1 \input texinfo @c -*-texinfo-*-
4 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
6 @c GNAT DOCUMENTATION o
10 @c Copyright (C) 1992-2004 Ada Core Technologies, Inc. o
12 @c GNAT is free software; you can redistribute it and/or modify it under o
13 @c terms of the GNU General Public License as published by the Free Soft- o
14 @c ware Foundation; either version 2, or (at your option) any later ver- o
15 @c sion. GNAT is distributed in the hope that it will be useful, but WITH- o
16 @c OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY o
17 @c or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License o
18 @c for more details. You should have received a copy of the GNU General o
19 @c Public License distributed with GNAT; see file COPYING. If not, write o
20 @c to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, o
21 @c MA 02111-1307, USA. o
23 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
25 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
27 @c GNAT_UGN Style Guide
29 @c 1. Always put a @noindent on the line before the first paragraph
30 @c after any of these commands:
42 @c 2. DO NOT use @example. Use @smallexample instead.
43 @c a) DO NOT use highlighting commands (@b{}, @i{}) inside an @smallexample
44 @c context. These can interfere with the readability of the texi
45 @c source file. Instead, use one of the following annotated
46 @c @smallexample commands, and preprocess the texi file with the
47 @c ada2texi tool (which generates appropriate highlighting):
48 @c @smallexample @c ada
49 @c @smallexample @c adanocomment
50 @c @smallexample @c projectfile
51 @c b) The "@c ada" markup will result in boldface for reserved words
52 @c and italics for comments
53 @c c) The "@c adanocomment" markup will result only in boldface for
54 @c reserved words (comments are left alone)
55 @c d) The "@c projectfile" markup is like "@c ada" except that the set
56 @c of reserved words include the new reserved words for project files
58 @c 3. Each @chapter, @section, @subsection, @subsubsection, etc.
59 @c command must be preceded by two empty lines
61 @c 4. The @item command should be on a line of its own if it is in an
62 @c @itemize or @enumerate command.
64 @c 5. When talking about ALI files use "ALI" (all uppercase), not "Ali"
67 @c 6. DO NOT put trailing spaces at the end of a line. Such spaces will
68 @c cause the document build to fail.
70 @c 7. DO NOT use @cartouche for examples that are longer than around 10 lines.
71 @c This command inhibits page breaks, so long examples in a @cartouche can
72 @c lead to large, ugly patches of empty space on a page.
74 @c NOTE: This file should be submitted to xgnatugn with either the vms flag
75 @c or the unw flag set. The unw flag covers topics for both Unix and
78 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
80 @setfilename gnat_ugn.info
83 @settitle GNAT User's Guide for Native Platforms / OpenVMS Alpha
87 @settitle GNAT User's Guide for Native Platforms / Unix and Windows
90 @include gcc-common.texi
92 @setchapternewpage odd
97 Copyright @copyright{} 1995-2004, Free Software Foundation
99 Permission is granted to copy, distribute and/or modify this document
100 under the terms of the GNU Free Documentation License, Version 1.2
101 or any later version published by the Free Software Foundation;
102 with the Invariant Sections being ``GNU Free Documentation License'', with the
103 Front-Cover Texts being
105 ``GNAT User's Guide for Native Platforms / OpenVMS Alpha'',
108 ``GNAT User's Guide for Native Platforms / Unix and Windows'',
110 and with no Back-Cover Texts.
111 A copy of the license is included in the section entitled
112 ``GNU Free Documentation License''.
117 @title GNAT User's Guide
118 @center @titlefont{for Native Platforms}
123 @titlefont{@i{Unix and Windows}}
126 @titlefont{@i{OpenVMS Alpha}}
131 @subtitle GNAT, The GNU Ada 95 Compiler
132 @subtitle GCC version @value{version-GCC}
134 @author Ada Core Technologies, Inc.
137 @vskip 0pt plus 1filll
145 @node Top, About This Guide, (dir), (dir)
146 @top GNAT User's Guide
150 GNAT User's Guide for Native Platforms / OpenVMS Alpha
155 GNAT User's Guide for Native Platforms / Unix and Windows
159 GNAT, The GNU Ada 95 Compiler@*
160 GCC version @value{version-GCC}@*
163 Ada Core Technologies, Inc.@*
167 * Getting Started with GNAT::
168 * The GNAT Compilation Model::
169 * Compiling Using gcc::
170 * Binding Using gnatbind::
171 * Linking Using gnatlink::
172 * The GNAT Make Program gnatmake::
173 * Improving Performance::
174 * Renaming Files Using gnatchop::
175 * Configuration Pragmas::
176 * Handling Arbitrary File Naming Conventions Using gnatname::
177 * GNAT Project Manager::
178 * The Cross-Referencing Tools gnatxref and gnatfind::
179 * The GNAT Pretty-Printer gnatpp::
180 * File Name Krunching Using gnatkr::
181 * Preprocessing Using gnatprep::
183 * The GNAT Run-Time Library Builder gnatlbr::
185 * The GNAT Library Browser gnatls::
186 * Cleaning Up Using gnatclean::
188 * GNAT and Libraries::
189 * Using the GNU make Utility::
191 * Finding Memory Problems::
192 * Creating Sample Bodies Using gnatstub::
193 * Other Utility Programs::
194 * Running and Debugging Ada Programs::
196 * Compatibility with DEC Ada::
198 * Platform-Specific Information for the Run-Time Libraries::
199 * Example of Binder Output File::
200 * Elaboration Order Handling in GNAT::
202 * Compatibility and Porting Guide::
204 * Microsoft Windows Topics::
206 * GNU Free Documentation License::
209 --- The Detailed Node Listing ---
213 * What This Guide Contains::
214 * What You Should Know before Reading This Guide::
215 * Related Information::
218 Getting Started with GNAT
221 * Running a Simple Ada Program::
222 * Running a Program with Multiple Units::
223 * Using the gnatmake Utility::
225 * Editing with Emacs::
228 * Introduction to GPS::
229 * Introduction to Glide and GVD::
232 The GNAT Compilation Model
234 * Source Representation::
235 * Foreign Language Representation::
236 * File Naming Rules::
237 * Using Other File Names::
238 * Alternative File Naming Schemes::
239 * Generating Object Files::
240 * Source Dependencies::
241 * The Ada Library Information Files::
242 * Binding an Ada Program::
243 * Mixed Language Programming::
244 * Building Mixed Ada & C++ Programs::
245 * Comparison between GNAT and C/C++ Compilation Models::
246 * Comparison between GNAT and Conventional Ada Library Models::
248 * Placement of temporary files::
251 Foreign Language Representation
254 * Other 8-Bit Codes::
255 * Wide Character Encodings::
257 Compiling Ada Programs With gcc
259 * Compiling Programs::
261 * Search Paths and the Run-Time Library (RTL)::
262 * Order of Compilation Issues::
267 * Output and Error Message Control::
268 * Warning Message Control::
269 * Debugging and Assertion Control::
271 * Stack Overflow Checking::
272 * Validity Checking::
274 * Using gcc for Syntax Checking::
275 * Using gcc for Semantic Checking::
276 * Compiling Ada 83 Programs::
277 * Character Set Control::
278 * File Naming Control::
279 * Subprogram Inlining Control::
280 * Auxiliary Output Control::
281 * Debugging Control::
282 * Exception Handling Control::
283 * Units to Sources Mapping Files::
284 * Integrated Preprocessing::
289 Binding Ada Programs With gnatbind
292 * Switches for gnatbind::
293 * Command-Line Access::
294 * Search Paths for gnatbind::
295 * Examples of gnatbind Usage::
297 Switches for gnatbind
299 * Consistency-Checking Modes::
300 * Binder Error Message Control::
301 * Elaboration Control::
303 * Binding with Non-Ada Main Programs::
304 * Binding Programs with No Main Subprogram::
306 Linking Using gnatlink
309 * Switches for gnatlink::
310 * Setting Stack Size from gnatlink::
311 * Setting Heap Size from gnatlink::
313 The GNAT Make Program gnatmake
316 * Switches for gnatmake::
317 * Mode Switches for gnatmake::
318 * Notes on the Command Line::
319 * How gnatmake Works::
320 * Examples of gnatmake Usage::
323 Improving Performance
324 * Performance Considerations::
325 * Reducing the Size of Ada Executables with gnatelim::
327 Performance Considerations
328 * Controlling Run-Time Checks::
329 * Use of Restrictions::
330 * Optimization Levels::
331 * Debugging Optimized Code::
332 * Inlining of Subprograms::
333 * Optimization and Strict Aliasing::
335 * Coverage Analysis::
338 Reducing the Size of Ada Executables with gnatelim
341 * Correcting the List of Eliminate Pragmas::
342 * Making Your Executables Smaller::
343 * Summary of the gnatelim Usage Cycle::
345 Renaming Files Using gnatchop
347 * Handling Files with Multiple Units::
348 * Operating gnatchop in Compilation Mode::
349 * Command Line for gnatchop::
350 * Switches for gnatchop::
351 * Examples of gnatchop Usage::
353 Configuration Pragmas
355 * Handling of Configuration Pragmas::
356 * The Configuration Pragmas Files::
358 Handling Arbitrary File Naming Conventions Using gnatname
360 * Arbitrary File Naming Conventions::
362 * Switches for gnatname::
363 * Examples of gnatname Usage::
368 * Examples of Project Files::
369 * Project File Syntax::
370 * Objects and Sources in Project Files::
371 * Importing Projects::
372 * Project Extension::
373 * External References in Project Files::
374 * Packages in Project Files::
375 * Variables from Imported Projects::
378 * Using Third-Party Libraries through Projects::
379 * Stand-alone Library Projects::
380 * Switches Related to Project Files::
381 * Tools Supporting Project Files::
382 * An Extended Example::
383 * Project File Complete Syntax::
386 The Cross-Referencing Tools gnatxref and gnatfind
388 * gnatxref Switches::
389 * gnatfind Switches::
390 * Project Files for gnatxref and gnatfind::
391 * Regular Expressions in gnatfind and gnatxref::
392 * Examples of gnatxref Usage::
393 * Examples of gnatfind Usage::
396 The GNAT Pretty-Printer gnatpp
398 * Switches for gnatpp::
402 File Name Krunching Using gnatkr
407 * Examples of gnatkr Usage::
409 Preprocessing Using gnatprep
412 * Switches for gnatprep::
413 * Form of Definitions File::
414 * Form of Input Text for gnatprep::
417 The GNAT Run-Time Library Builder gnatlbr
420 * Switches for gnatlbr::
421 * Examples of gnatlbr Usage::
424 The GNAT Library Browser gnatls
427 * Switches for gnatls::
428 * Examples of gnatls Usage::
430 Cleaning Up Using gnatclean
432 * Running gnatclean::
433 * Switches for gnatclean::
434 * Examples of gnatclean Usage::
440 * Creating an Ada Library::
441 * Installing an Ada Library::
442 * Using an Ada Library::
443 * Creating an Ada Library to be Used in a Non-Ada Context::
444 * Rebuilding the GNAT Run-Time Library::
446 Using the GNU make Utility
448 * Using gnatmake in a Makefile::
449 * Automatically Creating a List of Directories::
450 * Generating the Command Line Switches::
451 * Overcoming Command Line Length Limits::
454 Finding Memory Problems
459 * The GNAT Debug Pool Facility::
465 * Switches for gnatmem::
466 * Example of gnatmem Usage::
469 The GNAT Debug Pool Facility
471 Creating Sample Bodies Using gnatstub
474 * Switches for gnatstub::
476 Other Utility Programs
478 * Using Other Utility Programs with GNAT::
479 * The External Symbol Naming Scheme of GNAT::
481 * Ada Mode for Glide::
483 * Converting Ada Files to html with gnathtml::
485 Running and Debugging Ada Programs
487 * The GNAT Debugger GDB::
489 * Introduction to GDB Commands::
490 * Using Ada Expressions::
491 * Calling User-Defined Subprograms::
492 * Using the Next Command in a Function::
495 * Debugging Generic Units::
496 * GNAT Abnormal Termination or Failure to Terminate::
497 * Naming Conventions for GNAT Source Files::
498 * Getting Internal Debugging Information::
506 Compatibility with DEC Ada
508 * Ada 95 Compatibility::
509 * Differences in the Definition of Package System::
510 * Language-Related Features::
511 * The Package STANDARD::
512 * The Package SYSTEM::
513 * Tasking and Task-Related Features::
514 * Implementation of Tasks in DEC Ada for OpenVMS Alpha Systems::
515 * Pragmas and Pragma-Related Features::
516 * Library of Predefined Units::
518 * Main Program Definition::
519 * Implementation-Defined Attributes::
520 * Compiler and Run-Time Interfacing::
521 * Program Compilation and Library Management::
523 * Implementation Limits::
526 Language-Related Features
528 * Integer Types and Representations::
529 * Floating-Point Types and Representations::
530 * Pragmas Float_Representation and Long_Float::
531 * Fixed-Point Types and Representations::
532 * Record and Array Component Alignment::
534 * Other Representation Clauses::
536 Implementation of Tasks in DEC Ada for OpenVMS Alpha Systems
538 * Assigning Task IDs::
539 * Task IDs and Delays::
540 * Task-Related Pragmas::
541 * Scheduling and Task Priority::
543 * External Interrupts::
545 Pragmas and Pragma-Related Features
547 * Restrictions on the Pragma INLINE::
548 * Restrictions on the Pragma INTERFACE::
549 * Restrictions on the Pragma SYSTEM_NAME::
551 Library of Predefined Units
553 * Changes to DECLIB::
557 * Shared Libraries and Options Files::
561 Platform-Specific Information for the Run-Time Libraries
563 * Summary of Run-Time Configurations::
564 * Specifying a Run-Time Library::
565 * Choosing between Native and FSU Threads Libraries::
566 * Choosing the Scheduling Policy::
567 * Solaris-Specific Considerations::
568 * IRIX-Specific Considerations::
569 * Linux-Specific Considerations::
571 Example of Binder Output File
573 Elaboration Order Handling in GNAT
575 * Elaboration Code in Ada 95::
576 * Checking the Elaboration Order in Ada 95::
577 * Controlling the Elaboration Order in Ada 95::
578 * Controlling Elaboration in GNAT - Internal Calls::
579 * Controlling Elaboration in GNAT - External Calls::
580 * Default Behavior in GNAT - Ensuring Safety::
581 * Treatment of Pragma Elaborate::
582 * Elaboration Issues for Library Tasks::
583 * Mixing Elaboration Models::
584 * What to Do If the Default Elaboration Behavior Fails::
585 * Elaboration for Access-to-Subprogram Values::
586 * Summary of Procedures for Elaboration Control::
587 * Other Elaboration Order Considerations::
591 * Basic Assembler Syntax::
592 * A Simple Example of Inline Assembler::
593 * Output Variables in Inline Assembler::
594 * Input Variables in Inline Assembler::
595 * Inlining Inline Assembler Code::
596 * Other Asm Functionality::
597 * A Complete Example::
599 Compatibility and Porting Guide
601 * Compatibility with Ada 83::
602 * Implementation-dependent characteristics::
603 * Compatibility with DEC Ada 83::
604 * Compatibility with Other Ada 95 Systems::
605 * Representation Clauses::
608 Microsoft Windows Topics
610 * Using GNAT on Windows::
611 * CONSOLE and WINDOWS subsystems::
613 * Mixed-Language Programming on Windows::
614 * Windows Calling Conventions::
615 * Introduction to Dynamic Link Libraries (DLLs)::
616 * Using DLLs with GNAT::
617 * Building DLLs with GNAT::
618 * GNAT and Windows Resources::
620 * GNAT and COM/DCOM Objects::
628 @node About This Guide
629 @unnumbered About This Guide
633 This guide describes the use of of GNAT, a full language compiler for the Ada
634 95 programming language, implemented on HP OpenVMS Alpha platforms.
637 This guide describes the use of GNAT, a compiler and software development
638 toolset for the full Ada 95 programming language.
640 It describes the features of the compiler and tools, and details
641 how to use them to build Ada 95 applications.
644 * What This Guide Contains::
645 * What You Should Know before Reading This Guide::
646 * Related Information::
650 @node What This Guide Contains
651 @unnumberedsec What This Guide Contains
654 This guide contains the following chapters:
658 @ref{Getting Started with GNAT}, describes how to get started compiling
659 and running Ada programs with the GNAT Ada programming environment.
661 @ref{The GNAT Compilation Model}, describes the compilation model used
665 @ref{Compiling Using gcc}, describes how to compile
666 Ada programs with @code{gcc}, the Ada compiler.
669 @ref{Binding Using gnatbind}, describes how to
670 perform binding of Ada programs with @code{gnatbind}, the GNAT binding
674 @ref{Linking Using gnatlink},
675 describes @code{gnatlink}, a
676 program that provides for linking using the GNAT run-time library to
677 construct a program. @code{gnatlink} can also incorporate foreign language
678 object units into the executable.
681 @ref{The GNAT Make Program gnatmake}, describes @code{gnatmake}, a
682 utility that automatically determines the set of sources
683 needed by an Ada compilation unit, and executes the necessary compilations
687 @ref{Improving Performance}, shows various techniques for making your
688 Ada program run faster or take less space.
689 It discusses the effect of the compiler's optimization switch and
690 also describes the @command{gnatelim} tool.
693 @ref{Renaming Files Using gnatchop}, describes
694 @code{gnatchop}, a utility that allows you to preprocess a file that
695 contains Ada source code, and split it into one or more new files, one
696 for each compilation unit.
699 @ref{Configuration Pragmas}, describes the configuration pragmas
703 @ref{Handling Arbitrary File Naming Conventions Using gnatname},
704 shows how to override the default GNAT file naming conventions,
705 either for an individual unit or globally.
708 @ref{GNAT Project Manager}, describes how to use project files
709 to organize large projects.
712 @ref{The Cross-Referencing Tools gnatxref and gnatfind}, discusses
713 @code{gnatxref} and @code{gnatfind}, two tools that provide an easy
714 way to navigate through sources.
717 @ref{The GNAT Pretty-Printer gnatpp}, shows how to produce a reformatted
718 version of an Ada source file with control over casing, indentation,
719 comment placement, and other elements of program presentation style.
723 @ref{File Name Krunching Using gnatkr}, describes the @code{gnatkr}
724 file name krunching utility, used to handle shortened
725 file names on operating systems with a limit on the length of names.
728 @ref{Preprocessing Using gnatprep}, describes @code{gnatprep}, a
729 preprocessor utility that allows a single source file to be used to
730 generate multiple or parameterized source files, by means of macro
735 @ref{The GNAT Run-Time Library Builder gnatlbr}, describes @command{gnatlbr},
736 a tool for rebuilding the GNAT run time with user-supplied
737 configuration pragmas.
741 @ref{The GNAT Library Browser gnatls}, describes @code{gnatls}, a
742 utility that displays information about compiled units, including dependences
743 on the corresponding sources files, and consistency of compilations.
746 @ref{Cleaning Up Using gnatclean}, describes @code{gnatclean}, a utility
747 to delete files that are produced by the compiler, binder and linker.
751 @ref{GNAT and Libraries}, describes the process of creating and using
752 Libraries with GNAT. It also describes how to recompile the GNAT run-time
756 @ref{Using the GNU make Utility}, describes some techniques for using
757 the GNAT toolset in Makefiles.
761 @ref{Finding Memory Problems}, describes
763 @command{gnatmem}, a utility that monitors dynamic allocation and deallocation
764 and helps detect ``memory leaks'', and
766 the GNAT Debug Pool facility, which helps detect incorrect memory references.
769 @ref{Creating Sample Bodies Using gnatstub}, discusses @code{gnatstub},
770 a utility that generates empty but compilable bodies for library units.
773 @ref{Other Utility Programs}, discusses several other GNAT utilities,
774 including @code{gnathtml}.
777 @ref{Running and Debugging Ada Programs}, describes how to run and debug
782 @ref{Compatibility with DEC Ada}, details the compatibility of GNAT with
783 DEC Ada 83 @footnote{``DEC Ada'' refers to the legacy product originally
784 developed by Digital Equipment Corporation and currently supported by HP.}
789 @ref{Platform-Specific Information for the Run-Time Libraries},
790 describes the various run-time
791 libraries supported by GNAT on various platforms and explains how to
792 choose a particular library.
795 @ref{Example of Binder Output File}, shows the source code for the binder
796 output file for a sample program.
799 @ref{Elaboration Order Handling in GNAT}, describes how GNAT helps
800 you deal with elaboration order issues.
803 @ref{Inline Assembler}, shows how to use the inline assembly facility
807 @ref{Compatibility and Porting Guide}, includes sections on compatibility
808 of GNAT with other Ada 83 and Ada 95 compilation systems, to assist
809 in porting code from other environments.
813 @ref{Microsoft Windows Topics}, presents information relevant to the
814 Microsoft Windows platform.
819 @c *************************************************
820 @node What You Should Know before Reading This Guide
821 @c *************************************************
822 @unnumberedsec What You Should Know before Reading This Guide
824 @cindex Ada 95 Language Reference Manual
826 This user's guide assumes that you are familiar with Ada 95 language, as
827 described in the International Standard ANSI/ISO/IEC-8652:1995, January
830 @node Related Information
831 @unnumberedsec Related Information
834 For further information about related tools, refer to the following
839 @cite{GNAT Reference Manual}, which contains all reference
840 material for the GNAT implementation of Ada 95.
844 @cite{Using the GNAT Programming System}, which describes the GPS
845 integrated development environment.
848 @cite{GNAT Programming System Tutorial}, which introduces the
849 main GPS features through examples.
853 @cite{Ada 95 Language Reference Manual}, which contains all reference
854 material for the Ada 95 programming language.
857 @cite{Debugging with GDB}
859 , located in the GNU:[DOCS] directory,
861 contains all details on the use of the GNU source-level debugger.
864 @cite{GNU Emacs Manual}
866 , located in the GNU:[DOCS] directory if the EMACS kit is installed,
868 contains full information on the extensible editor and programming
875 @unnumberedsec Conventions
877 @cindex Typographical conventions
880 Following are examples of the typographical and graphic conventions used
885 @code{Functions}, @code{utility program names}, @code{standard names},
892 @file{File Names}, @file{button names}, and @file{field names}.
901 [optional information or parameters]
904 Examples are described by text
906 and then shown this way.
911 Commands that are entered by the user are preceded in this manual by the
912 characters @w{``@code{$ }''} (dollar sign followed by space). If your system
913 uses this sequence as a prompt, then the commands will appear exactly as
914 you see them in the manual. If your system uses some other prompt, then
915 the command will appear with the @code{$} replaced by whatever prompt
916 character you are using.
919 Full file names are shown with the ``@code{/}'' character
920 as the directory separator; e.g., @file{parent-dir/subdir/myfile.adb}.
921 If you are using GNAT on a Windows platform, please note that
922 the ``@code{\}'' character should be used instead.
927 @c ****************************
928 @node Getting Started with GNAT
929 @chapter Getting Started with GNAT
932 This chapter describes some simple ways of using GNAT to build
933 executable Ada programs.
935 @ref{Running GNAT}, through @ref{Using the gnatmake Utility},
936 show how to use the command line environment.
937 @ref{Introduction to Glide and GVD}, provides a brief
938 introduction to the visually-oriented IDE for GNAT.
939 Supplementing Glide on some platforms is GPS, the
940 GNAT Programming System, which offers a richer graphical
941 ``look and feel'', enhanced configurability, support for
942 development in other programming language, comprehensive
943 browsing features, and many other capabilities.
944 For information on GPS please refer to
945 @cite{Using the GNAT Programming System}.
950 * Running a Simple Ada Program::
951 * Running a Program with Multiple Units::
952 * Using the gnatmake Utility::
954 * Editing with Emacs::
957 * Introduction to GPS::
958 * Introduction to Glide and GVD::
963 @section Running GNAT
966 Three steps are needed to create an executable file from an Ada source
971 The source file(s) must be compiled.
973 The file(s) must be bound using the GNAT binder.
975 All appropriate object files must be linked to produce an executable.
979 All three steps are most commonly handled by using the @code{gnatmake}
980 utility program that, given the name of the main program, automatically
981 performs the necessary compilation, binding and linking steps.
984 @node Running a Simple Ada Program
985 @section Running a Simple Ada Program
988 Any text editor may be used to prepare an Ada program.
991 used, the optional Ada mode may be helpful in laying out the program.
994 program text is a normal text file. We will suppose in our initial
995 example that you have used your editor to prepare the following
996 standard format text file:
1000 with Ada.Text_IO; use Ada.Text_IO;
1003 Put_Line ("Hello WORLD!");
1009 This file should be named @file{hello.adb}.
1010 With the normal default file naming conventions, GNAT requires
1012 contain a single compilation unit whose file name is the
1014 with periods replaced by hyphens; the
1015 extension is @file{ads} for a
1016 spec and @file{adb} for a body.
1017 You can override this default file naming convention by use of the
1018 special pragma @code{Source_File_Name} (@pxref{Using Other File Names}).
1019 Alternatively, if you want to rename your files according to this default
1020 convention, which is probably more convenient if you will be using GNAT
1021 for all your compilations, then the @code{gnatchop} utility
1022 can be used to generate correctly-named source files
1023 (@pxref{Renaming Files Using gnatchop}).
1025 You can compile the program using the following command (@code{$} is used
1026 as the command prompt in the examples in this document):
1033 @code{gcc} is the command used to run the compiler. This compiler is
1034 capable of compiling programs in several languages, including Ada 95 and
1035 C. It assumes that you have given it an Ada program if the file extension is
1036 either @file{.ads} or @file{.adb}, and it will then call
1037 the GNAT compiler to compile the specified file.
1040 The @option{-c} switch is required. It tells @command{gcc} to only do a
1041 compilation. (For C programs, @command{gcc} can also do linking, but this
1042 capability is not used directly for Ada programs, so the @option{-c}
1043 switch must always be present.)
1046 This compile command generates a file
1047 @file{hello.o}, which is the object
1048 file corresponding to your Ada program. It also generates
1049 an ``Ada Library Information'' file @file{hello.ali},
1050 which contains additional information used to check
1051 that an Ada program is consistent.
1052 To build an executable file,
1053 use @code{gnatbind} to bind the program
1054 and @code{gnatlink} to link it. The
1055 argument to both @code{gnatbind} and @code{gnatlink} is the name of the
1056 @file{ALI} file, but the default extension of @file{.ali} can
1057 be omitted. This means that in the most common case, the argument
1058 is simply the name of the main program:
1066 A simpler method of carrying out these steps is to use
1068 a master program that invokes all the required
1069 compilation, binding and linking tools in the correct order. In particular,
1070 @command{gnatmake} automatically recompiles any sources that have been
1071 modified since they were last compiled, or sources that depend
1072 on such modified sources, so that ``version skew'' is avoided.
1073 @cindex Version skew (avoided by @command{gnatmake})
1076 $ gnatmake hello.adb
1080 The result is an executable program called @file{hello}, which can be
1083 @c The following should be removed (BMB 2001-01-23)
1085 @c $ ^./hello^$ RUN HELLO^
1086 @c @end smallexample
1093 assuming that the current directory is on the search path
1094 for executable programs.
1097 and, if all has gone well, you will see
1104 appear in response to this command.
1107 @c ****************************************
1108 @node Running a Program with Multiple Units
1109 @section Running a Program with Multiple Units
1112 Consider a slightly more complicated example that has three files: a
1113 main program, and the spec and body of a package:
1115 @smallexample @c ada
1118 package Greetings is
1123 with Ada.Text_IO; use Ada.Text_IO;
1124 package body Greetings is
1127 Put_Line ("Hello WORLD!");
1130 procedure Goodbye is
1132 Put_Line ("Goodbye WORLD!");
1149 Following the one-unit-per-file rule, place this program in the
1150 following three separate files:
1154 spec of package @code{Greetings}
1157 body of package @code{Greetings}
1160 body of main program
1164 To build an executable version of
1165 this program, we could use four separate steps to compile, bind, and link
1166 the program, as follows:
1170 $ gcc -c greetings.adb
1176 Note that there is no required order of compilation when using GNAT.
1177 In particular it is perfectly fine to compile the main program first.
1178 Also, it is not necessary to compile package specs in the case where
1179 there is an accompanying body; you only need to compile the body. If you want
1180 to submit these files to the compiler for semantic checking and not code
1181 generation, then use the
1182 @option{-gnatc} switch:
1185 $ gcc -c greetings.ads -gnatc
1189 Although the compilation can be done in separate steps as in the
1190 above example, in practice it is almost always more convenient
1191 to use the @code{gnatmake} tool. All you need to know in this case
1192 is the name of the main program's source file. The effect of the above four
1193 commands can be achieved with a single one:
1196 $ gnatmake gmain.adb
1200 In the next section we discuss the advantages of using @code{gnatmake} in
1203 @c *****************************
1204 @node Using the gnatmake Utility
1205 @section Using the @command{gnatmake} Utility
1208 If you work on a program by compiling single components at a time using
1209 @code{gcc}, you typically keep track of the units you modify. In order to
1210 build a consistent system, you compile not only these units, but also any
1211 units that depend on the units you have modified.
1212 For example, in the preceding case,
1213 if you edit @file{gmain.adb}, you only need to recompile that file. But if
1214 you edit @file{greetings.ads}, you must recompile both
1215 @file{greetings.adb} and @file{gmain.adb}, because both files contain
1216 units that depend on @file{greetings.ads}.
1218 @code{gnatbind} will warn you if you forget one of these compilation
1219 steps, so that it is impossible to generate an inconsistent program as a
1220 result of forgetting to do a compilation. Nevertheless it is tedious and
1221 error-prone to keep track of dependencies among units.
1222 One approach to handle the dependency-bookkeeping is to use a
1223 makefile. However, makefiles present maintenance problems of their own:
1224 if the dependencies change as you change the program, you must make
1225 sure that the makefile is kept up-to-date manually, which is also an
1226 error-prone process.
1228 The @code{gnatmake} utility takes care of these details automatically.
1229 Invoke it using either one of the following forms:
1232 $ gnatmake gmain.adb
1233 $ gnatmake ^gmain^GMAIN^
1237 The argument is the name of the file containing the main program;
1238 you may omit the extension. @code{gnatmake}
1239 examines the environment, automatically recompiles any files that need
1240 recompiling, and binds and links the resulting set of object files,
1241 generating the executable file, @file{^gmain^GMAIN.EXE^}.
1242 In a large program, it
1243 can be extremely helpful to use @code{gnatmake}, because working out by hand
1244 what needs to be recompiled can be difficult.
1246 Note that @code{gnatmake}
1247 takes into account all the Ada 95 rules that
1248 establish dependencies among units. These include dependencies that result
1249 from inlining subprogram bodies, and from
1250 generic instantiation. Unlike some other
1251 Ada make tools, @code{gnatmake} does not rely on the dependencies that were
1252 found by the compiler on a previous compilation, which may possibly
1253 be wrong when sources change. @code{gnatmake} determines the exact set of
1254 dependencies from scratch each time it is run.
1257 @node Editing with Emacs
1258 @section Editing with Emacs
1262 Emacs is an extensible self-documenting text editor that is available in a
1263 separate VMSINSTAL kit.
1265 Invoke Emacs by typing @kbd{Emacs} at the command prompt. To get started,
1266 click on the Emacs Help menu and run the Emacs Tutorial.
1267 In a character cell terminal, Emacs help is invoked with @kbd{Ctrl-h} (also
1268 written as @kbd{C-h}), and the tutorial by @kbd{C-h t}.
1270 Documentation on Emacs and other tools is available in Emacs under the
1271 pull-down menu button: @code{Help - Info}. After selecting @code{Info},
1272 use the middle mouse button to select a topic (e.g. Emacs).
1274 In a character cell terminal, do @kbd{C-h i} to invoke info, and then @kbd{m}
1275 (stands for menu) followed by the menu item desired, as in @kbd{m Emacs}, to
1276 get to the Emacs manual.
1277 Help on Emacs is also available by typing @kbd{HELP EMACS} at the DCL command
1280 The tutorial is highly recommended in order to learn the intricacies of Emacs,
1281 which is sufficiently extensible to provide for a complete programming
1282 environment and shell for the sophisticated user.
1286 @node Introduction to GPS
1287 @section Introduction to GPS
1288 @cindex GPS (GNAT Programming System)
1289 @cindex GNAT Programming System (GPS)
1291 Although the command line interface (@command{gnatmake}, etc.) alone
1292 is sufficient, a graphical Interactive Development
1293 Environment can make it easier for you to compose, navigate, and debug
1294 programs. This section describes the main features of GPS
1295 (``GNAT Programming System''), the GNAT graphical IDE.
1296 You will see how to use GPS to build and debug an executable, and
1297 you will also learn some of the basics of the GNAT ``project'' facility.
1299 GPS enables you to do much more than is presented here;
1300 e.g., you can produce a call graph, interface to a third-party
1301 Version Control System, and inspect the generated assembly language
1303 Indeed, GPS also supports languages other than Ada.
1304 Such additional information, and an explanation of all of the GPS menu
1305 items. may be found in the on-line help, which includes
1306 a user's guide and a tutorial (these are also accessible from the GNAT
1310 * Building a New Program with GPS::
1311 * Simple Debugging with GPS::
1315 @node Building a New Program with GPS
1316 @subsection Building a New Program with GPS
1318 GPS invokes the GNAT compilation tools using information
1319 contained in a @emph{project} (also known as a @emph{project file}):
1320 a collection of properties such
1321 as source directories, identities of main subprograms, tool switches, etc.,
1322 and their associated values.
1323 (See @ref{GNAT Project Manager}, for details.)
1324 In order to run GPS, you will need to either create a new project
1325 or else open an existing one.
1327 This section will explain how you can use GPS to create a project,
1328 to associate Ada source files with a project, and to build and run
1332 @item @emph{Creating a project}
1334 Invoke GPS, either from the command line or the platform's IDE.
1335 After it starts, GPS will display a ``Welcome'' screen with three
1340 @code{Start with default project in directory}
1343 @code{Create new project with wizard}
1346 @code{Open existing project}
1350 Select @code{Create new project with wizard} and press @code{OK}.
1351 A new window will appear. In the text box labeled with
1352 @code{Enter the name of the project to create}, type @file{sample}
1353 as the project name.
1354 In the next box, browse to choose the directory in which you
1355 would like to create the project file.
1356 After selecting an appropriate directory, press @code{Forward}.
1358 A window will appear with the title
1359 @code{Version Control System Configuration}.
1360 Simply press @code{Forward}.
1362 A window will appear with the title
1363 @code{Please select the source directories for this project}.
1364 The directory that you specified for the project file will be selected
1365 by default as the one to use for sources; simply press @code{Forward}.
1367 A window will appear with the title
1368 @code{Please select the build directory for this project}.
1369 The directory that you specified for the project file will be selected
1370 by default for object files and executables;
1371 simply press @code{Forward}.
1373 A window will appear with the title
1374 @code{Please select the main units for this project}.
1375 You will supply this information later, after creating the source file.
1376 Simply press @code{Forward} for now.
1378 A window will appear with the title
1379 @code{Please select the switches to build the project}.
1380 Press @code{Apply}. This will create a project file named
1381 @file{sample.prj} in the directory that you had specified.
1383 @item @emph{Creating and saving the source file}
1385 After you create the new project, a GPS window will appear, which is
1386 partitioned into two main sections:
1390 A @emph{Workspace area}, initially greyed out, which you will use for
1391 creating and editing source files
1394 Directly below, a @emph{Messages area}, which initially displays a
1395 ``Welcome'' message.
1396 (If the Messages area is not visible, drag its border upward to expand it.)
1400 Select @code{File} on the menu bar, and then the @code{New} command.
1401 The Workspace area will become white, and you can now
1402 enter the source program explicitly.
1403 Type the following text
1405 @smallexample @c ada
1407 with Ada.Text_IO; use Ada.Text_IO;
1410 Put_Line("Hello from GPS!");
1416 Select @code{File}, then @code{Save As}, and enter the source file name
1418 The file will be saved in the same directory you specified as the
1419 location of the default project file.
1422 @item @emph{Updating the project file}
1424 You need to add the new source file to the project.
1426 the @code{Project} menu and then @code{Edit project properties}.
1427 Click the @code{Main files} tab on the left, and then the
1429 Choose @file{hello.adb} from the list, and press @code{Open}.
1430 The project settings window will reflect this action.
1433 @item @emph{Building and running the program}
1435 In the main GPS window, now choose the @code{Build} menu, then @code{Make},
1436 and select @file{hello.adb}.
1437 The Messages window will display the resulting invocations of @command{gcc},
1438 @command{gnatbind}, and @command{gnatlink}
1439 (reflecting the default switch settings from the
1440 project file that you created) and then a ``successful compilation/build''
1443 To run the program, choose the @code{Build} menu, then @code{Run}, and
1444 select @command{hello}.
1445 An @emph{Arguments Selection} window will appear.
1446 There are no command line arguments, so just click @code{OK}.
1448 The Messages window will now display the program's output (the string
1449 @code{Hello from GPS}), and at the bottom of the GPS window a status
1450 update is displayed (@code{Run: hello}).
1451 Close the GPS window (or select @code{File}, then @code{Exit}) to
1452 terminate this GPS session.
1457 @node Simple Debugging with GPS
1458 @subsection Simple Debugging with GPS
1460 This section illustrates basic debugging techniques (setting breakpoints,
1461 examining/modifying variables, single stepping).
1464 @item @emph{Opening a project}
1466 Start GPS and select @code{Open existing project}; browse to
1467 specify the project file @file{sample.prj} that you had created in the
1470 @item @emph{Creating a source file}
1472 Select @code{File}, then @code{New}, and type in the following program:
1474 @smallexample @c ada
1476 with Ada.Text_IO; use Ada.Text_IO;
1477 procedure Example is
1478 Line : String (1..80);
1481 Put_Line("Type a line of text at each prompt; an empty line to exit");
1485 Put_Line (Line (1..N) );
1493 Select @code{File}, then @code{Save as}, and enter the file name
1496 @item @emph{Updating the project file}
1498 Add @code{Example} as a new main unit for the project:
1501 Select @code{Project}, then @code{Edit Project Properties}.
1504 Select the @code{Main files} tab, click @code{Add}, then
1505 select the file @file{example.adb} from the list, and
1507 You will see the file name appear in the list of main units
1513 @item @emph{Building/running the executable}
1515 To build the executable
1516 select @code{Build}, then @code{Make}, and then choose @file{example.adb}.
1518 Run the program to see its effect (in the Messages area).
1519 Each line that you enter is displayed; an empty line will
1520 cause the loop to exit and the program to terminate.
1522 @item @emph{Debugging the program}
1524 Note that the @option{-g} switches to @command{gcc} and @command{gnatlink},
1525 which are required for debugging, are on by default when you create
1527 Thus unless you intentionally remove these settings, you will be able
1528 to debug any program that you develop using GPS.
1531 @item @emph{Initializing}
1533 Select @code{Debug}, then @code{Initialize}, then @file{example}
1535 @item @emph{Setting a breakpoint}
1537 After performing the initialization step, you will observe a small
1538 icon to the right of each line number.
1539 This serves as a toggle for breakpoints; clicking the icon will
1540 set a breakpoint at the corresponding line (the icon will change to
1541 a red circle with an ``x''), and clicking it again
1542 will remove the breakpoint / reset the icon.
1544 For purposes of this example, set a breakpoint at line 10 (the
1545 statement @code{Put_Line@ (Line@ (1..N));}
1547 @item @emph{Starting program execution}
1549 Select @code{Debug}, then @code{Run}. When the
1550 @code{Program Arguments} window appears, click @code{OK}.
1551 A console window will appear; enter some line of text,
1552 e.g. @code{abcde}, at the prompt.
1553 The program will pause execution when it gets to the
1554 breakpoint, and the corresponding line is highlighted.
1556 @item @emph{Examining a variable}
1558 Move the mouse over one of the occurrences of the variable @code{N}.
1559 You will see the value (5) displayed, in ``tool tip'' fashion.
1560 Right click on @code{N}, select @code{Debug}, then select @code{Display N}.
1561 You will see information about @code{N} appear in the @code{Debugger Data}
1562 pane, showing the value as 5.
1565 @item @emph{Assigning a new value to a variable}
1567 Right click on the @code{N} in the @code{Debugger Data} pane, and
1568 select @code{Set value of N}.
1569 When the input window appears, enter the value @code{4} and click
1571 This value does not automatically appear in the @code{Debugger Data}
1572 pane; to see it, right click again on the @code{N} in the
1573 @code{Debugger Data} pane and select @code{Update value}.
1574 The new value, 4, will appear in red.
1576 @item @emph{Single stepping}
1578 Select @code{Debug}, then @code{Next}.
1579 This will cause the next statement to be executed, in this case the
1580 call of @code{Put_Line} with the string slice.
1581 Notice in the console window that the displayed string is simply
1582 @code{abcd} and not @code{abcde} which you had entered.
1583 This is because the upper bound of the slice is now 4 rather than 5.
1585 @item @emph{Removing a breakpoint}
1587 Toggle the breakpoint icon at line 10.
1589 @item @emph{Resuming execution from a breakpoint}
1591 Select @code{Debug}, then @code{Continue}.
1592 The program will reach the next iteration of the loop, and
1593 wait for input after displaying the prompt.
1594 This time, just hit the @kbd{Enter} key.
1595 The value of @code{N} will be 0, and the program will terminate.
1596 The console window will disappear.
1601 @node Introduction to Glide and GVD
1602 @section Introduction to Glide and GVD
1606 This section describes the main features of Glide,
1607 a GNAT graphical IDE, and also shows how to use the basic commands in GVD,
1608 the GNU Visual Debugger.
1609 These tools may be present in addition to, or in place of, GPS on some
1611 Additional information on Glide and GVD may be found
1612 in the on-line help for these tools.
1615 * Building a New Program with Glide::
1616 * Simple Debugging with GVD::
1617 * Other Glide Features::
1620 @node Building a New Program with Glide
1621 @subsection Building a New Program with Glide
1623 The simplest way to invoke Glide is to enter @command{glide}
1624 at the command prompt. It will generally be useful to issue this
1625 as a background command, thus allowing you to continue using
1626 your command window for other purposes while Glide is running:
1633 Glide will start up with an initial screen displaying the top-level menu items
1634 as well as some other information. The menu selections are as follows
1636 @item @code{Buffers}
1647 For this introductory example, you will need to create a new Ada source file.
1648 First, select the @code{Files} menu. This will pop open a menu with around
1649 a dozen or so items. To create a file, select the @code{Open file...} choice.
