces between @option{-P} and
11013 @option{@emph{projectfile}}.
11015 If you want to define (on the command line) an external variable that is
11016 queried by the project file, you must use the
11017 @option{^-X^/EXTERNAL_REFERENCE=^@emph{vbl}=@emph{value}} switch.
11018 The Project Manager parses and interprets the project file, and drives the
11019 invoked tool based on the project settings.
11021 The Project Manager supports a wide range of development strategies,
11022 for systems of all sizes. Here are some typical practices that are
11026 Using a common set of source files, but generating object files in different
11027 directories via different ^switch^switch^ settings
11029 Using a mostly-shared set of source files, but with different versions of
11034 The destination of an executable can be controlled inside a project file
11035 using the @option{^-o^-o^}
11037 In the absence of such a ^switch^switch^ either inside
11038 the project file or on the command line, any executable files generated by
11039 @command{gnatmake} are placed in the directory @code{Exec_Dir} specified
11040 in the project file. If no @code{Exec_Dir} is specified, they will be placed
11041 in the object directory of the project.
11043 You can use project files to achieve some of the effects of a source
11044 versioning system (for example, defining separate projects for
11045 the different sets of sources that comprise different releases) but the
11046 Project Manager is independent of any source configuration management tools
11047 that might be used by the developers.
11049 The next section introduces the main features of GNAT's project facility
11050 through a sequence of examples; subsequent sections will present the syntax
11051 and semantics in more detail. A more formal description of the project
11052 facility appears in the GNAT Reference Manual.
11054 @c *****************************
11055 @c * Examples of Project Files *
11056 @c *****************************
11058 @node Examples of Project Files
11059 @section Examples of Project Files
11061 This section illustrates some of the typical uses of project files and
11062 explains their basic structure and behavior.
11065 * Common Sources with Different ^Switches^Switches^ and Directories::
11066 * Using External Variables::
11067 * Importing Other Projects::
11068 * Extending a Project::
11071 @node Common Sources with Different ^Switches^Switches^ and Directories
11072 @subsection Common Sources with Different ^Switches^Switches^ and Directories
11076 * Specifying the Object Directory::
11077 * Specifying the Exec Directory::
11078 * Project File Packages::
11079 * Specifying ^Switch^Switch^ Settings::
11080 * Main Subprograms::
11081 * Executable File Names::
11082 * Source File Naming Conventions::
11083 * Source Language(s)::
11087 Suppose that the Ada source files @file{pack.ads}, @file{pack.adb}, and
11088 @file{proc.adb} are in the @file{/common} directory. The file
11089 @file{proc.adb} contains an Ada main subprogram @code{Proc} that @code{with}s
11090 package @code{Pack}. We want to compile these source files under two sets
11091 of ^switches^switches^:
11094 When debugging, we want to pass the @option{-g} switch to @command{gnatmake},
11095 and the @option{^-gnata^-gnata^},
11096 @option{^-gnato^-gnato^},
11097 and @option{^-gnatE^-gnatE^} switches to the
11098 compiler; the compiler's output is to appear in @file{/common/debug}
11100 When preparing a release version, we want to pass the @option{^-O2^O2^} switch
11101 to the compiler; the compiler's output is to appear in @file{/common/release}
11105 The GNAT project files shown below, respectively @file{debug.gpr} and
11106 @file{release.gpr} in the @file{/common} directory, achieve these effects.
11119 ^/common/debug^[COMMON.DEBUG]^
11124 ^/common/release^[COMMON.RELEASE]^
11129 Here are the corresponding project files:
11131 @smallexample @c projectfile
11134 for Object_Dir use "debug";
11135 for Main use ("proc");
11138 for ^Default_Switches^Default_Switches^ ("Ada")
11140 for Executable ("proc.adb") use "proc1";
11145 package Compiler is
11146 for ^Default_Switches^Default_Switches^ ("Ada")
11147 use ("-fstack-check",
11150 "^-gnatE^-gnatE^");
11156 @smallexample @c projectfile
11159 for Object_Dir use "release";
11160 for Exec_Dir use ".";
11161 for Main use ("proc");
11163 package Compiler is
11164 for ^Default_Switches^Default_Switches^ ("Ada")
11172 The name of the project defined by @file{debug.gpr} is @code{"Debug"} (case
11173 insensitive), and analogously the project defined by @file{release.gpr} is
11174 @code{"Release"}. For consistency the file should have the same name as the
11175 project, and the project file's extension should be @code{"gpr"}. These
11176 conventions are not required, but a warning is issued if they are not followed.
11178 If the current directory is @file{^/temp^[TEMP]^}, then the command
11180 gnatmake ^-P/common/debug.gpr^/PROJECT_FILE=[COMMON]DEBUG^
11184 generates object and ALI files in @file{^/common/debug^[COMMON.DEBUG]^},
11185 as well as the @code{^proc1^PROC1.EXE^} executable,
11186 using the ^switch^switch^ settings defined in the project file.
11188 Likewise, the command
11190 gnatmake ^-P/common/release.gpr^/PROJECT_FILE=[COMMON]RELEASE^
11194 generates object and ALI files in @file{^/common/release^[COMMON.RELEASE]^},
11195 and the @code{^proc^PROC.EXE^}
11196 executable in @file{^/common^[COMMON]^},
11197 using the ^switch^switch^ settings from the project file.
11200 @unnumberedsubsubsec Source Files
11203 If a project file does not explicitly specify a set of source directories or
11204 a set of source files, then by default the project's source files are the
11205 Ada source files in the project file directory. Thus @file{pack.ads},
11206 @file{pack.adb}, and @file{proc.adb} are the source files for both projects.
11208 @node Specifying the Object Directory
11209 @unnumberedsubsubsec Specifying the Object Directory
11212 Several project properties are modeled by Ada-style @emph{attributes};
11213 a property is defined by supplying the equivalent of an Ada attribute
11214 definition clause in the project file.
11215 A project's object directory is another such a property; the corresponding
11216 attribute is @code{Object_Dir}, and its value is also a string expression,
11217 specified either as absolute or relative. In the later case,
11218 it is relative to the project file directory. Thus the compiler's
11219 output is directed to @file{^/common/debug^[COMMON.DEBUG]^}
11220 (for the @code{Debug} project)
11221 and to @file{^/common/release^[COMMON.RELEASE]^}
11222 (for the @code{Release} project).
11223 If @code{Object_Dir} is not specified, then the default is the project file
11226 @node Specifying the Exec Directory
11227 @unnumberedsubsubsec Specifying the Exec Directory
11230 A project's exec directory is another property; the corresponding
11231 attribute is @code{Exec_Dir}, and its value is also a string expression,
11232 either specified as relative or absolute. If @code{Exec_Dir} is not specified,
11233 then the default is the object directory (which may also be the project file
11234 directory if attribute @code{Object_Dir} is not specified). Thus the executable
11235 is placed in @file{^/common/debug^[COMMON.DEBUG]^}
11236 for the @code{Debug} project (attribute @code{Exec_Dir} not specified)
11237 and in @file{^/common^[COMMON]^} for the @code{Release} project.
11239 @node Project File Packages
11240 @unnumberedsubsubsec Project File Packages
11243 A GNAT tool that is integrated with the Project Manager is modeled by a
11244 corresponding package in the project file. In the example above,
11245 The @code{Debug} project defines the packages @code{Builder}
11246 (for @command{gnatmake}) and @code{Compiler};
11247 the @code{Release} project defines only the @code{Compiler} package.
11249 The Ada-like package syntax is not to be taken literally. Although packages in
11250 project files bear a surface resemblance to packages in Ada source code, the
11251 notation is simply a way to convey a grouping of properties for a named
11252 entity. Indeed, the package names permitted in project files are restricted
11253 to a predefined set, corresponding to the project-aware tools, and the contents
11254 of packages are limited to a small set of constructs.
11255 The packages in the example above contain attribute definitions.
11257 @node Specifying ^Switch^Switch^ Settings
11258 @unnumberedsubsubsec Specifying ^Switch^Switch^ Settings
11261 ^Switch^Switch^ settings for a project-aware tool can be specified through
11262 attributes in the package that corresponds to the tool.
11263 The example above illustrates one of the relevant attributes,
11264 @code{^Default_Switches^Default_Switches^}, which is defined in packages
11265 in both project files.
11266 Unlike simple attributes like @code{Source_Dirs},
11267 @code{^Default_Switches^Default_Switches^} is
11268 known as an @emph{associative array}. When you define this attribute, you must
11269 supply an ``index'' (a literal string), and the effect of the attribute
11270 definition is to set the value of the array at the specified index.
11271 For the @code{^Default_Switches^Default_Switches^} attribute,
11272 the index is a programming language (in our case, Ada),
11273 and the value specified (after @code{use}) must be a list
11274 of string expressions.
11276 The attributes permitted in project files are restricted to a predefined set.
11277 Some may appear at project level, others in packages.
11278 For any attribute that is an associative array, the index must always be a
11279 literal string, but the restrictions on this string (e.g., a file name or a
11280 language name) depend on the individual attribute.
11281 Also depending on the attribute, its specified value will need to be either a
11282 string or a string list.
11284 In the @code{Debug} project, we set the switches for two tools,
11285 @command{gnatmake} and the compiler, and thus we include the two corresponding
11286 packages; each package defines the @code{^Default_Switches^Default_Switches^}
11287 attribute with index @code{"Ada"}.
11288 Note that the package corresponding to
11289 @command{gnatmake} is named @code{Builder}. The @code{Release} project is
11290 similar, but only includes the @code{Compiler} package.
11292 In project @code{Debug} above, the ^switches^switches^ starting with
11293 @option{-gnat} that are specified in package @code{Compiler}
11294 could have been placed in package @code{Builder}, since @command{gnatmake}
11295 transmits all such ^switches^switches^ to the compiler.
11297 @node Main Subprograms
11298 @unnumberedsubsubsec Main Subprograms
11301 One of the specifiable properties of a project is a list of files that contain
11302 main subprograms. This property is captured in the @code{Main} attribute,
11303 whose value is a list of strings. If a project defines the @code{Main}
11304 attribute, it is not necessary to identify the main subprogram(s) when
11305 invoking @command{gnatmake} (@pxref{gnatmake and Project Files}).
11307 @node Executable File Names
11308 @unnumberedsubsubsec Executable File Names
11311 By default, the executable file name corresponding to a main source is
11312 deduced from the main source file name. Through the attributes
11313 @code{Executable} and @code{Executable_Suffix} of package @code{Builder},
11314 it is possible to change this default.
11315 In project @code{Debug} above, the executable file name
11316 for main source @file{^proc.adb^PROC.ADB^} is
11317 @file{^proc1^PROC1.EXE^}.
11318 Attribute @code{Executable_Suffix}, when specified, may change the suffix
11319 of the executable files, when no attribute @code{Executable} applies:
11320 its value replace the platform-specific executable suffix.
11321 Attributes @code{Executable} and @code{Executable_Suffix} are the only ways to
11322 specify a non default executable file name when several mains are built at once
11323 in a single @command{gnatmake} command.
11325 @node Source File Naming Conventions
11326 @unnumberedsubsubsec Source File Naming Conventions
11329 Since the project files above do not specify any source file naming
11330 conventions, the GNAT defaults are used. The mechanism for defining source
11331 file naming conventions -- a package named @code{Naming} --
11332 is described below (@pxref{Naming Schemes}).
11334 @node Source Language(s)
11335 @unnumberedsubsubsec Source Language(s)
11338 Since the project files do not specify a @code{Languages} attribute, by
11339 default the GNAT tools assume that the language of the project file is Ada.
11340 More generally, a project can comprise source files
11341 in Ada, C, and/or other languages.
11343 @node Using External Variables
11344 @subsection Using External Variables
11347 Instead of supplying different project files for debug and release, we can
11348 define a single project file that queries an external variable (set either
11349 on the command line or via an ^environment variable^logical name^) in order to
11350 conditionally define the appropriate settings. Again, assume that the
11351 source files @file{pack.ads}, @file{pack.adb}, and @file{proc.adb} are
11352 located in directory @file{^/common^[COMMON]^}. The following project file,
11353 @file{build.gpr}, queries the external variable named @code{STYLE} and
11354 defines an object directory and ^switch^switch^ settings based on whether
11355 the value is @code{"deb"} (debug) or @code{"rel"} (release), and where
11356 the default is @code{"deb"}.
11358 @smallexample @c projectfile
11361 for Main use ("proc");
11363 type Style_Type is ("deb", "rel");
11364 Style : Style_Type := external ("STYLE", "deb");
11368 for Object_Dir use "debug";
11371 for Object_Dir use "release";
11372 for Exec_Dir use ".";
11381 for ^Default_Switches^Default_Switches^ ("Ada")
11383 for Executable ("proc") use "proc1";
11392 package Compiler is
11396 for ^Default_Switches^Default_Switches^ ("Ada")
11397 use ("^-gnata^-gnata^",
11399 "^-gnatE^-gnatE^");
11402 for ^Default_Switches^Default_Switches^ ("Ada")
11413 @code{Style_Type} is an example of a @emph{string type}, which is the project
11414 file analog of an Ada enumeration type but whose components are string literals
11415 rather than identifiers. @code{Style} is declared as a variable of this type.
11417 The form @code{external("STYLE", "deb")} is known as an
11418 @emph{external reference}; its first argument is the name of an
11419 @emph{external variable}, and the second argument is a default value to be
11420 used if the external variable doesn't exist. You can define an external
11421 variable on the command line via the @option{^-X^/EXTERNAL_REFERENCE^} switch,
11422 or you can use ^an environment variable^a logical name^
11423 as an external variable.
11425 Each @code{case} construct is expanded by the Project Manager based on the
11426 value of @code{Style}. Thus the command
11429 gnatmake -P/common/build.gpr -XSTYLE=deb
11435 gnatmake /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=deb
11440 is equivalent to the @command{gnatmake} invocation using the project file
11441 @file{debug.gpr} in the earlier example. So is the command
11443 gnatmake ^-P/common/build.gpr^/PROJECT_FILE=[COMMON]BUILD.GPR^
11447 since @code{"deb"} is the default for @code{STYLE}.
11453 gnatmake -P/common/build.gpr -XSTYLE=rel
11459 GNAT MAKE /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=rel
11464 is equivalent to the @command{gnatmake} invocation using the project file
11465 @file{release.gpr} in the earlier example.
11467 @node Importing Other Projects
11468 @subsection Importing Other Projects
11469 @cindex @code{ADA_PROJECT_PATH}
11472 A compilation unit in a source file in one project may depend on compilation
11473 units in source files in other projects. To compile this unit under
11474 control of a project file, the
11475 dependent project must @emph{import} the projects containing the needed source
11477 This effect is obtained using syntax similar to an Ada @code{with} clause,
11478 but where @code{with}ed entities are strings that denote project files.
11480 As an example, suppose that the two projects @code{GUI_Proj} and
11481 @code{Comm_Proj} are defined in the project files @file{gui_proj.gpr} and
11482 @file{comm_proj.gpr} in directories @file{^/gui^[GUI]^}
11483 and @file{^/comm^[COMM]^}, respectively.
11484 Suppose that the source files for @code{GUI_Proj} are
11485 @file{gui.ads} and @file{gui.adb}, and that the source files for
11486 @code{Comm_Proj} are @file{comm.ads} and @file{comm.adb}, where each set of
11487 files is located in its respective project file directory. Schematically:
11506 We want to develop an application in directory @file{^/app^[APP]^} that
11507 @code{with} the packages @code{GUI} and @code{Comm}, using the properties of
11508 the corresponding project files (e.g. the ^switch^switch^ settings
11509 and object directory).
11510 Skeletal code for a main procedure might be something like the following:
11512 @smallexample @c ada
11515 procedure App_Main is
11524 Here is a project file, @file{app_proj.gpr}, that achieves the desired
11527 @smallexample @c projectfile
11529 with "/gui/gui_proj", "/comm/comm_proj";
11530 project App_Proj is
11531 for Main use ("app_main");
11537 Building an executable is achieved through the command:
11539 gnatmake ^-P/app/app_proj^/PROJECT_FILE=[APP]APP_PROJ^
11542 which will generate the @code{^app_main^APP_MAIN.EXE^} executable
11543 in the directory where @file{app_proj.gpr} resides.
11545 If an imported project file uses the standard extension (@code{^gpr^GPR^}) then
11546 (as illustrated above) the @code{with} clause can omit the extension.
11548 Our example specified an absolute path for each imported project file.
11549 Alternatively, the directory name of an imported object can be omitted
11553 The imported project file is in the same directory as the importing project
11556 You have defined ^an environment variable^a logical name^
11557 that includes the directory containing
11558 the needed project file. The syntax of @code{ADA_PROJECT_PATH} is the same as
11559 the syntax of @code{ADA_INCLUDE_PATH} and @code{ADA_OBJECTS_PATH}: a list of
11560 directory names separated by colons (semicolons on Windows).
11564 Thus, if we define @code{ADA_PROJECT_PATH} to include @file{^/gui^[GUI]^} and
11565 @file{^/comm^[COMM]^}, then our project file @file{app_proj.gpr} can be written
11568 @smallexample @c projectfile
11570 with "gui_proj", "comm_proj";
11571 project App_Proj is
11572 for Main use ("app_main");
11578 Importing other projects can create ambiguities.
11579 For example, the same unit might be present in different imported projects, or
11580 it might be present in both the importing project and in an imported project.
11581 Both of these conditions are errors. Note that in the current version of
11582 the Project Manager, it is illegal to have an ambiguous unit even if the
11583 unit is never referenced by the importing project. This restriction may be
11584 relaxed in a future release.
11586 @node Extending a Project
11587 @subsection Extending a Project
11590 In large software systems it is common to have multiple
11591 implementations of a common interface; in Ada terms, multiple versions of a
11592 package body for the same specification. For example, one implementation
11593 might be safe for use in tasking programs, while another might only be used
11594 in sequential applications. This can be modeled in GNAT using the concept
11595 of @emph{project extension}. If one project (the ``child'') @emph{extends}
11596 another project (the ``parent'') then by default all source files of the
11597 parent project are inherited by the child, but the child project can
11598 override any of the parent's source files with new versions, and can also
11599 add new files. This facility is the project analog of a type extension in
11600 Object-Oriented Programming. Project hierarchies are permitted (a child
11601 project may be the parent of yet another project), and a project that
11602 inherits one project can also import other projects.
11604 As an example, suppose that directory @file{^/seq^[SEQ]^} contains the project
11605 file @file{seq_proj.gpr} as well as the source files @file{pack.ads},
11606 @file{pack.adb}, and @file{proc.adb}:
11619 Note that the project file can simply be empty (that is, no attribute or
11620 package is defined):
11622 @smallexample @c projectfile
11624 project Seq_Proj is
11630 implying that its source files are all the Ada source files in the project
11633 Suppose we want to supply an alternate version of @file{pack.adb}, in
11634 directory @file{^/tasking^[TASKING]^}, but use the existing versions of
11635 @file{pack.ads} and @file{proc.adb}. We can define a project
11636 @code{Tasking_Proj} that inherits @code{Seq_Proj}:
11640 ^/tasking^[TASKING]^
11646 project Tasking_Proj extends "/seq/seq_proj" is
11652 The version of @file{pack.adb} used in a build depends on which project file
11655 Note that we could have obtained the desired behavior using project import
11656 rather than project inheritance; a @code{base} project would contain the
11657 sources for @file{pack.ads} and @file{proc.adb}, a sequential project would
11658 import @code{base} and add @file{pack.adb}, and likewise a tasking project
11659 would import @code{base} and add a different version of @file{pack.adb}. The
11660 choice depends on whether other sources in the original project need to be
11661 overridden. If they do, then project extension is necessary, otherwise,
11662 importing is sufficient.
11665 In a project file that extends another project file, it is possible to
11666 indicate that an inherited source is not part of the sources of the extending
11667 project. This is necessary sometimes when a package spec has been overloaded
11668 and no longer requires a body: in this case, it is necessary to indicate that
11669 the inherited body is not part of the sources of the project, otherwise there
11670 will be a compilation error when compiling the spec.
11672 For that purpose, the attribute @code{Locally_Removed_Files} is used.
11673 Its value is a string list: a list of file names.
11675 @smallexample @c @projectfile
11676 project B extends "a" is
11677 for Source_Files use ("pkg.ads");
11678 -- New spec of Pkg does not need a completion
11679 for Locally_Removed_Files use ("pkg.adb");
11683 Attribute @code{Locally_Removed_Files} may also be used to check if a source
11684 is still needed: if it is possible to build using @command{gnatmake} when such
11685 a source is put in attribute @code{Locally_Removed_Files} of a project P, then
11686 it is possible to remove the source completely from a system that includes
11689 @c ***********************
11690 @c * Project File Syntax *
11691 @c ***********************
11693 @node Project File Syntax
11694 @section Project File Syntax
11703 * Associative Array Attributes::
11704 * case Constructions::
11708 This section describes the structure of project files.
11710 A project may be an @emph{independent project}, entirely defined by a single
11711 project file. Any Ada source file in an independent project depends only
11712 on the predefined library and other Ada source files in the same project.
11715 A project may also @dfn{depend on} other projects, in either or both of
11716 the following ways:
11718 @item It may import any number of projects
11719 @item It may extend at most one other project
11723 The dependence relation is a directed acyclic graph (the subgraph reflecting
11724 the ``extends'' relation is a tree).
11726 A project's @dfn{immediate sources} are the source files directly defined by
11727 that project, either implicitly by residing in the project file's directory,
11728 or explicitly through any of the source-related attributes described below.
11729 More generally, a project @var{proj}'s @dfn{sources} are the immediate sources
11730 of @var{proj} together with the immediate sources (unless overridden) of any
11731 project on which @var{proj} depends (either directly or indirectly).
11734 @subsection Basic Syntax
11737 As seen in the earlier examples, project files have an Ada-like syntax.
11738 The minimal project file is:
11739 @smallexample @c projectfile
11748 The identifier @code{Empty} is the name of the project.
11749 This project name must be present after the reserved
11750 word @code{end} at the end of the project file, followed by a semi-colon.
11752 Any name in a project file, such as the project name or a variable name,
11753 has the same syntax as an Ada identifier.
11755 The reserved words of project files are the Ada reserved words plus
11756 @code{extends}, @code{external}, and @code{project}. Note that the only Ada
11757 reserved words currently used in project file syntax are:
11785 Comments in project files have the same syntax as in Ada, two consecutives
11786 hyphens through the end of the line.
11789 @subsection Packages
11792 A project file may contain @emph{packages}. The name of a package must be one
11793 of the identifiers from the following list. A package
11794 with a given name may only appear once in a project file. Package names are
11795 case insensitive. The following package names are legal:
11811 @code{Cross_Reference}
11815 @code{Pretty_Printer}
11825 @code{Language_Processing}
11829 In its simplest form, a package may be empty:
11831 @smallexample @c projectfile
11841 A package may contain @emph{attribute declarations},
11842 @emph{variable declarations} and @emph{case constructions}, as will be
11845 When there is ambiguity between a project name and a package name,
11846 the name always designates the project. To avoid possible confusion, it is
11847 always a good idea to avoid naming a project with one of the
11848 names allowed for packages or any name that starts with @code{gnat}.
11851 @subsection Expressions
11854 An @emph{expression} is either a @emph{string expression} or a
11855 @emph{string list expression}.
11857 A @emph{string expression} is either a @emph{simple string expression} or a
11858 @emph{compound string expression}.
11860 A @emph{simple string expression} is one of the following:
11862 @item A literal string; e.g.@code{"comm/my_proj.gpr"}
11863 @item A string-valued variable reference (@pxref{Variables})
11864 @item A string-valued attribute reference (@pxref{Attributes})
11865 @item An external reference (@pxref{External References in Project Files})
11869 A @emph{compound string expression} is a concatenation of string expressions,
11870 using the operator @code{"&"}
11872 Path & "/" & File_Name & ".ads"
11876 A @emph{string list expression} is either a
11877 @emph{simple string list expression} or a
11878 @emph{compound string list expression}.
11880 A @emph{simple string list expression} is one of the following:
11882 @item A parenthesized list of zero or more string expressions,
11883 separated by commas
11885 File_Names := (File_Name, "gnat.adc", File_Name & ".orig");
11888 @item A string list-valued variable reference
11889 @item A string list-valued attribute reference
11893 A @emph{compound string list expression} is the concatenation (using
11894 @code{"&"}) of a simple string list expression and an expression. Note that
11895 each term in a compound string list expression, except the first, may be
11896 either a string expression or a string list expression.
11898 @smallexample @c projectfile
11900 File_Name_List := () & File_Name; -- One string in this list
11901 Extended_File_Name_List := File_Name_List & (File_Name & ".orig");
11903 Big_List := File_Name_List & Extended_File_Name_List;
11904 -- Concatenation of two string lists: three strings
11905 Illegal_List := "gnat.adc" & Extended_File_Name_List;
11906 -- Illegal: must start with a string list
11911 @subsection String Types
11914 A @emph{string type declaration} introduces a discrete set of string literals.
11915 If a string variable is declared to have this type, its value
11916 is restricted to the given set of literals.
11918 Here is an example of a string type declaration:
11920 @smallexample @c projectfile
11921 type OS is ("NT", "nt", "Unix", "GNU/Linux", "other OS");
11925 Variables of a string type are called @emph{typed variables}; all other
11926 variables are called @emph{untyped variables}. Typed variables are
11927 particularly useful in @code{case} constructions, to support conditional
11928 attribute declarations.
11929 (@pxref{case Constructions}).
11931 The string literals in the list are case sensitive and must all be different.
11932 They may include any graphic characters allowed in Ada, including spaces.
11934 A string type may only be declared at the project level, not inside a package.
11936 A string type may be referenced by its name if it has been declared in the same
11937 project file, or by an expanded name whose prefix is the name of the project
11938 in which it is declared.
11941 @subsection Variables
11944 A variable may be declared at the project file level, or within a package.
11945 Here are some examples of variable declarations:
11947 @smallexample @c projectfile
11949 This_OS : OS := external ("OS"); -- a typed variable declaration
11950 That_OS := "GNU/Linux"; -- an untyped variable declaration
11955 The syntax of a @emph{typed variable declaration} is identical to the Ada
11956 syntax for an object declaration. By contrast, the syntax of an untyped
11957 variable declaration is identical to an Ada assignment statement. In fact,
11958 variable declarations in project files have some of the characteristics of
11959 an assignment, in that successive declarations for the same variable are
11960 allowed. Untyped variable declarations do establish the expected kind of the
11961 variable (string or string list), and successive declarations for it must
11962 respect the initial kind.
11965 A string variable declaration (typed or untyped) declares a variable
11966 whose value is a string. This variable may be used as a string expression.
11967 @smallexample @c projectfile
11968 File_Name := "readme.txt";
11969 Saved_File_Name := File_Name & ".saved";
11973 A string list variable declaration declares a variable whose value is a list
11974 of strings. The list may contain any number (zero or more) of strings.
11976 @smallexample @c projectfile
11978 List_With_One_Element := ("^-gnaty^-gnaty^");
11979 List_With_Two_Elements := List_With_One_Element & "^-gnatg^-gnatg^";
11980 Long_List := ("main.ada", "pack1_.ada", "pack1.ada", "pack2_.ada"
11981 "pack2.ada", "util_.ada", "util.ada");
11985 The same typed variable may not be declared more than once at project level,
11986 and it may not be declared more than once in any package; it is in effect
11989 The same untyped variable may be declared several times. Declarations are
11990 elaborated in the order in which they appear, so the new value replaces
11991 the old one, and any subsequent reference to the variable uses the new value.
11992 However, as noted above, if a variable has been declared as a string, all
11994 declarations must give it a string value. Similarly, if a variable has
11995 been declared as a string list, all subsequent declarations
11996 must give it a string list value.
11998 A @emph{variable reference} may take several forms:
12001 @item The simple variable name, for a variable in the current package (if any)
12002 or in the current project
12003 @item An expanded name, whose prefix is a context name.
12007 A @emph{context} may be one of the following:
12010 @item The name of an existing package in the current project
12011 @item The name of an imported project of the current project
12012 @item The name of an ancestor project (i.e., a project extended by the current
12013 project, either directly or indirectly)
12014 @item An expanded name whose prefix is an imported/parent project name, and
12015 whose selector is a package name in that project.
12019 A variable reference may be used in an expression.
12022 @subsection Attributes
12025 A project (and its packages) may have @emph{attributes} that define
12026 the project's properties. Some attributes have values that are strings;
12027 others have values that are string lists.
12029 There are two categories of attributes: @emph{simple attributes}
12030 and @emph{associative arrays} (@pxref{Associative Array Attributes}).
12032 Legal project attribute names, and attribute names for each legal package are
12033 listed below. Attributes names are case-insensitive.
12035 The following attributes are defined on projects (all are simple attributes):
12037 @multitable @columnfractions .4 .3
12038 @item @emph{Attribute Name}
12040 @item @code{Source_Files}
12042 @item @code{Source_Dirs}
12044 @item @code{Source_List_File}
12046 @item @code{Object_Dir}
12048 @item @code{Exec_Dir}
12050 @item @code{Locally_Removed_Files}
12052 @item @code{Languages}
12056 @item @code{Library_Dir}
12058 @item @code{Library_Name}
12060 @item @code{Library_Kind}
12062 @item @code{Library_Version}
12064 @item @code{Library_Interface}
12066 @item @code{Library_Auto_Init}
12068 @item @code{Library_Options}
12070 @item @code{Library_Src_Dir}
12072 @item @code{Library_ALI_Dir}
12074 @item @code{Library_GCC}
12076 @item @code{Library_Symbol_File}
12078 @item @code{Library_Symbol_Policy}
12080 @item @code{Library_Reference_Symbol_File}
12082 @item @code{Externally_Built}
12087 The following attributes are defined for package @code{Naming}
12088 (@pxref{Naming Schemes}):
12090 @multitable @columnfractions .4 .2 .2 .2
12091 @item Attribute Name @tab Category @tab Index @tab Value
12092 @item @code{Spec_Suffix}
12093 @tab associative array
12096 @item @code{Body_Suffix}
12097 @tab associative array
12100 @item @code{Separate_Suffix}
12101 @tab simple attribute
12104 @item @code{Casing}
12105 @tab simple attribute
12108 @item @code{Dot_Replacement}
12109 @tab simple attribute
12113 @tab associative array
12117 @tab associative array
12120 @item @code{Specification_Exceptions}
12121 @tab associative array
12124 @item @code{Implementation_Exceptions}
12125 @tab associative array
12131 The following attributes are defined for packages @code{Builder},
12132 @code{Compiler}, @code{Binder},
12133 @code{Linker}, @code{Cross_Reference}, and @code{Finder}
12134 (@pxref{^Switches^Switches^ and Project Files}).
12136 @multitable @columnfractions .4 .2 .2 .2
12137 @item Attribute Name @tab Category @tab Index @tab Value
12138 @item @code{^Default_Switches^Default_Switches^}
12139 @tab associative array
12142 @item @code{^Switches^Switches^}
12143 @tab associative array
12149 In addition, package @code{Compiler} has a single string attribute
12150 @code{Local_Configuration_Pragmas} and package @code{Builder} has a single
12151 string attribute @code{Global_Configuration_Pragmas}.
12154 Each simple attribute has a default value: the empty string (for string-valued
12155 attributes) and the empty list (for string list-valued attributes).
12157 An attribute declaration defines a new value for an attribute.
12159 Examples of simple attribute declarations:
12161 @smallexample @c projectfile
12162 for Object_Dir use "objects";
12163 for Source_Dirs use ("units", "test/drivers");
12167 The syntax of a @dfn{simple attribute declaration} is similar to that of an
12168 attribute definition clause in Ada.
12170 Attributes references may be appear in expressions.
12171 The general form for such a reference is @code{<entity>'<attribute>}:
12172 Associative array attributes are functions. Associative
12173 array attribute references must have an argument that is a string literal.
12177 @smallexample @c projectfile
12179 Naming'Dot_Replacement
12180 Imported_Project'Source_Dirs
12181 Imported_Project.Naming'Casing
12182 Builder'^Default_Switches^Default_Switches^("Ada")
12186 The prefix of an attribute may be:
12188 @item @code{project} for an attribute of the current project
12189 @item The name of an existing package of the current project
12190 @item The name of an imported project
12191 @item The name of a parent project that is extended by the current project
12192 @item An expanded name whose prefix is imported/parent project name,
12193 and whose selector is a package name
12198 @smallexample @c projectfile
12201 for Source_Dirs use project'Source_Dirs & "units";
12202 for Source_Dirs use project'Source_Dirs & "test/drivers"
12208 In the first attribute declaration, initially the attribute @code{Source_Dirs}
12209 has the default value: an empty string list. After this declaration,
12210 @code{Source_Dirs} is a string list of one element: @code{"units"}.
12211 After the second attribute declaration @code{Source_Dirs} is a string list of
12212 two elements: @code{"units"} and @code{"test/drivers"}.
12214 Note: this example is for illustration only. In practice,
12215 the project file would contain only one attribute declaration:
12217 @smallexample @c projectfile
12218 for Source_Dirs use ("units", "test/drivers");
12221 @node Associative Array Attributes
12222 @subsection Associative Array Attributes
12225 Some attributes are defined as @emph{associative arrays}. An associative
12226 array may be regarded as a function that takes a string as a parameter
12227 and delivers a string or string list value as its result.
12229 Here are some examples of single associative array attribute associations:
12231 @smallexample @c projectfile
12232 for Body ("main") use "Main.ada";
12233 for ^Switches^Switches^ ("main.ada")
12235 "^-gnatv^-gnatv^");
12236 for ^Switches^Switches^ ("main.ada")
12237 use Builder'^Switches^Switches^ ("main.ada")
12242 Like untyped variables and simple attributes, associative array attributes
12243 may be declared several times. Each declaration supplies a new value for the
12244 attribute, and replaces the previous setting.
12247 An associative array attribute may be declared as a full associative array
12248 declaration, with the value of the same attribute in an imported or extended
12251 @smallexample @c projectfile
12253 for Default_Switches use Default.Builder'Default_Switches;
12258 In this example, @code{Default} must be either a project imported by the
12259 current project, or the project that the current project extends. If the
12260 attribute is in a package (in this case, in package @code{Builder}), the same
12261 package needs to be specified.
12264 A full associative array declaration replaces any other declaration for the
12265 attribute, including other full associative array declaration. Single
12266 associative array associations may be declare after a full associative
12267 declaration, modifying the value for a single association of the attribute.
12269 @node case Constructions
12270 @subsection @code{case} Constructions
12273 A @code{case} construction is used in a project file to effect conditional
12275 Here is a typical example:
12277 @smallexample @c projectfile
12280 type OS_Type is ("GNU/Linux", "Unix", "NT", "VMS");
12282 OS : OS_Type := external ("OS", "GNU/Linux");
12286 package Compiler is
12288 when "GNU/Linux" | "Unix" =>
12289 for ^Default_Switches^Default_Switches^ ("Ada")
12290 use ("^-gnath^-gnath^");
12292 for ^Default_Switches^Default_Switches^ ("Ada")
12293 use ("^-gnatP^-gnatP^");
12302 The syntax of a @code{case} construction is based on the Ada case statement
12303 (although there is no @code{null} construction for empty alternatives).
12305 The case expression must be a typed string variable.
12306 Each alternative comprises the reserved word @code{when}, either a list of
12307 literal strings separated by the @code{"|"} character or the reserved word
12308 @code{others}, and the @code{"=>"} token.
12309 Each literal string must belong to the string type that is the type of the
12311 An @code{others} alternative, if present, must occur last.
12313 After each @code{=>}, there are zero or more constructions. The only
12314 constructions allowed in a case construction are other case constructions,
12315 attribute declarations and variable declarations. String type declarations and
12316 package declarations are not allowed. Variable declarations are restricted to
12317 variables that have already been declared before the case construction.
12319 The value of the case variable is often given by an external reference
12320 (@pxref{External References in Project Files}).
12322 @c ****************************************
12323 @c * Objects and Sources in Project Files *
12324 @c ****************************************
12326 @node Objects and Sources in Project Files
12327 @section Objects and Sources in Project Files
12330 * Object Directory::
12332 * Source Directories::
12333 * Source File Names::
12337 Each project has exactly one object directory and one or more source
12338 directories. The source directories must contain at least one source file,
12339 unless the project file explicitly specifies that no source files are present
12340 (@pxref{Source File Names}).
12342 @node Object Directory
12343 @subsection Object Directory
12346 The object directory for a project is the directory containing the compiler's
12347 output (such as @file{ALI} files and object files) for the project's immediate
12350 The object directory is given by the value of the attribute @code{Object_Dir}
12351 in the project file.
12353 @smallexample @c projectfile
12354 for Object_Dir use "objects";
12358 The attribute @var{Object_Dir} has a string value, the path name of the object
12359 directory. The path name may be absolute or relative to the directory of the
12360 project file. This directory must already exist, and be readable and writable.
12362 By default, when the attribute @code{Object_Dir} is not given an explicit value
12363 or when its value is the empty string, the object directory is the same as the
12364 directory containing the project file.
12366 @node Exec Directory
12367 @subsection Exec Directory
12370 The exec directory for a project is the directory containing the executables
12371 for the project's main subprograms.
12373 The exec directory is given by the value of the attribute @code{Exec_Dir}
12374 in the project file.
12376 @smallexample @c projectfile
12377 for Exec_Dir use "executables";
12381 The attribute @var{Exec_Dir} has a string value, the path name of the exec
12382 directory. The path name may be absolute or relative to the directory of the
12383 project file. This directory must already exist, and be writable.
12385 By default, when the attribute @code{Exec_Dir} is not given an explicit value
12386 or when its value is the empty string, the exec directory is the same as the
12387 object directory of the project file.
12389 @node Source Directories
12390 @subsection Source Directories
12393 The source directories of a project are specified by the project file
12394 attribute @code{Source_Dirs}.
12396 This attribute's value is a string list. If the attribute is not given an
12397 explicit value, then there is only one source directory, the one where the
12398 project file resides.
12400 A @code{Source_Dirs} attribute that is explicitly defined to be the empty list,
12403 @smallexample @c projectfile
12404 for Source_Dirs use ();
12408 indicates that the project contains no source files.
12410 Otherwise, each string in the string list designates one or more
12411 source directories.
12413 @smallexample @c projectfile
12414 for Source_Dirs use ("sources", "test/drivers");
12418 If a string in the list ends with @code{"/**"}, then the directory whose path
12419 name precedes the two asterisks, as well as all its subdirectories
12420 (recursively), are source directories.
12422 @smallexample @c projectfile
12423 for Source_Dirs use ("/system/sources/**");
12427 Here the directory @code{/system/sources} and all of its subdirectories
12428 (recursively) are source directories.
12430 To specify that the source directories are the directory of the project file
12431 and all of its subdirectories, you can declare @code{Source_Dirs} as follows:
12432 @smallexample @c projectfile
12433 for Source_Dirs use ("./**");
12437 Each of the source directories must exist and be readable.
12439 @node Source File Names
12440 @subsection Source File Names
12443 In a project that contains source files, their names may be specified by the
12444 attributes @code{Source_Files} (a string list) or @code{Source_List_File}
12445 (a string). Source file names never include any directory information.
12447 If the attribute @code{Source_Files} is given an explicit value, then each
12448 element of the list is a source file name.
12450 @smallexample @c projectfile
12451 for Source_Files use ("main.adb");
12452 for Source_Files use ("main.adb", "pack1.ads", "pack2.adb");
12456 If the attribute @code{Source_Files} is not given an explicit value,
12457 but the attribute @code{Source_List_File} is given a string value,
12458 then the source file names are contained in the text file whose path name
12459 (absolute or relative to the directory of the project file) is the
12460 value of the attribute @code{Source_List_File}.
12462 Each line in the file that is not empty or is not a comment
12463 contains a source file name.
12465 @smallexample @c projectfile
12466 for Source_List_File use "source_list.txt";
12470 By default, if neither the attribute @code{Source_Files} nor the attribute
12471 @code{Source_List_File} is given an explicit value, then each file in the
12472 source directories that conforms to the project's naming scheme
12473 (@pxref{Naming Schemes}) is an immediate source of the project.
12475 A warning is issued if both attributes @code{Source_Files} and
12476 @code{Source_List_File} are given explicit values. In this case, the attribute
12477 @code{Source_Files} prevails.
12479 Each source file name must be the name of one existing source file
12480 in one of the source directories.
12482 A @code{Source_Files} attribute whose value is an empty list
12483 indicates that there are no source files in the project.
12485 If the order of the source directories is known statically, that is if
12486 @code{"/**"} is not used in the string list @code{Source_Dirs}, then there may
12487 be several files with the same source file name. In this case, only the file
12488 in the first directory is considered as an immediate source of the project
12489 file. If the order of the source directories is not known statically, it is
12490 an error to have several files with the same source file name.
12492 Projects can be specified to have no Ada source
12493 files: the value of (@code{Source_Dirs} or @code{Source_Files} may be an empty
12494 list, or the @code{"Ada"} may be absent from @code{Languages}:
12496 @smallexample @c projectfile
12497 for Source_Dirs use ();
12498 for Source_Files use ();
12499 for Languages use ("C", "C++");
12503 Otherwise, a project must contain at least one immediate source.
12505 Projects with no source files are useful as template packages
12506 (@pxref{Packages in Project Files}) for other projects; in particular to
12507 define a package @code{Naming} (@pxref{Naming Schemes}).
12509 @c ****************************
12510 @c * Importing Projects *
12511 @c ****************************
12513 @node Importing Projects
12514 @section Importing Projects
12515 @cindex @code{ADA_PROJECT_PATH}
12518 An immediate source of a project P may depend on source files that
12519 are neither immediate sources of P nor in the predefined library.
12520 To get this effect, P must @emph{import} the projects that contain the needed
12523 @smallexample @c projectfile
12525 with "project1", "utilities.gpr";
12526 with "/namings/apex.gpr";
12533 As can be seen in this example, the syntax for importing projects is similar
12534 to the syntax for importing compilation units in Ada. However, project files
12535 use literal strings instead of names, and the @code{with} clause identifies
12536 project files rather than packages.
12538 Each literal string is the file name or path name (absolute or relative) of a
12539 project file. If a string corresponds to a file name, with no path or a
12540 relative path, then its location is determined by the @emph{project path}. The
12541 latter can be queried using @code{gnatls -v}. It contains:
12545 In first position, the directory containing the current project file.
12547 In last position, the default project directory. This default project directory
12548 is part of the GNAT installation and is the standard place to install project
12549 files giving access to standard support libraries.
12551 @ref{Installing a library}
12555 In between, all the directories referenced in the
12556 ^environment variable^logical name^ @env{ADA_PROJECT_PATH} if it exists.
12560 If a relative pathname is used, as in
12562 @smallexample @c projectfile
12567 then the full path for the project is constructed by concatenating this
12568 relative path to those in the project path, in order, until a matching file is
12569 found. Any symbolic link will be fully resolved in the directory of the
12570 importing project file before the imported project file is examined.
12572 If the @code{with}'ed project file name does not have an extension,
12573 the default is @file{^.gpr^.GPR^}. If a file with this extension is not found,
12574 then the file name as specified in the @code{with} clause (no extension) will
12575 be used. In the above example, if a file @code{project1.gpr} is found, then it
12576 will be used; otherwise, if a file @code{^project1^PROJECT1^} exists
12577 then it will be used; if neither file exists, this is an error.
12579 A warning is issued if the name of the project file does not match the
12580 name of the project; this check is case insensitive.
12582 Any source file that is an immediate source of the imported project can be
12583 used by the immediate sources of the importing project, transitively. Thus
12584 if @code{A} imports @code{B}, and @code{B} imports @code{C}, the immediate
12585 sources of @code{A} may depend on the immediate sources of @code{C}, even if
12586 @code{A} does not import @code{C} explicitly. However, this is not recommended,
12587 because if and when @code{B} ceases to import @code{C}, some sources in
12588 @code{A} will no longer compile.
12590 A side effect of this capability is that normally cyclic dependencies are not
12591 permitted: if @code{A} imports @code{B} (directly or indirectly) then @code{B}
12592 is not allowed to import @code{A}. However, there are cases when cyclic
12593 dependencies would be beneficial. For these cases, another form of import
12594 between projects exists, the @code{limited with}: a project @code{A} that
12595 imports a project @code{B} with a straight @code{with} may also be imported,
12596 directly or indirectly, by @code{B} on the condition that imports from @code{B}
12597 to @code{A} include at least one @code{limited with}.
12599 @smallexample @c 0projectfile
12605 limited with "../a/a.gpr";
12613 limited with "../a/a.gpr";
12619 In the above legal example, there are two project cycles:
12622 @item A -> C -> D -> A
12626 In each of these cycle there is one @code{limited with}: import of @code{A}
12627 from @code{B} and import of @code{A} from @code{D}.
12629 The difference between straight @code{with} and @code{limited with} is that
12630 the name of a project imported with a @code{limited with} cannot be used in the
12631 project that imports it. In particular, its packages cannot be renamed and
12632 its variables cannot be referred to.
12634 An exception to the above rules for @code{limited with} is that for the main
12635 project specified to @command{gnatmake} or to the @command{GNAT} driver a
12636 @code{limited with} is equivalent to a straight @code{with}. For example,
12637 in the example above, projects @code{B} and @code{D} could not be main
12638 projects for @command{gnatmake} or to the @command{GNAT} driver, because they
12639 each have a @code{limited with} that is the only one in a cycle of importing
12642 @c *********************
12643 @c * Project Extension *
12644 @c *********************
12646 @node Project Extension
12647 @section Project Extension
12650 During development of a large system, it is sometimes necessary to use
12651 modified versions of some of the source files, without changing the original
12652 sources. This can be achieved through the @emph{project extension} facility.
12654 @smallexample @c projectfile
12655 project Modified_Utilities extends "/baseline/utilities.gpr" is ...
12659 A project extension declaration introduces an extending project
12660 (the @emph{child}) and a project being extended (the @emph{parent}).
12662 By default, a child project inherits all the sources of its parent.
12663 However, inherited sources can be overridden: a unit in a parent is hidden
12664 by a unit of the same name in the child.
12666 Inherited sources are considered to be sources (but not immediate sources)
12667 of the child project; see @ref{Project File Syntax}.
12669 An inherited source file retains any switches specified in the parent project.
12671 For example if the project @code{Utilities} contains the specification and the
12672 body of an Ada package @code{Util_IO}, then the project
12673 @code{Modified_Utilities} can contain a new body for package @code{Util_IO}.
12674 The original body of @code{Util_IO} will not be considered in program builds.
12675 However, the package specification will still be found in the project
12678 A child project can have only one parent but it may import any number of other
12681 A project is not allowed to import directly or indirectly at the same time a
12682 child project and any of its ancestors.
12684 @c *******************************
12685 @c * Project Hierarchy Extension *
12686 @c *******************************
12688 @node Project Hierarchy Extension
12689 @section Project Hierarchy Extension
12692 When extending a large system spanning multiple projects, it is often
12693 inconvenient to extend every project in the hierarchy that is impacted by a
12694 small change introduced. In such cases, it is possible to create a virtual
12695 extension of entire hierarchy using @code{extends all} relationship.
12697 When the project is extended using @code{extends all} inheritance, all projects
12698 that are imported by it, both directly and indirectly, are considered virtually
12699 extended. That is, the Project Manager creates "virtual projects"
12700 that extend every project in the hierarchy; all these virtual projects have
12701 no sources of their own and have as object directory the object directory of
12702 the root of "extending all" project.
12704 It is possible to explicitly extend one or more projects in the hierarchy
12705 in order to modify the sources. These extending projects must be imported by
12706 the "extending all" project, which will replace the corresponding virtual
12707 projects with the explicit ones.
12709 When building such a project hierarchy extension, the Project Manager will
12710 ensure that both modified sources and sources in virtual extending projects
12711 that depend on them, are recompiled.
12713 By means of example, consider the following hierarchy of projects.
12717 project A, containing package P1
12719 project B importing A and containing package P2 which depends on P1
12721 project C importing B and containing package P3 which depends on P2
12725 We want to modify packages P1 and P3.
12727 This project hierarchy will need to be extended as follows:
12731 Create project A1 that extends A, placing modified P1 there:
12733 @smallexample @c 0projectfile
12734 project A1 extends "(...)/A" is
12739 Create project C1 that "extends all" C and imports A1, placing modified
12742 @smallexample @c 0projectfile
12744 project C1 extends all "(...)/C" is
12749 When you build project C1, your entire modified project space will be
12750 recompiled, including the virtual project B1 that has been impacted by the
12751 "extending all" inheritance of project C.
12753 Note that if a Library Project in the hierarchy is virtually extended,
12754 the virtual project that extends the Library Project is not a Library Project.
12756 @c ****************************************
12757 @c * External References in Project Files *
12758 @c ****************************************
12760 @node External References in Project Files
12761 @section External References in Project Files
12764 A project file may contain references to external variables; such references
12765 are called @emph{external references}.
12767 An external variable is either defined as part of the environment (an
12768 environment variable in Unix, for example) or else specified on the command
12769 line via the @option{^-X^/EXTERNAL_REFERENCE=^@emph{vbl}=@emph{value}} switch.
12770 If both, then the command line value is used.
12772 The value of an external reference is obtained by means of the built-in
12773 function @code{external}, which returns a string value.
12774 This function has two forms:
12776 @item @code{external (external_variable_name)}
12777 @item @code{external (external_variable_name, default_value)}
12781 Each parameter must be a string literal. For example:
12783 @smallexample @c projectfile
12785 external ("OS", "GNU/Linux")
12789 In the form with one parameter, the function returns the value of
12790 the external variable given as parameter. If this name is not present in the
12791 environment, the function returns an empty string.
12793 In the form with two string parameters, the second argument is
12794 the value returned when the variable given as the first argument is not
12795 present in the environment. In the example above, if @code{"OS"} is not
12796 the name of ^an environment variable^a logical name^ and is not passed on
12797 the command line, then the returned value is @code{"GNU/Linux"}.
12799 An external reference may be part of a string expression or of a string
12800 list expression, and can therefore appear in a variable declaration or
12801 an attribute declaration.
12803 @smallexample @c projectfile
12805 type Mode_Type is ("Debug", "Release");
12806 Mode : Mode_Type := external ("MODE");
12813 @c *****************************
12814 @c * Packages in Project Files *
12815 @c *****************************
12817 @node Packages in Project Files
12818 @section Packages in Project Files
12821 A @emph{package} defines the settings for project-aware tools within a
12823 For each such tool one can declare a package; the names for these
12824 packages are preset (@pxref{Packages}).
12825 A package may contain variable declarations, attribute declarations, and case
12828 @smallexample @c projectfile
12831 package Builder is -- used by gnatmake
12832 for ^Default_Switches^Default_Switches^ ("Ada")
12841 The syntax of package declarations mimics that of package in Ada.
12843 Most of the packages have an attribute
12844 @code{^Default_Switches^Default_Switches^}.
12845 This attribute is an associative array, and its value is a string list.
12846 The index of the associative array is the name of a programming language (case
12847 insensitive). This attribute indicates the ^switch^switch^
12848 or ^switches^switches^ to be used
12849 with the corresponding tool.
12851 Some packages also have another attribute, @code{^Switches^Switches^},
12852 an associative array whose value is a string list.
12853 The index is the name of a source file.
12854 This attribute indicates the ^switch^switch^
12855 or ^switches^switches^ to be used by the corresponding
12856 tool when dealing with this specific file.
12858 Further information on these ^switch^switch^-related attributes is found in
12859 @ref{^Switches^Switches^ and Project Files}.
12861 A package may be declared as a @emph{renaming} of another package; e.g., from
12862 the project file for an imported project.
12864 @smallexample @c projectfile
12866 with "/global/apex.gpr";
12868 package Naming renames Apex.Naming;
12875 Packages that are renamed in other project files often come from project files
12876 that have no sources: they are just used as templates. Any modification in the
12877 template will be reflected automatically in all the project files that rename
12878 a package from the template.
12880 In addition to the tool-oriented packages, you can also declare a package
12881 named @code{Naming} to establish specialized source file naming conventions
12882 (@pxref{Naming Schemes}).
12884 @c ************************************
12885 @c * Variables from Imported Projects *
12886 @c ************************************
12888 @node Variables from Imported Projects
12889 @section Variables from Imported Projects
12892 An attribute or variable defined in an imported or parent project can
12893 be used in expressions in the importing / extending project.
12894 Such an attribute or variable is denoted by an expanded name whose prefix
12895 is either the name of the project or the expanded name of a package within
12898 @smallexample @c projectfile
12901 project Main extends "base" is
12902 Var1 := Imported.Var;
12903 Var2 := Base.Var & ".new";
12908 for ^Default_Switches^Default_Switches^ ("Ada")
12909 use Imported.Builder'Ada_^Switches^Switches^ &
12910 "^-gnatg^-gnatg^" &
12916 package Compiler is
12917 for ^Default_Switches^Default_Switches^ ("Ada")
12918 use Base.Compiler'Ada_^Switches^Switches^;
12929 The value of @code{Var1} is a copy of the variable @code{Var} defined
12930 in the project file @file{"imported.gpr"}
12932 the value of @code{Var2} is a copy of the value of variable @code{Var}
12933 defined in the project file @file{base.gpr}, concatenated with @code{".new"}
12935 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
12936 @code{Builder} is a string list that includes in its value a copy of the value
12937 of @code{Ada_^Switches^Switches^} defined in the @code{Builder} package
12938 in project file @file{imported.gpr} plus two new elements:
12939 @option{"^-gnatg^-gnatg^"}
12940 and @option{"^-v^-v^"};
12942 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
12943 @code{Compiler} is a copy of the variable @code{Ada_^Switches^Switches^}
12944 defined in the @code{Compiler} package in project file @file{base.gpr},
12945 the project being extended.
12948 @c ******************
12949 @c * Naming Schemes *
12950 @c ******************
12952 @node Naming Schemes
12953 @section Naming Schemes
12956 Sometimes an Ada software system is ported from a foreign compilation
12957 environment to GNAT, and the file names do not use the default GNAT
12958 conventions. Instead of changing all the file names (which for a variety
12959 of reasons might not be possible), you can define the relevant file
12960 naming scheme in the @code{Naming} package in your project file.
12963 Note that the use of pragmas described in
12964 @ref{Alternative File Naming Schemes} by mean of a configuration
12965 pragmas file is not supported when using project files. You must use
12966 the features described in this paragraph. You can however use specify
12967 other configuration pragmas (@pxref{Specifying Configuration Pragmas}).
12970 For example, the following
12971 package models the Apex file naming rules:
12973 @smallexample @c projectfile
12976 for Casing use "lowercase";
12977 for Dot_Replacement use ".";
12978 for Spec_Suffix ("Ada") use ".1.ada";
12979 for Body_Suffix ("Ada") use ".2.ada";
12986 For example, the following package models the HP Ada file naming rules:
12988 @smallexample @c projectfile
12991 for Casing use "lowercase";
12992 for Dot_Replacement use "__";
12993 for Spec_Suffix ("Ada") use "_.^ada^ada^";
12994 for Body_Suffix ("Ada") use ".^ada^ada^";
13000 (Note that @code{Casing} is @code{"lowercase"} because GNAT gets the file
13001 names in lower case)
13005 You can define the following attributes in package @code{Naming}:
13010 This must be a string with one of the three values @code{"lowercase"},
13011 @code{"uppercase"} or @code{"mixedcase"}; these strings are case insensitive.
13014 If @var{Casing} is not specified, then the default is @code{"lowercase"}.
13016 @item @var{Dot_Replacement}
13017 This must be a string whose value satisfies the following conditions:
13020 @item It must not be empty
13021 @item It cannot start or end with an alphanumeric character
13022 @item It cannot be a single underscore
13023 @item It cannot start with an underscore followed by an alphanumeric
13024 @item It cannot contain a dot @code{'.'} except if the entire string
13029 If @code{Dot_Replacement} is not specified, then the default is @code{"-"}.
13031 @item @var{Spec_Suffix}
13032 This is an associative array (indexed by the programming language name, case
13033 insensitive) whose value is a string that must satisfy the following
13037 @item It must not be empty
13038 @item It must include at least one dot
13041 If @code{Spec_Suffix ("Ada")} is not specified, then the default is
13042 @code{"^.ads^.ADS^"}.
13044 @item @var{Body_Suffix}
13045 This is an associative array (indexed by the programming language name, case
13046 insensitive) whose value is a string that must satisfy the following
13050 @item It must not be empty
13051 @item It must include at least one dot
13052 @item It cannot end with the same string as @code{Spec_Suffix ("Ada")}
13055 If @code{Body_Suffix ("Ada")} is not specified, then the default is
13056 @code{"^.adb^.ADB^"}.
13058 @item @var{Separate_Suffix}
13059 This must be a string whose value satisfies the same conditions as
13060 @code{Body_Suffix}.
13063 If @code{Separate_Suffix ("Ada")} is not specified, then it defaults to same
13064 value as @code{Body_Suffix ("Ada")}.
13068 You can use the associative array attribute @code{Spec} to define
13069 the source file name for an individual Ada compilation unit's spec. The array
13070 index must be a string literal that identifies the Ada unit (case insensitive).
13071 The value of this attribute must be a string that identifies the file that
13072 contains this unit's spec (case sensitive or insensitive depending on the
13075 @smallexample @c projectfile
13076 for Spec ("MyPack.MyChild") use "mypack.mychild.spec";
13081 You can use the associative array attribute @code{Body} to
13082 define the source file name for an individual Ada compilation unit's body
13083 (possibly a subunit). The array index must be a string literal that identifies
13084 the Ada unit (case insensitive). The value of this attribute must be a string
13085 that identifies the file that contains this unit's body or subunit (case
13086 sensitive or insensitive depending on the operating system).
13088 @smallexample @c projectfile
13089 for Body ("MyPack.MyChild") use "mypack.mychild.body";
13093 @c ********************
13094 @c * Library Projects *
13095 @c ********************
13097 @node Library Projects
13098 @section Library Projects
13101 @emph{Library projects} are projects whose object code is placed in a library.
13102 (Note that this facility is not yet supported on all platforms)
13104 To create a library project, you need to define in its project file
13105 two project-level attributes: @code{Library_Name} and @code{Library_Dir}.
13106 Additionally, you may define other library-related attributes such as
13107 @code{Library_Kind}, @code{Library_Version}, @code{Library_Interface},
13108 @code{Library_Auto_Init}, @code{Library_Options} and @code{Library_GCC}.
13110 The @code{Library_Name} attribute has a string value. There is no restriction
13111 on the name of a library. It is the responsibility of the developer to
13112 choose a name that will be accepted by the platform. It is recommended to
13113 choose names that could be Ada identifiers; such names are almost guaranteed
13114 to be acceptable on all platforms.
13116 The @code{Library_Dir} attribute has a string value that designates the path
13117 (absolute or relative) of the directory where the library will reside.
13118 It must designate an existing directory, and this directory must be writable,
13119 different from the project's object directory and from any source directory
13120 in the project tree.
13122 If both @code{Library_Name} and @code{Library_Dir} are specified and
13123 are legal, then the project file defines a library project. The optional
13124 library-related attributes are checked only for such project files.
13126 The @code{Library_Kind} attribute has a string value that must be one of the
13127 following (case insensitive): @code{"static"}, @code{"dynamic"} or
13128 @code{"relocatable"} (which is a synonym for @code{"dynamic"}). If this
13129 attribute is not specified, the library is a static library, that is
13130 an archive of object files that can be potentially linked into a
13131 static executable. Otherwise, the library may be dynamic or
13132 relocatable, that is a library that is loaded only at the start of execution.
13134 If you need to build both a static and a dynamic library, you should use two
13135 different object directories, since in some cases some extra code needs to
13136 be generated for the latter. For such cases, it is recommended to either use
13137 two different project files, or a single one which uses external variables
13138 to indicate what kind of library should be build.
13140 The @code{Library_ALI_Dir} attribute may be specified to indicate the
13141 directory where the ALI files of the library will be copied. When it is
13142 not specified, the ALI files are copied to the directory specified in
13143 attribute @code{Library_Dir}. The directory specified by @code{Library_ALI_Dir}
13144 must be writable and different from the project's object directory and from
13145 any source directory in the project tree.
13147 The @code{Library_Version} attribute has a string value whose interpretation
13148 is platform dependent. It has no effect on VMS and Windows. On Unix, it is
13149 used only for dynamic/relocatable libraries as the internal name of the
13150 library (the @code{"soname"}). If the library file name (built from the
13151 @code{Library_Name}) is different from the @code{Library_Version}, then the
13152 library file will be a symbolic link to the actual file whose name will be
13153 @code{Library_Version}.
13157 @smallexample @c projectfile
13163 for Library_Dir use "lib_dir";
13164 for Library_Name use "dummy";
13165 for Library_Kind use "relocatable";
13166 for Library_Version use "libdummy.so." & Version;
13173 Directory @file{lib_dir} will contain the internal library file whose name
13174 will be @file{libdummy.so.1}, and @file{libdummy.so} will be a symbolic link to
13175 @file{libdummy.so.1}.
13177 When @command{gnatmake} detects that a project file
13178 is a library project file, it will check all immediate sources of the project
13179 and rebuild the library if any of the sources have been recompiled.
13181 Standard project files can import library project files. In such cases,
13182 the libraries will only be rebuilt if some of its sources are recompiled
13183 because they are in the closure of some other source in an importing project.
13184 Sources of the library project files that are not in such a closure will
13185 not be checked, unless the full library is checked, because one of its sources
13186 needs to be recompiled.
13188 For instance, assume the project file @code{A} imports the library project file
13189 @code{L}. The immediate sources of A are @file{a1.adb}, @file{a2.ads} and
13190 @file{a2.adb}. The immediate sources of L are @file{l1.ads}, @file{l1.adb},
13191 @file{l2.ads}, @file{l2.adb}.
13193 If @file{l1.adb} has been modified, then the library associated with @code{L}
13194 will be rebuilt when compiling all the immediate sources of @code{A} only
13195 if @file{a1.ads}, @file{a2.ads} or @file{a2.adb} includes a statement
13198 To be sure that all the sources in the library associated with @code{L} are
13199 up to date, and that all the sources of project @code{A} are also up to date,
13200 the following two commands needs to be used:
13207 When a library is built or rebuilt, an attempt is made first to delete all
13208 files in the library directory.
13209 All @file{ALI} files will also be copied from the object directory to the
13210 library directory. To build executables, @command{gnatmake} will use the
13211 library rather than the individual object files.
13214 It is also possible to create library project files for third-party libraries
13215 that are precompiled and cannot be compiled locally thanks to the
13216 @code{externally_built} attribute. (See @ref{Installing a library}).
13219 @c *******************************
13220 @c * Stand-alone Library Projects *
13221 @c *******************************
13223 @node Stand-alone Library Projects
13224 @section Stand-alone Library Projects
13227 A Stand-alone Library is a library that contains the necessary code to
13228 elaborate the Ada units that are included in the library. A Stand-alone
13229 Library is suitable to be used in an executable when the main is not
13230 in Ada. However, Stand-alone Libraries may also be used with an Ada main
13233 A Stand-alone Library Project is a Library Project where the library is
13234 a Stand-alone Library.
13236 To be a Stand-alone Library Project, in addition to the two attributes
13237 that make a project a Library Project (@code{Library_Name} and
13238 @code{Library_Dir}, see @ref{Library Projects}), the attribute
13239 @code{Library_Interface} must be defined.
13241 @smallexample @c projectfile
13243 for Library_Dir use "lib_dir";
13244 for Library_Name use "dummy";
13245 for Library_Interface use ("int1", "int1.child");
13249 Attribute @code{Library_Interface} has a non empty string list value,
13250 each string in the list designating a unit contained in an immediate source
13251 of the project file.
13253 When a Stand-alone Library is built, first the binder is invoked to build
13254 a package whose name depends on the library name
13255 (^b~dummy.ads/b^B$DUMMY.ADS/B^ in the example above).
13256 This binder-generated package includes initialization and
13257 finalization procedures whose
13258 names depend on the library name (dummyinit and dummyfinal in the example
13259 above). The object corresponding to this package is included in the library.
13261 A dynamic or relocatable Stand-alone Library is automatically initialized
13262 if automatic initialization of Stand-alone Libraries is supported on the
13263 platform and if attribute @code{Library_Auto_Init} is not specified or
13264 is specified with the value "true". A static Stand-alone Library is never
13265 automatically initialized.
13267 Single string attribute @code{Library_Auto_Init} may be specified with only
13268 two possible values: "false" or "true" (case-insensitive). Specifying
13269 "false" for attribute @code{Library_Auto_Init} will prevent automatic
13270 initialization of dynamic or relocatable libraries.
13272 When a non automatically initialized Stand-alone Library is used
13273 in an executable, its initialization procedure must be called before
13274 any service of the library is used.
13275 When the main subprogram is in Ada, it may mean that the initialization
13276 procedure has to be called during elaboration of another package.
13278 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
13279 (those that are listed in attribute @code{Library_Interface}) are copied to
13280 the Library Directory. As a consequence, only the Interface Units may be
13281 imported from Ada units outside of the library. If other units are imported,
13282 the binding phase will fail.
13284 When a Stand-Alone Library is bound, the switches that are specified in
13285 the attribute @code{Default_Switches ("Ada")} in package @code{Binder} are
13286 used in the call to @command{gnatbind}.
13288 The string list attribute @code{Library_Options} may be used to specified
13289 additional switches to the call to @command{gcc} to link the library.
13291 The attribute @code{Library_Src_Dir}, may be specified for a
13292 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
13293 single string value. Its value must be the path (absolute or relative to the
13294 project directory) of an existing directory. This directory cannot be the
13295 object directory or one of the source directories, but it can be the same as
13296 the library directory. The sources of the Interface
13297 Units of the library, necessary to an Ada client of the library, will be
13298 copied to the designated directory, called Interface Copy directory.
13299 These sources includes the specs of the Interface Units, but they may also
13300 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
13301 are used, or when there is a generic units in the spec. Before the sources
13302 are copied to the Interface Copy directory, an attempt is made to delete all
13303 files in the Interface Copy directory.
13305 @c *************************************
13306 @c * Switches Related to Project Files *
13307 @c *************************************
13308 @node Switches Related to Project Files
13309 @section Switches Related to Project Files
13312 The following switches are used by GNAT tools that support project files:
13316 @item ^-P^/PROJECT_FILE=^@var{project}
13317 @cindex @option{^-P^/PROJECT_FILE^} (any tool supporting project files)
13318 Indicates the name of a project file. This project file will be parsed with
13319 the verbosity indicated by @option{^-vP^MESSAGE_PROJECT_FILES=^@emph{x}},
13320 if any, and using the external references indicated
13321 by @option{^-X^/EXTERNAL_REFERENCE^} switches, if any.
13323 There may zero, one or more spaces between @option{-P} and @var{project}.
13327 There must be only one @option{^-P^/PROJECT_FILE^} switch on the command line.
13330 Since the Project Manager parses the project file only after all the switches
13331 on the command line are checked, the order of the switches
13332 @option{^-P^/PROJECT_FILE^},
13333 @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}}
13334 or @option{^-X^/EXTERNAL_REFERENCE^} is not significant.
13336 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
13337 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (any tool supporting project files)
13338 Indicates that external variable @var{name} has the value @var{value}.
13339 The Project Manager will use this value for occurrences of
13340 @code{external(name)} when parsing the project file.
13344 If @var{name} or @var{value} includes a space, then @var{name=value} should be
13345 put between quotes.
13353 Several @option{^-X^/EXTERNAL_REFERENCE^} switches can be used simultaneously.
13354 If several @option{^-X^/EXTERNAL_REFERENCE^} switches specify the same
13355 @var{name}, only the last one is used.
13358 An external variable specified with a @option{^-X^/EXTERNAL_REFERENCE^} switch
13359 takes precedence over the value of the same name in the environment.
13361 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
13362 @cindex @code{^-vP^/MESSAGES_PROJECT_FILE^} (any tool supporting project files)
13363 @c Previous line uses code vs option command, to stay less than 80 chars
13364 Indicates the verbosity of the parsing of GNAT project files.
13367 @option{-vP0} means Default;
13368 @option{-vP1} means Medium;
13369 @option{-vP2} means High.
13373 There are three possible options for this qualifier: DEFAULT, MEDIUM and
13378 The default is ^Default^DEFAULT^: no output for syntactically correct
13381 If several @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}} switches are present,
13382 only the last one is used.
13386 @c **********************************
13387 @c * Tools Supporting Project Files *
13388 @c **********************************
13390 @node Tools Supporting Project Files
13391 @section Tools Supporting Project Files
13394 * gnatmake and Project Files::
13395 * The GNAT Driver and Project Files::
13398 @node gnatmake and Project Files
13399 @subsection gnatmake and Project Files
13402 This section covers several topics related to @command{gnatmake} and
13403 project files: defining ^switches^switches^ for @command{gnatmake}
13404 and for the tools that it invokes; specifying configuration pragmas;
13405 the use of the @code{Main} attribute; building and rebuilding library project
13409 * ^Switches^Switches^ and Project Files::
13410 * Specifying Configuration Pragmas::
13411 * Project Files and Main Subprograms::
13412 * Library Project Files::
13415 @node ^Switches^Switches^ and Project Files
13416 @subsubsection ^Switches^Switches^ and Project Files
13419 It is not currently possible to specify VMS style qualifiers in the project
13420 files; only Unix style ^switches^switches^ may be specified.
13424 For each of the packages @code{Builder}, @code{Compiler}, @code{Binder}, and
13425 @code{Linker}, you can specify a @code{^Default_Switches^Default_Switches^}
13426 attribute, a @code{^Switches^Switches^} attribute, or both;
13427 as their names imply, these ^switch^switch^-related
13428 attributes affect the ^switches^switches^ that are used for each of these GNAT
13430 @command{gnatmake} is invoked. As will be explained below, these
13431 component-specific ^switches^switches^ precede
13432 the ^switches^switches^ provided on the @command{gnatmake} command line.
13434 The @code{^Default_Switches^Default_Switches^} attribute is an associative
13435 array indexed by language name (case insensitive) whose value is a string list.
13438 @smallexample @c projectfile
13440 package Compiler is
13441 for ^Default_Switches^Default_Switches^ ("Ada")
13442 use ("^-gnaty^-gnaty^",
13449 The @code{^Switches^Switches^} attribute is also an associative array,
13450 indexed by a file name (which may or may not be case sensitive, depending
13451 on the operating system) whose value is a string list. For example:
13453 @smallexample @c projectfile
13456 for ^Switches^Switches^ ("main1.adb")
13458 for ^Switches^Switches^ ("main2.adb")
13465 For the @code{Builder} package, the file names must designate source files
13466 for main subprograms. For the @code{Binder} and @code{Linker} packages, the
13467 file names must designate @file{ALI} or source files for main subprograms.
13468 In each case just the file name without an explicit extension is acceptable.
13470 For each tool used in a program build (@command{gnatmake}, the compiler, the
13471 binder, and the linker), the corresponding package @dfn{contributes} a set of
13472 ^switches^switches^ for each file on which the tool is invoked, based on the
13473 ^switch^switch^-related attributes defined in the package.
13474 In particular, the ^switches^switches^
13475 that each of these packages contributes for a given file @var{f} comprise:
13479 the value of attribute @code{^Switches^Switches^ (@var{f})},
13480 if it is specified in the package for the given file,
13482 otherwise, the value of @code{^Default_Switches^Default_Switches^ ("Ada")},
13483 if it is specified in the package.
13487 If neither of these attributes is defined in the package, then the package does
13488 not contribute any ^switches^switches^ for the given file.
13490 When @command{gnatmake} is invoked on a file, the ^switches^switches^ comprise
13491 two sets, in the following order: those contributed for the file
13492 by the @code{Builder} package;
13493 and the switches passed on the command line.
13495 When @command{gnatmake} invokes a tool (compiler, binder, linker) on a file,
13496 the ^switches^switches^ passed to the tool comprise three sets,
13497 in the following order:
13501 the applicable ^switches^switches^ contributed for the file
13502 by the @code{Builder} package in the project file supplied on the command line;
13505 those contributed for the file by the package (in the relevant project file --
13506 see below) corresponding to the tool; and
13509 the applicable switches passed on the command line.
13513 The term @emph{applicable ^switches^switches^} reflects the fact that
13514 @command{gnatmake} ^switches^switches^ may or may not be passed to individual
13515 tools, depending on the individual ^switch^switch^.
13517 @command{gnatmake} may invoke the compiler on source files from different
13518 projects. The Project Manager will use the appropriate project file to
13519 determine the @code{Compiler} package for each source file being compiled.
13520 Likewise for the @code{Binder} and @code{Linker} packages.
13522 As an example, consider the following package in a project file:
13524 @smallexample @c projectfile
13527 package Compiler is
13528 for ^Default_Switches^Default_Switches^ ("Ada")
13530 for ^Switches^Switches^ ("a.adb")
13532 for ^Switches^Switches^ ("b.adb")
13534 "^-gnaty^-gnaty^");
13541 If @command{gnatmake} is invoked with this project file, and it needs to
13542 compile, say, the files @file{a.adb}, @file{b.adb}, and @file{c.adb}, then
13543 @file{a.adb} will be compiled with the ^switch^switch^
13544 @option{^-O1^-O1^},
13545 @file{b.adb} with ^switches^switches^
13547 and @option{^-gnaty^-gnaty^},
13548 and @file{c.adb} with @option{^-g^-g^}.
13550 The following example illustrates the ordering of the ^switches^switches^
13551 contributed by different packages:
13553 @smallexample @c projectfile
13557 for ^Switches^Switches^ ("main.adb")
13565 package Compiler is
13566 for ^Switches^Switches^ ("main.adb")
13574 If you issue the command:
13577 gnatmake ^-Pproj2^/PROJECT_FILE=PROJ2^ -O0 main
13581 then the compiler will be invoked on @file{main.adb} with the following
13582 sequence of ^switches^switches^
13585 ^-g -O1 -O2 -O0^-g -O1 -O2 -O0^
13588 with the last @option{^-O^-O^}
13589 ^switch^switch^ having precedence over the earlier ones;
13590 several other ^switches^switches^
13591 (such as @option{^-c^-c^}) are added implicitly.
13593 The ^switches^switches^
13595 and @option{^-O1^-O1^} are contributed by package
13596 @code{Builder}, @option{^-O2^-O2^} is contributed
13597 by the package @code{Compiler}
13598 and @option{^-O0^-O0^} comes from the command line.
13600 The @option{^-g^-g^}
13601 ^switch^switch^ will also be passed in the invocation of
13602 @command{Gnatlink.}
13604 A final example illustrates switch contributions from packages in different
13607 @smallexample @c projectfile
13610 for Source_Files use ("pack.ads", "pack.adb");
13611 package Compiler is
13612 for ^Default_Switches^Default_Switches^ ("Ada")
13613 use ("^-gnata^-gnata^");
13621 for Source_Files use ("foo_main.adb", "bar_main.adb");
13623 for ^Switches^Switches^ ("foo_main.adb")
13631 -- Ada source file:
13633 procedure Foo_Main is
13641 gnatmake ^-PProj4^/PROJECT_FILE=PROJ4^ foo_main.adb -cargs -gnato
13645 then the ^switches^switches^ passed to the compiler for @file{foo_main.adb} are
13646 @option{^-g^-g^} (contributed by the package @code{Proj4.Builder}) and
13647 @option{^-gnato^-gnato^} (passed on the command line).
13648 When the imported package @code{Pack} is compiled, the ^switches^switches^ used
13649 are @option{^-g^-g^} from @code{Proj4.Builder},
13650 @option{^-gnata^-gnata^} (contributed from package @code{Proj3.Compiler},
13651 and @option{^-gnato^-gnato^} from the command line.
13654 When using @command{gnatmake} with project files, some ^switches^switches^ or
13655 arguments may be expressed as relative paths. As the working directory where
13656 compilation occurs may change, these relative paths are converted to absolute
13657 paths. For the ^switches^switches^ found in a project file, the relative paths
13658 are relative to the project file directory, for the switches on the command
13659 line, they are relative to the directory where @command{gnatmake} is invoked.
13660 The ^switches^switches^ for which this occurs are:
13666 ^-aI^-aI^, as well as all arguments that are not switches (arguments to
13668 ^-o^-o^, object files specified in package @code{Linker} or after
13669 -largs on the command line). The exception to this rule is the ^switch^switch^
13670 ^--RTS=^--RTS=^ for which a relative path argument is never converted.
13672 @node Specifying Configuration Pragmas
13673 @subsubsection Specifying Configuration Pragmas
13675 When using @command{gnatmake} with project files, if there exists a file
13676 @file{gnat.adc} that contains configuration pragmas, this file will be
13679 Configuration pragmas can be defined by means of the following attributes in
13680 project files: @code{Global_Configuration_Pragmas} in package @code{Builder}
13681 and @code{Local_Configuration_Pragmas} in package @code{Compiler}.
13683 Both these attributes are single string attributes. Their values is the path
13684 name of a file containing configuration pragmas. If a path name is relative,
13685 then it is relative to the project directory of the project file where the
13686 attribute is defined.
13688 When compiling a source, the configuration pragmas used are, in order,
13689 those listed in the file designated by attribute
13690 @code{Global_Configuration_Pragmas} in package @code{Builder} of the main
13691 project file, if it is specified, and those listed in the file designated by
13692 attribute @code{Local_Configuration_Pragmas} in package @code{Compiler} of
13693 the project file of the source, if it exists.
13695 @node Project Files and Main Subprograms
13696 @subsubsection Project Files and Main Subprograms
13699 When using a project file, you can invoke @command{gnatmake}
13700 with one or several main subprograms, by specifying their source files on the
13704 gnatmake ^-P^/PROJECT_FILE=^prj main1 main2 main3
13708 Each of these needs to be a source file of the same project, except
13709 when the switch ^-u^/UNIQUE^ is used.
13712 When ^-u^/UNIQUE^ is not used, all the mains need to be sources of the
13713 same project, one of the project in the tree rooted at the project specified
13714 on the command line. The package @code{Builder} of this common project, the
13715 "main project" is the one that is considered by @command{gnatmake}.
13718 When ^-u^/UNIQUE^ is used, the specified source files may be in projects
13719 imported directly or indirectly by the project specified on the command line.
13720 Note that if such a source file is not part of the project specified on the
13721 command line, the ^switches^switches^ found in package @code{Builder} of the
13722 project specified on the command line, if any, that are transmitted
13723 to the compiler will still be used, not those found in the project file of
13727 When using a project file, you can also invoke @command{gnatmake} without
13728 explicitly specifying any main, and the effect depends on whether you have
13729 defined the @code{Main} attribute. This attribute has a string list value,
13730 where each element in the list is the name of a source file (the file
13731 extension is optional) that contains a unit that can be a main subprogram.
13733 If the @code{Main} attribute is defined in a project file as a non-empty
13734 string list and the switch @option{^-u^/UNIQUE^} is not used on the command
13735 line, then invoking @command{gnatmake} with this project file but without any
13736 main on the command line is equivalent to invoking @command{gnatmake} with all
13737 the file names in the @code{Main} attribute on the command line.
13740 @smallexample @c projectfile
13743 for Main use ("main1", "main2", "main3");
13749 With this project file, @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^"}
13751 @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^ main1 main2 main3"}.
13753 When the project attribute @code{Main} is not specified, or is specified
13754 as an empty string list, or when the switch @option{-u} is used on the command
13755 line, then invoking @command{gnatmake} with no main on the command line will
13756 result in all immediate sources of the project file being checked, and
13757 potentially recompiled. Depending on the presence of the switch @option{-u},
13758 sources from other project files on which the immediate sources of the main
13759 project file depend are also checked and potentially recompiled. In other
13760 words, the @option{-u} switch is applied to all of the immediate sources of the
13763 When no main is specified on the command line and attribute @code{Main} exists
13764 and includes several mains, or when several mains are specified on the
13765 command line, the default ^switches^switches^ in package @code{Builder} will
13766 be used for all mains, even if there are specific ^switches^switches^
13767 specified for one or several mains.
13769 But the ^switches^switches^ from package @code{Binder} or @code{Linker} will be
13770 the specific ^switches^switches^ for each main, if they are specified.
13772 @node Library Project Files
13773 @subsubsection Library Project Files
13776 When @command{gnatmake} is invoked with a main project file that is a library
13777 project file, it is not allowed to specify one or more mains on the command
13781 When a library project file is specified, switches ^-b^/ACTION=BIND^ and
13782 ^-l^/ACTION=LINK^ have special meanings.
13785 @item ^-b^/ACTION=BIND^ is only allowed for stand-alone libraries. It indicates
13786 to @command{gnatmake} that @command{gnatbind} should be invoked for the
13789 @item ^-l^/ACTION=LINK^ may be used for all library projects. It indicates
13790 to @command{gnatmake} that the binder generated file should be compiled
13791 (in the case of a stand-alone library) and that the library should be built.
13795 @node The GNAT Driver and Project Files
13796 @subsection The GNAT Driver and Project Files
13799 A number of GNAT tools, other than @command{^gnatmake^gnatmake^}
13800 can benefit from project files:
13801 @command{^gnatbind^gnatbind^},
13802 @command{^gnatcheck^gnatcheck^}),
13803 @command{^gnatclean^gnatclean^}),
13804 @command{^gnatelim^gnatelim^},
13805 @command{^gnatfind^gnatfind^},
13806 @command{^gnatlink^gnatlink^},
13807 @command{^gnatls^gnatls^},
13808 @command{^gnatmetric^gnatmetric^},
13809 @command{^gnatpp^gnatpp^},
13810 @command{^gnatstub^gnatstub^},
13811 and @command{^gnatxref^gnatxref^}. However, none of these tools can be invoked
13812 directly with a project file switch (@option{^-P^/PROJECT_FILE=^}).
13813 They must be invoked through the @command{gnat} driver.
13815 The @command{gnat} driver is a wrapper that accepts a number of commands and
13816 calls the corresponding tool. It was designed initially for VMS platforms (to
13817 convert VMS qualifiers to Unix-style switches), but it is now available on all
13820 On non-VMS platforms, the @command{gnat} driver accepts the following commands
13821 (case insensitive):
13825 BIND to invoke @command{^gnatbind^gnatbind^}
13827 CHOP to invoke @command{^gnatchop^gnatchop^}
13829 CLEAN to invoke @command{^gnatclean^gnatclean^}
13831 COMP or COMPILE to invoke the compiler
13833 ELIM to invoke @command{^gnatelim^gnatelim^}
13835 FIND to invoke @command{^gnatfind^gnatfind^}
13837 KR or KRUNCH to invoke @command{^gnatkr^gnatkr^}
13839 LINK to invoke @command{^gnatlink^gnatlink^}
13841 LS or LIST to invoke @command{^gnatls^gnatls^}
13843 MAKE to invoke @command{^gnatmake^gnatmake^}
13845 NAME to invoke @command{^gnatname^gnatname^}
13847 PREP or PREPROCESS to invoke @command{^gnatprep^gnatprep^}
13849 PP or PRETTY to invoke @command{^gnatpp^gnatpp^}
13851 METRIC to invoke @command{^gnatmetric^gnatmetric^}
13853 STUB to invoke @command{^gnatstub^gnatstub^}
13855 XREF to invoke @command{^gnatxref^gnatxref^}
13859 (note that the compiler is invoked using the command
13860 @command{^gnatmake -f -u -c^gnatmake -f -u -c^}).
13863 On non VMS platforms, between @command{gnat} and the command, two
13864 special switches may be used:
13868 @command{-v} to display the invocation of the tool.
13870 @command{-dn} to prevent the @command{gnat} driver from removing
13871 the temporary files it has created. These temporary files are
13872 configuration files and temporary file list files.
13876 The command may be followed by switches and arguments for the invoked
13880 gnat bind -C main.ali
13886 Switches may also be put in text files, one switch per line, and the text
13887 files may be specified with their path name preceded by '@@'.
13890 gnat bind @@args.txt main.ali
13894 In addition, for commands BIND, COMP or COMPILE, FIND, ELIM, LS or LIST, LINK,
13895 METRIC, PP or PRETTY, STUB and XREF, the project file related switches
13896 (@option{^-P^/PROJECT_FILE^},
13897 @option{^-X^/EXTERNAL_REFERENCE^} and
13898 @option{^-vP^/MESSAGES_PROJECT_FILE=^x}) may be used in addition to
13899 the switches of the invoking tool.
13902 When GNAT PP or GNAT PRETTY is used with a project file, but with no source
13903 specified on the command line, it invokes @command{^gnatpp^gnatpp^} with all
13904 the immediate sources of the specified project file.
13907 When GNAT METRIC is used with a project file, but with no source
13908 specified on the command line, it invokes @command{^gnatmetric^gnatmetric^}
13909 with all the immediate sources of the specified project file and with
13910 @option{^-d^/DIRECTORY^} with the parameter pointing to the object directory
13914 In addition, when GNAT PP, GNAT PRETTY or GNAT METRIC is used with
13915 a project file, no source is specified on the command line and
13916 switch ^-U^/ALL_PROJECTS^ is specified on the command line, then
13917 the underlying tool (^gnatpp^gnatpp^ or
13918 ^gnatmetric^gnatmetric^) is invoked for all sources of all projects,
13919 not only for the immediate sources of the main project.
13921 (-U stands for Universal or Union of the project files of the project tree)
13925 For each of the following commands, there is optionally a corresponding
13926 package in the main project.
13930 package @code{Binder} for command BIND (invoking @code{^gnatbind^gnatbind^})
13933 package @code{Check} for command CHECK (invoking
13934 @code{^gnatcheck^gnatcheck^})
13937 package @code{Compiler} for command COMP or COMPILE (invoking the compiler)
13940 package @code{Cross_Reference} for command XREF (invoking
13941 @code{^gnatxref^gnatxref^})
13944 package @code{Eliminate} for command ELIM (invoking
13945 @code{^gnatelim^gnatelim^})
13948 package @code{Finder} for command FIND (invoking @code{^gnatfind^gnatfind^})
13951 package @code{Gnatls} for command LS or LIST (invoking @code{^gnatls^gnatls^})
13954 package @code{Gnatstub} for command STUB
13955 (invoking @code{^gnatstub^gnatstub^})
13958 package @code{Linker} for command LINK (invoking @code{^gnatlink^gnatlink^})
13961 package @code{Metrics} for command METRIC
13962 (invoking @code{^gnatmetric^gnatmetric^})
13965 package @code{Pretty_Printer} for command PP or PRETTY
13966 (invoking @code{^gnatpp^gnatpp^})
13971 Package @code{Gnatls} has a unique attribute @code{^Switches^Switches^},
13972 a simple variable with a string list value. It contains ^switches^switches^
13973 for the invocation of @code{^gnatls^gnatls^}.
13975 @smallexample @c projectfile
13979 for ^Switches^Switches^
13988 All other packages have two attribute @code{^Switches^Switches^} and
13989 @code{^Default_Switches^Default_Switches^}.
13992 @code{^Switches^Switches^} is an associative array attribute, indexed by the
13993 source file name, that has a string list value: the ^switches^switches^ to be
13994 used when the tool corresponding to the package is invoked for the specific
13998 @code{^Default_Switches^Default_Switches^} is an associative array attribute,
13999 indexed by the programming language that has a string list value.
14000 @code{^Default_Switches^Default_Switches^ ("Ada")} contains the
14001 ^switches^switches^ for the invocation of the tool corresponding
14002 to the package, except if a specific @code{^Switches^Switches^} attribute
14003 is specified for the source file.
14005 @smallexample @c projectfile
14009 for Source_Dirs use ("./**");
14012 for ^Switches^Switches^ use
14019 package Compiler is
14020 for ^Default_Switches^Default_Switches^ ("Ada")
14021 use ("^-gnatv^-gnatv^",
14022 "^-gnatwa^-gnatwa^");
14028 for ^Default_Switches^Default_Switches^ ("Ada")
14036 for ^Default_Switches^Default_Switches^ ("Ada")
14038 for ^Switches^Switches^ ("main.adb")
14047 for ^Default_Switches^Default_Switches^ ("Ada")
14054 package Cross_Reference is
14055 for ^Default_Switches^Default_Switches^ ("Ada")
14060 end Cross_Reference;
14066 With the above project file, commands such as
14069 ^gnat comp -Pproj main^GNAT COMP /PROJECT_FILE=PROJ MAIN^
14070 ^gnat ls -Pproj main^GNAT LIST /PROJECT_FILE=PROJ MAIN^
14071 ^gnat xref -Pproj main^GNAT XREF /PROJECT_FILE=PROJ MAIN^
14072 ^gnat bind -Pproj main.ali^GNAT BIND /PROJECT_FILE=PROJ MAIN.ALI^
14073 ^gnat link -Pproj main.ali^GNAT LINK /PROJECT_FILE=PROJ MAIN.ALI^
14077 will set up the environment properly and invoke the tool with the switches
14078 found in the package corresponding to the tool:
14079 @code{^Default_Switches^Default_Switches^ ("Ada")} for all tools,
14080 except @code{^Switches^Switches^ ("main.adb")}
14081 for @code{^gnatlink^gnatlink^}.
14082 It is also possible to invoke some of the tools,
14083 @code{^gnatcheck^gnatcheck^}),
14084 @code{^gnatmetric^gnatmetric^}),
14085 and @code{^gnatpp^gnatpp^})
14086 on a set of project units thanks to the combination of the switches
14087 @code{-P}, @code{-U} and possibly the main unit when one is interested
14088 in its closure. For instance,
14092 will compute the metrics for all the immediate units of project
14095 gnat metric -Pproj -U
14097 will compute the metrics for all the units of the closure of projects
14098 rooted at @code{proj}.
14100 gnat metric -Pproj -U main_unit
14102 will compute the metrics for the closure of units rooted at
14103 @code{main_unit}. This last possibility relies implicitly
14104 on @command{gnatbind}'s option @option{-R}.
14106 @c **********************
14107 @node An Extended Example
14108 @section An Extended Example
14111 Suppose that we have two programs, @var{prog1} and @var{prog2},
14112 whose sources are in corresponding directories. We would like
14113 to build them with a single @command{gnatmake} command, and we want to place
14114 their object files into @file{build} subdirectories of the source directories.
14115 Furthermore, we want to have to have two separate subdirectories
14116 in @file{build} -- @file{release} and @file{debug} -- which will contain
14117 the object files compiled with different set of compilation flags.
14119 In other words, we have the following structure:
14136 Here are the project files that we must place in a directory @file{main}
14137 to maintain this structure:
14141 @item We create a @code{Common} project with a package @code{Compiler} that
14142 specifies the compilation ^switches^switches^:
14147 @b{project} Common @b{is}
14149 @b{for} Source_Dirs @b{use} (); -- No source files
14153 @b{type} Build_Type @b{is} ("release", "debug");
14154 Build : Build_Type := External ("BUILD", "debug");
14157 @b{package} Compiler @b{is}
14158 @b{case} Build @b{is}
14159 @b{when} "release" =>
14160 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
14161 @b{use} ("^-O2^-O2^");
14162 @b{when} "debug" =>
14163 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
14164 @b{use} ("^-g^-g^");
14172 @item We create separate projects for the two programs:
14179 @b{project} Prog1 @b{is}
14181 @b{for} Source_Dirs @b{use} ("prog1");
14182 @b{for} Object_Dir @b{use} "prog1/build/" & Common.Build;
14184 @b{package} Compiler @b{renames} Common.Compiler;
14195 @b{project} Prog2 @b{is}
14197 @b{for} Source_Dirs @b{use} ("prog2");
14198 @b{for} Object_Dir @b{use} "prog2/build/" & Common.Build;
14200 @b{package} Compiler @b{renames} Common.Compiler;
14206 @item We create a wrapping project @code{Main}:
14215 @b{project} Main @b{is}
14217 @b{package} Compiler @b{renames} Common.Compiler;
14223 @item Finally we need to create a dummy procedure that @code{with}s (either
14224 explicitly or implicitly) all the sources of our two programs.
14229 Now we can build the programs using the command
14232 gnatmake ^-P^/PROJECT_FILE=^main dummy
14236 for the Debug mode, or
14240 gnatmake -Pmain -XBUILD=release
14246 GNAT MAKE /PROJECT_FILE=main /EXTERNAL_REFERENCE=BUILD=release
14251 for the Release mode.
14253 @c ********************************
14254 @c * Project File Complete Syntax *
14255 @c ********************************
14257 @node Project File Complete Syntax
14258 @section Project File Complete Syntax
14262 context_clause project_declaration
14268 @b{with} path_name @{ , path_name @} ;
14273 project_declaration ::=
14274 simple_project_declaration | project_extension
14276 simple_project_declaration ::=
14277 @b{project} <project_>simple_name @b{is}
14278 @{declarative_item@}
14279 @b{end} <project_>simple_name;
14281 project_extension ::=
14282 @b{project} <project_>simple_name @b{extends} path_name @b{is}
14283 @{declarative_item@}
14284 @b{end} <project_>simple_name;
14286 declarative_item ::=
14287 package_declaration |
14288 typed_string_declaration |
14289 other_declarative_item
14291 package_declaration ::=
14292 package_specification | package_renaming
14294 package_specification ::=
14295 @b{package} package_identifier @b{is}
14296 @{simple_declarative_item@}
14297 @b{end} package_identifier ;
14299 package_identifier ::=
14300 @code{Naming} | @code{Builder} | @code{Compiler} | @code{Binder} |
14301 @code{Linker} | @code{Finder} | @code{Cross_Reference} |
14302 @code{^gnatls^gnatls^} | @code{IDE} | @code{Pretty_Printer}
14304 package_renaming ::==
14305 @b{package} package_identifier @b{renames}
14306 <project_>simple_name.package_identifier ;
14308 typed_string_declaration ::=
14309 @b{type} <typed_string_>_simple_name @b{is}
14310 ( string_literal @{, string_literal@} );
14312 other_declarative_item ::=
14313 attribute_declaration |
14314 typed_variable_declaration |
14315 variable_declaration |
14318 attribute_declaration ::=
14319 full_associative_array_declaration |
14320 @b{for} attribute_designator @b{use} expression ;
14322 full_associative_array_declaration ::=
14323 @b{for} <associative_array_attribute_>simple_name @b{use}
14324 <project_>simple_name [ . <package_>simple_Name ] ' <attribute_>simple_name ;
14326 attribute_designator ::=
14327 <simple_attribute_>simple_name |
14328 <associative_array_attribute_>simple_name ( string_literal )
14330 typed_variable_declaration ::=
14331 <typed_variable_>simple_name : <typed_string_>name := string_expression ;
14333 variable_declaration ::=
14334 <variable_>simple_name := expression;
14344 attribute_reference
14350 ( <string_>expression @{ , <string_>expression @} )
14353 @b{external} ( string_literal [, string_literal] )
14355 attribute_reference ::=
14356 attribute_prefix ' <simple_attribute_>simple_name [ ( literal_string ) ]
14358 attribute_prefix ::=
14360 <project_>simple_name | package_identifier |
14361 <project_>simple_name . package_identifier
14363 case_construction ::=
14364 @b{case} <typed_variable_>name @b{is}
14369 @b{when} discrete_choice_list =>
14370 @{case_construction | attribute_declaration@}
14372 discrete_choice_list ::=
14373 string_literal @{| string_literal@} |
14377 simple_name @{. simple_name@}
14380 identifier (same as Ada)
14384 @node The Cross-Referencing Tools gnatxref and gnatfind
14385 @chapter The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
14390 The compiler generates cross-referencing information (unless
14391 you set the @samp{-gnatx} switch), which are saved in the @file{.ali} files.
14392 This information indicates where in the source each entity is declared and
14393 referenced. Note that entities in package Standard are not included, but
14394 entities in all other predefined units are included in the output.
14396 Before using any of these two tools, you need to compile successfully your
14397 application, so that GNAT gets a chance to generate the cross-referencing
14400 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
14401 information to provide the user with the capability to easily locate the
14402 declaration and references to an entity. These tools are quite similar,
14403 the difference being that @code{gnatfind} is intended for locating
14404 definitions and/or references to a specified entity or entities, whereas
14405 @code{gnatxref} is oriented to generating a full report of all
14408 To use these tools, you must not compile your application using the
14409 @option{-gnatx} switch on the @command{gnatmake} command line
14410 (@pxref{The GNAT Make Program gnatmake}). Otherwise, cross-referencing
14411 information will not be generated.
14413 Note: to invoke @code{gnatxref} or @code{gnatfind} with a project file,
14414 use the @code{gnat} driver (see @ref{The GNAT Driver and Project Files}).
14417 * gnatxref Switches::
14418 * gnatfind Switches::
14419 * Project Files for gnatxref and gnatfind::
14420 * Regular Expressions in gnatfind and gnatxref::
14421 * Examples of gnatxref Usage::
14422 * Examples of gnatfind Usage::
14425 @node gnatxref Switches
14426 @section @code{gnatxref} Switches
14429 The command invocation for @code{gnatxref} is:
14431 $ gnatxref [switches] sourcefile1 [sourcefile2 ...]
14438 @item sourcefile1, sourcefile2
14439 identifies the source files for which a report is to be generated. The
14440 ``with''ed units will be processed too. You must provide at least one file.
14442 These file names are considered to be regular expressions, so for instance
14443 specifying @file{source*.adb} is the same as giving every file in the current
14444 directory whose name starts with @file{source} and whose extension is
14447 You shouldn't specify any directory name, just base names. @command{gnatxref}
14448 and @command{gnatfind} will be able to locate these files by themselves using
14449 the source path. If you specify directories, no result is produced.
14454 The switches can be :
14457 @item ^-a^/ALL_FILES^
14458 @cindex @option{^-a^/ALL_FILES^} (@command{gnatxref})
14459 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
14460 the read-only files found in the library search path. Otherwise, these files
14461 will be ignored. This option can be used to protect Gnat sources or your own
14462 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
14463 much faster, and their output much smaller. Read-only here refers to access
14464 or permissions status in the file system for the current user.
14467 @cindex @option{-aIDIR} (@command{gnatxref})
14468 When looking for source files also look in directory DIR. The order in which
14469 source file search is undertaken is the same as for @command{gnatmake}.
14472 @cindex @option{-aODIR} (@command{gnatxref})
14473 When searching for library and object files, look in directory
14474 DIR. The order in which library files are searched is the same as for
14475 @command{gnatmake}.
14478 @cindex @option{-nostdinc} (@command{gnatxref})
14479 Do not look for sources in the system default directory.
14482 @cindex @option{-nostdlib} (@command{gnatxref})
14483 Do not look for library files in the system default directory.
14485 @item --RTS=@var{rts-path}
14486 @cindex @option{--RTS} (@command{gnatxref})
14487 Specifies the default location of the runtime library. Same meaning as the
14488 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
14490 @item ^-d^/DERIVED_TYPES^
14491 @cindex @option{^-d^/DERIVED_TYPES^} (@command{gnatxref})
14492 If this switch is set @code{gnatxref} will output the parent type
14493 reference for each matching derived types.
14495 @item ^-f^/FULL_PATHNAME^
14496 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatxref})
14497 If this switch is set, the output file names will be preceded by their
14498 directory (if the file was found in the search path). If this switch is
14499 not set, the directory will not be printed.
14501 @item ^-g^/IGNORE_LOCALS^
14502 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatxref})
14503 If this switch is set, information is output only for library-level
14504 entities, ignoring local entities. The use of this switch may accelerate
14505 @code{gnatfind} and @code{gnatxref}.
14508 @cindex @option{-IDIR} (@command{gnatxref})
14509 Equivalent to @samp{-aODIR -aIDIR}.
14512 @cindex @option{-pFILE} (@command{gnatxref})
14513 Specify a project file to use @xref{Project Files}.
14514 If you need to use the @file{.gpr}
14515 project files, you should use gnatxref through the GNAT driver
14516 (@command{gnat xref -Pproject}).
14518 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
14519 project file in the current directory.
14521 If a project file is either specified or found by the tools, then the content
14522 of the source directory and object directory lines are added as if they
14523 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^}
14524 and @samp{^-aO^OBJECT_SEARCH^}.
14526 Output only unused symbols. This may be really useful if you give your
14527 main compilation unit on the command line, as @code{gnatxref} will then
14528 display every unused entity and 'with'ed package.
14532 Instead of producing the default output, @code{gnatxref} will generate a
14533 @file{tags} file that can be used by vi. For examples how to use this
14534 feature, see @ref{Examples of gnatxref Usage}. The tags file is output
14535 to the standard output, thus you will have to redirect it to a file.
14541 All these switches may be in any order on the command line, and may even
14542 appear after the file names. They need not be separated by spaces, thus
14543 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
14544 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
14546 @node gnatfind Switches
14547 @section @code{gnatfind} Switches
14550 The command line for @code{gnatfind} is:
14553 $ gnatfind [switches] pattern[:sourcefile[:line[:column]]]
14562 An entity will be output only if it matches the regular expression found
14563 in @samp{pattern}, see @ref{Regular Expressions in gnatfind and gnatxref}.
14565 Omitting the pattern is equivalent to specifying @samp{*}, which
14566 will match any entity. Note that if you do not provide a pattern, you
14567 have to provide both a sourcefile and a line.
14569 Entity names are given in Latin-1, with uppercase/lowercase equivalence
14570 for matching purposes. At the current time there is no support for
14571 8-bit codes other than Latin-1, or for wide characters in identifiers.
14574 @code{gnatfind} will look for references, bodies or declarations
14575 of symbols referenced in @file{sourcefile}, at line @samp{line}
14576 and column @samp{column}. See @ref{Examples of gnatfind Usage}
14577 for syntax examples.
14580 is a decimal integer identifying the line number containing
14581 the reference to the entity (or entities) to be located.
14584 is a decimal integer identifying the exact location on the
14585 line of the first character of the identifier for the
14586 entity reference. Columns are numbered from 1.
14588 @item file1 file2 ...
14589 The search will be restricted to these source files. If none are given, then
14590 the search will be done for every library file in the search path.
14591 These file must appear only after the pattern or sourcefile.
14593 These file names are considered to be regular expressions, so for instance
14594 specifying 'source*.adb' is the same as giving every file in the current
14595 directory whose name starts with 'source' and whose extension is 'adb'.
14597 The location of the spec of the entity will always be displayed, even if it
14598 isn't in one of file1, file2,... The occurrences of the entity in the
14599 separate units of the ones given on the command line will also be displayed.
14601 Note that if you specify at least one file in this part, @code{gnatfind} may
14602 sometimes not be able to find the body of the subprograms...
14607 At least one of 'sourcefile' or 'pattern' has to be present on
14610 The following switches are available:
14614 @item ^-a^/ALL_FILES^
14615 @cindex @option{^-a^/ALL_FILES^} (@command{gnatfind})
14616 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
14617 the read-only files found in the library search path. Otherwise, these files
14618 will be ignored. This option can be used to protect Gnat sources or your own
14619 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
14620 much faster, and their output much smaller. Read-only here refers to access
14621 or permission status in the file system for the current user.
14624 @cindex @option{-aIDIR} (@command{gnatfind})
14625 When looking for source files also look in directory DIR. The order in which
14626 source file search is undertaken is the same as for @command{gnatmake}.
14629 @cindex @option{-aODIR} (@command{gnatfind})
14630 When searching for library and object files, look in directory
14631 DIR. The order in which library files are searched is the same as for
14632 @command{gnatmake}.
14635 @cindex @option{-nostdinc} (@command{gnatfind})
14636 Do not look for sources in the system default directory.
14639 @cindex @option{-nostdlib} (@command{gnatfind})
14640 Do not look for library files in the system default directory.
14642 @item --RTS=@var{rts-path}
14643 @cindex @option{--RTS} (@command{gnatfind})
14644 Specifies the default location of the runtime library. Same meaning as the
14645 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
14647 @item ^-d^/DERIVED_TYPE_INFORMATION^
14648 @cindex @option{^-d^/DERIVED_TYPE_INFORMATION^} (@code{gnatfind})
14649 If this switch is set, then @code{gnatfind} will output the parent type
14650 reference for each matching derived types.
14652 @item ^-e^/EXPRESSIONS^
14653 @cindex @option{^-e^/EXPRESSIONS^} (@command{gnatfind})
14654 By default, @code{gnatfind} accept the simple regular expression set for
14655 @samp{pattern}. If this switch is set, then the pattern will be
14656 considered as full Unix-style regular expression.
14658 @item ^-f^/FULL_PATHNAME^
14659 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatfind})
14660 If this switch is set, the output file names will be preceded by their
14661 directory (if the file was found in the search path). If this switch is
14662 not set, the directory will not be printed.
14664 @item ^-g^/IGNORE_LOCALS^
14665 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatfind})
14666 If this switch is set, information is output only for library-level
14667 entities, ignoring local entities. The use of this switch may accelerate
14668 @code{gnatfind} and @code{gnatxref}.
14671 @cindex @option{-IDIR} (@command{gnatfind})
14672 Equivalent to @samp{-aODIR -aIDIR}.
14675 @cindex @option{-pFILE} (@command{gnatfind})
14676 Specify a project file (@pxref{Project Files}) to use.
14677 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
14678 project file in the current directory.
14680 If a project file is either specified or found by the tools, then the content
14681 of the source directory and object directory lines are added as if they
14682 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^} and
14683 @samp{^-aO^/OBJECT_SEARCH^}.
14685 @item ^-r^/REFERENCES^
14686 @cindex @option{^-r^/REFERENCES^} (@command{gnatfind})
14687 By default, @code{gnatfind} will output only the information about the
14688 declaration, body or type completion of the entities. If this switch is
14689 set, the @code{gnatfind} will locate every reference to the entities in
14690 the files specified on the command line (or in every file in the search
14691 path if no file is given on the command line).
14693 @item ^-s^/PRINT_LINES^
14694 @cindex @option{^-s^/PRINT_LINES^} (@command{gnatfind})
14695 If this switch is set, then @code{gnatfind} will output the content
14696 of the Ada source file lines were the entity was found.
14698 @item ^-t^/TYPE_HIERARCHY^
14699 @cindex @option{^-t^/TYPE_HIERARCHY^} (@command{gnatfind})
14700 If this switch is set, then @code{gnatfind} will output the type hierarchy for
14701 the specified type. It act like -d option but recursively from parent
14702 type to parent type. When this switch is set it is not possible to
14703 specify more than one file.
14708 All these switches may be in any order on the command line, and may even
14709 appear after the file names. They need not be separated by spaces, thus
14710 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
14711 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
14713 As stated previously, gnatfind will search in every directory in the
14714 search path. You can force it to look only in the current directory if
14715 you specify @code{*} at the end of the command line.
14717 @node Project Files for gnatxref and gnatfind
14718 @section Project Files for @command{gnatxref} and @command{gnatfind}
14721 Project files allow a programmer to specify how to compile its
14722 application, where to find sources, etc. These files are used
14724 primarily by GPS, but they can also be used
14727 @code{gnatxref} and @code{gnatfind}.
14729 A project file name must end with @file{.gpr}. If a single one is
14730 present in the current directory, then @code{gnatxref} and @code{gnatfind} will
14731 extract the information from it. If multiple project files are found, none of
14732 them is read, and you have to use the @samp{-p} switch to specify the one
14735 The following lines can be included, even though most of them have default
14736 values which can be used in most cases.
14737 The lines can be entered in any order in the file.
14738 Except for @file{src_dir} and @file{obj_dir}, you can only have one instance of
14739 each line. If you have multiple instances, only the last one is taken into
14744 [default: @code{"^./^[]^"}]
14745 specifies a directory where to look for source files. Multiple @code{src_dir}
14746 lines can be specified and they will be searched in the order they
14750 [default: @code{"^./^[]^"}]
14751 specifies a directory where to look for object and library files. Multiple
14752 @code{obj_dir} lines can be specified, and they will be searched in the order
14755 @item comp_opt=SWITCHES
14756 [default: @code{""}]
14757 creates a variable which can be referred to subsequently by using
14758 the @code{$@{comp_opt@}} notation. This is intended to store the default
14759 switches given to @command{gnatmake} and @command{gcc}.
14761 @item bind_opt=SWITCHES
14762 [default: @code{""}]
14763 creates a variable which can be referred to subsequently by using
14764 the @samp{$@{bind_opt@}} notation. This is intended to store the default
14765 switches given to @command{gnatbind}.
14767 @item link_opt=SWITCHES
14768 [default: @code{""}]
14769 creates a variable which can be referred to subsequently by using
14770 the @samp{$@{link_opt@}} notation. This is intended to store the default
14771 switches given to @command{gnatlink}.
14773 @item main=EXECUTABLE
14774 [default: @code{""}]
14775 specifies the name of the executable for the application. This variable can
14776 be referred to in the following lines by using the @samp{$@{main@}} notation.
14779 @item comp_cmd=COMMAND
14780 [default: @code{"GNAT COMPILE /SEARCH=$@{src_dir@} /DEBUG /TRY_SEMANTICS"}]
14783 @item comp_cmd=COMMAND
14784 [default: @code{"gcc -c -I$@{src_dir@} -g -gnatq"}]
14786 specifies the command used to compile a single file in the application.
14789 @item make_cmd=COMMAND
14790 [default: @code{"GNAT MAKE $@{main@}
14791 /SOURCE_SEARCH=$@{src_dir@} /OBJECT_SEARCH=$@{obj_dir@}
14792 /DEBUG /TRY_SEMANTICS /COMPILER_QUALIFIERS $@{comp_opt@}
14793 /BINDER_QUALIFIERS $@{bind_opt@} /LINKER_QUALIFIERS $@{link_opt@}"}]
14796 @item make_cmd=COMMAND
14797 [default: @code{"gnatmake $@{main@} -aI$@{src_dir@}
14798 -aO$@{obj_dir@} -g -gnatq -cargs $@{comp_opt@}
14799 -bargs $@{bind_opt@} -largs $@{link_opt@}"}]
14801 specifies the command used to recompile the whole application.
14803 @item run_cmd=COMMAND
14804 [default: @code{"$@{main@}"}]
14805 specifies the command used to run the application.
14807 @item debug_cmd=COMMAND
14808 [default: @code{"gdb $@{main@}"}]
14809 specifies the command used to debug the application
14814 @command{gnatxref} and @command{gnatfind} only take into account the
14815 @code{src_dir} and @code{obj_dir} lines, and ignore the others.
14817 @node Regular Expressions in gnatfind and gnatxref
14818 @section Regular Expressions in @code{gnatfind} and @code{gnatxref}
14821 As specified in the section about @command{gnatfind}, the pattern can be a
14822 regular expression. Actually, there are to set of regular expressions
14823 which are recognized by the program :
14826 @item globbing patterns
14827 These are the most usual regular expression. They are the same that you
14828 generally used in a Unix shell command line, or in a DOS session.
14830 Here is a more formal grammar :
14837 term ::= elmt -- matches elmt
14838 term ::= elmt elmt -- concatenation (elmt then elmt)
14839 term ::= * -- any string of 0 or more characters
14840 term ::= ? -- matches any character
14841 term ::= [char @{char@}] -- matches any character listed
14842 term ::= [char - char] -- matches any character in range
14846 @item full regular expression
14847 The second set of regular expressions is much more powerful. This is the
14848 type of regular expressions recognized by utilities such a @file{grep}.
14850 The following is the form of a regular expression, expressed in Ada
14851 reference manual style BNF is as follows
14858 regexp ::= term @{| term@} -- alternation (term or term ...)
14860 term ::= item @{item@} -- concatenation (item then item)
14862 item ::= elmt -- match elmt
14863 item ::= elmt * -- zero or more elmt's
14864 item ::= elmt + -- one or more elmt's
14865 item ::= elmt ? -- matches elmt or nothing
14868 elmt ::= nschar -- matches given character
14869 elmt ::= [nschar @{nschar@}] -- matches any character listed
14870 elmt ::= [^^^ nschar @{nschar@}] -- matches any character not listed
14871 elmt ::= [char - char] -- matches chars in given range
14872 elmt ::= \ char -- matches given character
14873 elmt ::= . -- matches any single character
14874 elmt ::= ( regexp ) -- parens used for grouping
14876 char ::= any character, including special characters
14877 nschar ::= any character except ()[].*+?^^^
14881 Following are a few examples :
14885 will match any of the two strings 'abcde' and 'fghi'.
14888 will match any string like 'abd', 'abcd', 'abccd', 'abcccd', and so on
14891 will match any string which has only lowercase characters in it (and at
14892 least one character
14897 @node Examples of gnatxref Usage
14898 @section Examples of @code{gnatxref} Usage
14900 @subsection General Usage
14903 For the following examples, we will consider the following units :
14905 @smallexample @c ada
14911 3: procedure Foo (B : in Integer);
14918 1: package body Main is
14919 2: procedure Foo (B : in Integer) is
14930 2: procedure Print (B : Integer);
14939 The first thing to do is to recompile your application (for instance, in
14940 that case just by doing a @samp{gnatmake main}, so that GNAT generates
14941 the cross-referencing information.
14942 You can then issue any of the following commands:
14944 @item gnatxref main.adb
14945 @code{gnatxref} generates cross-reference information for main.adb
14946 and every unit 'with'ed by main.adb.
14948 The output would be:
14956 Decl: main.ads 3:20
14957 Body: main.adb 2:20
14958 Ref: main.adb 4:13 5:13 6:19
14961 Ref: main.adb 6:8 7:8
14971 Decl: main.ads 3:15
14972 Body: main.adb 2:15
14975 Body: main.adb 1:14
14978 Ref: main.adb 6:12 7:12
14982 that is the entity @code{Main} is declared in main.ads, line 2, column 9,
14983 its body is in main.adb, line 1, column 14 and is not referenced any where.
14985 The entity @code{Print} is declared in bar.ads, line 2, column 15 and it
14986 it referenced in main.adb, line 6 column 12 and line 7 column 12.
14988 @item gnatxref package1.adb package2.ads
14989 @code{gnatxref} will generates cross-reference information for
14990 package1.adb, package2.ads and any other package 'with'ed by any
14996 @subsection Using gnatxref with vi
14998 @code{gnatxref} can generate a tags file output, which can be used
14999 directly from @file{vi}. Note that the standard version of @file{vi}
15000 will not work properly with overloaded symbols. Consider using another
15001 free implementation of @file{vi}, such as @file{vim}.
15004 $ gnatxref -v gnatfind.adb > tags
15008 will generate the tags file for @code{gnatfind} itself (if the sources
15009 are in the search path!).
15011 From @file{vi}, you can then use the command @samp{:tag @i{entity}}
15012 (replacing @i{entity} by whatever you are looking for), and vi will
15013 display a new file with the corresponding declaration of entity.
15016 @node Examples of gnatfind Usage
15017 @section Examples of @code{gnatfind} Usage
15021 @item gnatfind ^-f^/FULL_PATHNAME^ xyz:main.adb
15022 Find declarations for all entities xyz referenced at least once in
15023 main.adb. The references are search in every library file in the search
15026 The directories will be printed as well (as the @samp{^-f^/FULL_PATHNAME^}
15029 The output will look like:
15031 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
15032 ^directory/^[directory]^main.adb:24:10: xyz <= body
15033 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
15037 that is to say, one of the entities xyz found in main.adb is declared at
15038 line 12 of main.ads (and its body is in main.adb), and another one is
15039 declared at line 45 of foo.ads
15041 @item gnatfind ^-fs^/FULL_PATHNAME/SOURCE_LINE^ xyz:main.adb
15042 This is the same command as the previous one, instead @code{gnatfind} will
15043 display the content of the Ada source file lines.
15045 The output will look like:
15048 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
15050 ^directory/^[directory]^main.adb:24:10: xyz <= body
15052 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
15057 This can make it easier to find exactly the location your are looking
15060 @item gnatfind ^-r^/REFERENCES^ "*x*":main.ads:123 foo.adb
15061 Find references to all entities containing an x that are
15062 referenced on line 123 of main.ads.
15063 The references will be searched only in main.ads and foo.adb.
15065 @item gnatfind main.ads:123
15066 Find declarations and bodies for all entities that are referenced on
15067 line 123 of main.ads.
15069 This is the same as @code{gnatfind "*":main.adb:123}.
15071 @item gnatfind ^mydir/^[mydir]^main.adb:123:45
15072 Find the declaration for the entity referenced at column 45 in
15073 line 123 of file main.adb in directory mydir. Note that it
15074 is usual to omit the identifier name when the column is given,
15075 since the column position identifies a unique reference.
15077 The column has to be the beginning of the identifier, and should not
15078 point to any character in the middle of the identifier.
15082 @c *********************************
15083 @node The GNAT Pretty-Printer gnatpp
15084 @chapter The GNAT Pretty-Printer @command{gnatpp}
15086 @cindex Pretty-Printer
15089 ^The @command{gnatpp} tool^GNAT PRETTY^ is an ASIS-based utility
15090 for source reformatting / pretty-printing.
15091 It takes an Ada source file as input and generates a reformatted
15093 You can specify various style directives via switches; e.g.,
15094 identifier case conventions, rules of indentation, and comment layout.
15096 To produce a reformatted file, @command{gnatpp} generates and uses the ASIS
15097 tree for the input source and thus requires the input to be syntactically and
15098 semantically legal.
15099 If this condition is not met, @command{gnatpp} will terminate with an
15100 error message; no output file will be generated.
15102 If the source files presented to @command{gnatpp} contain
15103 preprocessing directives, then the output file will
15104 correspond to the generated source after all
15105 preprocessing is carried out. There is no way
15106 using @command{gnatpp} to obtain pretty printed files that
15107 include the preprocessing directives.
15109 If the compilation unit
15110 contained in the input source depends semantically upon units located
15111 outside the current directory, you have to provide the source search path
15112 when invoking @command{gnatpp}, if these units are contained in files with
15113 names that do not follow the GNAT file naming rules, you have to provide
15114 the configuration file describing the corresponding naming scheme;
15115 see the description of the @command{gnatpp}
15116 switches below. Another possibility is to use a project file and to
15117 call @command{gnatpp} through the @command{gnat} driver
15119 The @command{gnatpp} command has the form
15122 $ gnatpp [@var{switches}] @var{filename}
15129 @var{switches} is an optional sequence of switches defining such properties as
15130 the formatting rules, the source search path, and the destination for the
15134 @var{filename} is the name (including the extension) of the source file to
15135 reformat; ``wildcards'' or several file names on the same gnatpp command are
15136 allowed. The file name may contain path information; it does not have to
15137 follow the GNAT file naming rules
15141 * Switches for gnatpp::
15142 * Formatting Rules::
15145 @node Switches for gnatpp
15146 @section Switches for @command{gnatpp}
15149 The following subsections describe the various switches accepted by
15150 @command{gnatpp}, organized by category.
15153 You specify a switch by supplying a name and generally also a value.
15154 In many cases the values for a switch with a given name are incompatible with
15156 (for example the switch that controls the casing of a reserved word may have
15157 exactly one value: upper case, lower case, or
15158 mixed case) and thus exactly one such switch can be in effect for an
15159 invocation of @command{gnatpp}.
15160 If more than one is supplied, the last one is used.
15161 However, some values for the same switch are mutually compatible.
15162 You may supply several such switches to @command{gnatpp}, but then
15163 each must be specified in full, with both the name and the value.
15164 Abbreviated forms (the name appearing once, followed by each value) are
15166 For example, to set
15167 the alignment of the assignment delimiter both in declarations and in
15168 assignment statements, you must write @option{-A2A3}
15169 (or @option{-A2 -A3}), but not @option{-A23}.
15173 In many cases the set of options for a given qualifier are incompatible with
15174 each other (for example the qualifier that controls the casing of a reserved
15175 word may have exactly one option, which specifies either upper case, lower
15176 case, or mixed case), and thus exactly one such option can be in effect for
15177 an invocation of @command{gnatpp}.
15178 If more than one is supplied, the last one is used.
15179 However, some qualifiers have options that are mutually compatible,
15180 and then you may then supply several such options when invoking
15184 In most cases, it is obvious whether or not the
15185 ^values for a switch with a given name^options for a given qualifier^
15186 are compatible with each other.
15187 When the semantics might not be evident, the summaries below explicitly
15188 indicate the effect.
15191 * Alignment Control::
15193 * Construct Layout Control::
15194 * General Text Layout Control::
15195 * Other Formatting Options::
15196 * Setting the Source Search Path::
15197 * Output File Control::
15198 * Other gnatpp Switches::
15201 @node Alignment Control
15202 @subsection Alignment Control
15203 @cindex Alignment control in @command{gnatpp}
15206 Programs can be easier to read if certain constructs are vertically aligned.
15207 By default all alignments are set ON.
15208 Through the @option{^-A0^/ALIGN=OFF^} switch you may reset the default to
15209 OFF, and then use one or more of the other
15210 ^@option{-A@var{n}} switches^@option{/ALIGN} options^
15211 to activate alignment for specific constructs.
15214 @cindex @option{^-A@var{n}^/ALIGN^} (@command{gnatpp})
15218 Set all alignments to ON
15221 @item ^-A0^/ALIGN=OFF^
15222 Set all alignments to OFF
15224 @item ^-A1^/ALIGN=COLONS^
15225 Align @code{:} in declarations
15227 @item ^-A2^/ALIGN=DECLARATIONS^
15228 Align @code{:=} in initializations in declarations
15230 @item ^-A3^/ALIGN=STATEMENTS^
15231 Align @code{:=} in assignment statements
15233 @item ^-A4^/ALIGN=ARROWS^
15234 Align @code{=>} in associations
15236 @item ^-A5^/ALIGN=COMPONENT_CLAUSES^
15237 Align @code{at} keywords in the component clauses in record
15238 representation clauses
15242 The @option{^-A^/ALIGN^} switches are mutually compatible; any combination
15245 @node Casing Control
15246 @subsection Casing Control
15247 @cindex Casing control in @command{gnatpp}
15250 @command{gnatpp} allows you to specify the casing for reserved words,
15251 pragma names, attribute designators and identifiers.
15252 For identifiers you may define a
15253 general rule for name casing but also override this rule
15254 via a set of dictionary files.
15256 Three types of casing are supported: lower case, upper case, and mixed case.
15257 Lower and upper case are self-explanatory (but since some letters in
15258 Latin1 and other GNAT-supported character sets
15259 exist only in lower-case form, an upper case conversion will have no
15261 ``Mixed case'' means that the first letter, and also each letter immediately
15262 following an underscore, are converted to their uppercase forms;
15263 all the other letters are converted to their lowercase forms.
15266 @cindex @option{^-a@var{x}^/ATTRIBUTE^} (@command{gnatpp})
15267 @item ^-aL^/ATTRIBUTE_CASING=LOWER_CASE^
15268 Attribute designators are lower case
15270 @item ^-aU^/ATTRIBUTE_CASING=UPPER_CASE^
15271 Attribute designators are upper case
15273 @item ^-aM^/ATTRIBUTE_CASING=MIXED_CASE^
15274 Attribute designators are mixed case (this is the default)
15276 @cindex @option{^-k@var{x}^/KEYWORD_CASING^} (@command{gnatpp})
15277 @item ^-kL^/KEYWORD_CASING=LOWER_CASE^
15278 Keywords (technically, these are known in Ada as @emph{reserved words}) are
15279 lower case (this is the default)
15281 @item ^-kU^/KEYWORD_CASING=UPPER_CASE^
15282 Keywords are upper case
15284 @cindex @option{^-n@var{x}^/NAME_CASING^} (@command{gnatpp})
15285 @item ^-nD^/NAME_CASING=AS_DECLARED^
15286 Name casing for defining occurrences are as they appear in the source file
15287 (this is the default)
15289 @item ^-nU^/NAME_CASING=UPPER_CASE^
15290 Names are in upper case
15292 @item ^-nL^/NAME_CASING=LOWER_CASE^
15293 Names are in lower case
15295 @item ^-nM^/NAME_CASING=MIXED_CASE^
15296 Names are in mixed case
15298 @cindex @option{^-p@var{x}^/PRAGMA_CASING^} (@command{gnatpp})
15299 @item ^-pL^/PRAGMA_CASING=LOWER_CASE^
15300 Pragma names are lower case
15302 @item ^-pU^/PRAGMA_CASING=UPPER_CASE^
15303 Pragma names are upper case
15305 @item ^-pM^/PRAGMA_CASING=MIXED_CASE^
15306 Pragma names are mixed case (this is the default)
15308 @item ^-D@var{file}^/DICTIONARY=@var{file}^
15309 @cindex @option{^-D^/DICTIONARY^} (@command{gnatpp})
15310 Use @var{file} as a @emph{dictionary file} that defines
15311 the casing for a set of specified names,
15312 thereby overriding the effect on these names by
15313 any explicit or implicit
15314 ^-n^/NAME_CASING^ switch.
15315 To supply more than one dictionary file,
15316 use ^several @option{-D} switches^a list of files as options^.
15319 @option{gnatpp} implicitly uses a @emph{default dictionary file}
15320 to define the casing for the Ada predefined names and
15321 the names declared in the GNAT libraries.
15323 @item ^-D-^/SPECIFIC_CASING^
15324 @cindex @option{^-D-^/SPECIFIC_CASING^} (@command{gnatpp})
15325 Do not use the default dictionary file;
15326 instead, use the casing
15327 defined by a @option{^-n^/NAME_CASING^} switch and any explicit
15332 The structure of a dictionary file, and details on the conventions
15333 used in the default dictionary file, are defined in @ref{Name Casing}.
15335 The @option{^-D-^/SPECIFIC_CASING^} and
15336 @option{^-D@var{file}^/DICTIONARY=@var{file}^} switches are mutually
15339 @node Construct Layout Control
15340 @subsection Construct Layout Control
15341 @cindex Layout control in @command{gnatpp}
15344 This group of @command{gnatpp} switches controls the layout of comments and
15345 complex syntactic constructs. See @ref{Formatting Comments} for details
15349 @cindex @option{^-c@var{n}^/COMMENTS_LAYOUT^} (@command{gnatpp})
15350 @item ^-c0^/COMMENTS_LAYOUT=UNTOUCHED^
15351 All the comments remain unchanged
15353 @item ^-c1^/COMMENTS_LAYOUT=DEFAULT^
15354 GNAT-style comment line indentation (this is the default).
15356 @item ^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^
15357 Reference-manual comment line indentation.
15359 @item ^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^
15360 GNAT-style comment beginning
15362 @item ^-c4^/COMMENTS_LAYOUT=REFORMAT^
15363 Reformat comment blocks
15365 @cindex @option{^-l@var{n}^/CONSTRUCT_LAYOUT^} (@command{gnatpp})
15366 @item ^-l1^/CONSTRUCT_LAYOUT=GNAT^
15367 GNAT-style layout (this is the default)
15369 @item ^-l2^/CONSTRUCT_LAYOUT=COMPACT^
15372 @item ^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^
15375 @cindex @option{^-N^/NOTABS^} (@command{gnatpp})
15377 All the VT characters are removed from the comment text. All the HT characters
15378 are expanded with the sequences of space characters to get to the next tab
15381 @cindex @option{^--no-separate-is^/NO_SEPARATE_IS^} (@command{gnatpp})
15382 @item ^--no-separate-is^/NO_SEPARATE_IS^
15383 Do not place the keyword @code{is} on a separate line in a subprogram body in
15384 case if the specification occupies more then one line.
15390 The @option{-c1} and @option{-c2} switches are incompatible.
15391 The @option{-c3} and @option{-c4} switches are compatible with each other and
15392 also with @option{-c1} and @option{-c2}. The @option{-c0} switch disables all
15393 the other comment formatting switches.
15395 The @option{-l1}, @option{-l2}, and @option{-l3} switches are incompatible.
15400 For the @option{/COMMENTS_LAYOUT} qualifier:
15403 The @option{DEFAULT} and @option{STANDARD_INDENT} options are incompatible.
15405 The @option{GNAT_BEGINNING} and @option{REFORMAT} options are compatible with
15406 each other and also with @option{DEFAULT} and @option{STANDARD_INDENT}.
15410 The @option{GNAT}, @option{COMPACT}, and @option{UNCOMPACT} options for the
15411 @option{/CONSTRUCT_LAYOUT} qualifier are incompatible.
15414 @node General Text Layout Control
15415 @subsection General Text Layout Control
15418 These switches allow control over line length and indentation.
15421 @item ^-M@i{nnn}^/LINE_LENGTH_MAX=@i{nnn}^
15422 @cindex @option{^-M^/LINE_LENGTH^} (@command{gnatpp})
15423 Maximum line length, @i{nnn} from 32 ..256, the default value is 79
15425 @item ^-i@i{nnn}^/INDENTATION_LEVEL=@i{nnn}^
15426 @cindex @option{^-i^/INDENTATION_LEVEL^} (@command{gnatpp})
15427 Indentation level, @i{nnn} from 1 .. 9, the default value is 3
15429 @item ^-cl@i{nnn}^/CONTINUATION_INDENT=@i{nnn}^
15430 @cindex @option{^-cl^/CONTINUATION_INDENT^} (@command{gnatpp})
15431 Indentation level for continuation lines (relative to the line being
15432 continued), @i{nnn} from 1 .. 9.
15434 value is one less then the (normal) indentation level, unless the
15435 indentation is set to 1 (in which case the default value for continuation
15436 line indentation is also 1)
15439 @node Other Formatting Options
15440 @subsection Other Formatting Options
15443 These switches control the inclusion of missing end/exit labels, and
15444 the indentation level in @b{case} statements.
15447 @item ^-e^/NO_MISSED_LABELS^
15448 @cindex @option{^-e^/NO_MISSED_LABELS^} (@command{gnatpp})
15449 Do not insert missing end/exit labels. An end label is the name of
15450 a construct that may optionally be repeated at the end of the
15451 construct's declaration;
15452 e.g., the names of packages, subprograms, and tasks.
15453 An exit label is the name of a loop that may appear as target
15454 of an exit statement within the loop.
15455 By default, @command{gnatpp} inserts these end/exit labels when
15456 they are absent from the original source. This option suppresses such
15457 insertion, so that the formatted source reflects the original.
15459 @item ^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^
15460 @cindex @option{^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^} (@command{gnatpp})
15461 Insert a Form Feed character after a pragma Page.
15463 @item ^-T@i{nnn}^/MAX_INDENT=@i{nnn}^
15464 @cindex @option{^-T^/MAX_INDENT^} (@command{gnatpp})
15465 Do not use an additional indentation level for @b{case} alternatives
15466 and variants if there are @i{nnn} or more (the default
15468 If @i{nnn} is 0, an additional indentation level is
15469 used for @b{case} alternatives and variants regardless of their number.
15472 @node Setting the Source Search Path
15473 @subsection Setting the Source Search Path
15476 To define the search path for the input source file, @command{gnatpp}
15477 uses the same switches as the GNAT compiler, with the same effects.
15480 @item ^-I^/SEARCH=^@var{dir}
15481 @cindex @option{^-I^/SEARCH^} (@code{gnatpp})
15482 The same as the corresponding gcc switch
15484 @item ^-I-^/NOCURRENT_DIRECTORY^
15485 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatpp})
15486 The same as the corresponding gcc switch
15488 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE^=@var{path}
15489 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@code{gnatpp})
15490 The same as the corresponding gcc switch
15492 @item ^--RTS^/RUNTIME_SYSTEM^=@var{path}
15493 @cindex @option{^--RTS^/RUNTIME_SYSTEM^} (@code{gnatpp})
15494 The same as the corresponding gcc switch
15498 @node Output File Control
15499 @subsection Output File Control
15502 By default the output is sent to the file whose name is obtained by appending
15503 the ^@file{.pp}^@file{$PP}^ suffix to the name of the input file
15504 (if the file with this name already exists, it is unconditionally overwritten).
15505 Thus if the input file is @file{^my_ada_proc.adb^MY_ADA_PROC.ADB^} then
15506 @command{gnatpp} will produce @file{^my_ada_proc.adb.pp^MY_ADA_PROC.ADB$PP^}
15508 The output may be redirected by the following switches:
15511 @item ^-pipe^/STANDARD_OUTPUT^
15512 @cindex @option{^-pipe^/STANDARD_OUTPUT^} (@code{gnatpp})
15513 Send the output to @code{Standard_Output}
15515 @item ^-o @var{output_file}^/OUTPUT=@var{output_file}^
15516 @cindex @option{^-o^/OUTPUT^} (@code{gnatpp})
15517 Write the output into @var{output_file}.
15518 If @var{output_file} already exists, @command{gnatpp} terminates without
15519 reading or processing the input file.
15521 @item ^-of ^/FORCED_OUTPUT=^@var{output_file}
15522 @cindex @option{^-of^/FORCED_OUTPUT^} (@code{gnatpp})
15523 Write the output into @var{output_file}, overwriting the existing file
15524 (if one is present).
15526 @item ^-r^/REPLACE^
15527 @cindex @option{^-r^/REPLACE^} (@code{gnatpp})
15528 Replace the input source file with the reformatted output, and copy the
15529 original input source into the file whose name is obtained by appending the
15530 ^@file{.npp}^@file{$NPP}^ suffix to the name of the input file.
15531 If a file with this name already exists, @command{gnatpp} terminates without
15532 reading or processing the input file.
15534 @item ^-rf^/OVERRIDING_REPLACE^
15535 @cindex @option{^-rf^/OVERRIDING_REPLACE^} (@code{gnatpp})
15536 Like @option{^-r^/REPLACE^} except that if the file with the specified name
15537 already exists, it is overwritten.
15539 @item ^-rnb^/NO_BACKUP^
15540 @cindex @option{^-rnb^/NO_BACKUP^} (@code{gnatpp})
15541 Replace the input source file with the reformatted output without
15542 creating any backup copy of the input source.
15544 @item ^--eol=@var{xxx}^/END_OF_LINE=@var{xxx}^
15545 @cindex @option{^--eol^/END_OF_LINE^} (@code{gnatpp})
15546 Specifies the format of the reformatted output file. The @var{xxx}
15547 ^string specified with the switch^option^ may be either
15549 @item ``@option{^dos^DOS^}'' MS DOS style, lines end with CR LF characters
15550 @item ``@option{^crlf^CRLF^}''
15551 the same as @option{^crlf^CRLF^}
15552 @item ``@option{^unix^UNIX^}'' UNIX style, lines end with LF character
15553 @item ``@option{^lf^LF^}''
15554 the same as @option{^unix^UNIX^}
15557 @item ^-W^/RESULT_ENCODING=^@var{e}
15558 @cindex @option{^-W^/RESULT_ENCODING=^} (@command{gnatpp})
15559 Specify the wide character encoding method used to write the code in the
15561 @var{e} is one of the following:
15569 Upper half encoding
15571 @item ^s^SHIFT_JIS^
15581 Brackets encoding (default value)
15587 Options @option{^-pipe^/STANDARD_OUTPUT^},
15588 @option{^-o^/OUTPUT^} and
15589 @option{^-of^/FORCED_OUTPUT^} are allowed only if the call to gnatpp
15590 contains only one file to reformat.
15592 @option{^--eol^/END_OF_LINE^}
15594 @option{^-W^/RESULT_ENCODING^}
15595 cannot be used together
15596 with @option{^-pipe^/STANDARD_OUTPUT^} option.
15598 @node Other gnatpp Switches
15599 @subsection Other @code{gnatpp} Switches
15602 The additional @command{gnatpp} switches are defined in this subsection.
15605 @item ^-files @var{filename}^/FILES=@var{output_file}^
15606 @cindex @option{^-files^/FILES^} (@code{gnatpp})
15607 Take the argument source files from the specified file. This file should be an
15608 ordinary textual file containing file names separated by spaces or
15609 line breaks. You can use this switch more then once in the same call to
15610 @command{gnatpp}. You also can combine this switch with explicit list of
15613 @item ^-v^/VERBOSE^
15614 @cindex @option{^-v^/VERBOSE^} (@code{gnatpp})
15616 @command{gnatpp} generates version information and then
15617 a trace of the actions it takes to produce or obtain the ASIS tree.
15619 @item ^-w^/WARNINGS^
15620 @cindex @option{^-w^/WARNINGS^} (@code{gnatpp})
15622 @command{gnatpp} generates a warning whenever it cannot provide
15623 a required layout in the result source.
15626 @node Formatting Rules
15627 @section Formatting Rules
15630 The following subsections show how @command{gnatpp} treats ``white space'',
15631 comments, program layout, and name casing.
15632 They provide the detailed descriptions of the switches shown above.
15635 * White Space and Empty Lines::
15636 * Formatting Comments::
15637 * Construct Layout::
15641 @node White Space and Empty Lines
15642 @subsection White Space and Empty Lines
15645 @command{gnatpp} does not have an option to control space characters.
15646 It will add or remove spaces according to the style illustrated by the
15647 examples in the @cite{Ada Reference Manual}.
15649 The only format effectors
15650 (see @cite{Ada Reference Manual}, paragraph 2.1(13))
15651 that will appear in the output file are platform-specific line breaks,
15652 and also format effectors within (but not at the end of) comments.
15653 In particular, each horizontal tab character that is not inside
15654 a comment will be treated as a space and thus will appear in the
15655 output file as zero or more spaces depending on
15656 the reformatting of the line in which it appears.
15657 The only exception is a Form Feed character, which is inserted after a
15658 pragma @code{Page} when @option{-ff} is set.
15660 The output file will contain no lines with trailing ``white space'' (spaces,
15663 Empty lines in the original source are preserved
15664 only if they separate declarations or statements.
15665 In such contexts, a
15666 sequence of two or more empty lines is replaced by exactly one empty line.
15667 Note that a blank line will be removed if it separates two ``comment blocks''
15668 (a comment block is a sequence of whole-line comments).
15669 In order to preserve a visual separation between comment blocks, use an
15670 ``empty comment'' (a line comprising only hyphens) rather than an empty line.
15671 Likewise, if for some reason you wish to have a sequence of empty lines,
15672 use a sequence of empty comments instead.
15674 @node Formatting Comments
15675 @subsection Formatting Comments
15678 Comments in Ada code are of two kinds:
15681 a @emph{whole-line comment}, which appears by itself (possibly preceded by
15682 ``white space'') on a line
15685 an @emph{end-of-line comment}, which follows some other Ada lexical element
15690 The indentation of a whole-line comment is that of either
15691 the preceding or following line in
15692 the formatted source, depending on switch settings as will be described below.
15694 For an end-of-line comment, @command{gnatpp} leaves the same number of spaces
15695 between the end of the preceding Ada lexical element and the beginning
15696 of the comment as appear in the original source,
15697 unless either the comment has to be split to
15698 satisfy the line length limitation, or else the next line contains a
15699 whole line comment that is considered a continuation of this end-of-line
15700 comment (because it starts at the same position).
15702 cases, the start of the end-of-line comment is moved right to the nearest
15703 multiple of the indentation level.
15704 This may result in a ``line overflow'' (the right-shifted comment extending
15705 beyond the maximum line length), in which case the comment is split as
15708 There is a difference between @option{^-c1^/COMMENTS_LAYOUT=DEFAULT^}
15709 (GNAT-style comment line indentation)
15710 and @option{^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^}
15711 (reference-manual comment line indentation).
15712 With reference-manual style, a whole-line comment is indented as if it
15713 were a declaration or statement at the same place
15714 (i.e., according to the indentation of the preceding line(s)).
15715 With GNAT style, a whole-line comment that is immediately followed by an
15716 @b{if} or @b{case} statement alternative, a record variant, or the reserved
15717 word @b{begin}, is indented based on the construct that follows it.
15720 @smallexample @c ada
15732 Reference-manual indentation produces:
15734 @smallexample @c ada
15746 while GNAT-style indentation produces:
15748 @smallexample @c ada
15760 The @option{^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^} switch
15761 (GNAT style comment beginning) has the following
15766 For each whole-line comment that does not end with two hyphens,
15767 @command{gnatpp} inserts spaces if necessary after the starting two hyphens
15768 to ensure that there are at least two spaces between these hyphens and the
15769 first non-blank character of the comment.
15773 For an end-of-line comment, if in the original source the next line is a
15774 whole-line comment that starts at the same position
15775 as the end-of-line comment,
15776 then the whole-line comment (and all whole-line comments
15777 that follow it and that start at the same position)
15778 will start at this position in the output file.
15781 That is, if in the original source we have:
15783 @smallexample @c ada
15786 A := B + C; -- B must be in the range Low1..High1
15787 -- C must be in the range Low2..High2
15788 --B+C will be in the range Low1+Low2..High1+High2
15794 Then in the formatted source we get
15796 @smallexample @c ada
15799 A := B + C; -- B must be in the range Low1..High1
15800 -- C must be in the range Low2..High2
15801 -- B+C will be in the range Low1+Low2..High1+High2
15807 A comment that exceeds the line length limit will be split.
15809 @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} (reformat comment blocks) is set and
15810 the line belongs to a reformattable block, splitting the line generates a
15811 @command{gnatpp} warning.
15812 The @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} switch specifies that whole-line
15813 comments may be reformatted in typical
15814 word processor style (that is, moving words between lines and putting as
15815 many words in a line as possible).
15817 @node Construct Layout
15818 @subsection Construct Layout
15821 In several cases the suggested layout in the Ada Reference Manual includes
15822 an extra level of indentation that many programmers prefer to avoid. The
15823 affected cases include:
15827 @item Record type declaration (RM 3.8)
15829 @item Record representation clause (RM 13.5.1)
15831 @item Loop statement in case if a loop has a statement identifier (RM 5.6)
15833 @item Block statement in case if a block has a statement identifier (RM 5.6)
15837 In compact mode (when GNAT style layout or compact layout is set),
15838 the pretty printer uses one level of indentation instead
15839 of two. This is achieved in the record definition and record representation
15840 clause cases by putting the @code{record} keyword on the same line as the
15841 start of the declaration or representation clause, and in the block and loop
15842 case by putting the block or loop header on the same line as the statement
15846 The difference between GNAT style @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^}
15847 and compact @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^}
15848 layout on the one hand, and uncompact layout
15849 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} on the other hand,
15850 can be illustrated by the following examples:
15854 @multitable @columnfractions .5 .5
15855 @item @i{GNAT style, compact layout} @tab @i{Uncompact layout}
15858 @smallexample @c ada
15865 @smallexample @c ada
15874 @smallexample @c ada
15876 a at 0 range 0 .. 31;
15877 b at 4 range 0 .. 31;
15881 @smallexample @c ada
15884 a at 0 range 0 .. 31;
15885 b at 4 range 0 .. 31;
15890 @smallexample @c ada
15898 @smallexample @c ada
15908 @smallexample @c ada
15909 Clear : for J in 1 .. 10 loop
15914 @smallexample @c ada
15916 for J in 1 .. 10 loop
15927 GNAT style, compact layout Uncompact layout
15929 type q is record type q is
15930 a : integer; record
15931 b : integer; a : integer;
15932 end record; b : integer;
15935 for q use record for q use
15936 a at 0 range 0 .. 31; record
15937 b at 4 range 0 .. 31; a at 0 range 0 .. 31;
15938 end record; b at 4 range 0 .. 31;
15941 Block : declare Block :
15942 A : Integer := 3; declare
15943 begin A : Integer := 3;
15945 end Block; Proc (A, A);
15948 Clear : for J in 1 .. 10 loop Clear :
15949 A (J) := 0; for J in 1 .. 10 loop
15950 end loop Clear; A (J) := 0;
15957 A further difference between GNAT style layout and compact layout is that
15958 GNAT style layout inserts empty lines as separation for
15959 compound statements, return statements and bodies.
15962 @subsection Name Casing
15965 @command{gnatpp} always converts the usage occurrence of a (simple) name to
15966 the same casing as the corresponding defining identifier.
15968 You control the casing for defining occurrences via the
15969 @option{^-n^/NAME_CASING^} switch.
15971 With @option{-nD} (``as declared'', which is the default),
15974 With @option{/NAME_CASING=AS_DECLARED}, which is the default,
15976 defining occurrences appear exactly as in the source file
15977 where they are declared.
15978 The other ^values for this switch^options for this qualifier^ ---
15979 @option{^-nU^UPPER_CASE^},
15980 @option{^-nL^LOWER_CASE^},
15981 @option{^-nM^MIXED_CASE^} ---
15983 ^upper, lower, or mixed case, respectively^the corresponding casing^.
15984 If @command{gnatpp} changes the casing of a defining
15985 occurrence, it analogously changes the casing of all the
15986 usage occurrences of this name.
15988 If the defining occurrence of a name is not in the source compilation unit
15989 currently being processed by @command{gnatpp}, the casing of each reference to
15990 this name is changed according to the value of the @option{^-n^/NAME_CASING^}
15991 switch (subject to the dictionary file mechanism described below).
15992 Thus @command{gnatpp} acts as though the @option{^-n^/NAME_CASING^} switch
15994 casing for the defining occurrence of the name.
15996 Some names may need to be spelled with casing conventions that are not
15997 covered by the upper-, lower-, and mixed-case transformations.
15998 You can arrange correct casing by placing such names in a
15999 @emph{dictionary file},
16000 and then supplying a @option{^-D^/DICTIONARY^} switch.
16001 The casing of names from dictionary files overrides
16002 any @option{^-n^/NAME_CASING^} switch.
16004 To handle the casing of Ada predefined names and the names from GNAT libraries,
16005 @command{gnatpp} assumes a default dictionary file.
16006 The name of each predefined entity is spelled with the same casing as is used
16007 for the entity in the @cite{Ada Reference Manual}.
16008 The name of each entity in the GNAT libraries is spelled with the same casing
16009 as is used in the declaration of that entity.
16011 The @w{@option{^-D-^/SPECIFIC_CASING^}} switch suppresses the use of the
16012 default dictionary file.
16013 Instead, the casing for predefined and GNAT-defined names will be established
16014 by the @option{^-n^/NAME_CASING^} switch or explicit dictionary files.
16015 For example, by default the names @code{Ada.Text_IO} and @code{GNAT.OS_Lib}
16016 will appear as just shown,
16017 even in the presence of a @option{^-nU^/NAME_CASING=UPPER_CASE^} switch.
16018 To ensure that even such names are rendered in uppercase,
16019 additionally supply the @w{@option{^-D-^/SPECIFIC_CASING^}} switch
16020 (or else, less conveniently, place these names in upper case in a dictionary
16023 A dictionary file is
16024 a plain text file; each line in this file can be either a blank line
16025 (containing only space characters and ASCII.HT characters), an Ada comment
16026 line, or the specification of exactly one @emph{casing schema}.
16028 A casing schema is a string that has the following syntax:
16032 @var{casing_schema} ::= @var{identifier} | *@var{simple_identifier}*
16034 @var{simple_identifier} ::= @var{letter}@{@var{letter_or_digit}@}
16039 (See @cite{Ada Reference Manual}, Section 2.3) for the definition of the
16040 @var{identifier} lexical element and the @var{letter_or_digit} category.)
16042 The casing schema string can be followed by white space and/or an Ada-style
16043 comment; any amount of white space is allowed before the string.
16045 If a dictionary file is passed as
16047 the value of a @option{-D@var{file}} switch
16050 an option to the @option{/DICTIONARY} qualifier
16053 simple name and every identifier, @command{gnatpp} checks if the dictionary
16054 defines the casing for the name or for some of its parts (the term ``subword''
16055 is used below to denote the part of a name which is delimited by ``_'' or by
16056 the beginning or end of the word and which does not contain any ``_'' inside):
16060 if the whole name is in the dictionary, @command{gnatpp} uses for this name
16061 the casing defined by the dictionary; no subwords are checked for this word
16064 for every subword @command{gnatpp} checks if the dictionary contains the
16065 corresponding string of the form @code{*@var{simple_identifier}*},
16066 and if it does, the casing of this @var{simple_identifier} is used
16070 if the whole name does not contain any ``_'' inside, and if for this name
16071 the dictionary contains two entries - one of the form @var{identifier},
16072 and another - of the form *@var{simple_identifier}*, then the first one
16073 is applied to define the casing of this name
16076 if more than one dictionary file is passed as @command{gnatpp} switches, each
16077 dictionary adds new casing exceptions and overrides all the existing casing
16078 exceptions set by the previous dictionaries
16081 when @command{gnatpp} checks if the word or subword is in the dictionary,
16082 this check is not case sensitive
16086 For example, suppose we have the following source to reformat:
16088 @smallexample @c ada
16091 name1 : integer := 1;
16092 name4_name3_name2 : integer := 2;
16093 name2_name3_name4 : Boolean;
16096 name2_name3_name4 := name4_name3_name2 > name1;
16102 And suppose we have two dictionaries:
16119 If @command{gnatpp} is called with the following switches:
16123 @command{gnatpp -nM -D dict1 -D dict2 test.adb}
16126 @command{gnatpp test.adb /NAME_CASING=MIXED_CASE /DICTIONARY=(dict1, dict2)}
16131 then we will get the following name casing in the @command{gnatpp} output:
16133 @smallexample @c ada
16136 NAME1 : Integer := 1;
16137 Name4_NAME3_Name2 : Integer := 2;
16138 Name2_NAME3_Name4 : Boolean;
16141 Name2_NAME3_Name4 := Name4_NAME3_Name2 > NAME1;
16146 @c *********************************
16147 @node The GNAT Metric Tool gnatmetric
16148 @chapter The GNAT Metric Tool @command{gnatmetric}
16150 @cindex Metric tool
16153 ^The @command{gnatmetric} tool^@command{GNAT METRIC}^ is an ASIS-based utility
16154 for computing various program metrics.
16155 It takes an Ada source file as input and generates a file containing the
16156 metrics data as output. Various switches control which
16157 metrics are computed and output.