1650 Depending on the platform, you may see a pop-up window where you can browse
1651 to an appropriate directory and then enter the file name, or else simply
1652 see a line at the bottom of the Glide window where you can likewise enter
1653 the file name. Note that in Glide, when you attempt to open a non-existent
1654 file, the effect is to create a file with that name. For this example enter
1655 @file{hello.adb} as the name of the file.
1657 A new buffer will now appear, occupying the entire Glide window,
1658 with the file name at the top. The menu selections are slightly different
1659 from the ones you saw on the opening screen; there is an @code{Entities} item,
1660 and in place of @code{Glide} there is now an @code{Ada} item. Glide uses
1661 the file extension to identify the source language, so @file{adb} indicates
1664 You will enter some of the source program lines explicitly,
1665 and use the syntax-oriented template mechanism to enter other lines.
1666 First, type the following text:
1668 with Ada.Text_IO; use Ada.Text_IO;
1674 Observe that Glide uses different colors to distinguish reserved words from
1675 identifiers. Also, after the @code{procedure Hello is} line, the cursor is
1676 automatically indented in anticipation of declarations. When you enter
1677 @code{begin}, Glide recognizes that there are no declarations and thus places
1678 @code{begin} flush left. But after the @code{begin} line the cursor is again
1679 indented, where the statement(s) will be placed.
1681 The main part of the program will be a @code{for} loop. Instead of entering
1682 the text explicitly, however, use a statement template. Select the @code{Ada}
1683 item on the top menu bar, move the mouse to the @code{Statements} item,
1684 and you will see a large selection of alternatives. Choose @code{for loop}.
1685 You will be prompted (at the bottom of the buffer) for a loop name;
1686 simply press the @key{Enter} key since a loop name is not needed.
1687 You should see the beginning of a @code{for} loop appear in the source
1688 program window. You will now be prompted for the name of the loop variable;
1689 enter a line with the identifier @code{ind} (lower case). Note that,
1690 by default, Glide capitalizes the name (you can override such behavior
1691 if you wish, although this is outside the scope of this introduction).
1692 Next, Glide prompts you for the loop range; enter a line containing
1693 @code{1..5} and you will see this also appear in the source program,
1694 together with the remaining elements of the @code{for} loop syntax.
1696 Next enter the statement (with an intentional error, a missing semicolon)
1697 that will form the body of the loop:
1699 Put_Line("Hello, World" & Integer'Image(I))
1703 Finally, type @code{end Hello;} as the last line in the program.
1704 Now save the file: choose the @code{File} menu item, and then the
1705 @code{Save buffer} selection. You will see a message at the bottom
1706 of the buffer confirming that the file has been saved.
1708 You are now ready to attempt to build the program. Select the @code{Ada}
1709 item from the top menu bar. Although we could choose simply to compile
1710 the file, we will instead attempt to do a build (which invokes
1711 @command{gnatmake}) since, if the compile is successful, we want to build
1712 an executable. Thus select @code{Ada build}. This will fail because of the
1713 compilation error, and you will notice that the Glide window has been split:
1714 the top window contains the source file, and the bottom window contains the
1715 output from the GNAT tools. Glide allows you to navigate from a compilation
1716 error to the source file position corresponding to the error: click the
1717 middle mouse button (or simultaneously press the left and right buttons,
1718 on a two-button mouse) on the diagnostic line in the tool window. The
1719 focus will shift to the source window, and the cursor will be positioned
1720 on the character at which the error was detected.
1722 Correct the error: type in a semicolon to terminate the statement.
1723 Although you can again save the file explicitly, you can also simply invoke
1724 @code{Ada} @result{} @code{Build} and you will be prompted to save the file.
1725 This time the build will succeed; the tool output window shows you the
1726 options that are supplied by default. The GNAT tools' output (e.g.
1727 object and ALI files, executable) will go in the directory from which
1730 To execute the program, choose @code{Ada} and then @code{Run}.
1731 You should see the program's output displayed in the bottom window:
1741 @node Simple Debugging with GVD
1742 @subsection Simple Debugging with GVD
1745 This section describes how to set breakpoints, examine/modify variables,
1746 and step through execution.
1748 In order to enable debugging, you need to pass the @option{-g} switch
1749 to both the compiler and to @command{gnatlink}. If you are using
1750 the command line, passing @option{-g} to @command{gnatmake} will have
1751 this effect. You can then launch GVD, e.g. on the @code{hello} program,
1752 by issuing the command:
1759 If you are using Glide, then @option{-g} is passed to the relevant tools
1760 by default when you do a build. Start the debugger by selecting the
1761 @code{Ada} menu item, and then @code{Debug}.
1763 GVD comes up in a multi-part window. One pane shows the names of files
1764 comprising your executable; another pane shows the source code of the current
1765 unit (initially your main subprogram), another pane shows the debugger output
1766 and user interactions, and the fourth pane (the data canvas at the top
1767 of the window) displays data objects that you have selected.
1769 To the left of the source file pane, you will notice green dots adjacent
1770 to some lines. These are lines for which object code exists and where
1771 breakpoints can thus be set. You set/reset a breakpoint by clicking
1772 the green dot. When a breakpoint is set, the dot is replaced by an @code{X}
1773 in a red circle. Clicking the circle toggles the breakpoint off,
1774 and the red circle is replaced by the green dot.
1776 For this example, set a breakpoint at the statement where @code{Put_Line}
1779 Start program execution by selecting the @code{Run} button on the top menu bar.
1780 (The @code{Start} button will also start your program, but it will
1781 cause program execution to break at the entry to your main subprogram.)
1782 Evidence of reaching the breakpoint will appear: the source file line will be
1783 highlighted, and the debugger interactions pane will display
1786 You can examine the values of variables in several ways. Move the mouse
1787 over an occurrence of @code{Ind} in the @code{for} loop, and you will see
1788 the value (now @code{1}) displayed. Alternatively, right-click on @code{Ind}
1789 and select @code{Display Ind}; a box showing the variable's name and value
1790 will appear in the data canvas.
1792 Although a loop index is a constant with respect to Ada semantics,
1793 you can change its value in the debugger. Right-click in the box
1794 for @code{Ind}, and select the @code{Set Value of Ind} item.
1795 Enter @code{2} as the new value, and press @command{OK}.
1796 The box for @code{Ind} shows the update.
1798 Press the @code{Step} button on the top menu bar; this will step through
1799 one line of program text (the invocation of @code{Put_Line}), and you can
1800 observe the effect of having modified @code{Ind} since the value displayed
1803 Remove the breakpoint, and resume execution by selecting the @code{Cont}
1804 button. You will see the remaining output lines displayed in the debugger
1805 interaction window, along with a message confirming normal program
1808 @node Other Glide Features
1809 @subsection Other Glide Features
1812 You may have observed that some of the menu selections contain abbreviations;
1813 e.g., @code{(C-x C-f)} for @code{Open file...} in the @code{Files} menu.
1814 These are @emph{shortcut keys} that you can use instead of selecting
1815 menu items. The @key{C} stands for @key{Ctrl}; thus @code{(C-x C-f)} means
1816 @key{Ctrl-x} followed by @key{Ctrl-f}, and this sequence can be used instead
1817 of selecting @code{Files} and then @code{Open file...}.
1819 To abort a Glide command, type @key{Ctrl-g}.
1821 If you want Glide to start with an existing source file, you can either
1822 launch Glide as above and then open the file via @code{Files} @result{}
1823 @code{Open file...}, or else simply pass the name of the source file
1824 on the command line:
1831 While you are using Glide, a number of @emph{buffers} exist.
1832 You create some explicitly; e.g., when you open/create a file.
1833 Others arise as an effect of the commands that you issue; e.g., the buffer
1834 containing the output of the tools invoked during a build. If a buffer
1835 is hidden, you can bring it into a visible window by first opening
1836 the @code{Buffers} menu and then selecting the desired entry.
1838 If a buffer occupies only part of the Glide screen and you want to expand it
1839 to fill the entire screen, then click in the buffer and then select
1840 @code{Files} @result{} @code{One Window}.
1842 If a window is occupied by one buffer and you want to split the window
1843 to bring up a second buffer, perform the following steps:
1845 @item Select @code{Files} @result{} @code{Split Window};
1846 this will produce two windows each of which holds the original buffer
1847 (these are not copies, but rather different views of the same buffer contents)
1849 @item With the focus in one of the windows,
1850 select the desired buffer from the @code{Buffers} menu
1854 To exit from Glide, choose @code{Files} @result{} @code{Exit}.
1857 @node The GNAT Compilation Model
1858 @chapter The GNAT Compilation Model
1859 @cindex GNAT compilation model
1860 @cindex Compilation model
1863 * Source Representation::
1864 * Foreign Language Representation::
1865 * File Naming Rules::
1866 * Using Other File Names::
1867 * Alternative File Naming Schemes::
1868 * Generating Object Files::
1869 * Source Dependencies::
1870 * The Ada Library Information Files::
1871 * Binding an Ada Program::
1872 * Mixed Language Programming::
1873 * Building Mixed Ada & C++ Programs::
1874 * Comparison between GNAT and C/C++ Compilation Models::
1875 * Comparison between GNAT and Conventional Ada Library Models::
1877 * Placement of temporary files::
1882 This chapter describes the compilation model used by GNAT. Although
1883 similar to that used by other languages, such as C and C++, this model
1884 is substantially different from the traditional Ada compilation models,
1885 which are based on a library. The model is initially described without
1886 reference to the library-based model. If you have not previously used an
1887 Ada compiler, you need only read the first part of this chapter. The
1888 last section describes and discusses the differences between the GNAT
1889 model and the traditional Ada compiler models. If you have used other
1890 Ada compilers, this section will help you to understand those
1891 differences, and the advantages of the GNAT model.
1893 @node Source Representation
1894 @section Source Representation
1898 Ada source programs are represented in standard text files, using
1899 Latin-1 coding. Latin-1 is an 8-bit code that includes the familiar
1900 7-bit ASCII set, plus additional characters used for
1901 representing foreign languages (@pxref{Foreign Language Representation}
1902 for support of non-USA character sets). The format effector characters
1903 are represented using their standard ASCII encodings, as follows:
1908 Vertical tab, @code{16#0B#}
1912 Horizontal tab, @code{16#09#}
1916 Carriage return, @code{16#0D#}
1920 Line feed, @code{16#0A#}
1924 Form feed, @code{16#0C#}
1928 Source files are in standard text file format. In addition, GNAT will
1929 recognize a wide variety of stream formats, in which the end of physical
1930 physical lines is marked by any of the following sequences:
1931 @code{LF}, @code{CR}, @code{CR-LF}, or @code{LF-CR}. This is useful
1932 in accommodating files that are imported from other operating systems.
1934 @cindex End of source file
1935 @cindex Source file, end
1937 The end of a source file is normally represented by the physical end of
1938 file. However, the control character @code{16#1A#} (@code{SUB}) is also
1939 recognized as signalling the end of the source file. Again, this is
1940 provided for compatibility with other operating systems where this
1941 code is used to represent the end of file.
1943 Each file contains a single Ada compilation unit, including any pragmas
1944 associated with the unit. For example, this means you must place a
1945 package declaration (a package @dfn{spec}) and the corresponding body in
1946 separate files. An Ada @dfn{compilation} (which is a sequence of
1947 compilation units) is represented using a sequence of files. Similarly,
1948 you will place each subunit or child unit in a separate file.
1950 @node Foreign Language Representation
1951 @section Foreign Language Representation
1954 GNAT supports the standard character sets defined in Ada 95 as well as
1955 several other non-standard character sets for use in localized versions
1956 of the compiler (@pxref{Character Set Control}).
1959 * Other 8-Bit Codes::
1960 * Wide Character Encodings::
1968 The basic character set is Latin-1. This character set is defined by ISO
1969 standard 8859, part 1. The lower half (character codes @code{16#00#}
1970 ... @code{16#7F#)} is identical to standard ASCII coding, but the upper half
1971 is used to represent additional characters. These include extended letters
1972 used by European languages, such as French accents, the vowels with umlauts
1973 used in German, and the extra letter A-ring used in Swedish.
1975 @findex Ada.Characters.Latin_1
1976 For a complete list of Latin-1 codes and their encodings, see the source
1977 file of library unit @code{Ada.Characters.Latin_1} in file
1978 @file{a-chlat1.ads}.
1979 You may use any of these extended characters freely in character or
1980 string literals. In addition, the extended characters that represent
1981 letters can be used in identifiers.
1983 @node Other 8-Bit Codes
1984 @subsection Other 8-Bit Codes
1987 GNAT also supports several other 8-bit coding schemes:
1990 @item ISO 8859-2 (Latin-2)
1993 Latin-2 letters allowed in identifiers, with uppercase and lowercase
1996 @item ISO 8859-3 (Latin-3)
1999 Latin-3 letters allowed in identifiers, with uppercase and lowercase
2002 @item ISO 8859-4 (Latin-4)
2005 Latin-4 letters allowed in identifiers, with uppercase and lowercase
2008 @item ISO 8859-5 (Cyrillic)
2011 ISO 8859-5 letters (Cyrillic) allowed in identifiers, with uppercase and
2012 lowercase equivalence.
2014 @item ISO 8859-15 (Latin-9)
2017 ISO 8859-15 (Latin-9) letters allowed in identifiers, with uppercase and
2018 lowercase equivalence
2020 @item IBM PC (code page 437)
2021 @cindex code page 437
2022 This code page is the normal default for PCs in the U.S. It corresponds
2023 to the original IBM PC character set. This set has some, but not all, of
2024 the extended Latin-1 letters, but these letters do not have the same
2025 encoding as Latin-1. In this mode, these letters are allowed in
2026 identifiers with uppercase and lowercase equivalence.
2028 @item IBM PC (code page 850)
2029 @cindex code page 850
2030 This code page is a modification of 437 extended to include all the
2031 Latin-1 letters, but still not with the usual Latin-1 encoding. In this
2032 mode, all these letters are allowed in identifiers with uppercase and
2033 lowercase equivalence.
2035 @item Full Upper 8-bit
2036 Any character in the range 80-FF allowed in identifiers, and all are
2037 considered distinct. In other words, there are no uppercase and lowercase
2038 equivalences in this range. This is useful in conjunction with
2039 certain encoding schemes used for some foreign character sets (e.g.
2040 the typical method of representing Chinese characters on the PC).
2043 No upper-half characters in the range 80-FF are allowed in identifiers.
2044 This gives Ada 83 compatibility for identifier names.
2048 For precise data on the encodings permitted, and the uppercase and lowercase
2049 equivalences that are recognized, see the file @file{csets.adb} in
2050 the GNAT compiler sources. You will need to obtain a full source release
2051 of GNAT to obtain this file.
2053 @node Wide Character Encodings
2054 @subsection Wide Character Encodings
2057 GNAT allows wide character codes to appear in character and string
2058 literals, and also optionally in identifiers, by means of the following
2059 possible encoding schemes:
2064 In this encoding, a wide character is represented by the following five
2072 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
2073 characters (using uppercase letters) of the wide character code. For
2074 example, ESC A345 is used to represent the wide character with code
2076 This scheme is compatible with use of the full Wide_Character set.
2078 @item Upper-Half Coding
2079 @cindex Upper-Half Coding
2080 The wide character with encoding @code{16#abcd#} where the upper bit is on
2081 (in other words, ``a'' is in the range 8-F) is represented as two bytes,
2082 @code{16#ab#} and @code{16#cd#}. The second byte cannot be a format control
2083 character, but is not required to be in the upper half. This method can
2084 be also used for shift-JIS or EUC, where the internal coding matches the
2087 @item Shift JIS Coding
2088 @cindex Shift JIS Coding
2089 A wide character is represented by a two-character sequence,
2091 @code{16#cd#}, with the restrictions described for upper-half encoding as
2092 described above. The internal character code is the corresponding JIS
2093 character according to the standard algorithm for Shift-JIS
2094 conversion. Only characters defined in the JIS code set table can be
2095 used with this encoding method.
2099 A wide character is represented by a two-character sequence
2101 @code{16#cd#}, with both characters being in the upper half. The internal
2102 character code is the corresponding JIS character according to the EUC
2103 encoding algorithm. Only characters defined in the JIS code set table
2104 can be used with this encoding method.
2107 A wide character is represented using
2108 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
2109 10646-1/Am.2. Depending on the character value, the representation
2110 is a one, two, or three byte sequence:
2115 16#0000#-16#007f#: 2#0xxxxxxx#
2116 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
2117 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
2122 where the xxx bits correspond to the left-padded bits of the
2123 16-bit character value. Note that all lower half ASCII characters
2124 are represented as ASCII bytes and all upper half characters and
2125 other wide characters are represented as sequences of upper-half
2126 (The full UTF-8 scheme allows for encoding 31-bit characters as
2127 6-byte sequences, but in this implementation, all UTF-8 sequences
2128 of four or more bytes length will be treated as illegal).
2129 @item Brackets Coding
2130 In this encoding, a wide character is represented by the following eight
2138 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
2139 characters (using uppercase letters) of the wide character code. For
2140 example, [``A345''] is used to represent the wide character with code
2141 @code{16#A345#}. It is also possible (though not required) to use the
2142 Brackets coding for upper half characters. For example, the code
2143 @code{16#A3#} can be represented as @code{[``A3'']}.
2145 This scheme is compatible with use of the full Wide_Character set,
2146 and is also the method used for wide character encoding in the standard
2147 ACVC (Ada Compiler Validation Capability) test suite distributions.
2152 Note: Some of these coding schemes do not permit the full use of the
2153 Ada 95 character set. For example, neither Shift JIS, nor EUC allow the
2154 use of the upper half of the Latin-1 set.
2156 @node File Naming Rules
2157 @section File Naming Rules
2160 The default file name is determined by the name of the unit that the
2161 file contains. The name is formed by taking the full expanded name of
2162 the unit and replacing the separating dots with hyphens and using
2163 ^lowercase^uppercase^ for all letters.
2165 An exception arises if the file name generated by the above rules starts
2166 with one of the characters
2173 and the second character is a
2174 minus. In this case, the character ^tilde^dollar sign^ is used in place
2175 of the minus. The reason for this special rule is to avoid clashes with
2176 the standard names for child units of the packages System, Ada,
2177 Interfaces, and GNAT, which use the prefixes
2186 The file extension is @file{.ads} for a spec and
2187 @file{.adb} for a body. The following list shows some
2188 examples of these rules.
2195 @item arith_functions.ads
2196 Arith_Functions (package spec)
2197 @item arith_functions.adb
2198 Arith_Functions (package body)
2200 Func.Spec (child package spec)
2202 Func.Spec (child package body)
2204 Sub (subunit of Main)
2205 @item ^a~bad.adb^A$BAD.ADB^
2206 A.Bad (child package body)
2210 Following these rules can result in excessively long
2211 file names if corresponding
2212 unit names are long (for example, if child units or subunits are
2213 heavily nested). An option is available to shorten such long file names
2214 (called file name ``krunching''). This may be particularly useful when
2215 programs being developed with GNAT are to be used on operating systems
2216 with limited file name lengths. @xref{Using gnatkr}.
2218 Of course, no file shortening algorithm can guarantee uniqueness over
2219 all possible unit names; if file name krunching is used, it is your
2220 responsibility to ensure no name clashes occur. Alternatively you
2221 can specify the exact file names that you want used, as described
2222 in the next section. Finally, if your Ada programs are migrating from a
2223 compiler with a different naming convention, you can use the gnatchop
2224 utility to produce source files that follow the GNAT naming conventions.
2225 (For details @pxref{Renaming Files Using gnatchop}.)
2227 Note: in the case of @code{Windows NT/XP} or @code{OpenVMS} operating
2228 systems, case is not significant. So for example on @code{Windows XP}
2229 if the canonical name is @code{main-sub.adb}, you can use the file name
2230 @code{Main-Sub.adb} instead. However, case is significant for other
2231 operating systems, so for example, if you want to use other than
2232 canonically cased file names on a Unix system, you need to follow
2233 the procedures described in the next section.
2235 @node Using Other File Names
2236 @section Using Other File Names
2240 In the previous section, we have described the default rules used by
2241 GNAT to determine the file name in which a given unit resides. It is
2242 often convenient to follow these default rules, and if you follow them,
2243 the compiler knows without being explicitly told where to find all
2246 However, in some cases, particularly when a program is imported from
2247 another Ada compiler environment, it may be more convenient for the
2248 programmer to specify which file names contain which units. GNAT allows
2249 arbitrary file names to be used by means of the Source_File_Name pragma.
2250 The form of this pragma is as shown in the following examples:
2251 @cindex Source_File_Name pragma
2253 @smallexample @c ada
2255 pragma Source_File_Name (My_Utilities.Stacks,
2256 Spec_File_Name => "myutilst_a.ada");
2257 pragma Source_File_name (My_Utilities.Stacks,
2258 Body_File_Name => "myutilst.ada");
2263 As shown in this example, the first argument for the pragma is the unit
2264 name (in this example a child unit). The second argument has the form
2265 of a named association. The identifier
2266 indicates whether the file name is for a spec or a body;
2267 the file name itself is given by a string literal.
2269 The source file name pragma is a configuration pragma, which means that
2270 normally it will be placed in the @file{gnat.adc}
2271 file used to hold configuration
2272 pragmas that apply to a complete compilation environment.
2273 For more details on how the @file{gnat.adc} file is created and used
2274 @pxref{Handling of Configuration Pragmas}
2275 @cindex @file{gnat.adc}
2278 GNAT allows completely arbitrary file names to be specified using the
2279 source file name pragma. However, if the file name specified has an
2280 extension other than @file{.ads} or @file{.adb} it is necessary to use
2281 a special syntax when compiling the file. The name in this case must be
2282 preceded by the special sequence @code{-x} followed by a space and the name
2283 of the language, here @code{ada}, as in:
2286 $ gcc -c -x ada peculiar_file_name.sim
2291 @code{gnatmake} handles non-standard file names in the usual manner (the
2292 non-standard file name for the main program is simply used as the
2293 argument to gnatmake). Note that if the extension is also non-standard,
2294 then it must be included in the gnatmake command, it may not be omitted.
2296 @node Alternative File Naming Schemes
2297 @section Alternative File Naming Schemes
2298 @cindex File naming schemes, alternative
2301 In the previous section, we described the use of the @code{Source_File_Name}
2302 pragma to allow arbitrary names to be assigned to individual source files.
2303 However, this approach requires one pragma for each file, and especially in
2304 large systems can result in very long @file{gnat.adc} files, and also create
2305 a maintenance problem.
2307 GNAT also provides a facility for specifying systematic file naming schemes
2308 other than the standard default naming scheme previously described. An
2309 alternative scheme for naming is specified by the use of
2310 @code{Source_File_Name} pragmas having the following format:
2311 @cindex Source_File_Name pragma
2313 @smallexample @c ada
2314 pragma Source_File_Name (
2315 Spec_File_Name => FILE_NAME_PATTERN
2316 [,Casing => CASING_SPEC]
2317 [,Dot_Replacement => STRING_LITERAL]);
2319 pragma Source_File_Name (
2320 Body_File_Name => FILE_NAME_PATTERN
2321 [,Casing => CASING_SPEC]
2322 [,Dot_Replacement => STRING_LITERAL]);
2324 pragma Source_File_Name (
2325 Subunit_File_Name => FILE_NAME_PATTERN
2326 [,Casing => CASING_SPEC]
2327 [,Dot_Replacement => STRING_LITERAL]);
2329 FILE_NAME_PATTERN ::= STRING_LITERAL
2330 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
2334 The @code{FILE_NAME_PATTERN} string shows how the file name is constructed.
2335 It contains a single asterisk character, and the unit name is substituted
2336 systematically for this asterisk. The optional parameter
2337 @code{Casing} indicates
2338 whether the unit name is to be all upper-case letters, all lower-case letters,
2339 or mixed-case. If no
2340 @code{Casing} parameter is used, then the default is all
2341 ^lower-case^upper-case^.
2343 The optional @code{Dot_Replacement} string is used to replace any periods
2344 that occur in subunit or child unit names. If no @code{Dot_Replacement}
2345 argument is used then separating dots appear unchanged in the resulting
2347 Although the above syntax indicates that the
2348 @code{Casing} argument must appear
2349 before the @code{Dot_Replacement} argument, but it
2350 is also permissible to write these arguments in the opposite order.
2352 As indicated, it is possible to specify different naming schemes for
2353 bodies, specs, and subunits. Quite often the rule for subunits is the
2354 same as the rule for bodies, in which case, there is no need to give
2355 a separate @code{Subunit_File_Name} rule, and in this case the
2356 @code{Body_File_name} rule is used for subunits as well.
2358 The separate rule for subunits can also be used to implement the rather
2359 unusual case of a compilation environment (e.g. a single directory) which
2360 contains a subunit and a child unit with the same unit name. Although
2361 both units cannot appear in the same partition, the Ada Reference Manual
2362 allows (but does not require) the possibility of the two units coexisting
2363 in the same environment.
2365 The file name translation works in the following steps:
2370 If there is a specific @code{Source_File_Name} pragma for the given unit,
2371 then this is always used, and any general pattern rules are ignored.
2374 If there is a pattern type @code{Source_File_Name} pragma that applies to
2375 the unit, then the resulting file name will be used if the file exists. If
2376 more than one pattern matches, the latest one will be tried first, and the
2377 first attempt resulting in a reference to a file that exists will be used.
2380 If no pattern type @code{Source_File_Name} pragma that applies to the unit
2381 for which the corresponding file exists, then the standard GNAT default
2382 naming rules are used.
2387 As an example of the use of this mechanism, consider a commonly used scheme
2388 in which file names are all lower case, with separating periods copied
2389 unchanged to the resulting file name, and specs end with @file{.1.ada}, and
2390 bodies end with @file{.2.ada}. GNAT will follow this scheme if the following
2393 @smallexample @c ada
2394 pragma Source_File_Name
2395 (Spec_File_Name => "*.1.ada");
2396 pragma Source_File_Name
2397 (Body_File_Name => "*.2.ada");
2401 The default GNAT scheme is actually implemented by providing the following
2402 default pragmas internally:
2404 @smallexample @c ada
2405 pragma Source_File_Name
2406 (Spec_File_Name => "*.ads", Dot_Replacement => "-");
2407 pragma Source_File_Name
2408 (Body_File_Name => "*.adb", Dot_Replacement => "-");
2412 Our final example implements a scheme typically used with one of the
2413 Ada 83 compilers, where the separator character for subunits was ``__''
2414 (two underscores), specs were identified by adding @file{_.ADA}, bodies
2415 by adding @file{.ADA}, and subunits by
2416 adding @file{.SEP}. All file names were
2417 upper case. Child units were not present of course since this was an
2418 Ada 83 compiler, but it seems reasonable to extend this scheme to use
2419 the same double underscore separator for child units.
2421 @smallexample @c ada
2422 pragma Source_File_Name
2423 (Spec_File_Name => "*_.ADA",
2424 Dot_Replacement => "__",
2425 Casing = Uppercase);
2426 pragma Source_File_Name
2427 (Body_File_Name => "*.ADA",
2428 Dot_Replacement => "__",
2429 Casing = Uppercase);
2430 pragma Source_File_Name
2431 (Subunit_File_Name => "*.SEP",
2432 Dot_Replacement => "__",
2433 Casing = Uppercase);
2436 @node Generating Object Files
2437 @section Generating Object Files
2440 An Ada program consists of a set of source files, and the first step in
2441 compiling the program is to generate the corresponding object files.
2442 These are generated by compiling a subset of these source files.
2443 The files you need to compile are the following:
2447 If a package spec has no body, compile the package spec to produce the
2448 object file for the package.
2451 If a package has both a spec and a body, compile the body to produce the
2452 object file for the package. The source file for the package spec need
2453 not be compiled in this case because there is only one object file, which
2454 contains the code for both the spec and body of the package.
2457 For a subprogram, compile the subprogram body to produce the object file
2458 for the subprogram. The spec, if one is present, is as usual in a
2459 separate file, and need not be compiled.
2463 In the case of subunits, only compile the parent unit. A single object
2464 file is generated for the entire subunit tree, which includes all the
2468 Compile child units independently of their parent units
2469 (though, of course, the spec of all the ancestor unit must be present in order
2470 to compile a child unit).
2474 Compile generic units in the same manner as any other units. The object
2475 files in this case are small dummy files that contain at most the
2476 flag used for elaboration checking. This is because GNAT always handles generic
2477 instantiation by means of macro expansion. However, it is still necessary to
2478 compile generic units, for dependency checking and elaboration purposes.
2482 The preceding rules describe the set of files that must be compiled to
2483 generate the object files for a program. Each object file has the same
2484 name as the corresponding source file, except that the extension is
2487 You may wish to compile other files for the purpose of checking their
2488 syntactic and semantic correctness. For example, in the case where a
2489 package has a separate spec and body, you would not normally compile the
2490 spec. However, it is convenient in practice to compile the spec to make
2491 sure it is error-free before compiling clients of this spec, because such
2492 compilations will fail if there is an error in the spec.
2494 GNAT provides an option for compiling such files purely for the
2495 purposes of checking correctness; such compilations are not required as
2496 part of the process of building a program. To compile a file in this
2497 checking mode, use the @option{-gnatc} switch.
2499 @node Source Dependencies
2500 @section Source Dependencies
2503 A given object file clearly depends on the source file which is compiled
2504 to produce it. Here we are using @dfn{depends} in the sense of a typical
2505 @code{make} utility; in other words, an object file depends on a source
2506 file if changes to the source file require the object file to be
2508 In addition to this basic dependency, a given object may depend on
2509 additional source files as follows:
2513 If a file being compiled @code{with}'s a unit @var{X}, the object file
2514 depends on the file containing the spec of unit @var{X}. This includes
2515 files that are @code{with}'ed implicitly either because they are parents
2516 of @code{with}'ed child units or they are run-time units required by the
2517 language constructs used in a particular unit.
2520 If a file being compiled instantiates a library level generic unit, the
2521 object file depends on both the spec and body files for this generic
2525 If a file being compiled instantiates a generic unit defined within a
2526 package, the object file depends on the body file for the package as
2527 well as the spec file.
2531 @cindex @option{-gnatn} switch
2532 If a file being compiled contains a call to a subprogram for which
2533 pragma @code{Inline} applies and inlining is activated with the
2534 @option{-gnatn} switch, the object file depends on the file containing the
2535 body of this subprogram as well as on the file containing the spec. Note
2536 that for inlining to actually occur as a result of the use of this switch,
2537 it is necessary to compile in optimizing mode.
2539 @cindex @option{-gnatN} switch
2540 The use of @option{-gnatN} activates a more extensive inlining optimization
2541 that is performed by the front end of the compiler. This inlining does
2542 not require that the code generation be optimized. Like @option{-gnatn},
2543 the use of this switch generates additional dependencies.
2545 @option{-gnatN} automatically implies @option{-gnatn} so it is not necessary
2546 to specify both options.
2549 If an object file O depends on the proper body of a subunit through inlining
2550 or instantiation, it depends on the parent unit of the subunit. This means that
2551 any modification of the parent unit or one of its subunits affects the
2555 The object file for a parent unit depends on all its subunit body files.
2558 The previous two rules meant that for purposes of computing dependencies and
2559 recompilation, a body and all its subunits are treated as an indivisible whole.
2562 These rules are applied transitively: if unit @code{A} @code{with}'s
2563 unit @code{B}, whose elaboration calls an inlined procedure in package
2564 @code{C}, the object file for unit @code{A} will depend on the body of
2565 @code{C}, in file @file{c.adb}.
2567 The set of dependent files described by these rules includes all the
2568 files on which the unit is semantically dependent, as described in the
2569 Ada 95 Language Reference Manual. However, it is a superset of what the
2570 ARM describes, because it includes generic, inline, and subunit dependencies.
2572 An object file must be recreated by recompiling the corresponding source
2573 file if any of the source files on which it depends are modified. For
2574 example, if the @code{make} utility is used to control compilation,
2575 the rule for an Ada object file must mention all the source files on
2576 which the object file depends, according to the above definition.
2577 The determination of the necessary
2578 recompilations is done automatically when one uses @code{gnatmake}.
2581 @node The Ada Library Information Files
2582 @section The Ada Library Information Files
2583 @cindex Ada Library Information files
2584 @cindex @file{ALI} files
2587 Each compilation actually generates two output files. The first of these
2588 is the normal object file that has a @file{.o} extension. The second is a
2589 text file containing full dependency information. It has the same
2590 name as the source file, but an @file{.ali} extension.
2591 This file is known as the Ada Library Information (@file{ALI}) file.
2592 The following information is contained in the @file{ALI} file.
2596 Version information (indicates which version of GNAT was used to compile
2597 the unit(s) in question)
2600 Main program information (including priority and time slice settings,
2601 as well as the wide character encoding used during compilation).
2604 List of arguments used in the @code{gcc} command for the compilation
2607 Attributes of the unit, including configuration pragmas used, an indication
2608 of whether the compilation was successful, exception model used etc.
2611 A list of relevant restrictions applying to the unit (used for consistency)
2615 Categorization information (e.g. use of pragma @code{Pure}).
2618 Information on all @code{with}'ed units, including presence of
2619 @code{Elaborate} or @code{Elaborate_All} pragmas.
2622 Information from any @code{Linker_Options} pragmas used in the unit
2625 Information on the use of @code{Body_Version} or @code{Version}
2626 attributes in the unit.
2629 Dependency information. This is a list of files, together with
2630 time stamp and checksum information. These are files on which
2631 the unit depends in the sense that recompilation is required
2632 if any of these units are modified.
2635 Cross-reference data. Contains information on all entities referenced
2636 in the unit. Used by tools like @code{gnatxref} and @code{gnatfind} to
2637 provide cross-reference information.
2642 For a full detailed description of the format of the @file{ALI} file,
2643 see the source of the body of unit @code{Lib.Writ}, contained in file
2644 @file{lib-writ.adb} in the GNAT compiler sources.
2646 @node Binding an Ada Program
2647 @section Binding an Ada Program
2650 When using languages such as C and C++, once the source files have been
2651 compiled the only remaining step in building an executable program
2652 is linking the object modules together. This means that it is possible to
2653 link an inconsistent version of a program, in which two units have
2654 included different versions of the same header.
2656 The rules of Ada do not permit such an inconsistent program to be built.
2657 For example, if two clients have different versions of the same package,
2658 it is illegal to build a program containing these two clients.
2659 These rules are enforced by the GNAT binder, which also determines an
2660 elaboration order consistent with the Ada rules.
2662 The GNAT binder is run after all the object files for a program have
2663 been created. It is given the name of the main program unit, and from
2664 this it determines the set of units required by the program, by reading the
2665 corresponding ALI files. It generates error messages if the program is
2666 inconsistent or if no valid order of elaboration exists.
2668 If no errors are detected, the binder produces a main program, in Ada by
2669 default, that contains calls to the elaboration procedures of those
2670 compilation unit that require them, followed by
2671 a call to the main program. This Ada program is compiled to generate the
2672 object file for the main program. The name of
2673 the Ada file is @file{b~@var{xxx}.adb} (with the corresponding spec
2674 @file{b~@var{xxx}.ads}) where @var{xxx} is the name of the
2677 Finally, the linker is used to build the resulting executable program,
2678 using the object from the main program from the bind step as well as the
2679 object files for the Ada units of the program.
2681 @node Mixed Language Programming
2682 @section Mixed Language Programming
2683 @cindex Mixed Language Programming
2686 This section describes how to develop a mixed-language program,
2687 specifically one that comprises units in both Ada and C.
2690 * Interfacing to C::
2691 * Calling Conventions::
2694 @node Interfacing to C
2695 @subsection Interfacing to C
2697 Interfacing Ada with a foreign language such as C involves using
2698 compiler directives to import and/or export entity definitions in each
2699 language---using @code{extern} statements in C, for instance, and the
2700 @code{Import}, @code{Export}, and @code{Convention} pragmas in Ada. For
2701 a full treatment of these topics, read Appendix B, section 1 of the Ada
2702 95 Language Reference Manual.
2704 There are two ways to build a program using GNAT that contains some Ada
2705 sources and some foreign language sources, depending on whether or not
2706 the main subprogram is written in Ada. Here is a source example with
2707 the main subprogram in Ada:
2713 void print_num (int num)
2715 printf ("num is %d.\n", num);
2721 /* num_from_Ada is declared in my_main.adb */
2722 extern int num_from_Ada;
2726 return num_from_Ada;
2730 @smallexample @c ada
2732 procedure My_Main is
2734 -- Declare then export an Integer entity called num_from_Ada
2735 My_Num : Integer := 10;
2736 pragma Export (C, My_Num, "num_from_Ada");
2738 -- Declare an Ada function spec for Get_Num, then use
2739 -- C function get_num for the implementation.
2740 function Get_Num return Integer;
2741 pragma Import (C, Get_Num, "get_num");
2743 -- Declare an Ada procedure spec for Print_Num, then use
2744 -- C function print_num for the implementation.
2745 procedure Print_Num (Num : Integer);
2746 pragma Import (C, Print_Num, "print_num");
2749 Print_Num (Get_Num);
2755 To build this example, first compile the foreign language files to
2756 generate object files:
2763 Then, compile the Ada units to produce a set of object files and ALI
2766 gnatmake ^-c^/ACTIONS=COMPILE^ my_main.adb
2770 Run the Ada binder on the Ada main program:
2772 gnatbind my_main.ali
2776 Link the Ada main program, the Ada objects and the other language
2779 gnatlink my_main.ali file1.o file2.o
2783 The last three steps can be grouped in a single command:
2785 gnatmake my_main.adb -largs file1.o file2.o
2788 @cindex Binder output file
2790 If the main program is in a language other than Ada, then you may have
2791 more than one entry point into the Ada subsystem. You must use a special
2792 binder option to generate callable routines that initialize and
2793 finalize the Ada units (@pxref{Binding with Non-Ada Main Programs}).
2794 Calls to the initialization and finalization routines must be inserted
2795 in the main program, or some other appropriate point in the code. The
2796 call to initialize the Ada units must occur before the first Ada
2797 subprogram is called, and the call to finalize the Ada units must occur
2798 after the last Ada subprogram returns. The binder will place the
2799 initialization and finalization subprograms into the
2800 @file{b~@var{xxx}.adb} file where they can be accessed by your C
2801 sources. To illustrate, we have the following example:
2805 extern void adainit (void);
2806 extern void adafinal (void);
2807 extern int add (int, int);
2808 extern int sub (int, int);
2810 int main (int argc, char *argv[])
2816 /* Should print "21 + 7 = 28" */
2817 printf ("%d + %d = %d\n", a, b, add (a, b));
2818 /* Should print "21 - 7 = 14" */
2819 printf ("%d - %d = %d\n", a, b, sub (a, b));
2825 @smallexample @c ada
2828 function Add (A, B : Integer) return Integer;
2829 pragma Export (C, Add, "add");
2833 package body Unit1 is
2834 function Add (A, B : Integer) return Integer is
2842 function Sub (A, B : Integer) return Integer;
2843 pragma Export (C, Sub, "sub");
2847 package body Unit2 is
2848 function Sub (A, B : Integer) return Integer is
2857 The build procedure for this application is similar to the last
2858 example's. First, compile the foreign language files to generate object
2865 Next, compile the Ada units to produce a set of object files and ALI
2868 gnatmake ^-c^/ACTIONS=COMPILE^ unit1.adb
2869 gnatmake ^-c^/ACTIONS=COMPILE^ unit2.adb
2873 Run the Ada binder on every generated ALI file. Make sure to use the
2874 @option{-n} option to specify a foreign main program:
2876 gnatbind ^-n^/NOMAIN^ unit1.ali unit2.ali
2880 Link the Ada main program, the Ada objects and the foreign language
2881 objects. You need only list the last ALI file here:
2883 gnatlink unit2.ali main.o -o exec_file
2886 This procedure yields a binary executable called @file{exec_file}.
2889 @node Calling Conventions
2890 @subsection Calling Conventions
2891 @cindex Foreign Languages
2892 @cindex Calling Conventions
2893 GNAT follows standard calling sequence conventions and will thus interface
2894 to any other language that also follows these conventions. The following
2895 Convention identifiers are recognized by GNAT:
2898 @cindex Interfacing to Ada
2899 @cindex Other Ada compilers
2900 @cindex Convention Ada
2902 This indicates that the standard Ada calling sequence will be
2903 used and all Ada data items may be passed without any limitations in the
2904 case where GNAT is used to generate both the caller and callee. It is also
2905 possible to mix GNAT generated code and code generated by another Ada
2906 compiler. In this case, the data types should be restricted to simple
2907 cases, including primitive types. Whether complex data types can be passed
2908 depends on the situation. Probably it is safe to pass simple arrays, such
2909 as arrays of integers or floats. Records may or may not work, depending
2910 on whether both compilers lay them out identically. Complex structures
2911 involving variant records, access parameters, tasks, or protected types,
2912 are unlikely to be able to be passed.
2914 Note that in the case of GNAT running
2915 on a platform that supports DEC Ada 83, a higher degree of compatibility
2916 can be guaranteed, and in particular records are layed out in an identical
2917 manner in the two compilers. Note also that if output from two different
2918 compilers is mixed, the program is responsible for dealing with elaboration
2919 issues. Probably the safest approach is to write the main program in the
2920 version of Ada other than GNAT, so that it takes care of its own elaboration
2921 requirements, and then call the GNAT-generated adainit procedure to ensure
2922 elaboration of the GNAT components. Consult the documentation of the other
2923 Ada compiler for further details on elaboration.
2925 However, it is not possible to mix the tasking run time of GNAT and
2926 DEC Ada 83, All the tasking operations must either be entirely within
2927 GNAT compiled sections of the program, or entirely within DEC Ada 83
2928 compiled sections of the program.
2930 @cindex Interfacing to Assembly
2931 @cindex Convention Assembler
2933 Specifies assembler as the convention. In practice this has the
2934 same effect as convention Ada (but is not equivalent in the sense of being
2935 considered the same convention).
2937 @cindex Convention Asm
2940 Equivalent to Assembler.
2942 @cindex Interfacing to COBOL
2943 @cindex Convention COBOL
2946 Data will be passed according to the conventions described
2947 in section B.4 of the Ada 95 Reference Manual.
2950 @cindex Interfacing to C
2951 @cindex Convention C
2953 Data will be passed according to the conventions described
2954 in section B.3 of the Ada 95 Reference Manual.
2956 @findex C varargs function
2957 @cindex Intefacing to C varargs function
2958 @cindex varargs function intefacs
2959 @item C varargs function
2960 In C, @code{varargs} allows a function to take a variable number of
2961 arguments. There is no direct equivalent in this to Ada. One
2962 approach that can be used is to create a C wrapper for each
2963 different profile and then interface to this C wrapper. For
2964 example, to print an @code{int} value using @code{printf},
2965 create a C function @code{printfi} that takes two arguments, a
2966 pointer to a string and an int, and calls @code{printf}.
2967 Then in the Ada program, use pragma @code{Import} to
2968 interface to printfi.
2970 It may work on some platforms to directly interface to
2971 a @code{varargs} function by providing a specific Ada profile
2972 for a a particular call. However, this does not work on
2973 all platforms, since there is no guarantee that the
2974 calling sequence for a two argument normal C function
2975 is the same as for calling a @code{varargs} C function with
2976 the same two arguments.
2978 @cindex Convention Default
2983 @cindex Convention External
2989 @cindex Interfacing to C++
2990 @cindex Convention C++
2992 This stands for C++. For most purposes this is identical to C.
2993 See the separate description of the specialized GNAT pragmas relating to
2994 C++ interfacing for further details.
2997 @cindex Interfacing to Fortran
2998 @cindex Convention Fortran
3000 Data will be passed according to the conventions described
3001 in section B.5 of the Ada 95 Reference Manual.
3004 This applies to an intrinsic operation, as defined in the Ada 95
3005 Reference Manual. If a a pragma Import (Intrinsic) applies to a subprogram,
3006 this means that the body of the subprogram is provided by the compiler itself,
3007 usually by means of an efficient code sequence, and that the user does not
3008 supply an explicit body for it. In an application program, the pragma can
3009 only be applied to the following two sets of names, which the GNAT compiler
3014 Rotate_Left, Rotate_Right, Shift_Left, Shift_Right, Shift_Right_-
3015 Arithmetic. The corresponding subprogram declaration must have
3016 two formal parameters. The
3017 first one must be a signed integer type or a modular type with a binary
3018 modulus, and the second parameter must be of type Natural.
3019 The return type must be the same as the type of the first argument. The size
3020 of this type can only be 8, 16, 32, or 64.
3021 @item binary arithmetic operators: ``+'', ``-'', ``*'', ``/''
3022 The corresponding operator declaration must have parameters and result type
3023 that have the same root numeric type (for example, all three are long_float
3024 types). This simplifies the definition of operations that use type checking
3025 to perform dimensional checks:
3027 @smallexample @c ada
3028 type Distance is new Long_Float;
3029 type Time is new Long_Float;
3030 type Velocity is new Long_Float;
3031 function "/" (D : Distance; T : Time)
3033 pragma Import (Intrinsic, "/");
3037 This common idiom is often programmed with a generic definition and an
3038 explicit body. The pragma makes it simpler to introduce such declarations.
3039 It incurs no overhead in compilation time or code size, because it is
3040 implemented as a single machine instruction.
3046 @cindex Convention Stdcall
3048 This is relevant only to NT/Win95 implementations of GNAT,
3049 and specifies that the Stdcall calling sequence will be used, as defined
3053 @cindex Convention DLL
3055 This is equivalent to Stdcall.
3058 @cindex Convention Win32
3060 This is equivalent to Stdcall.
3064 @cindex Convention Stubbed
3066 This is a special convention that indicates that the compiler
3067 should provide a stub body that raises @code{Program_Error}.
3071 GNAT additionally provides a useful pragma @code{Convention_Identifier}
3072 that can be used to parametrize conventions and allow additional synonyms
3073 to be specified. For example if you have legacy code in which the convention
3074 identifier Fortran77 was used for Fortran, you can use the configuration
3077 @smallexample @c ada
3078 pragma Convention_Identifier (Fortran77, Fortran);
3082 And from now on the identifier Fortran77 may be used as a convention
3083 identifier (for example in an @code{Import} pragma) with the same
3086 @node Building Mixed Ada & C++ Programs
3087 @section Building Mixed Ada & C++ Programs
3090 A programmer inexperienced with mixed-language development may find that
3091 building an application containing both Ada and C++ code can be a
3092 challenge. As a matter of fact, interfacing with C++ has not been
3093 standardized in the Ada 95 Reference Manual due to the immaturity of --
3094 and lack of standards for -- C++ at the time. This section gives a few
3095 hints that should make this task easier. The first section addresses
3096 the differences regarding interfacing with C. The second section
3097 looks into the delicate problem of linking the complete application from
3098 its Ada and C++ parts. The last section gives some hints on how the GNAT
3099 run time can be adapted in order to allow inter-language dispatching
3100 with a new C++ compiler.
3103 * Interfacing to C++::
3104 * Linking a Mixed C++ & Ada Program::
3105 * A Simple Example::
3106 * Adapting the Run Time to a New C++ Compiler::
3109 @node Interfacing to C++
3110 @subsection Interfacing to C++
3113 GNAT supports interfacing with C++ compilers generating code that is
3114 compatible with the standard Application Binary Interface of the given
3118 Interfacing can be done at 3 levels: simple data, subprograms, and
3119 classes. In the first two cases, GNAT offers a specific @var{Convention
3120 CPP} that behaves exactly like @var{Convention C}. Usually, C++ mangles
3121 the names of subprograms, and currently, GNAT does not provide any help
3122 to solve the demangling problem. This problem can be addressed in two
3126 by modifying the C++ code in order to force a C convention using
3127 the @code{extern "C"} syntax.
3130 by figuring out the mangled name and use it as the Link_Name argument of
3135 Interfacing at the class level can be achieved by using the GNAT specific
3136 pragmas such as @code{CPP_Class} and @code{CPP_Virtual}. See the GNAT
3137 Reference Manual for additional information.
3139 @node Linking a Mixed C++ & Ada Program
3140 @subsection Linking a Mixed C++ & Ada Program
3143 Usually the linker of the C++ development system must be used to link
3144 mixed applications because most C++ systems will resolve elaboration
3145 issues (such as calling constructors on global class instances)
3146 transparently during the link phase. GNAT has been adapted to ease the
3147 use of a foreign linker for the last phase. Three cases can be
3152 Using GNAT and G++ (GNU C++ compiler) from the same GCC installation:
3153 The C++ linker can simply be called by using the C++ specific driver
3154 called @code{c++}. Note that this setup is not very common because it
3155 may involve recompiling the whole GCC tree from sources, which makes it
3156 harder to upgrade the compilation system for one language without
3157 destabilizing the other.
3162 $ gnatmake ada_unit -largs file1.o file2.o --LINK=c++
3166 Using GNAT and G++ from two different GCC installations: If both
3167 compilers are on the PATH, the previous method may be used. It is
3168 important to note that environment variables such as C_INCLUDE_PATH,
3169 GCC_EXEC_PREFIX, BINUTILS_ROOT, and GCC_ROOT will affect both compilers
3170 at the same time and may make one of the two compilers operate
3171 improperly if set during invocation of the wrong compiler. It is also
3172 very important that the linker uses the proper @file{libgcc.a} GCC
3173 library -- that is, the one from the C++ compiler installation. The
3174 implicit link command as suggested in the gnatmake command from the
3175 former example can be replaced by an explicit link command with the
3176 full-verbosity option in order to verify which library is used:
3179 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=c++
3181 If there is a problem due to interfering environment variables, it can
3182 be worked around by using an intermediate script. The following example
3183 shows the proper script to use when GNAT has not been installed at its
3184 default location and g++ has been installed at its default location:
3192 $ gnatlink -v -v ada_unit file1.o file2.o --LINK=./my_script
3196 Using a non-GNU C++ compiler: The commands previously described can be
3197 used to insure that the C++ linker is used. Nonetheless, you need to add
3198 the path to libgcc explicitly, since some libraries needed by GNAT are
3199 located in this directory:
3204 CC $* `gcc -print-libgcc-file-name`
3205 $ gnatlink ada_unit file1.o file2.o --LINK=./my_script
3208 Where CC is the name of the non-GNU C++ compiler.
3212 @node A Simple Example
3213 @subsection A Simple Example
3215 The following example, provided as part of the GNAT examples, shows how
3216 to achieve procedural interfacing between Ada and C++ in both
3217 directions. The C++ class A has two methods. The first method is exported
3218 to Ada by the means of an extern C wrapper function. The second method
3219 calls an Ada subprogram. On the Ada side, The C++ calls are modelled by
3220 a limited record with a layout comparable to the C++ class. The Ada
3221 subprogram, in turn, calls the C++ method. So, starting from the C++
3222 main program, the process passes back and forth between the two
3226 Here are the compilation commands:
3228 $ gnatmake -c simple_cpp_interface
3231 $ gnatbind -n simple_cpp_interface
3232 $ gnatlink simple_cpp_interface -o cpp_main --LINK=$(CPLUSPLUS)
3233 -lstdc++ ex7.o cpp_main.o
3237 Here are the corresponding sources:
3245 void adainit (void);
3246 void adafinal (void);
3247 void method1 (A *t);
3269 class A : public Origin @{
3271 void method1 (void);
3272 virtual void method2 (int v);
3282 extern "C" @{ void ada_method2 (A *t, int v);@}
3284 void A::method1 (void)
3287 printf ("in A::method1, a_value = %d \n",a_value);
3291 void A::method2 (int v)
3293 ada_method2 (this, v);
3294 printf ("in A::method2, a_value = %d \n",a_value);
3301 printf ("in A::A, a_value = %d \n",a_value);
3305 @b{package} @b{body} Simple_Cpp_Interface @b{is}
3307 @b{procedure} Ada_Method2 (This : @b{in} @b{out} A; V : Integer) @b{is}
3311 @b{end} Ada_Method2;
3313 @b{end} Simple_Cpp_Interface;
3315 @b{package} Simple_Cpp_Interface @b{is}
3316 @b{type} A @b{is} @b{limited}
3321 @b{pragma} Convention (C, A);
3323 @b{procedure} Method1 (This : @b{in} @b{out} A);
3324 @b{pragma} Import (C, Method1);
3326 @b{procedure} Ada_Method2 (This : @b{in} @b{out} A; V : Integer);
3327 @b{pragma} Export (C, Ada_Method2);
3329 @b{end} Simple_Cpp_Interface;
3332 @node Adapting the Run Time to a New C++ Compiler
3333 @subsection Adapting the Run Time to a New C++ Compiler
3335 GNAT offers the capability to derive Ada 95 tagged types directly from
3336 preexisting C++ classes and . See ``Interfacing with C++'' in the
3337 @cite{GNAT Reference Manual}. The mechanism used by GNAT for achieving
3339 has been made user configurable through a GNAT library unit
3340 @code{Interfaces.CPP}. The default version of this file is adapted to
3341 the GNU C++ compiler. Internal knowledge of the virtual
3342 table layout used by the new C++ compiler is needed to configure
3343 properly this unit. The Interface of this unit is known by the compiler
3344 and cannot be changed except for the value of the constants defining the
3345 characteristics of the virtual table: CPP_DT_Prologue_Size, CPP_DT_Entry_Size,
3346 CPP_TSD_Prologue_Size, CPP_TSD_Entry_Size. Read comments in the source
3347 of this unit for more details.
3349 @node Comparison between GNAT and C/C++ Compilation Models
3350 @section Comparison between GNAT and C/C++ Compilation Models
3353 The GNAT model of compilation is close to the C and C++ models. You can
3354 think of Ada specs as corresponding to header files in C. As in C, you
3355 don't need to compile specs; they are compiled when they are used. The
3356 Ada @code{with} is similar in effect to the @code{#include} of a C
3359 One notable difference is that, in Ada, you may compile specs separately
3360 to check them for semantic and syntactic accuracy. This is not always
3361 possible with C headers because they are fragments of programs that have
3362 less specific syntactic or semantic rules.
3364 The other major difference is the requirement for running the binder,
3365 which performs two important functions. First, it checks for
3366 consistency. In C or C++, the only defense against assembling
3367 inconsistent programs lies outside the compiler, in a makefile, for
3368 example. The binder satisfies the Ada requirement that it be impossible
3369 to construct an inconsistent program when the compiler is used in normal
3372 @cindex Elaboration order control
3373 The other important function of the binder is to deal with elaboration
3374 issues. There are also elaboration issues in C++ that are handled
3375 automatically. This automatic handling has the advantage of being
3376 simpler to use, but the C++ programmer has no control over elaboration.
3377 Where @code{gnatbind} might complain there was no valid order of
3378 elaboration, a C++ compiler would simply construct a program that
3379 malfunctioned at run time.
3381 @node Comparison between GNAT and Conventional Ada Library Models
3382 @section Comparison between GNAT and Conventional Ada Library Models
3385 This section is intended to be useful to Ada programmers who have
3386 previously used an Ada compiler implementing the traditional Ada library
3387 model, as described in the Ada 95 Language Reference Manual. If you
3388 have not used such a system, please go on to the next section.
3390 @cindex GNAT library
3391 In GNAT, there is no @dfn{library} in the normal sense. Instead, the set of
3392 source files themselves acts as the library. Compiling Ada programs does
3393 not generate any centralized information, but rather an object file and
3394 a ALI file, which are of interest only to the binder and linker.
3395 In a traditional system, the compiler reads information not only from
3396 the source file being compiled, but also from the centralized library.
3397 This means that the effect of a compilation depends on what has been
3398 previously compiled. In particular:
3402 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3403 to the version of the unit most recently compiled into the library.
3406 Inlining is effective only if the necessary body has already been
3407 compiled into the library.
3410 Compiling a unit may obsolete other units in the library.
3414 In GNAT, compiling one unit never affects the compilation of any other
3415 units because the compiler reads only source files. Only changes to source
3416 files can affect the results of a compilation. In particular:
3420 When a unit is @code{with}'ed, the unit seen by the compiler corresponds
3421 to the source version of the unit that is currently accessible to the
3426 Inlining requires the appropriate source files for the package or
3427 subprogram bodies to be available to the compiler. Inlining is always
3428 effective, independent of the order in which units are complied.
3431 Compiling a unit never affects any other compilations. The editing of
3432 sources may cause previous compilations to be out of date if they
3433 depended on the source file being modified.
3437 The most important result of these differences is that order of compilation
3438 is never significant in GNAT. There is no situation in which one is
3439 required to do one compilation before another. What shows up as order of
3440 compilation requirements in the traditional Ada library becomes, in
3441 GNAT, simple source dependencies; in other words, there is only a set
3442 of rules saying what source files must be present when a file is
3446 @node Placement of temporary files
3447 @section Placement of temporary files
3448 @cindex Temporary files (user control over placement)
3451 GNAT creates temporary files in the directory designated by the environment
3452 variable @env{TMPDIR}.
3453 (See the HP @emph{C RTL Reference Manual} on the function @code{getenv()}
3454 for detailed information on how environment variables are resolved.
3455 For most users the easiest way to make use of this feature is to simply
3456 define @env{TMPDIR} as a job level logical name).
3457 For example, if you wish to use a Ramdisk (assuming DECRAM is installed)
3458 for compiler temporary files, then you can include something like the
3459 following command in your @file{LOGIN.COM} file:
3462 $ define/job TMPDIR "/disk$scratchram/000000/temp/"
3466 If @env{TMPDIR} is not defined, then GNAT uses the directory designated by
3467 @env{TMP}; if @env{TMP} is not defined, then GNAT uses the directory
3468 designated by @env{TEMP}.
3469 If none of these environment variables are defined then GNAT uses the
3470 directory designated by the logical name @code{SYS$SCRATCH:}
3471 (by default the user's home directory). If all else fails
3472 GNAT uses the current directory for temporary files.
3476 @c *************************
3477 @node Compiling Using gcc
3478 @chapter Compiling Using @code{gcc}
3481 This chapter discusses how to compile Ada programs using the @code{gcc}
3482 command. It also describes the set of switches
3483 that can be used to control the behavior of the compiler.
3485 * Compiling Programs::
3486 * Switches for gcc::
3487 * Search Paths and the Run-Time Library (RTL)::
3488 * Order of Compilation Issues::
3492 @node Compiling Programs
3493 @section Compiling Programs
3496 The first step in creating an executable program is to compile the units
3497 of the program using the @code{gcc} command. You must compile the
3502 the body file (@file{.adb}) for a library level subprogram or generic
3506 the spec file (@file{.ads}) for a library level package or generic
3507 package that has no body
3510 the body file (@file{.adb}) for a library level package
3511 or generic package that has a body
3516 You need @emph{not} compile the following files
3521 the spec of a library unit which has a body
3528 because they are compiled as part of compiling related units. GNAT
3530 when the corresponding body is compiled, and subunits when the parent is
3533 @cindex cannot generate code
3534 If you attempt to compile any of these files, you will get one of the
3535 following error messages (where fff is the name of the file you compiled):
3538 cannot generate code for file @var{fff} (package spec)
3539 to check package spec, use -gnatc
3541 cannot generate code for file @var{fff} (missing subunits)
3542 to check parent unit, use -gnatc
3544 cannot generate code for file @var{fff} (subprogram spec)
3545 to check subprogram spec, use -gnatc
3547 cannot generate code for file @var{fff} (subunit)
3548 to check subunit, use -gnatc
3552 As indicated by the above error messages, if you want to submit
3553 one of these files to the compiler to check for correct semantics
3554 without generating code, then use the @option{-gnatc} switch.
3556 The basic command for compiling a file containing an Ada unit is
3559 $ gcc -c [@var{switches}] @file{file name}
3563 where @var{file name} is the name of the Ada file (usually
3565 @file{.ads} for a spec or @file{.adb} for a body).
3568 @option{-c} switch to tell @code{gcc} to compile, but not link, the file.
3570 The result of a successful compilation is an object file, which has the
3571 same name as the source file but an extension of @file{.o} and an Ada
3572 Library Information (ALI) file, which also has the same name as the
3573 source file, but with @file{.ali} as the extension. GNAT creates these
3574 two output files in the current directory, but you may specify a source
3575 file in any directory using an absolute or relative path specification
3576 containing the directory information.
3579 @code{gcc} is actually a driver program that looks at the extensions of
3580 the file arguments and loads the appropriate compiler. For example, the
3581 GNU C compiler is @file{cc1}, and the Ada compiler is @file{gnat1}.
3582 These programs are in directories known to the driver program (in some
3583 configurations via environment variables you set), but need not be in
3584 your path. The @code{gcc} driver also calls the assembler and any other
3585 utilities needed to complete the generation of the required object
3588 It is possible to supply several file names on the same @code{gcc}
3589 command. This causes @code{gcc} to call the appropriate compiler for
3590 each file. For example, the following command lists three separate
3591 files to be compiled:
3594 $ gcc -c x.adb y.adb z.c
3598 calls @code{gnat1} (the Ada compiler) twice to compile @file{x.adb} and
3599 @file{y.adb}, and @code{cc1} (the C compiler) once to compile @file{z.c}.
3600 The compiler generates three object files @file{x.o}, @file{y.o} and
3601 @file{z.o} and the two ALI files @file{x.ali} and @file{y.ali} from the
3602 Ada compilations. Any switches apply to all the files ^listed,^listed.^
3605 @option{-gnat@var{x}} switches, which apply only to Ada compilations.
3608 @node Switches for gcc
3609 @section Switches for @code{gcc}
3612 The @code{gcc} command accepts switches that control the
3613 compilation process. These switches are fully described in this section.
3614 First we briefly list all the switches, in alphabetical order, then we
3615 describe the switches in more detail in functionally grouped sections.
3618 * Output and Error Message Control::
3619 * Warning Message Control::
3620 * Debugging and Assertion Control::
3622 * Stack Overflow Checking::
3623 * Validity Checking::
3625 * Using gcc for Syntax Checking::
3626 * Using gcc for Semantic Checking::
3627 * Compiling Ada 83 Programs::
3628 * Character Set Control::
3629 * File Naming Control::
3630 * Subprogram Inlining Control::
3631 * Auxiliary Output Control::
3632 * Debugging Control::
3633 * Exception Handling Control::
3634 * Units to Sources Mapping Files::
3635 * Integrated Preprocessing::
3644 @cindex @option{-b} (@code{gcc})
3645 @item -b @var{target}
3646 Compile your program to run on @var{target}, which is the name of a
3647 system configuration. You must have a GNAT cross-compiler built if
3648 @var{target} is not the same as your host system.
3651 @cindex @option{-B} (@code{gcc})
3652 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
3653 from @var{dir} instead of the default location. Only use this switch
3654 when multiple versions of the GNAT compiler are available. See the
3655 @code{gcc} manual page for further details. You would normally use the
3656 @option{-b} or @option{-V} switch instead.
3659 @cindex @option{-c} (@code{gcc})
3660 Compile. Always use this switch when compiling Ada programs.
3662 Note: for some other languages when using @code{gcc}, notably in
3663 the case of C and C++, it is possible to use
3664 use @code{gcc} without a @option{-c} switch to
3665 compile and link in one step. In the case of GNAT, you
3666 cannot use this approach, because the binder must be run
3667 and @code{gcc} cannot be used to run the GNAT binder.
3671 @cindex @option{-fno-inline} (@code{gcc})
3672 Suppresses all back-end inlining, even if other optimization or inlining
3674 This includes suppression of inlining that results
3675 from the use of the pragma @code{Inline_Always}.
3676 See also @option{-gnatn} and @option{-gnatN}.
3678 @item -fno-strict-aliasing
3679 @cindex @option{-fno-strict-aliasing} (@code{gcc})
3680 Causes the compiler to avoid assumptions regarding non-aliasing
3681 of objects of different types. See section
3682 @pxref{Optimization and Strict Aliasing} for details.
3685 @cindex @option{-fstack-check} (@code{gcc})
3686 Activates stack checking.
3687 See @ref{Stack Overflow Checking} for details of the use of this option.
3690 @cindex @option{^-g^/DEBUG^} (@code{gcc})
3691 Generate debugging information. This information is stored in the object
3692 file and copied from there to the final executable file by the linker,
3693 where it can be read by the debugger. You must use the
3694 @option{^-g^/DEBUG^} switch if you plan on using the debugger.
3697 @cindex @option{-gnat83} (@code{gcc})
3698 Enforce Ada 83 restrictions.
3701 @cindex @option{-gnata} (@code{gcc})
3702 Assertions enabled. @code{Pragma Assert} and @code{pragma Debug} to be
3706 @cindex @option{-gnatA} (@code{gcc})
3707 Avoid processing @file{gnat.adc}. If a gnat.adc file is present,
3711 @cindex @option{-gnatb} (@code{gcc})
3712 Generate brief messages to @file{stderr} even if verbose mode set.
3715 @cindex @option{-gnatc} (@code{gcc})
3716 Check syntax and semantics only (no code generation attempted).
3719 @cindex @option{-gnatd} (@code{gcc})
3720 Specify debug options for the compiler. The string of characters after
3721 the @option{-gnatd} specify the specific debug options. The possible
3722 characters are 0-9, a-z, A-Z, optionally preceded by a dot. See
3723 compiler source file @file{debug.adb} for details of the implemented
3724 debug options. Certain debug options are relevant to applications
3725 programmers, and these are documented at appropriate points in this
3729 @cindex @option{-gnatD} (@code{gcc})
3730 Create expanded source files for source level debugging. This switch
3731 also suppress generation of cross-reference information
3732 (see @option{-gnatx}).
3734 @item -gnatec=@var{path}
3735 @cindex @option{-gnatec} (@code{gcc})
3736 Specify a configuration pragma file
3738 (the equal sign is optional)
3740 (see @ref{The Configuration Pragmas Files}).
3742 @item ^-gnateD^/DATA_PREPROCESSING=^symbol[=value]
3743 @cindex @option{-gnateD} (@code{gcc})
3744 Defines a symbol, associated with value, for preprocessing.
3745 (see @ref{Integrated Preprocessing})
3748 @cindex @option{-gnatef} (@code{gcc})
3749 Display full source path name in brief error messages.
3751 @item -gnatem=@var{path}
3752 @cindex @option{-gnatem} (@code{gcc})
3753 Specify a mapping file
3755 (the equal sign is optional)
3757 (see @ref{Units to Sources Mapping Files}).
3759 @item -gnatep=@var{file}
3760 @cindex @option{-gnatep} (@code{gcc})
3761 Specify a preprocessing data file
3763 (the equal sign is optional)
3765 (see @ref{Integrated Preprocessing}).
3768 @cindex @option{-gnatE} (@code{gcc})
3769 Full dynamic elaboration checks.
3772 @cindex @option{-gnatf} (@code{gcc})
3773 Full errors. Multiple errors per line, all undefined references, do not
3774 attempt to suppress cascaded errors.
3777 @cindex @option{-gnatF} (@code{gcc})
3778 Externals names are folded to all uppercase.
3781 @cindex @option{-gnatg} (@code{gcc})
3782 Internal GNAT implementation mode. This should not be used for
3783 applications programs, it is intended only for use by the compiler
3784 and its run-time library. For documentation, see the GNAT sources.
3785 Note that @option{-gnatg} implies @option{-gnatwu} so that warnings
3786 are generated on unreferenced entities, and all warnings are treated
3790 @cindex @option{-gnatG} (@code{gcc})
3791 List generated expanded code in source form.
3793 @item ^-gnath^/HELP^
3794 @cindex @option{^-gnath^/HELP^} (@code{gcc})
3795 Output usage information. The output is written to @file{stdout}.
3797 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
3798 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@code{gcc})
3799 Identifier character set
3801 (@var{c}=1/2/3/4/8/9/p/f/n/w).
3804 For details of the possible selections for @var{c},
3805 see @xref{Character Set Control}.
3808 @item -gnatk=@var{n}
3809 @cindex @option{-gnatk} (@code{gcc})
3810 Limit file names to @var{n} (1-999) characters ^(@code{k} = krunch)^^.
3813 @cindex @option{-gnatl} (@code{gcc})
3814 Output full source listing with embedded error messages.
3817 @cindex @option{-gnatL} (@code{gcc})
3818 Use the longjmp/setjmp method for exception handling
3820 @item -gnatm=@var{n}
3821 @cindex @option{-gnatm} (@code{gcc})
3822 Limit number of detected error or warning messages to @var{n}
3823 where @var{n} is in the range 1..999_999. The default setting if
3824 no switch is given is 9999. Compilation is terminated if this
3828 @cindex @option{-gnatn} (@code{gcc})
3829 Activate inlining for subprograms for which
3830 pragma @code{inline} is specified. This inlining is performed
3831 by the GCC back-end.
3834 @cindex @option{-gnatN} (@code{gcc})
3835 Activate front end inlining for subprograms for which
3836 pragma @code{Inline} is specified. This inlining is performed
3837 by the front end and will be visible in the
3838 @option{-gnatG} output.
3839 In some cases, this has proved more effective than the back end
3840 inlining resulting from the use of
3843 @option{-gnatN} automatically implies
3844 @option{-gnatn} so it is not necessary
3845 to specify both options. There are a few cases that the back-end inlining
3846 catches that cannot be dealt with in the front-end.
3849 @cindex @option{-gnato} (@code{gcc})
3850 Enable numeric overflow checking (which is not normally enabled by
3851 default). Not that division by zero is a separate check that is not
3852 controlled by this switch (division by zero checking is on by default).
3855 @cindex @option{-gnatp} (@code{gcc})
3856 Suppress all checks.
3859 @cindex @option{-gnatP} (@code{gcc})
3860 Enable polling. This is required on some systems (notably Windows NT) to
3861 obtain asynchronous abort and asynchronous transfer of control capability.
3862 See the description of pragma Polling in the GNAT Reference Manual for
3866 @cindex @option{-gnatq} (@code{gcc})
3867 Don't quit; try semantics, even if parse errors.
3870 @cindex @option{-gnatQ} (@code{gcc})
3871 Don't quit; generate @file{ALI} and tree files even if illegalities.
3873 @item ^-gnatR[0/1/2/3[s]]^/REPRESENTATION_INFO^
3874 @cindex @option{-gnatR} (@code{gcc})
3875 Output representation information for declared types and objects.
3878 @cindex @option{-gnats} (@code{gcc})
3882 @cindex @option{-gnatS} (@code{gcc})
3883 Print package Standard.
3886 @cindex @option{-gnatt} (@code{gcc})
3887 Generate tree output file.
3889 @item ^-gnatT^/TABLE_MULTIPLIER=^@var{nnn}
3890 @cindex @option{^-gnatT^/TABLE_MULTIPLIER^} (@code{gcc})
3891 All compiler tables start at @var{nnn} times usual starting size.
3894 @cindex @option{-gnatu} (@code{gcc})
3895 List units for this compilation.
3898 @cindex @option{-gnatU} (@code{gcc})
3899 Tag all error messages with the unique string ``error:''
3902 @cindex @option{-gnatv} (@code{gcc})
3903 Verbose mode. Full error output with source lines to @file{stdout}.
3906 @cindex @option{-gnatV} (@code{gcc})
3907 Control level of validity checking. See separate section describing
3910 @item ^-gnatw@var{xxx}^/WARNINGS=(@var{option}[,...])^
3911 @cindex @option{^-gnatw^/WARNINGS^} (@code{gcc})
3913 ^@var{xxx} is a string of option letters that^the list of options^ denotes
3914 the exact warnings that
3915 are enabled or disabled. (see @ref{Warning Message Control})
3917 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
3918 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@code{gcc})
3919 Wide character encoding method
3921 (@var{e}=n/h/u/s/e/8).
3924 (@var{e}=@code{BRACKETS, NONE, HEX, UPPER, SHIFT_JIS, EUC, UTF8})
3928 @cindex @option{-gnatx} (@code{gcc})
3929 Suppress generation of cross-reference information.
3931 @item ^-gnaty^/STYLE_CHECKS=(option,option..)^
3932 @cindex @option{^-gnaty^/STYLE_CHECKS^} (@code{gcc})
3933 Enable built-in style checks. (see @ref{Style Checking})
3935 @item ^-gnatz^/DISTRIBUTION_STUBS=^@var{m}
3936 @cindex @option{^-gnatz^/DISTRIBUTION_STUBS^} (@code{gcc})
3937 Distribution stub generation and compilation
3939 (@var{m}=r/c for receiver/caller stubs).
3942 (@var{m}=@code{RECEIVER} or @code{CALLER} to specify the type of stubs
3943 to be generated and compiled).
3947 Use the zero cost method for exception handling
3949 @item ^-I^/SEARCH=^@var{dir}
3950 @cindex @option{^-I^/SEARCH^} (@code{gcc})
3952 Direct GNAT to search the @var{dir} directory for source files needed by
3953 the current compilation
3954 (@pxref{Search Paths and the Run-Time Library (RTL)}).
3956 @item ^-I-^/NOCURRENT_DIRECTORY^
3957 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gcc})
3959 Except for the source file named in the command line, do not look for source
3960 files in the directory containing the source file named in the command line
3961 (@pxref{Search Paths and the Run-Time Library (RTL)}).
3965 @cindex @option{-mbig-switch} (@command{gcc})
3966 @cindex @code{case} statement (effect of @option{-mbig-switch} option)
3967 This standard gcc switch causes the compiler to use larger offsets in its
3968 jump table representation for @code{case} statements.
3969 This may result in less efficient code, but is sometimes necessary
3970 (for example on HP-UX targets)
3971 @cindex HP-UX and @option{-mbig-switch} option
3972 in order to compile large and/or nested @code{case} statements.
3975 @cindex @option{-o} (@code{gcc})
3976 This switch is used in @code{gcc} to redirect the generated object file
3977 and its associated ALI file. Beware of this switch with GNAT, because it may
3978 cause the object file and ALI file to have different names which in turn
3979 may confuse the binder and the linker.
3983 @cindex @option{-nostdinc} (@command{gcc})
3984 Inhibit the search of the default location for the GNAT Run Time
3985 Library (RTL) source files.
3988 @cindex @option{-nostdlib} (@command{gcc})
3989 Inhibit the search of the default location for the GNAT Run Time
3990 Library (RTL) ALI files.
3994 @cindex @option{-O} (@code{gcc})
3995 @var{n} controls the optimization level.
3999 No optimization, the default setting if no @option{-O} appears
4002 Normal optimization, the default if you specify @option{-O} without
4006 Extensive optimization
4009 Extensive optimization with automatic inlining of subprograms not
4010 specified by pragma @code{Inline}. This applies only to
4011 inlining within a unit. For details on control of inlining
4012 see @xref{Subprogram Inlining Control}.
4018 @cindex @option{/NOOPTIMIZE} (@code{GNAT COMPILE})
4019 Equivalent to @option{/OPTIMIZE=NONE}.
4020 This is the default behavior in the absence of an @option{/OPTMIZE}
4023 @item /OPTIMIZE[=(keyword[,...])]
4024 @cindex @option{/OPTIMIZE} (@code{GNAT COMPILE})
4025 Selects the level of optimization for your program. The supported
4026 keywords are as follows:
4029 Perform most optimizations, including those that
4031 This is the default if the @option{/OPTMIZE} qualifier is supplied
4032 without keyword options.
4035 Do not do any optimizations. Same as @code{/NOOPTIMIZE}.
4038 Perform some optimizations, but omit ones that are costly.
4041 Same as @code{SOME}.
4044 Full optimization, and also attempt automatic inlining of small
4045 subprograms within a unit even when pragma @code{Inline}
4046 is not specified (@pxref{Inlining of Subprograms}).
4049 Try to unroll loops. This keyword may be specified together with
4050 any keyword above other than @code{NONE}. Loop unrolling
4051 usually, but not always, improves the performance of programs.
4056 @item -pass-exit-codes
4057 @cindex @option{-pass-exit-codes} (@code{gcc})
4058 Catch exit codes from the compiler and use the most meaningful as
4062 @item --RTS=@var{rts-path}
4063 @cindex @option{--RTS} (@code{gcc})
4064 Specifies the default location of the runtime library. Same meaning as the
4065 equivalent @code{gnatmake} flag (see @ref{Switches for gnatmake}).
4068 @cindex @option{^-S^/ASM^} (@code{gcc})
4069 ^Used in place of @option{-c} to^Used to^
4070 cause the assembler source file to be
4071 generated, using @file{^.s^.S^} as the extension,
4072 instead of the object file.
4073 This may be useful if you need to examine the generated assembly code.
4076 @cindex @option{^-v^/VERBOSE^} (@code{gcc})
4077 Show commands generated by the @code{gcc} driver. Normally used only for
4078 debugging purposes or if you need to be sure what version of the
4079 compiler you are executing.
4083 @cindex @option{-V} (@code{gcc})
4084 Execute @var{ver} version of the compiler. This is the @code{gcc}
4085 version, not the GNAT version.
4091 You may combine a sequence of GNAT switches into a single switch. For
4092 example, the combined switch
4094 @cindex Combining GNAT switches
4100 is equivalent to specifying the following sequence of switches:
4103 -gnato -gnatf -gnati3
4108 @c NEED TO CHECK THIS FOR VMS
4111 The following restrictions apply to the combination of switches
4116 The switch @option{-gnatc} if combined with other switches must come
4117 first in the string.
4120 The switch @option{-gnats} if combined with other switches must come
4121 first in the string.
4125 @option{^-gnatz^/DISTRIBUTION_STUBS^}, @option{-gnatzc}, and @option{-gnatzr}
4126 may not be combined with any other switches.
4130 Once a ``y'' appears in the string (that is a use of the @option{-gnaty}
4131 switch), then all further characters in the switch are interpreted
4132 as style modifiers (see description of @option{-gnaty}).
4135 Once a ``d'' appears in the string (that is a use of the @option{-gnatd}
4136 switch), then all further characters in the switch are interpreted
4137 as debug flags (see description of @option{-gnatd}).
4140 Once a ``w'' appears in the string (that is a use of the @option{-gnatw}
4141 switch), then all further characters in the switch are interpreted
4142 as warning mode modifiers (see description of @option{-gnatw}).
4145 Once a ``V'' appears in the string (that is a use of the @option{-gnatV}
4146 switch), then all further characters in the switch are interpreted
4147 as validity checking options (see description of @option{-gnatV}).
4152 @node Output and Error Message Control
4153 @subsection Output and Error Message Control
4157 The standard default format for error messages is called ``brief format''.
4158 Brief format messages are written to @file{stderr} (the standard error
4159 file) and have the following form:
4162 e.adb:3:04: Incorrect spelling of keyword "function"
4163 e.adb:4:20: ";" should be "is"
4167 The first integer after the file name is the line number in the file,
4168 and the second integer is the column number within the line.
4169 @code{glide} can parse the error messages
4170 and point to the referenced character.
4171 The following switches provide control over the error message
4177 @cindex @option{-gnatv} (@code{gcc})
4180 The v stands for verbose.
4182 The effect of this setting is to write long-format error
4183 messages to @file{stdout} (the standard output file.
4184 The same program compiled with the
4185 @option{-gnatv} switch would generate:
4189 3. funcion X (Q : Integer)
4191 >>> Incorrect spelling of keyword "function"
4194 >>> ";" should be "is"
4199 The vertical bar indicates the location of the error, and the @samp{>>>}
4200 prefix can be used to search for error messages. When this switch is
4201 used the only source lines output are those with errors.
4204 @cindex @option{-gnatl} (@code{gcc})
4206 The @code{l} stands for list.
4208 This switch causes a full listing of
4209 the file to be generated. The output might look as follows:
4215 3. funcion X (Q : Integer)
4217 >>> Incorrect spelling of keyword "function"
4220 >>> ";" should be "is"
4232 When you specify the @option{-gnatv} or @option{-gnatl} switches and
4233 standard output is redirected, a brief summary is written to
4234 @file{stderr} (standard error) giving the number of error messages and
4235 warning messages generated.
4238 @cindex @option{-gnatU} (@code{gcc})
4239 This switch forces all error messages to be preceded by the unique
4240 string ``error:''. This means that error messages take a few more
4241 characters in space, but allows easy searching for and identification
4245 @cindex @option{-gnatb} (@code{gcc})
4247 The @code{b} stands for brief.
4249 This switch causes GNAT to generate the
4250 brief format error messages to @file{stderr} (the standard error
4251 file) as well as the verbose
4252 format message or full listing (which as usual is written to
4253 @file{stdout} (the standard output file).
4255 @item -gnatm^^=^@var{n}
4256 @cindex @option{-gnatm} (@code{gcc})
4258 The @code{m} stands for maximum.
4260 @var{n} is a decimal integer in the
4261 range of 1 to 999 and limits the number of error messages to be
4262 generated. For example, using @option{-gnatm2} might yield
4265 e.adb:3:04: Incorrect spelling of keyword "function"
4266 e.adb:5:35: missing ".."
4267 fatal error: maximum errors reached
4268 compilation abandoned
4272 @cindex @option{-gnatf} (@code{gcc})
4273 @cindex Error messages, suppressing
4275 The @code{f} stands for full.
4277 Normally, the compiler suppresses error messages that are likely to be
4278 redundant. This switch causes all error
4279 messages to be generated. In particular, in the case of
4280 references to undefined variables. If a given variable is referenced
4281 several times, the normal format of messages is
4283 e.adb:7:07: "V" is undefined (more references follow)
4287 where the parenthetical comment warns that there are additional
4288 references to the variable @code{V}. Compiling the same program with the
4289 @option{-gnatf} switch yields
4292 e.adb:7:07: "V" is undefined
4293 e.adb:8:07: "V" is undefined
4294 e.adb:8:12: "V" is undefined
4295 e.adb:8:16: "V" is undefined
4296 e.adb:9:07: "V" is undefined
4297 e.adb:9:12: "V" is undefined
4301 The @option{-gnatf} switch also generates additional information for
4302 some error messages. Some examples are:
4306 Full details on entities not available in high integrity mode
4308 Details on possibly non-portable unchecked conversion
4310 List possible interpretations for ambiguous calls
4312 Additional details on incorrect parameters
4317 @cindex @option{-gnatq} (@code{gcc})
4319 The @code{q} stands for quit (really ``don't quit'').
4321 In normal operation mode, the compiler first parses the program and
4322 determines if there are any syntax errors. If there are, appropriate
4323 error messages are generated and compilation is immediately terminated.
4325 GNAT to continue with semantic analysis even if syntax errors have been
4326 found. This may enable the detection of more errors in a single run. On
4327 the other hand, the semantic analyzer is more likely to encounter some
4328 internal fatal error when given a syntactically invalid tree.
4331 @cindex @option{-gnatQ} (@code{gcc})
4332 In normal operation mode, the @file{ALI} file is not generated if any
4333 illegalities are detected in the program. The use of @option{-gnatQ} forces
4334 generation of the @file{ALI} file. This file is marked as being in
4335 error, so it cannot be used for binding purposes, but it does contain
4336 reasonably complete cross-reference information, and thus may be useful
4337 for use by tools (e.g. semantic browsing tools or integrated development
4338 environments) that are driven from the @file{ALI} file. This switch
4339 implies @option{-gnatq}, since the semantic phase must be run to get a
4340 meaningful ALI file.
4342 In addition, if @option{-gnatt} is also specified, then the tree file is
4343 generated even if there are illegalities. It may be useful in this case
4344 to also specify @option{-gnatq} to ensure that full semantic processing
4345 occurs. The resulting tree file can be processed by ASIS, for the purpose
4346 of providing partial information about illegal units, but if the error
4347 causes the tree to be badly malformed, then ASIS may crash during the
4350 When @option{-gnatQ} is used and the generated @file{ALI} file is marked as
4351 being in error, @code{gnatmake} will attempt to recompile the source when it
4352 finds such an @file{ALI} file, including with switch @option{-gnatc}.
4354 Note that @option{-gnatQ} has no effect if @option{-gnats} is specified,
4355 since ALI files are never generated if @option{-gnats} is set.
4360 @node Warning Message Control
4361 @subsection Warning Message Control
4362 @cindex Warning messages
4364 In addition to error messages, which correspond to illegalities as defined
4365 in the Ada 95 Reference Manual, the compiler detects two kinds of warning
4368 First, the compiler considers some constructs suspicious and generates a
4369 warning message to alert you to a possible error. Second, if the
4370 compiler detects a situation that is sure to raise an exception at
4371 run time, it generates a warning message. The following shows an example
4372 of warning messages:
4374 e.adb:4:24: warning: creation of object may raise Storage_Error
4375 e.adb:10:17: warning: static value out of range
4376 e.adb:10:17: warning: "Constraint_Error" will be raised at run time
4380 GNAT considers a large number of situations as appropriate
4381 for the generation of warning messages. As always, warnings are not
4382 definite indications of errors. For example, if you do an out-of-range
4383 assignment with the deliberate intention of raising a
4384 @code{Constraint_Error} exception, then the warning that may be
4385 issued does not indicate an error. Some of the situations for which GNAT
4386 issues warnings (at least some of the time) are given in the following
4387 list. This list is not complete, and new warnings are often added to
4388 subsequent versions of GNAT. The list is intended to give a general idea
4389 of the kinds of warnings that are generated.