16159 @command{gnatmetric} generates and uses the ASIS
16160 tree for the input source and thus requires the input to be syntactically and
16161 semantically legal.
16162 If this condition is not met, @command{gnatmetric} will generate
16163 an error message; no metric information for this file will be
16164 computed and reported.
16166 If the compilation unit contained in the input source depends semantically
16167 upon units in files located outside the current directory, you have to provide
16168 the source search path when invoking @command{gnatmetric}.
16169 If it depends semantically upon units that are contained
16170 in files with names that do not follow the GNAT file naming rules, you have to
16171 provide the configuration file describing the corresponding naming scheme (see
16172 the description of the @command{gnatmetric} switches below.)
16173 Alternatively, you may use a project file and invoke @command{gnatmetric}
16174 through the @command{gnat} driver.
16176 The @command{gnatmetric} command has the form
16179 $ gnatmetric [@i{switches}] @{@i{filename}@} [@i{-cargs gcc_switches}]
16186 @i{switches} specify the metrics to compute and define the destination for
16190 Each @i{filename} is the name (including the extension) of a source
16191 file to process. ``Wildcards'' are allowed, and
16192 the file name may contain path information.
16193 If no @i{filename} is supplied, then the @i{switches} list must contain
16195 @option{-files} switch (@pxref{Other gnatmetric Switches}).
16196 Including both a @option{-files} switch and one or more
16197 @i{filename} arguments is permitted.
16200 @i{-cargs gcc_switches} is a list of switches for
16201 @command{gcc}. They will be passed on to all compiler invocations made by
16202 @command{gnatmetric} to generate the ASIS trees. Here you can provide
16203 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
16204 and use the @option{-gnatec} switch to set the configuration file.
16208 * Switches for gnatmetric::
16211 @node Switches for gnatmetric
16212 @section Switches for @command{gnatmetric}
16215 The following subsections describe the various switches accepted by
16216 @command{gnatmetric}, organized by category.
16219 * Output Files Control::
16220 * Disable Metrics For Local Units::
16221 * Line Metrics Control::
16222 * Syntax Metrics Control::
16223 * Complexity Metrics Control::
16224 * Other gnatmetric Switches::
16225 * Generate project-wide metrics::
16228 @node Output Files Control
16229 @subsection Output File Control
16230 @cindex Output file control in @command{gnatmetric}
16233 @command{gnatmetric} has two output formats. It can generate a
16234 textual (human-readable) form, and also XML. By default only textual
16235 output is generated.
16237 When generating the output in textual form, @command{gnatmetric} creates
16238 for each Ada source file a corresponding text file
16239 containing the computed metrics. By default, this file
16240 is placed in the same directory as where the source file is located, and
16241 its name is obtained
16242 by appending the ^@file{.metrix}^@file{$METRIX}^ suffix to the name of the
16245 All the output information generated in XML format is placed in a single
16246 file. By default this file is placed in the current directory and has the
16247 name ^@file{metrix.xml}^@file{METRIX$XML}^.
16249 Some of the computed metrics are summed over the units passed to
16250 @command{gnatmetric}; for example, the total number of lines of code.
16251 By default this information is sent to @file{stdout}, but a file
16252 can be specified with the @option{-og} switch.
16254 The following switches control the @command{gnatmetric} output:
16257 @cindex @option{^-x^/XML^} (@command{gnatmetric})
16259 Generate the XML output
16261 @cindex @option{^-nt^/NO_TEXT^} (@command{gnatmetric})
16262 @item ^-nt^/NO_TEXT^
16263 Do not generate the output in text form (implies @option{^-x^/XML^})
16265 @cindex @option{^-d^/DIRECTORY^} (@command{gnatmetric})
16266 @item ^-d @var{output_dir}^/DIRECTORY=@var{output_dir}^
16267 Put textual files with detailed metrics into @var{output_dir}
16269 @cindex @option{^-o^/SUFFIX_DETAILS^} (@command{gnatmetric})
16270 @item ^-o @var{file_suffix}^/SUFFIX_DETAILS=@var{file_suffix}^
16271 Use @var{file_suffix}, instead of ^@file{.metrix}^@file{$METRIX}^
16272 in the name of the output file.
16274 @cindex @option{^-og^/GLOBAL_OUTPUT^} (@command{gnatmetric})
16275 @item ^-og @var{file_name}^/GLOBAL_OUTPUT=@var{file_name}^
16276 Put global metrics into @var{file_name}
16278 @cindex @option{^-ox^/XML_OUTPUT^} (@command{gnatmetric})
16279 @item ^-ox @var{file_name}^/XML_OUTPUT=@var{file_name}^
16280 Put the XML output into @var{file_name} (also implies @option{^-x^/XML^})
16282 @cindex @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} (@command{gnatmetric})
16283 @item ^-sfn^/SHORT_SOURCE_FILE_NAME^
16284 Use ``short'' source file names in the output. (The @command{gnatmetric}
16285 output includes the name(s) of the Ada source file(s) from which the metrics
16286 are computed. By default each name includes the absolute path. The
16287 @option{^-sfn^/SHORT_SOURCE_FILE_NAME^} switch causes @command{gnatmetric}
16288 to exclude all directory information from the file names that are output.)
16292 @node Disable Metrics For Local Units
16293 @subsection Disable Metrics For Local Units
16294 @cindex Disable Metrics For Local Units in @command{gnatmetric}
16297 @command{gnatmetric} relies on the GNAT compilation model @minus{}
16299 unit per one source file. It computes line metrics for the whole source
16300 file, and it also computes syntax
16301 and complexity metrics for the file's outermost unit.
16303 By default, @command{gnatmetric} will also compute all metrics for certain
16304 kinds of locally declared program units:
16308 subprogram (and generic subprogram) bodies;
16311 package (and generic package) specifications and bodies;
16314 task object and type specifications and bodies;
16317 protected object and type specifications and bodies.
16321 These kinds of entities will be referred to as
16322 @emph{eligible local program units}, or simply @emph{eligible local units},
16323 @cindex Eligible local unit (for @command{gnatmetric})
16324 in the discussion below.
16326 Note that a subprogram declaration, generic instantiation,
16327 or renaming declaration only receives metrics
16328 computation when it appear as the outermost entity
16331 Suppression of metrics computation for eligible local units can be
16332 obtained via the following switch:
16335 @cindex @option{^-n@var{x}^/SUPPRESS^} (@command{gnatmetric})
16336 @item ^-nolocal^/SUPPRESS=LOCAL_DETAILS^
16337 Do not compute detailed metrics for eligible local program units
16341 @node Line Metrics Control
16342 @subsection Line Metrics Control
16343 @cindex Line metrics control in @command{gnatmetric}
16346 For any (legal) source file, and for each of its
16347 eligible local program units, @command{gnatmetric} computes the following
16352 the total number of lines;
16355 the total number of code lines (i.e., non-blank lines that are not comments)
16358 the number of comment lines
16361 the number of code lines containing end-of-line comments;
16364 the number of empty lines and lines containing only space characters and/or
16365 format effectors (blank lines)
16369 If @command{gnatmetric} is invoked on more than one source file, it sums the
16370 values of the line metrics for all the files being processed and then
16371 generates the cumulative results.
16373 By default, all the line metrics are computed and reported. You can use the
16374 following switches to select the specific line metrics to be computed and
16375 reported (if any of these parameters is set, only explicitly specified line
16376 metrics are computed).
16379 @cindex @option{^-la^/LINES_ALL^} (@command{gnatmetric})
16380 @item ^-la^/LINES_ALL^
16381 The number of all lines
16383 @cindex @option{^-lcode^/CODE_LINES^} (@command{gnatmetric})
16384 @item ^-lcode^/CODE_LINES^
16385 The number of code lines
16387 @cindex @option{^-lcomm^/COMMENT_LINES^} (@command{gnatmetric})
16388 @item ^-lcomm^/COMENT_LINES^
16389 The number of comment lines
16391 @cindex @option{^-leol^/MIXED_CODE_COMMENTS^} (@command{gnatmetric})
16392 @item ^-leol^/MIXED_CODE_COMMENTS^
16393 The number of code lines containing
16394 end-of-line comments
16396 @cindex @option{^-lb^/BLANK_LINES^} (@command{gnatmetric})
16397 @item ^-lb^/BLANK_LINES^
16398 The number of blank lines
16402 @node Syntax Metrics Control
16403 @subsection Syntax Metrics Control
16404 @cindex Syntax metrics control in @command{gnatmetric}
16407 @command{gnatmetric} computes various syntactic metrics for the
16408 outermost unit and for each eligible local unit:
16411 @item LSLOC (``Logical Source Lines Of Code'')
16412 The total number of declarations and the total number of statements
16414 @item Maximal static nesting level of inner program units
16416 @cite{Ada Reference Manual}, 10.1(1), ``A program unit is either a
16417 package, a task unit, a protected unit, a
16418 protected entry, a generic unit, or an explicitly declared subprogram other
16419 than an enumeration literal.''
16421 @item Maximal nesting level of composite syntactic constructs
16422 This corresponds to the notion of the
16423 maximum nesting level in the GNAT built-in style checks
16424 (@pxref{Style Checking})
16428 For the outermost unit in the file, @command{gnatmetric} additionally computes
16429 the following metrics:
16432 @item Public subprograms
16433 This metric is computed for package specifications. It is the
16434 number of subprograms and generic subprograms declared in the visible
16435 part (including in nested packages, protected objects, and
16438 @item All subprograms
16439 This metric is computed for bodies and subunits. The
16440 metric is equal to a total number of subprogram bodies in the compilation
16442 Neither generic instantiations nor renamings-as-a-body nor body stubs
16443 are counted. Any subprogram body is counted, independently of its nesting
16444 level and enclosing constructs. Generic bodies and bodies of protected
16445 subprograms are counted in the same way as ``usual'' subprogram bodies.
16448 This metric is computed for package specifications and
16449 generic package declarations. It is the total number of types
16450 that can be referenced from outside this compilation unit, plus the
16451 number of types from all the visible parts of all the visible generic packages.
16452 Generic formal types are not counted. Only types, not subtypes,
16456 Along with the total number of public types, the following
16457 types are counted and reported separately:
16464 Root tagged types (abstract, non-abstract, private, non-private). Type
16465 extensions are @emph{not} counted
16468 Private types (including private extensions)
16479 This metric is computed for any compilation unit. It is equal to the total
16480 number of the declarations of different types given in the compilation unit.
16481 The private and the corresponding full type declaration are counted as one
16482 type declaration. Incomplete type declarations and generic formal types
16484 No distinction is made among different kinds of types (abstract,
16485 private etc.); the total number of types is computed and reported.
16490 By default, all the syntax metrics are computed and reported. You can use the
16491 following switches to select specific syntax metrics;
16492 if any of these is set, only the explicitly specified metrics are computed.
16495 @cindex @option{^-ed^/DECLARATION_TOTAL^} (@command{gnatmetric})
16496 @item ^-ed^/DECLARATION_TOTAL^
16497 The total number of declarations
16499 @cindex @option{^-es^/STATEMENT_TOTAL^} (@command{gnatmetric})
16500 @item ^-es^/STATEMENT_TOTAL^
16501 The total number of statements
16503 @cindex @option{^-eps^/^} (@command{gnatmetric})
16504 @item ^-eps^/INT_SUBPROGRAMS^
16505 The number of public subprograms in a compilation unit
16507 @cindex @option{^-eas^/SUBPROGRAMS_ALL^} (@command{gnatmetric})
16508 @item ^-eas^/SUBPROGRAMS_ALL^
16509 The number of all the subprograms in a compilation unit
16511 @cindex @option{^-ept^/INT_TYPES^} (@command{gnatmetric})
16512 @item ^-ept^/INT_TYPES^
16513 The number of public types in a compilation unit
16515 @cindex @option{^-eat^/TYPES_ALL^} (@command{gnatmetric})
16516 @item ^-eat^/TYPES_ALL^
16517 The number of all the types in a compilation unit
16519 @cindex @option{^-enu^/PROGRAM_NESTING_MAX^} (@command{gnatmetric})
16520 @item ^-enu^/PROGRAM_NESTING_MAX^
16521 The maximal program unit nesting level
16523 @cindex @option{^-ec^/CONSTRUCT_NESTING_MAX^} (@command{gnatmetric})
16524 @item ^-ec^/CONSTRUCT_NESTING_MAX^
16525 The maximal construct nesting level
16529 @node Complexity Metrics Control
16530 @subsection Complexity Metrics Control
16531 @cindex Complexity metrics control in @command{gnatmetric}
16534 For a program unit that is an executable body (a subprogram body (including
16535 generic bodies), task body, entry body or a package body containing
16536 its own statement sequence ) @command{gnatmetric} computes the following
16537 complexity metrics:
16541 McCabe cyclomatic complexity;
16544 McCabe essential complexity;
16547 maximal loop nesting level
16552 The McCabe complexity metrics are defined
16553 in @url{http://www.mccabe.com/pdf/nist235r.pdf}
16555 According to McCabe, both control statements and short-circuit control forms
16556 should be taken into account when computing cyclomatic complexity. For each
16557 body, we compute three metric values:
16561 the complexity introduced by control
16562 statements only, without taking into account short-circuit forms,
16565 the complexity introduced by short-circuit control forms only, and
16569 cyclomatic complexity, which is the sum of these two values.
16573 When computing cyclomatic and essential complexity, @command{gnatmetric} skips
16574 the code in the exception handlers and in all the nested program units.
16576 By default, all the complexity metrics are computed and reported.
16577 For more finely-grained control you can use
16578 the following switches:
16581 @cindex @option{^-n@var{x}^/SUPPRESS^} (@command{gnatmetric})
16583 @item ^-nocc^/SUPPRESS=CYCLOMATIC_COMPLEXITY^
16584 Do not compute the McCabe Cyclomatic Complexity
16586 @item ^-noec^/SUPPRESS=ESSENTIAL_COMPLEXITY^
16587 Do not compute the Essential Complexity
16589 @item ^-nonl^/SUPPRESS=MAXIMAL_LOOP_NESTING^
16590 Do not compute maximal loop nesting level
16592 @item ^-ne^/SUPPRESS=EXITS_AS_GOTOS^
16593 Do not consider @code{exit} statements as @code{goto}s when
16594 computing Essential Complexity
16598 @node Other gnatmetric Switches
16599 @subsection Other @code{gnatmetric} Switches
16602 Additional @command{gnatmetric} switches are as follows:
16605 @item ^-files @var{filename}^/FILES=@var{filename}^
16606 @cindex @option{^-files^/FILES^} (@code{gnatmetric})
16607 Take the argument source files from the specified file. This file should be an
16608 ordinary textual file containing file names separated by spaces or
16609 line breaks. You can use this switch more then once in the same call to
16610 @command{gnatmetric}. You also can combine this switch with
16611 an explicit list of files.
16613 @item ^-v^/VERBOSE^
16614 @cindex @option{^-v^/VERBOSE^} (@code{gnatmetric})
16616 @command{gnatmetric} generates version information and then
16617 a trace of sources being processed.
16619 @item ^-dv^/DEBUG_OUTPUT^
16620 @cindex @option{^-dv^/DEBUG_OUTPUT^} (@code{gnatmetric})
16622 @command{gnatmetric} generates various messages useful to understand what
16623 happens during the metrics computation
16626 @cindex @option{^-q^/QUIET^} (@code{gnatmetric})
16630 @node Generate project-wide metrics
16631 @subsection Generate project-wide metrics
16633 In order to compute metrics on all units of a given project, one can use
16634 the @command{gnat} driver along with the @option{-P} option:
16638 If the project @code{proj} depends upon other projects, one can compute
16639 the metrics on the project closure using the @option{-U} option:
16641 gnat metric -Pproj -U
16643 Finally, if not all the units are relevant to a particular main
16644 program in the project closure, one can generate metrics for the set
16645 of units needed to create a given main program (unit closure) using
16646 the @option{-U} option followed by the name of the main unit:
16648 gnat metric -Pproj -U main
16652 @c ***********************************
16653 @node File Name Krunching Using gnatkr
16654 @chapter File Name Krunching Using @code{gnatkr}
16658 This chapter discusses the method used by the compiler to shorten
16659 the default file names chosen for Ada units so that they do not
16660 exceed the maximum length permitted. It also describes the
16661 @code{gnatkr} utility that can be used to determine the result of
16662 applying this shortening.
16666 * Krunching Method::
16667 * Examples of gnatkr Usage::
16671 @section About @code{gnatkr}
16674 The default file naming rule in GNAT
16675 is that the file name must be derived from
16676 the unit name. The exact default rule is as follows:
16679 Take the unit name and replace all dots by hyphens.
16681 If such a replacement occurs in the
16682 second character position of a name, and the first character is
16683 ^a, g, s, or i^A, G, S, or I^ then replace the dot by the character
16684 ^~ (tilde)^$ (dollar sign)^
16685 instead of a minus.
16687 The reason for this exception is to avoid clashes
16688 with the standard names for children of System, Ada, Interfaces,
16689 and GNAT, which use the prefixes ^s- a- i- and g-^S- A- I- and G-^
16692 The @option{^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{nn}}
16693 switch of the compiler activates a ``krunching''
16694 circuit that limits file names to nn characters (where nn is a decimal
16695 integer). For example, using OpenVMS,
16696 where the maximum file name length is
16697 39, the value of nn is usually set to 39, but if you want to generate
16698 a set of files that would be usable if ported to a system with some
16699 different maximum file length, then a different value can be specified.
16700 The default value of 39 for OpenVMS need not be specified.
16702 The @code{gnatkr} utility can be used to determine the krunched name for
16703 a given file, when krunched to a specified maximum length.
16706 @section Using @code{gnatkr}
16709 The @code{gnatkr} command has the form
16713 $ gnatkr @var{name} [@var{length}]
16719 $ gnatkr @var{name} /COUNT=nn
16724 @var{name} is the uncrunched file name, derived from the name of the unit
16725 in the standard manner described in the previous section (i.e. in particular
16726 all dots are replaced by hyphens). The file name may or may not have an
16727 extension (defined as a suffix of the form period followed by arbitrary
16728 characters other than period). If an extension is present then it will
16729 be preserved in the output. For example, when krunching @file{hellofile.ads}
16730 to eight characters, the result will be hellofil.ads.
16732 Note: for compatibility with previous versions of @code{gnatkr} dots may
16733 appear in the name instead of hyphens, but the last dot will always be
16734 taken as the start of an extension. So if @code{gnatkr} is given an argument
16735 such as @file{Hello.World.adb} it will be treated exactly as if the first
16736 period had been a hyphen, and for example krunching to eight characters
16737 gives the result @file{hellworl.adb}.
16739 Note that the result is always all lower case (except on OpenVMS where it is
16740 all upper case). Characters of the other case are folded as required.
16742 @var{length} represents the length of the krunched name. The default
16743 when no argument is given is ^8^39^ characters. A length of zero stands for
16744 unlimited, in other words do not chop except for system files where the
16745 implied crunching length is always eight characters.
16748 The output is the krunched name. The output has an extension only if the
16749 original argument was a file name with an extension.
16751 @node Krunching Method
16752 @section Krunching Method
16755 The initial file name is determined by the name of the unit that the file
16756 contains. The name is formed by taking the full expanded name of the
16757 unit and replacing the separating dots with hyphens and
16758 using ^lowercase^uppercase^
16759 for all letters, except that a hyphen in the second character position is
16760 replaced by a ^tilde^dollar sign^ if the first character is
16761 ^a, i, g, or s^A, I, G, or S^.
16762 The extension is @code{.ads} for a
16763 specification and @code{.adb} for a body.
16764 Krunching does not affect the extension, but the file name is shortened to
16765 the specified length by following these rules:
16769 The name is divided into segments separated by hyphens, tildes or
16770 underscores and all hyphens, tildes, and underscores are
16771 eliminated. If this leaves the name short enough, we are done.
16774 If the name is too long, the longest segment is located (left-most
16775 if there are two of equal length), and shortened by dropping
16776 its last character. This is repeated until the name is short enough.
16778 As an example, consider the krunching of @*@file{our-strings-wide_fixed.adb}
16779 to fit the name into 8 characters as required by some operating systems.
16782 our-strings-wide_fixed 22
16783 our strings wide fixed 19
16784 our string wide fixed 18
16785 our strin wide fixed 17
16786 our stri wide fixed 16
16787 our stri wide fixe 15
16788 our str wide fixe 14
16789 our str wid fixe 13
16795 Final file name: oustwifi.adb
16799 The file names for all predefined units are always krunched to eight
16800 characters. The krunching of these predefined units uses the following
16801 special prefix replacements:
16805 replaced by @file{^a^A^-}
16808 replaced by @file{^g^G^-}
16811 replaced by @file{^i^I^-}
16814 replaced by @file{^s^S^-}
16817 These system files have a hyphen in the second character position. That
16818 is why normal user files replace such a character with a
16819 ^tilde^dollar sign^, to
16820 avoid confusion with system file names.
16822 As an example of this special rule, consider
16823 @*@file{ada-strings-wide_fixed.adb}, which gets krunched as follows:
16826 ada-strings-wide_fixed 22
16827 a- strings wide fixed 18
16828 a- string wide fixed 17
16829 a- strin wide fixed 16
16830 a- stri wide fixed 15
16831 a- stri wide fixe 14
16832 a- str wide fixe 13
16838 Final file name: a-stwifi.adb
16842 Of course no file shortening algorithm can guarantee uniqueness over all
16843 possible unit names, and if file name krunching is used then it is your
16844 responsibility to ensure that no name clashes occur. The utility
16845 program @code{gnatkr} is supplied for conveniently determining the
16846 krunched name of a file.
16848 @node Examples of gnatkr Usage
16849 @section Examples of @code{gnatkr} Usage
16856 $ gnatkr very_long_unit_name.ads --> velounna.ads
16857 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
16858 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
16859 $ gnatkr grandparent-parent-child --> grparchi
16861 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
16862 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
16865 @node Preprocessing Using gnatprep
16866 @chapter Preprocessing Using @code{gnatprep}
16870 The @code{gnatprep} utility provides
16871 a simple preprocessing capability for Ada programs.
16872 It is designed for use with GNAT, but is not dependent on any special
16877 * Switches for gnatprep::
16878 * Form of Definitions File::
16879 * Form of Input Text for gnatprep::
16882 @node Using gnatprep
16883 @section Using @code{gnatprep}
16886 To call @code{gnatprep} use
16889 $ gnatprep [switches] infile outfile [deffile]
16896 is an optional sequence of switches as described in the next section.
16899 is the full name of the input file, which is an Ada source
16900 file containing preprocessor directives.
16903 is the full name of the output file, which is an Ada source
16904 in standard Ada form. When used with GNAT, this file name will
16905 normally have an ads or adb suffix.
16908 is the full name of a text file containing definitions of
16909 symbols to be referenced by the preprocessor. This argument is
16910 optional, and can be replaced by the use of the @option{-D} switch.
16914 @node Switches for gnatprep
16915 @section Switches for @code{gnatprep}
16920 @item ^-b^/BLANK_LINES^
16921 @cindex @option{^-b^/BLANK_LINES^} (@command{gnatprep})
16922 Causes both preprocessor lines and the lines deleted by
16923 preprocessing to be replaced by blank lines in the output source file,
16924 preserving line numbers in the output file.
16926 @item ^-c^/COMMENTS^
16927 @cindex @option{^-c^/COMMENTS^} (@command{gnatprep})
16928 Causes both preprocessor lines and the lines deleted
16929 by preprocessing to be retained in the output source as comments marked
16930 with the special string @code{"--! "}. This option will result in line numbers
16931 being preserved in the output file.
16933 @item ^-C^/REPLACE_IN_COMMENTS^
16934 @cindex @option{^-C^/REPLACE_IN_COMMENTS^} (@command{gnatprep})
16935 Causes comments to be scanned. Normally comments are ignored by gnatprep.
16936 If this option is specified, then comments are scanned and any $symbol
16937 substitutions performed as in program text. This is particularly useful
16938 when structured comments are used (e.g. when writing programs in the
16939 SPARK dialect of Ada). Note that this switch is not available when
16940 doing integrated preprocessing (it would be useless in this context
16941 since comments are ignored by the compiler in any case).
16943 @item ^-Dsymbol=value^/ASSOCIATE="symbol=value"^
16944 @cindex @option{^-D^/ASSOCIATE^} (@command{gnatprep})
16945 Defines a new symbol, associated with value. If no value is given on the
16946 command line, then symbol is considered to be @code{True}. This switch
16947 can be used in place of a definition file.
16951 @cindex @option{/REMOVE} (@command{gnatprep})
16952 This is the default setting which causes lines deleted by preprocessing
16953 to be entirely removed from the output file.
16956 @item ^-r^/REFERENCE^
16957 @cindex @option{^-r^/REFERENCE^} (@command{gnatprep})
16958 Causes a @code{Source_Reference} pragma to be generated that
16959 references the original input file, so that error messages will use
16960 the file name of this original file. The use of this switch implies
16961 that preprocessor lines are not to be removed from the file, so its
16962 use will force @option{^-b^/BLANK_LINES^} mode if
16963 @option{^-c^/COMMENTS^}
16964 has not been specified explicitly.
16966 Note that if the file to be preprocessed contains multiple units, then
16967 it will be necessary to @code{gnatchop} the output file from
16968 @code{gnatprep}. If a @code{Source_Reference} pragma is present
16969 in the preprocessed file, it will be respected by
16970 @code{gnatchop ^-r^/REFERENCE^}
16971 so that the final chopped files will correctly refer to the original
16972 input source file for @code{gnatprep}.
16974 @item ^-s^/SYMBOLS^
16975 @cindex @option{^-s^/SYMBOLS^} (@command{gnatprep})
16976 Causes a sorted list of symbol names and values to be
16977 listed on the standard output file.
16979 @item ^-u^/UNDEFINED^
16980 @cindex @option{^-u^/UNDEFINED^} (@command{gnatprep})
16981 Causes undefined symbols to be treated as having the value FALSE in the context
16982 of a preprocessor test. In the absence of this option, an undefined symbol in
16983 a @code{#if} or @code{#elsif} test will be treated as an error.
16989 Note: if neither @option{-b} nor @option{-c} is present,
16990 then preprocessor lines and
16991 deleted lines are completely removed from the output, unless -r is
16992 specified, in which case -b is assumed.
16995 @node Form of Definitions File
16996 @section Form of Definitions File
16999 The definitions file contains lines of the form
17006 where symbol is an identifier, following normal Ada (case-insensitive)
17007 rules for its syntax, and value is one of the following:
17011 Empty, corresponding to a null substitution
17013 A string literal using normal Ada syntax
17015 Any sequence of characters from the set
17016 (letters, digits, period, underline).
17020 Comment lines may also appear in the definitions file, starting with
17021 the usual @code{--},
17022 and comments may be added to the definitions lines.
17024 @node Form of Input Text for gnatprep
17025 @section Form of Input Text for @code{gnatprep}
17028 The input text may contain preprocessor conditional inclusion lines,
17029 as well as general symbol substitution sequences.
17031 The preprocessor conditional inclusion commands have the form
17036 #if @i{expression} [then]
17038 #elsif @i{expression} [then]
17040 #elsif @i{expression} [then]
17051 In this example, @i{expression} is defined by the following grammar:
17053 @i{expression} ::= <symbol>
17054 @i{expression} ::= <symbol> = "<value>"
17055 @i{expression} ::= <symbol> = <symbol>
17056 @i{expression} ::= <symbol> 'Defined
17057 @i{expression} ::= not @i{expression}
17058 @i{expression} ::= @i{expression} and @i{expression}
17059 @i{expression} ::= @i{expression} or @i{expression}
17060 @i{expression} ::= @i{expression} and then @i{expression}
17061 @i{expression} ::= @i{expression} or else @i{expression}
17062 @i{expression} ::= ( @i{expression} )
17066 For the first test (@i{expression} ::= <symbol>) the symbol must have
17067 either the value true or false, that is to say the right-hand of the
17068 symbol definition must be one of the (case-insensitive) literals
17069 @code{True} or @code{False}. If the value is true, then the
17070 corresponding lines are included, and if the value is false, they are
17073 The test (@i{expression} ::= <symbol> @code{'Defined}) is true only if
17074 the symbol has been defined in the definition file or by a @option{-D}
17075 switch on the command line. Otherwise, the test is false.
17077 The equality tests are case insensitive, as are all the preprocessor lines.
17079 If the symbol referenced is not defined in the symbol definitions file,
17080 then the effect depends on whether or not switch @option{-u}
17081 is specified. If so, then the symbol is treated as if it had the value
17082 false and the test fails. If this switch is not specified, then
17083 it is an error to reference an undefined symbol. It is also an error to
17084 reference a symbol that is defined with a value other than @code{True}
17087 The use of the @code{not} operator inverts the sense of this logical test.
17088 The @code{not} operator cannot be combined with the @code{or} or @code{and}
17089 operators, without parentheses. For example, "if not X or Y then" is not
17090 allowed, but "if (not X) or Y then" and "if not (X or Y) then" are.
17092 The @code{then} keyword is optional as shown
17094 The @code{#} must be the first non-blank character on a line, but
17095 otherwise the format is free form. Spaces or tabs may appear between
17096 the @code{#} and the keyword. The keywords and the symbols are case
17097 insensitive as in normal Ada code. Comments may be used on a
17098 preprocessor line, but other than that, no other tokens may appear on a
17099 preprocessor line. Any number of @code{elsif} clauses can be present,
17100 including none at all. The @code{else} is optional, as in Ada.
17102 The @code{#} marking the start of a preprocessor line must be the first
17103 non-blank character on the line, i.e. it must be preceded only by
17104 spaces or horizontal tabs.
17106 Symbol substitution outside of preprocessor lines is obtained by using
17114 anywhere within a source line, except in a comment or within a
17115 string literal. The identifier
17116 following the @code{$} must match one of the symbols defined in the symbol
17117 definition file, and the result is to substitute the value of the
17118 symbol in place of @code{$symbol} in the output file.
17120 Note that although the substitution of strings within a string literal
17121 is not possible, it is possible to have a symbol whose defined value is
17122 a string literal. So instead of setting XYZ to @code{hello} and writing:
17125 Header : String := "$XYZ";
17129 you should set XYZ to @code{"hello"} and write:
17132 Header : String := $XYZ;
17136 and then the substitution will occur as desired.
17139 @node The GNAT Run-Time Library Builder gnatlbr
17140 @chapter The GNAT Run-Time Library Builder @code{gnatlbr}
17142 @cindex Library builder
17145 @code{gnatlbr} is a tool for rebuilding the GNAT run time with user
17146 supplied configuration pragmas.
17149 * Running gnatlbr::
17150 * Switches for gnatlbr::
17151 * Examples of gnatlbr Usage::
17154 @node Running gnatlbr
17155 @section Running @code{gnatlbr}
17158 The @code{gnatlbr} command has the form
17161 $ GNAT LIBRARY /[CREATE | SET | DELETE]=directory [/CONFIG=file]
17164 @node Switches for gnatlbr
17165 @section Switches for @code{gnatlbr}
17168 @code{gnatlbr} recognizes the following switches:
17172 @item /CREATE=directory
17173 @cindex @code{/CREATE} (@code{gnatlbr})
17174 Create the new run-time library in the specified directory.
17176 @item /SET=directory
17177 @cindex @code{/SET} (@code{gnatlbr})
17178 Make the library in the specified directory the current run-time
17181 @item /DELETE=directory
17182 @cindex @code{/DELETE} (@code{gnatlbr})
17183 Delete the run-time library in the specified directory.
17186 @cindex @code{/CONFIG} (@code{gnatlbr})
17188 Use the configuration pragmas in the specified file when building
17192 Use the configuration pragmas in the specified file when compiling.
17196 @node Examples of gnatlbr Usage
17197 @section Example of @code{gnatlbr} Usage
17200 Contents of VAXFLOAT.ADC:
17201 pragma Float_Representation (VAX_Float);
17203 $ GNAT LIBRARY /CREATE=[.VAXFLOAT] /CONFIG=VAXFLOAT.ADC
17205 GNAT LIBRARY rebuilds the run-time library in directory [.VAXFLOAT]
17210 @node The GNAT Library Browser gnatls
17211 @chapter The GNAT Library Browser @code{gnatls}
17213 @cindex Library browser
17216 @code{gnatls} is a tool that outputs information about compiled
17217 units. It gives the relationship between objects, unit names and source
17218 files. It can also be used to check the source dependencies of a unit
17219 as well as various characteristics.
17221 Note: to invoke @code{gnatls} with a project file, use the @code{gnat}
17222 driver (see @ref{The GNAT Driver and Project Files}).
17226 * Switches for gnatls::
17227 * Examples of gnatls Usage::
17230 @node Running gnatls
17231 @section Running @code{gnatls}
17234 The @code{gnatls} command has the form
17237 $ gnatls switches @var{object_or_ali_file}
17241 The main argument is the list of object or @file{ali} files
17242 (@pxref{The Ada Library Information Files})
17243 for which information is requested.
17245 In normal mode, without additional option, @code{gnatls} produces a
17246 four-column listing. Each line represents information for a specific
17247 object. The first column gives the full path of the object, the second
17248 column gives the name of the principal unit in this object, the third
17249 column gives the status of the source and the fourth column gives the
17250 full path of the source representing this unit.
17251 Here is a simple example of use:
17255 ^./^[]^demo1.o demo1 DIF demo1.adb
17256 ^./^[]^demo2.o demo2 OK demo2.adb
17257 ^./^[]^hello.o h1 OK hello.adb
17258 ^./^[]^instr-child.o instr.child MOK instr-child.adb
17259 ^./^[]^instr.o instr OK instr.adb
17260 ^./^[]^tef.o tef DIF tef.adb
17261 ^./^[]^text_io_example.o text_io_example OK text_io_example.adb
17262 ^./^[]^tgef.o tgef DIF tgef.adb
17266 The first line can be interpreted as follows: the main unit which is
17268 object file @file{demo1.o} is demo1, whose main source is in
17269 @file{demo1.adb}. Furthermore, the version of the source used for the
17270 compilation of demo1 has been modified (DIF). Each source file has a status
17271 qualifier which can be:
17274 @item OK (unchanged)
17275 The version of the source file used for the compilation of the
17276 specified unit corresponds exactly to the actual source file.
17278 @item MOK (slightly modified)
17279 The version of the source file used for the compilation of the
17280 specified unit differs from the actual source file but not enough to
17281 require recompilation. If you use gnatmake with the qualifier
17282 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}, a file marked
17283 MOK will not be recompiled.
17285 @item DIF (modified)
17286 No version of the source found on the path corresponds to the source
17287 used to build this object.
17289 @item ??? (file not found)
17290 No source file was found for this unit.
17292 @item HID (hidden, unchanged version not first on PATH)
17293 The version of the source that corresponds exactly to the source used
17294 for compilation has been found on the path but it is hidden by another
17295 version of the same source that has been modified.
17299 @node Switches for gnatls
17300 @section Switches for @code{gnatls}
17303 @code{gnatls} recognizes the following switches:
17307 @item ^-a^/ALL_UNITS^
17308 @cindex @option{^-a^/ALL_UNITS^} (@code{gnatls})
17309 Consider all units, including those of the predefined Ada library.
17310 Especially useful with @option{^-d^/DEPENDENCIES^}.
17312 @item ^-d^/DEPENDENCIES^
17313 @cindex @option{^-d^/DEPENDENCIES^} (@code{gnatls})
17314 List sources from which specified units depend on.
17316 @item ^-h^/OUTPUT=OPTIONS^
17317 @cindex @option{^-h^/OUTPUT=OPTIONS^} (@code{gnatls})
17318 Output the list of options.
17320 @item ^-o^/OUTPUT=OBJECTS^
17321 @cindex @option{^-o^/OUTPUT=OBJECTS^} (@code{gnatls})
17322 Only output information about object files.
17324 @item ^-s^/OUTPUT=SOURCES^
17325 @cindex @option{^-s^/OUTPUT=SOURCES^} (@code{gnatls})
17326 Only output information about source files.
17328 @item ^-u^/OUTPUT=UNITS^
17329 @cindex @option{^-u^/OUTPUT=UNITS^} (@code{gnatls})
17330 Only output information about compilation units.
17332 @item ^-files^/FILES^=@var{file}
17333 @cindex @option{^-files^/FILES^} (@code{gnatls})
17334 Take as arguments the files listed in text file @var{file}.
17335 Text file @var{file} may contain empty lines that are ignored.
17336 Each non empty line should contain the name of an existing file.
17337 Several such switches may be specified simultaneously.
17339 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
17340 @itemx ^-aI^/SOURCE_SEARCH=^@var{dir}
17341 @itemx ^-I^/SEARCH=^@var{dir}
17342 @itemx ^-I-^/NOCURRENT_DIRECTORY^
17344 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatls})
17345 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatls})
17346 @cindex @option{^-I^/SEARCH^} (@code{gnatls})
17347 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatls})
17348 Source path manipulation. Same meaning as the equivalent @command{gnatmake}
17349 flags (@pxref{Switches for gnatmake}).
17351 @item --RTS=@var{rts-path}
17352 @cindex @option{--RTS} (@code{gnatls})
17353 Specifies the default location of the runtime library. Same meaning as the
17354 equivalent @command{gnatmake} flag (@pxref{Switches for gnatmake}).
17356 @item ^-v^/OUTPUT=VERBOSE^
17357 @cindex @option{^-v^/OUTPUT=VERBOSE^} (@code{gnatls})
17358 Verbose mode. Output the complete source, object and project paths. Do not use
17359 the default column layout but instead use long format giving as much as
17360 information possible on each requested units, including special
17361 characteristics such as:
17364 @item Preelaborable
17365 The unit is preelaborable in the Ada sense.
17368 No elaboration code has been produced by the compiler for this unit.
17371 The unit is pure in the Ada sense.
17373 @item Elaborate_Body
17374 The unit contains a pragma Elaborate_Body.
17377 The unit contains a pragma Remote_Types.
17379 @item Shared_Passive
17380 The unit contains a pragma Shared_Passive.
17383 This unit is part of the predefined environment and cannot be modified
17386 @item Remote_Call_Interface
17387 The unit contains a pragma Remote_Call_Interface.
17393 @node Examples of gnatls Usage
17394 @section Example of @code{gnatls} Usage
17398 Example of using the verbose switch. Note how the source and
17399 object paths are affected by the -I switch.
17402 $ gnatls -v -I.. demo1.o
17404 GNATLS 5.03w (20041123-34)
17405 Copyright 1997-2004 Free Software Foundation, Inc.
17407 Source Search Path:
17408 <Current_Directory>
17410 /home/comar/local/adainclude/
17412 Object Search Path:
17413 <Current_Directory>
17415 /home/comar/local/lib/gcc-lib/x86-linux/3.4.3/adalib/
17417 Project Search Path:
17418 <Current_Directory>
17419 /home/comar/local/lib/gnat/
17424 Kind => subprogram body
17425 Flags => No_Elab_Code
17426 Source => demo1.adb modified
17430 The following is an example of use of the dependency list.
17431 Note the use of the -s switch
17432 which gives a straight list of source files. This can be useful for
17433 building specialized scripts.
17436 $ gnatls -d demo2.o
17437 ./demo2.o demo2 OK demo2.adb
17443 $ gnatls -d -s -a demo1.o
17445 /home/comar/local/adainclude/ada.ads
17446 /home/comar/local/adainclude/a-finali.ads
17447 /home/comar/local/adainclude/a-filico.ads
17448 /home/comar/local/adainclude/a-stream.ads
17449 /home/comar/local/adainclude/a-tags.ads
17452 /home/comar/local/adainclude/gnat.ads
17453 /home/comar/local/adainclude/g-io.ads
17455 /home/comar/local/adainclude/system.ads
17456 /home/comar/local/adainclude/s-exctab.ads
17457 /home/comar/local/adainclude/s-finimp.ads
17458 /home/comar/local/adainclude/s-finroo.ads
17459 /home/comar/local/adainclude/s-secsta.ads
17460 /home/comar/local/adainclude/s-stalib.ads
17461 /home/comar/local/adainclude/s-stoele.ads
17462 /home/comar/local/adainclude/s-stratt.ads
17463 /home/comar/local/adainclude/s-tasoli.ads
17464 /home/comar/local/adainclude/s-unstyp.ads
17465 /home/comar/local/adainclude/unchconv.ads
17471 GNAT LIST /DEPENDENCIES /OUTPUT=SOURCES /ALL_UNITS DEMO1.ADB
17473 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]ada.ads
17474 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-finali.ads
17475 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-filico.ads
17476 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-stream.ads
17477 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-tags.ads
17481 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]gnat.ads
17482 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]g-io.ads
17484 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]system.ads
17485 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-exctab.ads
17486 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finimp.ads
17487 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finroo.ads
17488 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-secsta.ads
17489 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stalib.ads
17490 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stoele.ads
17491 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stratt.ads
17492 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-tasoli.ads
17493 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-unstyp.ads
17494 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]unchconv.ads
17498 @node Cleaning Up Using gnatclean
17499 @chapter Cleaning Up Using @code{gnatclean}
17501 @cindex Cleaning tool
17504 @code{gnatclean} is a tool that allows the deletion of files produced by the
17505 compiler, binder and linker, including ALI files, object files, tree files,
17506 expanded source files, library files, interface copy source files, binder
17507 generated files and executable files.
17510 * Running gnatclean::
17511 * Switches for gnatclean::
17512 @c * Examples of gnatclean Usage::
17515 @node Running gnatclean
17516 @section Running @code{gnatclean}
17519 The @code{gnatclean} command has the form:
17522 $ gnatclean switches @var{names}
17526 @var{names} is a list of source file names. Suffixes @code{.^ads^ADS^} and
17527 @code{^adb^ADB^} may be omitted. If a project file is specified using switch
17528 @code{^-P^/PROJECT_FILE=^}, then @var{names} may be completely omitted.
17531 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
17532 if switch @code{^-c^/COMPILER_FILES_ONLY^} is not specified, by the binder and
17533 the linker. In informative-only mode, specified by switch
17534 @code{^-n^/NODELETE^}, the list of files that would have been deleted in
17535 normal mode is listed, but no file is actually deleted.
17537 @node Switches for gnatclean
17538 @section Switches for @code{gnatclean}
17541 @code{gnatclean} recognizes the following switches:
17545 @item ^-c^/COMPILER_FILES_ONLY^
17546 @cindex @option{^-c^/COMPILER_FILES_ONLY^} (@code{gnatclean})
17547 Only attempt to delete the files produced by the compiler, not those produced
17548 by the binder or the linker. The files that are not to be deleted are library
17549 files, interface copy files, binder generated files and executable files.
17551 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
17552 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatclean})
17553 Indicate that ALI and object files should normally be found in directory
17556 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
17557 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatclean})
17558 When using project files, if some errors or warnings are detected during
17559 parsing and verbose mode is not in effect (no use of switch
17560 ^-v^/VERBOSE^), then error lines start with the full path name of the project
17561 file, rather than its simple file name.
17564 @cindex @option{^-h^/HELP^} (@code{gnatclean})
17565 Output a message explaining the usage of @code{^gnatclean^gnatclean^}.
17567 @item ^-n^/NODELETE^
17568 @cindex @option{^-n^/NODELETE^} (@code{gnatclean})
17569 Informative-only mode. Do not delete any files. Output the list of the files
17570 that would have been deleted if this switch was not specified.
17572 @item ^-P^/PROJECT_FILE=^@var{project}
17573 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatclean})
17574 Use project file @var{project}. Only one such switch can be used.
17575 When cleaning a project file, the files produced by the compilation of the
17576 immediate sources or inherited sources of the project files are to be
17577 deleted. This is not depending on the presence or not of executable names
17578 on the command line.
17581 @cindex @option{^-q^/QUIET^} (@code{gnatclean})
17582 Quiet output. If there are no errors, do not output anything, except in
17583 verbose mode (switch ^-v^/VERBOSE^) or in informative-only mode
17584 (switch ^-n^/NODELETE^).
17586 @item ^-r^/RECURSIVE^
17587 @cindex @option{^-r^/RECURSIVE^} (@code{gnatclean})
17588 When a project file is specified (using switch ^-P^/PROJECT_FILE=^),
17589 clean all imported and extended project files, recursively. If this switch
17590 is not specified, only the files related to the main project file are to be
17591 deleted. This switch has no effect if no project file is specified.
17593 @item ^-v^/VERBOSE^
17594 @cindex @option{^-v^/VERBOSE^} (@code{gnatclean})
17597 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
17598 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (@code{gnatclean})
17599 Indicates the verbosity of the parsing of GNAT project files.
17600 @xref{Switches Related to Project Files}.
17602 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
17603 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (@code{gnatclean})
17604 Indicates that external variable @var{name} has the value @var{value}.
17605 The Project Manager will use this value for occurrences of
17606 @code{external(name)} when parsing the project file.
17607 @xref{Switches Related to Project Files}.
17609 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
17610 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatclean})
17611 When searching for ALI and object files, look in directory
17614 @item ^-I^/SEARCH=^@var{dir}
17615 @cindex @option{^-I^/SEARCH^} (@code{gnatclean})
17616 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}}.
17618 @item ^-I-^/NOCURRENT_DIRECTORY^
17619 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatclean})
17620 @cindex Source files, suppressing search
17621 Do not look for ALI or object files in the directory
17622 where @code{gnatclean} was invoked.
17626 @c @node Examples of gnatclean Usage
17627 @c @section Examples of @code{gnatclean} Usage
17630 @node GNAT and Libraries
17631 @chapter GNAT and Libraries
17632 @cindex Library, building, installing, using
17635 This chapter describes how to build and use libraries with GNAT, and also shows
17636 how to recompile the GNAT run-time library. You should be familiar with the
17637 Project Manager facility (@pxref{GNAT Project Manager}) before reading this
17641 * Introduction to Libraries in GNAT::
17642 * General Ada Libraries::
17643 * Stand-alone Ada Libraries::
17644 * Rebuilding the GNAT Run-Time Library::
17647 @node Introduction to Libraries in GNAT
17648 @section Introduction to Libraries in GNAT
17651 A library is, conceptually, a collection of objects which does not have its
17652 own main thread of execution, but rather provides certain services to the
17653 applications that use it. A library can be either statically linked with the
17654 application, in which case its code is directly included in the application,
17655 or, on platforms that support it, be dynamically linked, in which case
17656 its code is shared by all applications making use of this library.
17658 GNAT supports both types of libraries.
17659 In the static case, the compiled code can be provided in different ways. The
17660 simplest approach is to provide directly the set of objects resulting from
17661 compilation of the library source files. Alternatively, you can group the
17662 objects into an archive using whatever commands are provided by the operating
17663 system. For the latter case, the objects are grouped into a shared library.
17665 In the GNAT environment, a library has three types of components:
17671 @xref{The Ada Library Information Files}.
17673 Object files, an archive or a shared library.
17677 A GNAT library may expose all its source files, which is useful for
17678 documentation purposes. Alternatively, it may expose only the units needed by
17679 an external user to make use of the library. That is to say, the specs
17680 reflecting the library services along with all the units needed to compile
17681 those specs, which can include generic bodies or any body implementing an
17682 inlined routine. In the case of @emph{stand-alone libraries} those exposed
17683 units are called @emph{interface units} (@pxref{Stand-alone Ada Libraries}).
17685 All compilation units comprising an application, including those in a library,
17686 need to be elaborated in an order partially defined by Ada's semantics. GNAT
17687 computes the elaboration order from the @file{ALI} files and this is why they
17688 constitute a mandatory part of GNAT libraries. Except in the case of
17689 @emph{stand-alone libraries}, where a specific library elaboration routine is
17690 produced independently of the application(s) using the library.
17692 @node General Ada Libraries
17693 @section General Ada Libraries
17696 * Building a library::
17697 * Installing a library::
17698 * Using a library::
17701 @node Building a library
17702 @subsection Building a library
17705 The easiest way to build a library is to use the Project Manager,
17706 which supports a special type of project called a @emph{Library Project}
17707 (@pxref{Library Projects}).
17709 A project is considered a library project, when two project-level attributes
17710 are defined in it: @code{Library_Name} and @code{Library_Dir}. In order to
17711 control different aspects of library configuration, additional optional
17712 project-level attributes can be specified:
17715 This attribute controls whether the library is to be static or dynamic
17717 @item Library_Version
17718 This attribute specifies the library version; this value is used
17719 during dynamic linking of shared libraries to determine if the currently
17720 installed versions of the binaries are compatible.
17722 @item Library_Options
17724 These attributes specify additional low-level options to be used during
17725 library generation, and redefine the actual application used to generate
17730 The GNAT Project Manager takes full care of the library maintenance task,
17731 including recompilation of the source files for which objects do not exist
17732 or are not up to date, assembly of the library archive, and installation of
17733 the library (i.e., copying associated source, object and @file{ALI} files
17734 to the specified location).
17736 Here is a simple library project file:
17737 @smallexample @c ada
17739 for Source_Dirs use ("src1", "src2");
17740 for Object_Dir use "obj";
17741 for Library_Name use "mylib";
17742 for Library_Dir use "lib";
17743 for Library_Kind use "dynamic";
17748 and the compilation command to build and install the library:
17750 @smallexample @c ada
17751 $ gnatmake -Pmy_lib
17755 It is not entirely trivial to perform manually all the steps required to
17756 produce a library. We recommend that you use the GNAT Project Manager
17757 for this task. In special cases where this is not desired, the necessary
17758 steps are discussed below.
17760 There are various possibilities for compiling the units that make up the
17761 library: for example with a Makefile (@pxref{Using the GNU make Utility}) or
17762 with a conventional script. For simple libraries, it is also possible to create
17763 a dummy main program which depends upon all the packages that comprise the
17764 interface of the library. This dummy main program can then be given to
17765 @command{gnatmake}, which will ensure that all necessary objects are built.
17767 After this task is accomplished, you should follow the standard procedure
17768 of the underlying operating system to produce the static or shared library.
17770 Here is an example of such a dummy program:
17771 @smallexample @c ada
17773 with My_Lib.Service1;
17774 with My_Lib.Service2;
17775 with My_Lib.Service3;
17776 procedure My_Lib_Dummy is
17784 Here are the generic commands that will build an archive or a shared library.
17787 # compiling the library
17788 $ gnatmake -c my_lib_dummy.adb
17790 # we don't need the dummy object itself
17791 $ rm my_lib_dummy.o my_lib_dummy.ali
17793 # create an archive with the remaining objects
17794 $ ar rc libmy_lib.a *.o
17795 # some systems may require "ranlib" to be run as well
17797 # or create a shared library
17798 $ gcc -shared -o libmy_lib.so *.o
17799 # some systems may require the code to have been compiled with -fPIC
17801 # remove the object files that are now in the library
17804 # Make the ALI files read-only so that gnatmake will not try to
17805 # regenerate the objects that are in the library
17810 Please note that the library must have a name of the form @file{libxxx.a} or
17811 @file{libxxx.so} (or @file{libxxx.dll} on Windows) in order to be accessed by
17812 the directive @option{-lxxx} at link time.
17814 @node Installing a library
17815 @subsection Installing a library
17816 @cindex @code{ADA_PROJECT_PATH}
17819 If you use project files, library installation is part of the library build
17820 process. Thus no further action is needed in order to make use of the
17821 libraries that are built as part of the general application build. A usable
17822 version of the library is installed in the directory specified by the
17823 @code{Library_Dir} attribute of the library project file.
17825 You may want to install a library in a context different from where the library
17826 is built. This situation arises with third party suppliers, who may want
17827 to distribute a library in binary form where the user is not expected to be
17828 able to recompile the library. The simplest option in this case is to provide
17829 a project file slightly different from the one used to build the library, by
17830 using the @code{externally_built} attribute. For instance, the project
17831 file used to build the library in the previous section can be changed into the
17832 following one when the library is installed:
17834 @smallexample @c projectfile
17836 for Source_Dirs use ("src1", "src2");
17837 for Library_Name use "mylib";
17838 for Library_Dir use "lib";
17839 for Library_Kind use "dynamic";
17840 for Externally_Built use "true";
17845 This project file assumes that the directories @file{src1},
17846 @file{src2}, and @file{lib} exist in
17847 the directory containing the project file. The @code{externally_built}
17848 attribute makes it clear to the GNAT builder that it should not attempt to
17849 recompile any of the units from this library. It allows the library provider to
17850 restrict the source set to the minimum necessary for clients to make use of the
17851 library as described in the first section of this chapter. It is the
17852 responsibility of the library provider to install the necessary sources, ALI
17853 files and libraries in the directories mentioned in the project file. For
17854 convenience, the user's library project file should be installed in a location
17855 that will be searched automatically by the GNAT
17856 builder. These are the directories referenced in the @code{ADA_PROJECT_PATH}
17857 environment variable (@pxref{Importing Projects}), and also the default GNAT
17858 library location that can be queried with @command{gnatls -v} and is usually of
17859 the form $gnat_install_root/lib/gnat.
17861 When project files are not an option, it is also possible, but not recommended,
17862 to install the library so that the sources needed to use the library are on the
17863 Ada source path and the ALI files & libraries be on the Ada Object path (see
17864 @ref{Search Paths and the Run-Time Library (RTL)}. Alternatively, the system
17865 administrator can place general-purpose libraries in the default compiler
17866 paths, by specifying the libraries' location in the configuration files
17867 @file{ada_source_path} and @file{ada_object_path}. These configuration files
17868 must be located in the GNAT installation tree at the same place as the gcc spec
17869 file. The location of the gcc spec file can be determined as follows:
17875 The configuration files mentioned above have a simple format: each line
17876 must contain one unique directory name.
17877 Those names are added to the corresponding path
17878 in their order of appearance in the file. The names can be either absolute
17879 or relative; in the latter case, they are relative to where theses files
17882 The files @file{ada_source_path} and @file{ada_object_path} might not be
17884 GNAT installation, in which case, GNAT will look for its run-time library in
17885 the directories @file{adainclude} (for the sources) and @file{adalib} (for the
17886 objects and @file{ALI} files). When the files exist, the compiler does not
17887 look in @file{adainclude} and @file{adalib}, and thus the
17888 @file{ada_source_path} file
17889 must contain the location for the GNAT run-time sources (which can simply
17890 be @file{adainclude}). In the same way, the @file{ada_object_path} file must
17891 contain the location for the GNAT run-time objects (which can simply
17894 You can also specify a new default path to the run-time library at compilation
17895 time with the switch @option{--RTS=rts-path}. You can thus choose / change
17896 the run-time library you want your program to be compiled with. This switch is
17897 recognized by @command{gcc}, @command{gnatmake}, @command{gnatbind},
17898 @command{gnatls}, @command{gnatfind} and @command{gnatxref}.
17900 It is possible to install a library before or after the standard GNAT
17901 library, by reordering the lines in the configuration files. In general, a
17902 library must be installed before the GNAT library if it redefines
17905 @node Using a library
17906 @subsection Using a library
17908 @noindent Once again, the project facility greatly simplifies the use of
17909 libraries. In this context, using a library is just a matter of adding a
17910 @code{with} clause in the user project. For instance, to make use of the
17911 library @code{My_Lib} shown in examples in earlier sections, you can
17914 @smallexample @c projectfile
17921 Even if you have a third-party, non-Ada library, you can still use GNAT's
17922 Project Manager facility to provide a wrapper for it. For example, the
17923 following project, when @code{with}ed by your main project, will link with the
17924 third-party library @file{liba.a}:
17926 @smallexample @c projectfile
17929 for Externally_Built use "true";
17930 for Source_Files use ();
17931 for Library_Dir use "lib";
17932 for Library_Name use "a";
17933 for Library_Kind use "static";
17937 This is an alternative to the use of @code{pragma Linker_Options}. It is
17938 especially interesting in the context of systems with several interdependent
17939 static libraries where finding a proper linker order is not easy and best be
17940 left to the tools having visibility over project dependence information.
17943 In order to use an Ada library manually, you need to make sure that this
17944 library is on both your source and object path
17945 (see @ref{Search Paths and the Run-Time Library (RTL)}
17946 and @ref{Search Paths for gnatbind}). Furthermore, when the objects are grouped
17947 in an archive or a shared library, you need to specify the desired
17948 library at link time.
17950 For example, you can use the library @file{mylib} installed in
17951 @file{/dir/my_lib_src} and @file{/dir/my_lib_obj} with the following commands:
17954 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \
17959 This can be expressed more simply:
17964 when the following conditions are met:
17967 @file{/dir/my_lib_src} has been added by the user to the environment
17968 variable @code{ADA_INCLUDE_PATH}, or by the administrator to the file
17969 @file{ada_source_path}
17971 @file{/dir/my_lib_obj} has been added by the user to the environment
17972 variable @code{ADA_OBJECTS_PATH}, or by the administrator to the file
17973 @file{ada_object_path}
17975 a pragma @code{Linker_Options} has been added to one of the sources.
17978 @smallexample @c ada
17979 pragma Linker_Options ("-lmy_lib");
17983 @node Stand-alone Ada Libraries
17984 @section Stand-alone Ada Libraries
17985 @cindex Stand-alone library, building, using
17988 * Introduction to Stand-alone Libraries::
17989 * Building a Stand-alone Library::
17990 * Creating a Stand-alone Library to be used in a non-Ada context::
17991 * Restrictions in Stand-alone Libraries::
17994 @node Introduction to Stand-alone Libraries
17995 @subsection Introduction to Stand-alone Libraries
17998 A Stand-alone Library (abbreviated ``SAL'') is a library that contains the
18000 elaborate the Ada units that are included in the library. In contrast with
18001 an ordinary library, which consists of all sources, objects and @file{ALI}
18003 library, a SAL may specify a restricted subset of compilation units
18004 to serve as a library interface. In this case, the fully
18005 self-sufficient set of files will normally consist of an objects
18006 archive, the sources of interface units' specs, and the @file{ALI}
18007 files of interface units.
18008 If an interface spec contains a generic unit or an inlined subprogram,
18010 source must also be provided; if the units that must be provided in the source
18011 form depend on other units, the source and @file{ALI} files of those must
18014 The main purpose of a SAL is to minimize the recompilation overhead of client
18015 applications when a new version of the library is installed. Specifically,
18016 if the interface sources have not changed, client applications do not need to
18017 be recompiled. If, furthermore, a SAL is provided in the shared form and its
18018 version, controlled by @code{Library_Version} attribute, is not changed,
18019 then the clients do not need to be relinked.
18021 SALs also allow the library providers to minimize the amount of library source
18022 text exposed to the clients. Such ``information hiding'' might be useful or
18023 necessary for various reasons.
18025 Stand-alone libraries are also well suited to be used in an executable whose
18026 main routine is not written in Ada.
18028 @node Building a Stand-alone Library
18029 @subsection Building a Stand-alone Library
18032 GNAT's Project facility provides a simple way of building and installing
18033 stand-alone libraries; see @ref{Stand-alone Library Projects}.
18034 To be a Stand-alone Library Project, in addition to the two attributes
18035 that make a project a Library Project (@code{Library_Name} and
18036 @code{Library_Dir}; see @ref{Library Projects}), the attribute
18037 @code{Library_Interface} must be defined. For example:
18039 @smallexample @c projectfile
18041 for Library_Dir use "lib_dir";
18042 for Library_Name use "dummy";
18043 for Library_Interface use ("int1", "int1.child");
18048 Attribute @code{Library_Interface} has a non-empty string list value,
18049 each string in the list designating a unit contained in an immediate source
18050 of the project file.
18052 When a Stand-alone Library is built, first the binder is invoked to build
18053 a package whose name depends on the library name
18054 (@file{^b~dummy.ads/b^B$DUMMY.ADS/B^} in the example above).
18055 This binder-generated package includes initialization and
18056 finalization procedures whose
18057 names depend on the library name (@code{dummyinit} and @code{dummyfinal}
18059 above). The object corresponding to this package is included in the library.
18061 You must ensure timely (e.g., prior to any use of interfaces in the SAL)
18062 calling of these procedures if a static SAL is built, or if a shared SAL
18064 with the project-level attribute @code{Library_Auto_Init} set to
18067 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
18068 (those that are listed in attribute @code{Library_Interface}) are copied to
18069 the Library Directory. As a consequence, only the Interface Units may be
18070 imported from Ada units outside of the library. If other units are imported,
18071 the binding phase will fail.
18073 The attribute @code{Library_Src_Dir} may be specified for a
18074 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
18075 single string value. Its value must be the path (absolute or relative to the
18076 project directory) of an existing directory. This directory cannot be the
18077 object directory or one of the source directories, but it can be the same as
18078 the library directory. The sources of the Interface
18079 Units of the library that are needed by an Ada client of the library will be
18080 copied to the designated directory, called the Interface Copy directory.
18081 These sources include the specs of the Interface Units, but they may also
18082 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
18083 are used, or when there is a generic unit in the spec. Before the sources
18084 are copied to the Interface Copy directory, an attempt is made to delete all
18085 files in the Interface Copy directory.
18087 Building stand-alone libraries by hand is somewhat tedious, but for those
18088 occasions when it is necessary here are the steps that you need to perform:
18091 Compile all library sources.
18094 Invoke the binder with the switch @option{-n} (No Ada main program),
18095 with all the @file{ALI} files of the interfaces, and
18096 with the switch @option{-L} to give specific names to the @code{init}
18097 and @code{final} procedures. For example:
18099 gnatbind -n int1.ali int2.ali -Lsal1
18103 Compile the binder generated file:
18109 Link the dynamic library with all the necessary object files,
18110 indicating to the linker the names of the @code{init} (and possibly
18111 @code{final}) procedures for automatic initialization (and finalization).
18112 The built library should be placed in a directory different from
18113 the object directory.
18116 Copy the @code{ALI} files of the interface to the library directory,
18117 add in this copy an indication that it is an interface to a SAL
18118 (i.e. add a word @option{SL} on the line in the @file{ALI} file that starts
18119 with letter ``P'') and make the modified copy of the @file{ALI} file
18124 Using SALs is not different from using other libraries
18125 (see @ref{Using a library}).
18127 @node Creating a Stand-alone Library to be used in a non-Ada context
18128 @subsection Creating a Stand-alone Library to be used in a non-Ada context
18131 It is easy to adapt the SAL build procedure discussed above for use of a SAL in
18134 The only extra step required is to ensure that library interface subprograms
18135 are compatible with the main program, by means of @code{pragma Export}
18136 or @code{pragma Convention}.
18138 Here is an example of simple library interface for use with C main program:
18140 @smallexample @c ada
18141 package Interface is
18143 procedure Do_Something;
18144 pragma Export (C, Do_Something, "do_something");
18146 procedure Do_Something_Else;
18147 pragma Export (C, Do_Something_Else, "do_something_else");
18153 On the foreign language side, you must provide a ``foreign'' view of the
18154 library interface; remember that it should contain elaboration routines in
18155 addition to interface subprograms.
18157 The example below shows the content of @code{mylib_interface.h} (note
18158 that there is no rule for the naming of this file, any name can be used)
18160 /* the library elaboration procedure */
18161 extern void mylibinit (void);
18163 /* the library finalization procedure */
18164 extern void mylibfinal (void);
18166 /* the interface exported by the library */
18167 extern void do_something (void);
18168 extern void do_something_else (void);
18172 Libraries built as explained above can be used from any program, provided
18173 that the elaboration procedures (named @code{mylibinit} in the previous
18174 example) are called before the library services are used. Any number of
18175 libraries can be used simultaneously, as long as the elaboration
18176 procedure of each library is called.
18178 Below is an example of a C program that uses the @code{mylib} library.
18181 #include "mylib_interface.h"
18186 /* First, elaborate the library before using it */
18189 /* Main program, using the library exported entities */
18191 do_something_else ();
18193 /* Library finalization at the end of the program */
18200 Note that invoking any library finalization procedure generated by
18201 @code{gnatbind} shuts down the Ada run-time environment.
18203 finalization of all Ada libraries must be performed at the end of the program.
18204 No call to these libraries or to the Ada run-time library should be made
18205 after the finalization phase.
18207 @node Restrictions in Stand-alone Libraries
18208 @subsection Restrictions in Stand-alone Libraries
18211 The pragmas listed below should be used with caution inside libraries,
18212 as they can create incompatibilities with other Ada libraries:
18214 @item pragma @code{Locking_Policy}
18215 @item pragma @code{Queuing_Policy}
18216 @item pragma @code{Task_Dispatching_Policy}
18217 @item pragma @code{Unreserve_All_Interrupts}
18221 When using a library that contains such pragmas, the user must make sure
18222 that all libraries use the same pragmas with the same values. Otherwise,
18223 @code{Program_Error} will
18224 be raised during the elaboration of the conflicting
18225 libraries. The usage of these pragmas and its consequences for the user
18226 should therefore be well documented.
18228 Similarly, the traceback in the exception occurrence mechanism should be
18229 enabled or disabled in a consistent manner across all libraries.
18230 Otherwise, Program_Error will be raised during the elaboration of the
18231 conflicting libraries.
18233 If the @code{Version} or @code{Body_Version}
18234 attributes are used inside a library, then you need to
18235 perform a @code{gnatbind} step that specifies all @file{ALI} files in all
18236 libraries, so that version identifiers can be properly computed.
18237 In practice these attributes are rarely used, so this is unlikely
18238 to be a consideration.
18240 @node Rebuilding the GNAT Run-Time Library
18241 @section Rebuilding the GNAT Run-Time Library
18242 @cindex GNAT Run-Time Library, rebuilding
18243 @cindex Building the GNAT Run-Time Library
18244 @cindex Rebuilding the GNAT Run-Time Library
18245 @cindex Run-Time Library, rebuilding
18248 It may be useful to recompile the GNAT library in various contexts, the
18249 most important one being the use of partition-wide configuration pragmas
18250 such as @code{Normalize_Scalars}. A special Makefile called
18251 @code{Makefile.adalib} is provided to that effect and can be found in
18252 the directory containing the GNAT library. The location of this
18253 directory depends on the way the GNAT environment has been installed and can
18254 be determined by means of the command:
18261 The last entry in the object search path usually contains the
18262 gnat library. This Makefile contains its own documentation and in
18263 particular the set of instructions needed to rebuild a new library and
18266 @node Using the GNU make Utility
18267 @chapter Using the GNU @code{make} Utility
18271 This chapter offers some examples of makefiles that solve specific
18272 problems. It does not explain how to write a makefile (see the GNU make
18273 documentation), nor does it try to replace the @command{gnatmake} utility
18274 (@pxref{The GNAT Make Program gnatmake}).
18276 All the examples in this section are specific to the GNU version of
18277 make. Although @code{make} is a standard utility, and the basic language
18278 is the same, these examples use some advanced features found only in
18282 * Using gnatmake in a Makefile::
18283 * Automatically Creating a List of Directories::
18284 * Generating the Command Line Switches::
18285 * Overcoming Command Line Length Limits::
18288 @node Using gnatmake in a Makefile
18289 @section Using gnatmake in a Makefile
18294 Complex project organizations can be handled in a very powerful way by
18295 using GNU make combined with gnatmake. For instance, here is a Makefile
18296 which allows you to build each subsystem of a big project into a separate
18297 shared library. Such a makefile allows you to significantly reduce the link
18298 time of very big applications while maintaining full coherence at
18299 each step of the build process.
18301 The list of dependencies are handled automatically by
18302 @command{gnatmake}. The Makefile is simply used to call gnatmake in each of
18303 the appropriate directories.
18305 Note that you should also read the example on how to automatically
18306 create the list of directories
18307 (@pxref{Automatically Creating a List of Directories})
18308 which might help you in case your project has a lot of subdirectories.
18313 @font@heightrm=cmr8
18316 ## This Makefile is intended to be used with the following directory
18318 ## - The sources are split into a series of csc (computer software components)
18319 ## Each of these csc is put in its own directory.
18320 ## Their name are referenced by the directory names.
18321 ## They will be compiled into shared library (although this would also work
18322 ## with static libraries
18323 ## - The main program (and possibly other packages that do not belong to any
18324 ## csc is put in the top level directory (where the Makefile is).
18325 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
18326 ## \_ second_csc (sources) __ lib (will contain the library)
18328 ## Although this Makefile is build for shared library, it is easy to modify
18329 ## to build partial link objects instead (modify the lines with -shared and
18332 ## With this makefile, you can change any file in the system or add any new
18333 ## file, and everything will be recompiled correctly (only the relevant shared
18334 ## objects will be recompiled, and the main program will be re-linked).
18336 # The list of computer software component for your project. This might be
18337 # generated automatically.
18340 # Name of the main program (no extension)
18343 # If we need to build objects with -fPIC, uncomment the following line
18346 # The following variable should give the directory containing libgnat.so
18347 # You can get this directory through 'gnatls -v'. This is usually the last
18348 # directory in the Object_Path.
18351 # The directories for the libraries
18352 # (This macro expands the list of CSC to the list of shared libraries, you
18353 # could simply use the expanded form :
18354 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
18355 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
18357 $@{MAIN@}: objects $@{LIB_DIR@}
18358 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
18359 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
18362 # recompile the sources
18363 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
18365 # Note: In a future version of GNAT, the following commands will be simplified
18366 # by a new tool, gnatmlib
18368 mkdir -p $@{dir $@@ @}
18369 cd $@{dir $@@ @}; gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
18370 cd $@{dir $@@ @}; cp -f ../*.ali .
18372 # The dependencies for the modules
18373 # Note that we have to force the expansion of *.o, since in some cases
18374 # make won't be able to do it itself.
18375 aa/lib/libaa.so: $@{wildcard aa/*.o@}
18376 bb/lib/libbb.so: $@{wildcard bb/*.o@}
18377 cc/lib/libcc.so: $@{wildcard cc/*.o@}
18379 # Make sure all of the shared libraries are in the path before starting the
18382 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
18385 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
18386 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
18387 $@{RM@} $@{CSC_LIST:%=%/*.o@}
18388 $@{RM@} *.o *.ali $@{MAIN@}
18391 @node Automatically Creating a List of Directories
18392 @section Automatically Creating a List of Directories
18395 In most makefiles, you will have to specify a list of directories, and
18396 store it in a variable. For small projects, it is often easier to
18397 specify each of them by hand, since you then have full control over what
18398 is the proper order for these directories, which ones should be
18401 However, in larger projects, which might involve hundreds of
18402 subdirectories, it might be more convenient to generate this list
18405 The example below presents two methods. The first one, although less
18406 general, gives you more control over the list. It involves wildcard
18407 characters, that are automatically expanded by @code{make}. Its
18408 shortcoming is that you need to explicitly specify some of the
18409 organization of your project, such as for instance the directory tree
18410 depth, whether some directories are found in a separate tree,...
18412 The second method is the most general one. It requires an external
18413 program, called @code{find}, which is standard on all Unix systems. All
18414 the directories found under a given root directory will be added to the
18420 @font@heightrm=cmr8
18423 # The examples below are based on the following directory hierarchy:
18424 # All the directories can contain any number of files
18425 # ROOT_DIRECTORY -> a -> aa -> aaa
18428 # -> b -> ba -> baa
18431 # This Makefile creates a variable called DIRS, that can be reused any time
18432 # you need this list (see the other examples in this section)
18434 # The root of your project's directory hierarchy
18438 # First method: specify explicitly the list of directories
18439 # This allows you to specify any subset of all the directories you need.
18442 DIRS := a/aa/ a/ab/ b/ba/
18445 # Second method: use wildcards
18446 # Note that the argument(s) to wildcard below should end with a '/'.
18447 # Since wildcards also return file names, we have to filter them out
18448 # to avoid duplicate directory names.
18449 # We thus use make's @code{dir} and @code{sort} functions.
18450 # It sets DIRs to the following value (note that the directories aaa and baa
18451 # are not given, unless you change the arguments to wildcard).
18452 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
18455 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
18456 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
18459 # Third method: use an external program
18460 # This command is much faster if run on local disks, avoiding NFS slowdowns.
18461 # This is the most complete command: it sets DIRs to the following value:
18462 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
18465 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
18469 @node Generating the Command Line Switches
18470 @section Generating the Command Line Switches
18473 Once you have created the list of directories as explained in the
18474 previous section (@pxref{Automatically Creating a List of Directories}),
18475 you can easily generate the command line arguments to pass to gnatmake.
18477 For the sake of completeness, this example assumes that the source path
18478 is not the same as the object path, and that you have two separate lists
18482 # see "Automatically creating a list of directories" to create
18487 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
18488 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
18491 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
18494 @node Overcoming Command Line Length Limits
18495 @section Overcoming Command Line Length Limits
18498 One problem that might be encountered on big projects is that many
18499 operating systems limit the length of the command line. It is thus hard to give
18500 gnatmake the list of source and object directories.
18502 This example shows how you can set up environment variables, which will
18503 make @command{gnatmake} behave exactly as if the directories had been
18504 specified on the command line, but have a much higher length limit (or
18505 even none on most systems).
18507 It assumes that you have created a list of directories in your Makefile,
18508 using one of the methods presented in
18509 @ref{Automatically Creating a List of Directories}.
18510 For the sake of completeness, we assume that the object
18511 path (where the ALI files are found) is different from the sources patch.
18513 Note a small trick in the Makefile below: for efficiency reasons, we
18514 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
18515 expanded immediately by @code{make}. This way we overcome the standard
18516 make behavior which is to expand the variables only when they are
18519 On Windows, if you are using the standard Windows command shell, you must
18520 replace colons with semicolons in the assignments to these variables.
18525 @font@heightrm=cmr8
18528 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECT_PATH.
18529 # This is the same thing as putting the -I arguments on the command line.
18530 # (the equivalent of using -aI on the command line would be to define
18531 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECT_PATH).
18532 # You can of course have different values for these variables.
18534 # Note also that we need to keep the previous values of these variables, since
18535 # they might have been set before running 'make' to specify where the GNAT
18536 # library is installed.
18538 # see "Automatically creating a list of directories" to create these
18544 space:=$@{empty@} $@{empty@}
18545 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
18546 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
18547 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
18548 ADA_OBJECT_PATH += $@{OBJECT_LIST@}
18549 export ADA_INCLUDE_PATH
18550 export ADA_OBJECT_PATH
18557 @node Memory Management Issues
18558 @chapter Memory Management Issues
18561 This chapter describes some useful memory pools provided in the GNAT library
18562 and in particular the GNAT Debug Pool facility, which can be used to detect
18563 incorrect uses of access values (including ``dangling references'').