4393 Possible infinitely recursive calls
4396 Out-of-range values being assigned
4399 Possible order of elaboration problems
4405 Fixed-point type declarations with a null range
4408 Variables that are never assigned a value
4411 Variables that are referenced before being initialized
4414 Task entries with no corresponding @code{accept} statement
4417 Duplicate accepts for the same task entry in a @code{select}
4420 Objects that take too much storage
4423 Unchecked conversion between types of differing sizes
4426 Missing @code{return} statement along some execution path in a function
4429 Incorrect (unrecognized) pragmas
4432 Incorrect external names
4435 Allocation from empty storage pool
4438 Potentially blocking operation in protected type
4441 Suspicious parenthesization of expressions
4444 Mismatching bounds in an aggregate
4447 Attempt to return local value by reference
4451 Premature instantiation of a generic body
4454 Attempt to pack aliased components
4457 Out of bounds array subscripts
4460 Wrong length on string assignment
4463 Violations of style rules if style checking is enabled
4466 Unused @code{with} clauses
4469 @code{Bit_Order} usage that does not have any effect
4472 @code{Standard.Duration} used to resolve universal fixed expression
4475 Dereference of possibly null value
4478 Declaration that is likely to cause storage error
4481 Internal GNAT unit @code{with}'ed by application unit
4484 Values known to be out of range at compile time
4487 Unreferenced labels and variables
4490 Address overlays that could clobber memory
4493 Unexpected initialization when address clause present
4496 Bad alignment for address clause
4499 Useless type conversions
4502 Redundant assignment statements and other redundant constructs
4505 Useless exception handlers
4508 Accidental hiding of name by child unit
4512 Access before elaboration detected at compile time
4515 A range in a @code{for} loop that is known to be null or might be null
4520 The following switches are available to control the handling of
4526 @emph{Activate all optional errors.}
4527 @cindex @option{-gnatwa} (@code{gcc})
4528 This switch activates most optional warning messages, see remaining list
4529 in this section for details on optional warning messages that can be
4530 individually controlled. The warnings that are not turned on by this
4532 @option{-gnatwd} (implicit dereferencing),
4533 @option{-gnatwh} (hiding),
4534 and @option{-gnatwl} (elaboration warnings).
4535 All other optional warnings are turned on.
4538 @emph{Suppress all optional errors.}
4539 @cindex @option{-gnatwA} (@code{gcc})
4540 This switch suppresses all optional warning messages, see remaining list
4541 in this section for details on optional warning messages that can be
4542 individually controlled.
4545 @emph{Activate warnings on conditionals.}
4546 @cindex @option{-gnatwc} (@code{gcc})
4547 @cindex Conditionals, constant
4548 This switch activates warnings for conditional expressions used in
4549 tests that are known to be True or False at compile time. The default
4550 is that such warnings are not generated.
4551 Note that this warning does
4552 not get issued for the use of boolean variables or constants whose
4553 values are known at compile time, since this is a standard technique
4554 for conditional compilation in Ada, and this would generate too many
4555 ``false positive'' warnings.
4556 This warning can also be turned on using @option{-gnatwa}.
4559 @emph{Suppress warnings on conditionals.}
4560 @cindex @option{-gnatwC} (@code{gcc})
4561 This switch suppresses warnings for conditional expressions used in
4562 tests that are known to be True or False at compile time.
4565 @emph{Activate warnings on implicit dereferencing.}
4566 @cindex @option{-gnatwd} (@code{gcc})
4567 If this switch is set, then the use of a prefix of an access type
4568 in an indexed component, slice, or selected component without an
4569 explicit @code{.all} will generate a warning. With this warning
4570 enabled, access checks occur only at points where an explicit
4571 @code{.all} appears in the source code (assuming no warnings are
4572 generated as a result of this switch). The default is that such
4573 warnings are not generated.
4574 Note that @option{-gnatwa} does not affect the setting of
4575 this warning option.
4578 @emph{Suppress warnings on implicit dereferencing.}
4579 @cindex @option{-gnatwD} (@code{gcc})
4580 @cindex Implicit dereferencing
4581 @cindex Dereferencing, implicit
4582 This switch suppresses warnings for implicit dereferences in
4583 indexed components, slices, and selected components.
4586 @emph{Treat warnings as errors.}
4587 @cindex @option{-gnatwe} (@code{gcc})
4588 @cindex Warnings, treat as error
4589 This switch causes warning messages to be treated as errors.
4590 The warning string still appears, but the warning messages are counted
4591 as errors, and prevent the generation of an object file.
4594 @emph{Activate warnings on unreferenced formals.}
4595 @cindex @option{-gnatwf} (@code{gcc})
4596 @cindex Formals, unreferenced
4597 This switch causes a warning to be generated if a formal parameter
4598 is not referenced in the body of the subprogram. This warning can
4599 also be turned on using @option{-gnatwa} or @option{-gnatwu}.
4602 @emph{Suppress warnings on unreferenced formals.}
4603 @cindex @option{-gnatwF} (@code{gcc})
4604 This switch suppresses warnings for unreferenced formal
4605 parameters. Note that the
4606 combination @option{-gnatwu} followed by @option{-gnatwF} has the
4607 effect of warning on unreferenced entities other than subprogram
4611 @emph{Activate warnings on unrecognized pragmas.}
4612 @cindex @option{-gnatwg} (@code{gcc})
4613 @cindex Pragmas, unrecognized
4614 This switch causes a warning to be generated if an unrecognized
4615 pragma is encountered. Apart from issuing this warning, the
4616 pragma is ignored and has no effect. This warning can
4617 also be turned on using @option{-gnatwa}. The default
4618 is that such warnings are issued (satisfying the Ada Reference
4619 Manual requirement that such warnings appear).
4622 @emph{Suppress warnings on unrecognized pragmas.}
4623 @cindex @option{-gnatwG} (@code{gcc})
4624 This switch suppresses warnings for unrecognized pragmas.
4627 @emph{Activate warnings on hiding.}
4628 @cindex @option{-gnatwh} (@code{gcc})
4629 @cindex Hiding of Declarations
4630 This switch activates warnings on hiding declarations.
4631 A declaration is considered hiding
4632 if it is for a non-overloadable entity, and it declares an entity with the
4633 same name as some other entity that is directly or use-visible. The default
4634 is that such warnings are not generated.
4635 Note that @option{-gnatwa} does not affect the setting of this warning option.
4638 @emph{Suppress warnings on hiding.}
4639 @cindex @option{-gnatwH} (@code{gcc})
4640 This switch suppresses warnings on hiding declarations.
4643 @emph{Activate warnings on implementation units.}
4644 @cindex @option{-gnatwi} (@code{gcc})
4645 This switch activates warnings for a @code{with} of an internal GNAT
4646 implementation unit, defined as any unit from the @code{Ada},
4647 @code{Interfaces}, @code{GNAT},
4648 ^^@code{DEC},^ or @code{System}
4649 hierarchies that is not
4650 documented in either the Ada Reference Manual or the GNAT
4651 Programmer's Reference Manual. Such units are intended only
4652 for internal implementation purposes and should not be @code{with}'ed
4653 by user programs. The default is that such warnings are generated
4654 This warning can also be turned on using @option{-gnatwa}.
4657 @emph{Disable warnings on implementation units.}
4658 @cindex @option{-gnatwI} (@code{gcc})
4659 This switch disables warnings for a @code{with} of an internal GNAT
4660 implementation unit.
4663 @emph{Activate warnings on obsolescent features (Annex J).}
4664 @cindex @option{-gnatwj} (@code{gcc})
4665 @cindex Features, obsolescent
4666 @cindex Obsolescent features
4667 If this warning option is activated, then warnings are generated for
4668 calls to subprograms marked with @code{pragma Obsolescent} and
4669 for use of features in Annex J of the Ada Reference Manual. In the
4670 case of Annex J, not all features are flagged. In particular use
4671 of the renamed packages (like @code{Text_IO}) and use of package
4672 @code{ASCII} are not flagged, since these are very common and
4673 would generate many annoying positive warnings. The default is that
4674 such warnings are not generated.
4677 @emph{Suppress warnings on obsolescent features (Annex J).}
4678 @cindex @option{-gnatwJ} (@code{gcc})
4679 This switch disables warnings on use of obsolescent features.
4682 @emph{Activate warnings on variables that could be constants.}
4683 @cindex @option{-gnatwk} (@code{gcc})
4684 This switch activates warnings for variables that are initialized but
4685 never modified, and then could be declared constants.
4688 @emph{Suppress warnings on variables that could be constants.}
4689 @cindex @option{-gnatwK} (@code{gcc})
4690 This switch disables warnings on variables that could be declared constants.
4693 @emph{Activate warnings for missing elaboration pragmas.}
4694 @cindex @option{-gnatwl} (@code{gcc})
4695 @cindex Elaboration, warnings
4696 This switch activates warnings on missing
4697 @code{pragma Elaborate_All} statements.
4698 See the section in this guide on elaboration checking for details on
4699 when such pragma should be used. Warnings are also generated if you
4700 are using the static mode of elaboration, and a @code{pragma Elaborate}
4701 is encountered. The default is that such warnings
4703 This warning is not automatically turned on by the use of @option{-gnatwa}.
4706 @emph{Suppress warnings for missing elaboration pragmas.}
4707 @cindex @option{-gnatwL} (@code{gcc})
4708 This switch suppresses warnings on missing pragma Elaborate_All statements.
4709 See the section in this guide on elaboration checking for details on
4710 when such pragma should be used.
4713 @emph{Activate warnings on modified but unreferenced variables.}
4714 @cindex @option{-gnatwm} (@code{gcc})
4715 This switch activates warnings for variables that are assigned (using
4716 an initialization value or with one or more assignment statements) but
4717 whose value is never read. The warning is suppressed for volatile
4718 variables and also for variables that are renamings of other variables
4719 or for which an address clause is given.
4720 This warning can also be turned on using @option{-gnatwa}.
4723 @emph{Disable warnings on modified but unreferenced variables.}
4724 @cindex @option{-gnatwM} (@code{gcc})
4725 This switch disables warnings for variables that are assigned or
4726 initialized, but never read.
4729 @emph{Set normal warnings mode.}
4730 @cindex @option{-gnatwn} (@code{gcc})
4731 This switch sets normal warning mode, in which enabled warnings are
4732 issued and treated as warnings rather than errors. This is the default
4733 mode. the switch @option{-gnatwn} can be used to cancel the effect of
4734 an explicit @option{-gnatws} or
4735 @option{-gnatwe}. It also cancels the effect of the
4736 implicit @option{-gnatwe} that is activated by the
4737 use of @option{-gnatg}.
4740 @emph{Activate warnings on address clause overlays.}
4741 @cindex @option{-gnatwo} (@code{gcc})
4742 @cindex Address Clauses, warnings
4743 This switch activates warnings for possibly unintended initialization
4744 effects of defining address clauses that cause one variable to overlap
4745 another. The default is that such warnings are generated.
4746 This warning can also be turned on using @option{-gnatwa}.
4749 @emph{Suppress warnings on address clause overlays.}
4750 @cindex @option{-gnatwO} (@code{gcc})
4751 This switch suppresses warnings on possibly unintended initialization
4752 effects of defining address clauses that cause one variable to overlap
4756 @emph{Activate warnings on ineffective pragma Inlines.}
4757 @cindex @option{-gnatwp} (@code{gcc})
4758 @cindex Inlining, warnings
4759 This switch activates warnings for failure of front end inlining
4760 (activated by @option{-gnatN}) to inline a particular call. There are
4761 many reasons for not being able to inline a call, including most
4762 commonly that the call is too complex to inline.
4763 This warning can also be turned on using @option{-gnatwa}.
4766 @emph{Suppress warnings on ineffective pragma Inlines.}
4767 @cindex @option{-gnatwP} (@code{gcc})
4768 This switch suppresses warnings on ineffective pragma Inlines. If the
4769 inlining mechanism cannot inline a call, it will simply ignore the
4773 @emph{Activate warnings on redundant constructs.}
4774 @cindex @option{-gnatwr} (@code{gcc})
4775 This switch activates warnings for redundant constructs. The following
4776 is the current list of constructs regarded as redundant:
4777 This warning can also be turned on using @option{-gnatwa}.
4781 Assignment of an item to itself.
4783 Type conversion that converts an expression to its own type.
4785 Use of the attribute @code{Base} where @code{typ'Base} is the same
4788 Use of pragma @code{Pack} when all components are placed by a record
4789 representation clause.
4791 Exception handler containing only a reraise statement (raise with no
4792 operand) which has no effect.
4794 Use of the operator abs on an operand that is known at compile time
4797 Use of an unnecessary extra level of parentheses (C-style) around conditions
4798 in @code{if} statements, @code{while} statements and @code{exit} statements.
4800 Comparison of boolean expressions to an explicit True value.
4804 @emph{Suppress warnings on redundant constructs.}
4805 @cindex @option{-gnatwR} (@code{gcc})
4806 This switch suppresses warnings for redundant constructs.
4809 @emph{Suppress all warnings.}
4810 @cindex @option{-gnatws} (@code{gcc})
4811 This switch completely suppresses the
4812 output of all warning messages from the GNAT front end.
4813 Note that it does not suppress warnings from the @code{gcc} back end.
4814 To suppress these back end warnings as well, use the switch @option{-w}
4815 in addition to @option{-gnatws}.
4818 @emph{Activate warnings on unused entities.}
4819 @cindex @option{-gnatwu} (@code{gcc})
4820 This switch activates warnings to be generated for entities that
4821 are declared but not referenced, and for units that are @code{with}'ed
4823 referenced. In the case of packages, a warning is also generated if
4824 no entities in the package are referenced. This means that if the package
4825 is referenced but the only references are in @code{use}
4826 clauses or @code{renames}
4827 declarations, a warning is still generated. A warning is also generated
4828 for a generic package that is @code{with}'ed but never instantiated.
4829 In the case where a package or subprogram body is compiled, and there
4830 is a @code{with} on the corresponding spec
4831 that is only referenced in the body,
4832 a warning is also generated, noting that the
4833 @code{with} can be moved to the body. The default is that
4834 such warnings are not generated.
4835 This switch also activates warnings on unreferenced formals
4836 (it is includes the effect of @option{-gnatwf}).
4837 This warning can also be turned on using @option{-gnatwa}.
4840 @emph{Suppress warnings on unused entities.}
4841 @cindex @option{-gnatwU} (@code{gcc})
4842 This switch suppresses warnings for unused entities and packages.
4843 It also turns off warnings on unreferenced formals (and thus includes
4844 the effect of @option{-gnatwF}).
4847 @emph{Activate warnings on unassigned variables.}
4848 @cindex @option{-gnatwv} (@code{gcc})
4849 @cindex Unassigned variable warnings
4850 This switch activates warnings for access to variables which
4851 may not be properly initialized. The default is that
4852 such warnings are generated.
4855 @emph{Suppress warnings on unassigned variables.}
4856 @cindex @option{-gnatwV} (@code{gcc})
4857 This switch suppresses warnings for access to variables which
4858 may not be properly initialized.
4861 @emph{Activate warnings on Export/Import pragmas.}
4862 @cindex @option{-gnatwx} (@code{gcc})
4863 @cindex Export/Import pragma warnings
4864 This switch activates warnings on Export/Import pragmas when
4865 the compiler detects a possible conflict between the Ada and
4866 foreign language calling sequences. For example, the use of
4867 default parameters in a convention C procedure is dubious
4868 because the C compiler cannot supply the proper default, so
4869 a warning is issued. The default is that such warnings are
4873 @emph{Suppress warnings on Export/Import pragmas.}
4874 @cindex @option{-gnatwX} (@code{gcc})
4875 This switch suppresses warnings on Export/Import pragmas.
4876 The sense of this is that you are telling the compiler that
4877 you know what you are doing in writing the pragma, and it
4878 should not complain at you.
4881 @emph{Activate warnings on unchecked conversions.}
4882 @cindex @option{-gnatwz} (@code{gcc})
4883 @cindex Unchecked_Conversion warnings
4884 This switch activates warnings for unchecked conversions
4885 where the types are known at compile time to have different
4887 is that such warnings are generated.
4890 @emph{Suppress warnings on unchecked conversions.}
4891 @cindex @option{-gnatwZ} (@code{gcc})
4892 This switch suppresses warnings for unchecked conversions
4893 where the types are known at compile time to have different
4896 @item ^-Wuninitialized^WARNINGS=UNINITIALIZED^
4897 @cindex @option{-Wuninitialized}
4898 The warnings controlled by the @option{-gnatw} switch are generated by the
4899 front end of the compiler. In some cases, the @option{^gcc^GCC^} back end
4900 can provide additional warnings. One such useful warning is provided by
4901 @option{^-Wuninitialized^WARNINGS=UNINITIALIZED^}. This must be used in
4902 conjunction with tunrning on optimization mode. This causes the flow
4903 analysis circuits of the back end optimizer to output additional
4904 warnings about uninitialized variables.
4906 @item ^-w^/NO_BACK_END_WARNINGS^
4908 This switch suppresses warnings from the @option{^gcc^GCC^} back end. It may
4909 be used in conjunction with @option{-gnatws} to ensure that all warnings
4910 are suppressed during the entire compilation process.
4916 A string of warning parameters can be used in the same parameter. For example:
4923 will turn on all optional warnings except for elaboration pragma warnings,
4924 and also specify that warnings should be treated as errors.
4926 When no switch @option{^-gnatw^/WARNINGS^} is used, this is equivalent to:
4952 @node Debugging and Assertion Control
4953 @subsection Debugging and Assertion Control
4957 @cindex @option{-gnata} (@code{gcc})
4963 The pragmas @code{Assert} and @code{Debug} normally have no effect and
4964 are ignored. This switch, where @samp{a} stands for assert, causes
4965 @code{Assert} and @code{Debug} pragmas to be activated.
4967 The pragmas have the form:
4971 @b{pragma} Assert (@var{Boolean-expression} [,
4972 @var{static-string-expression}])
4973 @b{pragma} Debug (@var{procedure call})
4978 The @code{Assert} pragma causes @var{Boolean-expression} to be tested.
4979 If the result is @code{True}, the pragma has no effect (other than
4980 possible side effects from evaluating the expression). If the result is
4981 @code{False}, the exception @code{Assert_Failure} declared in the package
4982 @code{System.Assertions} is
4983 raised (passing @var{static-string-expression}, if present, as the
4984 message associated with the exception). If no string expression is
4985 given the default is a string giving the file name and line number
4988 The @code{Debug} pragma causes @var{procedure} to be called. Note that
4989 @code{pragma Debug} may appear within a declaration sequence, allowing
4990 debugging procedures to be called between declarations.
4993 @item /DEBUG[=debug-level]
4995 Specifies how much debugging information is to be included in
4996 the resulting object file where 'debug-level' is one of the following:
4999 Include both debugger symbol records and traceback
5001 This is the default setting.
5003 Include both debugger symbol records and traceback in
5006 Excludes both debugger symbol records and traceback
5007 the object file. Same as /NODEBUG.
5009 Includes only debugger symbol records in the object
5010 file. Note that this doesn't include traceback information.
5015 @node Validity Checking
5016 @subsection Validity Checking
5017 @findex Validity Checking
5020 The Ada 95 Reference Manual has specific requirements for checking
5021 for invalid values. In particular, RM 13.9.1 requires that the
5022 evaluation of invalid values (for example from unchecked conversions),
5023 not result in erroneous execution. In GNAT, the result of such an
5024 evaluation in normal default mode is to either use the value
5025 unmodified, or to raise Constraint_Error in those cases where use
5026 of the unmodified value would cause erroneous execution. The cases
5027 where unmodified values might lead to erroneous execution are case
5028 statements (where a wild jump might result from an invalid value),
5029 and subscripts on the left hand side (where memory corruption could
5030 occur as a result of an invalid value).
5032 The @option{-gnatV^@var{x}^^} switch allows more control over the validity
5035 The @code{x} argument is a string of letters that
5036 indicate validity checks that are performed or not performed in addition
5037 to the default checks described above.
5040 The options allowed for this qualifier
5041 indicate validity checks that are performed or not performed in addition
5042 to the default checks described above.
5049 @emph{All validity checks.}
5050 @cindex @option{-gnatVa} (@code{gcc})
5051 All validity checks are turned on.
5053 That is, @option{-gnatVa} is
5054 equivalent to @option{gnatVcdfimorst}.
5058 @emph{Validity checks for copies.}
5059 @cindex @option{-gnatVc} (@code{gcc})
5060 The right hand side of assignments, and the initializing values of
5061 object declarations are validity checked.
5064 @emph{Default (RM) validity checks.}
5065 @cindex @option{-gnatVd} (@code{gcc})
5066 Some validity checks are done by default following normal Ada semantics
5068 A check is done in case statements that the expression is within the range
5069 of the subtype. If it is not, Constraint_Error is raised.
5070 For assignments to array components, a check is done that the expression used
5071 as index is within the range. If it is not, Constraint_Error is raised.
5072 Both these validity checks may be turned off using switch @option{-gnatVD}.
5073 They are turned on by default. If @option{-gnatVD} is specified, a subsequent
5074 switch @option{-gnatVd} will leave the checks turned on.
5075 Switch @option{-gnatVD} should be used only if you are sure that all such
5076 expressions have valid values. If you use this switch and invalid values
5077 are present, then the program is erroneous, and wild jumps or memory
5078 overwriting may occur.
5081 @emph{Validity checks for floating-point values.}
5082 @cindex @option{-gnatVf} (@code{gcc})
5083 In the absence of this switch, validity checking occurs only for discrete
5084 values. If @option{-gnatVf} is specified, then validity checking also applies
5085 for floating-point values, and NaN's and infinities are considered invalid,
5086 as well as out of range values for constrained types. Note that this means
5087 that standard @code{IEEE} infinity mode is not allowed. The exact contexts
5088 in which floating-point values are checked depends on the setting of other
5089 options. For example,
5090 @option{^-gnatVif^VALIDITY_CHECKING=(IN_PARAMS,FLOATS)^} or
5091 @option{^-gnatVfi^VALIDITY_CHECKING=(FLOATS,IN_PARAMS)^}
5092 (the order does not matter) specifies that floating-point parameters of mode
5093 @code{in} should be validity checked.
5096 @emph{Validity checks for @code{in} mode parameters}
5097 @cindex @option{-gnatVi} (@code{gcc})
5098 Arguments for parameters of mode @code{in} are validity checked in function
5099 and procedure calls at the point of call.
5102 @emph{Validity checks for @code{in out} mode parameters.}
5103 @cindex @option{-gnatVm} (@code{gcc})
5104 Arguments for parameters of mode @code{in out} are validity checked in
5105 procedure calls at the point of call. The @code{'m'} here stands for
5106 modify, since this concerns parameters that can be modified by the call.
5107 Note that there is no specific option to test @code{out} parameters,
5108 but any reference within the subprogram will be tested in the usual
5109 manner, and if an invalid value is copied back, any reference to it
5110 will be subject to validity checking.
5113 @emph{No validity checks.}
5114 @cindex @option{-gnatVn} (@code{gcc})
5115 This switch turns off all validity checking, including the default checking
5116 for case statements and left hand side subscripts. Note that the use of
5117 the switch @option{-gnatp} suppresses all run-time checks, including
5118 validity checks, and thus implies @option{-gnatVn}. When this switch
5119 is used, it cancels any other @option{-gnatV} previously issued.
5122 @emph{Validity checks for operator and attribute operands.}
5123 @cindex @option{-gnatVo} (@code{gcc})
5124 Arguments for predefined operators and attributes are validity checked.
5125 This includes all operators in package @code{Standard},
5126 the shift operators defined as intrinsic in package @code{Interfaces}
5127 and operands for attributes such as @code{Pos}. Checks are also made
5128 on individual component values for composite comparisons.
5131 @emph{Validity checks for parameters.}
5132 @cindex @option{-gnatVp} (@code{gcc})
5133 This controls the treatment of parameters within a subprogram (as opposed
5134 to @option{-gnatVi} and @option{-gnatVm} which control validity testing
5135 of parameters on a call. If either of these call options is used, then
5136 normally an assumption is made within a subprogram that the input arguments
5137 have been validity checking at the point of call, and do not need checking
5138 again within a subprogram). If @option{-gnatVp} is set, then this assumption
5139 is not made, and parameters are not assumed to be valid, so their validity
5140 will be checked (or rechecked) within the subprogram.
5143 @emph{Validity checks for function returns.}
5144 @cindex @option{-gnatVr} (@code{gcc})
5145 The expression in @code{return} statements in functions is validity
5149 @emph{Validity checks for subscripts.}
5150 @cindex @option{-gnatVs} (@code{gcc})
5151 All subscripts expressions are checked for validity, whether they appear
5152 on the right side or left side (in default mode only left side subscripts
5153 are validity checked).
5156 @emph{Validity checks for tests.}
5157 @cindex @option{-gnatVt} (@code{gcc})
5158 Expressions used as conditions in @code{if}, @code{while} or @code{exit}
5159 statements are checked, as well as guard expressions in entry calls.
5164 The @option{-gnatV} switch may be followed by
5165 ^a string of letters^a list of options^
5166 to turn on a series of validity checking options.
5168 @option{^-gnatVcr^/VALIDITY_CHECKING=(COPIES, RETURNS)^}
5169 specifies that in addition to the default validity checking, copies and
5170 function return expressions are to be validity checked.
5171 In order to make it easier
5172 to specify the desired combination of effects,
5174 the upper case letters @code{CDFIMORST} may
5175 be used to turn off the corresponding lower case option.
5178 the prefix @code{NO} on an option turns off the corresponding validity
5181 @item @code{NOCOPIES}
5182 @item @code{NODEFAULT}
5183 @item @code{NOFLOATS}
5184 @item @code{NOIN_PARAMS}
5185 @item @code{NOMOD_PARAMS}
5186 @item @code{NOOPERANDS}
5187 @item @code{NORETURNS}
5188 @item @code{NOSUBSCRIPTS}
5189 @item @code{NOTESTS}
5193 @option{^-gnatVaM^/VALIDITY_CHECKING=(ALL, NOMOD_PARAMS)^}
5194 turns on all validity checking options except for
5195 checking of @code{@b{in out}} procedure arguments.
5197 The specification of additional validity checking generates extra code (and
5198 in the case of @option{-gnatVa} the code expansion can be substantial.
5199 However, these additional checks can be very useful in detecting
5200 uninitialized variables, incorrect use of unchecked conversion, and other
5201 errors leading to invalid values. The use of pragma @code{Initialize_Scalars}
5202 is useful in conjunction with the extra validity checking, since this
5203 ensures that wherever possible uninitialized variables have invalid values.
5205 See also the pragma @code{Validity_Checks} which allows modification of
5206 the validity checking mode at the program source level, and also allows for
5207 temporary disabling of validity checks.
5210 @node Style Checking
5211 @subsection Style Checking
5212 @findex Style checking
5215 The @option{-gnaty^x^(option,option,...)^} switch
5216 @cindex @option{-gnaty} (@code{gcc})
5217 causes the compiler to
5218 enforce specified style rules. A limited set of style rules has been used
5219 in writing the GNAT sources themselves. This switch allows user programs
5220 to activate all or some of these checks. If the source program fails a
5221 specified style check, an appropriate warning message is given, preceded by
5222 the character sequence ``(style)''.
5224 @code{(option,option,...)} is a sequence of keywords
5227 The string @var{x} is a sequence of letters or digits
5229 indicating the particular style
5230 checks to be performed. The following checks are defined:
5235 @emph{Specify indentation level.}
5236 If a digit from 1-9 appears
5237 ^in the string after @option{-gnaty}^as an option for /STYLE_CHECKS^
5238 then proper indentation is checked, with the digit indicating the
5239 indentation level required.
5240 The general style of required indentation is as specified by
5241 the examples in the Ada Reference Manual. Full line comments must be
5242 aligned with the @code{--} starting on a column that is a multiple of
5243 the alignment level.
5246 @emph{Check attribute casing.}
5247 If the ^letter a^word ATTRIBUTE^ appears in the string after @option{-gnaty}
5248 then attribute names, including the case of keywords such as @code{digits}
5249 used as attributes names, must be written in mixed case, that is, the
5250 initial letter and any letter following an underscore must be uppercase.
5251 All other letters must be lowercase.
5254 @emph{Blanks not allowed at statement end.}
5255 If the ^letter b^word BLANKS^ appears in the string after @option{-gnaty} then
5256 trailing blanks are not allowed at the end of statements. The purpose of this
5257 rule, together with h (no horizontal tabs), is to enforce a canonical format
5258 for the use of blanks to separate source tokens.
5261 @emph{Check comments.}
5262 If the ^letter c^word COMMENTS^ appears in the string after @option{-gnaty}
5263 then comments must meet the following set of rules:
5268 The ``@code{--}'' that starts the column must either start in column one,
5269 or else at least one blank must precede this sequence.
5272 Comments that follow other tokens on a line must have at least one blank
5273 following the ``@code{--}'' at the start of the comment.
5276 Full line comments must have two blanks following the ``@code{--}'' that
5277 starts the comment, with the following exceptions.
5280 A line consisting only of the ``@code{--}'' characters, possibly preceded
5281 by blanks is permitted.
5284 A comment starting with ``@code{--x}'' where @code{x} is a special character
5286 This allows proper processing of the output generated by specialized tools
5287 including @command{gnatprep} (where ``@code{--!}'' is used) and the SPARK
5289 language (where ``@code{--#}'' is used). For the purposes of this rule, a
5290 special character is defined as being in one of the ASCII ranges
5291 @code{16#21#..16#2F#} or @code{16#3A#..16#3F#}.
5292 Note that this usage is not permitted
5293 in GNAT implementation units (i.e. when @option{-gnatg} is used).
5296 A line consisting entirely of minus signs, possibly preceded by blanks, is
5297 permitted. This allows the construction of box comments where lines of minus
5298 signs are used to form the top and bottom of the box.
5301 If a comment starts and ends with ``@code{--}'' is permitted as long as at
5302 least one blank follows the initial ``@code{--}''. Together with the preceding
5303 rule, this allows the construction of box comments, as shown in the following
5306 ---------------------------
5307 -- This is a box comment --
5308 -- with two text lines. --
5309 ---------------------------
5314 @emph{Check end/exit labels.}
5315 If the ^letter e^word END^ appears in the string after @option{-gnaty} then
5316 optional labels on @code{end} statements ending subprograms and on
5317 @code{exit} statements exiting named loops, are required to be present.
5320 @emph{No form feeds or vertical tabs.}
5321 If the ^letter f^word VTABS^ appears in the string after @option{-gnaty} then
5322 neither form feeds nor vertical tab characters are not permitted
5326 @emph{No horizontal tabs.}
5327 If the ^letter h^word HTABS^ appears in the string after @option{-gnaty} then
5328 horizontal tab characters are not permitted in the source text.
5329 Together with the b (no blanks at end of line) check, this
5330 enforces a canonical form for the use of blanks to separate
5334 @emph{Check if-then layout.}
5335 If the ^letter i^word IF_THEN^ appears in the string after @option{-gnaty},
5336 then the keyword @code{then} must appear either on the same
5337 line as corresponding @code{if}, or on a line on its own, lined
5338 up under the @code{if} with at least one non-blank line in between
5339 containing all or part of the condition to be tested.
5342 @emph{Check keyword casing.}
5343 If the ^letter k^word KEYWORD^ appears in the string after @option{-gnaty} then
5344 all keywords must be in lower case (with the exception of keywords
5345 such as @code{digits} used as attribute names to which this check
5349 @emph{Check layout.}
5350 If the ^letter l^word LAYOUT^ appears in the string after @option{-gnaty} then
5351 layout of statement and declaration constructs must follow the
5352 recommendations in the Ada Reference Manual, as indicated by the
5353 form of the syntax rules. For example an @code{else} keyword must
5354 be lined up with the corresponding @code{if} keyword.
5356 There are two respects in which the style rule enforced by this check
5357 option are more liberal than those in the Ada Reference Manual. First
5358 in the case of record declarations, it is permissible to put the
5359 @code{record} keyword on the same line as the @code{type} keyword, and
5360 then the @code{end} in @code{end record} must line up under @code{type}.
5361 For example, either of the following two layouts is acceptable:
5363 @smallexample @c ada
5379 Second, in the case of a block statement, a permitted alternative
5380 is to put the block label on the same line as the @code{declare} or
5381 @code{begin} keyword, and then line the @code{end} keyword up under
5382 the block label. For example both the following are permitted:
5384 @smallexample @c ada
5402 The same alternative format is allowed for loops. For example, both of
5403 the following are permitted:
5405 @smallexample @c ada
5407 Clear : while J < 10 loop
5418 @item ^m^LINE_LENGTH^
5419 @emph{Check maximum line length.}
5420 If the ^letter m^word LINE_LENGTH^ appears in the string after @option{-gnaty}
5421 then the length of source lines must not exceed 79 characters, including
5422 any trailing blanks. The value of 79 allows convenient display on an
5423 80 character wide device or window, allowing for possible special
5424 treatment of 80 character lines. Note that this count is of raw
5425 characters in the source text. This means that a tab character counts
5426 as one character in this count and a wide character sequence counts as
5427 several characters (however many are needed in the encoding).
5429 @item ^Mnnn^MAX_LENGTH=nnn^
5430 @emph{Set maximum line length.}
5431 If the sequence ^M^MAX_LENGTH=^nnn, where nnn is a decimal number, appears in
5432 the string after @option{-gnaty} then the length of lines must not exceed the
5435 @item ^n^STANDARD_CASING^
5436 @emph{Check casing of entities in Standard.}
5437 If the ^letter n^word STANDARD_CASING^ appears in the string
5438 after @option{-gnaty} then any identifier from Standard must be cased
5439 to match the presentation in the Ada Reference Manual (for example,
5440 @code{Integer} and @code{ASCII.NUL}).
5442 @item ^o^ORDERED_SUBPROGRAMS^
5443 @emph{Check order of subprogram bodies.}
5444 If the ^letter o^word ORDERED_SUBPROGRAMS^ appears in the string
5445 after @option{-gnaty} then all subprogram bodies in a given scope
5446 (e.g. a package body) must be in alphabetical order. The ordering
5447 rule uses normal Ada rules for comparing strings, ignoring casing
5448 of letters, except that if there is a trailing numeric suffix, then
5449 the value of this suffix is used in the ordering (e.g. Junk2 comes
5453 @emph{Check pragma casing.}
5454 If the ^letter p^word PRAGMA^ appears in the string after @option{-gnaty} then
5455 pragma names must be written in mixed case, that is, the
5456 initial letter and any letter following an underscore must be uppercase.
5457 All other letters must be lowercase.
5459 @item ^r^REFERENCES^
5460 @emph{Check references.}
5461 If the ^letter r^word REFERENCES^ appears in the string after @option{-gnaty}
5462 then all identifier references must be cased in the same way as the
5463 corresponding declaration. No specific casing style is imposed on
5464 identifiers. The only requirement is for consistency of references
5468 @emph{Check separate specs.}
5469 If the ^letter s^word SPECS^ appears in the string after @option{-gnaty} then
5470 separate declarations (``specs'') are required for subprograms (a
5471 body is not allowed to serve as its own declaration). The only
5472 exception is that parameterless library level procedures are
5473 not required to have a separate declaration. This exception covers
5474 the most frequent form of main program procedures.
5477 @emph{Check token spacing.}
5478 If the ^letter t^word TOKEN^ appears in the string after @option{-gnaty} then
5479 the following token spacing rules are enforced:
5484 The keywords @code{@b{abs}} and @code{@b{not}} must be followed by a space.
5487 The token @code{=>} must be surrounded by spaces.
5490 The token @code{<>} must be preceded by a space or a left parenthesis.
5493 Binary operators other than @code{**} must be surrounded by spaces.
5494 There is no restriction on the layout of the @code{**} binary operator.
5497 Colon must be surrounded by spaces.
5500 Colon-equal (assignment, initialization) must be surrounded by spaces.
5503 Comma must be the first non-blank character on the line, or be
5504 immediately preceded by a non-blank character, and must be followed
5508 If the token preceding a left parenthesis ends with a letter or digit, then
5509 a space must separate the two tokens.
5512 A right parenthesis must either be the first non-blank character on
5513 a line, or it must be preceded by a non-blank character.
5516 A semicolon must not be preceded by a space, and must not be followed by
5517 a non-blank character.
5520 A unary plus or minus may not be followed by a space.
5523 A vertical bar must be surrounded by spaces.
5527 In the above rules, appearing in column one is always permitted, that is,
5528 counts as meeting either a requirement for a required preceding space,
5529 or as meeting a requirement for no preceding space.
5531 Appearing at the end of a line is also always permitted, that is, counts
5532 as meeting either a requirement for a following space, or as meeting
5533 a requirement for no following space.
5538 If any of these style rules is violated, a message is generated giving
5539 details on the violation. The initial characters of such messages are
5540 always ``@code{(style)}''. Note that these messages are treated as warning
5541 messages, so they normally do not prevent the generation of an object
5542 file. The @option{-gnatwe} switch can be used to treat warning messages,
5543 including style messages, as fatal errors.
5547 @option{-gnaty} on its own (that is not
5548 followed by any letters or digits),
5549 is equivalent to @code{gnaty3abcefhiklmprst}, that is all checking
5550 options enabled with the exception of -gnatyo,
5553 /STYLE_CHECKS=ALL_BUILTIN enables all checking options with
5554 the exception of ORDERED_SUBPROGRAMS,
5556 with an indentation level of 3. This is the standard
5557 checking option that is used for the GNAT sources.
5566 clears any previously set style checks.
5568 @node Run-Time Checks
5569 @subsection Run-Time Checks
5570 @cindex Division by zero
5571 @cindex Access before elaboration
5572 @cindex Checks, division by zero
5573 @cindex Checks, access before elaboration
5576 If you compile with the default options, GNAT will insert many run-time
5577 checks into the compiled code, including code that performs range
5578 checking against constraints, but not arithmetic overflow checking for
5579 integer operations (including division by zero) or checks for access
5580 before elaboration on subprogram calls. All other run-time checks, as
5581 required by the Ada 95 Reference Manual, are generated by default.
5582 The following @code{gcc} switches refine this default behavior:
5587 @cindex @option{-gnatp} (@code{gcc})
5588 @cindex Suppressing checks
5589 @cindex Checks, suppressing
5591 Suppress all run-time checks as though @code{pragma Suppress (all_checks})
5592 had been present in the source. Validity checks are also suppressed (in
5593 other words @option{-gnatp} also implies @option{-gnatVn}.
5594 Use this switch to improve the performance
5595 of the code at the expense of safety in the presence of invalid data or
5599 @cindex @option{-gnato} (@code{gcc})
5600 @cindex Overflow checks
5601 @cindex Check, overflow
5602 Enables overflow checking for integer operations.
5603 This causes GNAT to generate slower and larger executable
5604 programs by adding code to check for overflow (resulting in raising
5605 @code{Constraint_Error} as required by standard Ada
5606 semantics). These overflow checks correspond to situations in which
5607 the true value of the result of an operation may be outside the base
5608 range of the result type. The following example shows the distinction:
5610 @smallexample @c ada
5611 X1 : Integer := Integer'Last;
5612 X2 : Integer range 1 .. 5 := 5;
5613 X3 : Integer := Integer'Last;
5614 X4 : Integer range 1 .. 5 := 5;
5615 F : Float := 2.0E+20;
5624 Here the first addition results in a value that is outside the base range
5625 of Integer, and hence requires an overflow check for detection of the
5626 constraint error. Thus the first assignment to @code{X1} raises a
5627 @code{Constraint_Error} exception only if @option{-gnato} is set.
5629 The second increment operation results in a violation
5630 of the explicit range constraint, and such range checks are always
5631 performed (unless specifically suppressed with a pragma @code{suppress}
5632 or the use of @option{-gnatp}).
5634 The two conversions of @code{F} both result in values that are outside
5635 the base range of type @code{Integer} and thus will raise
5636 @code{Constraint_Error} exceptions only if @option{-gnato} is used.
5637 The fact that the result of the second conversion is assigned to
5638 variable @code{X4} with a restricted range is irrelevant, since the problem
5639 is in the conversion, not the assignment.
5641 Basically the rule is that in the default mode (@option{-gnato} not
5642 used), the generated code assures that all integer variables stay
5643 within their declared ranges, or within the base range if there is
5644 no declared range. This prevents any serious problems like indexes
5645 out of range for array operations.
5647 What is not checked in default mode is an overflow that results in
5648 an in-range, but incorrect value. In the above example, the assignments
5649 to @code{X1}, @code{X2}, @code{X3} all give results that are within the
5650 range of the target variable, but the result is wrong in the sense that
5651 it is too large to be represented correctly. Typically the assignment
5652 to @code{X1} will result in wrap around to the largest negative number.
5653 The conversions of @code{F} will result in some @code{Integer} value
5654 and if that integer value is out of the @code{X4} range then the
5655 subsequent assignment would generate an exception.
5657 @findex Machine_Overflows
5658 Note that the @option{-gnato} switch does not affect the code generated
5659 for any floating-point operations; it applies only to integer
5661 For floating-point, GNAT has the @code{Machine_Overflows}
5662 attribute set to @code{False} and the normal mode of operation is to
5663 generate IEEE NaN and infinite values on overflow or invalid operations
5664 (such as dividing 0.0 by 0.0).
5666 The reason that we distinguish overflow checking from other kinds of
5667 range constraint checking is that a failure of an overflow check can
5668 generate an incorrect value, but cannot cause erroneous behavior. This
5669 is unlike the situation with a constraint check on an array subscript,
5670 where failure to perform the check can result in random memory description,
5671 or the range check on a case statement, where failure to perform the check
5672 can cause a wild jump.
5674 Note again that @option{-gnato} is off by default, so overflow checking is
5675 not performed in default mode. This means that out of the box, with the
5676 default settings, GNAT does not do all the checks expected from the
5677 language description in the Ada Reference Manual. If you want all constraint
5678 checks to be performed, as described in this Manual, then you must
5679 explicitly use the -gnato switch either on the @code{gnatmake} or
5683 @cindex @option{-gnatE} (@code{gcc})
5684 @cindex Elaboration checks
5685 @cindex Check, elaboration
5686 Enables dynamic checks for access-before-elaboration
5687 on subprogram calls and generic instantiations.