18565 It also describes the @command{gnatmem} tool, which can be used to track down
18570 * Some Useful Memory Pools::
18571 * The GNAT Debug Pool Facility::
18573 * The gnatmem Tool::
18577 @node Some Useful Memory Pools
18578 @section Some Useful Memory Pools
18579 @findex Memory Pool
18580 @cindex storage, pool
18583 The @code{System.Pool_Global} package offers the Unbounded_No_Reclaim_Pool
18584 storage pool. Allocations use the standard system call @code{malloc} while
18585 deallocations use the standard system call @code{free}. No reclamation is
18586 performed when the pool goes out of scope. For performance reasons, the
18587 standard default Ada allocators/deallocators do not use any explicit storage
18588 pools but if they did, they could use this storage pool without any change in
18589 behavior. That is why this storage pool is used when the user
18590 manages to make the default implicit allocator explicit as in this example:
18591 @smallexample @c ada
18592 type T1 is access Something;
18593 -- no Storage pool is defined for T2
18594 type T2 is access Something_Else;
18595 for T2'Storage_Pool use T1'Storage_Pool;
18596 -- the above is equivalent to
18597 for T2'Storage_Pool use System.Pool_Global.Global_Pool_Object;
18601 The @code{System.Pool_Local} package offers the Unbounded_Reclaim_Pool storage
18602 pool. The allocation strategy is similar to @code{Pool_Local}'s
18603 except that the all
18604 storage allocated with this pool is reclaimed when the pool object goes out of
18605 scope. This pool provides a explicit mechanism similar to the implicit one
18606 provided by several Ada 83 compilers for allocations performed through a local
18607 access type and whose purpose was to reclaim memory when exiting the
18608 scope of a given local access. As an example, the following program does not
18609 leak memory even though it does not perform explicit deallocation:
18611 @smallexample @c ada
18612 with System.Pool_Local;
18613 procedure Pooloc1 is
18614 procedure Internal is
18615 type A is access Integer;
18616 X : System.Pool_Local.Unbounded_Reclaim_Pool;
18617 for A'Storage_Pool use X;
18620 for I in 1 .. 50 loop
18625 for I in 1 .. 100 loop
18632 The @code{System.Pool_Size} package implements the Stack_Bounded_Pool used when
18633 @code{Storage_Size} is specified for an access type.
18634 The whole storage for the pool is
18635 allocated at once, usually on the stack at the point where the access type is
18636 elaborated. It is automatically reclaimed when exiting the scope where the
18637 access type is defined. This package is not intended to be used directly by the
18638 user and it is implicitly used for each such declaration:
18640 @smallexample @c ada
18641 type T1 is access Something;
18642 for T1'Storage_Size use 10_000;
18645 @node The GNAT Debug Pool Facility
18646 @section The GNAT Debug Pool Facility
18648 @cindex storage, pool, memory corruption
18651 The use of unchecked deallocation and unchecked conversion can easily
18652 lead to incorrect memory references. The problems generated by such
18653 references are usually difficult to tackle because the symptoms can be
18654 very remote from the origin of the problem. In such cases, it is
18655 very helpful to detect the problem as early as possible. This is the
18656 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
18658 In order to use the GNAT specific debugging pool, the user must
18659 associate a debug pool object with each of the access types that may be
18660 related to suspected memory problems. See Ada Reference Manual 13.11.
18661 @smallexample @c ada
18662 type Ptr is access Some_Type;
18663 Pool : GNAT.Debug_Pools.Debug_Pool;
18664 for Ptr'Storage_Pool use Pool;
18668 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
18669 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
18670 allow the user to redefine allocation and deallocation strategies. They
18671 also provide a checkpoint for each dereference, through the use of
18672 the primitive operation @code{Dereference} which is implicitly called at
18673 each dereference of an access value.
18675 Once an access type has been associated with a debug pool, operations on
18676 values of the type may raise four distinct exceptions,
18677 which correspond to four potential kinds of memory corruption:
18680 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
18682 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
18684 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
18686 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage }
18690 For types associated with a Debug_Pool, dynamic allocation is performed using
18691 the standard GNAT allocation routine. References to all allocated chunks of
18692 memory are kept in an internal dictionary. Several deallocation strategies are
18693 provided, whereupon the user can choose to release the memory to the system,
18694 keep it allocated for further invalid access checks, or fill it with an easily
18695 recognizable pattern for debug sessions. The memory pattern is the old IBM
18696 hexadecimal convention: @code{16#DEADBEEF#}.
18698 See the documentation in the file g-debpoo.ads for more information on the
18699 various strategies.
18701 Upon each dereference, a check is made that the access value denotes a
18702 properly allocated memory location. Here is a complete example of use of
18703 @code{Debug_Pools}, that includes typical instances of memory corruption:
18704 @smallexample @c ada
18708 with Gnat.Io; use Gnat.Io;
18709 with Unchecked_Deallocation;
18710 with Unchecked_Conversion;
18711 with GNAT.Debug_Pools;
18712 with System.Storage_Elements;
18713 with Ada.Exceptions; use Ada.Exceptions;
18714 procedure Debug_Pool_Test is
18716 type T is access Integer;
18717 type U is access all T;
18719 P : GNAT.Debug_Pools.Debug_Pool;
18720 for T'Storage_Pool use P;
18722 procedure Free is new Unchecked_Deallocation (Integer, T);
18723 function UC is new Unchecked_Conversion (U, T);
18726 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
18736 Put_Line (Integer'Image(B.all));
18738 when E : others => Put_Line ("raised: " & Exception_Name (E));
18743 when E : others => Put_Line ("raised: " & Exception_Name (E));
18747 Put_Line (Integer'Image(B.all));
18749 when E : others => Put_Line ("raised: " & Exception_Name (E));
18754 when E : others => Put_Line ("raised: " & Exception_Name (E));
18757 end Debug_Pool_Test;
18761 The debug pool mechanism provides the following precise diagnostics on the
18762 execution of this erroneous program:
18765 Total allocated bytes : 0
18766 Total deallocated bytes : 0
18767 Current Water Mark: 0
18771 Total allocated bytes : 8
18772 Total deallocated bytes : 0
18773 Current Water Mark: 8
18776 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
18777 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
18778 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
18779 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
18781 Total allocated bytes : 8
18782 Total deallocated bytes : 4
18783 Current Water Mark: 4
18788 @node The gnatmem Tool
18789 @section The @command{gnatmem} Tool
18793 The @code{gnatmem} utility monitors dynamic allocation and
18794 deallocation activity in a program, and displays information about
18795 incorrect deallocations and possible sources of memory leaks.
18796 It provides three type of information:
18799 General information concerning memory management, such as the total
18800 number of allocations and deallocations, the amount of allocated
18801 memory and the high water mark, i.e. the largest amount of allocated
18802 memory in the course of program execution.
18805 Backtraces for all incorrect deallocations, that is to say deallocations
18806 which do not correspond to a valid allocation.
18809 Information on each allocation that is potentially the origin of a memory
18814 * Running gnatmem::
18815 * Switches for gnatmem::
18816 * Example of gnatmem Usage::
18819 @node Running gnatmem
18820 @subsection Running @code{gnatmem}
18823 @code{gnatmem} makes use of the output created by the special version of
18824 allocation and deallocation routines that record call information. This
18825 allows to obtain accurate dynamic memory usage history at a minimal cost to
18826 the execution speed. Note however, that @code{gnatmem} is not supported on
18827 all platforms (currently, it is supported on AIX, HP-UX, GNU/Linux,
18828 Solaris and Windows NT/2000/XP (x86).
18831 The @code{gnatmem} command has the form
18834 $ gnatmem [switches] user_program
18838 The program must have been linked with the instrumented version of the
18839 allocation and deallocation routines. This is done by linking with the
18840 @file{libgmem.a} library. For correct symbolic backtrace information,
18841 the user program should be compiled with debugging options
18842 (see @ref{Switches for gcc}). For example to build @file{my_program}:
18845 $ gnatmake -g my_program -largs -lgmem
18849 As library @file{libgmem.a} contains an alternate body for package
18850 @code{System.Memory}, @file{s-memory.adb} should not be compiled and linked
18851 when an executable is linked with library @file{libgmem.a}. It is then not
18852 recommended to use @command{gnatmake} with switch @option{^-a^/ALL_FILES^}.
18855 When @file{my_program} is executed, the file @file{gmem.out} is produced.
18856 This file contains information about all allocations and deallocations
18857 performed by the program. It is produced by the instrumented allocations and
18858 deallocations routines and will be used by @code{gnatmem}.
18860 In order to produce symbolic backtrace information for allocations and
18861 deallocations performed by the GNAT run-time library, you need to use a
18862 version of that library that has been compiled with the @option{-g} switch
18863 (see @ref{Rebuilding the GNAT Run-Time Library}).
18865 Gnatmem must be supplied with the @file{gmem.out} file and the executable to
18866 examine. If the location of @file{gmem.out} file was not explicitly supplied by
18867 @code{-i} switch, gnatmem will assume that this file can be found in the
18868 current directory. For example, after you have executed @file{my_program},
18869 @file{gmem.out} can be analyzed by @code{gnatmem} using the command:
18872 $ gnatmem my_program
18876 This will produce the output with the following format:
18878 *************** debut cc
18880 $ gnatmem my_program
18884 Total number of allocations : 45
18885 Total number of deallocations : 6
18886 Final Water Mark (non freed mem) : 11.29 Kilobytes
18887 High Water Mark : 11.40 Kilobytes
18892 Allocation Root # 2
18893 -------------------
18894 Number of non freed allocations : 11
18895 Final Water Mark (non freed mem) : 1.16 Kilobytes
18896 High Water Mark : 1.27 Kilobytes
18898 my_program.adb:23 my_program.alloc
18904 The first block of output gives general information. In this case, the
18905 Ada construct ``@code{@b{new}}'' was executed 45 times, and only 6 calls to an
18906 Unchecked_Deallocation routine occurred.
18909 Subsequent paragraphs display information on all allocation roots.
18910 An allocation root is a specific point in the execution of the program
18911 that generates some dynamic allocation, such as a ``@code{@b{new}}''
18912 construct. This root is represented by an execution backtrace (or subprogram
18913 call stack). By default the backtrace depth for allocations roots is 1, so
18914 that a root corresponds exactly to a source location. The backtrace can
18915 be made deeper, to make the root more specific.
18917 @node Switches for gnatmem
18918 @subsection Switches for @code{gnatmem}
18921 @code{gnatmem} recognizes the following switches:
18926 @cindex @option{-q} (@code{gnatmem})
18927 Quiet. Gives the minimum output needed to identify the origin of the
18928 memory leaks. Omits statistical information.
18931 @cindex @var{N} (@code{gnatmem})
18932 N is an integer literal (usually between 1 and 10) which controls the
18933 depth of the backtraces defining allocation root. The default value for
18934 N is 1. The deeper the backtrace, the more precise the localization of
18935 the root. Note that the total number of roots can depend on this
18936 parameter. This parameter must be specified @emph{before} the name of the
18937 executable to be analyzed, to avoid ambiguity.
18940 @cindex @option{-b} (@code{gnatmem})
18941 This switch has the same effect as just depth parameter.
18943 @item -i @var{file}
18944 @cindex @option{-i} (@code{gnatmem})
18945 Do the @code{gnatmem} processing starting from @file{file}, rather than
18946 @file{gmem.out} in the current directory.
18949 @cindex @option{-m} (@code{gnatmem})
18950 This switch causes @code{gnatmem} to mask the allocation roots that have less
18951 than n leaks. The default value is 1. Specifying the value of 0 will allow to
18952 examine even the roots that didn't result in leaks.
18955 @cindex @option{-s} (@code{gnatmem})
18956 This switch causes @code{gnatmem} to sort the allocation roots according to the
18957 specified order of sort criteria, each identified by a single letter. The
18958 currently supported criteria are @code{n, h, w} standing respectively for
18959 number of unfreed allocations, high watermark, and final watermark
18960 corresponding to a specific root. The default order is @code{nwh}.
18964 @node Example of gnatmem Usage
18965 @subsection Example of @code{gnatmem} Usage
18968 The following example shows the use of @code{gnatmem}
18969 on a simple memory-leaking program.
18970 Suppose that we have the following Ada program:
18972 @smallexample @c ada
18975 with Unchecked_Deallocation;
18976 procedure Test_Gm is
18978 type T is array (1..1000) of Integer;
18979 type Ptr is access T;
18980 procedure Free is new Unchecked_Deallocation (T, Ptr);
18983 procedure My_Alloc is
18988 procedure My_DeAlloc is
18996 for I in 1 .. 5 loop
18997 for J in I .. 5 loop
19008 The program needs to be compiled with debugging option and linked with
19009 @code{gmem} library:
19012 $ gnatmake -g test_gm -largs -lgmem
19016 Then we execute the program as usual:
19023 Then @code{gnatmem} is invoked simply with
19029 which produces the following output (result may vary on different platforms):
19034 Total number of allocations : 18
19035 Total number of deallocations : 5
19036 Final Water Mark (non freed mem) : 53.00 Kilobytes
19037 High Water Mark : 56.90 Kilobytes
19039 Allocation Root # 1
19040 -------------------
19041 Number of non freed allocations : 11
19042 Final Water Mark (non freed mem) : 42.97 Kilobytes
19043 High Water Mark : 46.88 Kilobytes
19045 test_gm.adb:11 test_gm.my_alloc
19047 Allocation Root # 2
19048 -------------------
19049 Number of non freed allocations : 1
19050 Final Water Mark (non freed mem) : 10.02 Kilobytes
19051 High Water Mark : 10.02 Kilobytes
19053 s-secsta.adb:81 system.secondary_stack.ss_init
19055 Allocation Root # 3
19056 -------------------
19057 Number of non freed allocations : 1
19058 Final Water Mark (non freed mem) : 12 Bytes
19059 High Water Mark : 12 Bytes
19061 s-secsta.adb:181 system.secondary_stack.ss_init
19065 Note that the GNAT run time contains itself a certain number of
19066 allocations that have no corresponding deallocation,
19067 as shown here for root #2 and root
19068 #3. This is a normal behavior when the number of non freed allocations
19069 is one, it allocates dynamic data structures that the run time needs for
19070 the complete lifetime of the program. Note also that there is only one
19071 allocation root in the user program with a single line back trace:
19072 test_gm.adb:11 test_gm.my_alloc, whereas a careful analysis of the
19073 program shows that 'My_Alloc' is called at 2 different points in the
19074 source (line 21 and line 24). If those two allocation roots need to be
19075 distinguished, the backtrace depth parameter can be used:
19078 $ gnatmem 3 test_gm
19082 which will give the following output:
19087 Total number of allocations : 18
19088 Total number of deallocations : 5
19089 Final Water Mark (non freed mem) : 53.00 Kilobytes
19090 High Water Mark : 56.90 Kilobytes
19092 Allocation Root # 1
19093 -------------------
19094 Number of non freed allocations : 10
19095 Final Water Mark (non freed mem) : 39.06 Kilobytes
19096 High Water Mark : 42.97 Kilobytes
19098 test_gm.adb:11 test_gm.my_alloc
19099 test_gm.adb:24 test_gm
19100 b_test_gm.c:52 main
19102 Allocation Root # 2
19103 -------------------
19104 Number of non freed allocations : 1
19105 Final Water Mark (non freed mem) : 10.02 Kilobytes
19106 High Water Mark : 10.02 Kilobytes
19108 s-secsta.adb:81 system.secondary_stack.ss_init
19109 s-secsta.adb:283 <system__secondary_stack___elabb>
19110 b_test_gm.c:33 adainit
19112 Allocation Root # 3
19113 -------------------
19114 Number of non freed allocations : 1
19115 Final Water Mark (non freed mem) : 3.91 Kilobytes
19116 High Water Mark : 3.91 Kilobytes
19118 test_gm.adb:11 test_gm.my_alloc
19119 test_gm.adb:21 test_gm
19120 b_test_gm.c:52 main
19122 Allocation Root # 4
19123 -------------------
19124 Number of non freed allocations : 1
19125 Final Water Mark (non freed mem) : 12 Bytes
19126 High Water Mark : 12 Bytes
19128 s-secsta.adb:181 system.secondary_stack.ss_init
19129 s-secsta.adb:283 <system__secondary_stack___elabb>
19130 b_test_gm.c:33 adainit
19134 The allocation root #1 of the first example has been split in 2 roots #1
19135 and #3 thanks to the more precise associated backtrace.
19139 @node Stack Related Facilities
19140 @chapter Stack Related Facilities
19143 This chapter describes some useful tools associated with stack
19144 checking and analysis. In
19145 particular, it deals with dynamic and static stack usage measurements.
19148 * Stack Overflow Checking::
19149 * Static Stack Usage Analysis::
19150 * Dynamic Stack Usage Analysis::
19153 @node Stack Overflow Checking
19154 @section Stack Overflow Checking
19155 @cindex Stack Overflow Checking
19156 @cindex -fstack-check
19159 For most operating systems, @command{gcc} does not perform stack overflow
19160 checking by default. This means that if the main environment task or
19161 some other task exceeds the available stack space, then unpredictable
19162 behavior will occur. Most native systems offer some level of protection by
19163 adding a guard page at the end of each task stack. This mechanism is usually
19164 not enough for dealing properly with stack overflow situations because
19165 a large local variable could ``jump'' above the guard page.
19166 Furthermore, when the
19167 guard page is hit, there may not be any space left on the stack for executing
19168 the exception propagation code. Enabling stack checking avoids
19171 To activate stack checking, compile all units with the gcc option
19172 @option{-fstack-check}. For example:
19175 gcc -c -fstack-check package1.adb
19179 Units compiled with this option will generate extra instructions to check
19180 that any use of the stack (for procedure calls or for declaring local
19181 variables in declare blocks) does not exceed the available stack space.
19182 If the space is exceeded, then a @code{Storage_Error} exception is raised.
19184 For declared tasks, the stack size is controlled by the size
19185 given in an applicable @code{Storage_Size} pragma or by the value specified
19186 at bind time with @option{-d} (@pxref{Switches for gnatbind}) or is set to
19187 the default size as defined in the GNAT runtime otherwise.
19189 For the environment task, the stack size depends on
19190 system defaults and is unknown to the compiler. Stack checking
19191 may still work correctly if a fixed
19192 size stack is allocated, but this cannot be guaranteed.
19193 To ensure that a clean exception is signalled for stack
19194 overflow, set the environment variable
19195 @code{GNAT_STACK_LIMIT} to indicate the maximum
19196 stack area that can be used, as in:
19197 @cindex GNAT_STACK_LIMIT
19200 SET GNAT_STACK_LIMIT 1600
19204 The limit is given in kilobytes, so the above declaration would
19205 set the stack limit of the environment task to 1.6 megabytes.
19206 Note that the only purpose of this usage is to limit the amount
19207 of stack used by the environment task. If it is necessary to
19208 increase the amount of stack for the environment task, then this
19209 is an operating systems issue, and must be addressed with the
19210 appropriate operating systems commands.
19212 @node Static Stack Usage Analysis
19213 @section Static Stack Usage Analysis
19214 @cindex Static Stack Usage Analysis
19215 @cindex -fstack-usage
19218 A unit compiled with @option{-fstack-usage} will generate an extra file
19220 the maximum amount of stack used, on a per-function basis.
19221 The file has the same
19222 basename as the target object file with a @file{.su} extension.
19223 Each line of this file is made up of three fields:
19227 The name of the function.
19231 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
19234 The second field corresponds to the size of the known part of the function
19237 The qualifier @code{static} means that the function frame size
19239 It usually means that all local variables have a static size.
19240 In this case, the second field is a reliable measure of the function stack
19243 The qualifier @code{dynamic} means that the function frame size is not static.
19244 It happens mainly when some local variables have a dynamic size. When this
19245 qualifier appears alone, the second field is not a reliable measure
19246 of the function stack analysis. When it is qualified with @code{bounded}, it
19247 means that the second field is a reliable maximum of the function stack
19250 @node Dynamic Stack Usage Analysis
19251 @section Dynamic Stack Usage Analysis
19254 It is possible to measure the maximum amount of stack used by a task, by
19255 adding a switch to @command{gnatbind}, as:
19258 $ gnatbind -u0 file
19262 With this option, at each task termination, its stack usage is output on
19264 It is not always convenient to output the stack usage when the program
19265 is still running. Hence, it is possible to delay this output until program
19266 termination. for a given number of tasks specified as the argument of the
19267 @code{-u} option. For instance:
19270 $ gnatbind -u100 file
19274 will buffer the stack usage information of the first 100 tasks to terminate and
19275 output this info at program termination. Results are displayed in four
19279 Index | Task Name | Stack Size | Actual Use [min - max]
19286 is a number associated with each task.
19289 is the name of the task analyzed.
19292 is the maximum size for the stack.
19295 is the measure done by the stack analyzer. In order to prevent overflow,
19296 the stack is not entirely analyzed, and it's not possible to know exactly how
19297 much has actually been used. The real amount of stack used is between the min
19303 The environment task stack, e.g. the stack that contains the main unit, is
19304 only processed when the environment variable GNAT_STACK_LIMIT is set.
19307 @c *********************************
19309 @c *********************************
19310 @node Verifying Properties Using gnatcheck
19311 @chapter Verifying Properties Using @command{gnatcheck}
19313 @cindex @command{gnatcheck}
19316 The @command{gnatcheck} tool is an ASIS-based utility that checks properties
19317 of Ada source files according to a given set of semantic rules.
19320 In order to check compliance with a given rule, @command{gnatcheck} has to
19321 semantically analyze the Ada sources.
19322 Therefore, checks can only be performed on
19323 legal Ada units. Moreover, when a unit depends semantically upon units located
19324 outside the current directory, the source search path has to be provided when
19325 calling @command{gnatcheck}, either through a specified project file or
19326 through @command{gnatcheck} switches as described below.
19328 A number of rules are predefined in @command{gnatcheck} and are described
19329 later in this chapter.
19330 You can also add new rules, by modifying the @command{gnatcheck} code and
19331 rebuilding the tool. In order to add a simple rule making some local checks,
19332 a small amount of straightforward ASIS-based programming is usually needed.
19334 Project support for @command{gnatcheck} is provided by the GNAT
19335 driver (see @ref{The GNAT Driver and Project Files}).
19337 Invoking @command{gnatcheck} on the command line has the form:
19340 $ gnatcheck [@i{switches}] @{@i{filename}@}
19341 [^-files^/FILES^=@{@i{arg_list_filename}@}]
19342 [-cargs @i{gcc_switches}] [-rules @i{rule_options}]
19349 @i{switches} specify the general tool options
19352 Each @i{filename} is the name (including the extension) of a source
19353 file to process. ``Wildcards'' are allowed, and
19354 the file name may contain path information.
19357 Each @i{arg_list_filename} is the name (including the extension) of a text
19358 file containing the names of the source files to process, separated by spaces
19362 @i{gcc_switches} is a list of switches for
19363 @command{gcc}. They will be passed on to all compiler invocations made by
19364 @command{gnatcheck} to generate the ASIS trees. Here you can provide
19365 @option{^-I^/INCLUDE_DIRS=^} switches to form the source search path,
19366 and use the @option{-gnatec} switch to set the configuration file.
19369 @i{rule_options} is a list of options for controlling a set of
19370 rules to be checked by @command{gnatcheck} (@pxref{gnatcheck Rule Options}).
19374 Either a @i{filename} or an @i{arg_list_filename} must be supplied.
19377 * Format of the Report File::
19378 * General gnatcheck Switches::
19379 * gnatcheck Rule Options::
19380 * Adding the Results of Compiler Checks to gnatcheck Output::
19381 * Project-Wide Checks::
19382 * Predefined Rules::
19385 @node Format of the Report File
19386 @section Format of the Report File
19387 @cindex Report file (for @code{gnatcheck})
19390 The @command{gnatcheck} tool outputs on @file{stdout} all messages concerning
19392 It also creates, in the current
19393 directory, a text file named @file{^gnatcheck.out^GNATCHECK.OUT^} that
19394 contains the complete report of the last gnatcheck run. This report contains:
19396 @item a list of the Ada source files being checked,
19397 @item a list of enabled and disabled rules,
19398 @item a list of the diagnostic messages, ordered in three different ways
19399 and collected in three separate
19400 sections. Section 1 contains the raw list of diagnostic messages. It
19401 corresponds to the output going to @file{stdout}. Section 2 contains
19402 messages ordered by rules.
19403 Section 3 contains messages ordered by source files.
19406 @node General gnatcheck Switches
19407 @section General @command{gnatcheck} Switches
19410 The following switches control the general @command{gnatcheck} behavior
19414 @cindex @option{^-a^/ALL^} (@command{gnatcheck})
19416 Process all units including those with read-only ALI files such as
19417 those from GNAT Run-Time library.
19421 @cindex @option{-d} (@command{gnatcheck})
19426 @cindex @option{-dd} (@command{gnatcheck})
19428 Progress indicator mode (for use in GPS)
19431 @cindex @option{^-h^/HELP^} (@command{gnatcheck})
19433 List the predefined and user-defined rules. For more details see
19434 @ref{Predefined Rules}.
19436 @cindex @option{^-l^/LOCS^} (@command{gnatcheck})
19438 Use full source locations references in the report file. For a construct from
19439 a generic instantiation a full source location is a chain from the location
19440 of this construct in the generic unit to the place where this unit is
19443 @cindex @option{^-q^/QUIET^} (@command{gnatcheck})
19445 Quiet mode. All the diagnoses about rule violations are placed in the
19446 @command{gnatcheck} report file only, without duplicating in @file{stdout}.
19448 @cindex @option{^-s^/SHORT^} (@command{gnatcheck})
19450 Short format of the report file (no version information, no list of applied
19451 rules, no list of checked sources is included)
19453 @cindex @option{^-s1^/COMPILER_STYLE^} (@command{gnatcheck})
19454 @item ^-s1^/COMPILER_STYLE^
19455 Include the compiler-style section in the report file
19457 @cindex @option{^-s2^/BY_RULES^} (@command{gnatcheck})
19458 @item ^-s2^/BY_RULES^
19459 Include the section containing diagnoses ordered by rules in the report file
19461 @cindex @option{^-s3^/BY_FILES_BY_RULES^} (@command{gnatcheck})
19462 @item ^-s3^/BY_FILES_BY_RULES^
19463 Include the section containing diagnoses ordered by files and then by rules
19466 @cindex @option{^-v^/VERBOSE^} (@command{gnatcheck})
19467 @item ^-v^/VERBOSE^
19468 Verbose mode; @command{gnatcheck} generates version information and then
19469 a trace of sources being processed.
19474 Note that if any of the options @option{^-s1^/COMPILER_STYLE^},
19475 @option{^-s2^/BY_RULES^} or
19476 @option{^-s3^/BY_FILES_BY_RULES^} is specified,
19477 then the @command{gnatcheck} report file will only contain sections
19478 explicitly denoted by these options.
19480 @node gnatcheck Rule Options
19481 @section @command{gnatcheck} Rule Options
19484 The following options control the processing performed by
19485 @command{gnatcheck}.
19488 @cindex @option{+ALL} (@command{gnatcheck})
19490 Turn all the rule checks ON.
19492 @cindex @option{-ALL} (@command{gnatcheck})
19494 Turn all the rule checks OFF.
19496 @cindex @option{+R} (@command{gnatcheck})
19497 @item +R@i{rule_id[:param]}
19498 Turn on the check for a specified rule with the specified parameter, if any.
19499 @i{rule_id} must be the identifier of one of the currently implemented rules
19500 (use @option{^-h^/HELP^} for the list of implemented rules). Rule identifiers
19501 are not case-sensitive. The @i{param} item must
19502 be a string representing a valid parameter(s) for the specified rule.
19503 If it contains any space characters then this string must be enclosed in
19506 @cindex @option{-R} (@command{gnatcheck})
19507 @item -R@i{rule_id[:param]}
19508 Turn off the check for a specified rule with the specified parameter, if any.
19510 @cindex @option{-from} (@command{gnatcheck})
19511 @item -from=@i{rule_option_filename}
19512 Read the rule options from the text file @i{rule_option_filename}, referred as
19513 ``rule file'' below.
19518 The default behavior is that all the rule checks are enabled, except for
19519 the checks performed by the compiler.
19521 and the checks associated with the
19525 A rule file is a text file containing a set of rule options.
19526 @cindex Rule file (for @code{gnatcheck})
19527 The file may contain empty lines and Ada-style comments (comment
19528 lines and end-of-line comments). The rule file has free format; that is,
19529 you do not have to start a new rule option on a new line.
19531 A rule file may contain other @option{-from=@i{rule_option_filename}}
19532 options, each such option being replaced with the content of the
19533 corresponding rule file during the rule files processing. In case a
19534 cycle is detected (that is, @i{rule_file_1} reads rule options from
19535 @i{rule_file_2}, and @i{rule_file_2} reads (directly or indirectly)
19536 rule options from @i{rule_file_1}), the processing
19537 of rule files is interrupted and a part of their content is ignored.
19540 @node Adding the Results of Compiler Checks to gnatcheck Output
19541 @section Adding the Results of Compiler Checks to @command{gnatcheck} Output
19544 The @command{gnatcheck} tool can include in the generated diagnostic messages
19546 the report file the results of the checks performed by the compiler. Though
19547 disabled by default, this effect may be obtained by using @option{+R} with
19548 the following rule identifiers and parameters:
19552 To record restrictions violations (that are performed by the compiler if the
19553 pragma @code{Restrictions} or @code{Restriction_Warnings} are given),
19555 @code{Restrictions} with the same parameters as pragma
19556 @code{Restrictions} or @code{Restriction_Warnings}.
19559 To record compiler style checks, use the rule named
19560 @code{Style_Checks}. A parameter of this rule can be either @code{All_Checks},
19561 which enables all the style checks, or a string that has exactly the same
19562 structure and semantics as the @code{string_LITERAL} parameter of GNAT pragma
19563 @code{Style_Checks} (for further information about this pragma, please
19564 refer to the @cite{@value{EDITION} Reference Manual}).
19567 To record compiler warnings (@pxref{Warning Message Control}), use the rule
19568 named @code{Warnings} with a parameter that is a valid
19569 @i{static_string_expression} argument of GNAT pragma @code{Warnings}
19570 (for further information about this pragma, please
19571 refer to the @cite{@value{EDITION} Reference Manual}).
19575 @node Project-Wide Checks
19576 @section Project-Wide Checks
19577 @cindex Project-wide checks (for @command{gnatcheck})
19580 In order to perform checks on all units of a given project, you can use
19581 the GNAT driver along with the @option{-P} option:
19583 gnat check -Pproj -rules -from=my_rules
19587 If the project @code{proj} depends upon other projects, you can perform
19588 checks on the project closure using the @option{-U} option:
19590 gnat check -Pproj -U -rules -from=my_rules
19594 Finally, if not all the units are relevant to a particular main
19595 program in the project closure, you can perform checks for the set
19596 of units needed to create a given main program (unit closure) using
19597 the @option{-U} option followed by the name of the main unit:
19599 gnat check -Pproj -U main -rules -from=my_rules
19603 @node Predefined Rules
19604 @section Predefined Rules
19605 @cindex Predefined rules (for @command{gnatcheck})
19608 @c (Jan 2007) Since the global rules are still under development and are not
19609 @c documented, there is no point in explaining the difference between
19610 @c global and local rules
19612 A rule in @command{gnatcheck} is either local or global.
19613 A @emph{local rule} is a rule that applies to a well-defined section
19614 of a program and that can be checked by analyzing only this section.
19615 A @emph{global rule} requires analysis of some global properties of the
19616 whole program (mostly related to the program call graph).
19617 As of @value{NOW}, the implementation of global rules should be
19618 considered to be at a preliminary stage. You can use the
19619 @option{+GLOBAL} option to enable all the global rules, and the
19620 @option{-GLOBAL} rule option to disable all the global rules.
19622 All the global rules in the list below are
19623 so indicated by marking them ``GLOBAL''.
19624 This +GLOBAL and -GLOBAL options are not
19625 included in the list of gnatcheck options above, because at the moment they
19626 are considered as a temporary debug options.
19628 @command{gnatcheck} performs rule checks for generic
19629 instances only for global rules. This limitation may be relaxed in a later
19634 The following subsections document the rules implemented in
19635 @command{gnatcheck}.
19636 The subsection title is the same as the rule identifier, which may be
19637 used as a parameter of the @option{+R} or @option{-R} options.
19641 * Abstract_Type_Declarations::
19642 * Anonymous_Arrays::
19643 * Anonymous_Subtypes::
19645 * Boolean_Relational_Operators::
19647 * Ceiling_Violations::
19649 * Controlled_Type_Declarations::
19650 * Declarations_In_Blocks::
19651 * Default_Parameters::
19652 * Discriminated_Records::
19653 * Enumeration_Ranges_In_CASE_Statements::
19654 * Exceptions_As_Control_Flow::
19655 * EXIT_Statements_With_No_Loop_Name::
19656 * Expanded_Loop_Exit_Names::
19657 * Explicit_Full_Discrete_Ranges::
19658 * Float_Equality_Checks::
19659 * Forbidden_Pragmas::
19660 * Function_Style_Procedures::
19661 * Generics_In_Subprograms::
19662 * GOTO_Statements::
19663 * Implicit_IN_Mode_Parameters::
19664 * Implicit_SMALL_For_Fixed_Point_Types::
19665 * Improperly_Located_Instantiations::
19666 * Improper_Returns::
19667 * Library_Level_Subprograms::
19670 * Improperly_Called_Protected_Entries::
19672 * Misnamed_Identifiers::
19673 * Multiple_Entries_In_Protected_Definitions::
19675 * Non_Qualified_Aggregates::
19676 * Non_Short_Circuit_Operators::
19677 * Non_SPARK_Attributes::
19678 * Non_Tagged_Derived_Types::
19679 * Non_Visible_Exceptions::
19680 * Numeric_Literals::
19681 * OTHERS_In_Aggregates::
19682 * OTHERS_In_CASE_Statements::
19683 * OTHERS_In_Exception_Handlers::
19684 * Outer_Loop_Exits::
19685 * Overloaded_Operators::
19686 * Overly_Nested_Control_Structures::
19687 * Parameters_Out_Of_Order::
19688 * Positional_Actuals_For_Defaulted_Generic_Parameters::
19689 * Positional_Actuals_For_Defaulted_Parameters::
19690 * Positional_Components::
19691 * Positional_Generic_Parameters::
19692 * Positional_Parameters::
19693 * Predefined_Numeric_Types::
19694 * Raising_External_Exceptions::
19695 * Raising_Predefined_Exceptions::
19698 * Side_Effect_Functions::
19701 * Unassigned_OUT_Parameters::
19702 * Uncommented_BEGIN_In_Package_Bodies::
19703 * Unconstrained_Array_Returns::
19704 * Universal_Ranges::
19705 * Unnamed_Blocks_And_Loops::
19707 * Unused_Subprograms::
19709 * USE_PACKAGE_Clauses::
19710 * Volatile_Objects_Without_Address_Clauses::
19714 @node Abstract_Type_Declarations
19715 @subsection @code{Abstract_Type_Declarations}
19716 @cindex @code{Abstract_Type_Declarations} rule (for @command{gnatcheck})
19719 Flag all declarations of abstract types. For an abstract private
19720 type, both the private and full type declarations are flagged.
19722 This rule has no parameters.
19725 @node Anonymous_Arrays
19726 @subsection @code{Anonymous_Arrays}
19727 @cindex @code{Anonymous_Arrays} rule (for @command{gnatcheck})
19730 Flag all anonymous array type definitions (by Ada semantics these can only
19731 occur in object declarations).
19733 This rule has no parameters.
19735 @node Anonymous_Subtypes
19736 @subsection @code{Anonymous_Subtypes}
19737 @cindex @code{Anonymous_Subtypes} rule (for @command{gnatcheck})
19740 Flag all uses of anonymous subtypes. A use of an anonymous subtype is
19741 any instance of a subtype indication with a constraint, other than one
19742 that occurs immediately within a subtype declaration. Any use of a range
19743 other than as a constraint used immediately within a subtype declaration
19744 is considered as an anonymous subtype.
19746 An effect of this rule is that @code{for} loops such as the following are
19747 flagged (since @code{1..N} is formally a ``range''):
19749 @smallexample @c ada
19750 for I in 1 .. N loop
19756 Declaring an explicit subtype solves the problem:
19758 @smallexample @c ada
19759 subtype S is Integer range 1..N;
19767 This rule has no parameters.
19770 @subsection @code{Blocks}
19771 @cindex @code{Blocks} rule (for @command{gnatcheck})
19774 Flag each block statement.
19776 This rule has no parameters.
19778 @node Boolean_Relational_Operators
19779 @subsection @code{Boolean_Relational_Operators}
19780 @cindex @code{Boolean_Relational_Operators} rule (for @command{gnatcheck})
19783 Flag each call to a predefined relational operator (``<'', ``>'', ``<='',
19784 ``>='', ``='' and ``/='') for the predefined @code{Boolean} type.
19785 (This rule is useful in enforcing the SPARK language restrictions.)
19787 Calls to predefined relational operators of any type derived from
19788 @code{Standard.Boolean} are not detected. Calls to user-defined functions
19789 with these designators, and uses of operators that are renamings
19790 of the predefined relational operators for @code{Standard.Boolean},
19791 are likewise not detected.
19793 This rule has no parameters.
19796 @node Ceiling_Violations
19797 @subsection @code{Ceiling_Violations} (under construction, GLOBAL)
19798 @cindex @code{Ceiling_Violations} rule (for @command{gnatcheck})
19801 Flag invocations of a protected operation by a task whose priority exceeds
19802 the protected object's ceiling.
19804 As of @value{NOW}, this rule has the following limitations:
19809 We consider only pragmas Priority and Interrupt_Priority as means to define
19810 a task/protected operation priority. We do not consider the effect of using
19811 Ada.Dynamic_Priorities.Set_Priority procedure;
19814 We consider only base task priorities, and no priority inheritance. That is,
19815 we do not make a difference between calls issued during task activation and
19816 execution of the sequence of statements from task body;
19819 Any situation when the priority of protected operation caller is set by a
19820 dynamic expression (that is, the corresponding Priority or
19821 Interrupt_Priority pragma has a non-static expression as an argument) we
19822 treat as a priority inconsistency (and, therefore, detect this situation).
19826 At the moment the notion of the main subprogram is not implemented in
19827 gnatcheck, so any pragma Priority in a library level subprogram body (in case
19828 if this subprogram can be a main subprogram of a partition) changes the
19829 priority of an environment task. So if we have more then one such pragma in
19830 the set of processed sources, the pragma that is processed last, defines the
19831 priority of an environment task.
19833 This rule has no parameters.
19836 @node Controlled_Type_Declarations
19837 @subsection @code{Controlled_Type_Declarations}
19838 @cindex @code{Controlled_Type_Declarations} rule (for @command{gnatcheck})
19841 Flag all declarations of controlled types. A declaration of a private type
19842 is flagged if its full declaration declares a controlled type. A declaration
19843 of a derived type is flagged if its ancestor type is controlled. Subtype
19844 declarations are not checked. A declaration of a type that itself is not a
19845 descendant of a type declared in @code{Ada.Finalization} but has a controlled
19846 component is not checked.
19848 This rule has no parameters.
19852 @node Declarations_In_Blocks
19853 @subsection @code{Declarations_In_Blocks}
19854 @cindex @code{Declarations_In_Blocks} rule (for @command{gnatcheck})
19857 Flag all block statements containing local declarations. A @code{declare}
19858 block with an empty @i{declarative_part} or with a @i{declarative part}
19859 containing only pragmas and/or @code{use} clauses is not flagged.
19861 This rule has no parameters.
19864 @node Default_Parameters
19865 @subsection @code{Default_Parameters}
19866 @cindex @code{Default_Parameters} rule (for @command{gnatcheck})
19869 Flag all default expressions for subprogram parameters. Parameter
19870 declarations of formal and generic subprograms are also checked.
19872 This rule has no parameters.
19875 @node Discriminated_Records
19876 @subsection @code{Discriminated_Records}
19877 @cindex @code{Discriminated_Records} rule (for @command{gnatcheck})
19880 Flag all declarations of record types with discriminants. Only the
19881 declarations of record and record extension types are checked. Incomplete,
19882 formal, private, derived and private extension type declarations are not
19883 checked. Task and protected type declarations also are not checked.
19885 This rule has no parameters.
19888 @node Enumeration_Ranges_In_CASE_Statements
19889 @subsection @code{Enumeration_Ranges_In_CASE_Statements}
19890 @cindex @code{Enumeration_Ranges_In_CASE_Statements} (for @command{gnatcheck})
19893 Flag each use of a range of enumeration literals as a choice in a
19894 @code{case} statement.
19895 All forms for specifying a range (explicit ranges
19896 such as @code{A .. B}, subtype marks and @code{'Range} attributes) are flagged.
19897 An enumeration range is
19898 flagged even if contains exactly one enumeration value or no values at all. A
19899 type derived fom an enumeration type is considered as an enumeration type.
19901 This rule helps prevent maintenance problems arising from adding an
19902 enumeration value to a type and having it implicitly handled by an existing
19903 @code{case} statement with an enumeration range that includes the new literal.
19905 This rule has no parameters.
19908 @node Exceptions_As_Control_Flow
19909 @subsection @code{Exceptions_As_Control_Flow}
19910 @cindex @code{Exceptions_As_Control_Flow} (for @command{gnatcheck})
19913 Flag each place where an exception is explictly raised and handled in the
19914 same subprogram body. A @code{raise} statement in an exception handler,
19915 package body, task body or entry body is not flagged.
19917 The rule has no parameters.
19919 @node EXIT_Statements_With_No_Loop_Name
19920 @subsection @code{EXIT_Statements_With_No_Loop_Name}
19921 @cindex @code{EXIT_Statements_With_No_Loop_Name} (for @command{gnatcheck})
19924 Flag each @code{exit} statement that does not specify the name of the loop
19927 The rule has no parameters.
19930 @node Expanded_Loop_Exit_Names
19931 @subsection @code{Expanded_Loop_Exit_Names}
19932 @cindex @code{Expanded_Loop_Exit_Names} rule (for @command{gnatcheck})
19935 Flag all expanded loop names in @code{exit} statements.
19937 This rule has no parameters.
19939 @node Explicit_Full_Discrete_Ranges
19940 @subsection @code{Explicit_Full_Discrete_Ranges}
19941 @cindex @code{Explicit_Full_Discrete_Ranges} rule (for @command{gnatcheck})
19944 Flag each discrete range that has the form @code{A'First .. A'Last}.
19946 This rule has no parameters.
19948 @node Float_Equality_Checks
19949 @subsection @code{Float_Equality_Checks}
19950 @cindex @code{Float_Equality_Checks} rule (for @command{gnatcheck})
19953 Flag all calls to the predefined equality operations for floating-point types.
19954 Both ``@code{=}'' and ``@code{/=}'' operations are checked.
19955 User-defined equality operations are not flagged, nor are ``@code{=}''
19956 and ``@code{/=}'' operations for fixed-point types.
19958 This rule has no parameters.
19961 @node Forbidden_Pragmas
19962 @subsection @code{Forbidden_Pragmas}
19963 @cindex @code{Forbidden_Pragmas} rule (for @command{gnatcheck})
19966 Flag each use of the specified pragmas. The pragmas to be detected
19967 are named in the rule's parameters.
19969 This rule has the following parameters:
19972 @item For the @option{+R} option
19975 @item @emph{Pragma_Name}
19976 Adds the specified pragma to the set of pragmas to be
19977 checked and sets the checks for all the specified pragmas
19978 ON. @emph{Pragma_Name} is treated as a name of a pragma. If it
19979 does not correspond to any pragma name defined in the Ada
19980 standard or to the name of a GNAT-specific pragma defined
19981 in the GNAT Reference Manual, it is treated as the name of
19985 All the GNAT-specific pragmas are detected; this sets
19986 the checks for all the specified pragmas ON.
19989 All pragmas are detected; this sets the rule ON.
19992 @item For the @option{-R} option
19994 @item @emph{Pragma_Name}
19995 Removes the specified pragma from the set of pragmas to be
19996 checked without affecting checks for
19997 other pragmas. @emph{Pragma_Name} is treated as a name
19998 of a pragma. If it does not correspond to any pragma
19999 defined in the Ada standard or to any name defined in the
20000 GNAT Reference Manual,
20001 this option is treated as turning OFF detection of all
20005 Turn OFF detection of all GNAT-specific pragmas
20008 Clear the list of the pragmas to be detected and
20014 Parameters are not case sensitive. If @emph{Pragma_Name} does not have
20015 the syntax of an Ada identifier and therefore can not be considered
20016 as a pragma name, a diagnostic message is generated and the corresponding
20017 parameter is ignored.
20019 When more then one parameter is given in the same rule option, the parameters
20020 must be separated by a comma.
20022 If more then one option for this rule is specified for the @command{gnatcheck}
20023 call, a new option overrides the previous one(s).
20025 The @option{+R} option with no parameters turns the rule ON with the set of
20026 pragmas to be detected defined by the previous rule options.
20027 (By default this set is empty, so if the only option specified for the rule is
20028 @option{+RForbidden_Pragmas} (with
20029 no parameter), then the rule is enabled, but it does not detect anything).
20030 The @option{-R} option with no parameter turns the rule OFF, but it does not
20031 affect the set of pragmas to be detected.
20036 @node Function_Style_Procedures
20037 @subsection @code{Function_Style_Procedures}
20038 @cindex @code{Function_Style_Procedures} rule (for @command{gnatcheck})
20041 Flag each procedure that can be rewritten as a function. A procedure can be
20042 converted into a function if it has exactly one parameter of mode @code{out}
20043 and no parameters of mode @code{in out}. Procedure declarations,
20044 formal procedure declarations. and generic procedure declarations are always
20046 bodies and body stubs are flagged only if they do not have corresponding
20047 separate declarations. Procedure renamings and procedure instantiations are
20050 If a procedure can be rewritten as a fucntion, but its @code{out} parameter is
20051 of a limited type, it is not flagged.
20053 Protected procedures are not flagged. Null procedures also are not flagged.
20055 This rule has no parameters.
20058 @node Generics_In_Subprograms
20059 @subsection @code{Generics_In_Subprograms}
20060 @cindex @code{Generics_In_Subprograms} rule (for @command{gnatcheck})
20063 Flag each declaration of a generic unit in a supbrogram. Generic
20064 declarations in the bodies of generic subprograms are also flagged.
20065 A generic unit nested in another generic unit is not flagged.
20066 If a generic unit is
20067 declared in a local package that is declared in a subprogram body, the
20068 generic unit is flagged.
20070 This rule has no parameters.
20073 @node GOTO_Statements
20074 @subsection @code{GOTO_Statements}
20075 @cindex @code{GOTO_Statements} rule (for @command{gnatcheck})
20078 Flag each occurrence of a @code{goto} statement.
20080 This rule has no parameters.
20083 @node Implicit_IN_Mode_Parameters
20084 @subsection @code{Implicit_IN_Mode_Parameters}
20085 @cindex @code{Implicit_IN_Mode_Parameters} rule (for @command{gnatcheck})
20088 Flag each occurrence of a formal parameter with an implicit @code{in} mode.
20089 Note that @code{access} parameters, although they technically behave
20090 like @code{in} parameters, are not flagged.
20092 This rule has no parameters.
20095 @node Implicit_SMALL_For_Fixed_Point_Types
20096 @subsection @code{Implicit_SMALL_For_Fixed_Point_Types}
20097 @cindex @code{Implicit_SMALL_For_Fixed_Point_Types} rule (for @command{gnatcheck})
20100 Flag each fixed point type declaration that lacks an explicit
20101 representation clause to define its @code{'Small} value.
20102 Since @code{'Small} can be defined only for ordinary fixed point types,
20103 decimal fixed point type declarations are not checked.
20105 This rule has no parameters.
20108 @node Improperly_Located_Instantiations
20109 @subsection @code{Improperly_Located_Instantiations}
20110 @cindex @code{Improperly_Located_Instantiations} rule (for @command{gnatcheck})
20113 Flag all generic instantiations in library-level package specifications
20114 (including library generic packages) and in all subprogram bodies.
20116 Instantiations in task and entry bodies are not flagged. Instantiations in the
20117 bodies of protected subprograms are flagged.
20119 This rule has no parameters.
20123 @node Improper_Returns
20124 @subsection @code{Improper_Returns}
20125 @cindex @code{Improper_Returns} rule (for @command{gnatcheck})
20128 Flag each explicit @code{return} statement in procedures, and
20129 multiple @code{return} statements in functions.
20130 Diagnostic messages are generated for all @code{return} statements
20131 in a procedure (thus each procedure must be written so that it
20132 returns implicitly at the end of its statement part),
20133 and for all @code{return} statements in a function after the first one.
20134 This rule supports the stylistic convention that each subprogram
20135 should have no more than one point of normal return.
20137 This rule has no parameters.
20140 @node Library_Level_Subprograms
20141 @subsection @code{Library_Level_Subprograms}
20142 @cindex @code{Library_Level_Subprograms} rule (for @command{gnatcheck})
20145 Flag all library-level subprograms (including generic subprogram instantiations).
20147 This rule has no parameters.
20150 @node Local_Packages
20151 @subsection @code{Local_Packages}
20152 @cindex @code{Local_Packages} rule (for @command{gnatcheck})
20155 Flag all local packages declared in package and generic package
20157 Local packages in bodies are not flagged.
20159 This rule has no parameters.
20162 @node Improperly_Called_Protected_Entries
20163 @subsection @code{Improperly_Called_Protected_Entries} (under construction, GLOBAL)
20164 @cindex @code{Improperly_Called_Protected_Entries} rule (for @command{gnatcheck})
20167 Flag each protected entry that can be called from more than one task.
20169 This rule has no parameters.
20173 @node Misnamed_Identifiers
20174 @subsection @code{Misnamed_Identifiers}
20175 @cindex @code{Misnamed_Identifiers} rule (for @command{gnatcheck})
20178 Flag the declaration of each identifier that does not have a suffix
20179 corresponding to the kind of entity being declared.
20180 The following declarations are checked:
20187 constant declarations (but not number declarations)
20190 package renaming declarations (but not generic package renaming
20195 This rule may have parameters. When used without parameters, the rule enforces
20196 the following checks:
20200 type-defining names end with @code{_T}, unless the type is an access type,
20201 in which case the suffix must be @code{_A}
20203 constant names end with @code{_C}
20205 names defining package renamings end with @code{_R}
20209 For a private or incomplete type declaration the following checks are
20210 made for the defining name suffix:
20214 For an incomplete type declaration: if the corresponding full type
20215 declaration is available, the defining identifier from the full type
20216 declaration is checked, but the defining identifier from the incomplete type
20217 declaration is not; otherwise the defining identifier from the incomplete
20218 type declaration is checked against the suffix specified for type
20222 For a private type declaration (including private extensions), the defining
20223 identifier from the private type declaration is checked against the type
20224 suffix (even if the corresponding full declaration is an access type
20225 declaration), and the defining identifier from the corresponding full type
20226 declaration is not checked.
20230 For a deferred constant, the defining name in the corresponding full constant
20231 declaration is not checked.
20233 Defining names of formal types are not checked.
20235 The rule may have the following parameters:
20239 For the @option{+R} option:
20242 Sets the default listed above for all the names to be checked.
20244 @item Type_Suffix=@emph{string}
20245 Specifies the suffix for a type name.
20247 @item Access_Suffix=@emph{string}
20248 Specifies the suffix for an access type name. If
20249 this parameter is set, it overrides for access
20250 types the suffix set by the @code{Type_Suffix} parameter.
20252 @item Constant_Suffix=@emph{string}
20253 Specifies the suffix for a constant name.
20255 @item Renaming_Suffix=@emph{string}
20256 Specifies the suffix for a package renaming name.
20260 For the @option{-R} option:
20263 Remove all the suffixes specified for the
20264 identifier suffix checks, whether by default or
20265 as specified by other rule parameters. All the
20266 checks for this rule are disabled as a result.
20269 Removes the suffix specified for types. This
20270 disables checks for types but does not disable
20271 any other checks for this rule (including the
20272 check for access type names if @code{Access_Suffix} is
20275 @item Access_Suffix
20276 Removes the suffix specified for access types.
20277 This disables checks for access type names but
20278 does not disable any other checks for this rule.
20279 If @code{Type_Suffix} is set, access type names are
20280 checked as ordinary type names.
20282 @item Constant_Suffix
20283 Removes the suffix specified for constants. This
20284 disables checks for constant names but does not
20285 disable any other checks for this rule.
20287 @item Renaming_Suffix
20288 Removes the suffix specified for package
20289 renamings. This disables checks for package
20290 renamings but does not disable any other checks
20296 If more than one parameter is used, parameters must be separated by commas.
20298 If more than one option is specified for the @command{gnatcheck} invocation,
20299 a new option overrides the previous one(s).
20301 The @option{+RMisnamed_Identifiers} option (with no parameter) enables
20303 name suffixes specified by previous options used for this rule.
20305 The @option{-RMisnamed_Identifiers} option (with no parameter) disables
20306 all the checks but keeps
20307 all the suffixes specified by previous options used for this rule.
20309 The @emph{string} value must be a valid suffix for an Ada identifier (after
20310 trimming all the leading and trailing space characters, if any).
20311 Parameters are not case sensitive, except the @emph{string} part.
20313 If any error is detected in a rule parameter, the parameter is ignored.
20314 In such a case the options that are set for the rule are not
20319 @node Multiple_Entries_In_Protected_Definitions
20320 @subsection @code{Multiple_Entries_In_Protected_Definitions}
20321 @cindex @code{Multiple_Entries_In_Protected_Definitions} rule (for @command{gnatcheck})
20324 Flag each protected definition (i.e., each protected object/type declaration)
20325 that defines more than one entry.
20326 Diagnostic messages are generated for all the entry declarations
20327 except the first one. An entry family is counted as one entry. Entries from
20328 the private part of the protected definition are also checked.
20330 This rule has no parameters.
20333 @subsection @code{Name_Clashes}
20334 @cindex @code{Name_Clashes} rule (for @command{gnatcheck})
20337 Check that certain names are not used as defining identifiers. To activate
20338 this rule, you need to supply a reference to the dictionary file(s) as a rule
20339 parameter(s) (more then one dictionary file can be specified). If no
20340 dictionary file is set, this rule will not cause anything to be flagged.
20341 Only defining occurrences, not references, are checked.
20342 The check is not case-sensitive.
20344 This rule is enabled by default, but without setting any corresponding
20345 dictionary file(s); thus the default effect is to do no checks.
20347 A dictionary file is a plain text file. The maximum line length for this file
20348 is 1024 characters. If the line is longer then this limit, extra characters
20351 Each line can be either an empty line, a comment line, or a line containing
20352 a list of identifiers separated by space or HT characters.
20353 A comment is an Ada-style comment (from @code{--} to end-of-line).
20354 Identifiers must follow the Ada syntax for identifiers.
20355 A line containing one or more identifiers may end with a comment.
20357 @node Non_Qualified_Aggregates
20358 @subsection @code{Non_Qualified_Aggregates}
20359 @cindex @code{Non_Qualified_Aggregates} rule (for @command{gnatcheck})
20362 Flag each non-qualified aggregate.
20363 A non-qualified aggregate is an
20364 aggregate that is not the expression of a qualified expression. A
20365 string literal is not considered an aggregate, but an array
20366 aggregate of a string type is considered as a normal aggregate.
20368 This rule has no parameters.
20371 @node Non_Short_Circuit_Operators
20372 @subsection @code{Non_Short_Circuit_Operators}
20373 @cindex @code{Non_Short_Circuit_Operators} rule (for @command{gnatcheck})
20376 Flag all calls to predefined @code{and} and @code{or} operators for
20377 any boolean type. Calls to
20378 user-defined @code{and} and @code{or} and to operators defined by renaming
20379 declarations are not flagged. Calls to predefined @code{and} and @code{or}
20380 operators for modular types or boolean array types are not flagged.
20382 This rule has no parameters.
20386 @node Non_SPARK_Attributes
20387 @subsection @code{Non_SPARK_Attributes}
20388 @cindex @code{Non_SPARK_Attributes} rule (for @command{gnatcheck})
20391 The SPARK language defines the following subset of Ada 95 attribute
20392 designators as those that can be used in SPARK programs. The use of
20393 any other attribute is flagged.
20396 @item @code{'Adjacent}
20399 @item @code{'Ceiling}
20400 @item @code{'Component_Size}
20401 @item @code{'Compose}
20402 @item @code{'Copy_Sign}
20403 @item @code{'Delta}
20404 @item @code{'Denorm}
20405 @item @code{'Digits}
20406 @item @code{'Exponent}
20407 @item @code{'First}
20408 @item @code{'Floor}
20410 @item @code{'Fraction}
20412 @item @code{'Leading_Part}
20413 @item @code{'Length}
20414 @item @code{'Machine}
20415 @item @code{'Machine_Emax}
20416 @item @code{'Machine_Emin}
20417 @item @code{'Machine_Mantissa}
20418 @item @code{'Machine_Overflows}
20419 @item @code{'Machine_Radix}
20420 @item @code{'Machine_Rounds}
20423 @item @code{'Model}
20424 @item @code{'Model_Emin}
20425 @item @code{'Model_Epsilon}
20426 @item @code{'Model_Mantissa}
20427 @item @code{'Model_Small}
20428 @item @code{'Modulus}
20431 @item @code{'Range}
20432 @item @code{'Remainder}
20433 @item @code{'Rounding}
20434 @item @code{'Safe_First}
20435 @item @code{'Safe_Last}
20436 @item @code{'Scaling}
20437 @item @code{'Signed_Zeros}
20439 @item @code{'Small}
20441 @item @code{'Truncation}
20442 @item @code{'Unbiased_Rounding}
20444 @item @code{'Valid}
20448 This rule has no parameters.
20451 @node Non_Tagged_Derived_Types
20452 @subsection @code{Non_Tagged_Derived_Types}
20453 @cindex @code{Non_Tagged_Derived_Types} rule (for @command{gnatcheck})
20456 Flag all derived type declarations that do not have a record extension part.
20458 This rule has no parameters.
20462 @node Non_Visible_Exceptions
20463 @subsection @code{Non_Visible_Exceptions}
20464 @cindex @code{Non_Visible_Exceptions} rule (for @command{gnatcheck})
20467 Flag constructs leading to the possibility of propagating an exception
20468 out of the scope in which the exception is declared.
20469 Two cases are detected:
20473 An exception declaration in a subprogram body, task body or block
20474 statement is flagged if the body or statement does not contain a handler for
20475 that exception or a handler with an @code{others} choice.
20478 A @code{raise} statement in an exception handler of a subprogram body,
20479 task body or block statement is flagged if it (re)raises a locally
20480 declared exception. This may occur under the following circumstances:
20483 it explicitly raises a locally declared exception, or
20485 it does not specify an exception name (i.e., it is simply @code{raise;})
20486 and the enclosing handler contains a locally declared exception in its
20492 Renamings of local exceptions are not flagged.
20494 This rule has no parameters.
20497 @node Numeric_Literals
20498 @subsection @code{Numeric_Literals}
20499 @cindex @code{Numeric_Literals} rule (for @command{gnatcheck})
20502 Flag each use of a numeric literal in an index expression, and in any
20503 circumstance except for the following:
20507 a literal occurring in the initialization expression for a constant
20508 declaration or a named number declaration, or
20511 an integer literal that is less than or equal to a value
20512 specified by the @option{N} rule parameter.
20516 This rule may have the following parameters for the @option{+R} option:
20520 @emph{N} is an integer literal used as the maximal value that is not flagged
20521 (i.e., integer literals not exceeding this value are allowed)
20524 All integer literals are flagged
20528 If no parameters are set, the maximum unflagged value is 1.
20530 The last specified check limit (or the fact that there is no limit at
20531 all) is used when multiple @option{+R} options appear.
20533 The @option{-R} option for this rule has no parameters.
20534 It disables the rule but retains the last specified maximum unflagged value.
20535 If the @option{+R} option subsequently appears, this value is used as the
20536 threshold for the check.
20539 @node OTHERS_In_Aggregates
20540 @subsection @code{OTHERS_In_Aggregates}
20541 @cindex @code{OTHERS_In_Aggregates} rule (for @command{gnatcheck})
20544 Flag each use of an @code{others} choice in extension aggregates.
20545 In record and array aggregates, an @code{others} choice is flagged unless
20546 it is used to refer to all components, or to all but one component.
20548 If, in case of a named array aggregate, there are two associations, one
20549 with an @code{others} choice and another with a discrete range, the
20550 @code{others} choice is flagged even if the discrete range specifies
20551 exactly one component; for example, @code{(1..1 => 0, others => 1)}.
20553 This rule has no parameters.
20555 @node OTHERS_In_CASE_Statements
20556 @subsection @code{OTHERS_In_CASE_Statements}
20557 @cindex @code{OTHERS_In_CASE_Statements} rule (for @command{gnatcheck})
20560 Flag any use of an @code{others} choice in a @code{case} statement.
20562 This rule has no parameters.
20564 @node OTHERS_In_Exception_Handlers
20565 @subsection @code{OTHERS_In_Exception_Handlers}
20566 @cindex @code{OTHERS_In_Exception_Handlers} rule (for @command{gnatcheck})
20569 Flag any use of an @code{others} choice in an exception handler.
20571 This rule has no parameters.
20574 @node Outer_Loop_Exits
20575 @subsection @code{Outer_Loop_Exits}
20576 @cindex @code{Outer_Loop_Exits} rule (for @command{gnatcheck})
20579 Flag each @code{exit} statement containing a loop name that is not the name
20580 of the immediately enclosing @code{loop} statement.
20582 This rule has no parameters.
20585 @node Overloaded_Operators
20586 @subsection @code{Overloaded_Operators}
20587 @cindex @code{Overloaded_Operators} rule (for @command{gnatcheck})
20590 Flag each function declaration that overloads an operator symbol.
20591 A function body is checked only if the body does not have a
20592 separate spec. Formal functions are also checked. For a
20593 renaming declaration, only renaming-as-declaration is checked
20595 This rule has no parameters.
20598 @node Overly_Nested_Control_Structures
20599 @subsection @code{Overly_Nested_Control_Structures}
20600 @cindex @code{Overly_Nested_Control_Structures} rule (for @command{gnatcheck})
20603 Flag each control structure whose nesting level exceeds the value provided
20604 in the rule parameter.
20606 The control structures checked are the following:
20609 @item @code{if} statement
20610 @item @code{case} statement
20611 @item @code{loop} statement
20612 @item Selective accept statement
20613 @item Timed entry call statement
20614 @item Conditional entry call
20615 @item Asynchronous select statement
20619 The rule may have the following parameter for the @option{+R} option:
20623 Positive integer specifying the maximal control structure nesting
20624 level that is not flagged
20628 If the parameter for the @option{+R} option is not a positive integer,
20629 the parameter is ignored and the rule is turned ON with the most recently
20630 specified maximal non-flagged nesting level.
20632 If more then one option is specified for the gnatcheck call, the later option and
20633 new parameter override the previous one(s).
20635 A @option{+R} option with no parameter turns the rule ON using the maximal
20636 non-flagged nesting level specified by the most recent @option{+R} option with
20637 a parameter, or the value 4 if there is no such previous @option{+R} option.
20641 @node Parameters_Out_Of_Order
20642 @subsection @code{Parameters_Out_Of_Order}
20643 @cindex @code{Parameters_Out_Of_Order} rule (for @command{gnatcheck})
20646 Flag each subprogram and entry declaration whose formal parameters are not
20647 ordered according to the following scheme:
20651 @item @code{in} and @code{access} parameters first,
20652 then @code{in out} parameters,
20653 and then @code{out} parameters;
20655 @item for @code{in} mode, parameters with default initialization expressions
20660 Only the first violation of the described order is flagged.
20662 The following constructs are checked:
20665 @item subprogram declarations (including null procedures);
20666 @item generic subprogram declarations;
20667 @item formal subprogram declarations;
20668 @item entry declarations;
20669 @item subprogram bodies and subprogram body stubs that do not
20670 have separate specifications
20674 Subprogram renamings are not checked.
20676 This rule has no parameters.
20679 @node Positional_Actuals_For_Defaulted_Generic_Parameters
20680 @subsection @code{Positional_Actuals_For_Defaulted_Generic_Parameters}
20681 @cindex @code{Positional_Actuals_For_Defaulted_Generic_Parameters} rule (for @command{gnatcheck})
20684 Flag each generic actual parameter corresponding to a generic formal
20685 parameter with a default initialization, if positional notation is used.
20687 This rule has no parameters.
20689 @node Positional_Actuals_For_Defaulted_Parameters
20690 @subsection @code{Positional_Actuals_For_Defaulted_Parameters}
20691 @cindex @code{Positional_Actuals_For_Defaulted_Parameters} rule (for @command{gnatcheck})
20694 Flag each actual parameter to a subprogram or entry call where the
20695 corresponding formal parameter has a default expression, if positional
20698 This rule has no parameters.
20700 @node Positional_Components
20701 @subsection @code{Positional_Components}
20702 @cindex @code{Positional_Components} rule (for @command{gnatcheck})
20705 Flag each array, record and extension aggregate that includes positional
20708 This rule has no parameters.
20711 @node Positional_Generic_Parameters
20712 @subsection @code{Positional_Generic_Parameters}
20713 @cindex @code{Positional_Generic_Parameters} rule (for @command{gnatcheck})
20716 Flag each instantiation using positional parameter notation.
20718 This rule has no parameters.
20721 @node Positional_Parameters
20722 @subsection @code{Positional_Parameters}
20723 @cindex @code{Positional_Parameters} rule (for @command{gnatcheck})
20726 Flag each subprogram or entry call using positional parameter notation,
20727 except for the following:
20731 Invocations of prefix or infix operators are not flagged
20733 If the called subprogram or entry has only one formal parameter,
20734 the call is not flagged;
20736 If a subprogram call uses the @emph{Object.Operation} notation, then
20739 the first parameter (that is, @emph{Object}) is not flagged;
20741 if the called subprogram has only two parameters, the second parameter
20742 of the call is not flagged;
20747 This rule has no parameters.
20752 @node Predefined_Numeric_Types
20753 @subsection @code{Predefined_Numeric_Types}
20754 @cindex @code{Predefined_Numeric_Types} rule (for @command{gnatcheck})
20757 Flag each explicit use of the name of any numeric type or subtype defined
20758 in package @code{Standard}.
20760 The rationale for this rule is to detect when the
20761 program may depend on platform-specific characteristics of the implementation
20762 of the predefined numeric types. Note that this rule is over-pessimistic;
20763 for example, a program that uses @code{String} indexing
20764 likely needs a variable of type @code{Integer}.
20765 Another example is the flagging of predefined numeric types with explicit
20768 @smallexample @c ada
20769 subtype My_Integer is Integer range Left .. Right;
20770 Vy_Var : My_Integer;
20774 This rule detects only numeric types and subtypes defined in
20775 @code{Standard}. The use of numeric types and subtypes defined in other
20776 predefined packages (such as @code{System.Any_Priority} or
20777 @code{Ada.Text_IO.Count}) is not flagged
20779 This rule has no parameters.
20783 @node Raising_External_Exceptions
20784 @subsection @code{Raising_External_Exceptions}
20785 @cindex @code{Raising_External_Exceptions} rule (for @command{gnatcheck})
20788 Flag any @code{raise} statement, in a program unit declared in a library
20789 package or in a generic library package, for an exception that is
20790 neither a predefined exception nor an exception that is also declared (or
20791 renamed) in the visible part of the package.
20793 This rule has no parameters.
20797 @node Raising_Predefined_Exceptions
20798 @subsection @code{Raising_Predefined_Exceptions}
20799 @cindex @code{Raising_Predefined_Exceptions} rule (for @command{gnatcheck})
20802 Flag each @code{raise} statement that raises a predefined exception
20803 (i.e., one of the exceptions @code{Constraint_Error}, @code{Numeric_Error},
20804 @code{Program_Error}, @code{Storage_Error}, or @code{Tasking_Error}).
20806 This rule has no parameters.
20811 @subsection @code{Recursion} (under construction, GLOBAL)
20812 @cindex @code{Recursion} rule (for @command{gnatcheck})
20815 Flag recursive subprograms (cycles in the call graph). Declarations, and not
20816 calls, of recursive subprograms are detected.
20818 This rule has no parameters.
20822 @node Side_Effect_Functions
20823 @subsection @code{Side_Effect_Functions} (under construction, GLOBAL)
20824 @cindex @code{Side_Effect_Functions} rule (for @command{gnatcheck})
20827 Flag functions with side effects.
20829 We define a side effect as changing any data object that is not local for the
20830 body of this function.
20832 At the moment, we do NOT consider a side effect any input-output operations
20833 (changing a state or a content of any file).
20835 We do not consider protected functions for this rule (???)
20837 There are the following sources of side effect:
20840 @item Explicit (or direct) side-effect:
20844 direct assignment to a non-local variable;
20847 direct call to an entity that is known to change some data object that is
20848 not local for the body of this function (Note, that if F1 calls F2 and F2
20849 does have a side effect, this does not automatically mean that F1 also
20850 have a side effect, because it may be the case that F2 is declared in
20851 F1's body and it changes some data object that is global for F2, but
20855 @item Indirect side-effect:
20858 Subprogram calls implicitly issued by:
20861 computing initialization expressions from type declarations as a part
20862 of object elaboration or allocator evaluation;
20864 computing implicit parameters of subprogram or entry calls or generic
20869 activation of a task that change some non-local data object (directly or
20873 elaboration code of a package that is a result of a package instantiation;
20876 controlled objects;
20879 @item Situations when we can suspect a side-effect, but the full static check
20880 is either impossible or too hard:
20883 assignment to access variables or to the objects pointed by access
20887 call to a subprogram pointed by access-to-subprogram value
20895 This rule has no parameters.
20899 @subsection @code{Slices}
20900 @cindex @code{Slices} rule (for @command{gnatcheck})
20903 Flag all uses of array slicing
20905 This rule has no parameters.
20908 @node Unassigned_OUT_Parameters
20909 @subsection @code{Unassigned_OUT_Parameters}
20910 @cindex @code{Unassigned_OUT_Parameters} rule (for @command{gnatcheck})
20913 Flags procedures' @code{out} parameters that are not assigned, and
20914 identifies the contexts in which the assignments are missing.
20916 An @code{out} parameter is flagged in the statements in the procedure
20917 body's handled sequence of statements (before the procedure body's
20918 @code{exception} part, if any) if this sequence of statements contains
20919 no assignments to the parameter.
20921 An @code{out} parameter is flagged in an exception handler in the exception
20922 part of the procedure body's handled sequence of statements if the handler
20923 contains no assignment to the parameter.
20925 Bodies of generic procedures are also considered.
20927 The following are treated as assignments to an @code{out} parameter:
20931 an assignment statement, with the parameter or some component as the target;
20934 passing the parameter (or one of its components) as an @code{out} or
20935 @code{in out} parameter.
20939 This rule does not have any parameters.
20943 @node Uncommented_BEGIN_In_Package_Bodies
20944 @subsection @code{Uncommented_BEGIN_In_Package_Bodies}
20945 @cindex @code{Uncommented_BEGIN_In_Package_Bodies} rule (for @command{gnatcheck})
20948 Flags each package body with declarations and a statement part that does not
20949 include a trailing comment on the line containing the @code{begin} keyword;
20950 this trailing comment needs to specify the package name and nothing else.
20951 The @code{begin} is not flagged if the package body does not
20952 contain any declarations.
20954 If the @code{begin} keyword is placed on the
20955 same line as the last declaration or the first statement, it is flagged
20956 independently of whether the line contains a trailing comment. The
20957 diagnostic message is attached to the line containing the first statement.
20959 This rule has no parameters.
20962 @node Unconstrained_Array_Returns
20963 @subsection @code{Unconstrained_Array_Returns}
20964 @cindex @code{Unconstrained_Array_Returns} rule (for @command{gnatcheck})
20967 Flag each function returning an unconstrained array. Function declarations,
20968 function bodies (and body stubs) having no separate specifications,
20969 and generic function instantiations are checked.
20970 Generic function declarations, function calls and function renamings are
20973 This rule has no parameters.
20975 @node Universal_Ranges
20976 @subsection @code{Universal_Ranges}
20977 @cindex @code{Universal_Ranges} rule (for @command{gnatcheck})
20980 Flag discrete ranges that are a part of an index constraint, constrained
20981 array definition, or @code{for}-loop parameter specification, and whose bounds
20982 are both of type @i{universal_integer}. Ranges that have at least one
20983 bound of a specific type (such as @code{1 .. N}, where @code{N} is a variable
20984 or an expression of non-universal type) are not flagged.
20986 This rule has no parameters.
20989 @node Unnamed_Blocks_And_Loops
20990 @subsection @code{Unnamed_Blocks_And_Loops}
20991 @cindex @code{Unnamed_Blocks_And_Loops} rule (for @command{gnatcheck})
20994 Flag each unnamed block statement and loop statement.
20996 The rule has no parameters.
21001 @node Unused_Subprograms
21002 @subsection @code{Unused_Subprograms} (under construction, GLOBAL)
21003 @cindex @code{Unused_Subprograms} rule (for @command{gnatcheck})
21006 Flag all unused subprograms.
21008 This rule has no parameters.
21014 @node USE_PACKAGE_Clauses
21015 @subsection @code{USE_PACKAGE_Clauses}
21016 @cindex @code{USE_PACKAGE_Clauses} rule (for @command{gnatcheck})
21019 Flag all @code{use} clauses for packages; @code{use type} clauses are
21022 This rule has no parameters.
21026 @node Volatile_Objects_Without_Address_Clauses
21027 @subsection @code{Volatile_Objects_Without_Address_Clauses}
21028 @cindex @code{Volatile_Objects_Without_Address_Clauses} rule (for @command{gnatcheck})
21031 Flag each volatile object that does not have an address clause.
21033 The following check is made: if the pragma @code{Volatile} is applied to a
21034 data object or to its type, then an address clause must
21035 be supplied for this object.
21037 This rule does not check the components of data objects,
21038 array components that are volatile as a result of the pragma
21039 @code{Volatile_Components}, or objects that are volatile because
21040 they are atomic as a result of pragmas @code{Atomic} or
21041 @code{Atomic_Components}.
21043 Only variable declarations, and not constant declarations, are checked.
21045 This rule has no parameters.
21048 @c *********************************
21049 @node Creating Sample Bodies Using gnatstub
21050 @chapter Creating Sample Bodies Using @command{gnatstub}
21054 @command{gnatstub} creates body stubs, that is, empty but compilable bodies
21055 for library unit declarations.
21057 Note: to invoke @code{gnatstub} with a project file, use the @code{gnat}
21058 driver (see @ref{The GNAT Driver and Project Files}).
21060 To create a body stub, @command{gnatstub} has to compile the library
21061 unit declaration. Therefore, bodies can be created only for legal
21062 library units. Moreover, if a library unit depends semantically upon
21063 units located outside the current directory, you have to provide
21064 the source search path when calling @command{gnatstub}, see the description
21065 of @command{gnatstub} switches below.