5688 For full details of the effect and use of this switch,
5689 @xref{Compiling Using gcc}.
5694 The setting of these switches only controls the default setting of the
5695 checks. You may modify them using either @code{Suppress} (to remove
5696 checks) or @code{Unsuppress} (to add back suppressed checks) pragmas in
5699 @node Stack Overflow Checking
5700 @subsection Stack Overflow Checking
5701 @cindex Stack Overflow Checking
5702 @cindex -fstack-check
5705 For most operating systems, @code{gcc} does not perform stack overflow
5706 checking by default. This means that if the main environment task or
5707 some other task exceeds the available stack space, then unpredictable
5708 behavior will occur.
5710 To activate stack checking, compile all units with the gcc option
5711 @option{-fstack-check}. For example:
5714 gcc -c -fstack-check package1.adb
5718 Units compiled with this option will generate extra instructions to check
5719 that any use of the stack (for procedure calls or for declaring local
5720 variables in declare blocks) do not exceed the available stack space.
5721 If the space is exceeded, then a @code{Storage_Error} exception is raised.
5723 For declared tasks, the stack size is always controlled by the size
5724 given in an applicable @code{Storage_Size} pragma (or is set to
5725 the default size if no pragma is used.
5727 For the environment task, the stack size depends on
5728 system defaults and is unknown to the compiler. The stack
5729 may even dynamically grow on some systems, precluding the
5730 normal Ada semantics for stack overflow. In the worst case,
5731 unbounded stack usage, causes unbounded stack expansion
5732 resulting in the system running out of virtual memory.
5734 The stack checking may still work correctly if a fixed
5735 size stack is allocated, but this cannot be guaranteed.
5736 To ensure that a clean exception is signalled for stack
5737 overflow, set the environment variable
5738 @code{GNAT_STACK_LIMIT} to indicate the maximum
5739 stack area that can be used, as in:
5740 @cindex GNAT_STACK_LIMIT
5743 SET GNAT_STACK_LIMIT 1600
5747 The limit is given in kilobytes, so the above declaration would
5748 set the stack limit of the environment task to 1.6 megabytes.
5749 Note that the only purpose of this usage is to limit the amount
5750 of stack used by the environment task. If it is necessary to
5751 increase the amount of stack for the environment task, then this
5752 is an operating systems issue, and must be addressed with the
5753 appropriate operating systems commands.
5756 @node Using gcc for Syntax Checking
5757 @subsection Using @code{gcc} for Syntax Checking
5760 @cindex @option{-gnats} (@code{gcc})
5764 The @code{s} stands for ``syntax''.
5767 Run GNAT in syntax checking only mode. For
5768 example, the command
5771 $ gcc -c -gnats x.adb
5775 compiles file @file{x.adb} in syntax-check-only mode. You can check a
5776 series of files in a single command
5778 , and can use wild cards to specify such a group of files.
5779 Note that you must specify the @option{-c} (compile
5780 only) flag in addition to the @option{-gnats} flag.
5783 You may use other switches in conjunction with @option{-gnats}. In
5784 particular, @option{-gnatl} and @option{-gnatv} are useful to control the
5785 format of any generated error messages.
5787 When the source file is empty or contains only empty lines and/or comments,
5788 the output is a warning:
5791 $ gcc -c -gnats -x ada toto.txt
5792 toto.txt:1:01: warning: empty file, contains no compilation units
5796 Otherwise, the output is simply the error messages, if any. No object file or
5797 ALI file is generated by a syntax-only compilation. Also, no units other
5798 than the one specified are accessed. For example, if a unit @code{X}
5799 @code{with}'s a unit @code{Y}, compiling unit @code{X} in syntax
5800 check only mode does not access the source file containing unit
5803 @cindex Multiple units, syntax checking
5804 Normally, GNAT allows only a single unit in a source file. However, this
5805 restriction does not apply in syntax-check-only mode, and it is possible
5806 to check a file containing multiple compilation units concatenated
5807 together. This is primarily used by the @code{gnatchop} utility
5808 (@pxref{Renaming Files Using gnatchop}).
5812 @node Using gcc for Semantic Checking
5813 @subsection Using @code{gcc} for Semantic Checking
5816 @cindex @option{-gnatc} (@code{gcc})
5820 The @code{c} stands for ``check''.
5822 Causes the compiler to operate in semantic check mode,
5823 with full checking for all illegalities specified in the
5824 Ada 95 Reference Manual, but without generation of any object code
5825 (no object file is generated).
5827 Because dependent files must be accessed, you must follow the GNAT
5828 semantic restrictions on file structuring to operate in this mode:
5832 The needed source files must be accessible
5833 (@pxref{Search Paths and the Run-Time Library (RTL)}).
5836 Each file must contain only one compilation unit.
5839 The file name and unit name must match (@pxref{File Naming Rules}).
5842 The output consists of error messages as appropriate. No object file is
5843 generated. An @file{ALI} file is generated for use in the context of
5844 cross-reference tools, but this file is marked as not being suitable
5845 for binding (since no object file is generated).
5846 The checking corresponds exactly to the notion of
5847 legality in the Ada 95 Reference Manual.
5849 Any unit can be compiled in semantics-checking-only mode, including
5850 units that would not normally be compiled (subunits,
5851 and specifications where a separate body is present).
5854 @node Compiling Ada 83 Programs
5855 @subsection Compiling Ada 83 Programs
5857 @cindex Ada 83 compatibility
5859 @cindex @option{-gnat83} (@code{gcc})
5860 @cindex ACVC, Ada 83 tests
5863 Although GNAT is primarily an Ada 95 compiler, it accepts this switch to
5864 specify that an Ada 83 program is to be compiled in Ada 83 mode. If you specify
5865 this switch, GNAT rejects most Ada 95 extensions and applies Ada 83 semantics
5866 where this can be done easily.
5867 It is not possible to guarantee this switch does a perfect
5868 job; for example, some subtle tests, such as are
5869 found in earlier ACVC tests (and that have been removed from the ACATS suite
5870 for Ada 95), might not compile correctly.
5871 Nevertheless, this switch may be useful in some circumstances, for example
5872 where, due to contractual reasons, legacy code needs to be maintained
5873 using only Ada 83 features.
5875 With few exceptions (most notably the need to use @code{<>} on
5876 @cindex Generic formal parameters
5877 unconstrained generic formal parameters, the use of the new Ada 95
5878 reserved words, and the use of packages
5879 with optional bodies), it is not necessary to use the
5880 @option{-gnat83} switch when compiling Ada 83 programs, because, with rare
5881 exceptions, Ada 95 is upwardly compatible with Ada 83. This
5882 means that a correct Ada 83 program is usually also a correct Ada 95
5884 For further information, please refer to @ref{Compatibility and Porting Guide}.
5888 @node Character Set Control
5889 @subsection Character Set Control
5891 @item ^-gnati^/IDENTIFIER_CHARACTER_SET=^@var{c}
5892 @cindex @option{^-gnati^/IDENTIFIER_CHARACTER_SET^} (@code{gcc})
5895 Normally GNAT recognizes the Latin-1 character set in source program
5896 identifiers, as described in the Ada 95 Reference Manual.
5898 GNAT to recognize alternate character sets in identifiers. @var{c} is a
5899 single character ^^or word^ indicating the character set, as follows:
5903 ISO 8859-1 (Latin-1) identifiers
5906 ISO 8859-2 (Latin-2) letters allowed in identifiers
5909 ISO 8859-3 (Latin-3) letters allowed in identifiers
5912 ISO 8859-4 (Latin-4) letters allowed in identifiers
5915 ISO 8859-5 (Cyrillic) letters allowed in identifiers
5918 ISO 8859-15 (Latin-9) letters allowed in identifiers
5921 IBM PC letters (code page 437) allowed in identifiers
5924 IBM PC letters (code page 850) allowed in identifiers
5926 @item ^f^FULL_UPPER^
5927 Full upper-half codes allowed in identifiers
5930 No upper-half codes allowed in identifiers
5933 Wide-character codes (that is, codes greater than 255)
5934 allowed in identifiers
5937 @xref{Foreign Language Representation}, for full details on the
5938 implementation of these character sets.
5940 @item ^-gnatW^/WIDE_CHARACTER_ENCODING=^@var{e}
5941 @cindex @option{^-gnatW^/WIDE_CHARACTER_ENCODING^} (@code{gcc})
5942 Specify the method of encoding for wide characters.
5943 @var{e} is one of the following:
5948 Hex encoding (brackets coding also recognized)
5951 Upper half encoding (brackets encoding also recognized)
5954 Shift/JIS encoding (brackets encoding also recognized)
5957 EUC encoding (brackets encoding also recognized)
5960 UTF-8 encoding (brackets encoding also recognized)
5963 Brackets encoding only (default value)
5965 For full details on the these encoding
5966 methods see @xref{Wide Character Encodings}.
5967 Note that brackets coding is always accepted, even if one of the other
5968 options is specified, so for example @option{-gnatW8} specifies that both
5969 brackets and @code{UTF-8} encodings will be recognized. The units that are
5970 with'ed directly or indirectly will be scanned using the specified
5971 representation scheme, and so if one of the non-brackets scheme is
5972 used, it must be used consistently throughout the program. However,
5973 since brackets encoding is always recognized, it may be conveniently
5974 used in standard libraries, allowing these libraries to be used with
5975 any of the available coding schemes.
5976 scheme. If no @option{-gnatW?} parameter is present, then the default
5977 representation is Brackets encoding only.
5979 Note that the wide character representation that is specified (explicitly
5980 or by default) for the main program also acts as the default encoding used
5981 for Wide_Text_IO files if not specifically overridden by a WCEM form
5985 @node File Naming Control
5986 @subsection File Naming Control
5989 @item ^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{n}
5990 @cindex @option{-gnatk} (@code{gcc})
5991 Activates file name ``krunching''. @var{n}, a decimal integer in the range
5992 1-999, indicates the maximum allowable length of a file name (not
5993 including the @file{.ads} or @file{.adb} extension). The default is not
5994 to enable file name krunching.
5996 For the source file naming rules, @xref{File Naming Rules}.
6000 @node Subprogram Inlining Control
6001 @subsection Subprogram Inlining Control
6006 @cindex @option{-gnatn} (@code{gcc})
6008 The @code{n} here is intended to suggest the first syllable of the
6011 GNAT recognizes and processes @code{Inline} pragmas. However, for the
6012 inlining to actually occur, optimization must be enabled. To enable
6013 inlining of subprograms specified by pragma @code{Inline},
6014 you must also specify this switch.
6015 In the absence of this switch, GNAT does not attempt
6016 inlining and does not need to access the bodies of
6017 subprograms for which @code{pragma Inline} is specified if they are not
6018 in the current unit.
6020 If you specify this switch the compiler will access these bodies,
6021 creating an extra source dependency for the resulting object file, and
6022 where possible, the call will be inlined.
6023 For further details on when inlining is possible
6024 see @xref{Inlining of Subprograms}.
6027 @cindex @option{-gnatN} (@code{gcc})
6028 The front end inlining activated by this switch is generally more extensive,
6029 and quite often more effective than the standard @option{-gnatn} inlining mode.
6030 It will also generate additional dependencies.
6032 @option{-gnatN} automatically implies @option{-gnatn} so it is not necessary
6033 to specify both options.
6036 @node Auxiliary Output Control
6037 @subsection Auxiliary Output Control
6041 @cindex @option{-gnatt} (@code{gcc})
6042 @cindex Writing internal trees
6043 @cindex Internal trees, writing to file
6044 Causes GNAT to write the internal tree for a unit to a file (with the
6045 extension @file{.adt}.
6046 This not normally required, but is used by separate analysis tools.
6048 these tools do the necessary compilations automatically, so you should
6049 not have to specify this switch in normal operation.
6052 @cindex @option{-gnatu} (@code{gcc})
6053 Print a list of units required by this compilation on @file{stdout}.
6054 The listing includes all units on which the unit being compiled depends
6055 either directly or indirectly.
6058 @item -pass-exit-codes
6059 @cindex @option{-pass-exit-codes} (@code{gcc})
6060 If this switch is not used, the exit code returned by @code{gcc} when
6061 compiling multiple files indicates whether all source files have
6062 been successfully used to generate object files or not.
6064 When @option{-pass-exit-codes} is used, @code{gcc} exits with an extended
6065 exit status and allows an integrated development environment to better
6066 react to a compilation failure. Those exit status are:
6070 There was an error in at least one source file.
6072 At least one source file did not generate an object file.
6074 The compiler died unexpectedly (internal error for example).
6076 An object file has been generated for every source file.
6081 @node Debugging Control
6082 @subsection Debugging Control
6086 @cindex Debugging options
6089 @cindex @option{-gnatd} (@code{gcc})
6090 Activate internal debugging switches. @var{x} is a letter or digit, or
6091 string of letters or digits, which specifies the type of debugging
6092 outputs desired. Normally these are used only for internal development
6093 or system debugging purposes. You can find full documentation for these
6094 switches in the body of the @code{Debug} unit in the compiler source
6095 file @file{debug.adb}.
6099 @cindex @option{-gnatG} (@code{gcc})
6100 This switch causes the compiler to generate auxiliary output containing
6101 a pseudo-source listing of the generated expanded code. Like most Ada
6102 compilers, GNAT works by first transforming the high level Ada code into
6103 lower level constructs. For example, tasking operations are transformed
6104 into calls to the tasking run-time routines. A unique capability of GNAT
6105 is to list this expanded code in a form very close to normal Ada source.
6106 This is very useful in understanding the implications of various Ada
6107 usage on the efficiency of the generated code. There are many cases in
6108 Ada (e.g. the use of controlled types), where simple Ada statements can
6109 generate a lot of run-time code. By using @option{-gnatG} you can identify
6110 these cases, and consider whether it may be desirable to modify the coding
6111 approach to improve efficiency.
6113 The format of the output is very similar to standard Ada source, and is
6114 easily understood by an Ada programmer. The following special syntactic
6115 additions correspond to low level features used in the generated code that
6116 do not have any exact analogies in pure Ada source form. The following
6117 is a partial list of these special constructions. See the specification
6118 of package @code{Sprint} in file @file{sprint.ads} for a full list.
6121 @item new @var{xxx} [storage_pool = @var{yyy}]
6122 Shows the storage pool being used for an allocator.
6124 @item at end @var{procedure-name};
6125 Shows the finalization (cleanup) procedure for a scope.
6127 @item (if @var{expr} then @var{expr} else @var{expr})
6128 Conditional expression equivalent to the @code{x?y:z} construction in C.
6130 @item @var{target}^^^(@var{source})
6131 A conversion with floating-point truncation instead of rounding.
6133 @item @var{target}?(@var{source})
6134 A conversion that bypasses normal Ada semantic checking. In particular
6135 enumeration types and fixed-point types are treated simply as integers.
6137 @item @var{target}?^^^(@var{source})
6138 Combines the above two cases.
6140 @item @var{x} #/ @var{y}
6141 @itemx @var{x} #mod @var{y}
6142 @itemx @var{x} #* @var{y}
6143 @itemx @var{x} #rem @var{y}
6144 A division or multiplication of fixed-point values which are treated as
6145 integers without any kind of scaling.
6147 @item free @var{expr} [storage_pool = @var{xxx}]
6148 Shows the storage pool associated with a @code{free} statement.
6150 @item freeze @var{typename} [@var{actions}]
6151 Shows the point at which @var{typename} is frozen, with possible
6152 associated actions to be performed at the freeze point.
6154 @item reference @var{itype}
6155 Reference (and hence definition) to internal type @var{itype}.
6157 @item @var{function-name}! (@var{arg}, @var{arg}, @var{arg})
6158 Intrinsic function call.
6160 @item @var{labelname} : label
6161 Declaration of label @var{labelname}.
6163 @item @var{expr} && @var{expr} && @var{expr} ... && @var{expr}
6164 A multiple concatenation (same effect as @var{expr} & @var{expr} &
6165 @var{expr}, but handled more efficiently).
6167 @item [constraint_error]
6168 Raise the @code{Constraint_Error} exception.
6170 @item @var{expression}'reference
6171 A pointer to the result of evaluating @var{expression}.
6173 @item @var{target-type}!(@var{source-expression})
6174 An unchecked conversion of @var{source-expression} to @var{target-type}.
6176 @item [@var{numerator}/@var{denominator}]
6177 Used to represent internal real literals (that) have no exact
6178 representation in base 2-16 (for example, the result of compile time
6179 evaluation of the expression 1.0/27.0).
6183 @cindex @option{-gnatD} (@code{gcc})
6184 When used in conjunction with @option{-gnatG}, this switch causes
6185 the expanded source, as described above for
6186 @option{-gnatG} to be written to files with names
6187 @file{^xxx.dg^XXX_DG^}, where @file{xxx} is the normal file name,
6188 instead of to the standard ooutput file. For
6189 example, if the source file name is @file{hello.adb}, then a file
6190 @file{^hello.adb.dg^HELLO.ADB_DG^} will be written. The debugging
6191 information generated by the @code{gcc} @option{^-g^/DEBUG^} switch
6192 will refer to the generated @file{^xxx.dg^XXX_DG^} file. This allows
6193 you to do source level debugging using the generated code which is
6194 sometimes useful for complex code, for example to find out exactly
6195 which part of a complex construction raised an exception. This switch
6196 also suppress generation of cross-reference information (see
6197 @option{-gnatx}) since otherwise the cross-reference information
6198 would refer to the @file{^.dg^.DG^} file, which would cause
6199 confusion since this is not the original source file.
6201 Note that @option{-gnatD} actually implies @option{-gnatG}
6202 automatically, so it is not necessary to give both options.
6203 In other words @option{-gnatD} is equivalent to @option{-gnatDG}).
6206 @item -gnatR[0|1|2|3[s]]
6207 @cindex @option{-gnatR} (@code{gcc})
6208 This switch controls output from the compiler of a listing showing
6209 representation information for declared types and objects. For
6210 @option{-gnatR0}, no information is output (equivalent to omitting
6211 the @option{-gnatR} switch). For @option{-gnatR1} (which is the default,
6212 so @option{-gnatR} with no parameter has the same effect), size and alignment
6213 information is listed for declared array and record types. For
6214 @option{-gnatR2}, size and alignment information is listed for all
6215 expression information for values that are computed at run time for
6216 variant records. These symbolic expressions have a mostly obvious
6217 format with #n being used to represent the value of the n'th
6218 discriminant. See source files @file{repinfo.ads/adb} in the
6219 @code{GNAT} sources for full details on the format of @option{-gnatR3}
6220 output. If the switch is followed by an s (e.g. @option{-gnatR2s}), then
6221 the output is to a file with the name @file{^file.rep^file_REP^} where
6222 file is the name of the corresponding source file.
6225 @item /REPRESENTATION_INFO
6226 @cindex @option{/REPRESENTATION_INFO} (@code{gcc})
6227 This qualifier controls output from the compiler of a listing showing
6228 representation information for declared types and objects. For
6229 @option{/REPRESENTATION_INFO=NONE}, no information is output
6230 (equivalent to omitting the @option{/REPRESENTATION_INFO} qualifier).
6231 @option{/REPRESENTATION_INFO} without option is equivalent to
6232 @option{/REPRESENTATION_INFO=ARRAYS}.
6233 For @option{/REPRESENTATION_INFO=ARRAYS}, size and alignment
6234 information is listed for declared array and record types. For
6235 @option{/REPRESENTATION_INFO=OBJECTS}, size and alignment information
6236 is listed for all expression information for values that are computed
6237 at run time for variant records. These symbolic expressions have a mostly
6238 obvious format with #n being used to represent the value of the n'th
6239 discriminant. See source files @file{REPINFO.ADS/ADB} in the
6240 @code{GNAT} sources for full details on the format of
6241 @option{/REPRESENTATION_INFO=SYMBOLIC} output.
6242 If _FILE is added at the end of an option
6243 (e.g. @option{/REPRESENTATION_INFO=ARRAYS_FILE}),
6244 then the output is to a file with the name @file{file_REP} where
6245 file is the name of the corresponding source file.
6249 @cindex @option{-gnatS} (@code{gcc})
6250 The use of the switch @option{-gnatS} for an
6251 Ada compilation will cause the compiler to output a
6252 representation of package Standard in a form very
6253 close to standard Ada. It is not quite possible to
6254 do this and remain entirely Standard (since new
6255 numeric base types cannot be created in standard
6256 Ada), but the output is easily
6257 readable to any Ada programmer, and is useful to
6258 determine the characteristics of target dependent
6259 types in package Standard.
6262 @cindex @option{-gnatx} (@code{gcc})
6263 Normally the compiler generates full cross-referencing information in
6264 the @file{ALI} file. This information is used by a number of tools,
6265 including @code{gnatfind} and @code{gnatxref}. The @option{-gnatx} switch
6266 suppresses this information. This saves some space and may slightly
6267 speed up compilation, but means that these tools cannot be used.
6270 @node Exception Handling Control
6271 @subsection Exception Handling Control
6274 GNAT uses two methods for handling exceptions at run-time. The
6275 @code{longjmp/setjmp} method saves the context when entering
6276 a frame with an exception handler. Then when an exception is
6277 raised, the context can be restored immediately, without the
6278 need for tracing stack frames. This method provides very fast
6279 exception propagation, but introduces significant overhead for
6280 the use of exception handlers, even if no exception is raised.
6282 The other approach is called ``zero cost'' exception handling.
6283 With this method, the compiler builds static tables to describe
6284 the exception ranges. No dynamic code is required when entering
6285 a frame containing an exception handler. When an exception is
6286 raised, the tables are used to control a back trace of the
6287 subprogram invocation stack to locate the required exception
6288 handler. This method has considerably poorer performance for
6289 the propagation of exceptions, but there is no overhead for
6290 exception handlers if no exception is raised.
6292 The following switches can be used to control which of the
6293 two exception handling methods is used.
6299 @cindex @option{-gnatL} (@code{gcc})
6300 This switch causes the longjmp/setjmp approach to be used
6301 for exception handling. If this is the default mechanism for the
6302 target (see below), then this has no effect. If the default
6303 mechanism for the target is zero cost exceptions, then
6304 this switch can be used to modify this default, but it must be
6305 used for all units in the partition, including all run-time
6306 library units. One way to achieve this is to use the
6307 @option{-a} and @option{-f} switches for @code{gnatmake}.
6308 This option is rarely used. One case in which it may be
6309 advantageous is if you have an application where exception
6310 raising is common and the overall performance of the
6311 application is improved by favoring exception propagation.
6314 @cindex @option{-gnatZ} (@code{gcc})
6315 @cindex Zero Cost Exceptions
6316 This switch causes the zero cost approach to be sed
6317 for exception handling. If this is the default mechanism for the
6318 target (see below), then this has no effect. If the default
6319 mechanism for the target is longjmp/setjmp exceptions, then
6320 this switch can be used to modify this default, but it must be
6321 used for all units in the partition, including all run-time
6322 library units. One way to achieve this is to use the
6323 @option{-a} and @option{-f} switches for @code{gnatmake}.
6324 This option can only be used if the zero cost approach
6325 is available for the target in use (see below).
6329 The @code{longjmp/setjmp} approach is available on all targets, but
6330 the @code{zero cost} approach is only available on selected targets.
6331 To determine whether zero cost exceptions can be used for a
6332 particular target, look at the private part of the file system.ads.
6333 Either @code{GCC_ZCX_Support} or @code{Front_End_ZCX_Support} must
6334 be True to use the zero cost approach. If both of these switches
6335 are set to False, this means that zero cost exception handling
6336 is not yet available for that target. The switch
6337 @code{ZCX_By_Default} indicates the default approach. If this
6338 switch is set to True, then the @code{zero cost} approach is
6341 @node Units to Sources Mapping Files
6342 @subsection Units to Sources Mapping Files
6346 @item -gnatem^^=^@var{path}
6347 @cindex @option{-gnatem} (@code{gcc})
6348 A mapping file is a way to communicate to the compiler two mappings:
6349 from unit names to file names (without any directory information) and from
6350 file names to path names (with full directory information). These mappings
6351 are used by the compiler to short-circuit the path search.
6353 The use of mapping files is not required for correct operation of the
6354 compiler, but mapping files can improve efficiency, particularly when
6355 sources are read over a slow network connection. In normal operation,
6356 you need not be concerned with the format or use of mapping files,
6357 and the @option{-gnatem} switch is not a switch that you would use
6358 explicitly. it is intended only for use by automatic tools such as
6359 @code{gnatmake} running under the project file facility. The
6360 description here of the format of mapping files is provided
6361 for completeness and for possible use by other tools.
6363 A mapping file is a sequence of sets of three lines. In each set,
6364 the first line is the unit name, in lower case, with ``@code{%s}''
6366 specifications and ``@code{%b}'' appended for bodies; the second line is the
6367 file name; and the third line is the path name.
6373 /gnat/project1/sources/main.2.ada
6376 When the switch @option{-gnatem} is specified, the compiler will create
6377 in memory the two mappings from the specified file. If there is any problem
6378 (non existent file, truncated file or duplicate entries), no mapping
6381 Several @option{-gnatem} switches may be specified; however, only the last
6382 one on the command line will be taken into account.
6384 When using a project file, @code{gnatmake} create a temporary mapping file
6385 and communicates it to the compiler using this switch.
6390 @node Integrated Preprocessing
6391 @subsection Integrated Preprocessing
6394 GNAT sources may be preprocessed immediately before compilation; the actual
6395 text of the source is not the text of the source file, but is derived from it
6396 through a process called preprocessing. Integrated preprocessing is specified
6397 through switches @option{-gnatep} and/or @option{-gnateD}. @option{-gnatep}
6398 indicates, through a text file, the preprocessing data to be used.
6399 @option{-gnateD} specifies or modifies the values of preprocessing symbol.
6402 It is recommended that @code{gnatmake} switch ^-s^/SWITCH_CHECK^ should be
6403 used when Integrated Preprocessing is used. The reason is that preprocessing
6404 with another Preprocessing Data file without changing the sources will
6405 not trigger recompilation without this switch.
6408 Note that @code{gnatmake} switch ^-m^/MINIMAL_RECOMPILATION^ will almost
6409 always trigger recompilation for sources that are preprocessed,
6410 because @code{gnatmake} cannot compute the checksum of the source after
6414 The actual preprocessing function is described in details in section
6415 @ref{Preprocessing Using gnatprep}. This section only describes how integrated
6416 preprocessing is triggered and parameterized.
6420 @item -gnatep=@var{file}
6421 @cindex @option{-gnatep} (@code{gcc})
6422 This switch indicates to the compiler the file name (without directory
6423 information) of the preprocessor data file to use. The preprocessor data file
6424 should be found in the source directories.
6427 A preprocessing data file is a text file with significant lines indicating
6428 how should be preprocessed either a specific source or all sources not
6429 mentioned in other lines. A significant line is a non empty, non comment line.
6430 Comments are similar to Ada comments.
6433 Each significant line starts with either a literal string or the character '*'.
6434 A literal string is the file name (without directory information) of the source
6435 to preprocess. A character '*' indicates the preprocessing for all the sources
6436 that are not specified explicitly on other lines (order of the lines is not
6437 significant). It is an error to have two lines with the same file name or two
6438 lines starting with the character '*'.
6441 After the file name or the character '*', another optional literal string
6442 indicating the file name of the definition file to be used for preprocessing.
6443 (see @ref{Form of Definitions File}. The definition files are found by the
6444 compiler in one of the source directories. In some cases, when compiling
6445 a source in a directory other than the current directory, if the definition
6446 file is in the current directory, it may be necessary to add the current
6447 directory as a source directory through switch ^-I.^/SEARCH=[]^, otherwise
6448 the compiler would not find the definition file.
6451 Then, optionally, ^switches^switches^ similar to those of @code{gnatprep} may
6452 be found. Those ^switches^switches^ are:
6457 Causes both preprocessor lines and the lines deleted by
6458 preprocessing to be replaced by blank lines, preserving the line number.
6459 This ^switch^switch^ is always implied; however, if specified after @option{-c}
6460 it cancels the effect of @option{-c}.
6463 Causes both preprocessor lines and the lines deleted
6464 by preprocessing to be retained as comments marked
6465 with the special string ``@code{--! }''.
6467 @item -Dsymbol=value
6468 Define or redefine a symbol, associated with value. A symbol is an Ada
6469 identifier, or an Ada reserved word, with the exception of @code{if},
6470 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
6471 @code{value} is either a literal string, an Ada identifier or any Ada reserved
6472 word. A symbol declared with this ^switch^switch^ replaces a symbol with the
6473 same name defined in a definition file.
6476 Causes a sorted list of symbol names and values to be
6477 listed on the standard output file.
6480 Causes undefined symbols to be treated as having the value @code{FALSE}
6482 of a preprocessor test. In the absence of this option, an undefined symbol in
6483 a @code{#if} or @code{#elsif} test will be treated as an error.
6488 Examples of valid lines in a preprocessor data file:
6491 "toto.adb" "prep.def" -u
6492 -- preprocess "toto.adb", using definition file "prep.def",
6493 -- undefined symbol are False.
6496 -- preprocess all other sources without a definition file;
6497 -- suppressed lined are commented; symbol VERSION has the value V101.
6499 "titi.adb" "prep2.def" -s
6500 -- preprocess "titi.adb", using definition file "prep2.def";
6501 -- list all symbols with their values.
6504 @item ^-gnateD^/DATA_PREPROCESSING=^symbol[=value]
6505 @cindex @option{-gnateD} (@code{gcc})
6506 Define or redefine a preprocessing symbol, associated with value. If no value
6507 is given on the command line, then the value of the symbol is @code{True}.
6508 A symbol is an identifier, following normal Ada (case-insensitive)
6509 rules for its syntax, and value is any sequence (including an empty sequence)
6510 of characters from the set (letters, digits, period, underline).
6511 Ada reserved words may be used as symbols, with the exceptions of @code{if},
6512 @code{else}, @code{elsif}, @code{end}, @code{and}, @code{or} and @code{then}.
6515 A symbol declared with this ^switch^switch^ on the command line replaces a
6516 symbol with the same name either in a definition file or specified with a
6517 ^switch^switch^ -D in the preprocessor data file.
6520 This switch is similar to switch @option{^-D^/ASSOCIATE^} of @code{gnatprep}.
6526 @subsection Return Codes
6527 @cindex Return Codes
6528 @cindex @option{/RETURN_CODES=VMS}
6531 On VMS, GNAT compiled programs return POSIX-style codes by default,
6532 e.g. @option{/RETURN_CODES=POSIX}.
6534 To enable VMS style return codes, GNAT LINK with the option
6535 @option{/RETURN_CODES=VMS}. For example:
6538 GNAT LINK MYMAIN.ALI /RETURN_CODES=VMS
6542 Programs built with /RETURN_CODES=VMS are suitable to be called in
6543 VMS DCL scripts. Programs compiled with the default /RETURN_CODES=POSIX
6544 are suitable for spawning with appropriate GNAT RTL routines.
6549 @node Search Paths and the Run-Time Library (RTL)
6550 @section Search Paths and the Run-Time Library (RTL)
6553 With the GNAT source-based library system, the compiler must be able to
6554 find source files for units that are needed by the unit being compiled.
6555 Search paths are used to guide this process.
6557 The compiler compiles one source file whose name must be given
6558 explicitly on the command line. In other words, no searching is done
6559 for this file. To find all other source files that are needed (the most
6560 common being the specs of units), the compiler examines the following
6561 directories, in the following order:
6565 The directory containing the source file of the main unit being compiled
6566 (the file name on the command line).
6569 Each directory named by an @option{^-I^/SOURCE_SEARCH^} switch given on the
6570 @code{gcc} command line, in the order given.
6573 @findex ADA_INCLUDE_PATH
6574 Each of the directories listed in the value of the
6575 @code{ADA_INCLUDE_PATH} ^environment variable^logical name^.
6577 Construct this value
6578 exactly as the @code{PATH} environment variable: a list of directory
6579 names separated by colons (semicolons when working with the NT version).
6582 Normally, define this value as a logical name containing a comma separated
6583 list of directory names.
6585 This variable can also be defined by means of an environment string
6586 (an argument to the DEC C exec* set of functions).
6590 DEFINE ANOTHER_PATH FOO:[BAG]
6591 DEFINE ADA_INCLUDE_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
6594 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
6595 first, followed by the standard Ada 95
6596 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADAINCLUDE].
6597 If this is not redefined, the user will obtain the DEC Ada 83 IO packages
6598 (Text_IO, Sequential_IO, etc)
6599 instead of the Ada95 packages. Thus, in order to get the Ada 95
6600 packages by default, ADA_INCLUDE_PATH must be redefined.
6604 @findex ADA_PRJ_INCLUDE_FILE
6605 Each of the directories listed in the text file whose name is given
6606 by the @code{ADA_PRJ_INCLUDE_FILE} ^environment variable^logical name^.
6609 @code{ADA_PRJ_INCLUDE_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
6610 driver when project files are used. It should not normally be set
6614 The content of the @file{ada_source_path} file which is part of the GNAT
6615 installation tree and is used to store standard libraries such as the
6616 GNAT Run Time Library (RTL) source files.
6618 @ref{Installing an Ada Library}
6623 Specifying the switch @option{^-I-^/NOCURRENT_DIRECTORY^}
6624 inhibits the use of the directory
6625 containing the source file named in the command line. You can still
6626 have this directory on your search path, but in this case it must be
6627 explicitly requested with a @option{^-I^/SOURCE_SEARCH^} switch.
6629 Specifying the switch @option{-nostdinc}
6630 inhibits the search of the default location for the GNAT Run Time
6631 Library (RTL) source files.
6633 The compiler outputs its object files and ALI files in the current
6636 Caution: The object file can be redirected with the @option{-o} switch;
6637 however, @code{gcc} and @code{gnat1} have not been coordinated on this
6638 so the @file{ALI} file will not go to the right place. Therefore, you should
6639 avoid using the @option{-o} switch.
6643 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
6644 children make up the GNAT RTL, together with the simple @code{System.IO}
6645 package used in the @code{"Hello World"} example. The sources for these units
6646 are needed by the compiler and are kept together in one directory. Not
6647 all of the bodies are needed, but all of the sources are kept together
6648 anyway. In a normal installation, you need not specify these directory
6649 names when compiling or binding. Either the environment variables or
6650 the built-in defaults cause these files to be found.
6652 In addition to the language-defined hierarchies (@code{System}, @code{Ada} and
6653 @code{Interfaces}), the GNAT distribution provides a fourth hierarchy,
6654 consisting of child units of @code{GNAT}. This is a collection of generally
6655 useful types, subprograms, etc. See the @cite{GNAT Reference Manual} for
6658 Besides simplifying access to the RTL, a major use of search paths is
6659 in compiling sources from multiple directories. This can make
6660 development environments much more flexible.
6663 @node Order of Compilation Issues
6664 @section Order of Compilation Issues
6667 If, in our earlier example, there was a spec for the @code{hello}
6668 procedure, it would be contained in the file @file{hello.ads}; yet this
6669 file would not have to be explicitly compiled. This is the result of the
6670 model we chose to implement library management. Some of the consequences
6671 of this model are as follows:
6675 There is no point in compiling specs (except for package
6676 specs with no bodies) because these are compiled as needed by clients. If
6677 you attempt a useless compilation, you will receive an error message.
6678 It is also useless to compile subunits because they are compiled as needed
6682 There are no order of compilation requirements: performing a
6683 compilation never obsoletes anything. The only way you can obsolete
6684 something and require recompilations is to modify one of the
6685 source files on which it depends.
6688 There is no library as such, apart from the ALI files
6689 (@pxref{The Ada Library Information Files}, for information on the format
6690 of these files). For now we find it convenient to create separate ALI files,
6691 but eventually the information therein may be incorporated into the object
6695 When you compile a unit, the source files for the specs of all units
6696 that it @code{with}'s, all its subunits, and the bodies of any generics it
6697 instantiates must be available (reachable by the search-paths mechanism
6698 described above), or you will receive a fatal error message.
6705 The following are some typical Ada compilation command line examples:
6708 @item $ gcc -c xyz.adb
6709 Compile body in file @file{xyz.adb} with all default options.
6712 @item $ gcc -c -O2 -gnata xyz-def.adb
6715 @item $ GNAT COMPILE /OPTIMIZE=ALL -gnata xyz-def.adb
6718 Compile the child unit package in file @file{xyz-def.adb} with extensive
6719 optimizations, and pragma @code{Assert}/@code{Debug} statements
6722 @item $ gcc -c -gnatc abc-def.adb
6723 Compile the subunit in file @file{abc-def.adb} in semantic-checking-only
6727 @node Binding Using gnatbind
6728 @chapter Binding Using @code{gnatbind}
6732 * Running gnatbind::
6733 * Switches for gnatbind::
6734 * Command-Line Access::
6735 * Search Paths for gnatbind::
6736 * Examples of gnatbind Usage::
6740 This chapter describes the GNAT binder, @code{gnatbind}, which is used
6741 to bind compiled GNAT objects. The @code{gnatbind} program performs
6742 four separate functions:
6746 Checks that a program is consistent, in accordance with the rules in
6747 Chapter 10 of the Ada 95 Reference Manual. In particular, error
6748 messages are generated if a program uses inconsistent versions of a
6752 Checks that an acceptable order of elaboration exists for the program
6753 and issues an error message if it cannot find an order of elaboration
6754 that satisfies the rules in Chapter 10 of the Ada 95 Language Manual.
6757 Generates a main program incorporating the given elaboration order.
6758 This program is a small Ada package (body and spec) that
6759 must be subsequently compiled
6760 using the GNAT compiler. The necessary compilation step is usually
6761 performed automatically by @code{gnatlink}. The two most important
6762 functions of this program
6763 are to call the elaboration routines of units in an appropriate order
6764 and to call the main program.
6767 Determines the set of object files required by the given main program.
6768 This information is output in the forms of comments in the generated program,
6769 to be read by the @code{gnatlink} utility used to link the Ada application.
6773 @node Running gnatbind
6774 @section Running @code{gnatbind}
6777 The form of the @code{gnatbind} command is
6780 $ gnatbind [@i{switches}] @i{mainprog}[.ali] [@i{switches}]
6784 where @file{@i{mainprog}.adb} is the Ada file containing the main program
6785 unit body. If no switches are specified, @code{gnatbind} constructs an Ada
6786 package in two files whose names are
6787 @file{b~@i{mainprog}.ads}, and @file{b~@i{mainprog}.adb}.
6788 For example, if given the
6789 parameter @file{hello.ali}, for a main program contained in file
6790 @file{hello.adb}, the binder output files would be @file{b~hello.ads}
6791 and @file{b~hello.adb}.
6793 When doing consistency checking, the binder takes into consideration
6794 any source files it can locate. For example, if the binder determines
6795 that the given main program requires the package @code{Pack}, whose
6797 file is @file{pack.ali} and whose corresponding source spec file is
6798 @file{pack.ads}, it attempts to locate the source file @file{pack.ads}
6799 (using the same search path conventions as previously described for the
6800 @code{gcc} command). If it can locate this source file, it checks that
6802 or source checksums of the source and its references to in @file{ALI} files
6803 match. In other words, any @file{ALI} files that mentions this spec must have
6804 resulted from compiling this version of the source file (or in the case
6805 where the source checksums match, a version close enough that the
6806 difference does not matter).
6808 @cindex Source files, use by binder
6809 The effect of this consistency checking, which includes source files, is
6810 that the binder ensures that the program is consistent with the latest
6811 version of the source files that can be located at bind time. Editing a
6812 source file without compiling files that depend on the source file cause
6813 error messages to be generated by the binder.
6815 For example, suppose you have a main program @file{hello.adb} and a
6816 package @code{P}, from file @file{p.ads} and you perform the following
6821 Enter @code{gcc -c hello.adb} to compile the main program.
6824 Enter @code{gcc -c p.ads} to compile package @code{P}.
6827 Edit file @file{p.ads}.
6830 Enter @code{gnatbind hello}.
6834 At this point, the file @file{p.ali} contains an out-of-date time stamp
6835 because the file @file{p.ads} has been edited. The attempt at binding
6836 fails, and the binder generates the following error messages:
6839 error: "hello.adb" must be recompiled ("p.ads" has been modified)
6840 error: "p.ads" has been modified and must be recompiled
6844 Now both files must be recompiled as indicated, and then the bind can
6845 succeed, generating a main program. You need not normally be concerned
6846 with the contents of this file, but for reference purposes a sample
6847 binder output file is given in @ref{Example of Binder Output File}.
6849 In most normal usage, the default mode of @command{gnatbind} which is to
6850 generate the main package in Ada, as described in the previous section.
6851 In particular, this means that any Ada programmer can read and understand
6852 the generated main program. It can also be debugged just like any other
6853 Ada code provided the @option{^-g^/DEBUG^} switch is used for
6854 @command{gnatbind} and @command{gnatlink}.
6856 However for some purposes it may be convenient to generate the main
6857 program in C rather than Ada. This may for example be helpful when you
6858 are generating a mixed language program with the main program in C. The
6859 GNAT compiler itself is an example.
6860 The use of the @option{^-C^/BIND_FILE=C^} switch
6861 for both @code{gnatbind} and @code{gnatlink} will cause the program to
6862 be generated in C (and compiled using the gnu C compiler).
6865 @node Switches for gnatbind
6866 @section Switches for @command{gnatbind}
6869 The following switches are available with @code{gnatbind}; details will
6870 be presented in subsequent sections.
6873 * Consistency-Checking Modes::
6874 * Binder Error Message Control::
6875 * Elaboration Control::
6877 * Binding with Non-Ada Main Programs::
6878 * Binding Programs with No Main Subprogram::
6883 @item ^-aO^/OBJECT_SEARCH^
6884 @cindex @option{^-aO^/OBJECT_SEARCH^} (@command{gnatbind})
6885 Specify directory to be searched for ALI files.
6887 @item ^-aI^/SOURCE_SEARCH^
6888 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
6889 Specify directory to be searched for source file.
6891 @item ^-A^/BIND_FILE=ADA^
6892 @cindex @option{^-A^/BIND_FILE=ADA^} (@command{gnatbind})
6893 Generate binder program in Ada (default)
6895 @item ^-b^/REPORT_ERRORS=BRIEF^
6896 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@command{gnatbind})
6897 Generate brief messages to @file{stderr} even if verbose mode set.
6899 @item ^-c^/NOOUTPUT^
6900 @cindex @option{^-c^/NOOUTPUT^} (@command{gnatbind})
6901 Check only, no generation of binder output file.
6903 @item ^-C^/BIND_FILE=C^
6904 @cindex @option{^-C^/BIND_FILE=C^} (@command{gnatbind})
6905 Generate binder program in C
6907 @item ^-e^/ELABORATION_DEPENDENCIES^
6908 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@command{gnatbind})
6909 Output complete list of elaboration-order dependencies.
6911 @item ^-E^/STORE_TRACEBACKS^
6912 @cindex @option{^-E^/STORE_TRACEBACKS^} (@command{gnatbind})
6913 Store tracebacks in exception occurrences when the target supports it.
6914 This is the default with the zero cost exception mechanism.
6916 @c The following may get moved to an appendix
6917 This option is currently supported on the following targets:
6918 all x86 ports, Solaris, Windows, HP-UX, AIX, PowerPC VxWorks and Alpha VxWorks.
6920 See also the packages @code{GNAT.Traceback} and
6921 @code{GNAT.Traceback.Symbolic} for more information.
6923 Note that on x86 ports, you must not use @option{-fomit-frame-pointer}
6927 @item ^-F^/FORCE_ELABS_FLAGS^
6928 @cindex @option{^-F^/FORCE_ELABS_FLAGS^} (@command{gnatbind})
6929 Force the checks of elaboration flags. @command{gnatbind} does not normally
6930 generate checks of elaboration flags for the main executable, except when
6931 a Stand-Alone Library is used. However, there are cases when this cannot be
6932 detected by gnatbind. An example is importing an interface of a Stand-Alone
6933 Library through a pragma Import and only specifying through a linker switch
6934 this Stand-Alone Library. This switch is used to guarantee that elaboration
6935 flag checks are generated.
6938 @cindex @option{^-h^/HELP^} (@command{gnatbind})
6939 Output usage (help) information
6942 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
6943 Specify directory to be searched for source and ALI files.
6945 @item ^-I-^/NOCURRENT_DIRECTORY^
6946 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@command{gnatbind})
6947 Do not look for sources in the current directory where @code{gnatbind} was
6948 invoked, and do not look for ALI files in the directory containing the
6949 ALI file named in the @code{gnatbind} command line.
6951 @item ^-l^/ORDER_OF_ELABORATION^
6952 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@command{gnatbind})
6953 Output chosen elaboration order.
6955 @item ^-Lxxx^/BUILD_LIBRARY=xxx^
6956 @cindex @option{^-L^/BUILD_LIBRARY^} (@command{gnatbind})
6957 Binds the units for library building. In this case the adainit and
6958 adafinal procedures (See @pxref{Binding with Non-Ada Main Programs})
6959 are renamed to ^xxxinit^XXXINIT^ and
6960 ^xxxfinal^XXXFINAL^.
6961 Implies ^-n^/NOCOMPILE^.
6963 (@pxref{GNAT and Libraries}, for more details.)
6966 On OpenVMS, these init and final procedures are exported in uppercase
6967 letters. For example if /BUILD_LIBRARY=toto is used, the exported name of
6968 the init procedure will be "TOTOINIT" and the exported name of the final
6969 procedure will be "TOTOFINAL".
6972 @item ^-Mxyz^/RENAME_MAIN=xyz^
6973 @cindex @option{^-M^/RENAME_MAIN^} (@command{gnatbind})
6974 Rename generated main program from main to xyz
6976 @item ^-m^/ERROR_LIMIT=^@var{n}
6977 @cindex @option{^-m^/ERROR_LIMIT^} (@command{gnatbind})
6978 Limit number of detected errors to @var{n}, where @var{n} is
6979 in the range 1..999_999. The default value if no switch is
6980 given is 9999. Binding is terminated if the limit is exceeded.
6982 Furthermore, under Windows, the sources pointed to by the libraries path
6983 set in the registry are not searched for.
6987 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
6991 @cindex @option{-nostdinc} (@command{gnatbind})
6992 Do not look for sources in the system default directory.
6995 @cindex @option{-nostdlib} (@command{gnatbind})
6996 Do not look for library files in the system default directory.
6998 @item --RTS=@var{rts-path}
6999 @cindex @option{--RTS} (@code{gnatbind})
7000 Specifies the default location of the runtime library. Same meaning as the
7001 equivalent @code{gnatmake} flag (see @ref{Switches for gnatmake}).
7003 @item ^-o ^/OUTPUT=^@var{file}
7004 @cindex @option{^-o ^/OUTPUT^} (@command{gnatbind})
7005 Name the output file @var{file} (default is @file{b~@var{xxx}.adb}).
7006 Note that if this option is used, then linking must be done manually,
7007 gnatlink cannot be used.
7009 @item ^-O^/OBJECT_LIST^
7010 @cindex @option{^-O^/OBJECT_LIST^} (@command{gnatbind})
7013 @item ^-p^/PESSIMISTIC_ELABORATION^
7014 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@command{gnatbind})
7015 Pessimistic (worst-case) elaboration order
7017 @item ^-s^/READ_SOURCES=ALL^
7018 @cindex @option{^-s^/READ_SOURCES=ALL^} (@command{gnatbind})
7019 Require all source files to be present.
7021 @item ^-S@var{xxx}^/INITIALIZE_SCALARS=@var{xxx}^
7022 @cindex @option{^-S^/INITIALIZE_SCALARS^} (@command{gnatbind})
7023 Specifies the value to be used when detecting uninitialized scalar
7024 objects with pragma Initialize_Scalars.
7025 The @var{xxx} ^string specified with the switch^option^ may be either
7027 @item ``@option{^in^INVALID^}'' requesting an invalid value where possible
7028 @item ``@option{^lo^LOW^}'' for the lowest possible value
7029 possible, and the low
7030 @item ``@option{^hi^HIGH^}'' for the highest possible value
7031 @item ``@option{xx}'' for a value consisting of repeated bytes with the
7032 value 16#xx# (i.e. xx is a string of two hexadecimal digits).
7035 In addition, you can specify @option{-Sev} to indicate that the value is
7036 to be set at run time. In this case, the program will look for an environment
7037 @cindex GNAT_INIT_SCALARS
7038 variable of the form @code{GNAT_INIT_SCALARS=xx}, where xx is one
7039 of @option{in/lo/hi/xx} with the same meanings as above.
7040 If no environment variable is found, or if it does not have a valid value,
7041 then the default is @option{in} (invalid values).
7045 @cindex @option{-static} (@code{gnatbind})
7046 Link against a static GNAT run time.
7049 @cindex @option{-shared} (@code{gnatbind})
7050 Link against a shared GNAT run time when available.
7053 @item ^-t^/NOTIME_STAMP_CHECK^
7054 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
7055 Tolerate time stamp and other consistency errors
7057 @item ^-T@var{n}^/TIME_SLICE=@var{n}^
7058 @cindex @option{^-T^/TIME_SLICE^} (@code{gnatbind})
7059 Set the time slice value to @var{n} milliseconds. If the system supports
7060 the specification of a specific time slice value, then the indicated value
7061 is used. If the system does not support specific time slice values, but
7062 does support some general notion of round-robin scheduling, then any
7063 non-zero value will activate round-robin scheduling.
7065 A value of zero is treated specially. It turns off time
7066 slicing, and in addition, indicates to the tasking run time that the
7067 semantics should match as closely as possible the Annex D
7068 requirements of the Ada RM, and in particular sets the default
7069 scheduling policy to @code{FIFO_Within_Priorities}.
7071 @item ^-v^/REPORT_ERRORS=VERBOSE^
7072 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
7073 Verbose mode. Write error messages, header, summary output to
7078 @cindex @option{-w} (@code{gnatbind})
7079 Warning mode (@var{x}=s/e for suppress/treat as error)
7083 @item /WARNINGS=NORMAL
7084 @cindex @option{/WARNINGS} (@code{gnatbind})
7085 Normal warnings mode. Warnings are issued but ignored
7087 @item /WARNINGS=SUPPRESS
7088 @cindex @option{/WARNINGS} (@code{gnatbind})
7089 All warning messages are suppressed
7091 @item /WARNINGS=ERROR
7092 @cindex @option{/WARNINGS} (@code{gnatbind})
7093 Warning messages are treated as fatal errors
7096 @item ^-x^/READ_SOURCES=NONE^
7097 @cindex @option{^-x^/READ_SOURCES^} (@code{gnatbind})
7098 Exclude source files (check object consistency only).
7101 @item /READ_SOURCES=AVAILABLE
7102 @cindex @option{/READ_SOURCES} (@code{gnatbind})
7103 Default mode, in which sources are checked for consistency only if
7107 @item ^-z^/ZERO_MAIN^
7108 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
7114 You may obtain this listing of switches by running @code{gnatbind} with
7119 @node Consistency-Checking Modes
7120 @subsection Consistency-Checking Modes
7123 As described earlier, by default @code{gnatbind} checks
7124 that object files are consistent with one another and are consistent
7125 with any source files it can locate. The following switches control binder
7130 @item ^-s^/READ_SOURCES=ALL^
7131 @cindex @option{^-s^/READ_SOURCES=ALL^} (@code{gnatbind})
7132 Require source files to be present. In this mode, the binder must be
7133 able to locate all source files that are referenced, in order to check
7134 their consistency. In normal mode, if a source file cannot be located it
7135 is simply ignored. If you specify this switch, a missing source
7138 @item ^-x^/READ_SOURCES=NONE^
7139 @cindex @option{^-x^/READ_SOURCES=NONE^} (@code{gnatbind})
7140 Exclude source files. In this mode, the binder only checks that ALI
7141 files are consistent with one another. Source files are not accessed.
7142 The binder runs faster in this mode, and there is still a guarantee that
7143 the resulting program is self-consistent.
7144 If a source file has been edited since it was last compiled, and you
7145 specify this switch, the binder will not detect that the object
7146 file is out of date with respect to the source file. Note that this is the
7147 mode that is automatically used by @code{gnatmake} because in this
7148 case the checking against sources has already been performed by
7149 @code{gnatmake} in the course of compilation (i.e. before binding).
7152 @item /READ_SOURCES=AVAILABLE
7153 @cindex @code{/READ_SOURCES=AVAILABLE} (@code{gnatbind})
7154 This is the default mode in which source files are checked if they are
7155 available, and ignored if they are not available.
7159 @node Binder Error Message Control
7160 @subsection Binder Error Message Control
7163 The following switches provide control over the generation of error
7164 messages from the binder:
7168 @item ^-v^/REPORT_ERRORS=VERBOSE^
7169 @cindex @option{^-v^/REPORT_ERRORS=VERBOSE^} (@code{gnatbind})
7170 Verbose mode. In the normal mode, brief error messages are generated to
7171 @file{stderr}. If this switch is present, a header is written
7172 to @file{stdout} and any error messages are directed to @file{stdout}.
7173 All that is written to @file{stderr} is a brief summary message.
7175 @item ^-b^/REPORT_ERRORS=BRIEF^
7176 @cindex @option{^-b^/REPORT_ERRORS=BRIEF^} (@code{gnatbind})
7177 Generate brief error messages to @file{stderr} even if verbose mode is
7178 specified. This is relevant only when used with the
7179 @option{^-v^/REPORT_ERRORS=VERBOSE^} switch.
7183 @cindex @option{-m} (@code{gnatbind})
7184 Limits the number of error messages to @var{n}, a decimal integer in the
7185 range 1-999. The binder terminates immediately if this limit is reached.
7188 @cindex @option{-M} (@code{gnatbind})
7189 Renames the generated main program from @code{main} to @code{xxx}.
7190 This is useful in the case of some cross-building environments, where
7191 the actual main program is separate from the one generated
7195 @item ^-ws^/WARNINGS=SUPPRESS^
7196 @cindex @option{^-ws^/WARNINGS=SUPPRESS^} (@code{gnatbind})
7198 Suppress all warning messages.
7200 @item ^-we^/WARNINGS=ERROR^
7201 @cindex @option{^-we^/WARNINGS=ERROR^} (@code{gnatbind})
7202 Treat any warning messages as fatal errors.
7205 @item /WARNINGS=NORMAL
7206 Standard mode with warnings generated, but warnings do not get treated
7210 @item ^-t^/NOTIME_STAMP_CHECK^
7211 @cindex @option{^-t^/NOTIME_STAMP_CHECK^} (@code{gnatbind})
7212 @cindex Time stamp checks, in binder
7213 @cindex Binder consistency checks
7214 @cindex Consistency checks, in binder
7215 The binder performs a number of consistency checks including:
7219 Check that time stamps of a given source unit are consistent
7221 Check that checksums of a given source unit are consistent
7223 Check that consistent versions of @code{GNAT} were used for compilation
7225 Check consistency of configuration pragmas as required
7229 Normally failure of such checks, in accordance with the consistency
7230 requirements of the Ada Reference Manual, causes error messages to be
7231 generated which abort the binder and prevent the output of a binder
7232 file and subsequent link to obtain an executable.
7234 The @option{^-t^/NOTIME_STAMP_CHECK^} switch converts these error messages
7235 into warnings, so that
7236 binding and linking can continue to completion even in the presence of such
7237 errors. The result may be a failed link (due to missing symbols), or a
7238 non-functional executable which has undefined semantics.
7239 @emph{This means that
7240 @option{^-t^/NOTIME_STAMP_CHECK^} should be used only in unusual situations,
7244 @node Elaboration Control
7245 @subsection Elaboration Control
7248 The following switches provide additional control over the elaboration
7249 order. For full details see @xref{Elaboration Order Handling in GNAT}.
7252 @item ^-p^/PESSIMISTIC_ELABORATION^
7253 @cindex @option{^-p^/PESSIMISTIC_ELABORATION^} (@code{gnatbind})
7254 Normally the binder attempts to choose an elaboration order that is
7255 likely to minimize the likelihood of an elaboration order error resulting
7256 in raising a @code{Program_Error} exception. This switch reverses the
7257 action of the binder, and requests that it deliberately choose an order
7258 that is likely to maximize the likelihood of an elaboration error.
7259 This is useful in ensuring portability and avoiding dependence on
7260 accidental fortuitous elaboration ordering.
7262 Normally it only makes sense to use the @option{^-p^/PESSIMISTIC_ELABORATION^}
7264 elaboration checking is used (@option{-gnatE} switch used for compilation).
7265 This is because in the default static elaboration mode, all necessary
7266 @code{Elaborate_All} pragmas are implicitly inserted.
7267 These implicit pragmas are still respected by the binder in
7268 @option{^-p^/PESSIMISTIC_ELABORATION^} mode, so a
7269 safe elaboration order is assured.
7272 @node Output Control
7273 @subsection Output Control
7276 The following switches allow additional control over the output
7277 generated by the binder.
7282 @item ^-A^/BIND_FILE=ADA^
7283 @cindex @option{^-A^/BIND_FILE=ADA^} (@code{gnatbind})
7284 Generate binder program in Ada (default). The binder program is named
7285 @file{b~@var{mainprog}.adb} by default. This can be changed with
7286 @option{^-o^/OUTPUT^} @code{gnatbind} option.
7288 @item ^-c^/NOOUTPUT^
7289 @cindex @option{^-c^/NOOUTPUT^} (@code{gnatbind})
7290 Check only. Do not generate the binder output file. In this mode the
7291 binder performs all error checks but does not generate an output file.
7293 @item ^-C^/BIND_FILE=C^
7294 @cindex @option{^-C^/BIND_FILE=C^} (@code{gnatbind})
7295 Generate binder program in C. The binder program is named
7296 @file{b_@var{mainprog}.c}.
7297 This can be changed with @option{^-o^/OUTPUT^} @code{gnatbind}
7300 @item ^-e^/ELABORATION_DEPENDENCIES^
7301 @cindex @option{^-e^/ELABORATION_DEPENDENCIES^} (@code{gnatbind})
7302 Output complete list of elaboration-order dependencies, showing the
7303 reason for each dependency. This output can be rather extensive but may
7304 be useful in diagnosing problems with elaboration order. The output is
7305 written to @file{stdout}.
7308 @cindex @option{^-h^/HELP^} (@code{gnatbind})
7309 Output usage information. The output is written to @file{stdout}.
7311 @item ^-K^/LINKER_OPTION_LIST^
7312 @cindex @option{^-K^/LINKER_OPTION_LIST^} (@code{gnatbind})
7313 Output linker options to @file{stdout}. Includes library search paths,
7314 contents of pragmas Ident and Linker_Options, and libraries added
7317 @item ^-l^/ORDER_OF_ELABORATION^
7318 @cindex @option{^-l^/ORDER_OF_ELABORATION^} (@code{gnatbind})
7319 Output chosen elaboration order. The output is written to @file{stdout}.
7321 @item ^-O^/OBJECT_LIST^
7322 @cindex @option{^-O^/OBJECT_LIST^} (@code{gnatbind})
7323 Output full names of all the object files that must be linked to provide
7324 the Ada component of the program. The output is written to @file{stdout}.
7325 This list includes the files explicitly supplied and referenced by the user
7326 as well as implicitly referenced run-time unit files. The latter are
7327 omitted if the corresponding units reside in shared libraries. The
7328 directory names for the run-time units depend on the system configuration.
7330 @item ^-o ^/OUTPUT=^@var{file}
7331 @cindex @option{^-o^/OUTPUT^} (@code{gnatbind})
7332 Set name of output file to @var{file} instead of the normal
7333 @file{b~@var{mainprog}.adb} default. Note that @var{file} denote the Ada
7334 binder generated body filename. In C mode you would normally give
7335 @var{file} an extension of @file{.c} because it will be a C source program.
7336 Note that if this option is used, then linking must be done manually.
7337 It is not possible to use gnatlink in this case, since it cannot locate
7340 @item ^-r^/RESTRICTION_LIST^
7341 @cindex @option{^-r^/RESTRICTION_LIST^} (@code{gnatbind})
7342 Generate list of @code{pragma Restrictions} that could be applied to
7343 the current unit. This is useful for code audit purposes, and also may
7344 be used to improve code generation in some cases.
7348 @node Binding with Non-Ada Main Programs
7349 @subsection Binding with Non-Ada Main Programs
7352 In our description so far we have assumed that the main
7353 program is in Ada, and that the task of the binder is to generate a
7354 corresponding function @code{main} that invokes this Ada main
7355 program. GNAT also supports the building of executable programs where
7356 the main program is not in Ada, but some of the called routines are
7357 written in Ada and compiled using GNAT (@pxref{Mixed Language Programming}).
7358 The following switch is used in this situation:
7362 @cindex @option{^-n^/NOMAIN^} (@code{gnatbind})
7363 No main program. The main program is not in Ada.
7367 In this case, most of the functions of the binder are still required,
7368 but instead of generating a main program, the binder generates a file
7369 containing the following callable routines:
7374 You must call this routine to initialize the Ada part of the program by
7375 calling the necessary elaboration routines. A call to @code{adainit} is
7376 required before the first call to an Ada subprogram.
7378 Note that it is assumed that the basic execution environment must be setup
7379 to be appropriate for Ada execution at the point where the first Ada
7380 subprogram is called. In particular, if the Ada code will do any
7381 floating-point operations, then the FPU must be setup in an appropriate
7382 manner. For the case of the x86, for example, full precision mode is
7383 required. The procedure GNAT.Float_Control.Reset may be used to ensure
7384 that the FPU is in the right state.
7388 You must call this routine to perform any library-level finalization
7389 required by the Ada subprograms. A call to @code{adafinal} is required
7390 after the last call to an Ada subprogram, and before the program
7395 If the @option{^-n^/NOMAIN^} switch
7396 @cindex @option{^-n^/NOMAIN^} (@command{gnatbind})
7397 @cindex Binder, multiple input files
7398 is given, more than one ALI file may appear on
7399 the command line for @code{gnatbind}. The normal @dfn{closure}
7400 calculation is performed for each of the specified units. Calculating
7401 the closure means finding out the set of units involved by tracing
7402 @code{with} references. The reason it is necessary to be able to
7403 specify more than one ALI file is that a given program may invoke two or
7404 more quite separate groups of Ada units.
7406 The binder takes the name of its output file from the last specified ALI
7407 file, unless overridden by the use of the @option{^-o file^/OUTPUT=file^}.
7408 @cindex @option{^-o^/OUTPUT^} (@command{gnatbind})
7409 The output is an Ada unit in source form that can
7410 be compiled with GNAT unless the -C switch is used in which case the
7411 output is a C source file, which must be compiled using the C compiler.
7412 This compilation occurs automatically as part of the @code{gnatlink}
7415 Currently the GNAT run time requires a FPU using 80 bits mode
7416 precision. Under targets where this is not the default it is required to
7417 call GNAT.Float_Control.Reset before using floating point numbers (this
7418 include float computation, float input and output) in the Ada code. A
7419 side effect is that this could be the wrong mode for the foreign code
7420 where floating point computation could be broken after this call.
7422 @node Binding Programs with No Main Subprogram
7423 @subsection Binding Programs with No Main Subprogram
7426 It is possible to have an Ada program which does not have a main
7427 subprogram. This program will call the elaboration routines of all the
7428 packages, then the finalization routines.
7430 The following switch is used to bind programs organized in this manner:
7433 @item ^-z^/ZERO_MAIN^
7434 @cindex @option{^-z^/ZERO_MAIN^} (@code{gnatbind})
7435 Normally the binder checks that the unit name given on the command line
7436 corresponds to a suitable main subprogram. When this switch is used,
7437 a list of ALI files can be given, and the execution of the program
7438 consists of elaboration of these units in an appropriate order.
7442 @node Command-Line Access
7443 @section Command-Line Access
7446 The package @code{Ada.Command_Line} provides access to the command-line
7447 arguments and program name. In order for this interface to operate
7448 correctly, the two variables
7460 are declared in one of the GNAT library routines. These variables must
7461 be set from the actual @code{argc} and @code{argv} values passed to the
7462 main program. With no @option{^n^/NOMAIN^} present, @code{gnatbind}
7463 generates the C main program to automatically set these variables.
7464 If the @option{^n^/NOMAIN^} switch is used, there is no automatic way to
7465 set these variables. If they are not set, the procedures in
7466 @code{Ada.Command_Line} will not be available, and any attempt to use
7467 them will raise @code{Constraint_Error}. If command line access is
7468 required, your main program must set @code{gnat_argc} and
7469 @code{gnat_argv} from the @code{argc} and @code{argv} values passed to
7473 @node Search Paths for gnatbind
7474 @section Search Paths for @code{gnatbind}
7477 The binder takes the name of an ALI file as its argument and needs to
7478 locate source files as well as other ALI files to verify object consistency.
7480 For source files, it follows exactly the same search rules as @code{gcc}
7481 (@pxref{Search Paths and the Run-Time Library (RTL)}). For ALI files the
7482 directories searched are:
7486 The directory containing the ALI file named in the command line, unless
7487 the switch @option{^-I-^/NOCURRENT_DIRECTORY^} is specified.
7490 All directories specified by @option{^-I^/SEARCH^}
7491 switches on the @code{gnatbind}
7492 command line, in the order given.
7495 @findex ADA_OBJECTS_PATH
7496 Each of the directories listed in the value of the
7497 @code{ADA_OBJECTS_PATH} ^environment variable^logical name^.
7499 Construct this value
7500 exactly as the @code{PATH} environment variable: a list of directory
7501 names separated by colons (semicolons when working with the NT version
7505 Normally, define this value as a logical name containing a comma separated
7506 list of directory names.
7508 This variable can also be defined by means of an environment string
7509 (an argument to the DEC C exec* set of functions).
7513 DEFINE ANOTHER_PATH FOO:[BAG]
7514 DEFINE ADA_OBJECTS_PATH ANOTHER_PATH,FOO:[BAM],FOO:[BAR]
7517 By default, the path includes GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
7518 first, followed by the standard Ada 95
7519 libraries in GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB].
7520 If this is not redefined, the user will obtain the DEC Ada 83 IO packages
7521 (Text_IO, Sequential_IO, etc)
7522 instead of the Ada95 packages. Thus, in order to get the Ada 95
7523 packages by default, ADA_OBJECTS_PATH must be redefined.
7527 @findex ADA_PRJ_OBJECTS_FILE
7528 Each of the directories listed in the text file whose name is given
7529 by the @code{ADA_PRJ_OBJECTS_FILE} ^environment variable^logical name^.
7532 @code{ADA_PRJ_OBJECTS_FILE} is normally set by gnatmake or by the ^gnat^GNAT^
7533 driver when project files are used. It should not normally be set
7537 The content of the @file{ada_object_path} file which is part of the GNAT
7538 installation tree and is used to store standard libraries such as the
7539 GNAT Run Time Library (RTL) unless the switch @option{-nostdlib} is
7542 @ref{Installing an Ada Library}
7547 In the binder the switch @option{^-I^/SEARCH^}
7548 @cindex @option{^-I^/SEARCH^} (@command{gnatbind})
7549 is used to specify both source and
7550 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
7551 @cindex @option{^-aI^/SOURCE_SEARCH^} (@command{gnatbind})
7552 instead if you want to specify
7553 source paths only, and @option{^-aO^/LIBRARY_SEARCH^}
7554 @cindex @option{^-aO^/LIBRARY_SEARCH^} (@command{gnatbind})
7555 if you want to specify library paths
7556 only. This means that for the binder
7557 @option{^-I^/SEARCH=^}@var{dir} is equivalent to
7558 @option{^-aI^/SOURCE_SEARCH=^}@var{dir}
7559 @option{^-aO^/OBJECT_SEARCH=^}@var{dir}.
7560 The binder generates the bind file (a C language source file) in the
7561 current working directory.
7567 The packages @code{Ada}, @code{System}, and @code{Interfaces} and their
7568 children make up the GNAT Run-Time Library, together with the package
7569 GNAT and its children, which contain a set of useful additional
7570 library functions provided by GNAT. The sources for these units are
7571 needed by the compiler and are kept together in one directory. The ALI
7572 files and object files generated by compiling the RTL are needed by the
7573 binder and the linker and are kept together in one directory, typically
7574 different from the directory containing the sources. In a normal
7575 installation, you need not specify these directory names when compiling
7576 or binding. Either the environment variables or the built-in defaults
7577 cause these files to be found.
7579 Besides simplifying access to the RTL, a major use of search paths is
7580 in compiling sources from multiple directories. This can make
7581 development environments much more flexible.
7583 @node Examples of gnatbind Usage
7584 @section Examples of @code{gnatbind} Usage
7587 This section contains a number of examples of using the GNAT binding
7588 utility @code{gnatbind}.
7591 @item gnatbind hello
7592 The main program @code{Hello} (source program in @file{hello.adb}) is
7593 bound using the standard switch settings. The generated main program is
7594 @file{b~hello.adb}. This is the normal, default use of the binder.
7597 @item gnatbind hello -o mainprog.adb
7600 @item gnatbind HELLO.ALI /OUTPUT=Mainprog.ADB
7602 The main program @code{Hello} (source program in @file{hello.adb}) is
7603 bound using the standard switch settings. The generated main program is
7604 @file{mainprog.adb} with the associated spec in
7605 @file{mainprog.ads}. Note that you must specify the body here not the
7606 spec, in the case where the output is in Ada. Note that if this option
7607 is used, then linking must be done manually, since gnatlink will not
7608 be able to find the generated file.
7611 @item gnatbind main -C -o mainprog.c -x
7614 @item gnatbind MAIN.ALI /BIND_FILE=C /OUTPUT=Mainprog.C /READ_SOURCES=NONE
7616 The main program @code{Main} (source program in
7617 @file{main.adb}) is bound, excluding source files from the
7618 consistency checking, generating
7619 the file @file{mainprog.c}.
7622 @item gnatbind -x main_program -C -o mainprog.c
7623 This command is exactly the same as the previous example. Switches may
7624 appear anywhere in the command line, and single letter switches may be
7625 combined into a single switch.
7629 @item gnatbind -n math dbase -C -o ada-control.c
7632 @item gnatbind /NOMAIN math dbase /BIND_FILE=C /OUTPUT=ada-control.c
7634 The main program is in a language other than Ada, but calls to
7635 subprograms in packages @code{Math} and @code{Dbase} appear. This call
7636 to @code{gnatbind} generates the file @file{ada-control.c} containing
7637 the @code{adainit} and @code{adafinal} routines to be called before and
7638 after accessing the Ada units.
7642 @c ------------------------------------
7643 @node Linking Using gnatlink
7644 @chapter Linking Using @code{gnatlink}
7645 @c ------------------------------------
7649 This chapter discusses @code{gnatlink}, a tool that links
7650 an Ada program and builds an executable file. This utility
7651 invokes the system linker ^(via the @code{gcc} command)^^
7652 with a correct list of object files and library references.
7653 @code{gnatlink} automatically determines the list of files and
7654 references for the Ada part of a program. It uses the binder file
7655 generated by the @command{gnatbind} to determine this list.
7658 * Running gnatlink::
7659 * Switches for gnatlink::
7660 * Setting Stack Size from gnatlink::
7661 * Setting Heap Size from gnatlink::
7664 @node Running gnatlink
7665 @section Running @code{gnatlink}
7668 The form of the @code{gnatlink} command is
7671 $ gnatlink [@var{switches}] @var{mainprog}[.ali]
7672 [@var{non-Ada objects}] [@var{linker options}]
7676 The arguments of @code{gnatlink} (switches, main @file{ALI} file,
7678 or linker options) may be in any order, provided that no non-Ada object may
7679 be mistaken for a main @file{ALI} file.
7680 Any file name @file{F} without the @file{.ali}
7681 extension will be taken as the main @file{ALI} file if a file exists
7682 whose name is the concatenation of @file{F} and @file{.ali}.
7685 @file{@var{mainprog}.ali} references the ALI file of the main program.
7686 The @file{.ali} extension of this file can be omitted. From this
7687 reference, @code{gnatlink} locates the corresponding binder file
7688 @file{b~@var{mainprog}.adb} and, using the information in this file along
7689 with the list of non-Ada objects and linker options, constructs a
7690 linker command file to create the executable.
7692 The arguments other than the @code{gnatlink} switches and the main @file{ALI}
7693 file are passed to the linker uninterpreted.
7694 They typically include the names of
7695 object files for units written in other languages than Ada and any library
7696 references required to resolve references in any of these foreign language
7697 units, or in @code{Import} pragmas in any Ada units.
7699 @var{linker options} is an optional list of linker specific
7701 The default linker called by gnatlink is @var{gcc} which in
7702 turn calls the appropriate system linker.
7703 Standard options for the linker such as @option{-lmy_lib} or
7704 @option{-Ldir} can be added as is.
7705 For options that are not recognized by
7706 @var{gcc} as linker options, use the @var{gcc} switches @option{-Xlinker} or
7708 Refer to the GCC documentation for
7709 details. Here is an example showing how to generate a linker map:
7713 $ gnatlink my_prog -Wl,-Map,MAPFILE
7718 <<Need example for VMS>>
7721 Using @var{linker options} it is possible to set the program stack and
7722 heap size. See @ref{Setting Stack Size from gnatlink}, and
7723 @ref{Setting Heap Size from gnatlink}.
7725 @code{gnatlink} determines the list of objects required by the Ada
7726 program and prepends them to the list of objects passed to the linker.
7727 @code{gnatlink} also gathers any arguments set by the use of
7728 @code{pragma Linker_Options} and adds them to the list of arguments
7729 presented to the linker.
7732 @code{gnatlink} accepts the following types of extra files on the command
7733 line: objects (.OBJ), libraries (.OLB), sharable images (.EXE), and
7734 options files (.OPT). These are recognized and handled according to their
7738 @node Switches for gnatlink
7739 @section Switches for @code{gnatlink}
7742 The following switches are available with the @code{gnatlink} utility:
7747 @item ^-A^/BIND_FILE=ADA^
7748 @cindex @option{^-A^/BIND_FILE=ADA^} (@code{gnatlink})
7749 The binder has generated code in Ada. This is the default.
7751 @item ^-C^/BIND_FILE=C^
7752 @cindex @option{^-C^/BIND_FILE=C^} (@code{gnatlink})
7753 If instead of generating a file in Ada, the binder has generated one in
7754 C, then the linker needs to know about it. Use this switch to signal
7755 to @code{gnatlink} that the binder has generated C code rather than
7758 @item ^-f^/FORCE_OBJECT_FILE_LIST^
7759 @cindex Command line length
7760 @cindex @option{^-f^/FORCE_OBJECT_FILE_LIST^} (@code{gnatlink})
7761 On some targets, the command line length is limited, and @code{gnatlink}
7762 will generate a separate file for the linker if the list of object files
7764 The @option{^-f^/FORCE_OBJECT_FILE_LIST^} switch forces this file
7765 to be generated even if
7766 the limit is not exceeded. This is useful in some cases to deal with
7767 special situations where the command line length is exceeded.
7770 @cindex Debugging information, including
7771 @cindex @option{^-g^/DEBUG^} (@code{gnatlink})
7772 The option to include debugging information causes the Ada bind file (in
7773 other words, @file{b~@var{mainprog}.adb}) to be compiled with
7774 @option{^-g^/DEBUG^}.
7775 In addition, the binder does not delete the @file{b~@var{mainprog}.adb},
7776 @file{b~@var{mainprog}.o} and @file{b~@var{mainprog}.ali} files.
7777 Without @option{^-g^/DEBUG^}, the binder removes these files by
7778 default. The same procedure apply if a C bind file was generated using
7779 @option{^-C^/BIND_FILE=C^} @code{gnatbind} option, in this case the filenames
7780 are @file{b_@var{mainprog}.c} and @file{b_@var{mainprog}.o}.
7782 @item ^-n^/NOCOMPILE^
7783 @cindex @option{^-n^/NOCOMPILE^} (@code{gnatlink})
7784 Do not compile the file generated by the binder. This may be used when
7785 a link is rerun with different options, but there is no need to recompile
7789 @cindex @option{^-v^/VERBOSE^} (@code{gnatlink})
7790 Causes additional information to be output, including a full list of the
7791 included object files. This switch option is most useful when you want
7792 to see what set of object files are being used in the link step.
7794 @item ^-v -v^/VERBOSE/VERBOSE^
7795 @cindex @option{^-v -v^/VERBOSE/VERBOSE^} (@code{gnatlink})
7796 Very verbose mode. Requests that the compiler operate in verbose mode when
7797 it compiles the binder file, and that the system linker run in verbose mode.
7799 @item ^-o ^/EXECUTABLE=^@var{exec-name}
7800 @cindex @option{^-o^/EXECUTABLE^} (@code{gnatlink})
7801 @var{exec-name} specifies an alternate name for the generated
7802 executable program. If this switch is omitted, the executable has the same
7803 name as the main unit. For example, @code{gnatlink try.ali} creates
7804 an executable called @file{^try^TRY.EXE^}.
7807 @item -b @var{target}
7808 @cindex @option{-b} (@code{gnatlink})
7809 Compile your program to run on @var{target}, which is the name of a
7810 system configuration. You must have a GNAT cross-compiler built if
7811 @var{target} is not the same as your host system.
7814 @cindex @option{-B} (@code{gnatlink})
7815 Load compiler executables (for example, @code{gnat1}, the Ada compiler)
7816 from @var{dir} instead of the default location. Only use this switch
7817 when multiple versions of the GNAT compiler are available. See the
7818 @code{gcc} manual page for further details. You would normally use the
7819 @option{-b} or @option{-V} switch instead.
7821 @item --GCC=@var{compiler_name}
7822 @cindex @option{--GCC=compiler_name} (@code{gnatlink})
7823 Program used for compiling the binder file. The default is
7824 `@code{gcc}'. You need to use quotes around @var{compiler_name} if
7825 @code{compiler_name} contains spaces or other separator characters. As
7826 an example @option{--GCC="foo -x -y"} will instruct @code{gnatlink} to use
7827 @code{foo -x -y} as your compiler. Note that switch @option{-c} is always
7828 inserted after your command name. Thus in the above example the compiler
7829 command that will be used by @code{gnatlink} will be @code{foo -c -x -y}.
7830 If several @option{--GCC=compiler_name} are used, only the last
7831 @var{compiler_name} is taken into account. However, all the additional
7832 switches are also taken into account. Thus,
7833 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
7834 @option{--GCC="bar -x -y -z -t"}.
7836 @item --LINK=@var{name}
7837 @cindex @option{--LINK=} (@code{gnatlink})
7838 @var{name} is the name of the linker to be invoked. This is especially
7839 useful in mixed language programs since languages such as C++ require
7840 their own linker to be used. When this switch is omitted, the default
7841 name for the linker is (@file{gcc}). When this switch is used, the
7842 specified linker is called instead of (@file{gcc}) with exactly the same
7843 parameters that would have been passed to (@file{gcc}) so if the desired
7844 linker requires different parameters it is necessary to use a wrapper
7845 script that massages the parameters before invoking the real linker. It
7846 may be useful to control the exact invocation by using the verbose
7852 @item /DEBUG=TRACEBACK
7853 @cindex @code{/DEBUG=TRACEBACK} (@code{gnatlink})
7854 This qualifier causes sufficient information to be included in the
7855 executable file to allow a traceback, but does not include the full
7856 symbol information needed by the debugger.
7858 @item /IDENTIFICATION="<string>"
7859 @code{"<string>"} specifies the string to be stored in the image file
7860 identification field in the image header.
7861 It overrides any pragma @code{Ident} specified string.
7863 @item /NOINHIBIT-EXEC
7864 Generate the executable file even if there are linker warnings.
7866 @item /NOSTART_FILES
7867 Don't link in the object file containing the ``main'' transfer address.
7868 Used when linking with a foreign language main program compiled with a
7872 Prefer linking with object libraries over sharable images, even without
7878 @node Setting Stack Size from gnatlink
7879 @section Setting Stack Size from @code{gnatlink}
7882 Under Windows systems, it is possible to specify the program stack size from
7883 @code{gnatlink} using either:
7887 @item using @option{-Xlinker} linker option
7890 $ gnatlink hello -Xlinker --stack=0x10000,0x1000
7893 This sets the stack reserve size to 0x10000 bytes and the stack commit
7894 size to 0x1000 bytes.
7896 @item using @option{-Wl} linker option
7899 $ gnatlink hello -Wl,--stack=0x1000000
7902 This sets the stack reserve size to 0x1000000 bytes. Note that with
7903 @option{-Wl} option it is not possible to set the stack commit size
7904 because the coma is a separator for this option.
7908 @node Setting Heap Size from gnatlink
7909 @section Setting Heap Size from @code{gnatlink}
7912 Under Windows systems, it is possible to specify the program heap size from
7913 @code{gnatlink} using either:
7917 @item using @option{-Xlinker} linker option
7920 $ gnatlink hello -Xlinker --heap=0x10000,0x1000
7923 This sets the heap reserve size to 0x10000 bytes and the heap commit
7924 size to 0x1000 bytes.
7926 @item using @option{-Wl} linker option
7929 $ gnatlink hello -Wl,--heap=0x1000000
7932 This sets the heap reserve size to 0x1000000 bytes. Note that with
7933 @option{-Wl} option it is not possible to set the heap commit size
7934 because the coma is a separator for this option.
7938 @node The GNAT Make Program gnatmake
7939 @chapter The GNAT Make Program @code{gnatmake}
7943 * Running gnatmake::
7944 * Switches for gnatmake::
7945 * Mode Switches for gnatmake::
7946 * Notes on the Command Line::
7947 * How gnatmake Works::
7948 * Examples of gnatmake Usage::
7951 A typical development cycle when working on an Ada program consists of
7952 the following steps:
7956 Edit some sources to fix bugs.
7962 Compile all sources affected.
7972 The third step can be tricky, because not only do the modified files
7973 @cindex Dependency rules
7974 have to be compiled, but any files depending on these files must also be
7975 recompiled. The dependency rules in Ada can be quite complex, especially
7976 in the presence of overloading, @code{use} clauses, generics and inlined
7979 @code{gnatmake} automatically takes care of the third and fourth steps
7980 of this process. It determines which sources need to be compiled,
7981 compiles them, and binds and links the resulting object files.
7983 Unlike some other Ada make programs, the dependencies are always
7984 accurately recomputed from the new sources. The source based approach of
7985 the GNAT compilation model makes this possible. This means that if
7986 changes to the source program cause corresponding changes in
7987 dependencies, they will always be tracked exactly correctly by
7990 @node Running gnatmake
7991 @section Running @code{gnatmake}
7994 The usual form of the @code{gnatmake} command is
7997 $ gnatmake [@var{switches}] @var{file_name}
7998 [@var{file_names}] [@var{mode_switches}]
8002 The only required argument is one @var{file_name}, which specifies
8003 a compilation unit that is a main program. Several @var{file_names} can be
8004 specified: this will result in several executables being built.
8005 If @code{switches} are present, they can be placed before the first
8006 @var{file_name}, between @var{file_names} or after the last @var{file_name}.
8007 If @var{mode_switches} are present, they must always be placed after
8008 the last @var{file_name} and all @code{switches}.
8010 If you are using standard file extensions (.adb and .ads), then the
8011 extension may be omitted from the @var{file_name} arguments. However, if
8012 you are using non-standard extensions, then it is required that the
8013 extension be given. A relative or absolute directory path can be
8014 specified in a @var{file_name}, in which case, the input source file will
8015 be searched for in the specified directory only. Otherwise, the input
8016 source file will first be searched in the directory where
8017 @code{gnatmake} was invoked and if it is not found, it will be search on
8018 the source path of the compiler as described in
8019 @ref{Search Paths and the Run-Time Library (RTL)}.
8021 All @code{gnatmake} output (except when you specify
8022 @option{^-M^/DEPENDENCIES_LIST^}) is to
8023 @file{stderr}. The output produced by the
8024 @option{^-M^/DEPENDENCIES_LIST^} switch is send to
8027 @node Switches for gnatmake
8028 @section Switches for @code{gnatmake}
8031 You may specify any of the following switches to @code{gnatmake}:
8036 @item --GCC=@var{compiler_name}
8037 @cindex @option{--GCC=compiler_name} (@code{gnatmake})
8038 Program used for compiling. The default is `@code{gcc}'. You need to use
8039 quotes around @var{compiler_name} if @code{compiler_name} contains
8040 spaces or other separator characters. As an example @option{--GCC="foo -x
8041 -y"} will instruct @code{gnatmake} to use @code{foo -x -y} as your
8042 compiler. Note that switch @option{-c} is always inserted after your
8043 command name. Thus in the above example the compiler command that will
8044 be used by @code{gnatmake} will be @code{foo -c -x -y}.