21068 * Running gnatstub::
21069 * Switches for gnatstub::
21072 @node Running gnatstub
21073 @section Running @command{gnatstub}
21076 @command{gnatstub} has the command-line interface of the form
21079 $ gnatstub [switches] filename [directory]
21086 is the name of the source file that contains a library unit declaration
21087 for which a body must be created. The file name may contain the path
21089 The file name does not have to follow the GNAT file name conventions. If the
21091 does not follow GNAT file naming conventions, the name of the body file must
21093 explicitly as the value of the @option{^-o^/BODY=^@var{body-name}} option.
21094 If the file name follows the GNAT file naming
21095 conventions and the name of the body file is not provided,
21098 of the body file from the argument file name by replacing the @file{.ads}
21100 with the @file{.adb} suffix.
21103 indicates the directory in which the body stub is to be placed (the default
21108 is an optional sequence of switches as described in the next section
21111 @node Switches for gnatstub
21112 @section Switches for @command{gnatstub}
21118 @cindex @option{^-f^/FULL^} (@command{gnatstub})
21119 If the destination directory already contains a file with the name of the
21121 for the argument spec file, replace it with the generated body stub.
21123 @item ^-hs^/HEADER=SPEC^
21124 @cindex @option{^-hs^/HEADER=SPEC^} (@command{gnatstub})
21125 Put the comment header (i.e., all the comments preceding the
21126 compilation unit) from the source of the library unit declaration
21127 into the body stub.
21129 @item ^-hg^/HEADER=GENERAL^
21130 @cindex @option{^-hg^/HEADER=GENERAL^} (@command{gnatstub})
21131 Put a sample comment header into the body stub.
21135 @cindex @option{-IDIR} (@command{gnatstub})
21137 @cindex @option{-I-} (@command{gnatstub})
21140 @item /NOCURRENT_DIRECTORY
21141 @cindex @option{/NOCURRENT_DIRECTORY} (@command{gnatstub})
21143 ^These switches have ^This switch has^ the same meaning as in calls to
21145 ^They define ^It defines ^ the source search path in the call to
21146 @command{gcc} issued
21147 by @command{gnatstub} to compile an argument source file.
21149 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE=^@var{PATH}
21150 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatstub})
21151 This switch has the same meaning as in calls to @command{gcc}.
21152 It defines the additional configuration file to be passed to the call to
21153 @command{gcc} issued
21154 by @command{gnatstub} to compile an argument source file.
21156 @item ^-gnatyM^/MAX_LINE_LENGTH=^@var{n}
21157 @cindex @option{^-gnatyM^/MAX_LINE_LENGTH^} (@command{gnatstub})
21158 (@var{n} is a non-negative integer). Set the maximum line length in the
21159 body stub to @var{n}; the default is 79. The maximum value that can be
21160 specified is 32767. Note that in the special case of configuration
21161 pragma files, the maximum is always 32767 regardless of whether or
21162 not this switch appears.
21164 @item ^-gnaty^/STYLE_CHECKS=^@var{n}
21165 @cindex @option{^-gnaty^/STYLE_CHECKS=^} (@command{gnatstub})
21166 (@var{n} is a non-negative integer from 1 to 9). Set the indentation level in
21167 the generated body sample to @var{n}.
21168 The default indentation is 3.
21170 @item ^-gnatyo^/ORDERED_SUBPROGRAMS^
21171 @cindex @option{^-gnato^/ORDERED_SUBPROGRAMS^} (@command{gnatstub})
21172 Order local bodies alphabetically. (By default local bodies are ordered
21173 in the same way as the corresponding local specs in the argument spec file.)
21175 @item ^-i^/INDENTATION=^@var{n}
21176 @cindex @option{^-i^/INDENTATION^} (@command{gnatstub})
21177 Same as @option{^-gnaty^/STYLE_CHECKS=^@var{n}}
21179 @item ^-k^/TREE_FILE=SAVE^
21180 @cindex @option{^-k^/TREE_FILE=SAVE^} (@command{gnatstub})
21181 Do not remove the tree file (i.e., the snapshot of the compiler internal
21182 structures used by @command{gnatstub}) after creating the body stub.
21184 @item ^-l^/LINE_LENGTH=^@var{n}
21185 @cindex @option{^-l^/LINE_LENGTH^} (@command{gnatstub})
21186 Same as @option{^-gnatyM^/MAX_LINE_LENGTH=^@var{n}}
21188 @item ^-o^/BODY=^@var{body-name}
21189 @cindex @option{^-o^/BODY^} (@command{gnatstub})
21190 Body file name. This should be set if the argument file name does not
21192 the GNAT file naming
21193 conventions. If this switch is omitted the default name for the body will be
21195 from the argument file name according to the GNAT file naming conventions.
21198 @cindex @option{^-q^/QUIET^} (@command{gnatstub})
21199 Quiet mode: do not generate a confirmation when a body is
21200 successfully created, and do not generate a message when a body is not
21204 @item ^-r^/TREE_FILE=REUSE^
21205 @cindex @option{^-r^/TREE_FILE=REUSE^} (@command{gnatstub})
21206 Reuse the tree file (if it exists) instead of creating it. Instead of
21207 creating the tree file for the library unit declaration, @command{gnatstub}
21208 tries to find it in the current directory and use it for creating
21209 a body. If the tree file is not found, no body is created. This option
21210 also implies @option{^-k^/SAVE^}, whether or not
21211 the latter is set explicitly.
21213 @item ^-t^/TREE_FILE=OVERWRITE^
21214 @cindex @option{^-t^/TREE_FILE=OVERWRITE^} (@command{gnatstub})
21215 Overwrite the existing tree file. If the current directory already
21216 contains the file which, according to the GNAT file naming rules should
21217 be considered as a tree file for the argument source file,
21219 will refuse to create the tree file needed to create a sample body
21220 unless this option is set.
21222 @item ^-v^/VERBOSE^
21223 @cindex @option{^-v^/VERBOSE^} (@command{gnatstub})
21224 Verbose mode: generate version information.
21228 @node Other Utility Programs
21229 @chapter Other Utility Programs
21232 This chapter discusses some other utility programs available in the Ada
21236 * Using Other Utility Programs with GNAT::
21237 * The External Symbol Naming Scheme of GNAT::
21238 * Converting Ada Files to html with gnathtml::
21239 * Installing gnathtml::
21246 @node Using Other Utility Programs with GNAT
21247 @section Using Other Utility Programs with GNAT
21250 The object files generated by GNAT are in standard system format and in
21251 particular the debugging information uses this format. This means
21252 programs generated by GNAT can be used with existing utilities that
21253 depend on these formats.
21256 In general, any utility program that works with C will also often work with
21257 Ada programs generated by GNAT. This includes software utilities such as
21258 gprof (a profiling program), @code{gdb} (the FSF debugger), and utilities such
21262 @node The External Symbol Naming Scheme of GNAT
21263 @section The External Symbol Naming Scheme of GNAT
21266 In order to interpret the output from GNAT, when using tools that are
21267 originally intended for use with other languages, it is useful to
21268 understand the conventions used to generate link names from the Ada
21271 All link names are in all lowercase letters. With the exception of library
21272 procedure names, the mechanism used is simply to use the full expanded
21273 Ada name with dots replaced by double underscores. For example, suppose
21274 we have the following package spec:
21276 @smallexample @c ada
21287 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
21288 the corresponding link name is @code{qrs__mn}.
21290 Of course if a @code{pragma Export} is used this may be overridden:
21292 @smallexample @c ada
21297 pragma Export (Var1, C, External_Name => "var1_name");
21299 pragma Export (Var2, C, Link_Name => "var2_link_name");
21306 In this case, the link name for @var{Var1} is whatever link name the
21307 C compiler would assign for the C function @var{var1_name}. This typically
21308 would be either @var{var1_name} or @var{_var1_name}, depending on operating
21309 system conventions, but other possibilities exist. The link name for
21310 @var{Var2} is @var{var2_link_name}, and this is not operating system
21314 One exception occurs for library level procedures. A potential ambiguity
21315 arises between the required name @code{_main} for the C main program,
21316 and the name we would otherwise assign to an Ada library level procedure
21317 called @code{Main} (which might well not be the main program).
21319 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
21320 names. So if we have a library level procedure such as
21322 @smallexample @c ada
21325 procedure Hello (S : String);
21331 the external name of this procedure will be @var{_ada_hello}.
21334 @node Converting Ada Files to html with gnathtml
21335 @section Converting Ada Files to HTML with @code{gnathtml}
21338 This @code{Perl} script allows Ada source files to be browsed using
21339 standard Web browsers. For installation procedure, see the section
21340 @xref{Installing gnathtml}.
21342 Ada reserved keywords are highlighted in a bold font and Ada comments in
21343 a blue font. Unless your program was compiled with the gcc @option{-gnatx}
21344 switch to suppress the generation of cross-referencing information, user
21345 defined variables and types will appear in a different color; you will
21346 be able to click on any identifier and go to its declaration.
21348 The command line is as follow:
21350 $ perl gnathtml.pl [^switches^options^] ada-files
21354 You can pass it as many Ada files as you want. @code{gnathtml} will generate
21355 an html file for every ada file, and a global file called @file{index.htm}.
21356 This file is an index of every identifier defined in the files.
21358 The available ^switches^options^ are the following ones :
21362 @cindex @option{-83} (@code{gnathtml})
21363 Only the Ada 83 subset of keywords will be highlighted.
21365 @item -cc @var{color}
21366 @cindex @option{-cc} (@code{gnathtml})
21367 This option allows you to change the color used for comments. The default
21368 value is green. The color argument can be any name accepted by html.
21371 @cindex @option{-d} (@code{gnathtml})
21372 If the Ada files depend on some other files (for instance through
21373 @code{with} clauses, the latter files will also be converted to html.
21374 Only the files in the user project will be converted to html, not the files
21375 in the run-time library itself.
21378 @cindex @option{-D} (@code{gnathtml})
21379 This command is the same as @option{-d} above, but @command{gnathtml} will
21380 also look for files in the run-time library, and generate html files for them.
21382 @item -ext @var{extension}
21383 @cindex @option{-ext} (@code{gnathtml})
21384 This option allows you to change the extension of the generated HTML files.
21385 If you do not specify an extension, it will default to @file{htm}.
21388 @cindex @option{-f} (@code{gnathtml})
21389 By default, gnathtml will generate html links only for global entities
21390 ('with'ed units, global variables and types,...). If you specify
21391 @option{-f} on the command line, then links will be generated for local
21394 @item -l @var{number}
21395 @cindex @option{-l} (@code{gnathtml})
21396 If this ^switch^option^ is provided and @var{number} is not 0, then
21397 @code{gnathtml} will number the html files every @var{number} line.
21400 @cindex @option{-I} (@code{gnathtml})
21401 Specify a directory to search for library files (@file{.ALI} files) and
21402 source files. You can provide several -I switches on the command line,
21403 and the directories will be parsed in the order of the command line.
21406 @cindex @option{-o} (@code{gnathtml})
21407 Specify the output directory for html files. By default, gnathtml will
21408 saved the generated html files in a subdirectory named @file{html/}.
21410 @item -p @var{file}
21411 @cindex @option{-p} (@code{gnathtml})
21412 If you are using Emacs and the most recent Emacs Ada mode, which provides
21413 a full Integrated Development Environment for compiling, checking,
21414 running and debugging applications, you may use @file{.gpr} files
21415 to give the directories where Emacs can find sources and object files.
21417 Using this ^switch^option^, you can tell gnathtml to use these files.
21418 This allows you to get an html version of your application, even if it
21419 is spread over multiple directories.
21421 @item -sc @var{color}
21422 @cindex @option{-sc} (@code{gnathtml})
21423 This ^switch^option^ allows you to change the color used for symbol
21425 The default value is red. The color argument can be any name accepted by html.
21427 @item -t @var{file}
21428 @cindex @option{-t} (@code{gnathtml})
21429 This ^switch^option^ provides the name of a file. This file contains a list of
21430 file names to be converted, and the effect is exactly as though they had
21431 appeared explicitly on the command line. This
21432 is the recommended way to work around the command line length limit on some
21437 @node Installing gnathtml
21438 @section Installing @code{gnathtml}
21441 @code{Perl} needs to be installed on your machine to run this script.
21442 @code{Perl} is freely available for almost every architecture and
21443 Operating System via the Internet.
21445 On Unix systems, you may want to modify the first line of the script
21446 @code{gnathtml}, to explicitly tell the Operating system where Perl
21447 is. The syntax of this line is :
21449 #!full_path_name_to_perl
21453 Alternatively, you may run the script using the following command line:
21456 $ perl gnathtml.pl [switches] files
21465 The GNAT distribution provides an Ada 95 template for the HP Language
21466 Sensitive Editor (LSE), a component of DECset. In order to
21467 access it, invoke LSE with the qualifier /ENVIRONMENT=GNU:[LIB]ADA95.ENV.
21474 GNAT supports The HP Performance Coverage Analyzer (PCA), a component
21475 of DECset. To use it proceed as outlined under ``HELP PCA'', except for running
21476 the collection phase with the /DEBUG qualifier.
21479 $ GNAT MAKE /DEBUG <PROGRAM_NAME>
21480 $ DEFINE LIB$DEBUG PCA$COLLECTOR
21481 $ RUN/DEBUG <PROGRAM_NAME>
21486 @node Running and Debugging Ada Programs
21487 @chapter Running and Debugging Ada Programs
21491 This chapter discusses how to debug Ada programs.
21493 It applies to GNAT on the Alpha OpenVMS platform;
21494 for I64 OpenVMS please refer to the @cite{OpenVMS Debugger Manual},
21495 since HP has implemented Ada support in the OpenVMS debugger on I64.
21498 An incorrect Ada program may be handled in three ways by the GNAT compiler:
21502 The illegality may be a violation of the static semantics of Ada. In
21503 that case GNAT diagnoses the constructs in the program that are illegal.
21504 It is then a straightforward matter for the user to modify those parts of
21508 The illegality may be a violation of the dynamic semantics of Ada. In
21509 that case the program compiles and executes, but may generate incorrect
21510 results, or may terminate abnormally with some exception.
21513 When presented with a program that contains convoluted errors, GNAT
21514 itself may terminate abnormally without providing full diagnostics on
21515 the incorrect user program.
21519 * The GNAT Debugger GDB::
21521 * Introduction to GDB Commands::
21522 * Using Ada Expressions::
21523 * Calling User-Defined Subprograms::
21524 * Using the Next Command in a Function::
21527 * Debugging Generic Units::
21528 * GNAT Abnormal Termination or Failure to Terminate::
21529 * Naming Conventions for GNAT Source Files::
21530 * Getting Internal Debugging Information::
21531 * Stack Traceback::
21537 @node The GNAT Debugger GDB
21538 @section The GNAT Debugger GDB
21541 @code{GDB} is a general purpose, platform-independent debugger that
21542 can be used to debug mixed-language programs compiled with @command{gcc},
21543 and in particular is capable of debugging Ada programs compiled with
21544 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
21545 complex Ada data structures.
21547 The manual @cite{Debugging with GDB}
21549 , located in the GNU:[DOCS] directory,
21551 contains full details on the usage of @code{GDB}, including a section on
21552 its usage on programs. This manual should be consulted for full
21553 details. The section that follows is a brief introduction to the
21554 philosophy and use of @code{GDB}.
21556 When GNAT programs are compiled, the compiler optionally writes debugging
21557 information into the generated object file, including information on
21558 line numbers, and on declared types and variables. This information is
21559 separate from the generated code. It makes the object files considerably
21560 larger, but it does not add to the size of the actual executable that
21561 will be loaded into memory, and has no impact on run-time performance. The
21562 generation of debug information is triggered by the use of the
21563 ^-g^/DEBUG^ switch in the gcc or gnatmake command used to carry out
21564 the compilations. It is important to emphasize that the use of these
21565 options does not change the generated code.
21567 The debugging information is written in standard system formats that
21568 are used by many tools, including debuggers and profilers. The format
21569 of the information is typically designed to describe C types and
21570 semantics, but GNAT implements a translation scheme which allows full
21571 details about Ada types and variables to be encoded into these
21572 standard C formats. Details of this encoding scheme may be found in
21573 the file exp_dbug.ads in the GNAT source distribution. However, the
21574 details of this encoding are, in general, of no interest to a user,
21575 since @code{GDB} automatically performs the necessary decoding.
21577 When a program is bound and linked, the debugging information is
21578 collected from the object files, and stored in the executable image of
21579 the program. Again, this process significantly increases the size of
21580 the generated executable file, but it does not increase the size of
21581 the executable program itself. Furthermore, if this program is run in
21582 the normal manner, it runs exactly as if the debug information were
21583 not present, and takes no more actual memory.
21585 However, if the program is run under control of @code{GDB}, the
21586 debugger is activated. The image of the program is loaded, at which
21587 point it is ready to run. If a run command is given, then the program
21588 will run exactly as it would have if @code{GDB} were not present. This
21589 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
21590 entirely non-intrusive until a breakpoint is encountered. If no
21591 breakpoint is ever hit, the program will run exactly as it would if no
21592 debugger were present. When a breakpoint is hit, @code{GDB} accesses
21593 the debugging information and can respond to user commands to inspect
21594 variables, and more generally to report on the state of execution.
21598 @section Running GDB
21601 This section describes how to initiate the debugger.
21602 @c The above sentence is really just filler, but it was otherwise
21603 @c clumsy to get the first paragraph nonindented given the conditional
21604 @c nature of the description
21607 The debugger can be launched from a @code{GPS} menu or
21608 directly from the command line. The description below covers the latter use.
21609 All the commands shown can be used in the @code{GPS} debug console window,
21610 but there are usually more GUI-based ways to achieve the same effect.
21613 The command to run @code{GDB} is
21616 $ ^gdb program^GDB PROGRAM^
21620 where @code{^program^PROGRAM^} is the name of the executable file. This
21621 activates the debugger and results in a prompt for debugger commands.
21622 The simplest command is simply @code{run}, which causes the program to run
21623 exactly as if the debugger were not present. The following section
21624 describes some of the additional commands that can be given to @code{GDB}.
21626 @c *******************************
21627 @node Introduction to GDB Commands
21628 @section Introduction to GDB Commands
21631 @code{GDB} contains a large repertoire of commands. The manual
21632 @cite{Debugging with GDB}
21634 (located in the GNU:[DOCS] directory)
21636 includes extensive documentation on the use
21637 of these commands, together with examples of their use. Furthermore,
21638 the command @var{help} invoked from within @code{GDB} activates a simple help
21639 facility which summarizes the available commands and their options.
21640 In this section we summarize a few of the most commonly
21641 used commands to give an idea of what @code{GDB} is about. You should create
21642 a simple program with debugging information and experiment with the use of
21643 these @code{GDB} commands on the program as you read through the
21647 @item set args @var{arguments}
21648 The @var{arguments} list above is a list of arguments to be passed to
21649 the program on a subsequent run command, just as though the arguments
21650 had been entered on a normal invocation of the program. The @code{set args}
21651 command is not needed if the program does not require arguments.
21654 The @code{run} command causes execution of the program to start from
21655 the beginning. If the program is already running, that is to say if
21656 you are currently positioned at a breakpoint, then a prompt will ask
21657 for confirmation that you want to abandon the current execution and
21660 @item breakpoint @var{location}
21661 The breakpoint command sets a breakpoint, that is to say a point at which
21662 execution will halt and @code{GDB} will await further
21663 commands. @var{location} is
21664 either a line number within a file, given in the format @code{file:linenumber},
21665 or it is the name of a subprogram. If you request that a breakpoint be set on
21666 a subprogram that is overloaded, a prompt will ask you to specify on which of
21667 those subprograms you want to breakpoint. You can also
21668 specify that all of them should be breakpointed. If the program is run
21669 and execution encounters the breakpoint, then the program
21670 stops and @code{GDB} signals that the breakpoint was encountered by
21671 printing the line of code before which the program is halted.
21673 @item breakpoint exception @var{name}
21674 A special form of the breakpoint command which breakpoints whenever
21675 exception @var{name} is raised.
21676 If @var{name} is omitted,
21677 then a breakpoint will occur when any exception is raised.
21679 @item print @var{expression}
21680 This will print the value of the given expression. Most simple
21681 Ada expression formats are properly handled by @code{GDB}, so the expression
21682 can contain function calls, variables, operators, and attribute references.
21685 Continues execution following a breakpoint, until the next breakpoint or the
21686 termination of the program.
21689 Executes a single line after a breakpoint. If the next statement
21690 is a subprogram call, execution continues into (the first statement of)
21691 the called subprogram.
21694 Executes a single line. If this line is a subprogram call, executes and
21695 returns from the call.
21698 Lists a few lines around the current source location. In practice, it
21699 is usually more convenient to have a separate edit window open with the
21700 relevant source file displayed. Successive applications of this command
21701 print subsequent lines. The command can be given an argument which is a
21702 line number, in which case it displays a few lines around the specified one.
21705 Displays a backtrace of the call chain. This command is typically
21706 used after a breakpoint has occurred, to examine the sequence of calls that
21707 leads to the current breakpoint. The display includes one line for each
21708 activation record (frame) corresponding to an active subprogram.
21711 At a breakpoint, @code{GDB} can display the values of variables local
21712 to the current frame. The command @code{up} can be used to
21713 examine the contents of other active frames, by moving the focus up
21714 the stack, that is to say from callee to caller, one frame at a time.
21717 Moves the focus of @code{GDB} down from the frame currently being
21718 examined to the frame of its callee (the reverse of the previous command),
21720 @item frame @var{n}
21721 Inspect the frame with the given number. The value 0 denotes the frame
21722 of the current breakpoint, that is to say the top of the call stack.
21727 The above list is a very short introduction to the commands that
21728 @code{GDB} provides. Important additional capabilities, including conditional
21729 breakpoints, the ability to execute command sequences on a breakpoint,
21730 the ability to debug at the machine instruction level and many other
21731 features are described in detail in @cite{Debugging with GDB}.
21732 Note that most commands can be abbreviated
21733 (for example, c for continue, bt for backtrace).
21735 @node Using Ada Expressions
21736 @section Using Ada Expressions
21737 @cindex Ada expressions
21740 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
21741 extensions. The philosophy behind the design of this subset is
21745 That @code{GDB} should provide basic literals and access to operations for
21746 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
21747 leaving more sophisticated computations to subprograms written into the
21748 program (which therefore may be called from @code{GDB}).
21751 That type safety and strict adherence to Ada language restrictions
21752 are not particularly important to the @code{GDB} user.
21755 That brevity is important to the @code{GDB} user.
21759 Thus, for brevity, the debugger acts as if there were
21760 implicit @code{with} and @code{use} clauses in effect for all user-written
21761 packages, thus making it unnecessary to fully qualify most names with
21762 their packages, regardless of context. Where this causes ambiguity,
21763 @code{GDB} asks the user's intent.
21765 For details on the supported Ada syntax, see @cite{Debugging with GDB}.
21767 @node Calling User-Defined Subprograms
21768 @section Calling User-Defined Subprograms
21771 An important capability of @code{GDB} is the ability to call user-defined
21772 subprograms while debugging. This is achieved simply by entering
21773 a subprogram call statement in the form:
21776 call subprogram-name (parameters)
21780 The keyword @code{call} can be omitted in the normal case where the
21781 @code{subprogram-name} does not coincide with any of the predefined
21782 @code{GDB} commands.
21784 The effect is to invoke the given subprogram, passing it the
21785 list of parameters that is supplied. The parameters can be expressions and
21786 can include variables from the program being debugged. The
21787 subprogram must be defined
21788 at the library level within your program, and @code{GDB} will call the
21789 subprogram within the environment of your program execution (which
21790 means that the subprogram is free to access or even modify variables
21791 within your program).
21793 The most important use of this facility is in allowing the inclusion of
21794 debugging routines that are tailored to particular data structures
21795 in your program. Such debugging routines can be written to provide a suitably
21796 high-level description of an abstract type, rather than a low-level dump
21797 of its physical layout. After all, the standard
21798 @code{GDB print} command only knows the physical layout of your
21799 types, not their abstract meaning. Debugging routines can provide information
21800 at the desired semantic level and are thus enormously useful.
21802 For example, when debugging GNAT itself, it is crucial to have access to
21803 the contents of the tree nodes used to represent the program internally.
21804 But tree nodes are represented simply by an integer value (which in turn
21805 is an index into a table of nodes).
21806 Using the @code{print} command on a tree node would simply print this integer
21807 value, which is not very useful. But the PN routine (defined in file
21808 treepr.adb in the GNAT sources) takes a tree node as input, and displays
21809 a useful high level representation of the tree node, which includes the
21810 syntactic category of the node, its position in the source, the integers
21811 that denote descendant nodes and parent node, as well as varied
21812 semantic information. To study this example in more detail, you might want to
21813 look at the body of the PN procedure in the stated file.
21815 @node Using the Next Command in a Function
21816 @section Using the Next Command in a Function
21819 When you use the @code{next} command in a function, the current source
21820 location will advance to the next statement as usual. A special case
21821 arises in the case of a @code{return} statement.
21823 Part of the code for a return statement is the ``epilog'' of the function.
21824 This is the code that returns to the caller. There is only one copy of
21825 this epilog code, and it is typically associated with the last return
21826 statement in the function if there is more than one return. In some
21827 implementations, this epilog is associated with the first statement
21830 The result is that if you use the @code{next} command from a return
21831 statement that is not the last return statement of the function you
21832 may see a strange apparent jump to the last return statement or to
21833 the start of the function. You should simply ignore this odd jump.
21834 The value returned is always that from the first return statement
21835 that was stepped through.
21837 @node Ada Exceptions
21838 @section Breaking on Ada Exceptions
21842 You can set breakpoints that trip when your program raises
21843 selected exceptions.
21846 @item break exception
21847 Set a breakpoint that trips whenever (any task in the) program raises
21850 @item break exception @var{name}
21851 Set a breakpoint that trips whenever (any task in the) program raises
21852 the exception @var{name}.
21854 @item break exception unhandled
21855 Set a breakpoint that trips whenever (any task in the) program raises an
21856 exception for which there is no handler.
21858 @item info exceptions
21859 @itemx info exceptions @var{regexp}
21860 The @code{info exceptions} command permits the user to examine all defined
21861 exceptions within Ada programs. With a regular expression, @var{regexp}, as
21862 argument, prints out only those exceptions whose name matches @var{regexp}.
21870 @code{GDB} allows the following task-related commands:
21874 This command shows a list of current Ada tasks, as in the following example:
21881 ID TID P-ID Thread Pri State Name
21882 1 8088000 0 807e000 15 Child Activation Wait main_task
21883 2 80a4000 1 80ae000 15 Accept/Select Wait b
21884 3 809a800 1 80a4800 15 Child Activation Wait a
21885 * 4 80ae800 3 80b8000 15 Running c
21889 In this listing, the asterisk before the first task indicates it to be the
21890 currently running task. The first column lists the task ID that is used
21891 to refer to tasks in the following commands.
21893 @item break @var{linespec} task @var{taskid}
21894 @itemx break @var{linespec} task @var{taskid} if @dots{}
21895 @cindex Breakpoints and tasks
21896 These commands are like the @code{break @dots{} thread @dots{}}.
21897 @var{linespec} specifies source lines.
21899 Use the qualifier @samp{task @var{taskid}} with a breakpoint command
21900 to specify that you only want @code{GDB} to stop the program when a
21901 particular Ada task reaches this breakpoint. @var{taskid} is one of the
21902 numeric task identifiers assigned by @code{GDB}, shown in the first
21903 column of the @samp{info tasks} display.
21905 If you do not specify @samp{task @var{taskid}} when you set a
21906 breakpoint, the breakpoint applies to @emph{all} tasks of your
21909 You can use the @code{task} qualifier on conditional breakpoints as
21910 well; in this case, place @samp{task @var{taskid}} before the
21911 breakpoint condition (before the @code{if}).
21913 @item task @var{taskno}
21914 @cindex Task switching
21916 This command allows to switch to the task referred by @var{taskno}. In
21917 particular, This allows to browse the backtrace of the specified
21918 task. It is advised to switch back to the original task before
21919 continuing execution otherwise the scheduling of the program may be
21924 For more detailed information on the tasking support,
21925 see @cite{Debugging with GDB}.
21927 @node Debugging Generic Units
21928 @section Debugging Generic Units
21929 @cindex Debugging Generic Units
21933 GNAT always uses code expansion for generic instantiation. This means that
21934 each time an instantiation occurs, a complete copy of the original code is
21935 made, with appropriate substitutions of formals by actuals.
21937 It is not possible to refer to the original generic entities in
21938 @code{GDB}, but it is always possible to debug a particular instance of
21939 a generic, by using the appropriate expanded names. For example, if we have
21941 @smallexample @c ada
21946 generic package k is
21947 procedure kp (v1 : in out integer);
21951 procedure kp (v1 : in out integer) is
21957 package k1 is new k;
21958 package k2 is new k;
21960 var : integer := 1;
21973 Then to break on a call to procedure kp in the k2 instance, simply
21977 (gdb) break g.k2.kp
21981 When the breakpoint occurs, you can step through the code of the
21982 instance in the normal manner and examine the values of local variables, as for
21985 @node GNAT Abnormal Termination or Failure to Terminate
21986 @section GNAT Abnormal Termination or Failure to Terminate
21987 @cindex GNAT Abnormal Termination or Failure to Terminate
21990 When presented with programs that contain serious errors in syntax
21992 GNAT may on rare occasions experience problems in operation, such
21994 segmentation fault or illegal memory access, raising an internal
21995 exception, terminating abnormally, or failing to terminate at all.
21996 In such cases, you can activate
21997 various features of GNAT that can help you pinpoint the construct in your
21998 program that is the likely source of the problem.
22000 The following strategies are presented in increasing order of
22001 difficulty, corresponding to your experience in using GNAT and your
22002 familiarity with compiler internals.
22006 Run @command{gcc} with the @option{-gnatf}. This first
22007 switch causes all errors on a given line to be reported. In its absence,
22008 only the first error on a line is displayed.
22010 The @option{-gnatdO} switch causes errors to be displayed as soon as they
22011 are encountered, rather than after compilation is terminated. If GNAT
22012 terminates prematurely or goes into an infinite loop, the last error
22013 message displayed may help to pinpoint the culprit.
22016 Run @command{gcc} with the @option{^-v (verbose)^/VERBOSE^} switch. In this
22017 mode, @command{gcc} produces ongoing information about the progress of the
22018 compilation and provides the name of each procedure as code is
22019 generated. This switch allows you to find which Ada procedure was being
22020 compiled when it encountered a code generation problem.
22023 @cindex @option{-gnatdc} switch
22024 Run @command{gcc} with the @option{-gnatdc} switch. This is a GNAT specific
22025 switch that does for the front-end what @option{^-v^VERBOSE^} does
22026 for the back end. The system prints the name of each unit,
22027 either a compilation unit or nested unit, as it is being analyzed.
22029 Finally, you can start
22030 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
22031 front-end of GNAT, and can be run independently (normally it is just
22032 called from @command{gcc}). You can use @code{gdb} on @code{gnat1} as you
22033 would on a C program (but @pxref{The GNAT Debugger GDB} for caveats). The
22034 @code{where} command is the first line of attack; the variable
22035 @code{lineno} (seen by @code{print lineno}), used by the second phase of
22036 @code{gnat1} and by the @command{gcc} backend, indicates the source line at
22037 which the execution stopped, and @code{input_file name} indicates the name of
22041 @node Naming Conventions for GNAT Source Files
22042 @section Naming Conventions for GNAT Source Files
22045 In order to examine the workings of the GNAT system, the following
22046 brief description of its organization may be helpful:
22050 Files with prefix @file{^sc^SC^} contain the lexical scanner.
22053 All files prefixed with @file{^par^PAR^} are components of the parser. The
22054 numbers correspond to chapters of the Ada Reference Manual. For example,
22055 parsing of select statements can be found in @file{par-ch9.adb}.
22058 All files prefixed with @file{^sem^SEM^} perform semantic analysis. The
22059 numbers correspond to chapters of the Ada standard. For example, all
22060 issues involving context clauses can be found in @file{sem_ch10.adb}. In
22061 addition, some features of the language require sufficient special processing
22062 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
22063 dynamic dispatching, etc.
22066 All files prefixed with @file{^exp^EXP^} perform normalization and
22067 expansion of the intermediate representation (abstract syntax tree, or AST).
22068 these files use the same numbering scheme as the parser and semantics files.
22069 For example, the construction of record initialization procedures is done in
22070 @file{exp_ch3.adb}.
22073 The files prefixed with @file{^bind^BIND^} implement the binder, which
22074 verifies the consistency of the compilation, determines an order of
22075 elaboration, and generates the bind file.
22078 The files @file{atree.ads} and @file{atree.adb} detail the low-level
22079 data structures used by the front-end.
22082 The files @file{sinfo.ads} and @file{sinfo.adb} detail the structure of
22083 the abstract syntax tree as produced by the parser.
22086 The files @file{einfo.ads} and @file{einfo.adb} detail the attributes of
22087 all entities, computed during semantic analysis.
22090 Library management issues are dealt with in files with prefix
22096 Ada files with the prefix @file{^a-^A-^} are children of @code{Ada}, as
22097 defined in Annex A.
22102 Files with prefix @file{^i-^I-^} are children of @code{Interfaces}, as
22103 defined in Annex B.
22107 Files with prefix @file{^s-^S-^} are children of @code{System}. This includes
22108 both language-defined children and GNAT run-time routines.
22112 Files with prefix @file{^g-^G-^} are children of @code{GNAT}. These are useful
22113 general-purpose packages, fully documented in their specifications. All
22114 the other @file{.c} files are modifications of common @command{gcc} files.
22117 @node Getting Internal Debugging Information
22118 @section Getting Internal Debugging Information
22121 Most compilers have internal debugging switches and modes. GNAT
22122 does also, except GNAT internal debugging switches and modes are not
22123 secret. A summary and full description of all the compiler and binder
22124 debug flags are in the file @file{debug.adb}. You must obtain the
22125 sources of the compiler to see the full detailed effects of these flags.
22127 The switches that print the source of the program (reconstructed from
22128 the internal tree) are of general interest for user programs, as are the
22130 the full internal tree, and the entity table (the symbol table
22131 information). The reconstructed source provides a readable version of the
22132 program after the front-end has completed analysis and expansion,
22133 and is useful when studying the performance of specific constructs.
22134 For example, constraint checks are indicated, complex aggregates
22135 are replaced with loops and assignments, and tasking primitives
22136 are replaced with run-time calls.
22138 @node Stack Traceback
22139 @section Stack Traceback
22141 @cindex stack traceback
22142 @cindex stack unwinding
22145 Traceback is a mechanism to display the sequence of subprogram calls that
22146 leads to a specified execution point in a program. Often (but not always)
22147 the execution point is an instruction at which an exception has been raised.
22148 This mechanism is also known as @i{stack unwinding} because it obtains
22149 its information by scanning the run-time stack and recovering the activation
22150 records of all active subprograms. Stack unwinding is one of the most
22151 important tools for program debugging.
22153 The first entry stored in traceback corresponds to the deepest calling level,
22154 that is to say the subprogram currently executing the instruction
22155 from which we want to obtain the traceback.
22157 Note that there is no runtime performance penalty when stack traceback
22158 is enabled, and no exception is raised during program execution.
22161 * Non-Symbolic Traceback::
22162 * Symbolic Traceback::
22165 @node Non-Symbolic Traceback
22166 @subsection Non-Symbolic Traceback
22167 @cindex traceback, non-symbolic
22170 Note: this feature is not supported on all platforms. See
22171 @file{GNAT.Traceback spec in g-traceb.ads} for a complete list of supported
22175 * Tracebacks From an Unhandled Exception::
22176 * Tracebacks From Exception Occurrences (non-symbolic)::
22177 * Tracebacks From Anywhere in a Program (non-symbolic)::
22180 @node Tracebacks From an Unhandled Exception
22181 @subsubsection Tracebacks From an Unhandled Exception
22184 A runtime non-symbolic traceback is a list of addresses of call instructions.
22185 To enable this feature you must use the @option{-E}
22186 @code{gnatbind}'s option. With this option a stack traceback is stored as part
22187 of exception information. You can retrieve this information using the
22188 @code{addr2line} tool.
22190 Here is a simple example:
22192 @smallexample @c ada
22198 raise Constraint_Error;
22213 $ gnatmake stb -bargs -E
22216 Execution terminated by unhandled exception
22217 Exception name: CONSTRAINT_ERROR
22219 Call stack traceback locations:
22220 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
22224 As we see the traceback lists a sequence of addresses for the unhandled
22225 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
22226 guess that this exception come from procedure P1. To translate these
22227 addresses into the source lines where the calls appear, the
22228 @code{addr2line} tool, described below, is invaluable. The use of this tool
22229 requires the program to be compiled with debug information.
22232 $ gnatmake -g stb -bargs -E
22235 Execution terminated by unhandled exception
22236 Exception name: CONSTRAINT_ERROR
22238 Call stack traceback locations:
22239 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
22241 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
22242 0x4011f1 0x77e892a4
22244 00401373 at d:/stb/stb.adb:5
22245 0040138B at d:/stb/stb.adb:10
22246 0040139C at d:/stb/stb.adb:14
22247 00401335 at d:/stb/b~stb.adb:104
22248 004011C4 at /build/.../crt1.c:200
22249 004011F1 at /build/.../crt1.c:222
22250 77E892A4 in ?? at ??:0
22254 The @code{addr2line} tool has several other useful options:
22258 to get the function name corresponding to any location
22260 @item --demangle=gnat
22261 to use the gnat decoding mode for the function names. Note that
22262 for binutils version 2.9.x the option is simply @option{--demangle}.
22266 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
22267 0x40139c 0x401335 0x4011c4 0x4011f1
22269 00401373 in stb.p1 at d:/stb/stb.adb:5
22270 0040138B in stb.p2 at d:/stb/stb.adb:10
22271 0040139C in stb at d:/stb/stb.adb:14
22272 00401335 in main at d:/stb/b~stb.adb:104
22273 004011C4 in <__mingw_CRTStartup> at /build/.../crt1.c:200
22274 004011F1 in <mainCRTStartup> at /build/.../crt1.c:222
22278 From this traceback we can see that the exception was raised in
22279 @file{stb.adb} at line 5, which was reached from a procedure call in
22280 @file{stb.adb} at line 10, and so on. The @file{b~std.adb} is the binder file,
22281 which contains the call to the main program.
22282 @xref{Running gnatbind}. The remaining entries are assorted runtime routines,
22283 and the output will vary from platform to platform.
22285 It is also possible to use @code{GDB} with these traceback addresses to debug
22286 the program. For example, we can break at a given code location, as reported
22287 in the stack traceback:
22293 Furthermore, this feature is not implemented inside Windows DLL. Only
22294 the non-symbolic traceback is reported in this case.
22297 (gdb) break *0x401373
22298 Breakpoint 1 at 0x401373: file stb.adb, line 5.
22302 It is important to note that the stack traceback addresses
22303 do not change when debug information is included. This is particularly useful
22304 because it makes it possible to release software without debug information (to
22305 minimize object size), get a field report that includes a stack traceback
22306 whenever an internal bug occurs, and then be able to retrieve the sequence
22307 of calls with the same program compiled with debug information.
22309 @node Tracebacks From Exception Occurrences (non-symbolic)
22310 @subsubsection Tracebacks From Exception Occurrences
22313 Non-symbolic tracebacks are obtained by using the @option{-E} binder argument.
22314 The stack traceback is attached to the exception information string, and can
22315 be retrieved in an exception handler within the Ada program, by means of the
22316 Ada facilities defined in @code{Ada.Exceptions}. Here is a simple example:
22318 @smallexample @c ada
22320 with Ada.Exceptions;
22325 use Ada.Exceptions;
22333 Text_IO.Put_Line (Exception_Information (E));
22347 This program will output:
22352 Exception name: CONSTRAINT_ERROR
22353 Message: stb.adb:12
22354 Call stack traceback locations:
22355 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
22358 @node Tracebacks From Anywhere in a Program (non-symbolic)
22359 @subsubsection Tracebacks From Anywhere in a Program
22362 It is also possible to retrieve a stack traceback from anywhere in a
22363 program. For this you need to
22364 use the @code{GNAT.Traceback} API. This package includes a procedure called
22365 @code{Call_Chain} that computes a complete stack traceback, as well as useful
22366 display procedures described below. It is not necessary to use the
22367 @option{-E gnatbind} option in this case, because the stack traceback mechanism
22368 is invoked explicitly.
22371 In the following example we compute a traceback at a specific location in
22372 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
22373 convert addresses to strings:
22375 @smallexample @c ada
22377 with GNAT.Traceback;
22378 with GNAT.Debug_Utilities;
22384 use GNAT.Traceback;
22387 TB : Tracebacks_Array (1 .. 10);
22388 -- We are asking for a maximum of 10 stack frames.
22390 -- Len will receive the actual number of stack frames returned.
22392 Call_Chain (TB, Len);
22394 Text_IO.Put ("In STB.P1 : ");
22396 for K in 1 .. Len loop
22397 Text_IO.Put (Debug_Utilities.Image (TB (K)));
22418 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
22419 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
22423 You can then get further information by invoking the @code{addr2line}
22424 tool as described earlier (note that the hexadecimal addresses
22425 need to be specified in C format, with a leading ``0x'').
22427 @node Symbolic Traceback
22428 @subsection Symbolic Traceback
22429 @cindex traceback, symbolic
22432 A symbolic traceback is a stack traceback in which procedure names are
22433 associated with each code location.
22436 Note that this feature is not supported on all platforms. See
22437 @file{GNAT.Traceback.Symbolic spec in g-trasym.ads} for a complete
22438 list of currently supported platforms.
22441 Note that the symbolic traceback requires that the program be compiled
22442 with debug information. If it is not compiled with debug information
22443 only the non-symbolic information will be valid.
22446 * Tracebacks From Exception Occurrences (symbolic)::
22447 * Tracebacks From Anywhere in a Program (symbolic)::
22450 @node Tracebacks From Exception Occurrences (symbolic)
22451 @subsubsection Tracebacks From Exception Occurrences
22453 @smallexample @c ada
22455 with GNAT.Traceback.Symbolic;
22461 raise Constraint_Error;
22478 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
22483 $ gnatmake -g .\stb -bargs -E -largs -lgnat -laddr2line -lintl
22486 0040149F in stb.p1 at stb.adb:8
22487 004014B7 in stb.p2 at stb.adb:13
22488 004014CF in stb.p3 at stb.adb:18
22489 004015DD in ada.stb at stb.adb:22
22490 00401461 in main at b~stb.adb:168
22491 004011C4 in __mingw_CRTStartup at crt1.c:200
22492 004011F1 in mainCRTStartup at crt1.c:222
22493 77E892A4 in ?? at ??:0
22497 In the above example the ``.\'' syntax in the @command{gnatmake} command
22498 is currently required by @command{addr2line} for files that are in
22499 the current working directory.
22500 Moreover, the exact sequence of linker options may vary from platform
22502 The above @option{-largs} section is for Windows platforms. By contrast,
22503 under Unix there is no need for the @option{-largs} section.
22504 Differences across platforms are due to details of linker implementation.
22506 @node Tracebacks From Anywhere in a Program (symbolic)
22507 @subsubsection Tracebacks From Anywhere in a Program
22510 It is possible to get a symbolic stack traceback
22511 from anywhere in a program, just as for non-symbolic tracebacks.
22512 The first step is to obtain a non-symbolic
22513 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
22514 information. Here is an example:
22516 @smallexample @c ada
22518 with GNAT.Traceback;
22519 with GNAT.Traceback.Symbolic;
22524 use GNAT.Traceback;
22525 use GNAT.Traceback.Symbolic;
22528 TB : Tracebacks_Array (1 .. 10);
22529 -- We are asking for a maximum of 10 stack frames.
22531 -- Len will receive the actual number of stack frames returned.
22533 Call_Chain (TB, Len);
22534 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
22547 @c ******************************
22549 @node Compatibility with HP Ada
22550 @chapter Compatibility with HP Ada
22551 @cindex Compatibility
22556 @cindex Compatibility between GNAT and HP Ada
22557 This chapter compares HP Ada (formerly known as ``DEC Ada'')
22558 for OpenVMS Alpha and GNAT for OpenVMS for Alpha and for I64.
22559 GNAT is highly compatible
22560 with HP Ada, and it should generally be straightforward to port code
22561 from the HP Ada environment to GNAT. However, there are a few language
22562 and implementation differences of which the user must be aware. These
22563 differences are discussed in this chapter. In
22564 addition, the operating environment and command structure for the
22565 compiler are different, and these differences are also discussed.
22567 For further details on these and other compatibility issues,
22568 see Appendix E of the HP publication
22569 @cite{HP Ada, Technical Overview and Comparison on HP Platforms}.
22571 Except where otherwise indicated, the description of GNAT for OpenVMS
22572 applies to both the Alpha and I64 platforms.
22574 For information on porting Ada code from GNAT on Alpha OpenVMS to GNAT on
22575 I64 OpenVMS, see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
22577 The discussion in this chapter addresses specifically the implementation
22578 of Ada 83 for HP OpenVMS Alpha Systems. In cases where the implementation
22579 of HP Ada differs between OpenVMS Alpha Systems and OpenVMS VAX Systems,
22580 GNAT always follows the Alpha implementation.
22582 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
22583 attributes are recognized, although only a subset of them can sensibly
22584 be implemented. The description of pragmas in the
22585 @cite{GNAT Reference Manual} indicates whether or not they are applicable
22586 to non-VMS systems.
22589 * Ada Language Compatibility::
22590 * Differences in the Definition of Package System::
22591 * Language-Related Features::
22592 * The Package STANDARD::
22593 * The Package SYSTEM::
22594 * Tasking and Task-Related Features::
22595 * Pragmas and Pragma-Related Features::
22596 * Library of Predefined Units::
22598 * Main Program Definition::
22599 * Implementation-Defined Attributes::
22600 * Compiler and Run-Time Interfacing::
22601 * Program Compilation and Library Management::
22603 * Implementation Limits::
22604 * Tools and Utilities::
22607 @node Ada Language Compatibility
22608 @section Ada Language Compatibility
22611 GNAT handles Ada 95 and Ada 2005 as well as Ada 83, whereas HP Ada is only
22612 for Ada 83. Ada 95 and Ada 2005 are almost completely upwards compatible
22613 with Ada 83, and therefore Ada 83 programs will compile
22614 and run under GNAT with
22615 no changes or only minor changes. The @cite{Annotated Ada Reference Manual}
22616 provides details on specific incompatibilities.
22618 GNAT provides the switch @option{/83} on the @command{GNAT COMPILE} command,
22619 as well as the pragma @code{ADA_83}, to force the compiler to
22620 operate in Ada 83 mode. This mode does not guarantee complete
22621 conformance to Ada 83, but in practice is sufficient to
22622 eliminate most sources of incompatibilities.
22623 In particular, it eliminates the recognition of the
22624 additional Ada 95 and Ada 2005 keywords, so that their use as identifiers
22625 in Ada 83 programs is legal, and handles the cases of packages
22626 with optional bodies, and generics that instantiate unconstrained
22627 types without the use of @code{(<>)}.
22629 @node Differences in the Definition of Package System
22630 @section Differences in the Definition of Package @code{System}
22633 An Ada compiler is allowed to add
22634 implementation-dependent declarations to package @code{System}.
22636 GNAT does not take advantage of this permission, and the version of
22637 @code{System} provided by GNAT exactly matches that defined in the Ada
22640 However, HP Ada adds an extensive set of declarations to package
22642 as fully documented in the HP Ada manuals. To minimize changes required
22643 for programs that make use of these extensions, GNAT provides the pragma
22644 @code{Extend_System} for extending the definition of package System. By using:
22645 @cindex pragma @code{Extend_System}
22646 @cindex @code{Extend_System} pragma
22648 @smallexample @c ada
22651 pragma Extend_System (Aux_DEC);
22657 the set of definitions in @code{System} is extended to include those in
22658 package @code{System.Aux_DEC}.
22659 @cindex @code{System.Aux_DEC} package
22660 @cindex @code{Aux_DEC} package (child of @code{System})
22661 These definitions are incorporated directly into package @code{System},
22662 as though they had been declared there. For a
22663 list of the declarations added, see the specification of this package,
22664 which can be found in the file @file{s-auxdec.ads} in the GNAT library.
22665 @cindex @file{s-auxdec.ads} file
22666 The pragma @code{Extend_System} is a configuration pragma, which means that
22667 it can be placed in the file @file{gnat.adc}, so that it will automatically
22668 apply to all subsequent compilations. See @ref{Configuration Pragmas},
22669 for further details.
22671 An alternative approach that avoids the use of the non-standard
22672 @code{Extend_System} pragma is to add a context clause to the unit that
22673 references these facilities:
22675 @smallexample @c ada
22677 with System.Aux_DEC;
22678 use System.Aux_DEC;
22683 The effect is not quite semantically identical to incorporating
22684 the declarations directly into package @code{System},
22685 but most programs will not notice a difference
22686 unless they use prefix notation (e.g. @code{System.Integer_8})
22687 to reference the entities directly in package @code{System}.
22688 For units containing such references,
22689 the prefixes must either be removed, or the pragma @code{Extend_System}
22692 @node Language-Related Features
22693 @section Language-Related Features
22696 The following sections highlight differences in types,
22697 representations of types, operations, alignment, and
22701 * Integer Types and Representations::
22702 * Floating-Point Types and Representations::
22703 * Pragmas Float_Representation and Long_Float::
22704 * Fixed-Point Types and Representations::
22705 * Record and Array Component Alignment::
22706 * Address Clauses::
22707 * Other Representation Clauses::
22710 @node Integer Types and Representations
22711 @subsection Integer Types and Representations
22714 The set of predefined integer types is identical in HP Ada and GNAT.
22715 Furthermore the representation of these integer types is also identical,
22716 including the capability of size clauses forcing biased representation.
22719 HP Ada for OpenVMS Alpha systems has defined the
22720 following additional integer types in package @code{System}:
22737 @code{LARGEST_INTEGER}
22741 In GNAT, the first four of these types may be obtained from the
22742 standard Ada package @code{Interfaces}.
22743 Alternatively, by use of the pragma @code{Extend_System}, identical
22744 declarations can be referenced directly in package @code{System}.
22745 On both GNAT and HP Ada, the maximum integer size is 64 bits.
22747 @node Floating-Point Types and Representations
22748 @subsection Floating-Point Types and Representations
22749 @cindex Floating-Point types
22752 The set of predefined floating-point types is identical in HP Ada and GNAT.
22753 Furthermore the representation of these floating-point
22754 types is also identical. One important difference is that the default
22755 representation for HP Ada is @code{VAX_Float}, but the default representation
22758 Specific types may be declared to be @code{VAX_Float} or IEEE, using the
22759 pragma @code{Float_Representation} as described in the HP Ada
22761 For example, the declarations:
22763 @smallexample @c ada
22765 type F_Float is digits 6;
22766 pragma Float_Representation (VAX_Float, F_Float);
22771 declares a type @code{F_Float} that will be represented in @code{VAX_Float}
22773 This set of declarations actually appears in @code{System.Aux_DEC},
22775 the full set of additional floating-point declarations provided in
22776 the HP Ada version of package @code{System}.
22777 This and similar declarations may be accessed in a user program
22778 by using pragma @code{Extend_System}. The use of this
22779 pragma, and the related pragma @code{Long_Float} is described in further
22780 detail in the following section.
22782 @node Pragmas Float_Representation and Long_Float
22783 @subsection Pragmas @code{Float_Representation} and @code{Long_Float}
22786 HP Ada provides the pragma @code{Float_Representation}, which
22787 acts as a program library switch to allow control over
22788 the internal representation chosen for the predefined
22789 floating-point types declared in the package @code{Standard}.
22790 The format of this pragma is as follows:
22792 @smallexample @c ada
22794 pragma Float_Representation(VAX_Float | IEEE_Float);
22799 This pragma controls the representation of floating-point
22804 @code{VAX_Float} specifies that floating-point
22805 types are represented by default with the VAX system hardware types
22806 @code{F-floating}, @code{D-floating}, @code{G-floating}.
22807 Note that the @code{H-floating}
22808 type was available only on VAX systems, and is not available
22809 in either HP Ada or GNAT.
22812 @code{IEEE_Float} specifies that floating-point
22813 types are represented by default with the IEEE single and
22814 double floating-point types.
22818 GNAT provides an identical implementation of the pragma
22819 @code{Float_Representation}, except that it functions as a
22820 configuration pragma. Note that the
22821 notion of configuration pragma corresponds closely to the
22822 HP Ada notion of a program library switch.
22824 When no pragma is used in GNAT, the default is @code{IEEE_Float},
22826 from HP Ada 83, where the default is @code{VAX_Float}. In addition, the
22827 predefined libraries in GNAT are built using @code{IEEE_Float}, so it is not
22828 advisable to change the format of numbers passed to standard library
22829 routines, and if necessary explicit type conversions may be needed.
22831 The use of @code{IEEE_Float} is recommended in GNAT since it is more
22832 efficient, and (given that it conforms to an international standard)
22833 potentially more portable.
22834 The situation in which @code{VAX_Float} may be useful is in interfacing
22835 to existing code and data that expect the use of @code{VAX_Float}.
22836 In such a situation use the predefined @code{VAX_Float}
22837 types in package @code{System}, as extended by
22838 @code{Extend_System}. For example, use @code{System.F_Float}
22839 to specify the 32-bit @code{F-Float} format.
22842 On OpenVMS systems, HP Ada provides the pragma @code{Long_Float}
22843 to allow control over the internal representation chosen
22844 for the predefined type @code{Long_Float} and for floating-point
22845 type declarations with digits specified in the range 7 .. 15.
22846 The format of this pragma is as follows:
22848 @smallexample @c ada
22850 pragma Long_Float (D_FLOAT | G_FLOAT);
22854 @node Fixed-Point Types and Representations
22855 @subsection Fixed-Point Types and Representations
22858 On HP Ada for OpenVMS Alpha systems, rounding is
22859 away from zero for both positive and negative numbers.
22860 Therefore, @code{+0.5} rounds to @code{1},
22861 and @code{-0.5} rounds to @code{-1}.
22863 On GNAT the results of operations
22864 on fixed-point types are in accordance with the Ada
22865 rules. In particular, results of operations on decimal
22866 fixed-point types are truncated.
22868 @node Record and Array Component Alignment
22869 @subsection Record and Array Component Alignment
22872 On HP Ada for OpenVMS Alpha, all non composite components
22873 are aligned on natural boundaries. For example, 1-byte
22874 components are aligned on byte boundaries, 2-byte
22875 components on 2-byte boundaries, 4-byte components on 4-byte
22876 byte boundaries, and so on. The OpenVMS Alpha hardware
22877 runs more efficiently with naturally aligned data.
22879 On GNAT, alignment rules are compatible
22880 with HP Ada for OpenVMS Alpha.
22882 @node Address Clauses
22883 @subsection Address Clauses
22886 In HP Ada and GNAT, address clauses are supported for
22887 objects and imported subprograms.
22888 The predefined type @code{System.Address} is a private type
22889 in both compilers on Alpha OpenVMS, with the same representation
22890 (it is simply a machine pointer). Addition, subtraction, and comparison
22891 operations are available in the standard Ada package
22892 @code{System.Storage_Elements}, or in package @code{System}
22893 if it is extended to include @code{System.Aux_DEC} using a
22894 pragma @code{Extend_System} as previously described.
22896 Note that code that @code{with}'s both this extended package @code{System}
22897 and the package @code{System.Storage_Elements} should not @code{use}
22898 both packages, or ambiguities will result. In general it is better
22899 not to mix these two sets of facilities. The Ada package was
22900 designed specifically to provide the kind of features that HP Ada
22901 adds directly to package @code{System}.
22903 The type @code{System.Address} is a 64-bit integer type in GNAT for
22904 I64 OpenVMS. For more information,
22905 see @ref{Transitioning to 64-Bit GNAT for OpenVMS}.
22907 GNAT is compatible with HP Ada in its handling of address
22908 clauses, except for some limitations in
22909 the form of address clauses for composite objects with
22910 initialization. Such address clauses are easily replaced
22911 by the use of an explicitly-defined constant as described
22912 in the Ada Reference Manual (13.1(22)). For example, the sequence
22915 @smallexample @c ada
22917 X, Y : Integer := Init_Func;
22918 Q : String (X .. Y) := "abc";
22920 for Q'Address use Compute_Address;
22925 will be rejected by GNAT, since the address cannot be computed at the time
22926 that @code{Q} is declared. To achieve the intended effect, write instead:
22928 @smallexample @c ada
22931 X, Y : Integer := Init_Func;
22932 Q_Address : constant Address := Compute_Address;
22933 Q : String (X .. Y) := "abc";
22935 for Q'Address use Q_Address;
22941 which will be accepted by GNAT (and other Ada compilers), and is also
22942 compatible with Ada 83. A fuller description of the restrictions
22943 on address specifications is found in the @cite{GNAT Reference Manual}.
22945 @node Other Representation Clauses
22946 @subsection Other Representation Clauses
22949 GNAT implements in a compatible manner all the representation
22950 clauses supported by HP Ada. In addition, GNAT
22951 implements the representation clause forms that were introduced in Ada 95,
22952 including @code{COMPONENT_SIZE} and @code{SIZE} clauses for objects.
22954 @node The Package STANDARD
22955 @section The Package @code{STANDARD}
22958 The package @code{STANDARD}, as implemented by HP Ada, is fully
22959 described in the @cite{Ada Reference Manual} and in the
22960 @cite{HP Ada Language Reference Manual}. As implemented by GNAT, the
22961 package @code{STANDARD} is described in the @cite{Ada Reference Manual}.
22963 In addition, HP Ada supports the Latin-1 character set in
22964 the type @code{CHARACTER}. GNAT supports the Latin-1 character set
22965 in the type @code{CHARACTER} and also Unicode (ISO 10646 BMP) in
22966 the type @code{WIDE_CHARACTER}.
22968 The floating-point types supported by GNAT are those
22969 supported by HP Ada, but the defaults are different, and are controlled by
22970 pragmas. See @ref{Floating-Point Types and Representations}, for details.
22972 @node The Package SYSTEM
22973 @section The Package @code{SYSTEM}
22976 HP Ada provides a specific version of the package
22977 @code{SYSTEM} for each platform on which the language is implemented.
22978 For the complete specification of the package @code{SYSTEM}, see
22979 Appendix F of the @cite{HP Ada Language Reference Manual}.
22981 On HP Ada, the package @code{SYSTEM} includes the following conversion
22984 @item @code{TO_ADDRESS(INTEGER)}
22986 @item @code{TO_ADDRESS(UNSIGNED_LONGWORD)}
22988 @item @code{TO_ADDRESS(}@i{universal_integer}@code{)}
22990 @item @code{TO_INTEGER(ADDRESS)}
22992 @item @code{TO_UNSIGNED_LONGWORD(ADDRESS)}
22994 @item Function @code{IMPORT_VALUE return UNSIGNED_LONGWORD} and the
22995 functions @code{IMPORT_ADDRESS} and @code{IMPORT_LARGEST_VALUE}
22999 By default, GNAT supplies a version of @code{SYSTEM} that matches
23000 the definition given in the @cite{Ada Reference Manual}.
23002 is a subset of the HP system definitions, which is as
23003 close as possible to the original definitions. The only difference
23004 is that the definition of @code{SYSTEM_NAME} is different:
23006 @smallexample @c ada
23008 type Name is (SYSTEM_NAME_GNAT);
23009 System_Name : constant Name := SYSTEM_NAME_GNAT;
23014 Also, GNAT adds the Ada declarations for
23015 @code{BIT_ORDER} and @code{DEFAULT_BIT_ORDER}.
23017 However, the use of the following pragma causes GNAT
23018 to extend the definition of package @code{SYSTEM} so that it
23019 encompasses the full set of HP-specific extensions,
23020 including the functions listed above:
23022 @smallexample @c ada
23024 pragma Extend_System (Aux_DEC);
23029 The pragma @code{Extend_System} is a configuration pragma that
23030 is most conveniently placed in the @file{gnat.adc} file. See the
23031 @cite{GNAT Reference Manual} for further details.
23033 HP Ada does not allow the recompilation of the package
23034 @code{SYSTEM}. Instead HP Ada provides several pragmas
23035 (@code{SYSTEM_NAME}, @code{STORAGE_UNIT}, and @code{MEMORY_SIZE})
23036 to modify values in the package @code{SYSTEM}.
23037 On OpenVMS Alpha systems, the pragma
23038 @code{SYSTEM_NAME} takes the enumeration literal @code{OPENVMS_AXP} as
23039 its single argument.
23041 GNAT does permit the recompilation of package @code{SYSTEM} using
23042 the special switch @option{-gnatg}, and this switch can be used if
23043 it is necessary to modify the definitions in @code{SYSTEM}. GNAT does
23044 not permit the specification of @code{SYSTEM_NAME}, @code{STORAGE_UNIT}
23045 or @code{MEMORY_SIZE} by any other means.
23047 On GNAT systems, the pragma @code{SYSTEM_NAME} takes the
23048 enumeration literal @code{SYSTEM_NAME_GNAT}.
23050 The definitions provided by the use of
23052 @smallexample @c ada
23053 pragma Extend_System (AUX_Dec);
23057 are virtually identical to those provided by the HP Ada 83 package
23058 @code{SYSTEM}. One important difference is that the name of the
23060 function for type @code{UNSIGNED_LONGWORD} is changed to
23061 @code{TO_ADDRESS_LONG}.
23062 See the @cite{GNAT Reference Manual} for a discussion of why this change was
23066 The version of @code{TO_ADDRESS} taking a @i{universal_integer} argument
23068 an extension to Ada 83 not strictly compatible with the reference manual.
23069 GNAT, in order to be exactly compatible with the standard,
23070 does not provide this capability. In HP Ada 83, the
23071 point of this definition is to deal with a call like:
23073 @smallexample @c ada
23074 TO_ADDRESS (16#12777#);
23078 Normally, according to Ada 83 semantics, one would expect this to be
23079 ambiguous, since it matches both the @code{INTEGER} and
23080 @code{UNSIGNED_LONGWORD} forms of @code{TO_ADDRESS}.
23081 However, in HP Ada 83, there is no ambiguity, since the
23082 definition using @i{universal_integer} takes precedence.
23084 In GNAT, since the version with @i{universal_integer} cannot be supplied,
23086 not possible to be 100% compatible. Since there are many programs using
23087 numeric constants for the argument to @code{TO_ADDRESS}, the decision in
23089 to change the name of the function in the @code{UNSIGNED_LONGWORD} case,
23090 so the declarations provided in the GNAT version of @code{AUX_Dec} are:
23092 @smallexample @c ada
23093 function To_Address (X : Integer) return Address;
23094 pragma Pure_Function (To_Address);
23096 function To_Address_Long (X : Unsigned_Longword) return Address;
23097 pragma Pure_Function (To_Address_Long);
23101 This means that programs using @code{TO_ADDRESS} for
23102 @code{UNSIGNED_LONGWORD} must change the name to @code{TO_ADDRESS_LONG}.
23104 @node Tasking and Task-Related Features
23105 @section Tasking and Task-Related Features
23108 This section compares the treatment of tasking in GNAT
23109 and in HP Ada for OpenVMS Alpha.
23110 The GNAT description applies to both Alpha and I64 OpenVMS.
23111 For detailed information on tasking in
23112 HP Ada, see the @cite{HP Ada Language Reference Manual} and the
23113 relevant run-time reference manual.
23116 * Implementation of Tasks in HP Ada for OpenVMS Alpha Systems::
23117 * Assigning Task IDs::
23118 * Task IDs and Delays::
23119 * Task-Related Pragmas::
23120 * Scheduling and Task Priority::
23122 * External Interrupts::
23125 @node Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
23126 @subsection Implementation of Tasks in HP Ada for OpenVMS Alpha Systems
23129 On OpenVMS Alpha systems, each Ada task (except a passive
23130 task) is implemented as a single stream of execution
23131 that is created and managed by the kernel. On these
23132 systems, HP Ada tasking support is based on DECthreads,
23133 an implementation of the POSIX standard for threads.
23135 Also, on OpenVMS Alpha systems, HP Ada tasks and foreign
23136 code that calls DECthreads routines can be used together.
23137 The interaction between Ada tasks and DECthreads routines
23138 can have some benefits. For example when on OpenVMS Alpha,
23139 HP Ada can call C code that is already threaded.
23141 GNAT uses the facilities of DECthreads,
23142 and Ada tasks are mapped to threads.
23144 @node Assigning Task IDs
23145 @subsection Assigning Task IDs
23148 The HP Ada Run-Time Library always assigns @code{%TASK 1} to
23149 the environment task that executes the main program. On
23150 OpenVMS Alpha systems, @code{%TASK 0} is often used for tasks
23151 that have been created but are not yet activated.
23153 On OpenVMS Alpha systems, task IDs are assigned at
23154 activation. On GNAT systems, task IDs are also assigned at
23155 task creation but do not have the same form or values as
23156 task ID values in HP Ada. There is no null task, and the
23157 environment task does not have a specific task ID value.
23159 @node Task IDs and Delays
23160 @subsection Task IDs and Delays
23163 On OpenVMS Alpha systems, tasking delays are implemented
23164 using Timer System Services. The Task ID is used for the
23165 identification of the timer request (the @code{REQIDT} parameter).
23166 If Timers are used in the application take care not to use
23167 @code{0} for the identification, because cancelling such a timer
23168 will cancel all timers and may lead to unpredictable results.
23170 @node Task-Related Pragmas
23171 @subsection Task-Related Pragmas
23174 Ada supplies the pragma @code{TASK_STORAGE}, which allows
23175 specification of the size of the guard area for a task
23176 stack. (The guard area forms an area of memory that has no
23177 read or write access and thus helps in the detection of
23178 stack overflow.) On OpenVMS Alpha systems, if the pragma
23179 @code{TASK_STORAGE} specifies a value of zero, a minimal guard
23180 area is created. In the absence of a pragma @code{TASK_STORAGE},
23181 a default guard area is created.
23183 GNAT supplies the following task-related pragmas:
23186 @item @code{TASK_INFO}
23188 This pragma appears within a task definition and
23189 applies to the task in which it appears. The argument
23190 must be of type @code{SYSTEM.TASK_INFO.TASK_INFO_TYPE}.
23192 @item @code{TASK_STORAGE}
23194 GNAT implements pragma @code{TASK_STORAGE} in the same way as
23196 Both HP Ada and GNAT supply the pragmas @code{PASSIVE},
23197 @code{SUPPRESS}, and @code{VOLATILE}.
23199 @node Scheduling and Task Priority
23200 @subsection Scheduling and Task Priority
23203 HP Ada implements the Ada language requirement that
23204 when two tasks are eligible for execution and they have
23205 different priorities, the lower priority task does not
23206 execute while the higher priority task is waiting. The HP
23207 Ada Run-Time Library keeps a task running until either the
23208 task is suspended or a higher priority task becomes ready.
23210 On OpenVMS Alpha systems, the default strategy is round-
23211 robin with preemption. Tasks of equal priority take turns
23212 at the processor. A task is run for a certain period of
23213 time and then placed at the tail of the ready queue for
23214 its priority level.
23216 HP Ada provides the implementation-defined pragma @code{TIME_SLICE},
23217 which can be used to enable or disable round-robin
23218 scheduling of tasks with the same priority.
23219 See the relevant HP Ada run-time reference manual for
23220 information on using the pragmas to control HP Ada task
23223 GNAT follows the scheduling rules of Annex D (Real-Time
23224 Annex) of the @cite{Ada Reference Manual}. In general, this
23225 scheduling strategy is fully compatible with HP Ada
23226 although it provides some additional constraints (as
23227 fully documented in Annex D).
23228 GNAT implements time slicing control in a manner compatible with
23229 HP Ada 83, by means of the pragma @code{Time_Slice}, whose semantics
23230 are identical to the HP Ada 83 pragma of the same name.
23231 Note that it is not possible to mix GNAT tasking and
23232 HP Ada 83 tasking in the same program, since the two run-time
23233 libraries are not compatible.
23235 @node The Task Stack
23236 @subsection The Task Stack
23239 In HP Ada, a task stack is allocated each time a
23240 non-passive task is activated. As soon as the task is
23241 terminated, the storage for the task stack is deallocated.
23242 If you specify a size of zero (bytes) with @code{T'STORAGE_SIZE},
23243 a default stack size is used. Also, regardless of the size
23244 specified, some additional space is allocated for task
23245 management purposes. On OpenVMS Alpha systems, at least
23246 one page is allocated.
23248 GNAT handles task stacks in a similar manner. In accordance with
23249 the Ada rules, it provides the pragma @code{STORAGE_SIZE} as
23250 an alternative method for controlling the task stack size.
23251 The specification of the attribute @code{T'STORAGE_SIZE} is also
23252 supported in a manner compatible with HP Ada.
23254 @node External Interrupts
23255 @subsection External Interrupts
23258 On HP Ada, external interrupts can be associated with task entries.
23259 GNAT is compatible with HP Ada in its handling of external interrupts.
23261 @node Pragmas and Pragma-Related Features
23262 @section Pragmas and Pragma-Related Features
23265 Both HP Ada and GNAT supply all language-defined pragmas
23266 as specified by the Ada 83 standard. GNAT also supplies all
23267 language-defined pragmas introduced by Ada 95 and Ada 2005.
23268 In addition, GNAT implements the implementation-defined pragmas
23272 @item @code{AST_ENTRY}
23274 @item @code{COMMON_OBJECT}
23276 @item @code{COMPONENT_ALIGNMENT}
23278 @item @code{EXPORT_EXCEPTION}
23280 @item @code{EXPORT_FUNCTION}
23282 @item @code{EXPORT_OBJECT}
23284 @item @code{EXPORT_PROCEDURE}
23286 @item @code{EXPORT_VALUED_PROCEDURE}
23288 @item @code{FLOAT_REPRESENTATION}
23292 @item @code{IMPORT_EXCEPTION}
23294 @item @code{IMPORT_FUNCTION}
23296 @item @code{IMPORT_OBJECT}
23298 @item @code{IMPORT_PROCEDURE}
23300 @item @code{IMPORT_VALUED_PROCEDURE}
23302 @item @code{INLINE_GENERIC}
23304 @item @code{INTERFACE_NAME}
23306 @item @code{LONG_FLOAT}
23308 @item @code{MAIN_STORAGE}
23310 @item @code{PASSIVE}
23312 @item @code{PSET_OBJECT}
23314 @item @code{SHARE_GENERIC}
23316 @item @code{SUPPRESS_ALL}
23318 @item @code{TASK_STORAGE}
23320 @item @code{TIME_SLICE}
23326 These pragmas are all fully implemented, with the exception of @code{TITLE},
23327 @code{PASSIVE}, and @code{SHARE_GENERIC}, which are
23328 recognized, but which have no
23329 effect in GNAT. The effect of @code{PASSIVE} may be obtained by the
23330 use of Ada protected objects. In GNAT, all generics are inlined.
23332 Unlike HP Ada, the GNAT ``@code{EXPORT_}@i{subprogram}'' pragmas require
23333 a separate subprogram specification which must appear before the
23336 GNAT also supplies a number of implementation-defined pragmas as follows:
23338 @item @code{ABORT_DEFER}
23340 @item @code{ADA_83}
23342 @item @code{ADA_95}
23344 @item @code{ADA_05}
23346 @item @code{ANNOTATE}
23348 @item @code{ASSERT}
23350 @item @code{C_PASS_BY_COPY}
23352 @item @code{CPP_CLASS}
23354 @item @code{CPP_CONSTRUCTOR}
23356 @item @code{CPP_DESTRUCTOR}
23360 @item @code{EXTEND_SYSTEM}
23362 @item @code{LINKER_ALIAS}
23364 @item @code{LINKER_SECTION}
23366 @item @code{MACHINE_ATTRIBUTE}
23368 @item @code{NO_RETURN}
23370 @item @code{PURE_FUNCTION}
23372 @item @code{SOURCE_FILE_NAME}
23374 @item @code{SOURCE_REFERENCE}
23376 @item @code{TASK_INFO}
23378 @item @code{UNCHECKED_UNION}
23380 @item @code{UNIMPLEMENTED_UNIT}
23382 @item @code{UNIVERSAL_DATA}
23384 @item @code{UNSUPPRESS}
23386 @item @code{WARNINGS}
23388 @item @code{WEAK_EXTERNAL}
23392 For full details on these GNAT implementation-defined pragmas, see
23393 the GNAT Reference Manual.
23396 * Restrictions on the Pragma INLINE::
23397 * Restrictions on the Pragma INTERFACE::
23398 * Restrictions on the Pragma SYSTEM_NAME::
23401 @node Restrictions on the Pragma INLINE
23402 @subsection Restrictions on Pragma @code{INLINE}
23405 HP Ada enforces the following restrictions on the pragma @code{INLINE}:
23407 @item Parameters cannot have a task type.
23409 @item Function results cannot be task types, unconstrained
23410 array types, or unconstrained types with discriminants.
23412 @item Bodies cannot declare the following:
23414 @item Subprogram body or stub (imported subprogram is allowed)
23418 @item Generic declarations
23420 @item Instantiations
23424 @item Access types (types derived from access types allowed)
23426 @item Array or record types
23428 @item Dependent tasks
23430 @item Direct recursive calls of subprogram or containing
23431 subprogram, directly or via a renaming
23437 In GNAT, the only restriction on pragma @code{INLINE} is that the
23438 body must occur before the call if both are in the same
23439 unit, and the size must be appropriately small. There are
23440 no other specific restrictions which cause subprograms to
23441 be incapable of being inlined.
23443 @node Restrictions on the Pragma INTERFACE
23444 @subsection Restrictions on Pragma @code{INTERFACE}
23447 The following restrictions on pragma @code{INTERFACE}
23448 are enforced by both HP Ada and GNAT:
23450 @item Languages accepted: Ada, Bliss, C, Fortran, Default.
23451 Default is the default on OpenVMS Alpha systems.
23453 @item Parameter passing: Language specifies default
23454 mechanisms but can be overridden with an @code{EXPORT} pragma.
23457 @item Ada: Use internal Ada rules.
23459 @item Bliss, C: Parameters must be mode @code{in}; cannot be
23460 record or task type. Result cannot be a string, an
23461 array, or a record.
23463 @item Fortran: Parameters cannot have a task type. Result cannot
23464 be a string, an array, or a record.
23469 GNAT is entirely upwards compatible with HP Ada, and in addition allows
23470 record parameters for all languages.
23472 @node Restrictions on the Pragma SYSTEM_NAME
23473 @subsection Restrictions on Pragma @code{SYSTEM_NAME}
23476 For HP Ada for OpenVMS Alpha, the enumeration literal
23477 for the type @code{NAME} is @code{OPENVMS_AXP}.