8045 If several @option{--GCC=compiler_name} are used, only the last
8046 @var{compiler_name} is taken into account. However, all the additional
8047 switches are also taken into account. Thus,
8048 @option{--GCC="foo -x -y" --GCC="bar -z -t"} is equivalent to
8049 @option{--GCC="bar -x -y -z -t"}.
8051 @item --GNATBIND=@var{binder_name}
8052 @cindex @option{--GNATBIND=binder_name} (@code{gnatmake})
8053 Program used for binding. The default is `@code{gnatbind}'. You need to
8054 use quotes around @var{binder_name} if @var{binder_name} contains spaces
8055 or other separator characters. As an example @option{--GNATBIND="bar -x
8056 -y"} will instruct @code{gnatmake} to use @code{bar -x -y} as your
8057 binder. Binder switches that are normally appended by @code{gnatmake} to
8058 `@code{gnatbind}' are now appended to the end of @code{bar -x -y}.
8060 @item --GNATLINK=@var{linker_name}
8061 @cindex @option{--GNATLINK=linker_name} (@code{gnatmake})
8062 Program used for linking. The default is `@code{gnatlink}'. You need to
8063 use quotes around @var{linker_name} if @var{linker_name} contains spaces
8064 or other separator characters. As an example @option{--GNATLINK="lan -x
8065 -y"} will instruct @code{gnatmake} to use @code{lan -x -y} as your
8066 linker. Linker switches that are normally appended by @code{gnatmake} to
8067 `@code{gnatlink}' are now appended to the end of @code{lan -x -y}.
8071 @item ^-a^/ALL_FILES^
8072 @cindex @option{^-a^/ALL_FILES^} (@code{gnatmake})
8073 Consider all files in the make process, even the GNAT internal system
8074 files (for example, the predefined Ada library files), as well as any
8075 locked files. Locked files are files whose ALI file is write-protected.
8077 @code{gnatmake} does not check these files,
8078 because the assumption is that the GNAT internal files are properly up
8079 to date, and also that any write protected ALI files have been properly
8080 installed. Note that if there is an installation problem, such that one
8081 of these files is not up to date, it will be properly caught by the
8083 You may have to specify this switch if you are working on GNAT
8084 itself. The switch @option{^-a^/ALL_FILES^} is also useful
8085 in conjunction with @option{^-f^/FORCE_COMPILE^}
8086 if you need to recompile an entire application,
8087 including run-time files, using special configuration pragmas,
8088 such as a @code{Normalize_Scalars} pragma.
8091 @code{gnatmake ^-a^/ALL_FILES^} compiles all GNAT
8094 @code{gcc -c -gnatpg} rather than @code{gcc -c}.
8097 the @code{/CHECKS=SUPPRESS_ALL /STYLE_CHECKS=GNAT} switch.
8100 @item ^-b^/ACTIONS=BIND^
8101 @cindex @option{^-b^/ACTIONS=BIND^} (@code{gnatmake})
8102 Bind only. Can be combined with @option{^-c^/ACTIONS=COMPILE^} to do
8103 compilation and binding, but no link.
8104 Can be combined with @option{^-l^/ACTIONS=LINK^}
8105 to do binding and linking. When not combined with
8106 @option{^-c^/ACTIONS=COMPILE^}
8107 all the units in the closure of the main program must have been previously
8108 compiled and must be up to date. The root unit specified by @var{file_name}
8109 may be given without extension, with the source extension or, if no GNAT
8110 Project File is specified, with the ALI file extension.
8112 @item ^-c^/ACTIONS=COMPILE^
8113 @cindex @option{^-c^/ACTIONS=COMPILE^} (@code{gnatmake})
8114 Compile only. Do not perform binding, except when @option{^-b^/ACTIONS=BIND^}
8115 is also specified. Do not perform linking, except if both
8116 @option{^-b^/ACTIONS=BIND^} and
8117 @option{^-l^/ACTIONS=LINK^} are also specified.
8118 If the root unit specified by @var{file_name} is not a main unit, this is the
8119 default. Otherwise @code{gnatmake} will attempt binding and linking
8120 unless all objects are up to date and the executable is more recent than
8124 @cindex @option{^-C^/MAPPING^} (@code{gnatmake})
8125 Use a temporary mapping file. A mapping file is a way to communicate to the
8126 compiler two mappings: from unit names to file names (without any directory
8127 information) and from file names to path names (with full directory
8128 information). These mappings are used by the compiler to short-circuit the path
8129 search. When @code{gnatmake} is invoked with this switch, it will create
8130 a temporary mapping file, initially populated by the project manager,
8131 if @option{^-P^/PROJECT_FILE^} is used, otherwise initially empty.
8132 Each invocation of the compiler will add the newly accessed sources to the
8133 mapping file. This will improve the source search during the next invocation
8136 @item ^-C=^/USE_MAPPING_FILE=^@var{file}
8137 @cindex @option{^-C=^/USE_MAPPING^} (@code{gnatmake})
8138 Use a specific mapping file. The file, specified as a path name (absolute or
8139 relative) by this switch, should already exist, otherwise the switch is
8140 ineffective. The specified mapping file will be communicated to the compiler.
8141 This switch is not compatible with a project file
8142 (^-P^/PROJECT_FILE=^@var{file}) or with multiple compiling processes
8143 (^-j^/PROCESSES=^nnn, when nnn is greater than 1).
8145 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
8146 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatmake})
8147 Put all object files and ALI file in directory @var{dir}.
8148 If the @option{^-D^/DIRECTORY_OBJECTS^} switch is not used, all object files
8149 and ALI files go in the current working directory.
8151 This switch cannot be used when using a project file.
8153 @item ^-f^/FORCE_COMPILE^
8154 @cindex @option{^-f^/FORCE_COMPILE^} (@code{gnatmake})
8155 Force recompilations. Recompile all sources, even though some object
8156 files may be up to date, but don't recompile predefined or GNAT internal
8157 files or locked files (files with a write-protected ALI file),
8158 unless the @option{^-a^/ALL_FILES^} switch is also specified.
8160 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
8161 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatmake})
8162 When using project files, if some errors or warnings are detected during
8163 parsing and verbose mode is not in effect (no use of switch
8164 ^-v^/VERBOSE^), then error lines start with the full path name of the project
8165 file, rather than its simple file name.
8167 @item ^-i^/IN_PLACE^
8168 @cindex @option{^-i^/IN_PLACE^} (@code{gnatmake})
8169 In normal mode, @code{gnatmake} compiles all object files and ALI files
8170 into the current directory. If the @option{^-i^/IN_PLACE^} switch is used,
8171 then instead object files and ALI files that already exist are overwritten
8172 in place. This means that once a large project is organized into separate
8173 directories in the desired manner, then @code{gnatmake} will automatically
8174 maintain and update this organization. If no ALI files are found on the
8175 Ada object path (@ref{Search Paths and the Run-Time Library (RTL)}),
8176 the new object and ALI files are created in the
8177 directory containing the source being compiled. If another organization
8178 is desired, where objects and sources are kept in different directories,
8179 a useful technique is to create dummy ALI files in the desired directories.
8180 When detecting such a dummy file, @code{gnatmake} will be forced to recompile
8181 the corresponding source file, and it will be put the resulting object
8182 and ALI files in the directory where it found the dummy file.
8184 @item ^-j^/PROCESSES=^@var{n}
8185 @cindex @option{^-j^/PROCESSES^} (@code{gnatmake})
8186 @cindex Parallel make
8187 Use @var{n} processes to carry out the (re)compilations. On a
8188 multiprocessor machine compilations will occur in parallel. In the
8189 event of compilation errors, messages from various compilations might
8190 get interspersed (but @code{gnatmake} will give you the full ordered
8191 list of failing compiles at the end). If this is problematic, rerun
8192 the make process with n set to 1 to get a clean list of messages.
8194 @item ^-k^/CONTINUE_ON_ERROR^
8195 @cindex @option{^-k^/CONTINUE_ON_ERROR^} (@code{gnatmake})
8196 Keep going. Continue as much as possible after a compilation error. To
8197 ease the programmer's task in case of compilation errors, the list of
8198 sources for which the compile fails is given when @code{gnatmake}
8201 If @code{gnatmake} is invoked with several @file{file_names} and with this
8202 switch, if there are compilation errors when building an executable,
8203 @code{gnatmake} will not attempt to build the following executables.
8205 @item ^-l^/ACTIONS=LINK^
8206 @cindex @option{^-l^/ACTIONS=LINK^} (@code{gnatmake})
8207 Link only. Can be combined with @option{^-b^/ACTIONS=BIND^} to binding
8208 and linking. Linking will not be performed if combined with
8209 @option{^-c^/ACTIONS=COMPILE^}
8210 but not with @option{^-b^/ACTIONS=BIND^}.
8211 When not combined with @option{^-b^/ACTIONS=BIND^}
8212 all the units in the closure of the main program must have been previously
8213 compiled and must be up to date, and the main program need to have been bound.
8214 The root unit specified by @var{file_name}
8215 may be given without extension, with the source extension or, if no GNAT
8216 Project File is specified, with the ALI file extension.
8218 @item ^-m^/MINIMAL_RECOMPILATION^
8219 @cindex @option{^-m^/MINIMAL_RECOMPILATION^} (@code{gnatmake})
8220 Specifies that the minimum necessary amount of recompilations
8221 be performed. In this mode @code{gnatmake} ignores time
8222 stamp differences when the only
8223 modifications to a source file consist in adding/removing comments,
8224 empty lines, spaces or tabs. This means that if you have changed the
8225 comments in a source file or have simply reformatted it, using this
8226 switch will tell gnatmake not to recompile files that depend on it
8227 (provided other sources on which these files depend have undergone no
8228 semantic modifications). Note that the debugging information may be
8229 out of date with respect to the sources if the @option{-m} switch causes
8230 a compilation to be switched, so the use of this switch represents a
8231 trade-off between compilation time and accurate debugging information.
8233 @item ^-M^/DEPENDENCIES_LIST^
8234 @cindex Dependencies, producing list
8235 @cindex @option{^-M^/DEPENDENCIES_LIST^} (@code{gnatmake})
8236 Check if all objects are up to date. If they are, output the object
8237 dependences to @file{stdout} in a form that can be directly exploited in
8238 a @file{Makefile}. By default, each source file is prefixed with its
8239 (relative or absolute) directory name. This name is whatever you
8240 specified in the various @option{^-aI^/SOURCE_SEARCH^}
8241 and @option{^-I^/SEARCH^} switches. If you use
8242 @code{gnatmake ^-M^/DEPENDENCIES_LIST^}
8243 @option{^-q^/QUIET^}
8244 (see below), only the source file names,
8245 without relative paths, are output. If you just specify the
8246 @option{^-M^/DEPENDENCIES_LIST^}
8247 switch, dependencies of the GNAT internal system files are omitted. This
8248 is typically what you want. If you also specify
8249 the @option{^-a^/ALL_FILES^} switch,
8250 dependencies of the GNAT internal files are also listed. Note that
8251 dependencies of the objects in external Ada libraries (see switch
8252 @option{^-aL^/SKIP_MISSING=^}@var{dir} in the following list)
8255 @item ^-n^/DO_OBJECT_CHECK^
8256 @cindex @option{^-n^/DO_OBJECT_CHECK^} (@code{gnatmake})
8257 Don't compile, bind, or link. Checks if all objects are up to date.
8258 If they are not, the full name of the first file that needs to be
8259 recompiled is printed.
8260 Repeated use of this option, followed by compiling the indicated source
8261 file, will eventually result in recompiling all required units.
8263 @item ^-o ^/EXECUTABLE=^@var{exec_name}
8264 @cindex @option{^-o^/EXECUTABLE^} (@code{gnatmake})
8265 Output executable name. The name of the final executable program will be
8266 @var{exec_name}. If the @option{^-o^/EXECUTABLE^} switch is omitted the default
8267 name for the executable will be the name of the input file in appropriate form
8268 for an executable file on the host system.
8270 This switch cannot be used when invoking @code{gnatmake} with several
8273 @item ^-P^/PROJECT_FILE=^@var{project}
8274 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatmake})
8275 Use project file @var{project}. Only one such switch can be used.
8276 See @ref{gnatmake and Project Files}.
8279 @cindex @option{^-q^/QUIET^} (@code{gnatmake})
8280 Quiet. When this flag is not set, the commands carried out by
8281 @code{gnatmake} are displayed.
8283 @item ^-s^/SWITCH_CHECK/^
8284 @cindex @option{^-s^/SWITCH_CHECK^} (@code{gnatmake})
8285 Recompile if compiler switches have changed since last compilation.
8286 All compiler switches but -I and -o are taken into account in the
8288 orders between different ``first letter'' switches are ignored, but
8289 orders between same switches are taken into account. For example,
8290 @option{-O -O2} is different than @option{-O2 -O}, but @option{-g -O}
8291 is equivalent to @option{-O -g}.
8293 This switch is recommended when Integrated Preprocessing is used.
8296 @cindex @option{^-u^/UNIQUE^} (@code{gnatmake})
8297 Unique. Recompile at most the main files. It implies -c. Combined with
8298 -f, it is equivalent to calling the compiler directly. Note that using
8299 ^-u^/UNIQUE^ with a project file and no main has a special meaning
8300 (see @ref{Project Files and Main Subprograms}).
8302 @item ^-U^/ALL_PROJECTS^
8303 @cindex @option{^-U^/ALL_PROJECTS^} (@code{gnatmake})
8304 When used without a project file or with one or several mains on the command
8305 line, is equivalent to ^-u^/UNIQUE^. When used with a project file and no main
8306 on the command line, all sources of all project files are checked and compiled
8307 if not up to date, and libraries are rebuilt, if necessary.
8310 @cindex @option{^-v^/REASONS^} (@code{gnatmake})
8311 Verbose. Displays the reason for all recompilations @code{gnatmake}
8312 decides are necessary.
8314 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
8315 Indicates the verbosity of the parsing of GNAT project files.
8316 See @ref{Switches Related to Project Files}.
8318 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
8319 Indicates that external variable @var{name} has the value @var{value}.
8320 The Project Manager will use this value for occurrences of
8321 @code{external(name)} when parsing the project file.
8322 See @ref{Switches Related to Project Files}.
8325 @cindex @option{^-z^/NOMAIN^} (@code{gnatmake})
8326 No main subprogram. Bind and link the program even if the unit name
8327 given on the command line is a package name. The resulting executable
8328 will execute the elaboration routines of the package and its closure,
8329 then the finalization routines.
8332 @cindex @option{^-g^/DEBUG^} (@code{gnatmake})
8333 Enable debugging. This switch is simply passed to the compiler and to the
8339 @item @code{gcc} @asis{switches}
8341 Any uppercase switch (other than @option{-A},
8343 @option{-S}) or any switch that is more than one character is passed to
8344 @code{gcc} (e.g. @option{-O}, @option{-gnato,} etc.)
8347 Any qualifier that cannot be recognized as a qualifier for @code{GNAT MAKE}
8348 but is recognizable as a valid qualifier for @code{GNAT COMPILE} is
8349 automatically treated as a compiler switch, and passed on to all
8350 compilations that are carried out.
8355 Source and library search path switches:
8359 @item ^-aI^/SOURCE_SEARCH=^@var{dir}
8360 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatmake})
8361 When looking for source files also look in directory @var{dir}.
8362 The order in which source files search is undertaken is
8363 described in @ref{Search Paths and the Run-Time Library (RTL)}.
8365 @item ^-aL^/SKIP_MISSING=^@var{dir}
8366 @cindex @option{^-aL^/SKIP_MISSING^} (@code{gnatmake})
8367 Consider @var{dir} as being an externally provided Ada library.
8368 Instructs @code{gnatmake} to skip compilation units whose @file{.ALI}
8369 files have been located in directory @var{dir}. This allows you to have
8370 missing bodies for the units in @var{dir} and to ignore out of date bodies
8371 for the same units. You still need to specify
8372 the location of the specs for these units by using the switches
8373 @option{^-aI^/SOURCE_SEARCH=^@var{dir}}
8374 or @option{^-I^/SEARCH=^@var{dir}}.
8375 Note: this switch is provided for compatibility with previous versions
8376 of @code{gnatmake}. The easier method of causing standard libraries
8377 to be excluded from consideration is to write-protect the corresponding
8380 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
8381 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatmake})
8382 When searching for library and object files, look in directory
8383 @var{dir}. The order in which library files are searched is described in
8384 @ref{Search Paths for gnatbind}.
8386 @item ^-A^/CONDITIONAL_SOURCE_SEARCH=^@var{dir}
8387 @cindex Search paths, for @code{gnatmake}
8388 @cindex @option{^-A^/CONDITIONAL_SOURCE_SEARCH^} (@code{gnatmake})
8389 Equivalent to @option{^-aL^/SKIP_MISSING=^@var{dir}
8390 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
8392 @item ^-I^/SEARCH=^@var{dir}
8393 @cindex @option{^-I^/SEARCH^} (@code{gnatmake})
8394 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}
8395 ^-aI^/SOURCE_SEARCH=^@var{dir}}.
8397 @item ^-I-^/NOCURRENT_DIRECTORY^
8398 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatmake})
8399 @cindex Source files, suppressing search
8400 Do not look for source files in the directory containing the source
8401 file named in the command line.
8402 Do not look for ALI or object files in the directory
8403 where @code{gnatmake} was invoked.
8405 @item ^-L^/LIBRARY_SEARCH=^@var{dir}
8406 @cindex @option{^-L^/LIBRARY_SEARCH^} (@code{gnatmake})
8407 @cindex Linker libraries
8408 Add directory @var{dir} to the list of directories in which the linker
8409 will search for libraries. This is equivalent to
8410 @option{-largs ^-L^/LIBRARY_SEARCH=^}@var{dir}.
8412 Furthermore, under Windows, the sources pointed to by the libraries path
8413 set in the registry are not searched for.
8417 @cindex @option{-nostdinc} (@code{gnatmake})
8418 Do not look for source files in the system default directory.
8421 @cindex @option{-nostdlib} (@code{gnatmake})
8422 Do not look for library files in the system default directory.
8424 @item --RTS=@var{rts-path}
8425 @cindex @option{--RTS} (@code{gnatmake})
8426 Specifies the default location of the runtime library. GNAT looks for the
8428 in the following directories, and stops as soon as a valid runtime is found
8429 (@file{adainclude} or @file{ada_source_path}, and @file{adalib} or
8430 @file{ada_object_path} present):
8433 @item <current directory>/$rts_path
8435 @item <default-search-dir>/$rts_path
8437 @item <default-search-dir>/rts-$rts_path
8441 The selected path is handled like a normal RTS path.
8445 @node Mode Switches for gnatmake
8446 @section Mode Switches for @code{gnatmake}
8449 The mode switches (referred to as @code{mode_switches}) allow the
8450 inclusion of switches that are to be passed to the compiler itself, the
8451 binder or the linker. The effect of a mode switch is to cause all
8452 subsequent switches up to the end of the switch list, or up to the next
8453 mode switch, to be interpreted as switches to be passed on to the
8454 designated component of GNAT.
8458 @item -cargs @var{switches}
8459 @cindex @option{-cargs} (@code{gnatmake})
8460 Compiler switches. Here @var{switches} is a list of switches
8461 that are valid switches for @code{gcc}. They will be passed on to
8462 all compile steps performed by @code{gnatmake}.
8464 @item -bargs @var{switches}
8465 @cindex @option{-bargs} (@code{gnatmake})
8466 Binder switches. Here @var{switches} is a list of switches
8467 that are valid switches for @code{gnatbind}. They will be passed on to
8468 all bind steps performed by @code{gnatmake}.
8470 @item -largs @var{switches}
8471 @cindex @option{-largs} (@code{gnatmake})
8472 Linker switches. Here @var{switches} is a list of switches
8473 that are valid switches for @code{gnatlink}. They will be passed on to
8474 all link steps performed by @code{gnatmake}.
8476 @item -margs @var{switches}
8477 @cindex @option{-margs} (@code{gnatmake})
8478 Make switches. The switches are directly interpreted by @code{gnatmake},
8479 regardless of any previous occurrence of @option{-cargs}, @option{-bargs}
8483 @node Notes on the Command Line
8484 @section Notes on the Command Line
8487 This section contains some additional useful notes on the operation
8488 of the @code{gnatmake} command.
8492 @cindex Recompilation, by @code{gnatmake}
8493 If @code{gnatmake} finds no ALI files, it recompiles the main program
8494 and all other units required by the main program.
8495 This means that @code{gnatmake}
8496 can be used for the initial compile, as well as during subsequent steps of
8497 the development cycle.
8500 If you enter @code{gnatmake @var{file}.adb}, where @file{@var{file}.adb}
8501 is a subunit or body of a generic unit, @code{gnatmake} recompiles
8502 @file{@var{file}.adb} (because it finds no ALI) and stops, issuing a
8506 In @code{gnatmake} the switch @option{^-I^/SEARCH^}
8507 is used to specify both source and
8508 library file paths. Use @option{^-aI^/SOURCE_SEARCH^}
8509 instead if you just want to specify
8510 source paths only and @option{^-aO^/OBJECT_SEARCH^}
8511 if you want to specify library paths
8515 @code{gnatmake} examines both an ALI file and its corresponding object file
8516 for consistency. If an ALI is more recent than its corresponding object,
8517 or if the object file is missing, the corresponding source will be recompiled.
8518 Note that @code{gnatmake} expects an ALI and the corresponding object file
8519 to be in the same directory.
8522 @code{gnatmake} will ignore any files whose ALI file is write-protected.
8523 This may conveniently be used to exclude standard libraries from
8524 consideration and in particular it means that the use of the
8525 @option{^-f^/FORCE_COMPILE^} switch will not recompile these files
8526 unless @option{^-a^/ALL_FILES^} is also specified.
8529 @code{gnatmake} has been designed to make the use of Ada libraries
8530 particularly convenient. Assume you have an Ada library organized
8531 as follows: @i{^obj-dir^[OBJ_DIR]^} contains the objects and ALI files for
8532 of your Ada compilation units,
8533 whereas @i{^include-dir^[INCLUDE_DIR]^} contains the
8534 specs of these units, but no bodies. Then to compile a unit
8535 stored in @code{main.adb}, which uses this Ada library you would just type
8539 $ gnatmake -aI@var{include-dir} -aL@var{obj-dir} main
8542 $ gnatmake /SOURCE_SEARCH=@i{[INCLUDE_DIR]}
8543 /SKIP_MISSING=@i{[OBJ_DIR]} main
8548 Using @code{gnatmake} along with the
8549 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}
8550 switch provides a mechanism for avoiding unnecessary rcompilations. Using
8552 you can update the comments/format of your
8553 source files without having to recompile everything. Note, however, that
8554 adding or deleting lines in a source files may render its debugging
8555 info obsolete. If the file in question is a spec, the impact is rather
8556 limited, as that debugging info will only be useful during the
8557 elaboration phase of your program. For bodies the impact can be more
8558 significant. In all events, your debugger will warn you if a source file
8559 is more recent than the corresponding object, and alert you to the fact
8560 that the debugging information may be out of date.
8563 @node How gnatmake Works
8564 @section How @code{gnatmake} Works
8567 Generally @code{gnatmake} automatically performs all necessary
8568 recompilations and you don't need to worry about how it works. However,
8569 it may be useful to have some basic understanding of the @code{gnatmake}
8570 approach and in particular to understand how it uses the results of
8571 previous compilations without incorrectly depending on them.
8573 First a definition: an object file is considered @dfn{up to date} if the
8574 corresponding ALI file exists and its time stamp predates that of the
8575 object file and if all the source files listed in the
8576 dependency section of this ALI file have time stamps matching those in
8577 the ALI file. This means that neither the source file itself nor any
8578 files that it depends on have been modified, and hence there is no need
8579 to recompile this file.
8581 @code{gnatmake} works by first checking if the specified main unit is up
8582 to date. If so, no compilations are required for the main unit. If not,
8583 @code{gnatmake} compiles the main program to build a new ALI file that
8584 reflects the latest sources. Then the ALI file of the main unit is
8585 examined to find all the source files on which the main program depends,
8586 and @code{gnatmake} recursively applies the above procedure on all these files.
8588 This process ensures that @code{gnatmake} only trusts the dependencies
8589 in an existing ALI file if they are known to be correct. Otherwise it
8590 always recompiles to determine a new, guaranteed accurate set of
8591 dependencies. As a result the program is compiled ``upside down'' from what may
8592 be more familiar as the required order of compilation in some other Ada
8593 systems. In particular, clients are compiled before the units on which
8594 they depend. The ability of GNAT to compile in any order is critical in
8595 allowing an order of compilation to be chosen that guarantees that
8596 @code{gnatmake} will recompute a correct set of new dependencies if
8599 When invoking @code{gnatmake} with several @var{file_names}, if a unit is
8600 imported by several of the executables, it will be recompiled at most once.
8602 Note: when using non-standard naming conventions
8603 (See @ref{Using Other File Names}), changing through a configuration pragmas
8604 file the version of a source and invoking @code{gnatmake} to recompile may
8605 have no effect, if the previous version of the source is still accessible
8606 by @code{gnatmake}. It may be necessary to use the switch ^-f^/FORCE_COMPILE^.
8608 @node Examples of gnatmake Usage
8609 @section Examples of @code{gnatmake} Usage
8612 @item gnatmake hello.adb
8613 Compile all files necessary to bind and link the main program
8614 @file{hello.adb} (containing unit @code{Hello}) and bind and link the
8615 resulting object files to generate an executable file @file{^hello^HELLO.EXE^}.
8617 @item gnatmake main1 main2 main3
8618 Compile all files necessary to bind and link the main programs
8619 @file{main1.adb} (containing unit @code{Main1}), @file{main2.adb}
8620 (containing unit @code{Main2}) and @file{main3.adb}
8621 (containing unit @code{Main3}) and bind and link the resulting object files
8622 to generate three executable files @file{^main1^MAIN1.EXE^},
8623 @file{^main2^MAIN2.EXE^}
8624 and @file{^main3^MAIN3.EXE^}.
8627 @item gnatmake -q Main_Unit -cargs -O2 -bargs -l
8631 @item gnatmake Main_Unit /QUIET
8632 /COMPILER_QUALIFIERS /OPTIMIZE=ALL
8633 /BINDER_QUALIFIERS /ORDER_OF_ELABORATION
8635 Compile all files necessary to bind and link the main program unit
8636 @code{Main_Unit} (from file @file{main_unit.adb}). All compilations will
8637 be done with optimization level 2 and the order of elaboration will be
8638 listed by the binder. @code{gnatmake} will operate in quiet mode, not
8639 displaying commands it is executing.
8643 @c *************************
8644 @node Improving Performance
8645 @chapter Improving Performance
8646 @cindex Improving performance
8649 This chapter presents several topics related to program performance.
8650 It first describes some of the tradeoffs that need to be considered
8651 and some of the techniques for making your program run faster.
8652 It then documents the @command{gnatelim} tool, which can reduce
8653 the size of program executables.
8657 * Performance Considerations::
8658 * Reducing the Size of Ada Executables with gnatelim::
8663 @c *****************************
8664 @node Performance Considerations
8665 @section Performance Considerations
8668 The GNAT system provides a number of options that allow a trade-off
8673 performance of the generated code
8676 speed of compilation
8679 minimization of dependences and recompilation
8682 the degree of run-time checking.
8686 The defaults (if no options are selected) aim at improving the speed
8687 of compilation and minimizing dependences, at the expense of performance
8688 of the generated code:
8695 no inlining of subprogram calls
8698 all run-time checks enabled except overflow and elaboration checks
8702 These options are suitable for most program development purposes. This
8703 chapter describes how you can modify these choices, and also provides
8704 some guidelines on debugging optimized code.
8707 * Controlling Run-Time Checks::
8708 * Use of Restrictions::
8709 * Optimization Levels::
8710 * Debugging Optimized Code::
8711 * Inlining of Subprograms::
8712 * Optimization and Strict Aliasing::
8714 * Coverage Analysis::
8718 @node Controlling Run-Time Checks
8719 @subsection Controlling Run-Time Checks
8722 By default, GNAT generates all run-time checks, except arithmetic overflow
8723 checking for integer operations and checks for access before elaboration on
8724 subprogram calls. The latter are not required in default mode, because all
8725 necessary checking is done at compile time.
8726 @cindex @option{-gnatp} (@code{gcc})
8727 @cindex @option{-gnato} (@code{gcc})
8728 Two gnat switches, @option{-gnatp} and @option{-gnato} allow this default to
8729 be modified. @xref{Run-Time Checks}.
8731 Our experience is that the default is suitable for most development
8734 We treat integer overflow specially because these
8735 are quite expensive and in our experience are not as important as other
8736 run-time checks in the development process. Note that division by zero
8737 is not considered an overflow check, and divide by zero checks are
8738 generated where required by default.
8740 Elaboration checks are off by default, and also not needed by default, since
8741 GNAT uses a static elaboration analysis approach that avoids the need for
8742 run-time checking. This manual contains a full chapter discussing the issue
8743 of elaboration checks, and if the default is not satisfactory for your use,
8744 you should read this chapter.
8746 For validity checks, the minimal checks required by the Ada Reference
8747 Manual (for case statements and assignments to array elements) are on
8748 by default. These can be suppressed by use of the @option{-gnatVn} switch.
8749 Note that in Ada 83, there were no validity checks, so if the Ada 83 mode
8750 is acceptable (or when comparing GNAT performance with an Ada 83 compiler),
8751 it may be reasonable to routinely use @option{-gnatVn}. Validity checks
8752 are also suppressed entirely if @option{-gnatp} is used.
8754 @cindex Overflow checks
8755 @cindex Checks, overflow
8758 @cindex pragma Suppress
8759 @cindex pragma Unsuppress
8760 Note that the setting of the switches controls the default setting of
8761 the checks. They may be modified using either @code{pragma Suppress} (to
8762 remove checks) or @code{pragma Unsuppress} (to add back suppressed
8763 checks) in the program source.
8765 @node Use of Restrictions
8766 @subsection Use of Restrictions
8769 The use of pragma Restrictions allows you to control which features are
8770 permitted in your program. Apart from the obvious point that if you avoid
8771 relatively expensive features like finalization (enforceable by the use
8772 of pragma Restrictions (No_Finalization), the use of this pragma does not
8773 affect the generated code in most cases.
8775 One notable exception to this rule is that the possibility of task abort
8776 results in some distributed overhead, particularly if finalization or
8777 exception handlers are used. The reason is that certain sections of code
8778 have to be marked as non-abortable.
8780 If you use neither the @code{abort} statement, nor asynchronous transfer
8781 of control (@code{select .. then abort}), then this distributed overhead
8782 is removed, which may have a general positive effect in improving
8783 overall performance. Especially code involving frequent use of tasking
8784 constructs and controlled types will show much improved performance.
8785 The relevant restrictions pragmas are
8788 pragma Restrictions (No_Abort_Statements);
8789 pragma Restrictions (Max_Asynchronous_Select_Nesting => 0);
8793 It is recommended that these restriction pragmas be used if possible. Note
8794 that this also means that you can write code without worrying about the
8795 possibility of an immediate abort at any point.
8797 @node Optimization Levels
8798 @subsection Optimization Levels
8799 @cindex @option{^-O^/OPTIMIZE^} (@code{gcc})
8802 The default is optimization off. This results in the fastest compile
8803 times, but GNAT makes absolutely no attempt to optimize, and the
8804 generated programs are considerably larger and slower than when
8805 optimization is enabled. You can use the
8807 @option{-O@var{n}} switch, where @var{n} is an integer from 0 to 3,
8810 @code{OPTIMIZE} qualifier
8812 to @code{gcc} to control the optimization level:
8815 @item ^-O0^/OPTIMIZE=NONE^
8816 No optimization (the default);
8817 generates unoptimized code but has
8818 the fastest compilation time.
8820 @item ^-O1^/OPTIMIZE=SOME^
8821 Medium level optimization;
8822 optimizes reasonably well but does not
8823 degrade compilation time significantly.
8825 @item ^-O2^/OPTIMIZE=ALL^
8827 @itemx /OPTIMIZE=DEVELOPMENT
8830 generates highly optimized code and has
8831 the slowest compilation time.
8833 @item ^-O3^/OPTIMIZE=INLINING^
8834 Full optimization as in @option{-O2},
8835 and also attempts automatic inlining of small
8836 subprograms within a unit (@pxref{Inlining of Subprograms}).
8840 Higher optimization levels perform more global transformations on the
8841 program and apply more expensive analysis algorithms in order to generate
8842 faster and more compact code. The price in compilation time, and the
8843 resulting improvement in execution time,
8844 both depend on the particular application and the hardware environment.
8845 You should experiment to find the best level for your application.
8847 Since the precise set of optimizations done at each level will vary from
8848 release to release (and sometime from target to target), it is best to think
8849 of the optimization settings in general terms.
8850 The @cite{Using GNU GCC} manual contains details about
8851 ^the @option{-O} settings and a number of @option{-f} options that^how to^
8852 individually enable or disable specific optimizations.
8854 Unlike some other compilation systems, ^@command{gcc}^GNAT^ has
8855 been tested extensively at all optimization levels. There are some bugs
8856 which appear only with optimization turned on, but there have also been
8857 bugs which show up only in @emph{unoptimized} code. Selecting a lower
8858 level of optimization does not improve the reliability of the code
8859 generator, which in practice is highly reliable at all optimization
8862 Note regarding the use of @option{-O3}: The use of this optimization level
8863 is generally discouraged with GNAT, since it often results in larger
8864 executables which run more slowly. See further discussion of this point
8865 in @pxref{Inlining of Subprograms}.
8868 @node Debugging Optimized Code
8869 @subsection Debugging Optimized Code
8870 @cindex Debugging optimized code
8871 @cindex Optimization and debugging
8874 Although it is possible to do a reasonable amount of debugging at
8876 non-zero optimization levels,
8877 the higher the level the more likely that
8880 @option{/OPTIMIZE} settings other than @code{NONE},
8881 such settings will make it more likely that
8883 source-level constructs will have been eliminated by optimization.
8884 For example, if a loop is strength-reduced, the loop
8885 control variable may be completely eliminated and thus cannot be
8886 displayed in the debugger.
8887 This can only happen at @option{-O2} or @option{-O3}.
8888 Explicit temporary variables that you code might be eliminated at
8889 ^level^setting^ @option{-O1} or higher.
8891 The use of the @option{^-g^/DEBUG^} switch,
8892 @cindex @option{^-g^/DEBUG^} (@code{gcc})
8893 which is needed for source-level debugging,
8894 affects the size of the program executable on disk,
8895 and indeed the debugging information can be quite large.
8896 However, it has no effect on the generated code (and thus does not
8897 degrade performance)
8899 Since the compiler generates debugging tables for a compilation unit before
8900 it performs optimizations, the optimizing transformations may invalidate some
8901 of the debugging data. You therefore need to anticipate certain
8902 anomalous situations that may arise while debugging optimized code.
8903 These are the most common cases:
8907 @i{The ``hopping Program Counter'':} Repeated @code{step} or @code{next}
8909 the PC bouncing back and forth in the code. This may result from any of
8910 the following optimizations:
8914 @i{Common subexpression elimination:} using a single instance of code for a
8915 quantity that the source computes several times. As a result you
8916 may not be able to stop on what looks like a statement.
8919 @i{Invariant code motion:} moving an expression that does not change within a
8920 loop, to the beginning of the loop.
8923 @i{Instruction scheduling:} moving instructions so as to
8924 overlap loads and stores (typically) with other code, or in
8925 general to move computations of values closer to their uses. Often
8926 this causes you to pass an assignment statement without the assignment
8927 happening and then later bounce back to the statement when the
8928 value is actually needed. Placing a breakpoint on a line of code
8929 and then stepping over it may, therefore, not always cause all the
8930 expected side-effects.
8934 @i{The ``big leap'':} More commonly known as @emph{cross-jumping}, in which
8935 two identical pieces of code are merged and the program counter suddenly
8936 jumps to a statement that is not supposed to be executed, simply because
8937 it (and the code following) translates to the same thing as the code
8938 that @emph{was} supposed to be executed. This effect is typically seen in
8939 sequences that end in a jump, such as a @code{goto}, a @code{return}, or
8940 a @code{break} in a C @code{^switch^switch^} statement.
8943 @i{The ``roving variable'':} The symptom is an unexpected value in a variable.
8944 There are various reasons for this effect:
8948 In a subprogram prologue, a parameter may not yet have been moved to its
8952 A variable may be dead, and its register re-used. This is
8953 probably the most common cause.
8956 As mentioned above, the assignment of a value to a variable may
8960 A variable may be eliminated entirely by value propagation or
8961 other means. In this case, GCC may incorrectly generate debugging
8962 information for the variable
8966 In general, when an unexpected value appears for a local variable or parameter
8967 you should first ascertain if that value was actually computed by
8968 your program, as opposed to being incorrectly reported by the debugger.
8970 array elements in an object designated by an access value
8971 are generally less of a problem, once you have ascertained that the access
8973 Typically, this means checking variables in the preceding code and in the
8974 calling subprogram to verify that the value observed is explainable from other
8975 values (one must apply the procedure recursively to those
8976 other values); or re-running the code and stopping a little earlier
8977 (perhaps before the call) and stepping to better see how the variable obtained
8978 the value in question; or continuing to step @emph{from} the point of the
8979 strange value to see if code motion had simply moved the variable's
8984 In light of such anomalies, a recommended technique is to use @option{-O0}
8985 early in the software development cycle, when extensive debugging capabilities
8986 are most needed, and then move to @option{-O1} and later @option{-O2} as
8987 the debugger becomes less critical.
8988 Whether to use the @option{^-g^/DEBUG^} switch in the release version is
8989 a release management issue.
8991 Note that if you use @option{-g} you can then use the @command{strip} program
8992 on the resulting executable,
8993 which removes both debugging information and global symbols.
8997 @node Inlining of Subprograms
8998 @subsection Inlining of Subprograms
9001 A call to a subprogram in the current unit is inlined if all the
9002 following conditions are met:
9006 The optimization level is at least @option{-O1}.
9009 The called subprogram is suitable for inlining: It must be small enough
9010 and not contain nested subprograms or anything else that @code{gcc}
9011 cannot support in inlined subprograms.
9014 The call occurs after the definition of the body of the subprogram.
9017 @cindex pragma Inline
9019 Either @code{pragma Inline} applies to the subprogram or it is
9020 small and automatic inlining (optimization level @option{-O3}) is
9025 Calls to subprograms in @code{with}'ed units are normally not inlined.
9026 To achieve this level of inlining, the following conditions must all be
9031 The optimization level is at least @option{-O1}.
9034 The called subprogram is suitable for inlining: It must be small enough
9035 and not contain nested subprograms or anything else @code{gcc} cannot
9036 support in inlined subprograms.
9039 The call appears in a body (not in a package spec).
9042 There is a @code{pragma Inline} for the subprogram.
9045 @cindex @option{-gnatn} (@code{gcc})
9046 The @option{^-gnatn^/INLINE^} switch
9047 is used in the @code{gcc} command line
9050 Note that specifying the @option{-gnatn} switch causes additional
9051 compilation dependencies. Consider the following:
9053 @smallexample @c ada
9073 With the default behavior (no @option{-gnatn} switch specified), the
9074 compilation of the @code{Main} procedure depends only on its own source,
9075 @file{main.adb}, and the spec of the package in file @file{r.ads}. This
9076 means that editing the body of @code{R} does not require recompiling
9079 On the other hand, the call @code{R.Q} is not inlined under these
9080 circumstances. If the @option{-gnatn} switch is present when @code{Main}
9081 is compiled, the call will be inlined if the body of @code{Q} is small
9082 enough, but now @code{Main} depends on the body of @code{R} in
9083 @file{r.adb} as well as on the spec. This means that if this body is edited,
9084 the main program must be recompiled. Note that this extra dependency
9085 occurs whether or not the call is in fact inlined by @code{gcc}.
9087 The use of front end inlining with @option{-gnatN} generates similar
9088 additional dependencies.
9090 @cindex @option{^-fno-inline^/INLINE=SUPPRESS^} (@code{gcc})
9091 Note: The @option{^-fno-inline^/INLINE=SUPPRESS^} switch
9092 can be used to prevent
9093 all inlining. This switch overrides all other conditions and ensures
9094 that no inlining occurs. The extra dependences resulting from
9095 @option{-gnatn} will still be active, even if
9096 this switch is used to suppress the resulting inlining actions.
9098 Note regarding the use of @option{-O3}: There is no difference in inlining
9099 behavior between @option{-O2} and @option{-O3} for subprograms with an explicit
9100 pragma @code{Inline} assuming the use of @option{-gnatn}
9101 or @option{-gnatN} (the switches that activate inlining). If you have used
9102 pragma @code{Inline} in appropriate cases, then it is usually much better
9103 to use @option{-O2} and @option{-gnatn} and avoid the use of @option{-O3} which
9104 in this case only has the effect of inlining subprograms you did not
9105 think should be inlined. We often find that the use of @option{-O3} slows
9106 down code by performing excessive inlining, leading to increased instruction
9107 cache pressure from the increased code size. So the bottom line here is
9108 that you should not automatically assume that @option{-O3} is better than
9109 @option{-O2}, and indeed you should use @option{-O3} only if tests show that
9110 it actually improves performance.
9112 @node Optimization and Strict Aliasing
9113 @subsection Optimization and Strict Aliasing
9115 @cindex Strict Aliasing
9116 @cindex No_Strict_Aliasing
9119 The strong typing capabilities of Ada allow an optimizer to generate
9120 efficient code in situations where other languages would be forced to
9121 make worst case assumptions preventing such optimizations. Consider
9122 the following example:
9124 @smallexample @c ada
9127 type Int1 is new Integer;
9128 type Int2 is new Integer;
9129 type Int1A is access Int1;
9130 type Int2A is access Int2;
9137 for J in Data'Range loop
9138 if Data (J) = Int1V.all then
9139 Int2V.all := Int2V.all + 1;
9148 In this example, since the variable @code{Int1V} can only access objects
9149 of type @code{Int1}, and @code{Int2V} can only access objects of type
9150 @code{Int2}, there is no possibility that the assignment to
9151 @code{Int2V.all} affects the value of @code{Int1V.all}. This means that
9152 the compiler optimizer can "know" that the value @code{Int1V.all} is constant
9153 for all iterations of the loop and avoid the extra memory reference
9154 required to dereference it each time through the loop.