23478 In GNAT, the enumeration
23479 literal for the type @code{NAME} is @code{SYSTEM_NAME_GNAT}.
23481 @node Library of Predefined Units
23482 @section Library of Predefined Units
23485 A library of predefined units is provided as part of the
23486 HP Ada and GNAT implementations. HP Ada does not provide
23487 the package @code{MACHINE_CODE} but instead recommends importing
23490 The GNAT versions of the HP Ada Run-Time Library (@code{ADA$PREDEFINED:})
23491 units are taken from the OpenVMS Alpha version, not the OpenVMS VAX
23493 The HP Ada Predefined Library units are modified to remove post-Ada 83
23494 incompatibilities and to make them interoperable with GNAT
23495 (@pxref{Changes to DECLIB}, for details).
23496 The units are located in the @file{DECLIB} directory.
23498 The GNAT RTL is contained in
23499 the @file{ADALIB} directory, and
23500 the default search path is set up to find @code{DECLIB} units in preference
23501 to @code{ADALIB} units with the same name (@code{TEXT_IO},
23502 @code{SEQUENTIAL_IO}, and @code{DIRECT_IO}, for example).
23505 * Changes to DECLIB::
23508 @node Changes to DECLIB
23509 @subsection Changes to @code{DECLIB}
23512 The changes made to the HP Ada predefined library for GNAT and post-Ada 83
23513 compatibility are minor and include the following:
23516 @item Adjusting the location of pragmas and record representation
23517 clauses to obey Ada 95 (and thus Ada 2005) rules
23519 @item Adding the proper notation to generic formal parameters
23520 that take unconstrained types in instantiation
23522 @item Adding pragma @code{ELABORATE_BODY} to package specifications
23523 that have package bodies not otherwise allowed
23525 @item Replacing occurrences of the identifier ``@code{PROTECTED}'' by
23526 ``@code{PROTECTD}''.
23527 Currently these are found only in the @code{STARLET} package spec.
23529 @item Changing @code{SYSTEM.ADDRESS} to @code{SYSTEM.SHORT_ADDRESS}
23530 where the address size is constrained to 32 bits.
23534 None of the above changes is visible to users.
23540 On OpenVMS Alpha, HP Ada provides the following strongly-typed bindings:
23543 @item Command Language Interpreter (CLI interface)
23545 @item DECtalk Run-Time Library (DTK interface)
23547 @item Librarian utility routines (LBR interface)
23549 @item General Purpose Run-Time Library (LIB interface)
23551 @item Math Run-Time Library (MTH interface)
23553 @item National Character Set Run-Time Library (NCS interface)
23555 @item Compiled Code Support Run-Time Library (OTS interface)
23557 @item Parallel Processing Run-Time Library (PPL interface)
23559 @item Screen Management Run-Time Library (SMG interface)
23561 @item Sort Run-Time Library (SOR interface)
23563 @item String Run-Time Library (STR interface)
23565 @item STARLET System Library
23568 @item X Window System Version 11R4 and 11R5 (X, XLIB interface)
23570 @item X Windows Toolkit (XT interface)
23572 @item X/Motif Version 1.1.3 and 1.2 (XM interface)
23576 GNAT provides implementations of these HP bindings in the @code{DECLIB}
23577 directory, on both the Alpha and I64 OpenVMS platforms.
23579 The X/Motif bindings used to build @code{DECLIB} are whatever versions are
23581 HP Ada @file{ADA$PREDEFINED} directory with extension @file{.ADC}.
23582 A pragma @code{Linker_Options} has been added to packages @code{Xm},
23583 @code{Xt}, and @code{X_Lib}
23584 causing the default X/Motif sharable image libraries to be linked in. This
23585 is done via options files named @file{xm.opt}, @file{xt.opt}, and
23586 @file{x_lib.opt} (also located in the @file{DECLIB} directory).
23588 It may be necessary to edit these options files to update or correct the
23589 library names if, for example, the newer X/Motif bindings from
23590 @file{ADA$EXAMPLES}
23591 had been (previous to installing GNAT) copied and renamed to supersede the
23592 default @file{ADA$PREDEFINED} versions.
23595 * Shared Libraries and Options Files::
23596 * Interfaces to C::
23599 @node Shared Libraries and Options Files
23600 @subsection Shared Libraries and Options Files
23603 When using the HP Ada
23604 predefined X and Motif bindings, the linking with their sharable images is
23605 done automatically by @command{GNAT LINK}.
23606 When using other X and Motif bindings, you need
23607 to add the corresponding sharable images to the command line for
23608 @code{GNAT LINK}. When linking with shared libraries, or with
23609 @file{.OPT} files, you must
23610 also add them to the command line for @command{GNAT LINK}.
23612 A shared library to be used with GNAT is built in the same way as other
23613 libraries under VMS. The VMS Link command can be used in standard fashion.
23615 @node Interfaces to C
23616 @subsection Interfaces to C
23620 provides the following Ada types and operations:
23623 @item C types package (@code{C_TYPES})
23625 @item C strings (@code{C_TYPES.NULL_TERMINATED})
23627 @item Other_types (@code{SHORT_INT})
23631 Interfacing to C with GNAT, you can use the above approach
23632 described for HP Ada or the facilities of Annex B of
23633 the @cite{Ada Reference Manual} (packages @code{INTERFACES.C},
23634 @code{INTERFACES.C.STRINGS} and @code{INTERFACES.C.POINTERS}). For more
23635 information, see the section ``Interfacing to C'' in the
23636 @cite{GNAT Reference Manual}.
23638 The @option{-gnatF} qualifier forces default and explicit
23639 @code{External_Name} parameters in pragmas @code{Import} and @code{Export}
23640 to be uppercased for compatibility with the default behavior
23641 of HP C. The qualifier has no effect on @code{Link_Name} parameters.
23643 @node Main Program Definition
23644 @section Main Program Definition
23647 The following section discusses differences in the
23648 definition of main programs on HP Ada and GNAT.
23649 On HP Ada, main programs are defined to meet the
23650 following conditions:
23652 @item Procedure with no formal parameters (returns @code{0} upon
23655 @item Procedure with no formal parameters (returns @code{42} when
23656 an unhandled exception is raised)
23658 @item Function with no formal parameters whose returned value
23659 is of a discrete type
23661 @item Procedure with one @code{out} formal of a discrete type for
23662 which a specification of pragma @code{EXPORT_VALUED_PROCEDURE}
23668 When declared with the pragma @code{EXPORT_VALUED_PROCEDURE},
23669 a main function or main procedure returns a discrete
23670 value whose size is less than 64 bits (32 on VAX systems),
23671 the value is zero- or sign-extended as appropriate.
23672 On GNAT, main programs are defined as follows:
23674 @item Must be a non-generic, parameterless subprogram that
23675 is either a procedure or function returning an Ada
23676 @code{STANDARD.INTEGER} (the predefined type)
23678 @item Cannot be a generic subprogram or an instantiation of a
23682 @node Implementation-Defined Attributes
23683 @section Implementation-Defined Attributes
23686 GNAT provides all HP Ada implementation-defined
23689 @node Compiler and Run-Time Interfacing
23690 @section Compiler and Run-Time Interfacing
23693 HP Ada provides the following qualifiers to pass options to the linker
23696 @item @option{/WAIT} and @option{/SUBMIT}
23698 @item @option{/COMMAND}
23700 @item @option{/[NO]MAP}
23702 @item @option{/OUTPUT=@i{file-spec}}
23704 @item @option{/[NO]DEBUG} and @option{/[NO]TRACEBACK}
23708 To pass options to the linker, GNAT provides the following
23712 @item @option{/EXECUTABLE=@i{exec-name}}
23714 @item @option{/VERBOSE}
23716 @item @option{/[NO]DEBUG} and @option{/[NO]TRACEBACK}
23720 For more information on these switches, see
23721 @ref{Switches for gnatlink}.
23722 In HP Ada, the command-line switch @option{/OPTIMIZE} is available
23723 to control optimization. HP Ada also supplies the
23726 @item @code{OPTIMIZE}
23728 @item @code{INLINE}
23730 @item @code{INLINE_GENERIC}
23732 @item @code{SUPPRESS_ALL}
23734 @item @code{PASSIVE}
23738 In GNAT, optimization is controlled strictly by command
23739 line parameters, as described in the corresponding section of this guide.
23740 The HP pragmas for control of optimization are
23741 recognized but ignored.
23743 Note that in GNAT, the default is optimization off, whereas in HP Ada
23744 the default is that optimization is turned on.
23746 @node Program Compilation and Library Management
23747 @section Program Compilation and Library Management
23750 HP Ada and GNAT provide a comparable set of commands to
23751 build programs. HP Ada also provides a program library,
23752 which is a concept that does not exist on GNAT. Instead,
23753 GNAT provides directories of sources that are compiled as
23756 The following table summarizes
23757 the HP Ada commands and provides
23758 equivalent GNAT commands. In this table, some GNAT
23759 equivalents reflect the fact that GNAT does not use the
23760 concept of a program library. Instead, it uses a model
23761 in which collections of source and object files are used
23762 in a manner consistent with other languages like C and
23763 Fortran. Therefore, standard system file commands are used
23764 to manipulate these elements. Those GNAT commands are marked with
23766 Note that, unlike HP Ada, none of the GNAT commands accepts wild cards.
23769 @multitable @columnfractions .35 .65
23771 @item @emph{HP Ada Command}
23772 @tab @emph{GNAT Equivalent / Description}
23774 @item @command{ADA}
23775 @tab @command{GNAT COMPILE}@*
23776 Invokes the compiler to compile one or more Ada source files.
23778 @item @command{ACS ATTACH}@*
23779 @tab [No equivalent]@*
23780 Switches control of terminal from current process running the program
23783 @item @command{ACS CHECK}
23784 @tab @command{GNAT MAKE /DEPENDENCY_LIST}@*
23785 Forms the execution closure of one
23786 or more compiled units and checks completeness and currency.
23788 @item @command{ACS COMPILE}
23789 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
23790 Forms the execution closure of one or
23791 more specified units, checks completeness and currency,
23792 identifies units that have revised source files, compiles same,
23793 and recompiles units that are or will become obsolete.
23794 Also completes incomplete generic instantiations.
23796 @item @command{ACS COPY FOREIGN}
23798 Copies a foreign object file into the program library as a
23801 @item @command{ACS COPY UNIT}
23803 Copies a compiled unit from one program library to another.
23805 @item @command{ACS CREATE LIBRARY}
23806 @tab Create /directory (*)@*
23807 Creates a program library.
23809 @item @command{ACS CREATE SUBLIBRARY}
23810 @tab Create /directory (*)@*
23811 Creates a program sublibrary.
23813 @item @command{ACS DELETE LIBRARY}
23815 Deletes a program library and its contents.
23817 @item @command{ACS DELETE SUBLIBRARY}
23819 Deletes a program sublibrary and its contents.
23821 @item @command{ACS DELETE UNIT}
23822 @tab Delete file (*)@*
23823 On OpenVMS systems, deletes one or more compiled units from
23824 the current program library.
23826 @item @command{ACS DIRECTORY}
23827 @tab Directory (*)@*
23828 On OpenVMS systems, lists units contained in the current
23831 @item @command{ACS ENTER FOREIGN}
23833 Allows the import of a foreign body as an Ada library
23834 specification and enters a reference to a pointer.
23836 @item @command{ACS ENTER UNIT}
23838 Enters a reference (pointer) from the current program library to
23839 a unit compiled into another program library.
23841 @item @command{ACS EXIT}
23842 @tab [No equivalent]@*
23843 Exits from the program library manager.
23845 @item @command{ACS EXPORT}
23847 Creates an object file that contains system-specific object code
23848 for one or more units. With GNAT, object files can simply be copied
23849 into the desired directory.
23851 @item @command{ACS EXTRACT SOURCE}
23853 Allows access to the copied source file for each Ada compilation unit
23855 @item @command{ACS HELP}
23856 @tab @command{HELP GNAT}@*
23857 Provides online help.
23859 @item @command{ACS LINK}
23860 @tab @command{GNAT LINK}@*
23861 Links an object file containing Ada units into an executable file.
23863 @item @command{ACS LOAD}
23865 Loads (partially compiles) Ada units into the program library.
23866 Allows loading a program from a collection of files into a library
23867 without knowing the relationship among units.
23869 @item @command{ACS MERGE}
23871 Merges into the current program library, one or more units from
23872 another library where they were modified.
23874 @item @command{ACS RECOMPILE}
23875 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
23876 Recompiles from external or copied source files any obsolete
23877 unit in the closure. Also, completes any incomplete generic
23880 @item @command{ACS REENTER}
23881 @tab @command{GNAT MAKE}@*
23882 Reenters current references to units compiled after last entered
23883 with the @command{ACS ENTER UNIT} command.
23885 @item @command{ACS SET LIBRARY}
23886 @tab Set default (*)@*
23887 Defines a program library to be the compilation context as well
23888 as the target library for compiler output and commands in general.
23890 @item @command{ACS SET PRAGMA}
23891 @tab Edit @file{gnat.adc} (*)@*
23892 Redefines specified values of the library characteristics
23893 @code{LONG_ FLOAT}, @code{MEMORY_SIZE}, @code{SYSTEM_NAME},
23894 and @code{Float_Representation}.
23896 @item @command{ACS SET SOURCE}
23897 @tab Define @code{ADA_INCLUDE_PATH} path (*)@*
23898 Defines the source file search list for the @command{ACS COMPILE} command.
23900 @item @command{ACS SHOW LIBRARY}
23901 @tab Directory (*)@*
23902 Lists information about one or more program libraries.
23904 @item @command{ACS SHOW PROGRAM}
23905 @tab [No equivalent]@*
23906 Lists information about the execution closure of one or
23907 more units in the program library.
23909 @item @command{ACS SHOW SOURCE}
23910 @tab Show logical @code{ADA_INCLUDE_PATH}@*
23911 Shows the source file search used when compiling units.
23913 @item @command{ACS SHOW VERSION}
23914 @tab Compile with @option{VERBOSE} option
23915 Displays the version number of the compiler and program library
23918 @item @command{ACS SPAWN}
23919 @tab [No equivalent]@*
23920 Creates a subprocess of the current process (same as @command{DCL SPAWN}
23923 @item @command{ACS VERIFY}
23924 @tab [No equivalent]@*
23925 Performs a series of consistency checks on a program library to
23926 determine whether the library structure and library files are in
23933 @section Input-Output
23936 On OpenVMS Alpha systems, HP Ada uses OpenVMS Record
23937 Management Services (RMS) to perform operations on
23941 HP Ada and GNAT predefine an identical set of input-
23942 output packages. To make the use of the
23943 generic @code{TEXT_IO} operations more convenient, HP Ada
23944 provides predefined library packages that instantiate the
23945 integer and floating-point operations for the predefined
23946 integer and floating-point types as shown in the following table.
23948 @multitable @columnfractions .45 .55
23949 @item @emph{Package Name} @tab Instantiation
23951 @item @code{INTEGER_TEXT_IO}
23952 @tab @code{INTEGER_IO(INTEGER)}
23954 @item @code{SHORT_INTEGER_TEXT_IO}
23955 @tab @code{INTEGER_IO(SHORT_INTEGER)}
23957 @item @code{SHORT_SHORT_INTEGER_TEXT_IO}
23958 @tab @code{INTEGER_IO(SHORT_SHORT_INTEGER)}
23960 @item @code{FLOAT_TEXT_IO}
23961 @tab @code{FLOAT_IO(FLOAT)}
23963 @item @code{LONG_FLOAT_TEXT_IO}
23964 @tab @code{FLOAT_IO(LONG_FLOAT)}
23968 The HP Ada predefined packages and their operations
23969 are implemented using OpenVMS Alpha files and input-output
23970 facilities. HP Ada supports asynchronous input-output on OpenVMS Alpha.
23971 Familiarity with the following is recommended:
23973 @item RMS file organizations and access methods
23975 @item OpenVMS file specifications and directories
23977 @item OpenVMS File Definition Language (FDL)
23981 GNAT provides I/O facilities that are completely
23982 compatible with HP Ada. The distribution includes the
23983 standard HP Ada versions of all I/O packages, operating
23984 in a manner compatible with HP Ada. In particular, the
23985 following packages are by default the HP Ada (Ada 83)
23986 versions of these packages rather than the renamings
23987 suggested in Annex J of the Ada Reference Manual:
23989 @item @code{TEXT_IO}
23991 @item @code{SEQUENTIAL_IO}
23993 @item @code{DIRECT_IO}
23997 The use of the standard child package syntax (for
23998 example, @code{ADA.TEXT_IO}) retrieves the post-Ada 83 versions of these
24000 GNAT provides HP-compatible predefined instantiations
24001 of the @code{TEXT_IO} packages, and also
24002 provides the standard predefined instantiations required
24003 by the @cite{Ada Reference Manual}.
24005 For further information on how GNAT interfaces to the file
24006 system or how I/O is implemented in programs written in
24007 mixed languages, see the chapter ``Implementation of the
24008 Standard I/O'' in the @cite{GNAT Reference Manual}.
24009 This chapter covers the following:
24011 @item Standard I/O packages
24013 @item @code{FORM} strings
24015 @item @code{ADA.DIRECT_IO}
24017 @item @code{ADA.SEQUENTIAL_IO}
24019 @item @code{ADA.TEXT_IO}
24021 @item Stream pointer positioning
24023 @item Reading and writing non-regular files
24025 @item @code{GET_IMMEDIATE}
24027 @item Treating @code{TEXT_IO} files as streams
24034 @node Implementation Limits
24035 @section Implementation Limits
24038 The following table lists implementation limits for HP Ada
24040 @multitable @columnfractions .60 .20 .20
24042 @item @emph{Compilation Parameter}
24047 @item In a subprogram or entry declaration, maximum number of
24048 formal parameters that are of an unconstrained record type
24053 @item Maximum identifier length (number of characters)
24058 @item Maximum number of characters in a source line
24063 @item Maximum collection size (number of bytes)
24068 @item Maximum number of discriminants for a record type
24073 @item Maximum number of formal parameters in an entry or
24074 subprogram declaration
24079 @item Maximum number of dimensions in an array type
24084 @item Maximum number of library units and subunits in a compilation.
24089 @item Maximum number of library units and subunits in an execution.
24094 @item Maximum number of objects declared with the pragma @code{COMMON_OBJECT}
24095 or @code{PSECT_OBJECT}
24100 @item Maximum number of enumeration literals in an enumeration type
24106 @item Maximum number of lines in a source file
24111 @item Maximum number of bits in any object
24116 @item Maximum size of the static portion of a stack frame (approximate)
24121 @node Tools and Utilities
24122 @section Tools and Utilities
24125 The following table lists some of the OpenVMS development tools
24126 available for HP Ada, and the corresponding tools for
24127 use with @value{EDITION} on Alpha and I64 platforms.
24128 Aside from the debugger, all the OpenVMS tools identified are part
24129 of the DECset package.
24132 @c Specify table in TeX since Texinfo does a poor job
24136 \settabs\+Language-Sensitive Editor\quad
24137 &Product with HP Ada\quad
24140 &\it Product with HP Ada
24141 & \it Product with GNAT Pro\cr
24143 \+Code Management System
24147 \+Language-Sensitive Editor
24149 & emacs or HP LSE (Alpha)\cr
24159 & OpenVMS Debug (I64)\cr
24161 \+Source Code Analyzer /
24178 \+Coverage Analyzer
24182 \+Module Management
24184 & Not applicable\cr
24194 @c This is the Texinfo version of the table. It renders poorly in pdf, hence
24195 @c the TeX version above for the printed version
24197 @c @multitable @columnfractions .3 .4 .4
24198 @multitable {Source Code Analyzer /}{Tool with HP Ada}{Tool with GNAT Pro}
24200 @tab @i{Tool with HP Ada}
24201 @tab @i{Tool with @value{EDITION}}
24202 @item Code Management@*System
24205 @item Language-Sensitive@*Editor
24207 @tab emacs or HP LSE (Alpha)
24216 @tab OpenVMS Debug (I64)
24217 @item Source Code Analyzer /@*Cross Referencer
24221 @tab HP Digital Test@*Manager (DTM)
24223 @item Performance and@*Coverage Analyzer
24226 @item Module Management@*System
24228 @tab Not applicable
24235 @c **************************************
24236 @node Platform-Specific Information for the Run-Time Libraries
24237 @appendix Platform-Specific Information for the Run-Time Libraries
24238 @cindex Tasking and threads libraries
24239 @cindex Threads libraries and tasking
24240 @cindex Run-time libraries (platform-specific information)
24243 The GNAT run-time implementation may vary with respect to both the
24244 underlying threads library and the exception handling scheme.
24245 For threads support, one or more of the following are supplied:
24247 @item @b{native threads library}, a binding to the thread package from
24248 the underlying operating system
24250 @item @b{pthreads library} (Sparc Solaris only), a binding to the Solaris
24251 POSIX thread package
24255 For exception handling, either or both of two models are supplied:
24257 @item @b{Zero-Cost Exceptions} (``ZCX''),@footnote{
24258 Most programs should experience a substantial speed improvement by
24259 being compiled with a ZCX run-time.
24260 This is especially true for
24261 tasking applications or applications with many exception handlers.}
24262 @cindex Zero-Cost Exceptions
24263 @cindex ZCX (Zero-Cost Exceptions)
24264 which uses binder-generated tables that
24265 are interrogated at run time to locate a handler
24267 @item @b{setjmp / longjmp} (``SJLJ''),
24268 @cindex setjmp/longjmp Exception Model
24269 @cindex SJLJ (setjmp/longjmp Exception Model)
24270 which uses dynamically-set data to establish
24271 the set of handlers
24275 This appendix summarizes which combinations of threads and exception support
24276 are supplied on various GNAT platforms.
24277 It then shows how to select a particular library either
24278 permanently or temporarily,
24279 explains the properties of (and tradeoffs among) the various threads
24280 libraries, and provides some additional
24281 information about several specific platforms.
24284 * Summary of Run-Time Configurations::
24285 * Specifying a Run-Time Library::
24286 * Choosing the Scheduling Policy::
24287 * Solaris-Specific Considerations::
24288 * Linux-Specific Considerations::
24289 * AIX-Specific Considerations::
24292 @node Summary of Run-Time Configurations
24293 @section Summary of Run-Time Configurations
24295 @multitable @columnfractions .30 .70
24296 @item @b{alpha-openvms}
24297 @item @code{@ @ }@i{rts-native (default)}
24298 @item @code{@ @ @ @ }Tasking @tab native VMS threads
24299 @item @code{@ @ @ @ }Exceptions @tab ZCX
24301 @item @b{alpha-tru64}
24302 @item @code{@ @ }@i{rts-native (default)}
24303 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
24304 @item @code{@ @ @ @ }Exceptions @tab ZCX
24306 @item @code{@ @ }@i{rts-sjlj}
24307 @item @code{@ @ @ @ }Tasking @tab native TRU64 threads
24308 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24310 @item @b{ia64-hp_linux}
24311 @item @code{@ @ }@i{rts-native (default)}
24312 @item @code{@ @ @ @ }Tasking @tab pthread library
24313 @item @code{@ @ @ @ }Exceptions @tab ZCX
24315 @item @b{ia64-hpux}
24316 @item @code{@ @ }@i{rts-native (default)}
24317 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24318 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24320 @item @b{ia64-openvms}
24321 @item @code{@ @ }@i{rts-native (default)}
24322 @item @code{@ @ @ @ }Tasking @tab native VMS threads
24323 @item @code{@ @ @ @ }Exceptions @tab ZCX
24325 @item @b{ia64-sgi_linux}
24326 @item @code{@ @ }@i{rts-native (default)}
24327 @item @code{@ @ @ @ }Tasking @tab pthread library
24328 @item @code{@ @ @ @ }Exceptions @tab ZCX
24330 @item @b{mips-irix}
24331 @item @code{@ @ }@i{rts-native (default)}
24332 @item @code{@ @ @ @ }Tasking @tab native IRIX threads
24333 @item @code{@ @ @ @ }Exceptions @tab ZCX
24336 @item @code{@ @ }@i{rts-native (default)}
24337 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24338 @item @code{@ @ @ @ }Exceptions @tab ZCX
24340 @item @code{@ @ }@i{rts-sjlj}
24341 @item @code{@ @ @ @ }Tasking @tab native HP-UX threads
24342 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24345 @item @code{@ @ }@i{rts-native (default)}
24346 @item @code{@ @ @ @ }Tasking @tab native AIX threads
24347 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24349 @item @b{ppc-darwin}
24350 @item @code{@ @ }@i{rts-native (default)}
24351 @item @code{@ @ @ @ }Tasking @tab native MacOS threads
24352 @item @code{@ @ @ @ }Exceptions @tab ZCX
24354 @item @b{sparc-solaris} @tab
24355 @item @code{@ @ }@i{rts-native (default)}
24356 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24357 @item @code{@ @ @ @ }Exceptions @tab ZCX
24359 @item @code{@ @ }@i{rts-pthread}
24360 @item @code{@ @ @ @ }Tasking @tab pthread library
24361 @item @code{@ @ @ @ }Exceptions @tab ZCX
24363 @item @code{@ @ }@i{rts-sjlj}
24364 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24365 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24367 @item @b{sparc64-solaris} @tab
24368 @item @code{@ @ }@i{rts-native (default)}
24369 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
24370 @item @code{@ @ @ @ }Exceptions @tab ZCX
24372 @item @b{x86-linux}
24373 @item @code{@ @ }@i{rts-native (default)}
24374 @item @code{@ @ @ @ }Tasking @tab pthread library
24375 @item @code{@ @ @ @ }Exceptions @tab ZCX
24377 @item @code{@ @ }@i{rts-sjlj}
24378 @item @code{@ @ @ @ }Tasking @tab pthread library
24379 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24382 @item @code{@ @ }@i{rts-native (default)}
24383 @item @code{@ @ @ @ }Tasking @tab native LynxOS threads
24384 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24386 @item @b{x86-solaris}
24387 @item @code{@ @ }@i{rts-native (default)}
24388 @item @code{@ @ @ @ }Tasking @tab native Solaris threads
24389 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24391 @item @b{x86-windows}
24392 @item @code{@ @ }@i{rts-native (default)}
24393 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
24394 @item @code{@ @ @ @ }Exceptions @tab ZCX
24396 @item @code{@ @ }@i{rts-sjlj (default)}
24397 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
24398 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24400 @item @b{x86_64-linux}
24401 @item @code{@ @ }@i{rts-native (default)}
24402 @item @code{@ @ @ @ }Tasking @tab pthread library
24403 @item @code{@ @ @ @ }Exceptions @tab ZCX
24405 @item @code{@ @ }@i{rts-sjlj}
24406 @item @code{@ @ @ @ }Tasking @tab pthread library
24407 @item @code{@ @ @ @ }Exceptions @tab SJLJ
24411 @node Specifying a Run-Time Library
24412 @section Specifying a Run-Time Library
24415 The @file{adainclude} subdirectory containing the sources of the GNAT
24416 run-time library, and the @file{adalib} subdirectory containing the
24417 @file{ALI} files and the static and/or shared GNAT library, are located
24418 in the gcc target-dependent area:
24421 target=$prefix/lib/gcc/gcc-@i{dumpmachine}/gcc-@i{dumpversion}/
24425 As indicated above, on some platforms several run-time libraries are supplied.
24426 These libraries are installed in the target dependent area and
24427 contain a complete source and binary subdirectory. The detailed description
24428 below explains the differences between the different libraries in terms of
24429 their thread support.
24431 The default run-time library (when GNAT is installed) is @emph{rts-native}.
24432 This default run time is selected by the means of soft links.
24433 For example on x86-linux:
24439 +--- adainclude----------+
24441 +--- adalib-----------+ |
24443 +--- rts-native | |
24445 | +--- adainclude <---+
24447 | +--- adalib <----+
24458 If the @i{rts-sjlj} library is to be selected on a permanent basis,
24459 these soft links can be modified with the following commands:
24463 $ rm -f adainclude adalib
24464 $ ln -s rts-sjlj/adainclude adainclude
24465 $ ln -s rts-sjlj/adalib adalib
24469 Alternatively, you can specify @file{rts-sjlj/adainclude} in the file
24470 @file{$target/ada_source_path} and @file{rts-sjlj/adalib} in
24471 @file{$target/ada_object_path}.
24473 Selecting another run-time library temporarily can be
24474 achieved by using the @option{--RTS} switch, e.g., @option{--RTS=sjlj}
24475 @cindex @option{--RTS} option
24477 @node Choosing the Scheduling Policy
24478 @section Choosing the Scheduling Policy
24481 When using a POSIX threads implementation, you have a choice of several
24482 scheduling policies: @code{SCHED_FIFO},
24483 @cindex @code{SCHED_FIFO} scheduling policy
24485 @cindex @code{SCHED_RR} scheduling policy
24486 and @code{SCHED_OTHER}.
24487 @cindex @code{SCHED_OTHER} scheduling policy
24488 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
24489 or @code{SCHED_RR} requires special (e.g., root) privileges.
24491 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
24493 @cindex @code{SCHED_FIFO} scheduling policy
24494 you can use one of the following:
24498 @code{pragma Time_Slice (0.0)}
24499 @cindex pragma Time_Slice
24501 the corresponding binder option @option{-T0}
24502 @cindex @option{-T0} option
24504 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
24505 @cindex pragma Task_Dispatching_Policy
24509 To specify @code{SCHED_RR},
24510 @cindex @code{SCHED_RR} scheduling policy
24511 you should use @code{pragma Time_Slice} with a
24512 value greater than @code{0.0}, or else use the corresponding @option{-T}
24515 @node Solaris-Specific Considerations
24516 @section Solaris-Specific Considerations
24517 @cindex Solaris Sparc threads libraries
24520 This section addresses some topics related to the various threads libraries
24524 * Solaris Threads Issues::
24527 @node Solaris Threads Issues
24528 @subsection Solaris Threads Issues
24531 GNAT under Solaris/Sparc 32 bits comes with an alternate tasking run-time
24532 library based on POSIX threads --- @emph{rts-pthread}.
24533 @cindex rts-pthread threads library
24534 This run-time library has the advantage of being mostly shared across all
24535 POSIX-compliant thread implementations, and it also provides under
24536 @w{Solaris 8} the @code{PTHREAD_PRIO_INHERIT}
24537 @cindex @code{PTHREAD_PRIO_INHERIT} policy (under rts-pthread)
24538 and @code{PTHREAD_PRIO_PROTECT}
24539 @cindex @code{PTHREAD_PRIO_PROTECT} policy (under rts-pthread)
24540 semantics that can be selected using the predefined pragma
24541 @code{Locking_Policy}
24542 @cindex pragma Locking_Policy (under rts-pthread)
24544 @code{Inheritance_Locking} and @code{Ceiling_Locking} as the policy.
24545 @cindex @code{Inheritance_Locking} (under rts-pthread)
24546 @cindex @code{Ceiling_Locking} (under rts-pthread)
24548 As explained above, the native run-time library is based on the Solaris thread
24549 library (@code{libthread}) and is the default library.
24551 When the Solaris threads library is used (this is the default), programs
24552 compiled with GNAT can automatically take advantage of
24553 and can thus execute on multiple processors.
24554 The user can alternatively specify a processor on which the program should run
24555 to emulate a single-processor system. The multiprocessor / uniprocessor choice
24557 setting the environment variable @code{GNAT_PROCESSOR}
24558 @cindex @code{GNAT_PROCESSOR} environment variable (on Sparc Solaris)
24559 to one of the following:
24563 Use the default configuration (run the program on all
24564 available processors) - this is the same as having
24565 @code{GNAT_PROCESSOR} unset
24568 Let the run-time implementation choose one processor and run the program on
24571 @item 0 .. Last_Proc
24572 Run the program on the specified processor.
24573 @code{Last_Proc} is equal to @code{_SC_NPROCESSORS_CONF - 1}
24574 (where @code{_SC_NPROCESSORS_CONF} is a system variable).
24577 @node Linux-Specific Considerations
24578 @section Linux-Specific Considerations
24579 @cindex Linux threads libraries
24582 On GNU/Linux without NPTL support (usually system with GNU C Library
24583 older than 2.3), the signal model is not POSIX compliant, which means
24584 that to send a signal to the process, you need to send the signal to all
24585 threads, e.g. by using @code{killpg()}.
24587 @node AIX-Specific Considerations
24588 @section AIX-Specific Considerations
24589 @cindex AIX resolver library
24592 On AIX, the resolver library initializes some internal structure on
24593 the first call to @code{get*by*} functions, which are used to implement
24594 @code{GNAT.Sockets.Get_Host_By_Name} and
24595 @code{GNAT.Sockets.Get_Host_By_Address}.
24596 If such initialization occurs within an Ada task, and the stack size for
24597 the task is the default size, a stack overflow may occur.
24599 To avoid this overflow, the user should either ensure that the first call
24600 to @code{GNAT.Sockets.Get_Host_By_Name} or
24601 @code{GNAT.Sockets.Get_Host_By_Addrss}
24602 occurs in the environment task, or use @code{pragma Storage_Size} to
24603 specify a sufficiently large size for the stack of the task that contains
24606 @c *******************************
24607 @node Example of Binder Output File
24608 @appendix Example of Binder Output File
24611 This Appendix displays the source code for @command{gnatbind}'s output
24612 file generated for a simple ``Hello World'' program.
24613 Comments have been added for clarification purposes.
24615 @smallexample @c adanocomment
24619 -- The package is called Ada_Main unless this name is actually used
24620 -- as a unit name in the partition, in which case some other unique
24624 package ada_main is
24626 Elab_Final_Code : Integer;
24627 pragma Import (C, Elab_Final_Code, "__gnat_inside_elab_final_code");
24629 -- The main program saves the parameters (argument count,
24630 -- argument values, environment pointer) in global variables
24631 -- for later access by other units including
24632 -- Ada.Command_Line.
24634 gnat_argc : Integer;
24635 gnat_argv : System.Address;
24636 gnat_envp : System.Address;
24638 -- The actual variables are stored in a library routine. This
24639 -- is useful for some shared library situations, where there
24640 -- are problems if variables are not in the library.
24642 pragma Import (C, gnat_argc);
24643 pragma Import (C, gnat_argv);
24644 pragma Import (C, gnat_envp);
24646 -- The exit status is similarly an external location
24648 gnat_exit_status : Integer;
24649 pragma Import (C, gnat_exit_status);
24651 GNAT_Version : constant String :=
24652 "GNAT Version: 6.0.0w (20061115)";
24653 pragma Export (C, GNAT_Version, "__gnat_version");
24655 -- This is the generated adafinal routine that performs
24656 -- finalization at the end of execution. In the case where
24657 -- Ada is the main program, this main program makes a call
24658 -- to adafinal at program termination.
24660 procedure adafinal;
24661 pragma Export (C, adafinal, "adafinal");
24663 -- This is the generated adainit routine that performs
24664 -- initialization at the start of execution. In the case
24665 -- where Ada is the main program, this main program makes
24666 -- a call to adainit at program startup.
24669 pragma Export (C, adainit, "adainit");
24671 -- This routine is called at the start of execution. It is
24672 -- a dummy routine that is used by the debugger to breakpoint
24673 -- at the start of execution.
24675 procedure Break_Start;
24676 pragma Import (C, Break_Start, "__gnat_break_start");
24678 -- This is the actual generated main program (it would be
24679 -- suppressed if the no main program switch were used). As
24680 -- required by standard system conventions, this program has
24681 -- the external name main.
24685 argv : System.Address;
24686 envp : System.Address)
24688 pragma Export (C, main, "main");
24690 -- The following set of constants give the version
24691 -- identification values for every unit in the bound
24692 -- partition. This identification is computed from all
24693 -- dependent semantic units, and corresponds to the
24694 -- string that would be returned by use of the
24695 -- Body_Version or Version attributes.
24697 type Version_32 is mod 2 ** 32;
24698 u00001 : constant Version_32 := 16#7880BEB3#;
24699 u00002 : constant Version_32 := 16#0D24CBD0#;
24700 u00003 : constant Version_32 := 16#3283DBEB#;
24701 u00004 : constant Version_32 := 16#2359F9ED#;
24702 u00005 : constant Version_32 := 16#664FB847#;
24703 u00006 : constant Version_32 := 16#68E803DF#;
24704 u00007 : constant Version_32 := 16#5572E604#;
24705 u00008 : constant Version_32 := 16#46B173D8#;
24706 u00009 : constant Version_32 := 16#156A40CF#;
24707 u00010 : constant Version_32 := 16#033DABE0#;
24708 u00011 : constant Version_32 := 16#6AB38FEA#;
24709 u00012 : constant Version_32 := 16#22B6217D#;
24710 u00013 : constant Version_32 := 16#68A22947#;
24711 u00014 : constant Version_32 := 16#18CC4A56#;
24712 u00015 : constant Version_32 := 16#08258E1B#;
24713 u00016 : constant Version_32 := 16#367D5222#;
24714 u00017 : constant Version_32 := 16#20C9ECA4#;
24715 u00018 : constant Version_32 := 16#50D32CB6#;
24716 u00019 : constant Version_32 := 16#39A8BB77#;
24717 u00020 : constant Version_32 := 16#5CF8FA2B#;
24718 u00021 : constant Version_32 := 16#2F1EB794#;
24719 u00022 : constant Version_32 := 16#31AB6444#;
24720 u00023 : constant Version_32 := 16#1574B6E9#;
24721 u00024 : constant Version_32 := 16#5109C189#;
24722 u00025 : constant Version_32 := 16#56D770CD#;
24723 u00026 : constant Version_32 := 16#02F9DE3D#;
24724 u00027 : constant Version_32 := 16#08AB6B2C#;
24725 u00028 : constant Version_32 := 16#3FA37670#;
24726 u00029 : constant Version_32 := 16#476457A0#;
24727 u00030 : constant Version_32 := 16#731E1B6E#;
24728 u00031 : constant Version_32 := 16#23C2E789#;
24729 u00032 : constant Version_32 := 16#0F1BD6A1#;
24730 u00033 : constant Version_32 := 16#7C25DE96#;
24731 u00034 : constant Version_32 := 16#39ADFFA2#;
24732 u00035 : constant Version_32 := 16#571DE3E7#;
24733 u00036 : constant Version_32 := 16#5EB646AB#;
24734 u00037 : constant Version_32 := 16#4249379B#;
24735 u00038 : constant Version_32 := 16#0357E00A#;
24736 u00039 : constant Version_32 := 16#3784FB72#;
24737 u00040 : constant Version_32 := 16#2E723019#;
24738 u00041 : constant Version_32 := 16#623358EA#;
24739 u00042 : constant Version_32 := 16#107F9465#;
24740 u00043 : constant Version_32 := 16#6843F68A#;
24741 u00044 : constant Version_32 := 16#63305874#;
24742 u00045 : constant Version_32 := 16#31E56CE1#;
24743 u00046 : constant Version_32 := 16#02917970#;
24744 u00047 : constant Version_32 := 16#6CCBA70E#;
24745 u00048 : constant Version_32 := 16#41CD4204#;
24746 u00049 : constant Version_32 := 16#572E3F58#;
24747 u00050 : constant Version_32 := 16#20729FF5#;
24748 u00051 : constant Version_32 := 16#1D4F93E8#;
24749 u00052 : constant Version_32 := 16#30B2EC3D#;
24750 u00053 : constant Version_32 := 16#34054F96#;
24751 u00054 : constant Version_32 := 16#5A199860#;
24752 u00055 : constant Version_32 := 16#0E7F912B#;
24753 u00056 : constant Version_32 := 16#5760634A#;
24754 u00057 : constant Version_32 := 16#5D851835#;
24756 -- The following Export pragmas export the version numbers
24757 -- with symbolic names ending in B (for body) or S
24758 -- (for spec) so that they can be located in a link. The
24759 -- information provided here is sufficient to track down
24760 -- the exact versions of units used in a given build.
24762 pragma Export (C, u00001, "helloB");
24763 pragma Export (C, u00002, "system__standard_libraryB");
24764 pragma Export (C, u00003, "system__standard_libraryS");
24765 pragma Export (C, u00004, "adaS");
24766 pragma Export (C, u00005, "ada__text_ioB");
24767 pragma Export (C, u00006, "ada__text_ioS");
24768 pragma Export (C, u00007, "ada__exceptionsB");
24769 pragma Export (C, u00008, "ada__exceptionsS");
24770 pragma Export (C, u00009, "gnatS");
24771 pragma Export (C, u00010, "gnat__heap_sort_aB");
24772 pragma Export (C, u00011, "gnat__heap_sort_aS");
24773 pragma Export (C, u00012, "systemS");
24774 pragma Export (C, u00013, "system__exception_tableB");
24775 pragma Export (C, u00014, "system__exception_tableS");
24776 pragma Export (C, u00015, "gnat__htableB");
24777 pragma Export (C, u00016, "gnat__htableS");
24778 pragma Export (C, u00017, "system__exceptionsS");
24779 pragma Export (C, u00018, "system__machine_state_operationsB");
24780 pragma Export (C, u00019, "system__machine_state_operationsS");
24781 pragma Export (C, u00020, "system__machine_codeS");
24782 pragma Export (C, u00021, "system__storage_elementsB");
24783 pragma Export (C, u00022, "system__storage_elementsS");
24784 pragma Export (C, u00023, "system__secondary_stackB");
24785 pragma Export (C, u00024, "system__secondary_stackS");
24786 pragma Export (C, u00025, "system__parametersB");
24787 pragma Export (C, u00026, "system__parametersS");
24788 pragma Export (C, u00027, "system__soft_linksB");
24789 pragma Export (C, u00028, "system__soft_linksS");
24790 pragma Export (C, u00029, "system__stack_checkingB");
24791 pragma Export (C, u00030, "system__stack_checkingS");
24792 pragma Export (C, u00031, "system__tracebackB");
24793 pragma Export (C, u00032, "system__tracebackS");
24794 pragma Export (C, u00033, "ada__streamsS");
24795 pragma Export (C, u00034, "ada__tagsB");
24796 pragma Export (C, u00035, "ada__tagsS");
24797 pragma Export (C, u00036, "system__string_opsB");
24798 pragma Export (C, u00037, "system__string_opsS");
24799 pragma Export (C, u00038, "interfacesS");
24800 pragma Export (C, u00039, "interfaces__c_streamsB");
24801 pragma Export (C, u00040, "interfaces__c_streamsS");
24802 pragma Export (C, u00041, "system__file_ioB");
24803 pragma Export (C, u00042, "system__file_ioS");
24804 pragma Export (C, u00043, "ada__finalizationB");
24805 pragma Export (C, u00044, "ada__finalizationS");
24806 pragma Export (C, u00045, "system__finalization_rootB");
24807 pragma Export (C, u00046, "system__finalization_rootS");
24808 pragma Export (C, u00047, "system__finalization_implementationB");
24809 pragma Export (C, u00048, "system__finalization_implementationS");
24810 pragma Export (C, u00049, "system__string_ops_concat_3B");
24811 pragma Export (C, u00050, "system__string_ops_concat_3S");
24812 pragma Export (C, u00051, "system__stream_attributesB");
24813 pragma Export (C, u00052, "system__stream_attributesS");
24814 pragma Export (C, u00053, "ada__io_exceptionsS");
24815 pragma Export (C, u00054, "system__unsigned_typesS");
24816 pragma Export (C, u00055, "system__file_control_blockS");
24817 pragma Export (C, u00056, "ada__finalization__list_controllerB");
24818 pragma Export (C, u00057, "ada__finalization__list_controllerS");
24820 -- BEGIN ELABORATION ORDER
24823 -- gnat.heap_sort_a (spec)
24824 -- gnat.heap_sort_a (body)
24825 -- gnat.htable (spec)
24826 -- gnat.htable (body)
24827 -- interfaces (spec)
24829 -- system.machine_code (spec)
24830 -- system.parameters (spec)
24831 -- system.parameters (body)
24832 -- interfaces.c_streams (spec)
24833 -- interfaces.c_streams (body)
24834 -- system.standard_library (spec)
24835 -- ada.exceptions (spec)
24836 -- system.exception_table (spec)
24837 -- system.exception_table (body)
24838 -- ada.io_exceptions (spec)
24839 -- system.exceptions (spec)
24840 -- system.storage_elements (spec)
24841 -- system.storage_elements (body)
24842 -- system.machine_state_operations (spec)
24843 -- system.machine_state_operations (body)
24844 -- system.secondary_stack (spec)
24845 -- system.stack_checking (spec)
24846 -- system.soft_links (spec)
24847 -- system.soft_links (body)
24848 -- system.stack_checking (body)
24849 -- system.secondary_stack (body)
24850 -- system.standard_library (body)
24851 -- system.string_ops (spec)
24852 -- system.string_ops (body)
24855 -- ada.streams (spec)
24856 -- system.finalization_root (spec)
24857 -- system.finalization_root (body)
24858 -- system.string_ops_concat_3 (spec)
24859 -- system.string_ops_concat_3 (body)
24860 -- system.traceback (spec)
24861 -- system.traceback (body)
24862 -- ada.exceptions (body)
24863 -- system.unsigned_types (spec)
24864 -- system.stream_attributes (spec)
24865 -- system.stream_attributes (body)
24866 -- system.finalization_implementation (spec)
24867 -- system.finalization_implementation (body)
24868 -- ada.finalization (spec)
24869 -- ada.finalization (body)
24870 -- ada.finalization.list_controller (spec)
24871 -- ada.finalization.list_controller (body)
24872 -- system.file_control_block (spec)
24873 -- system.file_io (spec)
24874 -- system.file_io (body)
24875 -- ada.text_io (spec)
24876 -- ada.text_io (body)
24878 -- END ELABORATION ORDER
24882 -- The following source file name pragmas allow the generated file
24883 -- names to be unique for different main programs. They are needed
24884 -- since the package name will always be Ada_Main.
24886 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
24887 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
24889 -- Generated package body for Ada_Main starts here
24891 package body ada_main is
24893 -- The actual finalization is performed by calling the
24894 -- library routine in System.Standard_Library.Adafinal
24896 procedure Do_Finalize;
24897 pragma Import (C, Do_Finalize, "system__standard_library__adafinal");
24904 procedure adainit is
24906 -- These booleans are set to True once the associated unit has
24907 -- been elaborated. It is also used to avoid elaborating the
24908 -- same unit twice.
24911 pragma Import (Ada, E040, "interfaces__c_streams_E");
24914 pragma Import (Ada, E008, "ada__exceptions_E");
24917 pragma Import (Ada, E014, "system__exception_table_E");
24920 pragma Import (Ada, E053, "ada__io_exceptions_E");
24923 pragma Import (Ada, E017, "system__exceptions_E");
24926 pragma Import (Ada, E024, "system__secondary_stack_E");
24929 pragma Import (Ada, E030, "system__stack_checking_E");
24932 pragma Import (Ada, E028, "system__soft_links_E");
24935 pragma Import (Ada, E035, "ada__tags_E");
24938 pragma Import (Ada, E033, "ada__streams_E");
24941 pragma Import (Ada, E046, "system__finalization_root_E");
24944 pragma Import (Ada, E048, "system__finalization_implementation_E");
24947 pragma Import (Ada, E044, "ada__finalization_E");
24950 pragma Import (Ada, E057, "ada__finalization__list_controller_E");
24953 pragma Import (Ada, E055, "system__file_control_block_E");
24956 pragma Import (Ada, E042, "system__file_io_E");
24959 pragma Import (Ada, E006, "ada__text_io_E");
24961 -- Set_Globals is a library routine that stores away the
24962 -- value of the indicated set of global values in global
24963 -- variables within the library.
24965 procedure Set_Globals
24966 (Main_Priority : Integer;
24967 Time_Slice_Value : Integer;
24968 WC_Encoding : Character;
24969 Locking_Policy : Character;
24970 Queuing_Policy : Character;
24971 Task_Dispatching_Policy : Character;
24972 Adafinal : System.Address;
24973 Unreserve_All_Interrupts : Integer;
24974 Exception_Tracebacks : Integer);
24975 @findex __gnat_set_globals
24976 pragma Import (C, Set_Globals, "__gnat_set_globals");
24978 -- SDP_Table_Build is a library routine used to build the
24979 -- exception tables. See unit Ada.Exceptions in files
24980 -- a-except.ads/adb for full details of how zero cost
24981 -- exception handling works. This procedure, the call to
24982 -- it, and the two following tables are all omitted if the
24983 -- build is in longjmp/setjump exception mode.
24985 @findex SDP_Table_Build
24986 @findex Zero Cost Exceptions
24987 procedure SDP_Table_Build
24988 (SDP_Addresses : System.Address;
24989 SDP_Count : Natural;
24990 Elab_Addresses : System.Address;
24991 Elab_Addr_Count : Natural);
24992 pragma Import (C, SDP_Table_Build, "__gnat_SDP_Table_Build");
24994 -- Table of Unit_Exception_Table addresses. Used for zero
24995 -- cost exception handling to build the top level table.
24997 ST : aliased constant array (1 .. 23) of System.Address := (
24999 Ada.Text_Io'UET_Address,
25000 Ada.Exceptions'UET_Address,
25001 Gnat.Heap_Sort_A'UET_Address,
25002 System.Exception_Table'UET_Address,
25003 System.Machine_State_Operations'UET_Address,
25004 System.Secondary_Stack'UET_Address,
25005 System.Parameters'UET_Address,
25006 System.Soft_Links'UET_Address,
25007 System.Stack_Checking'UET_Address,
25008 System.Traceback'UET_Address,
25009 Ada.Streams'UET_Address,
25010 Ada.Tags'UET_Address,
25011 System.String_Ops'UET_Address,
25012 Interfaces.C_Streams'UET_Address,
25013 System.File_Io'UET_Address,
25014 Ada.Finalization'UET_Address,
25015 System.Finalization_Root'UET_Address,
25016 System.Finalization_Implementation'UET_Address,
25017 System.String_Ops_Concat_3'UET_Address,
25018 System.Stream_Attributes'UET_Address,
25019 System.File_Control_Block'UET_Address,
25020 Ada.Finalization.List_Controller'UET_Address);
25022 -- Table of addresses of elaboration routines. Used for
25023 -- zero cost exception handling to make sure these
25024 -- addresses are included in the top level procedure
25027 EA : aliased constant array (1 .. 23) of System.Address := (
25028 adainit'Code_Address,
25029 Do_Finalize'Code_Address,
25030 Ada.Exceptions'Elab_Spec'Address,
25031 System.Exceptions'Elab_Spec'Address,
25032 Interfaces.C_Streams'Elab_Spec'Address,
25033 System.Exception_Table'Elab_Body'Address,
25034 Ada.Io_Exceptions'Elab_Spec'Address,
25035 System.Stack_Checking'Elab_Spec'Address,
25036 System.Soft_Links'Elab_Body'Address,
25037 System.Secondary_Stack'Elab_Body'Address,
25038 Ada.Tags'Elab_Spec'Address,
25039 Ada.Tags'Elab_Body'Address,
25040 Ada.Streams'Elab_Spec'Address,
25041 System.Finalization_Root'Elab_Spec'Address,
25042 Ada.Exceptions'Elab_Body'Address,
25043 System.Finalization_Implementation'Elab_Spec'Address,
25044 System.Finalization_Implementation'Elab_Body'Address,
25045 Ada.Finalization'Elab_Spec'Address,
25046 Ada.Finalization.List_Controller'Elab_Spec'Address,
25047 System.File_Control_Block'Elab_Spec'Address,
25048 System.File_Io'Elab_Body'Address,
25049 Ada.Text_Io'Elab_Spec'Address,
25050 Ada.Text_Io'Elab_Body'Address);
25052 -- Start of processing for adainit
25056 -- Call SDP_Table_Build to build the top level procedure
25057 -- table for zero cost exception handling (omitted in
25058 -- longjmp/setjump mode).
25060 SDP_Table_Build (ST'Address, 23, EA'Address, 23);
25062 -- Call Set_Globals to record various information for
25063 -- this partition. The values are derived by the binder
25064 -- from information stored in the ali files by the compiler.
25066 @findex __gnat_set_globals
25068 (Main_Priority => -1,
25069 -- Priority of main program, -1 if no pragma Priority used
25071 Time_Slice_Value => -1,
25072 -- Time slice from Time_Slice pragma, -1 if none used
25074 WC_Encoding => 'b',
25075 -- Wide_Character encoding used, default is brackets
25077 Locking_Policy => ' ',
25078 -- Locking_Policy used, default of space means not
25079 -- specified, otherwise it is the first character of
25080 -- the policy name.
25082 Queuing_Policy => ' ',
25083 -- Queuing_Policy used, default of space means not
25084 -- specified, otherwise it is the first character of
25085 -- the policy name.
25087 Task_Dispatching_Policy => ' ',
25088 -- Task_Dispatching_Policy used, default of space means
25089 -- not specified, otherwise first character of the
25092 Adafinal => System.Null_Address,
25093 -- Address of Adafinal routine, not used anymore
25095 Unreserve_All_Interrupts => 0,
25096 -- Set true if pragma Unreserve_All_Interrupts was used
25098 Exception_Tracebacks => 0);
25099 -- Indicates if exception tracebacks are enabled
25101 Elab_Final_Code := 1;
25103 -- Now we have the elaboration calls for all units in the partition.
25104 -- The Elab_Spec and Elab_Body attributes generate references to the
25105 -- implicit elaboration procedures generated by the compiler for
25106 -- each unit that requires elaboration.
25109 Interfaces.C_Streams'Elab_Spec;
25113 Ada.Exceptions'Elab_Spec;
25116 System.Exception_Table'Elab_Body;
25120 Ada.Io_Exceptions'Elab_Spec;
25124 System.Exceptions'Elab_Spec;
25128 System.Stack_Checking'Elab_Spec;
25131 System.Soft_Links'Elab_Body;
25136 System.Secondary_Stack'Elab_Body;
25140 Ada.Tags'Elab_Spec;
25143 Ada.Tags'Elab_Body;
25147 Ada.Streams'Elab_Spec;
25151 System.Finalization_Root'Elab_Spec;
25155 Ada.Exceptions'Elab_Body;
25159 System.Finalization_Implementation'Elab_Spec;
25162 System.Finalization_Implementation'Elab_Body;
25166 Ada.Finalization'Elab_Spec;
25170 Ada.Finalization.List_Controller'Elab_Spec;
25174 System.File_Control_Block'Elab_Spec;
25178 System.File_Io'Elab_Body;
25182 Ada.Text_Io'Elab_Spec;
25185 Ada.Text_Io'Elab_Body;
25189 Elab_Final_Code := 0;
25197 procedure adafinal is
25206 -- main is actually a function, as in the ANSI C standard,
25207 -- defined to return the exit status. The three parameters
25208 -- are the argument count, argument values and environment
25211 @findex Main Program
25214 argv : System.Address;
25215 envp : System.Address)
25218 -- The initialize routine performs low level system
25219 -- initialization using a standard library routine which
25220 -- sets up signal handling and performs any other
25221 -- required setup. The routine can be found in file
25224 @findex __gnat_initialize
25225 procedure initialize;
25226 pragma Import (C, initialize, "__gnat_initialize");
25228 -- The finalize routine performs low level system
25229 -- finalization using a standard library routine. The
25230 -- routine is found in file a-final.c and in the standard
25231 -- distribution is a dummy routine that does nothing, so
25232 -- really this is a hook for special user finalization.
25234 @findex __gnat_finalize
25235 procedure finalize;
25236 pragma Import (C, finalize, "__gnat_finalize");
25238 -- We get to the main program of the partition by using
25239 -- pragma Import because if we try to with the unit and
25240 -- call it Ada style, then not only do we waste time
25241 -- recompiling it, but also, we don't really know the right
25242 -- switches (e.g. identifier character set) to be used
25245 procedure Ada_Main_Program;
25246 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
25248 -- Start of processing for main
25251 -- Save global variables
25257 -- Call low level system initialization
25261 -- Call our generated Ada initialization routine
25265 -- This is the point at which we want the debugger to get
25270 -- Now we call the main program of the partition
25274 -- Perform Ada finalization
25278 -- Perform low level system finalization
25282 -- Return the proper exit status
25283 return (gnat_exit_status);
25286 -- This section is entirely comments, so it has no effect on the
25287 -- compilation of the Ada_Main package. It provides the list of
25288 -- object files and linker options, as well as some standard
25289 -- libraries needed for the link. The gnatlink utility parses
25290 -- this b~hello.adb file to read these comment lines to generate
25291 -- the appropriate command line arguments for the call to the
25292 -- system linker. The BEGIN/END lines are used for sentinels for
25293 -- this parsing operation.
25295 -- The exact file names will of course depend on the environment,
25296 -- host/target and location of files on the host system.
25298 @findex Object file list
25299 -- BEGIN Object file/option list
25302 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
25303 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
25304 -- END Object file/option list
25310 The Ada code in the above example is exactly what is generated by the
25311 binder. We have added comments to more clearly indicate the function
25312 of each part of the generated @code{Ada_Main} package.
25314 The code is standard Ada in all respects, and can be processed by any
25315 tools that handle Ada. In particular, it is possible to use the debugger
25316 in Ada mode to debug the generated @code{Ada_Main} package. For example,
25317 suppose that for reasons that you do not understand, your program is crashing
25318 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
25319 you can place a breakpoint on the call:
25321 @smallexample @c ada
25322 Ada.Text_Io'Elab_Body;
25326 and trace the elaboration routine for this package to find out where
25327 the problem might be (more usually of course you would be debugging
25328 elaboration code in your own application).
25330 @node Elaboration Order Handling in GNAT
25331 @appendix Elaboration Order Handling in GNAT
25332 @cindex Order of elaboration
25333 @cindex Elaboration control
25336 * Elaboration Code::
25337 * Checking the Elaboration Order::
25338 * Controlling the Elaboration Order::
25339 * Controlling Elaboration in GNAT - Internal Calls::
25340 * Controlling Elaboration in GNAT - External Calls::
25341 * Default Behavior in GNAT - Ensuring Safety::
25342 * Treatment of Pragma Elaborate::
25343 * Elaboration Issues for Library Tasks::
25344 * Mixing Elaboration Models::
25345 * What to Do If the Default Elaboration Behavior Fails::
25346 * Elaboration for Access-to-Subprogram Values::
25347 * Summary of Procedures for Elaboration Control::
25348 * Other Elaboration Order Considerations::
25352 This chapter describes the handling of elaboration code in Ada and
25353 in GNAT, and discusses how the order of elaboration of program units can
25354 be controlled in GNAT, either automatically or with explicit programming
25357 @node Elaboration Code
25358 @section Elaboration Code
25361 Ada provides rather general mechanisms for executing code at elaboration
25362 time, that is to say before the main program starts executing. Such code arises
25366 @item Initializers for variables.
25367 Variables declared at the library level, in package specs or bodies, can
25368 require initialization that is performed at elaboration time, as in:
25369 @smallexample @c ada
25371 Sqrt_Half : Float := Sqrt (0.5);
25375 @item Package initialization code
25376 Code in a @code{BEGIN-END} section at the outer level of a package body is
25377 executed as part of the package body elaboration code.
25379 @item Library level task allocators
25380 Tasks that are declared using task allocators at the library level
25381 start executing immediately and hence can execute at elaboration time.
25385 Subprogram calls are possible in any of these contexts, which means that
25386 any arbitrary part of the program may be executed as part of the elaboration
25387 code. It is even possible to write a program which does all its work at
25388 elaboration time, with a null main program, although stylistically this
25389 would usually be considered an inappropriate way to structure
25392 An important concern arises in the context of elaboration code:
25393 we have to be sure that it is executed in an appropriate order. What we
25394 have is a series of elaboration code sections, potentially one section
25395 for each unit in the program. It is important that these execute
25396 in the correct order. Correctness here means that, taking the above
25397 example of the declaration of @code{Sqrt_Half},
25398 if some other piece of
25399 elaboration code references @code{Sqrt_Half},
25400 then it must run after the
25401 section of elaboration code that contains the declaration of
25404 There would never be any order of elaboration problem if we made a rule
25405 that whenever you @code{with} a unit, you must elaborate both the spec and body
25406 of that unit before elaborating the unit doing the @code{with}'ing:
25408 @smallexample @c ada
25412 package Unit_2 is ...
25418 would require that both the body and spec of @code{Unit_1} be elaborated
25419 before the spec of @code{Unit_2}. However, a rule like that would be far too
25420 restrictive. In particular, it would make it impossible to have routines
25421 in separate packages that were mutually recursive.
25423 You might think that a clever enough compiler could look at the actual
25424 elaboration code and determine an appropriate correct order of elaboration,
25425 but in the general case, this is not possible. Consider the following
25428 In the body of @code{Unit_1}, we have a procedure @code{Func_1}
25430 the variable @code{Sqrt_1}, which is declared in the elaboration code
25431 of the body of @code{Unit_1}:
25433 @smallexample @c ada
25435 Sqrt_1 : Float := Sqrt (0.1);
25440 The elaboration code of the body of @code{Unit_1} also contains:
25442 @smallexample @c ada
25445 if expression_1 = 1 then
25446 Q := Unit_2.Func_2;
25453 @code{Unit_2} is exactly parallel,
25454 it has a procedure @code{Func_2} that references
25455 the variable @code{Sqrt_2}, which is declared in the elaboration code of
25456 the body @code{Unit_2}:
25458 @smallexample @c ada
25460 Sqrt_2 : Float := Sqrt (0.1);
25465 The elaboration code of the body of @code{Unit_2} also contains:
25467 @smallexample @c ada
25470 if expression_2 = 2 then
25471 Q := Unit_1.Func_1;
25478 Now the question is, which of the following orders of elaboration is
25503 If you carefully analyze the flow here, you will see that you cannot tell
25504 at compile time the answer to this question.
25505 If @code{expression_1} is not equal to 1,
25506 and @code{expression_2} is not equal to 2,
25507 then either order is acceptable, because neither of the function calls is
25508 executed. If both tests evaluate to true, then neither order is acceptable
25509 and in fact there is no correct order.
25511 If one of the two expressions is true, and the other is false, then one
25512 of the above orders is correct, and the other is incorrect. For example,
25513 if @code{expression_1} = 1 and @code{expression_2} /= 2,
25514 then the call to @code{Func_2}
25515 will occur, but not the call to @code{Func_1.}
25516 This means that it is essential
25517 to elaborate the body of @code{Unit_1} before
25518 the body of @code{Unit_2}, so the first
25519 order of elaboration is correct and the second is wrong.
25521 By making @code{expression_1} and @code{expression_2}
25522 depend on input data, or perhaps
25523 the time of day, we can make it impossible for the compiler or binder
25524 to figure out which of these expressions will be true, and hence it
25525 is impossible to guarantee a safe order of elaboration at run time.
25527 @node Checking the Elaboration Order
25528 @section Checking the Elaboration Order
25531 In some languages that involve the same kind of elaboration problems,
25532 e.g. Java and C++, the programmer is expected to worry about these
25533 ordering problems himself, and it is common to
25534 write a program in which an incorrect elaboration order gives
25535 surprising results, because it references variables before they
25537 Ada is designed to be a safe language, and a programmer-beware approach is
25538 clearly not sufficient. Consequently, the language provides three lines
25542 @item Standard rules
25543 Some standard rules restrict the possible choice of elaboration
25544 order. In particular, if you @code{with} a unit, then its spec is always
25545 elaborated before the unit doing the @code{with}. Similarly, a parent
25546 spec is always elaborated before the child spec, and finally
25547 a spec is always elaborated before its corresponding body.
25549 @item Dynamic elaboration checks
25550 @cindex Elaboration checks
25551 @cindex Checks, elaboration
25552 Dynamic checks are made at run time, so that if some entity is accessed
25553 before it is elaborated (typically by means of a subprogram call)
25554 then the exception (@code{Program_Error}) is raised.
25556 @item Elaboration control
25557 Facilities are provided for the programmer to specify the desired order
25561 Let's look at these facilities in more detail. First, the rules for
25562 dynamic checking. One possible rule would be simply to say that the
25563 exception is raised if you access a variable which has not yet been
25564 elaborated. The trouble with this approach is that it could require
25565 expensive checks on every variable reference. Instead Ada has two
25566 rules which are a little more restrictive, but easier to check, and
25570 @item Restrictions on calls
25571 A subprogram can only be called at elaboration time if its body
25572 has been elaborated. The rules for elaboration given above guarantee
25573 that the spec of the subprogram has been elaborated before the
25574 call, but not the body. If this rule is violated, then the
25575 exception @code{Program_Error} is raised.
25577 @item Restrictions on instantiations
25578 A generic unit can only be instantiated if the body of the generic
25579 unit has been elaborated. Again, the rules for elaboration given above
25580 guarantee that the spec of the generic unit has been elaborated
25581 before the instantiation, but not the body. If this rule is
25582 violated, then the exception @code{Program_Error} is raised.
25586 The idea is that if the body has been elaborated, then any variables
25587 it references must have been elaborated; by checking for the body being
25588 elaborated we guarantee that none of its references causes any
25589 trouble. As we noted above, this is a little too restrictive, because a
25590 subprogram that has no non-local references in its body may in fact be safe
25591 to call. However, it really would be unsafe to rely on this, because
25592 it would mean that the caller was aware of details of the implementation
25593 in the body. This goes against the basic tenets of Ada.
25595 A plausible implementation can be described as follows.
25596 A Boolean variable is associated with each subprogram
25597 and each generic unit. This variable is initialized to False, and is set to
25598 True at the point body is elaborated. Every call or instantiation checks the
25599 variable, and raises @code{Program_Error} if the variable is False.
25601 Note that one might think that it would be good enough to have one Boolean
25602 variable for each package, but that would not deal with cases of trying
25603 to call a body in the same package as the call
25604 that has not been elaborated yet.
25605 Of course a compiler may be able to do enough analysis to optimize away
25606 some of the Boolean variables as unnecessary, and @code{GNAT} indeed
25607 does such optimizations, but still the easiest conceptual model is to
25608 think of there being one variable per subprogram.
25610 @node Controlling the Elaboration Order
25611 @section Controlling the Elaboration Order
25614 In the previous section we discussed the rules in Ada which ensure
25615 that @code{Program_Error} is raised if an incorrect elaboration order is
25616 chosen. This prevents erroneous executions, but we need mechanisms to
25617 specify a correct execution and avoid the exception altogether.
25618 To achieve this, Ada provides a number of features for controlling
25619 the order of elaboration. We discuss these features in this section.
25621 First, there are several ways of indicating to the compiler that a given
25622 unit has no elaboration problems:
25625 @item packages that do not require a body
25626 A library package that does not require a body does not permit
25627 a body (this rule was introduced in Ada 95).
25628 Thus if we have a such a package, as in:
25630 @smallexample @c ada
25633 package Definitions is
25635 type m is new integer;
25637 type a is array (1 .. 10) of m;
25638 type b is array (1 .. 20) of m;
25646 A package that @code{with}'s @code{Definitions} may safely instantiate
25647 @code{Definitions.Subp} because the compiler can determine that there
25648 definitely is no package body to worry about in this case
25651 @cindex pragma Pure
25653 Places sufficient restrictions on a unit to guarantee that
25654 no call to any subprogram in the unit can result in an
25655 elaboration problem. This means that the compiler does not need
25656 to worry about the point of elaboration of such units, and in
25657 particular, does not need to check any calls to any subprograms
25660 @item pragma Preelaborate
25661 @findex Preelaborate
25662 @cindex pragma Preelaborate
25663 This pragma places slightly less stringent restrictions on a unit than
25665 but these restrictions are still sufficient to ensure that there
25666 are no elaboration problems with any calls to the unit.
25668 @item pragma Elaborate_Body
25669 @findex Elaborate_Body
25670 @cindex pragma Elaborate_Body
25671 This pragma requires that the body of a unit be elaborated immediately
25672 after its spec. Suppose a unit @code{A} has such a pragma,
25673 and unit @code{B} does
25674 a @code{with} of unit @code{A}. Recall that the standard rules require
25675 the spec of unit @code{A}
25676 to be elaborated before the @code{with}'ing unit; given the pragma in
25677 @code{A}, we also know that the body of @code{A}
25678 will be elaborated before @code{B}, so
25679 that calls to @code{A} are safe and do not need a check.
25684 unlike pragma @code{Pure} and pragma @code{Preelaborate},
25686 @code{Elaborate_Body} does not guarantee that the program is
25687 free of elaboration problems, because it may not be possible
25688 to satisfy the requested elaboration order.
25689 Let's go back to the example with @code{Unit_1} and @code{Unit_2}.
25691 marks @code{Unit_1} as @code{Elaborate_Body},
25692 and not @code{Unit_2,} then the order of
25693 elaboration will be:
25705 Now that means that the call to @code{Func_1} in @code{Unit_2}
25706 need not be checked,
25707 it must be safe. But the call to @code{Func_2} in
25708 @code{Unit_1} may still fail if
25709 @code{Expression_1} is equal to 1,
25710 and the programmer must still take
25711 responsibility for this not being the case.
25713 If all units carry a pragma @code{Elaborate_Body}, then all problems are
25714 eliminated, except for calls entirely within a body, which are
25715 in any case fully under programmer control. However, using the pragma
25716 everywhere is not always possible.
25717 In particular, for our @code{Unit_1}/@code{Unit_2} example, if
25718 we marked both of them as having pragma @code{Elaborate_Body}, then
25719 clearly there would be no possible elaboration order.
25721 The above pragmas allow a server to guarantee safe use by clients, and
25722 clearly this is the preferable approach. Consequently a good rule
25723 is to mark units as @code{Pure} or @code{Preelaborate} if possible,
25724 and if this is not possible,
25725 mark them as @code{Elaborate_Body} if possible.
25726 As we have seen, there are situations where neither of these
25727 three pragmas can be used.
25728 So we also provide methods for clients to control the
25729 order of elaboration of the servers on which they depend:
25732 @item pragma Elaborate (unit)
25734 @cindex pragma Elaborate
25735 This pragma is placed in the context clause, after a @code{with} clause,
25736 and it requires that the body of the named unit be elaborated before
25737 the unit in which the pragma occurs. The idea is to use this pragma
25738 if the current unit calls at elaboration time, directly or indirectly,
25739 some subprogram in the named unit.
25741 @item pragma Elaborate_All (unit)
25742 @findex Elaborate_All
25743 @cindex pragma Elaborate_All
25744 This is a stronger version of the Elaborate pragma. Consider the
25748 Unit A @code{with}'s unit B and calls B.Func in elab code
25749 Unit B @code{with}'s unit C, and B.Func calls C.Func
25753 Now if we put a pragma @code{Elaborate (B)}
25754 in unit @code{A}, this ensures that the
25755 body of @code{B} is elaborated before the call, but not the
25756 body of @code{C}, so
25757 the call to @code{C.Func} could still cause @code{Program_Error} to
25760 The effect of a pragma @code{Elaborate_All} is stronger, it requires
25761 not only that the body of the named unit be elaborated before the
25762 unit doing the @code{with}, but also the bodies of all units that the
25763 named unit uses, following @code{with} links transitively. For example,
25764 if we put a pragma @code{Elaborate_All (B)} in unit @code{A},
25766 not only that the body of @code{B} be elaborated before @code{A},
25768 body of @code{C}, because @code{B} @code{with}'s @code{C}.
25772 We are now in a position to give a usage rule in Ada for avoiding
25773 elaboration problems, at least if dynamic dispatching and access to
25774 subprogram values are not used. We will handle these cases separately
25777 The rule is simple. If a unit has elaboration code that can directly or
25778 indirectly make a call to a subprogram in a @code{with}'ed unit, or instantiate
25779 a generic package in a @code{with}'ed unit,
25780 then if the @code{with}'ed unit does not have
25781 pragma @code{Pure} or @code{Preelaborate}, then the client should have
25782 a pragma @code{Elaborate_All}
25783 for the @code{with}'ed unit. By following this rule a client is
25784 assured that calls can be made without risk of an exception.
25786 For generic subprogram instantiations, the rule can be relaxed to
25787 require only a pragma @code{Elaborate} since elaborating the body
25788 of a subprogram cannot cause any transitive elaboration (we are
25789 not calling the subprogram in this case, just elaborating its
25792 If this rule is not followed, then a program may be in one of four
25796 @item No order exists
25797 No order of elaboration exists which follows the rules, taking into
25798 account any @code{Elaborate}, @code{Elaborate_All},
25799 or @code{Elaborate_Body} pragmas. In
25800 this case, an Ada compiler must diagnose the situation at bind
25801 time, and refuse to build an executable program.
25803 @item One or more orders exist, all incorrect
25804 One or more acceptable elaboration orders exists, and all of them
25805 generate an elaboration order problem. In this case, the binder
25806 can build an executable program, but @code{Program_Error} will be raised
25807 when the program is run.
25809 @item Several orders exist, some right, some incorrect
25810 One or more acceptable elaboration orders exists, and some of them
25811 work, and some do not. The programmer has not controlled
25812 the order of elaboration, so the binder may or may not pick one of
25813 the correct orders, and the program may or may not raise an
25814 exception when it is run. This is the worst case, because it means
25815 that the program may fail when moved to another compiler, or even
25816 another version of the same compiler.
25818 @item One or more orders exists, all correct
25819 One ore more acceptable elaboration orders exist, and all of them
25820 work. In this case the program runs successfully. This state of
25821 affairs can be guaranteed by following the rule we gave above, but
25822 may be true even if the rule is not followed.
25826 Note that one additional advantage of following our rules on the use
25827 of @code{Elaborate} and @code{Elaborate_All}
25828 is that the program continues to stay in the ideal (all orders OK) state
25829 even if maintenance
25830 changes some bodies of some units. Conversely, if a program that does
25831 not follow this rule happens to be safe at some point, this state of affairs
25832 may deteriorate silently as a result of maintenance changes.
25834 You may have noticed that the above discussion did not mention
25835 the use of @code{Elaborate_Body}. This was a deliberate omission. If you
25836 @code{with} an @code{Elaborate_Body} unit, it still may be the case that
25837 code in the body makes calls to some other unit, so it is still necessary
25838 to use @code{Elaborate_All} on such units.
25840 @node Controlling Elaboration in GNAT - Internal Calls
25841 @section Controlling Elaboration in GNAT - Internal Calls
25844 In the case of internal calls, i.e. calls within a single package, the
25845 programmer has full control over the order of elaboration, and it is up
25846 to the programmer to elaborate declarations in an appropriate order. For
25849 @smallexample @c ada
25852 function One return Float;
25856 function One return Float is
25865 will obviously raise @code{Program_Error} at run time, because function
25866 One will be called before its body is elaborated. In this case GNAT will
25867 generate a warning that the call will raise @code{Program_Error}:
25873 2. function One return Float;
25875 4. Q : Float := One;
25877 >>> warning: cannot call "One" before body is elaborated
25878 >>> warning: Program_Error will be raised at run time
25881 6. function One return Float is
25894 Note that in this particular case, it is likely that the call is safe, because
25895 the function @code{One} does not access any global variables.
25896 Nevertheless in Ada, we do not want the validity of the check to depend on
25897 the contents of the body (think about the separate compilation case), so this
25898 is still wrong, as we discussed in the previous sections.
25900 The error is easily corrected by rearranging the declarations so that the
25901 body of @code{One} appears before the declaration containing the call
25902 (note that in Ada 95 and Ada 2005,
25903 declarations can appear in any order, so there is no restriction that
25904 would prevent this reordering, and if we write:
25906 @smallexample @c ada
25909 function One return Float;
25911 function One return Float is
25922 then all is well, no warning is generated, and no
25923 @code{Program_Error} exception
25925 Things are more complicated when a chain of subprograms is executed:
25927 @smallexample @c ada
25930 function A return Integer;
25931 function B return Integer;
25932 function C return Integer;
25934 function B return Integer is begin return A; end;
25935 function C return Integer is begin return B; end;
25939 function A return Integer is begin return 1; end;
25945 Now the call to @code{C}
25946 at elaboration time in the declaration of @code{X} is correct, because
25947 the body of @code{C} is already elaborated,
25948 and the call to @code{B} within the body of
25949 @code{C} is correct, but the call
25950 to @code{A} within the body of @code{B} is incorrect, because the body
25951 of @code{A} has not been elaborated, so @code{Program_Error}
25952 will be raised on the call to @code{A}.
25953 In this case GNAT will generate a
25954 warning that @code{Program_Error} may be
25955 raised at the point of the call. Let's look at the warning:
25961 2. function A return Integer;
25962 3. function B return Integer;
25963 4. function C return Integer;
25965 6. function B return Integer is begin return A; end;
25967 >>> warning: call to "A" before body is elaborated may
25968 raise Program_Error
25969 >>> warning: "B" called at line 7
25970 >>> warning: "C" called at line 9
25972 7. function C return Integer is begin return B; end;
25974 9. X : Integer := C;
25976 11. function A return Integer is begin return 1; end;
25986 Note that the message here says ``may raise'', instead of the direct case,
25987 where the message says ``will be raised''. That's because whether
25989 actually called depends in general on run-time flow of control.
25990 For example, if the body of @code{B} said
25992 @smallexample @c ada
25995 function B return Integer is
25997 if some-condition-depending-on-input-data then
26008 then we could not know until run time whether the incorrect call to A would
26009 actually occur, so @code{Program_Error} might
26010 or might not be raised. It is possible for a compiler to
26011 do a better job of analyzing bodies, to
26012 determine whether or not @code{Program_Error}
26013 might be raised, but it certainly
26014 couldn't do a perfect job (that would require solving the halting problem
26015 and is provably impossible), and because this is a warning anyway, it does
26016 not seem worth the effort to do the analysis. Cases in which it
26017 would be relevant are rare.
26019 In practice, warnings of either of the forms given
26020 above will usually correspond to
26021 real errors, and should be examined carefully and eliminated.
26022 In the rare case where a warning is bogus, it can be suppressed by any of
26023 the following methods:
26027 Compile with the @option{-gnatws} switch set
26030 Suppress @code{Elaboration_Check} for the called subprogram
26033 Use pragma @code{Warnings_Off} to turn warnings off for the call
26037 For the internal elaboration check case,
26038 GNAT by default generates the
26039 necessary run-time checks to ensure
26040 that @code{Program_Error} is raised if any
26041 call fails an elaboration check. Of course this can only happen if a
26042 warning has been issued as described above. The use of pragma
26043 @code{Suppress (Elaboration_Check)} may (but is not guaranteed to) suppress
26044 some of these checks, meaning that it may be possible (but is not
26045 guaranteed) for a program to be able to call a subprogram whose body
26046 is not yet elaborated, without raising a @code{Program_Error} exception.
26048 @node Controlling Elaboration in GNAT - External Calls
26049 @section Controlling Elaboration in GNAT - External Calls
26052 The previous section discussed the case in which the execution of a
26053 particular thread of elaboration code occurred entirely within a
26054 single unit. This is the easy case to handle, because a programmer
26055 has direct and total control over the order of elaboration, and
26056 furthermore, checks need only be generated in cases which are rare
26057 and which the compiler can easily detect.
26058 The situation is more complex when separate compilation is taken into account.
26059 Consider the following:
26061 @smallexample @c ada
26065 function Sqrt (Arg : Float) return Float;
26068 package body Math is
26069 function Sqrt (Arg : Float) return Float is
26078 X : Float := Math.Sqrt (0.5);
26091 where @code{Main} is the main program. When this program is executed, the
26092 elaboration code must first be executed, and one of the jobs of the
26093 binder is to determine the order in which the units of a program are
26094 to be elaborated. In this case we have four units: the spec and body
26096 the spec of @code{Stuff} and the body of @code{Main}).
26097 In what order should the four separate sections of elaboration code
26100 There are some restrictions in the order of elaboration that the binder
26101 can choose. In particular, if unit U has a @code{with}
26102 for a package @code{X}, then you
26103 are assured that the spec of @code{X}
26104 is elaborated before U , but you are
26105 not assured that the body of @code{X}
26106 is elaborated before U.
26107 This means that in the above case, the binder is allowed to choose the
26118 but that's not good, because now the call to @code{Math.Sqrt}
26119 that happens during
26120 the elaboration of the @code{Stuff}
26121 spec happens before the body of @code{Math.Sqrt} is
26122 elaborated, and hence causes @code{Program_Error} exception to be raised.
26123 At first glance, one might say that the binder is misbehaving, because
26124 obviously you want to elaborate the body of something you @code{with}
26126 that is not a general rule that can be followed in all cases. Consider
26128 @smallexample @c ada
26136 package body Y is ...
26139 package body X is ...
26145 This is a common arrangement, and, apart from the order of elaboration
26146 problems that might arise in connection with elaboration code, this works fine.
26147 A rule that says that you must first elaborate the body of anything you
26148 @code{with} cannot work in this case:
26149 the body of @code{X} @code{with}'s @code{Y},
26150 which means you would have to
26151 elaborate the body of @code{Y} first, but that @code{with}'s @code{X},
26153 you have to elaborate the body of @code{X} first, but ... and we have a
26154 loop that cannot be broken.
26156 It is true that the binder can in many cases guess an order of elaboration
26157 that is unlikely to cause a @code{Program_Error}
26158 exception to be raised, and it tries to do so (in the
26159 above example of @code{Math/Stuff/Spec}, the GNAT binder will
26161 elaborate the body of @code{Math} right after its spec, so all will be well).
26163 However, a program that blindly relies on the binder to be helpful can
26164 get into trouble, as we discussed in the previous sections, so
26166 provides a number of facilities for assisting the programmer in
26167 developing programs that are robust with respect to elaboration order.
26169 @node Default Behavior in GNAT - Ensuring Safety
26170 @section Default Behavior in GNAT - Ensuring Safety
26173 The default behavior in GNAT ensures elaboration safety. In its
26174 default mode GNAT implements the
26175 rule we previously described as the right approach. Let's restate it:
26179 @emph{If a unit has elaboration code that can directly or indirectly make a
26180 call to a subprogram in a @code{with}'ed unit, or instantiate a generic
26181 package in a @code{with}'ed unit, then if the @code{with}'ed unit
26182 does not have pragma @code{Pure} or
26183 @code{Preelaborate}, then the client should have an
26184 @code{Elaborate_All} pragma for the @code{with}'ed unit.}
26186 @emph{In the case of instantiating a generic subprogram, it is always
26187 sufficient to have only an @code{Elaborate} pragma for the
26188 @code{with}'ed unit.}
26192 By following this rule a client is assured that calls and instantiations
26193 can be made without risk of an exception.
26195 In this mode GNAT traces all calls that are potentially made from
26196 elaboration code, and puts in any missing implicit @code{Elaborate}
26197 and @code{Elaborate_All} pragmas.
26198 The advantage of this approach is that no elaboration problems
26199 are possible if the binder can find an elaboration order that is
26200 consistent with these implicit @code{Elaborate} and
26201 @code{Elaborate_All} pragmas. The
26202 disadvantage of this approach is that no such order may exist.
26204 If the binder does not generate any diagnostics, then it means that it has
26205 found an elaboration order that is guaranteed to be safe. However, the binder
26206 may still be relying on implicitly generated @code{Elaborate} and
26207 @code{Elaborate_All} pragmas so portability to other compilers than GNAT is not
26210 If it is important to guarantee portability, then the compilations should
26213 (warn on elaboration problems) switch. This will cause warning messages
26214 to be generated indicating the missing @code{Elaborate} and
26215 @code{Elaborate_All} pragmas.
26216 Consider the following source program:
26218 @smallexample @c ada
26223 m : integer := k.r;
26230 where it is clear that there
26231 should be a pragma @code{Elaborate_All}
26232 for unit @code{k}. An implicit pragma will be generated, and it is
26233 likely that the binder will be able to honor it. However, if you want
26234 to port this program to some other Ada compiler than GNAT.
26235 it is safer to include the pragma explicitly in the source. If this
26236 unit is compiled with the
26238 switch, then the compiler outputs a warning:
26245 3. m : integer := k.r;
26247 >>> warning: call to "r" may raise Program_Error
26248 >>> warning: missing pragma Elaborate_All for "k"
26256 and these warnings can be used as a guide for supplying manually
26257 the missing pragmas. It is usually a bad idea to use this warning
26258 option during development. That's because it will warn you when
26259 you need to put in a pragma, but cannot warn you when it is time
26260 to take it out. So the use of pragma @code{Elaborate_All} may lead to
26261 unnecessary dependencies and even false circularities.
26263 This default mode is more restrictive than the Ada Reference
26264 Manual, and it is possible to construct programs which will compile
26265 using the dynamic model described there, but will run into a
26266 circularity using the safer static model we have described.
26268 Of course any Ada compiler must be able to operate in a mode
26269 consistent with the requirements of the Ada Reference Manual,
26270 and in particular must have the capability of implementing the
26271 standard dynamic model of elaboration with run-time checks.
26273 In GNAT, this standard mode can be achieved either by the use of
26274 the @option{-gnatE} switch on the compiler (@command{gcc} or
26275 @command{gnatmake}) command, or by the use of the configuration pragma:
26277 @smallexample @c ada
26278 pragma Elaboration_Checks (RM);
26282 Either approach will cause the unit affected to be compiled using the
26283 standard dynamic run-time elaboration checks described in the Ada
26284 Reference Manual. The static model is generally preferable, since it
26285 is clearly safer to rely on compile and link time checks rather than
26286 run-time checks. However, in the case of legacy code, it may be
26287 difficult to meet the requirements of the static model. This
26288 issue is further discussed in
26289 @ref{What to Do If the Default Elaboration Behavior Fails}.
26291 Note that the static model provides a strict subset of the allowed
26292 behavior and programs of the Ada Reference Manual, so if you do
26293 adhere to the static model and no circularities exist,
26294 then you are assured that your program will
26295 work using the dynamic model, providing that you remove any
26296 pragma Elaborate statements from the source.
26298 @node Treatment of Pragma Elaborate
26299 @section Treatment of Pragma Elaborate
26300 @cindex Pragma Elaborate
26303 The use of @code{pragma Elaborate}
26304 should generally be avoided in Ada 95 and Ada 2005 programs,
26305 since there is no guarantee that transitive calls
26306 will be properly handled. Indeed at one point, this pragma was placed
26307 in Annex J (Obsolescent Features), on the grounds that it is never useful.
26309 Now that's a bit restrictive. In practice, the case in which
26310 @code{pragma Elaborate} is useful is when the caller knows that there
26311 are no transitive calls, or that the called unit contains all necessary
26312 transitive @code{pragma Elaborate} statements, and legacy code often
26313 contains such uses.
26315 Strictly speaking the static mode in GNAT should ignore such pragmas,
26316 since there is no assurance at compile time that the necessary safety
26317 conditions are met. In practice, this would cause GNAT to be incompatible
26318 with correctly written Ada 83 code that had all necessary
26319 @code{pragma Elaborate} statements in place. Consequently, we made the
26320 decision that GNAT in its default mode will believe that if it encounters
26321 a @code{pragma Elaborate} then the programmer knows what they are doing,
26322 and it will trust that no elaboration errors can occur.
26324 The result of this decision is two-fold. First to be safe using the
26325 static mode, you should remove all @code{pragma Elaborate} statements.
26326 Second, when fixing circularities in existing code, you can selectively
26327 use @code{pragma Elaborate} statements to convince the static mode of
26328 GNAT that it need not generate an implicit @code{pragma Elaborate_All}
26331 When using the static mode with @option{-gnatwl}, any use of
26332 @code{pragma Elaborate} will generate a warning about possible
26335 @node Elaboration Issues for Library Tasks
26336 @section Elaboration Issues for Library Tasks
26337 @cindex Library tasks, elaboration issues
26338 @cindex Elaboration of library tasks
26341 In this section we examine special elaboration issues that arise for
26342 programs that declare library level tasks.
26344 Generally the model of execution of an Ada program is that all units are
26345 elaborated, and then execution of the program starts. However, the
26346 declaration of library tasks definitely does not fit this model. The
26347 reason for this is that library tasks start as soon as they are declared
26348 (more precisely, as soon as the statement part of the enclosing package
26349 body is reached), that is to say before elaboration
26350 of the program is complete. This means that if such a task calls a
26351 subprogram, or an entry in another task, the callee may or may not be
26352 elaborated yet, and in the standard
26353 Reference Manual model of dynamic elaboration checks, you can even
26354 get timing dependent Program_Error exceptions, since there can be
26355 a race between the elaboration code and the task code.
26357 The static model of elaboration in GNAT seeks to avoid all such
26358 dynamic behavior, by being conservative, and the conservative
26359 approach in this particular case is to assume that all the code
26360 in a task body is potentially executed at elaboration time if
26361 a task is declared at the library level.
26363 This can definitely result in unexpected circularities. Consider
26364 the following example
26366 @smallexample @c ada
26372 type My_Int is new Integer;
26374 function Ident (M : My_Int) return My_Int;
26378 package body Decls is
26379 task body Lib_Task is
26385 function Ident (M : My_Int) return My_Int is
26393 procedure Put_Val (Arg : Decls.My_Int);
26397 package body Utils is
26398 procedure Put_Val (Arg : Decls.My_Int) is
26400 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
26407 Decls.Lib_Task.Start;
26412 If the above example is compiled in the default static elaboration
26413 mode, then a circularity occurs. The circularity comes from the call
26414 @code{Utils.Put_Val} in the task body of @code{Decls.Lib_Task}. Since
26415 this call occurs in elaboration code, we need an implicit pragma
26416 @code{Elaborate_All} for @code{Utils}. This means that not only must
26417 the spec and body of @code{Utils} be elaborated before the body
26418 of @code{Decls}, but also the spec and body of any unit that is
26419 @code{with'ed} by the body of @code{Utils} must also be elaborated before
26420 the body of @code{Decls}. This is the transitive implication of
26421 pragma @code{Elaborate_All} and it makes sense, because in general
26422 the body of @code{Put_Val} might have a call to something in a
26423 @code{with'ed} unit.
26425 In this case, the body of Utils (actually its spec) @code{with's}
26426 @code{Decls}. Unfortunately this means that the body of @code{Decls}
26427 must be elaborated before itself, in case there is a call from the
26428 body of @code{Utils}.
26430 Here is the exact chain of events we are worrying about:
26434 In the body of @code{Decls} a call is made from within the body of a library
26435 task to a subprogram in the package @code{Utils}. Since this call may
26436 occur at elaboration time (given that the task is activated at elaboration
26437 time), we have to assume the worst, i.e. that the
26438 call does happen at elaboration time.
26441 This means that the body and spec of @code{Util} must be elaborated before
26442 the body of @code{Decls} so that this call does not cause an access before
26446 Within the body of @code{Util}, specifically within the body of
26447 @code{Util.Put_Val} there may be calls to any unit @code{with}'ed
26451 One such @code{with}'ed package is package @code{Decls}, so there
26452 might be a call to a subprogram in @code{Decls} in @code{Put_Val}.
26453 In fact there is such a call in this example, but we would have to
26454 assume that there was such a call even if it were not there, since
26455 we are not supposed to write the body of @code{Decls} knowing what
26456 is in the body of @code{Utils}; certainly in the case of the
26457 static elaboration model, the compiler does not know what is in
26458 other bodies and must assume the worst.
26461 This means that the spec and body of @code{Decls} must also be
26462 elaborated before we elaborate the unit containing the call, but
26463 that unit is @code{Decls}! This means that the body of @code{Decls}
26464 must be elaborated before itself, and that's a circularity.
26468 Indeed, if you add an explicit pragma @code{Elaborate_All} for @code{Utils} in
26469 the body of @code{Decls} you will get a true Ada Reference Manual
26470 circularity that makes the program illegal.
26472 In practice, we have found that problems with the static model of
26473 elaboration in existing code often arise from library tasks, so
26474 we must address this particular situation.
26476 Note that if we compile and run the program above, using the dynamic model of
26477 elaboration (that is to say use the @option{-gnatE} switch),
26478 then it compiles, binds,
26479 links, and runs, printing the expected result of 2. Therefore in some sense
26480 the circularity here is only apparent, and we need to capture
26481 the properties of this program that distinguish it from other library-level
26482 tasks that have real elaboration problems.
26484 We have four possible answers to this question:
26489 Use the dynamic model of elaboration.
26491 If we use the @option{-gnatE} switch, then as noted above, the program works.
26492 Why is this? If we examine the task body, it is apparent that the task cannot
26494 @code{accept} statement until after elaboration has been completed, because
26495 the corresponding entry call comes from the main program, not earlier.
26496 This is why the dynamic model works here. But that's really giving
26497 up on a precise analysis, and we prefer to take this approach only if we cannot
26499 problem in any other manner. So let us examine two ways to reorganize
26500 the program to avoid the potential elaboration problem.
26503 Split library tasks into separate packages.
26505 Write separate packages, so that library tasks are isolated from
26506 other declarations as much as possible. Let us look at a variation on
26509 @smallexample @c ada
26517 package body Decls1 is
26518 task body Lib_Task is
26526 type My_Int is new Integer;
26527 function Ident (M : My_Int) return My_Int;
26531 package body Decls2 is
26532 function Ident (M : My_Int) return My_Int is
26540 procedure Put_Val (Arg : Decls2.My_Int);
26544 package body Utils is
26545 procedure Put_Val (Arg : Decls2.My_Int) is
26547 Text_IO.Put_Line (Decls2.My_Int'Image (Decls2.Ident (Arg)));
26554 Decls1.Lib_Task.Start;
26559 All we have done is to split @code{Decls} into two packages, one
26560 containing the library task, and one containing everything else. Now
26561 there is no cycle, and the program compiles, binds, links and executes
26562 using the default static model of elaboration.
26565 Declare separate task types.
26567 A significant part of the problem arises because of the use of the
26568 single task declaration form. This means that the elaboration of
26569 the task type, and the elaboration of the task itself (i.e. the
26570 creation of the task) happen at the same time. A good rule
26571 of style in Ada is to always create explicit task types. By
26572 following the additional step of placing task objects in separate
26573 packages from the task type declaration, many elaboration problems
26574 are avoided. Here is another modified example of the example program:
26576 @smallexample @c ada
26578 task type Lib_Task_Type is
26582 type My_Int is new Integer;
26584 function Ident (M : My_Int) return My_Int;
26588 package body Decls is
26589 task body Lib_Task_Type is
26595 function Ident (M : My_Int) return My_Int is
26603 procedure Put_Val (Arg : Decls.My_Int);
26607 package body Utils is
26608 procedure Put_Val (Arg : Decls.My_Int) is
26610 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
26616 Lib_Task : Decls.Lib_Task_Type;
26622 Declst.Lib_Task.Start;
26627 What we have done here is to replace the @code{task} declaration in
26628 package @code{Decls} with a @code{task type} declaration. Then we
26629 introduce a separate package @code{Declst} to contain the actual
26630 task object. This separates the elaboration issues for
26631 the @code{task type}
26632 declaration, which causes no trouble, from the elaboration issues
26633 of the task object, which is also unproblematic, since it is now independent
26634 of the elaboration of @code{Utils}.
26635 This separation of concerns also corresponds to
26636 a generally sound engineering principle of separating declarations
26637 from instances. This version of the program also compiles, binds, links,
26638 and executes, generating the expected output.
26641 Use No_Entry_Calls_In_Elaboration_Code restriction.
26642 @cindex No_Entry_Calls_In_Elaboration_Code
26644 The previous two approaches described how a program can be restructured
26645 to avoid the special problems caused by library task bodies. in practice,
26646 however, such restructuring may be difficult to apply to existing legacy code,
26647 so we must consider solutions that do not require massive rewriting.
26649 Let us consider more carefully why our original sample program works
26650 under the dynamic model of elaboration. The reason is that the code
26651 in the task body blocks immediately on the @code{accept}
26652 statement. Now of course there is nothing to prohibit elaboration
26653 code from making entry calls (for example from another library level task),
26654 so we cannot tell in isolation that
26655 the task will not execute the accept statement during elaboration.
26657 However, in practice it is very unusual to see elaboration code
26658 make any entry calls, and the pattern of tasks starting
26659 at elaboration time and then immediately blocking on @code{accept} or
26660 @code{select} statements is very common. What this means is that
26661 the compiler is being too pessimistic when it analyzes the
26662 whole package body as though it might be executed at elaboration
26665 If we know that the elaboration code contains no entry calls, (a very safe
26666 assumption most of the time, that could almost be made the default
26667 behavior), then we can compile all units of the program under control
26668 of the following configuration pragma:
26671 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
26675 This pragma can be placed in the @file{gnat.adc} file in the usual
26676 manner. If we take our original unmodified program and compile it
26677 in the presence of a @file{gnat.adc} containing the above pragma,
26678 then once again, we can compile, bind, link, and execute, obtaining
26679 the expected result. In the presence of this pragma, the compiler does
26680 not trace calls in a task body, that appear after the first @code{accept}
26681 or @code{select} statement, and therefore does not report a potential
26682 circularity in the original program.
26684 The compiler will check to the extent it can that the above
26685 restriction is not violated, but it is not always possible to do a
26686 complete check at compile time, so it is important to use this
26687 pragma only if the stated restriction is in fact met, that is to say
26688 no task receives an entry call before elaboration of all units is completed.
26692 @node Mixing Elaboration Models
26693 @section Mixing Elaboration Models
26695 So far, we have assumed that the entire program is either compiled
26696 using the dynamic model or static model, ensuring consistency. It
26697 is possible to mix the two models, but rules have to be followed
26698 if this mixing is done to ensure that elaboration checks are not
26701 The basic rule is that @emph{a unit compiled with the static model cannot
26702 be @code{with'ed} by a unit compiled with the dynamic model}. The
26703 reason for this is that in the static model, a unit assumes that
26704 its clients guarantee to use (the equivalent of) pragma
26705 @code{Elaborate_All} so that no elaboration checks are required
26706 in inner subprograms, and this assumption is violated if the
26707 client is compiled with dynamic checks.
26709 The precise rule is as follows. A unit that is compiled with dynamic
26710 checks can only @code{with} a unit that meets at least one of the
26711 following criteria:
26716 The @code{with'ed} unit is itself compiled with dynamic elaboration
26717 checks (that is with the @option{-gnatE} switch.
26720 The @code{with'ed} unit is an internal GNAT implementation unit from
26721 the System, Interfaces, Ada, or GNAT hierarchies.
26724 The @code{with'ed} unit has pragma Preelaborate or pragma Pure.
26727 The @code{with'ing} unit (that is the client) has an explicit pragma
26728 @code{Elaborate_All} for the @code{with'ed} unit.
26733 If this rule is violated, that is if a unit with dynamic elaboration
26734 checks @code{with's} a unit that does not meet one of the above four
26735 criteria, then the binder (@code{gnatbind}) will issue a warning
26736 similar to that in the following example:
26739 warning: "x.ads" has dynamic elaboration checks and with's
26740 warning: "y.ads" which has static elaboration checks
26744 These warnings indicate that the rule has been violated, and that as a result
26745 elaboration checks may be missed in the resulting executable file.
26746 This warning may be suppressed using the @option{-ws} binder switch
26747 in the usual manner.
26749 One useful application of this mixing rule is in the case of a subsystem
26750 which does not itself @code{with} units from the remainder of the
26751 application. In this case, the entire subsystem can be compiled with
26752 dynamic checks to resolve a circularity in the subsystem, while
26753 allowing the main application that uses this subsystem to be compiled
26754 using the more reliable default static model.
26756 @node What to Do If the Default Elaboration Behavior Fails
26757 @section What to Do If the Default Elaboration Behavior Fails
26760 If the binder cannot find an acceptable order, it outputs detailed
26761 diagnostics. For example:
26767 error: elaboration circularity detected
26768 info: "proc (body)" must be elaborated before "pack (body)"
26769 info: reason: Elaborate_All probably needed in unit "pack (body)"
26770 info: recompile "pack (body)" with -gnatwl
26771 info: for full details
26772 info: "proc (body)"
26773 info: is needed by its spec:
26774 info: "proc (spec)"
26775 info: which is withed by:
26776 info: "pack (body)"
26777 info: "pack (body)" must be elaborated before "proc (body)"
26778 info: reason: pragma Elaborate in unit "proc (body)"
26784 In this case we have a cycle that the binder cannot break. On the one
26785 hand, there is an explicit pragma Elaborate in @code{proc} for
26786 @code{pack}. This means that the body of @code{pack} must be elaborated
26787 before the body of @code{proc}. On the other hand, there is elaboration
26788 code in @code{pack} that calls a subprogram in @code{proc}. This means
26789 that for maximum safety, there should really be a pragma
26790 Elaborate_All in @code{pack} for @code{proc} which would require that
26791 the body of @code{proc} be elaborated before the body of
26792 @code{pack}. Clearly both requirements cannot be satisfied.
26793 Faced with a circularity of this kind, you have three different options.
26796 @item Fix the program
26797 The most desirable option from the point of view of long-term maintenance
26798 is to rearrange the program so that the elaboration problems are avoided.
26799 One useful technique is to place the elaboration code into separate
26800 child packages. Another is to move some of the initialization code to
26801 explicitly called subprograms, where the program controls the order
26802 of initialization explicitly. Although this is the most desirable option,
26803 it may be impractical and involve too much modification, especially in
26804 the case of complex legacy code.
26806 @item Perform dynamic checks
26807 If the compilations are done using the
26809 (dynamic elaboration check) switch, then GNAT behaves in a quite different
26810 manner. Dynamic checks are generated for all calls that could possibly result
26811 in raising an exception. With this switch, the compiler does not generate
26812 implicit @code{Elaborate} or @code{Elaborate_All} pragmas. The behavior then is
26813 exactly as specified in the @cite{Ada Reference Manual}.
26814 The binder will generate
26815 an executable program that may or may not raise @code{Program_Error}, and then
26816 it is the programmer's job to ensure that it does not raise an exception. Note
26817 that it is important to compile all units with the switch, it cannot be used
26820 @item Suppress checks
26821 The drawback of dynamic checks is that they generate a
26822 significant overhead at run time, both in space and time. If you
26823 are absolutely sure that your program cannot raise any elaboration
26824 exceptions, and you still want to use the dynamic elaboration model,
26825 then you can use the configuration pragma
26826 @code{Suppress (Elaboration_Check)} to suppress all such checks. For
26827 example this pragma could be placed in the @file{gnat.adc} file.
26829 @item Suppress checks selectively
26830 When you know that certain calls or instantiations in elaboration code cannot
26831 possibly lead to an elaboration error, and the binder nevertheless complains
26832 about implicit @code{Elaborate} and @code{Elaborate_All} pragmas that lead to
26833 elaboration circularities, it is possible to remove those warnings locally and
26834 obtain a program that will bind. Clearly this can be unsafe, and it is the
26835 responsibility of the programmer to make sure that the resulting program has no
26836 elaboration anomalies. The pragma @code{Suppress (Elaboration_Check)} can be
26837 used with different granularity to suppress warnings and break elaboration
26842 Place the pragma that names the called subprogram in the declarative part
26843 that contains the call.
26846 Place the pragma in the declarative part, without naming an entity. This
26847 disables warnings on all calls in the corresponding declarative region.
26850 Place the pragma in the package spec that declares the called subprogram,
26851 and name the subprogram. This disables warnings on all elaboration calls to
26855 Place the pragma in the package spec that declares the called subprogram,
26856 without naming any entity. This disables warnings on all elaboration calls to
26857 all subprograms declared in this spec.
26859 @item Use Pragma Elaborate
26860 As previously described in section @xref{Treatment of Pragma Elaborate},
26861 GNAT in static mode assumes that a @code{pragma} Elaborate indicates correctly
26862 that no elaboration checks are required on calls to the designated unit.
26863 There may be cases in which the caller knows that no transitive calls
26864 can occur, so that a @code{pragma Elaborate} will be sufficient in a
26865 case where @code{pragma Elaborate_All} would cause a circularity.
26869 These five cases are listed in order of decreasing safety, and therefore
26870 require increasing programmer care in their application. Consider the
26873 @smallexample @c adanocomment
26875 function F1 return Integer;
26880 function F2 return Integer;
26881 function Pure (x : integer) return integer;
26882 -- pragma Suppress (Elaboration_Check, On => Pure); -- (3)
26883 -- pragma Suppress (Elaboration_Check); -- (4)
26887 package body Pack1 is
26888 function F1 return Integer is
26892 Val : integer := Pack2.Pure (11); -- Elab. call (1)
26895 -- pragma Suppress(Elaboration_Check, Pack2.F2); -- (1)
26896 -- pragma Suppress(Elaboration_Check); -- (2)
26898 X1 := Pack2.F2 + 1; -- Elab. call (2)
26903 package body Pack2 is
26904 function F2 return Integer is
26908 function Pure (x : integer) return integer is
26910 return x ** 3 - 3 * x;
26914 with Pack1, Ada.Text_IO;
26917 Ada.Text_IO.Put_Line(Pack1.X1'Img); -- 101
26920 In the absence of any pragmas, an attempt to bind this program produces
26921 the following diagnostics:
26927 error: elaboration circularity detected
26928 info: "pack1 (body)" must be elaborated before "pack1 (body)"
26929 info: reason: Elaborate_All probably needed in unit "pack1 (body)"
26930 info: recompile "pack1 (body)" with -gnatwl for full details
26931 info: "pack1 (body)"
26932 info: must be elaborated along with its spec:
26933 info: "pack1 (spec)"
26934 info: which is withed by:
26935 info: "pack2 (body)"
26936 info: which must be elaborated along with its spec:
26937 info: "pack2 (spec)"
26938 info: which is withed by:
26939 info: "pack1 (body)"
26942 The sources of the circularity are the two calls to @code{Pack2.Pure} and
26943 @code{Pack2.F2} in the body of @code{Pack1}. We can see that the call to
26944 F2 is safe, even though F2 calls F1, because the call appears after the
26945 elaboration of the body of F1. Therefore the pragma (1) is safe, and will
26946 remove the warning on the call. It is also possible to use pragma (2)
26947 because there are no other potentially unsafe calls in the block.
26950 The call to @code{Pure} is safe because this function does not depend on the
26951 state of @code{Pack2}. Therefore any call to this function is safe, and it
26952 is correct to place pragma (3) in the corresponding package spec.
26955 Finally, we could place pragma (4) in the spec of @code{Pack2} to disable
26956 warnings on all calls to functions declared therein. Note that this is not
26957 necessarily safe, and requires more detailed examination of the subprogram
26958 bodies involved. In particular, a call to @code{F2} requires that @code{F1}
26959 be already elaborated.
26963 It is hard to generalize on which of these four approaches should be
26964 taken. Obviously if it is possible to fix the program so that the default
26965 treatment works, this is preferable, but this may not always be practical.
26966 It is certainly simple enough to use
26968 but the danger in this case is that, even if the GNAT binder
26969 finds a correct elaboration order, it may not always do so,
26970 and certainly a binder from another Ada compiler might not. A
26971 combination of testing and analysis (for which the warnings generated
26974 switch can be useful) must be used to ensure that the program is free
26975 of errors. One switch that is useful in this testing is the
26976 @option{^-p (pessimistic elaboration order)^/PESSIMISTIC_ELABORATION_ORDER^}
26979 Normally the binder tries to find an order that has the best chance of
26980 of avoiding elaboration problems. With this switch, the binder
26981 plays a devil's advocate role, and tries to choose the order that
26982 has the best chance of failing. If your program works even with this
26983 switch, then it has a better chance of being error free, but this is still
26986 For an example of this approach in action, consider the C-tests (executable
26987 tests) from the ACVC suite. If these are compiled and run with the default
26988 treatment, then all but one of them succeed without generating any error
26989 diagnostics from the binder. However, there is one test that fails, and
26990 this is not surprising, because the whole point of this test is to ensure
26991 that the compiler can handle cases where it is impossible to determine
26992 a correct order statically, and it checks that an exception is indeed
26993 raised at run time.
26995 This one test must be compiled and run using the
26997 switch, and then it passes. Alternatively, the entire suite can
26998 be run using this switch. It is never wrong to run with the dynamic
26999 elaboration switch if your code is correct, and we assume that the
27000 C-tests are indeed correct (it is less efficient, but efficiency is
27001 not a factor in running the ACVC tests.)
27003 @node Elaboration for Access-to-Subprogram Values
27004 @section Elaboration for Access-to-Subprogram Values
27005 @cindex Access-to-subprogram
27008 Access-to-subprogram types (introduced in Ada 95) complicate
27009 the handling of elaboration. The trouble is that it becomes
27010 impossible to tell at compile time which procedure
27011 is being called. This means that it is not possible for the binder
27012 to analyze the elaboration requirements in this case.
27014 If at the point at which the access value is created
27015 (i.e., the evaluation of @code{P'Access} for a subprogram @code{P}),
27016 the body of the subprogram is
27017 known to have been elaborated, then the access value is safe, and its use
27018 does not require a check. This may be achieved by appropriate arrangement
27019 of the order of declarations if the subprogram is in the current unit,
27020 or, if the subprogram is in another unit, by using pragma
27021 @code{Pure}, @code{Preelaborate}, or @code{Elaborate_Body}
27022 on the referenced unit.
27024 If the referenced body is not known to have been elaborated at the point
27025 the access value is created, then any use of the access value must do a
27026 dynamic check, and this dynamic check will fail and raise a
27027 @code{Program_Error} exception if the body has not been elaborated yet.
27028 GNAT will generate the necessary checks, and in addition, if the
27030 switch is set, will generate warnings that such checks are required.
27032 The use of dynamic dispatching for tagged types similarly generates
27033 a requirement for dynamic checks, and premature calls to any primitive
27034 operation of a tagged type before the body of the operation has been
27035 elaborated, will result in the raising of @code{Program_Error}.
27037 @node Summary of Procedures for Elaboration Control
27038 @section Summary of Procedures for Elaboration Control
27039 @cindex Elaboration control
27042 First, compile your program with the default options, using none of
27043 the special elaboration control switches. If the binder successfully
27044 binds your program, then you can be confident that, apart from issues
27045 raised by the use of access-to-subprogram types and dynamic dispatching,
27046 the program is free of elaboration errors. If it is important that the
27047 program be portable, then use the
27049 switch to generate warnings about missing @code{Elaborate} or
27050 @code{Elaborate_All} pragmas, and supply the missing pragmas.
27052 If the program fails to bind using the default static elaboration
27053 handling, then you can fix the program to eliminate the binder
27054 message, or recompile the entire program with the
27055 @option{-gnatE} switch to generate dynamic elaboration checks,
27056 and, if you are sure there really are no elaboration problems,
27057 use a global pragma @code{Suppress (Elaboration_Check)}.
27059 @node Other Elaboration Order Considerations
27060 @section Other Elaboration Order Considerations
27062 This section has been entirely concerned with the issue of finding a valid
27063 elaboration order, as defined by the Ada Reference Manual. In a case
27064 where several elaboration orders are valid, the task is to find one
27065 of the possible valid elaboration orders (and the static model in GNAT
27066 will ensure that this is achieved).
27068 The purpose of the elaboration rules in the Ada Reference Manual is to
27069 make sure that no entity is accessed before it has been elaborated. For
27070 a subprogram, this means that the spec and body must have been elaborated
27071 before the subprogram is called. For an object, this means that the object
27072 must have been elaborated before its value is read or written. A violation
27073 of either of these two requirements is an access before elaboration order,
27074 and this section has been all about avoiding such errors.
27076 In the case where more than one order of elaboration is possible, in the
27077 sense that access before elaboration errors are avoided, then any one of
27078 the orders is ``correct'' in the sense that it meets the requirements of
27079 the Ada Reference Manual, and no such error occurs.
27081 However, it may be the case for a given program, that there are
27082 constraints on the order of elaboration that come not from consideration
27083 of avoiding elaboration errors, but rather from extra-lingual logic
27084 requirements. Consider this example:
27086 @smallexample @c ada
27087 with Init_Constants;
27088 package Constants is
27093 package Init_Constants is
27094 procedure P; -- require a body
27095 end Init_Constants;
27098 package body Init_Constants is
27099 procedure P is begin null; end;
27103 end Init_Constants;
27107 Z : Integer := Constants.X + Constants.Y;
27111 with Text_IO; use Text_IO;
27114 Put_Line (Calc.Z'Img);
27119 In this example, there is more than one valid order of elaboration. For
27120 example both the following are correct orders:
27123 Init_Constants spec
27126 Init_Constants body
27131 Init_Constants spec
27132 Init_Constants body
27139 There is no language rule to prefer one or the other, both are correct
27140 from an order of elaboration point of view. But the programmatic effects
27141 of the two orders are very different. In the first, the elaboration routine
27142 of @code{Calc} initializes @code{Z} to zero, and then the main program
27143 runs with this value of zero. But in the second order, the elaboration
27144 routine of @code{Calc} runs after the body of Init_Constants has set
27145 @code{X} and @code{Y} and thus @code{Z} is set to 7 before @code{Main}
27148 One could perhaps by applying pretty clever non-artificial intelligence
27149 to the situation guess that it is more likely that the second order of
27150 elaboration is the one desired, but there is no formal linguistic reason
27151 to prefer one over the other. In fact in this particular case, GNAT will
27152 prefer the second order, because of the rule that bodies are elaborated
27153 as soon as possible, but it's just luck that this is what was wanted
27154 (if indeed the second order was preferred).
27156 If the program cares about the order of elaboration routines in a case like
27157 this, it is important to specify the order required. In this particular
27158 case, that could have been achieved by adding to the spec of Calc:
27160 @smallexample @c ada
27161 pragma Elaborate_All (Constants);
27165 which requires that the body (if any) and spec of @code{Constants},
27166 as well as the body and spec of any unit @code{with}'ed by
27167 @code{Constants} be elaborated before @code{Calc} is elaborated.
27169 Clearly no automatic method can always guess which alternative you require,
27170 and if you are working with legacy code that had constraints of this kind
27171 which were not properly specified by adding @code{Elaborate} or
27172 @code{Elaborate_All} pragmas, then indeed it is possible that two different
27173 compilers can choose different orders.
27175 However, GNAT does attempt to diagnose the common situation where there
27176 are uninitialized variables in the visible part of a package spec, and the
27177 corresponding package body has an elaboration block that directly or
27178 indirectly initialized one or more of these variables. This is the situation
27179 in which a pragma Elaborate_Body is usually desirable, and GNAT will generate
27180 a warning that suggests this addition if it detects this situation.
27182 The @code{gnatbind}
27183 @option{^-p^/PESSIMISTIC_ELABORATION^} switch may be useful in smoking
27184 out problems. This switch causes bodies to be elaborated as late as possible
27185 instead of as early as possible. In the example above, it would have forced
27186 the choice of the first elaboration order. If you get different results
27187 when using this switch, and particularly if one set of results is right,
27188 and one is wrong as far as you are concerned, it shows that you have some
27189 missing @code{Elaborate} pragmas. For the example above, we have the
27193 gnatmake -f -q main
27196 gnatmake -f -q main -bargs -p
27202 It is of course quite unlikely that both these results are correct, so
27203 it is up to you in a case like this to investigate the source of the
27204 difference, by looking at the two elaboration orders that are chosen,
27205 and figuring out which is correct, and then adding the necessary
27206 @code{Elaborate} or @code{Elaborate_All} pragmas to ensure the desired order.
27208 @node Inline Assembler
27209 @appendix Inline Assembler
27212 If you need to write low-level software that interacts directly
27213 with the hardware, Ada provides two ways to incorporate assembly
27214 language code into your program. First, you can import and invoke
27215 external routines written in assembly language, an Ada feature fully
27216 supported by GNAT@. However, for small sections of code it may be simpler
27217 or more efficient to include assembly language statements directly
27218 in your Ada source program, using the facilities of the implementation-defined
27219 package @code{System.Machine_Code}, which incorporates the gcc
27220 Inline Assembler. The Inline Assembler approach offers a number of advantages,
27221 including the following:
27224 @item No need to use non-Ada tools
27225 @item Consistent interface over different targets
27226 @item Automatic usage of the proper calling conventions
27227 @item Access to Ada constants and variables
27228 @item Definition of intrinsic routines
27229 @item Possibility of inlining a subprogram comprising assembler code
27230 @item Code optimizer can take Inline Assembler code into account
27233 This chapter presents a series of examples to show you how to use
27234 the Inline Assembler. Although it focuses on the Intel x86,
27235 the general approach applies also to other processors.
27236 It is assumed that you are familiar with Ada
27237 and with assembly language programming.
27240 * Basic Assembler Syntax::
27241 * A Simple Example of Inline Assembler::
27242 * Output Variables in Inline Assembler::
27243 * Input Variables in Inline Assembler::
27244 * Inlining Inline Assembler Code::
27245 * Other Asm Functionality::
27248 @c ---------------------------------------------------------------------------
27249 @node Basic Assembler Syntax
27250 @section Basic Assembler Syntax
27253 The assembler used by GNAT and gcc is based not on the Intel assembly
27254 language, but rather on a language that descends from the AT&T Unix
27255 assembler @emph{as} (and which is often referred to as ``AT&T syntax'').
27256 The following table summarizes the main features of @emph{as} syntax
27257 and points out the differences from the Intel conventions.
27258 See the gcc @emph{as} and @emph{gas} (an @emph{as} macro
27259 pre-processor) documentation for further information.
27262 @item Register names
27263 gcc / @emph{as}: Prefix with ``%''; for example @code{%eax}
27265 Intel: No extra punctuation; for example @code{eax}
27267 @item Immediate operand
27268 gcc / @emph{as}: Prefix with ``$''; for example @code{$4}
27270 Intel: No extra punctuation; for example @code{4}
27273 gcc / @emph{as}: Prefix with ``$''; for example @code{$loc}
27275 Intel: No extra punctuation; for example @code{loc}
27277 @item Memory contents
27278 gcc / @emph{as}: No extra punctuation; for example @code{loc}
27280 Intel: Square brackets; for example @code{[loc]}
27282 @item Register contents
27283 gcc / @emph{as}: Parentheses; for example @code{(%eax)}
27285 Intel: Square brackets; for example @code{[eax]}
27287 @item Hexadecimal numbers
27288 gcc / @emph{as}: Leading ``0x'' (C language syntax); for example @code{0xA0}
27290 Intel: Trailing ``h''; for example @code{A0h}
27293 gcc / @emph{as}: Explicit in op code; for example @code{movw} to move
27296 Intel: Implicit, deduced by assembler; for example @code{mov}
27298 @item Instruction repetition
27299 gcc / @emph{as}: Split into two lines; for example
27305 Intel: Keep on one line; for example @code{rep stosl}
27307 @item Order of operands
27308 gcc / @emph{as}: Source first; for example @code{movw $4, %eax}
27310 Intel: Destination first; for example @code{mov eax, 4}
27313 @c ---------------------------------------------------------------------------
27314 @node A Simple Example of Inline Assembler
27315 @section A Simple Example of Inline Assembler
27318 The following example will generate a single assembly language statement,
27319 @code{nop}, which does nothing. Despite its lack of run-time effect,
27320 the example will be useful in illustrating the basics of
27321 the Inline Assembler facility.
27323 @smallexample @c ada
27325 with System.Machine_Code; use System.Machine_Code;
27326 procedure Nothing is
27333 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
27334 here it takes one parameter, a @emph{template string} that must be a static
27335 expression and that will form the generated instruction.
27336 @code{Asm} may be regarded as a compile-time procedure that parses
27337 the template string and additional parameters (none here),
27338 from which it generates a sequence of assembly language instructions.
27340 The examples in this chapter will illustrate several of the forms
27341 for invoking @code{Asm}; a complete specification of the syntax
27342 is found in the @cite{GNAT Reference Manual}.
27344 Under the standard GNAT conventions, the @code{Nothing} procedure
27345 should be in a file named @file{nothing.adb}.
27346 You can build the executable in the usual way:
27350 However, the interesting aspect of this example is not its run-time behavior
27351 but rather the generated assembly code.
27352 To see this output, invoke the compiler as follows:
27354 gcc -c -S -fomit-frame-pointer -gnatp @file{nothing.adb}
27356 where the options are:
27360 compile only (no bind or link)
27362 generate assembler listing
27363 @item -fomit-frame-pointer
27364 do not set up separate stack frames
27366 do not add runtime checks
27369 This gives a human-readable assembler version of the code. The resulting
27370 file will have the same name as the Ada source file, but with a @code{.s}
27371 extension. In our example, the file @file{nothing.s} has the following
27376 .file "nothing.adb"
27378 ___gnu_compiled_ada:
27381 .globl __ada_nothing
27393 The assembly code you included is clearly indicated by
27394 the compiler, between the @code{#APP} and @code{#NO_APP}
27395 delimiters. The character before the 'APP' and 'NOAPP'
27396 can differ on different targets. For example, GNU/Linux uses '#APP' while
27397 on NT you will see '/APP'.
27399 If you make a mistake in your assembler code (such as using the
27400 wrong size modifier, or using a wrong operand for the instruction) GNAT
27401 will report this error in a temporary file, which will be deleted when
27402 the compilation is finished. Generating an assembler file will help
27403 in such cases, since you can assemble this file separately using the
27404 @emph{as} assembler that comes with gcc.
27406 Assembling the file using the command
27409 as @file{nothing.s}
27412 will give you error messages whose lines correspond to the assembler
27413 input file, so you can easily find and correct any mistakes you made.
27414 If there are no errors, @emph{as} will generate an object file
27415 @file{nothing.out}.
27417 @c ---------------------------------------------------------------------------
27418 @node Output Variables in Inline Assembler
27419 @section Output Variables in Inline Assembler
27422 The examples in this section, showing how to access the processor flags,
27423 illustrate how to specify the destination operands for assembly language
27426 @smallexample @c ada
27428 with Interfaces; use Interfaces;
27429 with Ada.Text_IO; use Ada.Text_IO;
27430 with System.Machine_Code; use System.Machine_Code;
27431 procedure Get_Flags is
27432 Flags : Unsigned_32;
27435 Asm ("pushfl" & LF & HT & -- push flags on stack
27436 "popl %%eax" & LF & HT & -- load eax with flags
27437 "movl %%eax, %0", -- store flags in variable
27438 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
27439 Put_Line ("Flags register:" & Flags'Img);
27444 In order to have a nicely aligned assembly listing, we have separated
27445 multiple assembler statements in the Asm template string with linefeed
27446 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
27447 The resulting section of the assembly output file is:
27454 movl %eax, -40(%ebp)
27459 It would have been legal to write the Asm invocation as:
27462 Asm ("pushfl popl %%eax movl %%eax, %0")
27465 but in the generated assembler file, this would come out as:
27469 pushfl popl %eax movl %eax, -40(%ebp)
27473 which is not so convenient for the human reader.
27475 We use Ada comments
27476 at the end of each line to explain what the assembler instructions
27477 actually do. This is a useful convention.
27479 When writing Inline Assembler instructions, you need to precede each register
27480 and variable name with a percent sign. Since the assembler already requires
27481 a percent sign at the beginning of a register name, you need two consecutive
27482 percent signs for such names in the Asm template string, thus @code{%%eax}.
27483 In the generated assembly code, one of the percent signs will be stripped off.
27485 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
27486 variables: operands you later define using @code{Input} or @code{Output}
27487 parameters to @code{Asm}.
27488 An output variable is illustrated in
27489 the third statement in the Asm template string:
27493 The intent is to store the contents of the eax register in a variable that can
27494 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
27495 necessarily work, since the compiler might optimize by using a register
27496 to hold Flags, and the expansion of the @code{movl} instruction would not be
27497 aware of this optimization. The solution is not to store the result directly
27498 but rather to advise the compiler to choose the correct operand form;
27499 that is the purpose of the @code{%0} output variable.
27501 Information about the output variable is supplied in the @code{Outputs}
27502 parameter to @code{Asm}:
27504 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
27507 The output is defined by the @code{Asm_Output} attribute of the target type;
27508 the general format is
27510 Type'Asm_Output (constraint_string, variable_name)
27513 The constraint string directs the compiler how
27514 to store/access the associated variable. In the example
27516 Unsigned_32'Asm_Output ("=m", Flags);
27518 the @code{"m"} (memory) constraint tells the compiler that the variable
27519 @code{Flags} should be stored in a memory variable, thus preventing
27520 the optimizer from keeping it in a register. In contrast,
27522 Unsigned_32'Asm_Output ("=r", Flags);
27524 uses the @code{"r"} (register) constraint, telling the compiler to
27525 store the variable in a register.
27527 If the constraint is preceded by the equal character (@strong{=}), it tells
27528 the compiler that the variable will be used to store data into it.
27530 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
27531 allowing the optimizer to choose whatever it deems best.
27533 There are a fairly large number of constraints, but the ones that are
27534 most useful (for the Intel x86 processor) are the following:
27540 global (i.e. can be stored anywhere)
27558 use one of eax, ebx, ecx or edx
27560 use one of eax, ebx, ecx, edx, esi or edi
27563 The full set of constraints is described in the gcc and @emph{as}
27564 documentation; note that it is possible to combine certain constraints
27565 in one constraint string.
27567 You specify the association of an output variable with an assembler operand
27568 through the @code{%}@emph{n} notation, where @emph{n} is a non-negative
27570 @smallexample @c ada
27572 Asm ("pushfl" & LF & HT & -- push flags on stack
27573 "popl %%eax" & LF & HT & -- load eax with flags
27574 "movl %%eax, %0", -- store flags in variable
27575 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
27579 @code{%0} will be replaced in the expanded code by the appropriate operand,
27581 the compiler decided for the @code{Flags} variable.
27583 In general, you may have any number of output variables:
27586 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
27588 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
27589 of @code{Asm_Output} attributes
27593 @smallexample @c ada
27595 Asm ("movl %%eax, %0" & LF & HT &
27596 "movl %%ebx, %1" & LF & HT &
27598 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
27599 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
27600 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
27604 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
27605 in the Ada program.
27607 As a variation on the @code{Get_Flags} example, we can use the constraints
27608 string to direct the compiler to store the eax register into the @code{Flags}
27609 variable, instead of including the store instruction explicitly in the
27610 @code{Asm} template string:
27612 @smallexample @c ada
27614 with Interfaces; use Interfaces;
27615 with Ada.Text_IO; use Ada.Text_IO;
27616 with System.Machine_Code; use System.Machine_Code;
27617 procedure Get_Flags_2 is
27618 Flags : Unsigned_32;
27621 Asm ("pushfl" & LF & HT & -- push flags on stack
27622 "popl %%eax", -- save flags in eax
27623 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
27624 Put_Line ("Flags register:" & Flags'Img);
27630 The @code{"a"} constraint tells the compiler that the @code{Flags}
27631 variable will come from the eax register. Here is the resulting code:
27639 movl %eax,-40(%ebp)
27644 The compiler generated the store of eax into Flags after
27645 expanding the assembler code.
27647 Actually, there was no need to pop the flags into the eax register;
27648 more simply, we could just pop the flags directly into the program variable:
27650 @smallexample @c ada
27652 with Interfaces; use Interfaces;
27653 with Ada.Text_IO; use Ada.Text_IO;
27654 with System.Machine_Code; use System.Machine_Code;
27655 procedure Get_Flags_3 is
27656 Flags : Unsigned_32;
27659 Asm ("pushfl" & LF & HT & -- push flags on stack
27660 "pop %0", -- save flags in Flags
27661 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
27662 Put_Line ("Flags register:" & Flags'Img);
27667 @c ---------------------------------------------------------------------------
27668 @node Input Variables in Inline Assembler
27669 @section Input Variables in Inline Assembler
27672 The example in this section illustrates how to specify the source operands
27673 for assembly language statements.
27674 The program simply increments its input value by 1:
27676 @smallexample @c ada
27678 with Interfaces; use Interfaces;
27679 with Ada.Text_IO; use Ada.Text_IO;
27680 with System.Machine_Code; use System.Machine_Code;
27681 procedure Increment is
27683 function Incr (Value : Unsigned_32) return Unsigned_32 is
27684 Result : Unsigned_32;
27687 Inputs => Unsigned_32'Asm_Input ("a", Value),
27688 Outputs => Unsigned_32'Asm_Output ("=a", Result));
27692 Value : Unsigned_32;
27696 Put_Line ("Value before is" & Value'Img);
27697 Value := Incr (Value);
27698 Put_Line ("Value after is" & Value'Img);
27703 The @code{Outputs} parameter to @code{Asm} specifies
27704 that the result will be in the eax register and that it is to be stored
27705 in the @code{Result} variable.
27707 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
27708 but with an @code{Asm_Input} attribute.
27709 The @code{"="} constraint, indicating an output value, is not present.
27711 You can have multiple input variables, in the same way that you can have more
27712 than one output variable.
27714 The parameter count (%0, %1) etc, now starts at the first input
27715 statement, and continues with the output statements.
27716 When both parameters use the same variable, the
27717 compiler will treat them as the same %n operand, which is the case here.
27719 Just as the @code{Outputs} parameter causes the register to be stored into the
27720 target variable after execution of the assembler statements, so does the
27721 @code{Inputs} parameter cause its variable to be loaded into the register
27722 before execution of the assembler statements.
27724 Thus the effect of the @code{Asm} invocation is:
27726 @item load the 32-bit value of @code{Value} into eax
27727 @item execute the @code{incl %eax} instruction
27728 @item store the contents of eax into the @code{Result} variable
27731 The resulting assembler file (with @option{-O2} optimization) contains:
27734 _increment__incr.1:
27747 @c ---------------------------------------------------------------------------
27748 @node Inlining Inline Assembler Code
27749 @section Inlining Inline Assembler Code
27752 For a short subprogram such as the @code{Incr} function in the previous
27753 section, the overhead of the call and return (creating / deleting the stack
27754 frame) can be significant, compared to the amount of code in the subprogram
27755 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
27756 which directs the compiler to expand invocations of the subprogram at the
27757 point(s) of call, instead of setting up a stack frame for out-of-line calls.
27758 Here is the resulting program:
27760 @smallexample @c ada
27762 with Interfaces; use Interfaces;
27763 with Ada.Text_IO; use Ada.Text_IO;
27764 with System.Machine_Code; use System.Machine_Code;
27765 procedure Increment_2 is
27767 function Incr (Value : Unsigned_32) return Unsigned_32 is
27768 Result : Unsigned_32;
27771 Inputs => Unsigned_32'Asm_Input ("a", Value),
27772 Outputs => Unsigned_32'Asm_Output ("=a", Result));
27775 pragma Inline (Increment);
27777 Value : Unsigned_32;
27781 Put_Line ("Value before is" & Value'Img);
27782 Value := Increment (Value);
27783 Put_Line ("Value after is" & Value'Img);
27788 Compile the program with both optimization (@option{-O2}) and inlining
27789 enabled (@option{-gnatpn} instead of @option{-gnatp}).
27791 The @code{Incr} function is still compiled as usual, but at the
27792 point in @code{Increment} where our function used to be called:
27797 call _increment__incr.1
27802 the code for the function body directly appears:
27815 thus saving the overhead of stack frame setup and an out-of-line call.
27817 @c ---------------------------------------------------------------------------
27818 @node Other Asm Functionality
27819 @section Other @code{Asm} Functionality
27822 This section describes two important parameters to the @code{Asm}
27823 procedure: @code{Clobber}, which identifies register usage;
27824 and @code{Volatile}, which inhibits unwanted optimizations.
27827 * The Clobber Parameter::
27828 * The Volatile Parameter::
27831 @c ---------------------------------------------------------------------------
27832 @node The Clobber Parameter
27833 @subsection The @code{Clobber} Parameter
27836 One of the dangers of intermixing assembly language and a compiled language
27837 such as Ada is that the compiler needs to be aware of which registers are
27838 being used by the assembly code. In some cases, such as the earlier examples,
27839 the constraint string is sufficient to indicate register usage (e.g.,
27841 the eax register). But more generally, the compiler needs an explicit
27842 identification of the registers that are used by the Inline Assembly
27845 Using a register that the compiler doesn't know about
27846 could be a side effect of an instruction (like @code{mull}
27847 storing its result in both eax and edx).
27848 It can also arise from explicit register usage in your
27849 assembly code; for example:
27852 Asm ("movl %0, %%ebx" & LF & HT &
27854 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
27855 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out));
27859 where the compiler (since it does not analyze the @code{Asm} template string)
27860 does not know you are using the ebx register.
27862 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
27863 to identify the registers that will be used by your assembly code:
27867 Asm ("movl %0, %%ebx" & LF & HT &
27869 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
27870 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
27875 The Clobber parameter is a static string expression specifying the
27876 register(s) you are using. Note that register names are @emph{not} prefixed
27877 by a percent sign. Also, if more than one register is used then their names
27878 are separated by commas; e.g., @code{"eax, ebx"}
27880 The @code{Clobber} parameter has several additional uses:
27882 @item Use ``register'' name @code{cc} to indicate that flags might have changed
27883 @item Use ``register'' name @code{memory} if you changed a memory location
27886 @c ---------------------------------------------------------------------------
27887 @node The Volatile Parameter
27888 @subsection The @code{Volatile} Parameter
27889 @cindex Volatile parameter
27892 Compiler optimizations in the presence of Inline Assembler may sometimes have
27893 unwanted effects. For example, when an @code{Asm} invocation with an input
27894 variable is inside a loop, the compiler might move the loading of the input
27895 variable outside the loop, regarding it as a one-time initialization.
27897 If this effect is not desired, you can disable such optimizations by setting
27898 the @code{Volatile} parameter to @code{True}; for example:
27900 @smallexample @c ada
27902 Asm ("movl %0, %%ebx" & LF & HT &
27904 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
27905 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
27911 By default, @code{Volatile} is set to @code{False} unless there is no
27912 @code{Outputs} parameter.
27914 Although setting @code{Volatile} to @code{True} prevents unwanted
27915 optimizations, it will also disable other optimizations that might be
27916 important for efficiency. In general, you should set @code{Volatile}
27917 to @code{True} only if the compiler's optimizations have created
27919 @c END OF INLINE ASSEMBLER CHAPTER
27920 @c ===============================
27922 @c ***********************************
27923 @c * Compatibility and Porting Guide *
27924 @c ***********************************
27925 @node Compatibility and Porting Guide
27926 @appendix Compatibility and Porting Guide
27929 This chapter describes the compatibility issues that may arise between
27930 GNAT and other Ada compilation systems (including those for Ada 83),
27931 and shows how GNAT can expedite porting
27932 applications developed in other Ada environments.
27935 * Compatibility with Ada 83::
27936 * Compatibility between Ada 95 and Ada 2005::
27937 * Implementation-dependent characteristics::
27938 * Compatibility with Other Ada Systems::
27939 * Representation Clauses::
27941 @c Brief section is only in non-VMS version
27942 @c Full chapter is in VMS version
27943 * Compatibility with HP Ada 83::
27946 * Transitioning to 64-Bit GNAT for OpenVMS::
27950 @node Compatibility with Ada 83
27951 @section Compatibility with Ada 83
27952 @cindex Compatibility (between Ada 83 and Ada 95 / Ada 2005)
27955 Ada 95 and Ada 2005 are highly upwards compatible with Ada 83. In
27956 particular, the design intention was that the difficulties associated
27957 with moving from Ada 83 to Ada 95 or Ada 2005 should be no greater than those
27958 that occur when moving from one Ada 83 system to another.
27960 However, there are a number of points at which there are minor
27961 incompatibilities. The @cite{Ada 95 Annotated Reference Manual} contains
27962 full details of these issues,
27963 and should be consulted for a complete treatment.
27965 following subsections treat the most likely issues to be encountered.
27968 * Legal Ada 83 programs that are illegal in Ada 95::
27969 * More deterministic semantics::
27970 * Changed semantics::
27971 * Other language compatibility issues::
27974 @node Legal Ada 83 programs that are illegal in Ada 95
27975 @subsection Legal Ada 83 programs that are illegal in Ada 95
27977 Some legal Ada 83 programs are illegal (i.e. they will fail to compile) in
27978 Ada 95 and thus also in Ada 2005:
27981 @item Character literals
27982 Some uses of character literals are ambiguous. Since Ada 95 has introduced
27983 @code{Wide_Character} as a new predefined character type, some uses of
27984 character literals that were legal in Ada 83 are illegal in Ada 95.
27986 @smallexample @c ada
27987 for Char in 'A' .. 'Z' loop ... end loop;
27991 The problem is that @code{'A'} and @code{'Z'} could be from either
27992 @code{Character} or @code{Wide_Character}. The simplest correction
27993 is to make the type explicit; e.g.:
27994 @smallexample @c ada
27995 for Char in Character range 'A' .. 'Z' loop ... end loop;
27998 @item New reserved words
27999 The identifiers @code{abstract}, @code{aliased}, @code{protected},
28000 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
28001 Existing Ada 83 code using any of these identifiers must be edited to
28002 use some alternative name.
28004 @item Freezing rules
28005 The rules in Ada 95 are slightly different with regard to the point at
28006 which entities are frozen, and representation pragmas and clauses are
28007 not permitted past the freeze point. This shows up most typically in
28008 the form of an error message complaining that a representation item
28009 appears too late, and the appropriate corrective action is to move
28010 the item nearer to the declaration of the entity to which it refers.
28012 A particular case is that representation pragmas
28015 extended HP Ada 83 compatibility pragmas such as @code{Export_Procedure})
28017 cannot be applied to a subprogram body. If necessary, a separate subprogram
28018 declaration must be introduced to which the pragma can be applied.
28020 @item Optional bodies for library packages
28021 In Ada 83, a package that did not require a package body was nevertheless
28022 allowed to have one. This lead to certain surprises in compiling large
28023 systems (situations in which the body could be unexpectedly ignored by the
28024 binder). In Ada 95, if a package does not require a body then it is not
28025 permitted to have a body. To fix this problem, simply remove a redundant
28026 body if it is empty, or, if it is non-empty, introduce a dummy declaration
28027 into the spec that makes the body required. One approach is to add a private
28028 part to the package declaration (if necessary), and define a parameterless
28029 procedure called @code{Requires_Body}, which must then be given a dummy
28030 procedure body in the package body, which then becomes required.
28031 Another approach (assuming that this does not introduce elaboration
28032 circularities) is to add an @code{Elaborate_Body} pragma to the package spec,
28033 since one effect of this pragma is to require the presence of a package body.
28035 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
28036 In Ada 95, the exception @code{Numeric_Error} is a renaming of
28037 @code{Constraint_Error}.
28038 This means that it is illegal to have separate exception handlers for
28039 the two exceptions. The fix is simply to remove the handler for the
28040 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
28041 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
28043 @item Indefinite subtypes in generics
28044 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String})
28045 as the actual for a generic formal private type, but then the instantiation
28046 would be illegal if there were any instances of declarations of variables
28047 of this type in the generic body. In Ada 95, to avoid this clear violation
28048 of the methodological principle known as the ``contract model'',
28049 the generic declaration explicitly indicates whether
28050 or not such instantiations are permitted. If a generic formal parameter
28051 has explicit unknown discriminants, indicated by using @code{(<>)} after the
28052 type name, then it can be instantiated with indefinite types, but no
28053 stand-alone variables can be declared of this type. Any attempt to declare
28054 such a variable will result in an illegality at the time the generic is
28055 declared. If the @code{(<>)} notation is not used, then it is illegal
28056 to instantiate the generic with an indefinite type.
28057 This is the potential incompatibility issue when porting Ada 83 code to Ada 95.
28058 It will show up as a compile time error, and
28059 the fix is usually simply to add the @code{(<>)} to the generic declaration.
28062 @node More deterministic semantics
28063 @subsection More deterministic semantics
28067 Conversions from real types to integer types round away from 0. In Ada 83
28068 the conversion Integer(2.5) could deliver either 2 or 3 as its value. This
28069 implementation freedom was intended to support unbiased rounding in
28070 statistical applications, but in practice it interfered with portability.
28071 In Ada 95 the conversion semantics are unambiguous, and rounding away from 0
28072 is required. Numeric code may be affected by this change in semantics.
28073 Note, though, that this issue is no worse than already existed in Ada 83
28074 when porting code from one vendor to another.
28077 The Real-Time Annex introduces a set of policies that define the behavior of
28078 features that were implementation dependent in Ada 83, such as the order in
28079 which open select branches are executed.
28082 @node Changed semantics
28083 @subsection Changed semantics
28086 The worst kind of incompatibility is one where a program that is legal in
28087 Ada 83 is also legal in Ada 95 but can have an effect in Ada 95 that was not
28088 possible in Ada 83. Fortunately this is extremely rare, but the one
28089 situation that you should be alert to is the change in the predefined type
28090 @code{Character} from 7-bit ASCII to 8-bit Latin-1.
28093 @item Range of type @code{Character}
28094 The range of @code{Standard.Character} is now the full 256 characters
28095 of Latin-1, whereas in most Ada 83 implementations it was restricted
28096 to 128 characters. Although some of the effects of
28097 this change will be manifest in compile-time rejection of legal
28098 Ada 83 programs it is possible for a working Ada 83 program to have
28099 a different effect in Ada 95, one that was not permitted in Ada 83.
28100 As an example, the expression
28101 @code{Character'Pos(Character'Last)} returned @code{127} in Ada 83 and now
28102 delivers @code{255} as its value.
28103 In general, you should look at the logic of any
28104 character-processing Ada 83 program and see whether it needs to be adapted
28105 to work correctly with Latin-1. Note that the predefined Ada 95 API has a
28106 character handling package that may be relevant if code needs to be adapted
28107 to account for the additional Latin-1 elements.
28108 The desirable fix is to
28109 modify the program to accommodate the full character set, but in some cases
28110 it may be convenient to define a subtype or derived type of Character that
28111 covers only the restricted range.
28115 @node Other language compatibility issues
28116 @subsection Other language compatibility issues
28119 @item @option{-gnat83} switch
28120 All implementations of GNAT provide a switch that causes GNAT to operate
28121 in Ada 83 mode. In this mode, some but not all compatibility problems
28122 of the type described above are handled automatically. For example, the
28123 new reserved words introduced in Ada 95 and Ada 2005 are treated simply
28124 as identifiers as in Ada 83.
28126 in practice, it is usually advisable to make the necessary modifications
28127 to the program to remove the need for using this switch.
28128 See @ref{Compiling Different Versions of Ada}.
28130 @item Support for removed Ada 83 pragmas and attributes
28131 A number of pragmas and attributes from Ada 83 were removed from Ada 95,
28132 generally because they were replaced by other mechanisms. Ada 95 and Ada 2005
28133 compilers are allowed, but not required, to implement these missing
28134 elements. In contrast with some other compilers, GNAT implements all
28135 such pragmas and attributes, eliminating this compatibility concern. These
28136 include @code{pragma Interface} and the floating point type attributes
28137 (@code{Emax}, @code{Mantissa}, etc.), among other items.
28141 @node Compatibility between Ada 95 and Ada 2005
28142 @section Compatibility between Ada 95 and Ada 2005
28143 @cindex Compatibility between Ada 95 and Ada 2005
28146 Although Ada 2005 was designed to be upwards compatible with Ada 95, there are
28147 a number of incompatibilities. Several are enumerated below;
28148 for a complete description please see the
28149 Annotated Ada 2005 Reference Manual, or section 9.1.1 in
28150 @cite{Rationale for Ada 2005}.
28153 @item New reserved words.
28154 The words @code{interface}, @code{overriding} and @code{synchronized} are
28155 reserved in Ada 2005.
28156 A pre-Ada 2005 program that uses any of these as an identifier will be
28159 @item New declarations in predefined packages.
28160 A number of packages in the predefined environment contain new declarations:
28161 @code{Ada.Exceptions}, @code{Ada.Real_Time}, @code{Ada.Strings},
28162 @code{Ada.Strings.Fixed}, @code{Ada.Strings.Bounded},
28163 @code{Ada.Strings.Unbounded}, @code{Ada.Strings.Wide_Fixed},
28164 @code{Ada.Strings.Wide_Bounded}, @code{Ada.Strings.Wide_Unbounded},
28165 @code{Ada.Tags}, @code{Ada.Text_IO}, and @code{Interfaces.C}.
28166 If an Ada 95 program does a @code{with} and @code{use} of any of these
28167 packages, the new declarations may cause name clashes.
28169 @item Access parameters.
28170 A nondispatching subprogram with an access parameter cannot be renamed
28171 as a dispatching operation. This was permitted in Ada 95.
28173 @item Access types, discriminants, and constraints.
28174 Rule changes in this area have led to some incompatibilities; for example,
28175 constrained subtypes of some access types are not permitted in Ada 2005.
28177 @item Aggregates for limited types.
28178 The allowance of aggregates for limited types in Ada 2005 raises the
28179 possibility of ambiguities in legal Ada 95 programs, since additional types
28180 now need to be considered in expression resolution.
28182 @item Fixed-point multiplication and division.
28183 Certain expressions involving ``*'' or ``/'' for a fixed-point type, which
28184 were legal in Ada 95 and invoked the predefined versions of these operations,
28186 The ambiguity may be resolved either by applying a type conversion to the
28187 expression, or by explicitly invoking the operation from package
28190 @item Return-by-reference types.
28191 The Ada 95 return-by-reference mechanism has been removed. Instead, the user
28192 can declare a function returning a value from an anonymous access type.
28196 @node Implementation-dependent characteristics
28197 @section Implementation-dependent characteristics
28199 Although the Ada language defines the semantics of each construct as
28200 precisely as practical, in some situations (for example for reasons of
28201 efficiency, or where the effect is heavily dependent on the host or target
28202 platform) the implementation is allowed some freedom. In porting Ada 83
28203 code to GNAT, you need to be aware of whether / how the existing code
28204 exercised such implementation dependencies. Such characteristics fall into
28205 several categories, and GNAT offers specific support in assisting the
28206 transition from certain Ada 83 compilers.
28209 * Implementation-defined pragmas::
28210 * Implementation-defined attributes::
28212 * Elaboration order::
28213 * Target-specific aspects::
28216 @node Implementation-defined pragmas
28217 @subsection Implementation-defined pragmas
28220 Ada compilers are allowed to supplement the language-defined pragmas, and
28221 these are a potential source of non-portability. All GNAT-defined pragmas
28222 are described in the GNAT Reference Manual, and these include several that
28223 are specifically intended to correspond to other vendors' Ada 83 pragmas.
28224 For migrating from VADS, the pragma @code{Use_VADS_Size} may be useful.
28226 compatibility with HP Ada 83, GNAT supplies the pragmas
28227 @code{Extend_System}, @code{Ident}, @code{Inline_Generic},
28228 @code{Interface_Name}, @code{Passive}, @code{Suppress_All},
28229 and @code{Volatile}.
28230 Other relevant pragmas include @code{External} and @code{Link_With}.
28231 Some vendor-specific
28232 Ada 83 pragmas (@code{Share_Generic}, @code{Subtitle}, and @code{Title}) are
28234 avoiding compiler rejection of units that contain such pragmas; they are not
28235 relevant in a GNAT context and hence are not otherwise implemented.
28237 @node Implementation-defined attributes
28238 @subsection Implementation-defined attributes
28240 Analogous to pragmas, the set of attributes may be extended by an
28241 implementation. All GNAT-defined attributes are described in the
28242 @cite{GNAT Reference Manual}, and these include several that are specifically
28244 to correspond to other vendors' Ada 83 attributes. For migrating from VADS,
28245 the attribute @code{VADS_Size} may be useful. For compatibility with HP
28246 Ada 83, GNAT supplies the attributes @code{Bit}, @code{Machine_Size} and
28250 @subsection Libraries
28252 Vendors may supply libraries to supplement the standard Ada API. If Ada 83
28253 code uses vendor-specific libraries then there are several ways to manage
28254 this in Ada 95 or Ada 2005:
28257 If the source code for the libraries (specifications and bodies) are
28258 available, then the libraries can be migrated in the same way as the
28261 If the source code for the specifications but not the bodies are
28262 available, then you can reimplement the bodies.
28264 Some features introduced by Ada 95 obviate the need for library support. For
28265 example most Ada 83 vendors supplied a package for unsigned integers. The
28266 Ada 95 modular type feature is the preferred way to handle this need, so
28267 instead of migrating or reimplementing the unsigned integer package it may
28268 be preferable to retrofit the application using modular types.
28271 @node Elaboration order
28272 @subsection Elaboration order
28274 The implementation can choose any elaboration order consistent with the unit
28275 dependency relationship. This freedom means that some orders can result in
28276 Program_Error being raised due to an ``Access Before Elaboration'': an attempt
28277 to invoke a subprogram its body has been elaborated, or to instantiate a
28278 generic before the generic body has been elaborated. By default GNAT
28279 attempts to choose a safe order (one that will not encounter access before
28280 elaboration problems) by implicitly inserting @code{Elaborate} or
28281 @code{Elaborate_All} pragmas where
28282 needed. However, this can lead to the creation of elaboration circularities
28283 and a resulting rejection of the program by gnatbind. This issue is
28284 thoroughly described in @ref{Elaboration Order Handling in GNAT}.
28285 In brief, there are several
28286 ways to deal with this situation:
28290 Modify the program to eliminate the circularities, e.g. by moving
28291 elaboration-time code into explicitly-invoked procedures
28293 Constrain the elaboration order by including explicit @code{Elaborate_Body} or
28294 @code{Elaborate} pragmas, and then inhibit the generation of implicit
28295 @code{Elaborate_All}
28296 pragmas either globally (as an effect of the @option{-gnatE} switch) or locally
28297 (by selectively suppressing elaboration checks via pragma
28298 @code{Suppress(Elaboration_Check)} when it is safe to do so).
28301 @node Target-specific aspects
28302 @subsection Target-specific aspects
28304 Low-level applications need to deal with machine addresses, data
28305 representations, interfacing with assembler code, and similar issues. If
28306 such an Ada 83 application is being ported to different target hardware (for
28307 example where the byte endianness has changed) then you will need to
28308 carefully examine the program logic; the porting effort will heavily depend
28309 on the robustness of the original design. Moreover, Ada 95 (and thus
28310 Ada 2005) are sometimes
28311 incompatible with typical Ada 83 compiler practices regarding implicit
28312 packing, the meaning of the Size attribute, and the size of access values.
28313 GNAT's approach to these issues is described in @ref{Representation Clauses}.
28315 @node Compatibility with Other Ada Systems
28316 @section Compatibility with Other Ada Systems
28319 If programs avoid the use of implementation dependent and
28320 implementation defined features, as documented in the @cite{Ada
28321 Reference Manual}, there should be a high degree of portability between
28322 GNAT and other Ada systems. The following are specific items which
28323 have proved troublesome in moving Ada 95 programs from GNAT to other Ada 95
28324 compilers, but do not affect porting code to GNAT@.
28325 (As of @value{NOW}, GNAT is the only compiler available for Ada 2005;
28326 the following issues may or may not arise for Ada 2005 programs
28327 when other compilers appear.)
28330 @item Ada 83 Pragmas and Attributes
28331 Ada 95 compilers are allowed, but not required, to implement the missing
28332 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
28333 GNAT implements all such pragmas and attributes, eliminating this as
28334 a compatibility concern, but some other Ada 95 compilers reject these
28335 pragmas and attributes.
28337 @item Specialized Needs Annexes
28338 GNAT implements the full set of special needs annexes. At the
28339 current time, it is the only Ada 95 compiler to do so. This means that
28340 programs making use of these features may not be portable to other Ada
28341 95 compilation systems.
28343 @item Representation Clauses
28344 Some other Ada 95 compilers implement only the minimal set of
28345 representation clauses required by the Ada 95 reference manual. GNAT goes
28346 far beyond this minimal set, as described in the next section.
28349 @node Representation Clauses
28350 @section Representation Clauses
28353 The Ada 83 reference manual was quite vague in describing both the minimal
28354 required implementation of representation clauses, and also their precise
28355 effects. Ada 95 (and thus also Ada 2005) are much more explicit, but the
28356 minimal set of capabilities required is still quite limited.
28358 GNAT implements the full required set of capabilities in
28359 Ada 95 and Ada 2005, but also goes much further, and in particular
28360 an effort has been made to be compatible with existing Ada 83 usage to the
28361 greatest extent possible.
28363 A few cases exist in which Ada 83 compiler behavior is incompatible with
28364 the requirements in Ada 95 (and thus also Ada 2005). These are instances of
28365 intentional or accidental dependence on specific implementation dependent
28366 characteristics of these Ada 83 compilers. The following is a list of
28367 the cases most likely to arise in existing Ada 83 code.
28370 @item Implicit Packing
28371 Some Ada 83 compilers allowed a Size specification to cause implicit
28372 packing of an array or record. This could cause expensive implicit
28373 conversions for change of representation in the presence of derived
28374 types, and the Ada design intends to avoid this possibility.
28375 Subsequent AI's were issued to make it clear that such implicit
28376 change of representation in response to a Size clause is inadvisable,
28377 and this recommendation is represented explicitly in the Ada 95 (and Ada 2005)
28378 Reference Manuals as implementation advice that is followed by GNAT@.
28379 The problem will show up as an error
28380 message rejecting the size clause. The fix is simply to provide
28381 the explicit pragma @code{Pack}, or for more fine tuned control, provide
28382 a Component_Size clause.
28384 @item Meaning of Size Attribute
28385 The Size attribute in Ada 95 (and Ada 2005) for discrete types is defined as
28386 the minimal number of bits required to hold values of the type. For example,
28387 on a 32-bit machine, the size of @code{Natural} will typically be 31 and not
28388 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
28389 some 32 in this situation. This problem will usually show up as a compile
28390 time error, but not always. It is a good idea to check all uses of the
28391 'Size attribute when porting Ada 83 code. The GNAT specific attribute
28392 Object_Size can provide a useful way of duplicating the behavior of
28393 some Ada 83 compiler systems.
28395 @item Size of Access Types
28396 A common assumption in Ada 83 code is that an access type is in fact a pointer,
28397 and that therefore it will be the same size as a System.Address value. This
28398 assumption is true for GNAT in most cases with one exception. For the case of
28399 a pointer to an unconstrained array type (where the bounds may vary from one
28400 value of the access type to another), the default is to use a ``fat pointer'',
28401 which is represented as two separate pointers, one to the bounds, and one to
28402 the array. This representation has a number of advantages, including improved
28403 efficiency. However, it may cause some difficulties in porting existing Ada 83
28404 code which makes the assumption that, for example, pointers fit in 32 bits on
28405 a machine with 32-bit addressing.
28407 To get around this problem, GNAT also permits the use of ``thin pointers'' for
28408 access types in this case (where the designated type is an unconstrained array
28409 type). These thin pointers are indeed the same size as a System.Address value.
28410 To specify a thin pointer, use a size clause for the type, for example:
28412 @smallexample @c ada
28413 type X is access all String;
28414 for X'Size use Standard'Address_Size;
28418 which will cause the type X to be represented using a single pointer.
28419 When using this representation, the bounds are right behind the array.
28420 This representation is slightly less efficient, and does not allow quite
28421 such flexibility in the use of foreign pointers or in using the
28422 Unrestricted_Access attribute to create pointers to non-aliased objects.
28423 But for any standard portable use of the access type it will work in
28424 a functionally correct manner and allow porting of existing code.
28425 Note that another way of forcing a thin pointer representation
28426 is to use a component size clause for the element size in an array,
28427 or a record representation clause for an access field in a record.
28431 @c This brief section is only in the non-VMS version
28432 @c The complete chapter on HP Ada is in the VMS version
28433 @node Compatibility with HP Ada 83
28434 @section Compatibility with HP Ada 83
28437 The VMS version of GNAT fully implements all the pragmas and attributes
28438 provided by HP Ada 83, as well as providing the standard HP Ada 83
28439 libraries, including Starlet. In addition, data layouts and parameter
28440 passing conventions are highly compatible. This means that porting
28441 existing HP Ada 83 code to GNAT in VMS systems should be easier than
28442 most other porting efforts. The following are some of the most
28443 significant differences between GNAT and HP Ada 83.
28446 @item Default floating-point representation
28447 In GNAT, the default floating-point format is IEEE, whereas in HP Ada 83,
28448 it is VMS format. GNAT does implement the necessary pragmas
28449 (Long_Float, Float_Representation) for changing this default.
28452 The package System in GNAT exactly corresponds to the definition in the
28453 Ada 95 reference manual, which means that it excludes many of the
28454 HP Ada 83 extensions. However, a separate package Aux_DEC is provided
28455 that contains the additional definitions, and a special pragma,
28456 Extend_System allows this package to be treated transparently as an
28457 extension of package System.
28460 The definitions provided by Aux_DEC are exactly compatible with those
28461 in the HP Ada 83 version of System, with one exception.
28462 HP Ada provides the following declarations:
28464 @smallexample @c ada
28465 TO_ADDRESS (INTEGER)
28466 TO_ADDRESS (UNSIGNED_LONGWORD)
28467 TO_ADDRESS (@i{universal_integer})
28471 The version of TO_ADDRESS taking a @i{universal integer} argument is in fact
28472 an extension to Ada 83 not strictly compatible with the reference manual.
28473 In GNAT, we are constrained to be exactly compatible with the standard,
28474 and this means we cannot provide this capability. In HP Ada 83, the
28475 point of this definition is to deal with a call like:
28477 @smallexample @c ada
28478 TO_ADDRESS (16#12777#);
28482 Normally, according to the Ada 83 standard, one would expect this to be
28483 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
28484 of TO_ADDRESS@. However, in HP Ada 83, there is no ambiguity, since the
28485 definition using @i{universal_integer} takes precedence.
28487 In GNAT, since the version with @i{universal_integer} cannot be supplied, it
28488 is not possible to be 100% compatible. Since there are many programs using
28489 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
28490 to change the name of the function in the UNSIGNED_LONGWORD case, so the
28491 declarations provided in the GNAT version of AUX_Dec are:
28493 @smallexample @c ada
28494 function To_Address (X : Integer) return Address;
28495 pragma Pure_Function (To_Address);
28497 function To_Address_Long (X : Unsigned_Longword)
28499 pragma Pure_Function (To_Address_Long);
28503 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
28504 change the name to TO_ADDRESS_LONG@.
28506 @item Task_Id values
28507 The Task_Id values assigned will be different in the two systems, and GNAT
28508 does not provide a specified value for the Task_Id of the environment task,
28509 which in GNAT is treated like any other declared task.
28513 For full details on these and other less significant compatibility issues,
28514 see appendix E of the HP publication entitled @cite{HP Ada, Technical
28515 Overview and Comparison on HP Platforms}.
28517 For GNAT running on other than VMS systems, all the HP Ada 83 pragmas and
28518 attributes are recognized, although only a subset of them can sensibly
28519 be implemented. The description of pragmas in the
28520 @cite{GNAT Reference Manual}
28521 indicates whether or not they are applicable to non-VMS systems.
28525 @node Transitioning to 64-Bit GNAT for OpenVMS
28526 @section Transitioning to 64-Bit @value{EDITION} for OpenVMS
28529 This section is meant to assist users of pre-2006 @value{EDITION}
28530 for Alpha OpenVMS who are transitioning to 64-bit @value{EDITION},
28531 the version of the GNAT technology supplied in 2006 and later for
28532 OpenVMS on both Alpha and I64.
28535 * Introduction to transitioning::
28536 * Migration of 32 bit code::
28537 * Taking advantage of 64 bit addressing::
28538 * Technical details::
28541 @node Introduction to transitioning
28542 @subsection Introduction
28545 64-bit @value{EDITION} for Open VMS has been designed to meet
28550 Providing a full conforming implementation of Ada 95 and Ada 2005
28553 Allowing maximum backward compatibility, thus easing migration of existing
28557 Supplying a path for exploiting the full 64-bit address range
28561 Ada's strong typing semantics has made it
28562 impractical to have different 32-bit and 64-bit modes. As soon as
28563 one object could possibly be outside the 32-bit address space, this
28564 would make it necessary for the @code{System.Address} type to be 64 bits.
28565 In particular, this would cause inconsistencies if 32-bit code is
28566 called from 64-bit code that raises an exception.
28568 This issue has been resolved by always using 64-bit addressing
28569 at the system level, but allowing for automatic conversions between
28570 32-bit and 64-bit addresses where required. Thus users who
28571 do not currently require 64-bit addressing capabilities, can
28572 recompile their code with only minimal changes (and indeed
28573 if the code is written in portable Ada, with no assumptions about
28574 the size of the @code{Address} type, then no changes at all are necessary).
28576 this approach provides a simple, gradual upgrade path to future
28577 use of larger memories than available for 32-bit systems.
28578 Also, newly written applications or libraries will by default
28579 be fully compatible with future systems exploiting 64-bit
28580 addressing capabilities.
28582 @ref{Migration of 32 bit code}, will focus on porting applications
28583 that do not require more than 2 GB of
28584 addressable memory. This code will be referred to as
28585 @emph{32-bit code}.
28586 For applications intending to exploit the full 64-bit address space,
28587 @ref{Taking advantage of 64 bit addressing},
28588 will consider further changes that may be required.
28589 Such code will be referred to below as @emph{64-bit code}.
28591 @node Migration of 32 bit code
28592 @subsection Migration of 32-bit code
28597 * Unchecked conversions::
28598 * Predefined constants::
28599 * Interfacing with C::
28600 * Experience with source compatibility::
28603 @node Address types
28604 @subsubsection Address types
28607 To solve the problem of mixing 64-bit and 32-bit addressing,
28608 while maintaining maximum backward compatibility, the following
28609 approach has been taken:
28613 @code{System.Address} always has a size of 64 bits
28616 @code{System.Short_Address} is a 32-bit subtype of @code{System.Address}
28620 Since @code{System.Short_Address} is a subtype of @code{System.Address},
28621 a @code{Short_Address}
28622 may be used where an @code{Address} is required, and vice versa, without
28623 needing explicit type conversions.
28624 By virtue of the Open VMS parameter passing conventions,
28626 and exported subprograms that have 32-bit address parameters are
28627 compatible with those that have 64-bit address parameters.
28628 (See @ref{Making code 64 bit clean} for details.)
28630 The areas that may need attention are those where record types have
28631 been defined that contain components of the type @code{System.Address}, and
28632 where objects of this type are passed to code expecting a record layout with
28635 Different compilers on different platforms cannot be
28636 expected to represent the same type in the same way,
28637 since alignment constraints
28638 and other system-dependent properties affect the compiler's decision.
28639 For that reason, Ada code
28640 generally uses representation clauses to specify the expected
28641 layout where required.
28643 If such a representation clause uses 32 bits for a component having
28644 the type @code{System.Address}, 64-bit @value{EDITION} for OpenVMS
28645 will detect that error and produce a specific diagnostic message.
28646 The developer should then determine whether the representation
28647 should be 64 bits or not and make either of two changes:
28648 change the size to 64 bits and leave the type as @code{System.Address}, or
28649 leave the size as 32 bits and change the type to @code{System.Short_Address}.
28650 Since @code{Short_Address} is a subtype of @code{Address}, no changes are
28651 required in any code setting or accessing the field; the compiler will
28652 automatically perform any needed conversions between address
28656 @subsubsection Access types
28659 By default, objects designated by access values are always
28660 allocated in the 32-bit
28661 address space. Thus legacy code will never contain
28662 any objects that are not addressable with 32-bit addresses, and
28663 the compiler will never raise exceptions as result of mixing
28664 32-bit and 64-bit addresses.
28666 However, the access values themselves are represented in 64 bits, for optimum
28667 performance and future compatibility with 64-bit code. As was
28668 the case with @code{System.Address}, the compiler will give an error message
28669 if an object or record component has a representation clause that
28670 requires the access value to fit in 32 bits. In such a situation,
28671 an explicit size clause for the access type, specifying 32 bits,
28672 will have the desired effect.
28674 General access types (declared with @code{access all}) can never be
28675 32 bits, as values of such types must be able to refer to any object
28676 of the designated type,
28677 including objects residing outside the 32-bit address range.
28678 Existing Ada 83 code will not contain such type definitions,
28679 however, since general access types were introduced in Ada 95.
28681 @node Unchecked conversions
28682 @subsubsection Unchecked conversions
28685 In the case of an @code{Unchecked_Conversion} where the source type is a
28686 64-bit access type or the type @code{System.Address}, and the target
28687 type is a 32-bit type, the compiler will generate a warning.
28688 Even though the generated code will still perform the required
28689 conversions, it is highly recommended in these cases to use
28690 respectively a 32-bit access type or @code{System.Short_Address}
28691 as the source type.
28693 @node Predefined constants
28694 @subsubsection Predefined constants
28697 The following table shows the correspondence between pre-2006 versions of
28698 @value{EDITION} on Alpha OpenVMS (``Old'') and 64-bit @value{EDITION}
28701 @multitable {@code{System.Short_Memory_Size}} {2**32} {2**64}
28702 @item @b{Constant} @tab @b{Old} @tab @b{New}
28703 @item @code{System.Word_Size} @tab 32 @tab 64
28704 @item @code{System.Memory_Size} @tab 2**32 @tab 2**64
28705 @item @code{System.Short_Memory_Size} @tab 2**32 @tab 2**32
28706 @item @code{System.Address_Size} @tab 32 @tab 64
28710 If you need to refer to the specific
28711 memory size of a 32-bit implementation, instead of the
28712 actual memory size, use @code{System.Short_Memory_Size}
28713 rather than @code{System.Memory_Size}.
28714 Similarly, references to @code{System.Address_Size} may need
28715 to be replaced by @code{System.Short_Address'Size}.
28716 The program @command{gnatfind} may be useful for locating
28717 references to the above constants, so that you can verify that they
28720 @node Interfacing with C
28721 @subsubsection Interfacing with C
28724 In order to minimize the impact of the transition to 64-bit addresses on
28725 legacy programs, some fundamental types in the @code{Interfaces.C}
28726 package hierarchy continue to be represented in 32 bits.
28727 These types are: @code{ptrdiff_t}, @code{size_t}, and @code{chars_ptr}.
28728 This eases integration with the default HP C layout choices, for example
28729 as found in the system routines in @code{DECC$SHR.EXE}.
28730 Because of this implementation choice, the type fully compatible with
28731 @code{chars_ptr} is now @code{Short_Address} and not @code{Address}.
28732 Depending on the context the compiler will issue a
28733 warning or an error when type @code{Address} is used, alerting the user to a
28734 potential problem. Otherwise 32-bit programs that use
28735 @code{Interfaces.C} should normally not require code modifications
28737 The other issue arising with C interfacing concerns pragma @code{Convention}.
28738 For VMS 64-bit systems, there is an issue of the appropriate default size
28739 of C convention pointers in the absence of an explicit size clause. The HP
28740 C compiler can choose either 32 or 64 bits depending on compiler options.
28741 GNAT chooses 32-bits rather than 64-bits in the default case where no size
28742 clause is given. This proves a better choice for porting 32-bit legacy
28743 applications. In order to have a 64-bit representation, it is necessary to
28744 specify a size representation clause. For example:
28746 @smallexample @c ada
28747 type int_star is access Interfaces.C.int;
28748 pragma Convention(C, int_star);
28749 for int_star'Size use 64; -- Necessary to get 64 and not 32 bits
28752 @node Experience with source compatibility
28753 @subsubsection Experience with source compatibility
28756 The Security Server and STARLET on I64 provide an interesting ``test case''
28757 for source compatibility issues, since it is in such system code
28758 where assumptions about @code{Address} size might be expected to occur.
28759 Indeed, there were a small number of occasions in the Security Server
28760 file @file{jibdef.ads}
28761 where a representation clause for a record type specified
28762 32 bits for a component of type @code{Address}.
28763 All of these errors were detected by the compiler.
28764 The repair was obvious and immediate; to simply replace @code{Address} by
28765 @code{Short_Address}.
28767 In the case of STARLET, there were several record types that should
28768 have had representation clauses but did not. In these record types
28769 there was an implicit assumption that an @code{Address} value occupied
28771 These compiled without error, but their usage resulted in run-time error
28772 returns from STARLET system calls.
28773 Future GNAT technology enhancements may include a tool that detects and flags
28774 these sorts of potential source code porting problems.
28776 @c ****************************************
28777 @node Taking advantage of 64 bit addressing
28778 @subsection Taking advantage of 64-bit addressing
28781 * Making code 64 bit clean::
28782 * Allocating memory from the 64 bit storage pool::
28783 * Restrictions on use of 64 bit objects::
28784 * Using 64 bit storage pools by default::
28785 * General access types::
28786 * STARLET and other predefined libraries::
28789 @node Making code 64 bit clean
28790 @subsubsection Making code 64-bit clean
28793 In order to prevent problems that may occur when (parts of) a
28794 system start using memory outside the 32-bit address range,
28795 we recommend some additional guidelines:
28799 For imported subprograms that take parameters of the
28800 type @code{System.Address}, ensure that these subprograms can
28801 indeed handle 64-bit addresses. If not, or when in doubt,
28802 change the subprogram declaration to specify
28803 @code{System.Short_Address} instead.
28806 Resolve all warnings related to size mismatches in
28807 unchecked conversions. Failing to do so causes
28808 erroneous execution if the source object is outside
28809 the 32-bit address space.
28812 (optional) Explicitly use the 32-bit storage pool
28813 for access types used in a 32-bit context, or use
28814 generic access types where possible
28815 (@pxref{Restrictions on use of 64 bit objects}).
28819 If these rules are followed, the compiler will automatically insert
28820 any necessary checks to ensure that no addresses or access values
28821 passed to 32-bit code ever refer to objects outside the 32-bit
28823 Any attempt to do this will raise @code{Constraint_Error}.
28825 @node Allocating memory from the 64 bit storage pool
28826 @subsubsection Allocating memory from the 64-bit storage pool
28829 For any access type @code{T} that potentially requires memory allocations
28830 beyond the 32-bit address space,
28831 use the following representation clause:
28833 @smallexample @c ada
28834 for T'Storage_Pool use System.Pool_64;
28837 @node Restrictions on use of 64 bit objects
28838 @subsubsection Restrictions on use of 64-bit objects
28841 Taking the address of an object allocated from a 64-bit storage pool,
28842 and then passing this address to a subprogram expecting
28843 @code{System.Short_Address},
28844 or assigning it to a variable of type @code{Short_Address}, will cause
28845 @code{Constraint_Error} to be raised. In case the code is not 64-bit clean
28846 (@pxref{Making code 64 bit clean}), or checks are suppressed,
28847 no exception is raised and execution
28848 will become erroneous.
28850 @node Using 64 bit storage pools by default
28851 @subsubsection Using 64-bit storage pools by default
28854 In some cases it may be desirable to have the compiler allocate
28855 from 64-bit storage pools by default. This may be the case for
28856 libraries that are 64-bit clean, but may be used in both 32-bit
28857 and 64-bit contexts. For these cases the following configuration
28858 pragma may be specified:
28860 @smallexample @c ada
28861 pragma Pool_64_Default;
28865 Any code compiled in the context of this pragma will by default
28866 use the @code{System.Pool_64} storage pool. This default may be overridden
28867 for a specific access type @code{T} by the representation clause:
28869 @smallexample @c ada
28870 for T'Storage_Pool use System.Pool_32;
28874 Any object whose address may be passed to a subprogram with a
28875 @code{Short_Address} argument, or assigned to a variable of type
28876 @code{Short_Address}, needs to be allocated from this pool.
28878 @node General access types
28879 @subsubsection General access types
28882 Objects designated by access values from a
28883 general access type (declared with @code{access all}) are never allocated
28884 from a 64-bit storage pool. Code that uses general access types will
28885 accept objects allocated in either 32-bit or 64-bit address spaces,
28886 but never allocate objects outside the 32-bit address space.
28887 Using general access types ensures maximum compatibility with both
28888 32-bit and 64-bit code.
28890 @node STARLET and other predefined libraries
28891 @subsubsection STARLET and other predefined libraries
28894 All code that comes as part of GNAT is 64-bit clean, but the
28895 restrictions given in @ref{Restrictions on use of 64 bit objects},
28896 still apply. Look at the package
28897 specifications to see in which contexts objects allocated
28898 in 64-bit address space are acceptable.
28900 @node Technical details
28901 @subsection Technical details
28904 64-bit @value{EDITION} for Open VMS takes advantage of the freedom given in the
28905 Ada standard with respect to the type of @code{System.Address}. Previous
28906 versions of GNAT Pro have defined this type as private and implemented it as a
28909 In order to allow defining @code{System.Short_Address} as a proper subtype,
28910 and to match the implicit sign extension in parameter passing,
28911 in 64-bit @value{EDITION} for Open VMS, @code{System.Address} is defined as a
28912 visible (i.e., non-private) integer type.
28913 Standard operations on the type, such as the binary operators ``+'', ``-'',
28914 etc., that take @code{Address} operands and return an @code{Address} result,
28915 have been hidden by declaring these
28916 @code{abstract}, a feature introduced in Ada 95 that helps avoid the potential
28917 ambiguities that would otherwise result from overloading.
28918 (Note that, although @code{Address} is a visible integer type,
28919 good programming practice dictates against exploiting the type's
28920 integer properties such as literals, since this will compromise
28923 Defining @code{Address} as a visible integer type helps achieve
28924 maximum compatibility for existing Ada code,
28925 without sacrificing the capabilities of the 64-bit architecture.
28928 @c ************************************************
28930 @node Microsoft Windows Topics
28931 @appendix Microsoft Windows Topics
28937 This chapter describes topics that are specific to the Microsoft Windows
28938 platforms (NT, 2000, and XP Professional).
28941 * Using GNAT on Windows::
28942 * Using a network installation of GNAT::
28943 * CONSOLE and WINDOWS subsystems::
28944 * Temporary Files::
28945 * Mixed-Language Programming on Windows::
28946 * Windows Calling Conventions::
28947 * Introduction to Dynamic Link Libraries (DLLs)::
28948 * Using DLLs with GNAT::
28949 * Building DLLs with GNAT::
28950 * Building DLLs with GNAT Project files::
28951 * Building DLLs with gnatdll::
28952 * GNAT and Windows Resources::
28953 * Debugging a DLL::
28954 * Setting Stack Size from gnatlink::
28955 * Setting Heap Size from gnatlink::
28958 @node Using GNAT on Windows
28959 @section Using GNAT on Windows
28962 One of the strengths of the GNAT technology is that its tool set
28963 (@command{gcc}, @command{gnatbind}, @command{gnatlink}, @command{gnatmake}, the
28964 @code{gdb} debugger, etc.) is used in the same way regardless of the
28967 On Windows this tool set is complemented by a number of Microsoft-specific
28968 tools that have been provided to facilitate interoperability with Windows
28969 when this is required. With these tools:
28974 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
28978 You can use any Dynamically Linked Library (DLL) in your Ada code (both
28979 relocatable and non-relocatable DLLs are supported).
28982 You can build Ada DLLs for use in other applications. These applications
28983 can be written in a language other than Ada (e.g., C, C++, etc). Again both
28984 relocatable and non-relocatable Ada DLLs are supported.
28987 You can include Windows resources in your Ada application.
28990 You can use or create COM/DCOM objects.
28994 Immediately below are listed all known general GNAT-for-Windows restrictions.
28995 Other restrictions about specific features like Windows Resources and DLLs
28996 are listed in separate sections below.
29001 It is not possible to use @code{GetLastError} and @code{SetLastError}
29002 when tasking, protected records, or exceptions are used. In these
29003 cases, in order to implement Ada semantics, the GNAT run-time system
29004 calls certain Win32 routines that set the last error variable to 0 upon
29005 success. It should be possible to use @code{GetLastError} and
29006 @code{SetLastError} when tasking, protected record, and exception
29007 features are not used, but it is not guaranteed to work.
29010 It is not possible to link against Microsoft libraries except for
29011 import libraries. The library must be built to be compatible with
29012 @file{MSVCRT.LIB} (/MD Microsoft compiler option), @file{LIBC.LIB} and
29013 @file{LIBCMT.LIB} (/ML or /MT Microsoft compiler options) are known to
29014 not be compatible with the GNAT runtime. Even if the library is
29015 compatible with @file{MSVCRT.LIB} it is not guaranteed to work.
29018 When the compilation environment is located on FAT32 drives, users may
29019 experience recompilations of the source files that have not changed if
29020 Daylight Saving Time (DST) state has changed since the last time files
29021 were compiled. NTFS drives do not have this problem.
29024 No components of the GNAT toolset use any entries in the Windows
29025 registry. The only entries that can be created are file associations and
29026 PATH settings, provided the user has chosen to create them at installation
29027 time, as well as some minimal book-keeping information needed to correctly
29028 uninstall or integrate different GNAT products.
29031 @node Using a network installation of GNAT
29032 @section Using a network installation of GNAT
29035 Make sure the system on which GNAT is installed is accessible from the
29036 current machine, i.e. the install location is shared over the network.
29037 Shared resources are accessed on Windows by means of UNC paths, which
29038 have the format @code{\\server\sharename\path}
29040 In order to use such a network installation, simply add the UNC path of the
29041 @file{bin} directory of your GNAT installation in front of your PATH. For
29042 example, if GNAT is installed in @file{\GNAT} directory of a share location
29043 called @file{c-drive} on a machine @file{LOKI}, the following command will
29046 @code{@ @ @ path \\loki\c-drive\gnat\bin;%path%}
29048 Be aware that every compilation using the network installation results in the
29049 transfer of large amounts of data across the network and will likely cause
29050 serious performance penalty.
29052 @node CONSOLE and WINDOWS subsystems
29053 @section CONSOLE and WINDOWS subsystems
29054 @cindex CONSOLE Subsystem
29055 @cindex WINDOWS Subsystem
29059 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
29060 (which is the default subsystem) will always create a console when
29061 launching the application. This is not something desirable when the
29062 application has a Windows GUI. To get rid of this console the
29063 application must be using the @code{WINDOWS} subsystem. To do so
29064 the @option{-mwindows} linker option must be specified.
29067 $ gnatmake winprog -largs -mwindows
29070 @node Temporary Files
29071 @section Temporary Files
29072 @cindex Temporary files
29075 It is possible to control where temporary files gets created by setting
29076 the TMP environment variable. The file will be created:
29079 @item Under the directory pointed to by the TMP environment variable if
29080 this directory exists.
29082 @item Under c:\temp, if the TMP environment variable is not set (or not
29083 pointing to a directory) and if this directory exists.
29085 @item Under the current working directory otherwise.
29089 This allows you to determine exactly where the temporary
29090 file will be created. This is particularly useful in networked
29091 environments where you may not have write access to some
29094 @node Mixed-Language Programming on Windows
29095 @section Mixed-Language Programming on Windows
29098 Developing pure Ada applications on Windows is no different than on
29099 other GNAT-supported platforms. However, when developing or porting an
29100 application that contains a mix of Ada and C/C++, the choice of your
29101 Windows C/C++ development environment conditions your overall
29102 interoperability strategy.
29104 If you use @command{gcc} to compile the non-Ada part of your application,
29105 there are no Windows-specific restrictions that affect the overall
29106 interoperability with your Ada code. If you plan to use
29107 Microsoft tools (e.g. Microsoft Visual C/C++), you should be aware of
29108 the following limitations:
29112 You cannot link your Ada code with an object or library generated with
29113 Microsoft tools if these use the @code{.tls} section (Thread Local
29114 Storage section) since the GNAT linker does not yet support this section.
29117 You cannot link your Ada code with an object or library generated with
29118 Microsoft tools if these use I/O routines other than those provided in
29119 the Microsoft DLL: @code{msvcrt.dll}. This is because the GNAT run time
29120 uses the services of @code{msvcrt.dll} for its I/Os. Use of other I/O
29121 libraries can cause a conflict with @code{msvcrt.dll} services. For
29122 instance Visual C++ I/O stream routines conflict with those in
29127 If you do want to use the Microsoft tools for your non-Ada code and hit one
29128 of the above limitations, you have two choices:
29132 Encapsulate your non Ada code in a DLL to be linked with your Ada
29133 application. In this case, use the Microsoft or whatever environment to
29134 build the DLL and use GNAT to build your executable
29135 (@pxref{Using DLLs with GNAT}).
29138 Or you can encapsulate your Ada code in a DLL to be linked with the
29139 other part of your application. In this case, use GNAT to build the DLL
29140 (@pxref{Building DLLs with GNAT}) and use the Microsoft or whatever
29141 environment to build your executable.
29144 @node Windows Calling Conventions
29145 @section Windows Calling Conventions
29150 * C Calling Convention::
29151 * Stdcall Calling Convention::
29152 * Win32 Calling Convention::
29153 * DLL Calling Convention::
29157 When a subprogram @code{F} (caller) calls a subprogram @code{G}
29158 (callee), there are several ways to push @code{G}'s parameters on the
29159 stack and there are several possible scenarios to clean up the stack
29160 upon @code{G}'s return. A calling convention is an agreed upon software
29161 protocol whereby the responsibilities between the caller (@code{F}) and
29162 the callee (@code{G}) are clearly defined. Several calling conventions
29163 are available for Windows:
29167 @code{C} (Microsoft defined)
29170 @code{Stdcall} (Microsoft defined)
29173 @code{Win32} (GNAT specific)
29176 @code{DLL} (GNAT specific)
29179 @node C Calling Convention
29180 @subsection @code{C} Calling Convention
29183 This is the default calling convention used when interfacing to C/C++
29184 routines compiled with either @command{gcc} or Microsoft Visual C++.
29186 In the @code{C} calling convention subprogram parameters are pushed on the
29187 stack by the caller from right to left. The caller itself is in charge of
29188 cleaning up the stack after the call. In addition, the name of a routine
29189 with @code{C} calling convention is mangled by adding a leading underscore.
29191 The name to use on the Ada side when importing (or exporting) a routine
29192 with @code{C} calling convention is the name of the routine. For
29193 instance the C function:
29196 int get_val (long);
29200 should be imported from Ada as follows:
29202 @smallexample @c ada
29204 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
29205 pragma Import (C, Get_Val, External_Name => "get_val");
29210 Note that in this particular case the @code{External_Name} parameter could
29211 have been omitted since, when missing, this parameter is taken to be the
29212 name of the Ada entity in lower case. When the @code{Link_Name} parameter
29213 is missing, as in the above example, this parameter is set to be the
29214 @code{External_Name} with a leading underscore.
29216 When importing a variable defined in C, you should always use the @code{C}
29217 calling convention unless the object containing the variable is part of a
29218 DLL (in which case you should use the @code{Stdcall} calling
29219 convention, @pxref{Stdcall Calling Convention}).
29221 @node Stdcall Calling Convention
29222 @subsection @code{Stdcall} Calling Convention
29225 This convention, which was the calling convention used for Pascal
29226 programs, is used by Microsoft for all the routines in the Win32 API for
29227 efficiency reasons. It must be used to import any routine for which this
29228 convention was specified.
29230 In the @code{Stdcall} calling convention subprogram parameters are pushed
29231 on the stack by the caller from right to left. The callee (and not the
29232 caller) is in charge of cleaning the stack on routine exit. In addition,
29233 the name of a routine with @code{Stdcall} calling convention is mangled by
29234 adding a leading underscore (as for the @code{C} calling convention) and a
29235 trailing @code{@@}@code{@i{nn}}, where @i{nn} is the overall size (in
29236 bytes) of the parameters passed to the routine.
29238 The name to use on the Ada side when importing a C routine with a
29239 @code{Stdcall} calling convention is the name of the C routine. The leading
29240 underscore and trailing @code{@@}@code{@i{nn}} are added automatically by
29241 the compiler. For instance the Win32 function:
29244 @b{APIENTRY} int get_val (long);
29248 should be imported from Ada as follows:
29250 @smallexample @c ada
29252 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
29253 pragma Import (Stdcall, Get_Val);
29254 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
29259 As for the @code{C} calling convention, when the @code{External_Name}
29260 parameter is missing, it is taken to be the name of the Ada entity in lower
29261 case. If instead of writing the above import pragma you write:
29263 @smallexample @c ada
29265 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
29266 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
29271 then the imported routine is @code{_retrieve_val@@4}. However, if instead
29272 of specifying the @code{External_Name} parameter you specify the
29273 @code{Link_Name} as in the following example:
29275 @smallexample @c ada
29277 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
29278 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
29283 then the imported routine is @code{retrieve_val}, that is, there is no
29284 decoration at all. No leading underscore and no Stdcall suffix
29285 @code{@@}@code{@i{nn}}.
29288 This is especially important as in some special cases a DLL's entry
29289 point name lacks a trailing @code{@@}@code{@i{nn}} while the exported
29290 name generated for a call has it.
29293 It is also possible to import variables defined in a DLL by using an
29294 import pragma for a variable. As an example, if a DLL contains a
29295 variable defined as:
29302 then, to access this variable from Ada you should write:
29304 @smallexample @c ada
29306 My_Var : Interfaces.C.int;
29307 pragma Import (Stdcall, My_Var);
29312 Note that to ease building cross-platform bindings this convention
29313 will be handled as a @code{C} calling convention on non Windows platforms.
29315 @node Win32 Calling Convention
29316 @subsection @code{Win32} Calling Convention
29319 This convention, which is GNAT-specific is fully equivalent to the
29320 @code{Stdcall} calling convention described above.
29322 @node DLL Calling Convention
29323 @subsection @code{DLL} Calling Convention
29326 This convention, which is GNAT-specific is fully equivalent to the
29327 @code{Stdcall} calling convention described above.
29329 @node Introduction to Dynamic Link Libraries (DLLs)
29330 @section Introduction to Dynamic Link Libraries (DLLs)
29334 A Dynamically Linked Library (DLL) is a library that can be shared by
29335 several applications running under Windows. A DLL can contain any number of
29336 routines and variables.
29338 One advantage of DLLs is that you can change and enhance them without
29339 forcing all the applications that depend on them to be relinked or
29340 recompiled. However, you should be aware than all calls to DLL routines are
29341 slower since, as you will understand below, such calls are indirect.
29343 To illustrate the remainder of this section, suppose that an application
29344 wants to use the services of a DLL @file{API.dll}. To use the services
29345 provided by @file{API.dll} you must statically link against the DLL or
29346 an import library which contains a jump table with an entry for each
29347 routine and variable exported by the DLL. In the Microsoft world this
29348 import library is called @file{API.lib}. When using GNAT this import
29349 library is called either @file{libAPI.a} or @file{libapi.a} (names are
29352 After you have linked your application with the DLL or the import library
29353 and you run your application, here is what happens:
29357 Your application is loaded into memory.
29360 The DLL @file{API.dll} is mapped into the address space of your
29361 application. This means that:
29365 The DLL will use the stack of the calling thread.
29368 The DLL will use the virtual address space of the calling process.
29371 The DLL will allocate memory from the virtual address space of the calling
29375 Handles (pointers) can be safely exchanged between routines in the DLL
29376 routines and routines in the application using the DLL.
29380 The entries in the jump table (from the import library @file{libAPI.a}
29381 or @file{API.lib} or automatically created when linking against a DLL)
29382 which is part of your application are initialized with the addresses
29383 of the routines and variables in @file{API.dll}.
29386 If present in @file{API.dll}, routines @code{DllMain} or
29387 @code{DllMainCRTStartup} are invoked. These routines typically contain
29388 the initialization code needed for the well-being of the routines and
29389 variables exported by the DLL.
29393 There is an additional point which is worth mentioning. In the Windows
29394 world there are two kind of DLLs: relocatable and non-relocatable
29395 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
29396 in the target application address space. If the addresses of two
29397 non-relocatable DLLs overlap and these happen to be used by the same
29398 application, a conflict will occur and the application will run
29399 incorrectly. Hence, when possible, it is always preferable to use and
29400 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
29401 supported by GNAT. Note that the @option{-s} linker option (see GNU Linker
29402 User's Guide) removes the debugging symbols from the DLL but the DLL can
29403 still be relocated.
29405 As a side note, an interesting difference between Microsoft DLLs and
29406 Unix shared libraries, is the fact that on most Unix systems all public
29407 routines are exported by default in a Unix shared library, while under
29408 Windows it is possible (but not required) to list exported routines in
29409 a definition file (@pxref{The Definition File}).
29411 @node Using DLLs with GNAT
29412 @section Using DLLs with GNAT
29415 * Creating an Ada Spec for the DLL Services::
29416 * Creating an Import Library::
29420 To use the services of a DLL, say @file{API.dll}, in your Ada application
29425 The Ada spec for the routines and/or variables you want to access in
29426 @file{API.dll}. If not available this Ada spec must be built from the C/C++
29427 header files provided with the DLL.
29430 The import library (@file{libAPI.a} or @file{API.lib}). As previously
29431 mentioned an import library is a statically linked library containing the
29432 import table which will be filled at load time to point to the actual
29433 @file{API.dll} routines. Sometimes you don't have an import library for the
29434 DLL you want to use. The following sections will explain how to build
29435 one. Note that this is optional.
29438 The actual DLL, @file{API.dll}.
29442 Once you have all the above, to compile an Ada application that uses the
29443 services of @file{API.dll} and whose main subprogram is @code{My_Ada_App},
29444 you simply issue the command
29447 $ gnatmake my_ada_app -largs -lAPI
29451 The argument @option{-largs -lAPI} at the end of the @command{gnatmake} command
29452 tells the GNAT linker to look first for a library named @file{API.lib}
29453 (Microsoft-style name) and if not found for a library named @file{libAPI.a}
29454 (GNAT-style name). Note that if the Ada package spec for @file{API.dll}
29455 contains the following pragma
29457 @smallexample @c ada
29458 pragma Linker_Options ("-lAPI");
29462 you do not have to add @option{-largs -lAPI} at the end of the
29463 @command{gnatmake} command.
29465 If any one of the items above is missing you will have to create it
29466 yourself. The following sections explain how to do so using as an
29467 example a fictitious DLL called @file{API.dll}.
29469 @node Creating an Ada Spec for the DLL Services
29470 @subsection Creating an Ada Spec for the DLL Services
29473 A DLL typically comes with a C/C++ header file which provides the
29474 definitions of the routines and variables exported by the DLL. The Ada
29475 equivalent of this header file is a package spec that contains definitions
29476 for the imported entities. If the DLL you intend to use does not come with
29477 an Ada spec you have to generate one such spec yourself. For example if
29478 the header file of @file{API.dll} is a file @file{api.h} containing the
29479 following two definitions:
29491 then the equivalent Ada spec could be:
29493 @smallexample @c ada
29496 with Interfaces.C.Strings;
29501 function Get (Str : C.Strings.Chars_Ptr) return C.int;
29504 pragma Import (C, Get);
29505 pragma Import (DLL, Some_Var);
29512 Note that a variable is
29513 @strong{always imported with a Stdcall convention}. A function
29514 can have @code{C} or @code{Stdcall} convention.
29515 (@pxref{Windows Calling Conventions}).
29517 @node Creating an Import Library
29518 @subsection Creating an Import Library
29519 @cindex Import library
29522 * The Definition File::
29523 * GNAT-Style Import Library::
29524 * Microsoft-Style Import Library::
29528 If a Microsoft-style import library @file{API.lib} or a GNAT-style
29529 import library @file{libAPI.a} is available with @file{API.dll} you
29530 can skip this section. You can also skip this section if
29531 @file{API.dll} is built with GNU tools as in this case it is possible
29532 to link directly against the DLL. Otherwise read on.
29534 @node The Definition File
29535 @subsubsection The Definition File
29536 @cindex Definition file
29540 As previously mentioned, and unlike Unix systems, the list of symbols
29541 that are exported from a DLL must be provided explicitly in Windows.
29542 The main goal of a definition file is precisely that: list the symbols
29543 exported by a DLL. A definition file (usually a file with a @code{.def}
29544 suffix) has the following structure:
29550 [DESCRIPTION @i{string}]
29560 @item LIBRARY @i{name}
29561 This section, which is optional, gives the name of the DLL.
29563 @item DESCRIPTION @i{string}
29564 This section, which is optional, gives a description string that will be
29565 embedded in the import library.
29568 This section gives the list of exported symbols (procedures, functions or
29569 variables). For instance in the case of @file{API.dll} the @code{EXPORTS}
29570 section of @file{API.def} looks like:
29584 Note that you must specify the correct suffix (@code{@@}@code{@i{nn}})
29585 (@pxref{Windows Calling Conventions}) for a Stdcall
29586 calling convention function in the exported symbols list.
29589 There can actually be other sections in a definition file, but these
29590 sections are not relevant to the discussion at hand.
29592 @node GNAT-Style Import Library
29593 @subsubsection GNAT-Style Import Library
29596 To create a static import library from @file{API.dll} with the GNAT tools
29597 you should proceed as follows:
29601 Create the definition file @file{API.def} (@pxref{The Definition File}).
29602 For that use the @code{dll2def} tool as follows:
29605 $ dll2def API.dll > API.def
29609 @code{dll2def} is a very simple tool: it takes as input a DLL and prints
29610 to standard output the list of entry points in the DLL. Note that if
29611 some routines in the DLL have the @code{Stdcall} convention
29612 (@pxref{Windows Calling Conventions}) with stripped @code{@@}@i{nn}
29613 suffix then you'll have to edit @file{api.def} to add it, and specify
29614 @code{-k} to @code{gnatdll} when creating the import library.
29617 Here are some hints to find the right @code{@@}@i{nn} suffix.
29621 If you have the Microsoft import library (.lib), it is possible to get
29622 the right symbols by using Microsoft @code{dumpbin} tool (see the
29623 corresponding Microsoft documentation for further details).
29626 $ dumpbin /exports api.lib
29630 If you have a message about a missing symbol at link time the compiler
29631 tells you what symbol is expected. You just have to go back to the
29632 definition file and add the right suffix.
29636 Build the import library @code{libAPI.a}, using @code{gnatdll}
29637 (@pxref{Using gnatdll}) as follows:
29640 $ gnatdll -e API.def -d API.dll
29644 @code{gnatdll} takes as input a definition file @file{API.def} and the
29645 name of the DLL containing the services listed in the definition file
29646 @file{API.dll}. The name of the static import library generated is
29647 computed from the name of the definition file as follows: if the
29648 definition file name is @i{xyz}@code{.def}, the import library name will
29649 be @code{lib}@i{xyz}@code{.a}. Note that in the previous example option
29650 @option{-e} could have been removed because the name of the definition
29651 file (before the ``@code{.def}'' suffix) is the same as the name of the
29652 DLL (@pxref{Using gnatdll} for more information about @code{gnatdll}).
29655 @node Microsoft-Style Import Library
29656 @subsubsection Microsoft-Style Import Library
29659 With GNAT you can either use a GNAT-style or Microsoft-style import
29660 library. A Microsoft import library is needed only if you plan to make an
29661 Ada DLL available to applications developed with Microsoft
29662 tools (@pxref{Mixed-Language Programming on Windows}).
29664 To create a Microsoft-style import library for @file{API.dll} you
29665 should proceed as follows:
29669 Create the definition file @file{API.def} from the DLL. For this use either
29670 the @code{dll2def} tool as described above or the Microsoft @code{dumpbin}
29671 tool (see the corresponding Microsoft documentation for further details).
29674 Build the actual import library using Microsoft's @code{lib} utility:
29677 $ lib -machine:IX86 -def:API.def -out:API.lib
29681 If you use the above command the definition file @file{API.def} must
29682 contain a line giving the name of the DLL:
29689 See the Microsoft documentation for further details about the usage of
29693 @node Building DLLs with GNAT
29694 @section Building DLLs with GNAT
29695 @cindex DLLs, building
29698 This section explain how to build DLLs using the GNAT built-in DLL
29699 support. With the following procedure it is straight forward to build
29700 and use DLLs with GNAT.
29704 @item building object files
29706 The first step is to build all objects files that are to be included
29707 into the DLL. This is done by using the standard @command{gnatmake} tool.
29709 @item building the DLL
29711 To build the DLL you must use @command{gcc}'s @code{-shared}
29712 option. It is quite simple to use this method:
29715 $ gcc -shared -o api.dll obj1.o obj2.o ...
29718 It is important to note that in this case all symbols found in the
29719 object files are automatically exported. It is possible to restrict
29720 the set of symbols to export by passing to @command{gcc} a definition
29721 file, @pxref{The Definition File}. For example:
29724 $ gcc -shared -o api.dll api.def obj1.o obj2.o ...
29727 If you use a definition file you must export the elaboration procedures
29728 for every package that required one. Elaboration procedures are named
29729 using the package name followed by "_E".
29731 @item preparing DLL to be used
29733 For the DLL to be used by client programs the bodies must be hidden
29734 from it and the .ali set with read-only attribute. This is very important
29735 otherwise GNAT will recompile all packages and will not actually use
29736 the code in the DLL. For example:
29740 $ copy *.ads *.ali api.dll apilib
29741 $ attrib +R apilib\*.ali
29746 At this point it is possible to use the DLL by directly linking
29747 against it. Note that you must use the GNAT shared runtime when using
29748 GNAT shared libraries. This is achieved by using @code{-shared} binder's
29752 $ gnatmake main -Iapilib -bargs -shared -largs -Lapilib -lAPI
29755 @node Building DLLs with GNAT Project files
29756 @section Building DLLs with GNAT Project files
29757 @cindex DLLs, building
29760 There is nothing specific to Windows in this area. @pxref{Library Projects}.
29762 @node Building DLLs with gnatdll
29763 @section Building DLLs with gnatdll
29764 @cindex DLLs, building
29767 * Limitations When Using Ada DLLs from Ada::
29768 * Exporting Ada Entities::
29769 * Ada DLLs and Elaboration::
29770 * Ada DLLs and Finalization::
29771 * Creating a Spec for Ada DLLs::
29772 * Creating the Definition File::
29777 Note that it is preferred to use the built-in GNAT DLL support
29778 (@pxref{Building DLLs with GNAT}) or GNAT Project files
29779 (@pxref{Building DLLs with GNAT Project files}) to build DLLs.
29781 This section explains how to build DLLs containing Ada code using
29782 @code{gnatdll}. These DLLs will be referred to as Ada DLLs in the
29783 remainder of this section.
29785 The steps required to build an Ada DLL that is to be used by Ada as well as
29786 non-Ada applications are as follows:
29790 You need to mark each Ada @i{entity} exported by the DLL with a @code{C} or
29791 @code{Stdcall} calling convention to avoid any Ada name mangling for the
29792 entities exported by the DLL (@pxref{Exporting Ada Entities}). You can
29793 skip this step if you plan to use the Ada DLL only from Ada applications.
29796 Your Ada code must export an initialization routine which calls the routine
29797 @code{adainit} generated by @command{gnatbind} to perform the elaboration of
29798 the Ada code in the DLL (@pxref{Ada DLLs and Elaboration}). The initialization
29799 routine exported by the Ada DLL must be invoked by the clients of the DLL
29800 to initialize the DLL.
29803 When useful, the DLL should also export a finalization routine which calls
29804 routine @code{adafinal} generated by @command{gnatbind} to perform the
29805 finalization of the Ada code in the DLL (@pxref{Ada DLLs and Finalization}).
29806 The finalization routine exported by the Ada DLL must be invoked by the
29807 clients of the DLL when the DLL services are no further needed.
29810 You must provide a spec for the services exported by the Ada DLL in each
29811 of the programming languages to which you plan to make the DLL available.
29814 You must provide a definition file listing the exported entities
29815 (@pxref{The Definition File}).
29818 Finally you must use @code{gnatdll} to produce the DLL and the import
29819 library (@pxref{Using gnatdll}).
29823 Note that a relocatable DLL stripped using the @code{strip}
29824 binutils tool will not be relocatable anymore. To build a DLL without
29825 debug information pass @code{-largs -s} to @code{gnatdll}. This
29826 restriction does not apply to a DLL built using a Library Project.
29827 @pxref{Library Projects}.
29829 @node Limitations When Using Ada DLLs from Ada
29830 @subsection Limitations When Using Ada DLLs from Ada
29833 When using Ada DLLs from Ada applications there is a limitation users
29834 should be aware of. Because on Windows the GNAT run time is not in a DLL of
29835 its own, each Ada DLL includes a part of the GNAT run time. Specifically,
29836 each Ada DLL includes the services of the GNAT run time that are necessary
29837 to the Ada code inside the DLL. As a result, when an Ada program uses an
29838 Ada DLL there are two independent GNAT run times: one in the Ada DLL and
29839 one in the main program.
29841 It is therefore not possible to exchange GNAT run-time objects between the
29842 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
29843 handles (e.g. @code{Text_IO.File_Type}), tasks types, protected objects
29846 It is completely safe to exchange plain elementary, array or record types,
29847 Windows object handles, etc.
29849 @node Exporting Ada Entities
29850 @subsection Exporting Ada Entities
29851 @cindex Export table
29854 Building a DLL is a way to encapsulate a set of services usable from any
29855 application. As a result, the Ada entities exported by a DLL should be
29856 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
29857 any Ada name mangling. As an example here is an Ada package
29858 @code{API}, spec and body, exporting two procedures, a function, and a
29861 @smallexample @c ada
29864 with Interfaces.C; use Interfaces;
29866 Count : C.int := 0;
29867 function Factorial (Val : C.int) return C.int;
29869 procedure Initialize_API;
29870 procedure Finalize_API;
29871 -- Initialization & Finalization routines. More in the next section.
29873 pragma Export (C, Initialize_API);
29874 pragma Export (C, Finalize_API);
29875 pragma Export (C, Count);
29876 pragma Export (C, Factorial);
29882 @smallexample @c ada
29885 package body API is
29886 function Factorial (Val : C.int) return C.int is
29889 Count := Count + 1;
29890 for K in 1 .. Val loop
29896 procedure Initialize_API is
29898 pragma Import (C, Adainit);
29901 end Initialize_API;
29903 procedure Finalize_API is
29904 procedure Adafinal;
29905 pragma Import (C, Adafinal);
29915 If the Ada DLL you are building will only be used by Ada applications
29916 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
29917 convention. As an example, the previous package could be written as
29920 @smallexample @c ada
29924 Count : Integer := 0;
29925 function Factorial (Val : Integer) return Integer;
29927 procedure Initialize_API;
29928 procedure Finalize_API;
29929 -- Initialization and Finalization routines.
29935 @smallexample @c ada
29938 package body API is
29939 function Factorial (Val : Integer) return Integer is
29940 Fact : Integer := 1;
29942 Count := Count + 1;
29943 for K in 1 .. Val loop
29950 -- The remainder of this package body is unchanged.
29957 Note that if you do not export the Ada entities with a @code{C} or
29958 @code{Stdcall} convention you will have to provide the mangled Ada names
29959 in the definition file of the Ada DLL
29960 (@pxref{Creating the Definition File}).
29962 @node Ada DLLs and Elaboration
29963 @subsection Ada DLLs and Elaboration
29964 @cindex DLLs and elaboration
29967 The DLL that you are building contains your Ada code as well as all the
29968 routines in the Ada library that are needed by it. The first thing a
29969 user of your DLL must do is elaborate the Ada code
29970 (@pxref{Elaboration Order Handling in GNAT}).
29972 To achieve this you must export an initialization routine
29973 (@code{Initialize_API} in the previous example), which must be invoked
29974 before using any of the DLL services. This elaboration routine must call
29975 the Ada elaboration routine @code{adainit} generated by the GNAT binder
29976 (@pxref{Binding with Non-Ada Main Programs}). See the body of
29977 @code{Initialize_Api} for an example. Note that the GNAT binder is
29978 automatically invoked during the DLL build process by the @code{gnatdll}
29979 tool (@pxref{Using gnatdll}).
29981 When a DLL is loaded, Windows systematically invokes a routine called
29982 @code{DllMain}. It would therefore be possible to call @code{adainit}
29983 directly from @code{DllMain} without having to provide an explicit
29984 initialization routine. Unfortunately, it is not possible to call
29985 @code{adainit} from the @code{DllMain} if your program has library level
29986 tasks because access to the @code{DllMain} entry point is serialized by
29987 the system (that is, only a single thread can execute ``through'' it at a
29988 time), which means that the GNAT run time will deadlock waiting for the
29989 newly created task to complete its initialization.
29991 @node Ada DLLs and Finalization
29992 @subsection Ada DLLs and Finalization
29993 @cindex DLLs and finalization
29996 When the services of an Ada DLL are no longer needed, the client code should
29997 invoke the DLL finalization routine, if available. The DLL finalization
29998 routine is in charge of releasing all resources acquired by the DLL. In the
29999 case of the Ada code contained in the DLL, this is achieved by calling
30000 routine @code{adafinal} generated by the GNAT binder
30001 (@pxref{Binding with Non-Ada Main Programs}).
30002 See the body of @code{Finalize_Api} for an
30003 example. As already pointed out the GNAT binder is automatically invoked
30004 during the DLL build process by the @code{gnatdll} tool
30005 (@pxref{Using gnatdll}).
30007 @node Creating a Spec for Ada DLLs
30008 @subsection Creating a Spec for Ada DLLs
30011 To use the services exported by the Ada DLL from another programming
30012 language (e.g. C), you have to translate the specs of the exported Ada
30013 entities in that language. For instance in the case of @code{API.dll},
30014 the corresponding C header file could look like:
30019 extern int *_imp__count;
30020 #define count (*_imp__count)
30021 int factorial (int);
30027 It is important to understand that when building an Ada DLL to be used by
30028 other Ada applications, you need two different specs for the packages
30029 contained in the DLL: one for building the DLL and the other for using
30030 the DLL. This is because the @code{DLL} calling convention is needed to
30031 use a variable defined in a DLL, but when building the DLL, the variable
30032 must have either the @code{Ada} or @code{C} calling convention. As an
30033 example consider a DLL comprising the following package @code{API}:
30035 @smallexample @c ada
30039 Count : Integer := 0;
30041 -- Remainder of the package omitted.
30048 After producing a DLL containing package @code{API}, the spec that
30049 must be used to import @code{API.Count} from Ada code outside of the
30052 @smallexample @c ada
30057 pragma Import (DLL, Count);
30063 @node Creating the Definition File
30064 @subsection Creating the Definition File
30067 The definition file is the last file needed to build the DLL. It lists
30068 the exported symbols. As an example, the definition file for a DLL
30069 containing only package @code{API} (where all the entities are exported
30070 with a @code{C} calling convention) is:
30085 If the @code{C} calling convention is missing from package @code{API},
30086 then the definition file contains the mangled Ada names of the above
30087 entities, which in this case are:
30096 api__initialize_api
30101 @node Using gnatdll
30102 @subsection Using @code{gnatdll}
30106 * gnatdll Example::
30107 * gnatdll behind the Scenes::
30112 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
30113 and non-Ada sources that make up your DLL have been compiled.
30114 @code{gnatdll} is actually in charge of two distinct tasks: build the
30115 static import library for the DLL and the actual DLL. The form of the
30116 @code{gnatdll} command is
30120 $ gnatdll [@var{switches}] @var{list-of-files} [-largs @var{opts}]
30125 where @i{list-of-files} is a list of ALI and object files. The object
30126 file list must be the exact list of objects corresponding to the non-Ada
30127 sources whose services are to be included in the DLL. The ALI file list
30128 must be the exact list of ALI files for the corresponding Ada sources
30129 whose services are to be included in the DLL. If @i{list-of-files} is
30130 missing, only the static import library is generated.
30133 You may specify any of the following switches to @code{gnatdll}:
30136 @item -a[@var{address}]
30137 @cindex @option{-a} (@code{gnatdll})
30138 Build a non-relocatable DLL at @var{address}. If @var{address} is not
30139 specified the default address @var{0x11000000} will be used. By default,
30140 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
30141 advise the reader to build relocatable DLL.
30143 @item -b @var{address}
30144 @cindex @option{-b} (@code{gnatdll})
30145 Set the relocatable DLL base address. By default the address is
30148 @item -bargs @var{opts}
30149 @cindex @option{-bargs} (@code{gnatdll})
30150 Binder options. Pass @var{opts} to the binder.
30152 @item -d @var{dllfile}
30153 @cindex @option{-d} (@code{gnatdll})
30154 @var{dllfile} is the name of the DLL. This switch must be present for
30155 @code{gnatdll} to do anything. The name of the generated import library is
30156 obtained algorithmically from @var{dllfile} as shown in the following
30157 example: if @var{dllfile} is @code{xyz.dll}, the import library name is
30158 @code{libxyz.a}. The name of the definition file to use (if not specified
30159 by option @option{-e}) is obtained algorithmically from @var{dllfile}
30160 as shown in the following example:
30161 if @var{dllfile} is @code{xyz.dll}, the definition
30162 file used is @code{xyz.def}.
30164 @item -e @var{deffile}
30165 @cindex @option{-e} (@code{gnatdll})
30166 @var{deffile} is the name of the definition file.
30169 @cindex @option{-g} (@code{gnatdll})
30170 Generate debugging information. This information is stored in the object
30171 file and copied from there to the final DLL file by the linker,
30172 where it can be read by the debugger. You must use the
30173 @option{-g} switch if you plan on using the debugger or the symbolic
30177 @cindex @option{-h} (@code{gnatdll})
30178 Help mode. Displays @code{gnatdll} switch usage information.
30181 @cindex @option{-I} (@code{gnatdll})
30182 Direct @code{gnatdll} to search the @var{dir} directory for source and
30183 object files needed to build the DLL.
30184 (@pxref{Search Paths and the Run-Time Library (RTL)}).
30187 @cindex @option{-k} (@code{gnatdll})
30188 Removes the @code{@@}@i{nn} suffix from the import library's exported
30189 names, but keeps them for the link names. You must specify this
30190 option if you want to use a @code{Stdcall} function in a DLL for which
30191 the @code{@@}@i{nn} suffix has been removed. This is the case for most
30192 of the Windows NT DLL for example. This option has no effect when
30193 @option{-n} option is specified.
30195 @item -l @var{file}
30196 @cindex @option{-l} (@code{gnatdll})
30197 The list of ALI and object files used to build the DLL are listed in
30198 @var{file}, instead of being given in the command line. Each line in
30199 @var{file} contains the name of an ALI or object file.
30202 @cindex @option{-n} (@code{gnatdll})
30203 No Import. Do not create the import library.
30206 @cindex @option{-q} (@code{gnatdll})
30207 Quiet mode. Do not display unnecessary messages.
30210 @cindex @option{-v} (@code{gnatdll})
30211 Verbose mode. Display extra information.
30213 @item -largs @var{opts}
30214 @cindex @option{-largs} (@code{gnatdll})
30215 Linker options. Pass @var{opts} to the linker.
30218 @node gnatdll Example
30219 @subsubsection @code{gnatdll} Example
30222 As an example the command to build a relocatable DLL from @file{api.adb}
30223 once @file{api.adb} has been compiled and @file{api.def} created is
30226 $ gnatdll -d api.dll api.ali
30230 The above command creates two files: @file{libapi.a} (the import
30231 library) and @file{api.dll} (the actual DLL). If you want to create
30232 only the DLL, just type:
30235 $ gnatdll -d api.dll -n api.ali
30239 Alternatively if you want to create just the import library, type:
30242 $ gnatdll -d api.dll
30245 @node gnatdll behind the Scenes
30246 @subsubsection @code{gnatdll} behind the Scenes
30249 This section details the steps involved in creating a DLL. @code{gnatdll}
30250 does these steps for you. Unless you are interested in understanding what
30251 goes on behind the scenes, you should skip this section.
30253 We use the previous example of a DLL containing the Ada package @code{API},
30254 to illustrate the steps necessary to build a DLL. The starting point is a
30255 set of objects that will make up the DLL and the corresponding ALI
30256 files. In the case of this example this means that @file{api.o} and
30257 @file{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
30262 @code{gnatdll} builds the base file (@file{api.base}). A base file gives
30263 the information necessary to generate relocation information for the
30269 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
30274 In addition to the base file, the @command{gnatlink} command generates an
30275 output file @file{api.jnk} which can be discarded. The @option{-mdll} switch
30276 asks @command{gnatlink} to generate the routines @code{DllMain} and
30277 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
30278 is loaded into memory.
30281 @code{gnatdll} uses @code{dlltool} (@pxref{Using dlltool}) to build the
30282 export table (@file{api.exp}). The export table contains the relocation
30283 information in a form which can be used during the final link to ensure
30284 that the Windows loader is able to place the DLL anywhere in memory.
30288 $ dlltool --dllname api.dll --def api.def --base-file api.base \
30289 --output-exp api.exp
30294 @code{gnatdll} builds the base file using the new export table. Note that
30295 @command{gnatbind} must be called once again since the binder generated file
30296 has been deleted during the previous call to @command{gnatlink}.
30301 $ gnatlink api -o api.jnk api.exp -mdll
30302 -Wl,--base-file,api.base
30307 @code{gnatdll} builds the new export table using the new base file and
30308 generates the DLL import library @file{libAPI.a}.
30312 $ dlltool --dllname api.dll --def api.def --base-file api.base \
30313 --output-exp api.exp --output-lib libAPI.a
30318 Finally @code{gnatdll} builds the relocatable DLL using the final export
30324 $ gnatlink api api.exp -o api.dll -mdll
30329 @node Using dlltool
30330 @subsubsection Using @code{dlltool}
30333 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
30334 DLLs and static import libraries. This section summarizes the most
30335 common @code{dlltool} switches. The form of the @code{dlltool} command
30339 $ dlltool [@var{switches}]
30343 @code{dlltool} switches include:
30346 @item --base-file @var{basefile}
30347 @cindex @option{--base-file} (@command{dlltool})
30348 Read the base file @var{basefile} generated by the linker. This switch
30349 is used to create a relocatable DLL.
30351 @item --def @var{deffile}
30352 @cindex @option{--def} (@command{dlltool})
30353 Read the definition file.
30355 @item --dllname @var{name}
30356 @cindex @option{--dllname} (@command{dlltool})
30357 Gives the name of the DLL. This switch is used to embed the name of the
30358 DLL in the static import library generated by @code{dlltool} with switch
30359 @option{--output-lib}.
30362 @cindex @option{-k} (@command{dlltool})
30363 Kill @code{@@}@i{nn} from exported names
30364 (@pxref{Windows Calling Conventions}
30365 for a discussion about @code{Stdcall}-style symbols.
30368 @cindex @option{--help} (@command{dlltool})
30369 Prints the @code{dlltool} switches with a concise description.
30371 @item --output-exp @var{exportfile}
30372 @cindex @option{--output-exp} (@command{dlltool})
30373 Generate an export file @var{exportfile}. The export file contains the
30374 export table (list of symbols in the DLL) and is used to create the DLL.
30376 @item --output-lib @i{libfile}
30377 @cindex @option{--output-lib} (@command{dlltool})
30378 Generate a static import library @var{libfile}.
30381 @cindex @option{-v} (@command{dlltool})
30384 @item --as @i{assembler-name}
30385 @cindex @option{--as} (@command{dlltool})
30386 Use @i{assembler-name} as the assembler. The default is @code{as}.
30389 @node GNAT and Windows Resources
30390 @section GNAT and Windows Resources
30391 @cindex Resources, windows
30394 * Building Resources::
30395 * Compiling Resources::
30396 * Using Resources::
30400 Resources are an easy way to add Windows specific objects to your
30401 application. The objects that can be added as resources include:
30430 This section explains how to build, compile and use resources.
30432 @node Building Resources
30433 @subsection Building Resources
30434 @cindex Resources, building
30437 A resource file is an ASCII file. By convention resource files have an
30438 @file{.rc} extension.
30439 The easiest way to build a resource file is to use Microsoft tools
30440 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
30441 @code{dlgedit.exe} to build dialogs.
30442 It is always possible to build an @file{.rc} file yourself by writing a
30445 It is not our objective to explain how to write a resource file. A
30446 complete description of the resource script language can be found in the
30447 Microsoft documentation.
30449 @node Compiling Resources
30450 @subsection Compiling Resources
30453 @cindex Resources, compiling
30456 This section describes how to build a GNAT-compatible (COFF) object file
30457 containing the resources. This is done using the Resource Compiler
30458 @code{windres} as follows:
30461 $ windres -i myres.rc -o myres.o
30465 By default @code{windres} will run @command{gcc} to preprocess the @file{.rc}
30466 file. You can specify an alternate preprocessor (usually named
30467 @file{cpp.exe}) using the @code{windres} @option{--preprocessor}
30468 parameter. A list of all possible options may be obtained by entering
30469 the command @code{windres} @option{--help}.
30471 It is also possible to use the Microsoft resource compiler @code{rc.exe}
30472 to produce a @file{.res} file (binary resource file). See the
30473 corresponding Microsoft documentation for further details. In this case
30474 you need to use @code{windres} to translate the @file{.res} file to a
30475 GNAT-compatible object file as follows:
30478 $ windres -i myres.res -o myres.o
30481 @node Using Resources
30482 @subsection Using Resources
30483 @cindex Resources, using
30486 To include the resource file in your program just add the
30487 GNAT-compatible object file for the resource(s) to the linker
30488 arguments. With @command{gnatmake} this is done by using the @option{-largs}
30492 $ gnatmake myprog -largs myres.o
30495 @node Debugging a DLL
30496 @section Debugging a DLL
30497 @cindex DLL debugging
30500 * Program and DLL Both Built with GCC/GNAT::
30501 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
30505 Debugging a DLL is similar to debugging a standard program. But
30506 we have to deal with two different executable parts: the DLL and the
30507 program that uses it. We have the following four possibilities:
30511 The program and the DLL are built with @code{GCC/GNAT}.
30513 The program is built with foreign tools and the DLL is built with
30516 The program is built with @code{GCC/GNAT} and the DLL is built with
30522 In this section we address only cases one and two above.
30523 There is no point in trying to debug
30524 a DLL with @code{GNU/GDB}, if there is no GDB-compatible debugging
30525 information in it. To do so you must use a debugger compatible with the
30526 tools suite used to build the DLL.
30528 @node Program and DLL Both Built with GCC/GNAT
30529 @subsection Program and DLL Both Built with GCC/GNAT
30532 This is the simplest case. Both the DLL and the program have @code{GDB}
30533 compatible debugging information. It is then possible to break anywhere in
30534 the process. Let's suppose here that the main procedure is named
30535 @code{ada_main} and that in the DLL there is an entry point named
30539 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) and
30540 program must have been built with the debugging information (see GNAT -g
30541 switch). Here are the step-by-step instructions for debugging it:
30544 @item Launch @code{GDB} on the main program.
30550 @item Start the program and stop at the beginning of the main procedure
30557 This step is required to be able to set a breakpoint inside the DLL. As long
30558 as the program is not run, the DLL is not loaded. This has the
30559 consequence that the DLL debugging information is also not loaded, so it is not
30560 possible to set a breakpoint in the DLL.
30562 @item Set a breakpoint inside the DLL
30565 (gdb) break ada_dll
30572 At this stage a breakpoint is set inside the DLL. From there on
30573 you can use the standard approach to debug the whole program
30574 (@pxref{Running and Debugging Ada Programs}).
30577 @c This used to work, probably because the DLLs were non-relocatable
30578 @c keep this section around until the problem is sorted out.
30580 To break on the @code{DllMain} routine it is not possible to follow
30581 the procedure above. At the time the program stop on @code{ada_main}
30582 the @code{DllMain} routine as already been called. Either you can use
30583 the procedure below @pxref{Debugging the DLL Directly} or this procedure:
30586 @item Launch @code{GDB} on the main program.
30592 @item Load DLL symbols
30595 (gdb) add-sym api.dll
30598 @item Set a breakpoint inside the DLL
30601 (gdb) break ada_dll.adb:45
30604 Note that at this point it is not possible to break using the routine symbol
30605 directly as the program is not yet running. The solution is to break
30606 on the proper line (break in @file{ada_dll.adb} line 45).
30608 @item Start the program
30617 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT
30618 @subsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
30621 * Debugging the DLL Directly::
30622 * Attaching to a Running Process::
30626 In this case things are slightly more complex because it is not possible to
30627 start the main program and then break at the beginning to load the DLL and the
30628 associated DLL debugging information. It is not possible to break at the
30629 beginning of the program because there is no @code{GDB} debugging information,
30630 and therefore there is no direct way of getting initial control. This
30631 section addresses this issue by describing some methods that can be used
30632 to break somewhere in the DLL to debug it.
30635 First suppose that the main procedure is named @code{main} (this is for
30636 example some C code built with Microsoft Visual C) and that there is a
30637 DLL named @code{test.dll} containing an Ada entry point named
30641 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) must have
30642 been built with debugging information (see GNAT -g option).
30644 @node Debugging the DLL Directly
30645 @subsubsection Debugging the DLL Directly
30649 Find out the executable starting address
30652 $ objdump --file-header main.exe
30655 The starting address is reported on the last line. For example:
30658 main.exe: file format pei-i386
30659 architecture: i386, flags 0x0000010a:
30660 EXEC_P, HAS_DEBUG, D_PAGED
30661 start address 0x00401010
30665 Launch the debugger on the executable.
30672 Set a breakpoint at the starting address, and launch the program.
30675 $ (gdb) break *0x00401010
30679 The program will stop at the given address.
30682 Set a breakpoint on a DLL subroutine.
30685 (gdb) break ada_dll.adb:45
30688 Or if you want to break using a symbol on the DLL, you need first to
30689 select the Ada language (language used by the DLL).
30692 (gdb) set language ada
30693 (gdb) break ada_dll
30697 Continue the program.
30704 This will run the program until it reaches the breakpoint that has been
30705 set. From that point you can use the standard way to debug a program
30706 as described in (@pxref{Running and Debugging Ada Programs}).
30711 It is also possible to debug the DLL by attaching to a running process.
30713 @node Attaching to a Running Process
30714 @subsubsection Attaching to a Running Process
30715 @cindex DLL debugging, attach to process
30718 With @code{GDB} it is always possible to debug a running process by
30719 attaching to it. It is possible to debug a DLL this way. The limitation
30720 of this approach is that the DLL must run long enough to perform the
30721 attach operation. It may be useful for instance to insert a time wasting
30722 loop in the code of the DLL to meet this criterion.
30726 @item Launch the main program @file{main.exe}.
30732 @item Use the Windows @i{Task Manager} to find the process ID. Let's say
30733 that the process PID for @file{main.exe} is 208.
30741 @item Attach to the running process to be debugged.
30747 @item Load the process debugging information.
30750 (gdb) symbol-file main.exe
30753 @item Break somewhere in the DLL.
30756 (gdb) break ada_dll
30759 @item Continue process execution.
30768 This last step will resume the process execution, and stop at
30769 the breakpoint we have set. From there you can use the standard
30770 approach to debug a program as described in
30771 (@pxref{Running and Debugging Ada Programs}).
30773 @node Setting Stack Size from gnatlink
30774 @section Setting Stack Size from @command{gnatlink}
30777 It is possible to specify the program stack size at link time. On modern
30778 versions of Windows, starting with XP, this is mostly useful to set the size of
30779 the main stack (environment task). The other task stacks are set with pragma
30780 Storage_Size or with the @command{gnatbind -d} command.
30782 Since older versions of Windows (2000, NT4, etc.) do not allow setting the
30783 reserve size of individual tasks, the link-time stack size applies to all
30784 tasks, and pragma Storage_Size has no effect.
30785 In particular, Stack Overflow checks are made against this
30786 link-time specified size.
30788 This setting can be done with
30789 @command{gnatlink} using either:
30793 @item using @option{-Xlinker} linker option
30796 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
30799 This sets the stack reserve size to 0x10000 bytes and the stack commit
30800 size to 0x1000 bytes.
30802 @item using @option{-Wl} linker option
30805 $ gnatlink hello -Wl,--stack=0x1000000
30808 This sets the stack reserve size to 0x1000000 bytes. Note that with
30809 @option{-Wl} option it is not possible to set the stack commit size
30810 because the coma is a separator for this option.
30814 @node Setting Heap Size from gnatlink
30815 @section Setting Heap Size from @command{gnatlink}
30818 Under Windows systems, it is possible to specify the program heap size from
30819 @command{gnatlink} using either:
30823 @item using @option{-Xlinker} linker option
30826 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
30829 This sets the heap reserve size to 0x10000 bytes and the heap commit
30830 size to 0x1000 bytes.
30832 @item using @option{-Wl} linker option
30835 $ gnatlink hello -Wl,--heap=0x1000000
30838 This sets the heap reserve size to 0x1000000 bytes. Note that with
30839 @option{-Wl} option it is not possible to set the heap commit size
30840 because the coma is a separator for this option.
30846 @c **********************************
30847 @c * GNU Free Documentation License *
30848 @c **********************************
30850 @c GNU Free Documentation License
30852 @node Index,,GNU Free Documentation License, Top
30858 @c Put table of contents at end, otherwise it precedes the "title page" in
30859 @c the .txt version
30860 @c Edit the pdf file to move the contents to the beginning, after the title