9156 This kind of optimziation, called strict aliasing analysis, is
9157 triggered by specifying an optimization level of @option{-O2} or
9158 higher and allows @code{GNAT} to generate more efficient code
9159 when access values are involved.
9161 However, although this optimization is always correct in terms of
9162 the formal semantics of the Ada Reference Manual, difficulties can
9163 arise if features like @code{Unchecked_Conversion} are used to break
9164 the typing system. Consider the following complete program example:
9166 @smallexample @c ada
9169 type int1 is new integer;
9170 type int2 is new integer;
9171 type a1 is access int1;
9172 type a2 is access int2;
9177 function to_a2 (Input : a1) return a2;
9180 with Unchecked_Conversion;
9182 function to_a2 (Input : a1) return a2 is
9184 new Unchecked_Conversion (a1, a2);
9186 return to_a2u (Input);
9192 with Text_IO; use Text_IO;
9194 v1 : a1 := new int1;
9195 v2 : a2 := to_a2 (v1);
9199 put_line (int1'image (v1.all));
9205 This program prints out 0 in @code{-O0} or @code{-O1}
9206 mode, but it prints out 1 in @code{-O2} mode. That's
9207 because in strict aliasing mode, the compiler can and
9208 does assume that the assignment to @code{v2.all} could not
9209 affect the value of @code{v1.all}, since different types
9212 This behavior is not a case of non-conformance with the standard, since
9213 the Ada RM specifies that an unchecked conversion where the resulting
9214 bit pattern is not a correct value of the target type can result in an
9215 abnormal value and attempting to reference an abnormal value makes the
9216 execution of a program erroneous. That's the case here since the result
9217 does not point to an object of type @code{int2}. This means that the
9218 effect is entirely unpredictable.
9220 However, although that explanation may satisfy a language
9221 lawyer, in practice an applications programmer expects an
9222 unchecked conversion involving pointers to create true
9223 aliases and the behavior of printing 1 seems plain wrong.
9224 In this case, the strict aliasing optimization is unwelcome.
9226 Indeed the compiler recognizes this possibility, and the
9227 unchecked conversion generates a warning:
9230 p2.adb:5:07: warning: possible aliasing problem with type "a2"
9231 p2.adb:5:07: warning: use -fno-strict-aliasing switch for references
9232 p2.adb:5:07: warning: or use "pragma No_Strict_Aliasing (a2);"
9236 Unfortunately the problem is recognized when compiling the body of
9237 package @code{p2}, but the actual "bad" code is generated while
9238 compiling the body of @code{m} and this latter compilation does not see
9239 the suspicious @code{Unchecked_Conversion}.
9241 As implied by the warning message, there are approaches you can use to
9242 avoid the unwanted strict aliasing optimization in a case like this.
9244 One possibility is to simply avoid the use of @code{-O2}, but
9245 that is a bit drastic, since it throws away a number of useful
9246 optimizations that do not involve strict aliasing assumptions.
9248 A less drastic approach is to compile the program using the
9249 option @code{-fno-strict-aliasing}. Actually it is only the
9250 unit containing the dereferencing of the suspicious pointer
9251 that needs to be compiled. So in this case, if we compile
9252 unit @code{m} with this switch, then we get the expected
9253 value of zero printed. Analyzing which units might need
9254 the switch can be painful, so a more reasonable approach
9255 is to compile the entire program with options @code{-O2}
9256 and @code{-fno-strict-aliasing}. If the performance is
9257 satisfactory with this combination of options, then the
9258 advantage is that the entire issue of possible "wrong"
9259 optimization due to strict aliasing is avoided.
9261 To avoid the use of compiler switches, the configuration
9262 pragma @code{No_Strict_Aliasing} with no parameters may be
9263 used to specify that for all access types, the strict
9264 aliasing optimization should be suppressed.
9266 However, these approaches are still overkill, in that they causes
9267 all manipulations of all access values to be deoptimized. A more
9268 refined approach is to concentrate attention on the specific
9269 access type identified as problematic.
9271 First, if a careful analysis of uses of the pointer shows
9272 that there are no possible problematic references, then
9273 the warning can be suppressed by bracketing the
9274 instantiation of @code{Unchecked_Conversion} to turn
9277 @smallexample @c ada
9278 pragma Warnings (Off);
9280 new Unchecked_Conversion (a1, a2);
9281 pragma Warnings (On);
9285 Of course that approach is not appropriate for this particular
9286 example, since indeed there is a problematic reference. In this
9287 case we can take one of two other approaches.
9289 The first possibility is to move the instantiation of unchecked
9290 conversion to the unit in which the type is declared. In
9291 this example, we would move the instantiation of
9292 @code{Unchecked_Conversion} from the body of package
9293 @code{p2} to the spec of package @code{p1}. Now the
9294 warning disappears. That's because any use of the
9295 access type knows there is a suspicious unchecked
9296 conversion, and the strict aliasing optimization
9297 is automatically suppressed for the type.
9299 If it is not practical to move the unchecked conversion to the same unit
9300 in which the destination access type is declared (perhaps because the
9301 source type is not visible in that unit), you may use pragma
9302 @code{No_Strict_Aliasing} for the type. This pragma must occur in the
9303 same declarative sequence as the declaration of the access type:
9305 @smallexample @c ada
9306 type a2 is access int2;
9307 pragma No_Strict_Aliasing (a2);
9311 Here again, the compiler now knows that the strict aliasing optimization
9312 should be suppressed for any reference to type @code{a2} and the
9313 expected behavior is obtained.
9315 Finally, note that although the compiler can generate warnings for
9316 simple cases of unchecked conversions, there are tricker and more
9317 indirect ways of creating type incorrect aliases which the compiler
9318 cannot detect. Examples are the use of address overlays and unchecked
9319 conversions involving composite types containing access types as
9320 components. In such cases, no warnings are generated, but there can
9321 still be aliasing problems. One safe coding practice is to forbid the
9322 use of address clauses for type overlaying, and to allow unchecked
9323 conversion only for primitive types. This is not really a significant
9324 restriction since any possible desired effect can be achieved by
9325 unchecked conversion of access values.
9328 @node Coverage Analysis
9329 @subsection Coverage Analysis
9332 GNAT supports the Digital Performance Coverage Analyzer (PCA), which allows
9333 the user to determine the distribution of execution time across a program,
9334 @pxref{Profiling} for details of usage.
9337 @node Reducing the Size of Ada Executables with gnatelim
9338 @section Reducing the Size of Ada Executables with @code{gnatelim}
9342 This section describes @command{gnatelim}, a tool which detects unused
9343 subprograms and helps the compiler to create a smaller executable for your
9348 * Running gnatelim::
9349 * Correcting the List of Eliminate Pragmas::
9350 * Making Your Executables Smaller::
9351 * Summary of the gnatelim Usage Cycle::
9354 @node About gnatelim
9355 @subsection About @code{gnatelim}
9358 When a program shares a set of Ada
9359 packages with other programs, it may happen that this program uses
9360 only a fraction of the subprograms defined in these packages. The code
9361 created for these unused subprograms increases the size of the executable.
9363 @code{gnatelim} tracks unused subprograms in an Ada program and
9364 outputs a list of GNAT-specific pragmas @code{Eliminate} marking all the
9365 subprograms that are declared but never called. By placing the list of
9366 @code{Eliminate} pragmas in the GNAT configuration file @file{gnat.adc} and
9367 recompiling your program, you may decrease the size of its executable,
9368 because the compiler will not generate the code for 'eliminated' subprograms.
9369 See GNAT Reference Manual for more information about this pragma.
9371 @code{gnatelim} needs as its input data the name of the main subprogram
9372 and a bind file for a main subprogram.
9374 To create a bind file for @code{gnatelim}, run @code{gnatbind} for
9375 the main subprogram. @code{gnatelim} can work with both Ada and C
9376 bind files; when both are present, it uses the Ada bind file.
9377 The following commands will build the program and create the bind file:
9380 $ gnatmake ^-c Main_Prog^/ACTIONS=COMPILE MAIN_PROG^
9381 $ gnatbind main_prog
9384 Note that @code{gnatelim} needs neither object nor ALI files.
9386 @node Running gnatelim
9387 @subsection Running @code{gnatelim}
9390 @code{gnatelim} has the following command-line interface:
9393 $ gnatelim [options] name
9397 @code{name} should be a name of a source file that contains the main subprogram
9398 of a program (partition).
9400 @code{gnatelim} has the following switches:
9405 @cindex @option{^-q^/QUIET^} (@command{gnatelim})
9406 Quiet mode: by default @code{gnatelim} outputs to the standard error
9407 stream the number of program units left to be processed. This option turns
9411 @cindex @option{^-v^/VERBOSE^} (@command{gnatelim})
9412 Verbose mode: @code{gnatelim} version information is printed as Ada
9413 comments to the standard output stream. Also, in addition to the number of
9414 program units left @code{gnatelim} will output the name of the current unit
9418 @cindex @option{^-a^/ALL^} (@command{gnatelim})
9419 Also look for subprograms from the GNAT run time that can be eliminated. Note
9420 that when @file{gnat.adc} is produced using this switch, the entire program
9421 must be recompiled with switch @option{^-a^/ALL_FILES^} to @code{gnatmake}.
9423 @item ^-I^/INCLUDE_DIRS=^@var{dir}
9424 @cindex @option{^-I^/INCLUDE_DIRS^} (@command{gnatelim})
9425 When looking for source files also look in directory @var{dir}. Specifying
9426 @option{^-I-^/INCLUDE_DIRS=-^} instructs @code{gnatelim} not to look for
9427 sources in the current directory.
9429 @item ^-b^/BIND_FILE=^@var{bind_file}
9430 @cindex @option{^-b^/BIND_FILE^} (@command{gnatelim})
9431 Specifies @var{bind_file} as the bind file to process. If not set, the name
9432 of the bind file is computed from the full expanded Ada name
9433 of a main subprogram.
9435 @item ^-C^/CONFIG_FILE=^@var{config_file}
9436 @cindex @option{^-C^/CONFIG_FILE^} (@command{gnatelim})
9437 Specifies a file @var{config_file} that contains configuration pragmas. The
9438 file must be specified with full path.
9440 @item ^--GCC^/COMPILER^=@var{compiler_name}
9441 @cindex @option{^-GCC^/COMPILER^} (@command{gnatelim})
9442 Instructs @code{gnatelim} to use specific @code{gcc} compiler instead of one
9443 available on the path.
9445 @item ^--GNATMAKE^/GNATMAKE^=@var{gnatmake_name}
9446 @cindex @option{^--GNATMAKE^/GNATMAKE^} (@command{gnatelim})
9447 Instructs @code{gnatelim} to use specific @code{gnatmake} instead of one
9448 available on the path.
9451 @cindex @option{-d@var{x}} (@command{gnatelim})
9452 Activate internal debugging switches. @var{x} is a letter or digit, or
9453 string of letters or digits, which specifies the type of debugging
9454 mode desired. Normally these are used only for internal development
9455 or system debugging purposes. You can find full documentation for these
9456 switches in the spec of the @code{Gnatelim} unit in the compiler
9457 source file @file{gnatelim.ads}.
9461 @code{gnatelim} sends its output to the standard output stream, and all the
9462 tracing and debug information is sent to the standard error stream.
9463 In order to produce a proper GNAT configuration file
9464 @file{gnat.adc}, redirection must be used:
9468 $ PIPE GNAT ELIM MAIN_PROG.ADB > GNAT.ADC
9471 $ gnatelim main_prog.adb > gnat.adc
9480 $ gnatelim main_prog.adb >> gnat.adc
9484 in order to append the @code{gnatelim} output to the existing contents of
9488 @node Correcting the List of Eliminate Pragmas
9489 @subsection Correcting the List of Eliminate Pragmas
9492 In some rare cases @code{gnatelim} may try to eliminate
9493 subprograms that are actually called in the program. In this case, the
9494 compiler will generate an error message of the form:
9497 file.adb:106:07: cannot call eliminated subprogram "My_Prog"
9501 You will need to manually remove the wrong @code{Eliminate} pragmas from
9502 the @file{gnat.adc} file. You should recompile your program
9503 from scratch after that, because you need a consistent @file{gnat.adc} file
9504 during the entire compilation.
9507 @node Making Your Executables Smaller
9508 @subsection Making Your Executables Smaller
9511 In order to get a smaller executable for your program you now have to
9512 recompile the program completely with the new @file{gnat.adc} file
9513 created by @code{gnatelim} in your current directory:
9516 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
9520 (Use the @option{^-f^/FORCE_COMPILE^} option for @command{gnatmake} to
9521 recompile everything
9522 with the set of pragmas @code{Eliminate} that you have obtained with
9523 @command{gnatelim}).
9525 Be aware that the set of @code{Eliminate} pragmas is specific to each
9526 program. It is not recommended to merge sets of @code{Eliminate}
9527 pragmas created for different programs in one @file{gnat.adc} file.
9529 @node Summary of the gnatelim Usage Cycle
9530 @subsection Summary of the gnatelim Usage Cycle
9533 Here is a quick summary of the steps to be taken in order to reduce
9534 the size of your executables with @code{gnatelim}. You may use
9535 other GNAT options to control the optimization level,
9536 to produce the debugging information, to set search path, etc.
9543 $ gnatmake ^-c main_prog^/ACTIONS=COMPILE MAIN_PROG^
9544 $ gnatbind main_prog
9548 Generate a list of @code{Eliminate} pragmas
9551 $ PIPE GNAT ELIM MAIN_PROG > GNAT.ADC
9554 $ gnatelim main_prog >[>] gnat.adc
9559 Recompile the application
9562 $ gnatmake ^-f main_prog^/FORCE_COMPILE MAIN_PROG^
9570 @c ********************************
9571 @node Renaming Files Using gnatchop
9572 @chapter Renaming Files Using @code{gnatchop}
9576 This chapter discusses how to handle files with multiple units by using
9577 the @code{gnatchop} utility. This utility is also useful in renaming
9578 files to meet the standard GNAT default file naming conventions.
9581 * Handling Files with Multiple Units::
9582 * Operating gnatchop in Compilation Mode::
9583 * Command Line for gnatchop::
9584 * Switches for gnatchop::
9585 * Examples of gnatchop Usage::
9588 @node Handling Files with Multiple Units
9589 @section Handling Files with Multiple Units
9592 The basic compilation model of GNAT requires that a file submitted to the
9593 compiler have only one unit and there be a strict correspondence
9594 between the file name and the unit name.
9596 The @code{gnatchop} utility allows both of these rules to be relaxed,
9597 allowing GNAT to process files which contain multiple compilation units
9598 and files with arbitrary file names. @code{gnatchop}
9599 reads the specified file and generates one or more output files,
9600 containing one unit per file. The unit and the file name correspond,
9601 as required by GNAT.
9603 If you want to permanently restructure a set of ``foreign'' files so that
9604 they match the GNAT rules, and do the remaining development using the
9605 GNAT structure, you can simply use @command{gnatchop} once, generate the
9606 new set of files and work with them from that point on.
9608 Alternatively, if you want to keep your files in the ``foreign'' format,
9609 perhaps to maintain compatibility with some other Ada compilation
9610 system, you can set up a procedure where you use @command{gnatchop} each
9611 time you compile, regarding the source files that it writes as temporary
9612 files that you throw away.
9615 @node Operating gnatchop in Compilation Mode
9616 @section Operating gnatchop in Compilation Mode
9619 The basic function of @code{gnatchop} is to take a file with multiple units
9620 and split it into separate files. The boundary between files is reasonably
9621 clear, except for the issue of comments and pragmas. In default mode, the
9622 rule is that any pragmas between units belong to the previous unit, except
9623 that configuration pragmas always belong to the following unit. Any comments
9624 belong to the following unit. These rules
9625 almost always result in the right choice of
9626 the split point without needing to mark it explicitly and most users will
9627 find this default to be what they want. In this default mode it is incorrect to
9628 submit a file containing only configuration pragmas, or one that ends in
9629 configuration pragmas, to @code{gnatchop}.
9631 However, using a special option to activate ``compilation mode'',
9633 can perform another function, which is to provide exactly the semantics
9634 required by the RM for handling of configuration pragmas in a compilation.
9635 In the absence of configuration pragmas (at the main file level), this
9636 option has no effect, but it causes such configuration pragmas to be handled
9637 in a quite different manner.
9639 First, in compilation mode, if @code{gnatchop} is given a file that consists of
9640 only configuration pragmas, then this file is appended to the
9641 @file{gnat.adc} file in the current directory. This behavior provides
9642 the required behavior described in the RM for the actions to be taken
9643 on submitting such a file to the compiler, namely that these pragmas
9644 should apply to all subsequent compilations in the same compilation
9645 environment. Using GNAT, the current directory, possibly containing a
9646 @file{gnat.adc} file is the representation
9647 of a compilation environment. For more information on the
9648 @file{gnat.adc} file, see the section on handling of configuration
9649 pragmas @pxref{Handling of Configuration Pragmas}.
9651 Second, in compilation mode, if @code{gnatchop}
9652 is given a file that starts with
9653 configuration pragmas, and contains one or more units, then these
9654 configuration pragmas are prepended to each of the chopped files. This
9655 behavior provides the required behavior described in the RM for the
9656 actions to be taken on compiling such a file, namely that the pragmas
9657 apply to all units in the compilation, but not to subsequently compiled
9660 Finally, if configuration pragmas appear between units, they are appended
9661 to the previous unit. This results in the previous unit being illegal,
9662 since the compiler does not accept configuration pragmas that follow
9663 a unit. This provides the required RM behavior that forbids configuration
9664 pragmas other than those preceding the first compilation unit of a
9667 For most purposes, @code{gnatchop} will be used in default mode. The
9668 compilation mode described above is used only if you need exactly
9669 accurate behavior with respect to compilations, and you have files
9670 that contain multiple units and configuration pragmas. In this
9671 circumstance the use of @code{gnatchop} with the compilation mode
9672 switch provides the required behavior, and is for example the mode
9673 in which GNAT processes the ACVC tests.
9675 @node Command Line for gnatchop
9676 @section Command Line for @code{gnatchop}
9679 The @code{gnatchop} command has the form:
9682 $ gnatchop switches @var{file name} [@var{file name} @var{file name} ...]
9687 The only required argument is the file name of the file to be chopped.
9688 There are no restrictions on the form of this file name. The file itself
9689 contains one or more Ada units, in normal GNAT format, concatenated
9690 together. As shown, more than one file may be presented to be chopped.
9692 When run in default mode, @code{gnatchop} generates one output file in
9693 the current directory for each unit in each of the files.
9695 @var{directory}, if specified, gives the name of the directory to which
9696 the output files will be written. If it is not specified, all files are
9697 written to the current directory.
9699 For example, given a
9700 file called @file{hellofiles} containing
9702 @smallexample @c ada
9707 with Text_IO; use Text_IO;
9720 $ gnatchop ^hellofiles^HELLOFILES.^
9724 generates two files in the current directory, one called
9725 @file{hello.ads} containing the single line that is the procedure spec,
9726 and the other called @file{hello.adb} containing the remaining text. The
9727 original file is not affected. The generated files can be compiled in
9731 When gnatchop is invoked on a file that is empty or that contains only empty
9732 lines and/or comments, gnatchop will not fail, but will not produce any
9735 For example, given a
9736 file called @file{toto.txt} containing
9738 @smallexample @c ada
9750 $ gnatchop ^toto.txt^TOT.TXT^
9754 will not produce any new file and will result in the following warnings:
9757 toto.txt:1:01: warning: empty file, contains no compilation units
9758 no compilation units found
9759 no source files written
9762 @node Switches for gnatchop
9763 @section Switches for @code{gnatchop}
9766 @command{gnatchop} recognizes the following switches:
9771 @item ^-c^/COMPILATION^
9772 @cindex @option{^-c^/COMPILATION^} (@code{gnatchop})
9773 Causes @code{gnatchop} to operate in compilation mode, in which
9774 configuration pragmas are handled according to strict RM rules. See
9775 previous section for a full description of this mode.
9779 This passes the given @option{-gnatxxx} switch to @code{gnat} which is
9780 used to parse the given file. Not all @code{xxx} options make sense,
9781 but for example, the use of @option{-gnati2} allows @code{gnatchop} to
9782 process a source file that uses Latin-2 coding for identifiers.
9786 Causes @code{gnatchop} to generate a brief help summary to the standard
9787 output file showing usage information.
9789 @item ^-k@var{mm}^/FILE_NAME_MAX_LENGTH=@var{mm}^
9790 @cindex @option{^-k^/FILE_NAME_MAX_LENGTH^} (@code{gnatchop})
9791 Limit generated file names to the specified number @code{mm}
9793 This is useful if the
9794 resulting set of files is required to be interoperable with systems
9795 which limit the length of file names.
9797 If no value is given, or
9798 if no @code{/FILE_NAME_MAX_LENGTH} qualifier is given,
9799 a default of 39, suitable for OpenVMS Alpha
9803 No space is allowed between the @option{-k} and the numeric value. The numeric
9804 value may be omitted in which case a default of @option{-k8},
9806 with DOS-like file systems, is used. If no @option{-k} switch
9808 there is no limit on the length of file names.
9811 @item ^-p^/PRESERVE^
9812 @cindex @option{^-p^/PRESERVE^} (@code{gnatchop})
9813 Causes the file ^modification^creation^ time stamp of the input file to be
9814 preserved and used for the time stamp of the output file(s). This may be
9815 useful for preserving coherency of time stamps in an environment where
9816 @code{gnatchop} is used as part of a standard build process.
9819 @cindex @option{^-q^/QUIET^} (@code{gnatchop})
9820 Causes output of informational messages indicating the set of generated
9821 files to be suppressed. Warnings and error messages are unaffected.
9823 @item ^-r^/REFERENCE^
9824 @cindex @option{^-r^/REFERENCE^} (@code{gnatchop})
9825 @findex Source_Reference
9826 Generate @code{Source_Reference} pragmas. Use this switch if the output
9827 files are regarded as temporary and development is to be done in terms
9828 of the original unchopped file. This switch causes
9829 @code{Source_Reference} pragmas to be inserted into each of the
9830 generated files to refers back to the original file name and line number.
9831 The result is that all error messages refer back to the original
9833 In addition, the debugging information placed into the object file (when
9834 the @option{^-g^/DEBUG^} switch of @code{gcc} or @code{gnatmake} is specified)
9835 also refers back to this original file so that tools like profilers and
9836 debuggers will give information in terms of the original unchopped file.
9838 If the original file to be chopped itself contains
9839 a @code{Source_Reference}
9840 pragma referencing a third file, then gnatchop respects
9841 this pragma, and the generated @code{Source_Reference} pragmas
9842 in the chopped file refer to the original file, with appropriate
9843 line numbers. This is particularly useful when @code{gnatchop}
9844 is used in conjunction with @code{gnatprep} to compile files that
9845 contain preprocessing statements and multiple units.
9848 @cindex @option{^-v^/VERBOSE^} (@code{gnatchop})
9849 Causes @code{gnatchop} to operate in verbose mode. The version
9850 number and copyright notice are output, as well as exact copies of
9851 the gnat1 commands spawned to obtain the chop control information.
9853 @item ^-w^/OVERWRITE^
9854 @cindex @option{^-w^/OVERWRITE^} (@code{gnatchop})
9855 Overwrite existing file names. Normally @code{gnatchop} regards it as a
9856 fatal error if there is already a file with the same name as a
9857 file it would otherwise output, in other words if the files to be
9858 chopped contain duplicated units. This switch bypasses this
9859 check, and causes all but the last instance of such duplicated
9860 units to be skipped.
9864 @cindex @option{--GCC=} (@code{gnatchop})
9865 Specify the path of the GNAT parser to be used. When this switch is used,
9866 no attempt is made to add the prefix to the GNAT parser executable.
9870 @node Examples of gnatchop Usage
9871 @section Examples of @code{gnatchop} Usage
9875 @item gnatchop /OVERWRITE HELLO_S.ADA [PRERELEASE.FILES]
9878 @item gnatchop -w hello_s.ada prerelease/files
9881 Chops the source file @file{hello_s.ada}. The output files will be
9882 placed in the directory @file{^prerelease/files^[PRERELEASE.FILES]^},
9884 files with matching names in that directory (no files in the current
9885 directory are modified).
9887 @item gnatchop ^archive^ARCHIVE.^
9888 Chops the source file @file{^archive^ARCHIVE.^}
9889 into the current directory. One
9890 useful application of @code{gnatchop} is in sending sets of sources
9891 around, for example in email messages. The required sources are simply
9892 concatenated (for example, using a ^Unix @code{cat}^VMS @code{APPEND/NEW}^
9894 @code{gnatchop} is used at the other end to reconstitute the original
9897 @item gnatchop file1 file2 file3 direc
9898 Chops all units in files @file{file1}, @file{file2}, @file{file3}, placing
9899 the resulting files in the directory @file{direc}. Note that if any units
9900 occur more than once anywhere within this set of files, an error message
9901 is generated, and no files are written. To override this check, use the
9902 @option{^-w^/OVERWRITE^} switch,
9903 in which case the last occurrence in the last file will
9904 be the one that is output, and earlier duplicate occurrences for a given
9905 unit will be skipped.
9908 @node Configuration Pragmas
9909 @chapter Configuration Pragmas
9910 @cindex Configuration pragmas
9911 @cindex Pragmas, configuration
9914 In Ada 95, configuration pragmas include those pragmas described as
9915 such in the Ada 95 Reference Manual, as well as
9916 implementation-dependent pragmas that are configuration pragmas. See the
9917 individual descriptions of pragmas in the GNAT Reference Manual for
9918 details on these additional GNAT-specific configuration pragmas. Most
9919 notably, the pragma @code{Source_File_Name}, which allows
9920 specifying non-default names for source files, is a configuration
9921 pragma. The following is a complete list of configuration pragmas
9922 recognized by @code{GNAT}:
9934 External_Name_Casing
9935 Float_Representation
9942 Propagate_Exceptions
9951 Task_Dispatching_Policy
9960 * Handling of Configuration Pragmas::
9961 * The Configuration Pragmas Files::
9964 @node Handling of Configuration Pragmas
9965 @section Handling of Configuration Pragmas
9967 Configuration pragmas may either appear at the start of a compilation
9968 unit, in which case they apply only to that unit, or they may apply to
9969 all compilations performed in a given compilation environment.
9971 GNAT also provides the @code{gnatchop} utility to provide an automatic
9972 way to handle configuration pragmas following the semantics for
9973 compilations (that is, files with multiple units), described in the RM.
9974 See section @pxref{Operating gnatchop in Compilation Mode} for details.
9975 However, for most purposes, it will be more convenient to edit the
9976 @file{gnat.adc} file that contains configuration pragmas directly,
9977 as described in the following section.
9979 @node The Configuration Pragmas Files
9980 @section The Configuration Pragmas Files
9981 @cindex @file{gnat.adc}
9984 In GNAT a compilation environment is defined by the current
9985 directory at the time that a compile command is given. This current
9986 directory is searched for a file whose name is @file{gnat.adc}. If
9987 this file is present, it is expected to contain one or more
9988 configuration pragmas that will be applied to the current compilation.
9989 However, if the switch @option{-gnatA} is used, @file{gnat.adc} is not
9992 Configuration pragmas may be entered into the @file{gnat.adc} file
9993 either by running @code{gnatchop} on a source file that consists only of
9994 configuration pragmas, or more conveniently by
9995 direct editing of the @file{gnat.adc} file, which is a standard format
9998 In addition to @file{gnat.adc}, one additional file containing configuration
9999 pragmas may be applied to the current compilation using the switch
10000 @option{-gnatec}@var{path}. @var{path} must designate an existing file that
10001 contains only configuration pragmas. These configuration pragmas are
10002 in addition to those found in @file{gnat.adc} (provided @file{gnat.adc}
10003 is present and switch @option{-gnatA} is not used).
10005 It is allowed to specify several switches @option{-gnatec}, however only
10006 the last one on the command line will be taken into account.
10008 If you are using project file, a separate mechanism is provided using
10009 project attributes, see @ref{Specifying Configuration Pragmas} for more
10013 Of special interest to GNAT OpenVMS Alpha is the following
10014 configuration pragma:
10016 @smallexample @c ada
10018 pragma Extend_System (Aux_DEC);
10023 In the presence of this pragma, GNAT adds to the definition of the
10024 predefined package SYSTEM all the additional types and subprograms that are
10025 defined in DEC Ada. See @pxref{Compatibility with DEC Ada} for details.
10028 @node Handling Arbitrary File Naming Conventions Using gnatname
10029 @chapter Handling Arbitrary File Naming Conventions Using @code{gnatname}
10030 @cindex Arbitrary File Naming Conventions
10033 * Arbitrary File Naming Conventions::
10034 * Running gnatname::
10035 * Switches for gnatname::
10036 * Examples of gnatname Usage::
10039 @node Arbitrary File Naming Conventions
10040 @section Arbitrary File Naming Conventions
10043 The GNAT compiler must be able to know the source file name of a compilation
10044 unit. When using the standard GNAT default file naming conventions
10045 (@code{.ads} for specs, @code{.adb} for bodies), the GNAT compiler
10046 does not need additional information.
10049 When the source file names do not follow the standard GNAT default file naming
10050 conventions, the GNAT compiler must be given additional information through
10051 a configuration pragmas file (see @ref{Configuration Pragmas})
10053 When the non standard file naming conventions are well-defined,
10054 a small number of pragmas @code{Source_File_Name} specifying a naming pattern
10055 (see @ref{Alternative File Naming Schemes}) may be sufficient. However,
10056 if the file naming conventions are irregular or arbitrary, a number
10057 of pragma @code{Source_File_Name} for individual compilation units
10059 To help maintain the correspondence between compilation unit names and
10060 source file names within the compiler,
10061 GNAT provides a tool @code{gnatname} to generate the required pragmas for a
10064 @node Running gnatname
10065 @section Running @code{gnatname}
10068 The usual form of the @code{gnatname} command is
10071 $ gnatname [@var{switches}] @var{naming_pattern} [@var{naming_patterns}]
10075 All of the arguments are optional. If invoked without any argument,
10076 @code{gnatname} will display its usage.
10079 When used with at least one naming pattern, @code{gnatname} will attempt to
10080 find all the compilation units in files that follow at least one of the
10081 naming patterns. To find these compilation units,
10082 @code{gnatname} will use the GNAT compiler in syntax-check-only mode on all
10086 One or several Naming Patterns may be given as arguments to @code{gnatname}.
10087 Each Naming Pattern is enclosed between double quotes.
10088 A Naming Pattern is a regular expression similar to the wildcard patterns
10089 used in file names by the Unix shells or the DOS prompt.
10092 Examples of Naming Patterns are
10101 For a more complete description of the syntax of Naming Patterns,
10102 see the second kind of regular expressions described in @file{g-regexp.ads}
10103 (the ``Glob'' regular expressions).
10106 When invoked with no switches, @code{gnatname} will create a configuration
10107 pragmas file @file{gnat.adc} in the current working directory, with pragmas
10108 @code{Source_File_Name} for each file that contains a valid Ada unit.
10110 @node Switches for gnatname
10111 @section Switches for @code{gnatname}
10114 Switches for @code{gnatname} must precede any specified Naming Pattern.
10117 You may specify any of the following switches to @code{gnatname}:
10122 @item ^-c^/CONFIG_FILE=^@file{file}
10123 @cindex @option{^-c^/CONFIG_FILE^} (@code{gnatname})
10124 Create a configuration pragmas file @file{file} (instead of the default
10127 There may be zero, one or more space between @option{-c} and
10130 @file{file} may include directory information. @file{file} must be
10131 writable. There may be only one switch @option{^-c^/CONFIG_FILE^}.
10132 When a switch @option{^-c^/CONFIG_FILE^} is
10133 specified, no switch @option{^-P^/PROJECT_FILE^} may be specified (see below).
10135 @item ^-d^/SOURCE_DIRS=^@file{dir}
10136 @cindex @option{^-d^/SOURCE_DIRS^} (@code{gnatname})
10137 Look for source files in directory @file{dir}. There may be zero, one or more
10138 spaces between @option{^-d^/SOURCE_DIRS=^} and @file{dir}.
10139 When a switch @option{^-d^/SOURCE_DIRS^}
10140 is specified, the current working directory will not be searched for source
10141 files, unless it is explicitly specified with a @option{^-d^/SOURCE_DIRS^}
10142 or @option{^-D^/DIR_FILES^} switch.
10143 Several switches @option{^-d^/SOURCE_DIRS^} may be specified.
10144 If @file{dir} is a relative path, it is relative to the directory of
10145 the configuration pragmas file specified with switch
10146 @option{^-c^/CONFIG_FILE^},
10147 or to the directory of the project file specified with switch
10148 @option{^-P^/PROJECT_FILE^} or,
10149 if neither switch @option{^-c^/CONFIG_FILE^}
10150 nor switch @option{^-P^/PROJECT_FILE^} are specified, it is relative to the
10151 current working directory. The directory
10152 specified with switch @option{^-d^/SOURCE_DIRS^} must exist and be readable.
10154 @item ^-D^/DIRS_FILE=^@file{file}
10155 @cindex @option{^-D^/DIRS_FILE^} (@code{gnatname})
10156 Look for source files in all directories listed in text file @file{file}.
10157 There may be zero, one or more spaces between @option{^-D^/DIRS_FILE=^}
10159 @file{file} must be an existing, readable text file.
10160 Each non empty line in @file{file} must be a directory.
10161 Specifying switch @option{^-D^/DIRS_FILE^} is equivalent to specifying as many
10162 switches @option{^-d^/SOURCE_DIRS^} as there are non empty lines in
10165 @item ^-f^/FOREIGN_PATTERN=^@file{pattern}
10166 @cindex @option{^-f^/FOREIGN_PATTERN^} (@code{gnatname})
10167 Foreign patterns. Using this switch, it is possible to add sources of languages
10168 other than Ada to the list of sources of a project file.
10169 It is only useful if a ^-P^/PROJECT_FILE^ switch is used.
10172 gnatname ^-Pprj -f"*.c"^/PROJECT_FILE=PRJ /FOREIGN_PATTERN=*.C^ "*.ada"
10175 will look for Ada units in all files with the @file{.ada} extension,
10176 and will add to the list of file for project @file{prj.gpr} the C files
10177 with extension ".^c^C^".
10180 @cindex @option{^-h^/HELP^} (@code{gnatname})
10181 Output usage (help) information. The output is written to @file{stdout}.
10183 @item ^-P^/PROJECT_FILE=^@file{proj}
10184 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatname})
10185 Create or update project file @file{proj}. There may be zero, one or more space
10186 between @option{-P} and @file{proj}. @file{proj} may include directory
10187 information. @file{proj} must be writable.
10188 There may be only one switch @option{^-P^/PROJECT_FILE^}.
10189 When a switch @option{^-P^/PROJECT_FILE^} is specified,
10190 no switch @option{^-c^/CONFIG_FILE^} may be specified.
10192 @item ^-v^/VERBOSE^
10193 @cindex @option{^-v^/VERBOSE^} (@code{gnatname})
10194 Verbose mode. Output detailed explanation of behavior to @file{stdout}.
10195 This includes name of the file written, the name of the directories to search
10196 and, for each file in those directories whose name matches at least one of
10197 the Naming Patterns, an indication of whether the file contains a unit,
10198 and if so the name of the unit.
10200 @item ^-v -v^/VERBOSE /VERBOSE^
10201 @cindex @option{^-v -v^/VERBOSE /VERBOSE^} (@code{gnatname})
10202 Very Verbose mode. In addition to the output produced in verbose mode,
10203 for each file in the searched directories whose name matches none of
10204 the Naming Patterns, an indication is given that there is no match.
10206 @item ^-x^/EXCLUDED_PATTERN=^@file{pattern}
10207 @cindex @option{^-x^/EXCLUDED_PATTERN^} (@code{gnatname})
10208 Excluded patterns. Using this switch, it is possible to exclude some files
10209 that would match the name patterns. For example,
10211 gnatname ^-x "*_nt.ada"^/EXCLUDED_PATTERN=*_nt.ada^ "*.ada"
10214 will look for Ada units in all files with the @file{.ada} extension,
10215 except those whose names end with @file{_nt.ada}.
10219 @node Examples of gnatname Usage
10220 @section Examples of @code{gnatname} Usage
10224 $ gnatname /CONFIG_FILE=[HOME.ME]NAMES.ADC /SOURCE_DIRS=SOURCES "[a-z]*.ada*"
10230 $ gnatname -c /home/me/names.adc -d sources "[a-z]*.ada*"
10235 In this example, the directory @file{^/home/me^[HOME.ME]^} must already exist
10236 and be writable. In addition, the directory
10237 @file{^/home/me/sources^[HOME.ME.SOURCES]^} (specified by
10238 @option{^-d sources^/SOURCE_DIRS=SOURCES^}) must exist and be readable.
10241 Note the optional spaces after @option{-c} and @option{-d}.
10246 $ gnatname -P/home/me/proj -x "*_nt_body.ada"
10247 -dsources -dsources/plus -Dcommon_dirs.txt "body_*" "spec_*"
10250 $ gnatname /PROJECT_FILE=[HOME.ME]PROJ
10251 /EXCLUDED_PATTERN=*_nt_body.ada
10252 /SOURCE_DIRS=(SOURCES,[SOURCES.PLUS])
10253 /DIRS_FILE=COMMON_DIRS.TXT "body_*" "spec_*"
10257 Note that several switches @option{^-d^/SOURCE_DIRS^} may be used,
10258 even in conjunction with one or several switches
10259 @option{^-D^/DIRS_FILE^}. Several Naming Patterns and one excluded pattern
10260 are used in this example.
10263 @c *****************************************
10264 @c * G N A T P r o j e c t M a n a g e r *
10265 @c *****************************************
10266 @node GNAT Project Manager
10267 @chapter GNAT Project Manager
10271 * Examples of Project Files::
10272 * Project File Syntax::
10273 * Objects and Sources in Project Files::
10274 * Importing Projects::
10275 * Project Extension::
10276 * External References in Project Files::
10277 * Packages in Project Files::
10278 * Variables from Imported Projects::
10280 * Library Projects::
10281 * Using Third-Party Libraries through Projects::
10282 * Stand-alone Library Projects::
10283 * Switches Related to Project Files::
10284 * Tools Supporting Project Files::
10285 * An Extended Example::
10286 * Project File Complete Syntax::
10289 @c ****************
10290 @c * Introduction *
10291 @c ****************
10294 @section Introduction
10297 This chapter describes GNAT's @emph{Project Manager}, a facility that allows
10298 you to manage complex builds involving a number of source files, directories,
10299 and compilation options for different system configurations. In particular,
10300 project files allow you to specify:
10303 The directory or set of directories containing the source files, and/or the
10304 names of the specific source files themselves
10306 The directory in which the compiler's output
10307 (@file{ALI} files, object files, tree files) is to be placed
10309 The directory in which the executable programs is to be placed
10311 ^Switch^Switch^ settings for any of the project-enabled tools
10312 (@command{gnatmake}, compiler, binder, linker, @code{gnatls}, @code{gnatxref},
10313 @code{gnatfind}); you can apply these settings either globally or to individual
10316 The source files containing the main subprogram(s) to be built
10318 The source programming language(s) (currently Ada and/or C)
10320 Source file naming conventions; you can specify these either globally or for
10321 individual compilation units
10328 @node Project Files
10329 @subsection Project Files
10332 Project files are written in a syntax close to that of Ada, using familiar
10333 notions such as packages, context clauses, declarations, default values,
10334 assignments, and inheritance. Finally, project files can be built
10335 hierarchically from other project files, simplifying complex system
10336 integration and project reuse.
10338 A @dfn{project} is a specific set of values for various compilation properties.
10339 The settings for a given project are described by means of
10340 a @dfn{project file}, which is a text file written in an Ada-like syntax.
10341 Property values in project files are either strings or lists of strings.
10342 Properties that are not explicitly set receive default values. A project
10343 file may interrogate the values of @dfn{external variables} (user-defined
10344 command-line switches or environment variables), and it may specify property
10345 settings conditionally, based on the value of such variables.
10347 In simple cases, a project's source files depend only on other source files
10348 in the same project, or on the predefined libraries. (@emph{Dependence} is
10350 the Ada technical sense; as in one Ada unit @code{with}ing another.) However,
10351 the Project Manager also allows more sophisticated arrangements,
10352 where the source files in one project depend on source files in other
10356 One project can @emph{import} other projects containing needed source files.
10358 You can organize GNAT projects in a hierarchy: a @emph{child} project
10359 can extend a @emph{parent} project, inheriting the parent's source files and
10360 optionally overriding any of them with alternative versions
10364 More generally, the Project Manager lets you structure large development
10365 efforts into hierarchical subsystems, where build decisions are delegated
10366 to the subsystem level, and thus different compilation environments
10367 (^switch^switch^ settings) used for different subsystems.
10369 The Project Manager is invoked through the
10370 @option{^-P^/PROJECT_FILE=^@emph{projectfile}}
10371 switch to @command{gnatmake} or to the @command{^gnat^GNAT^} front driver.
10373 There may be zero, one or more spaces between @option{-P} and
10374 @option{@emph{projectfile}}.
10376 If you want to define (on the command line) an external variable that is
10377 queried by the project file, you must use the
10378 @option{^-X^/EXTERNAT_REFERENCE=^@emph{vbl}=@emph{value}} switch.
10379 The Project Manager parses and interprets the project file, and drives the
10380 invoked tool based on the project settings.
10382 The Project Manager supports a wide range of development strategies,
10383 for systems of all sizes. Here are some typical practices that are
10387 Using a common set of source files, but generating object files in different
10388 directories via different ^switch^switch^ settings
10390 Using a mostly-shared set of source files, but with different versions of
10395 The destination of an executable can be controlled inside a project file
10396 using the @option{^-o^-o^}
10398 In the absence of such a ^switch^switch^ either inside
10399 the project file or on the command line, any executable files generated by
10400 @command{gnatmake} are placed in the directory @code{Exec_Dir} specified
10401 in the project file. If no @code{Exec_Dir} is specified, they will be placed
10402 in the object directory of the project.
10404 You can use project files to achieve some of the effects of a source
10405 versioning system (for example, defining separate projects for
10406 the different sets of sources that comprise different releases) but the
10407 Project Manager is independent of any source configuration management tools
10408 that might be used by the developers.
10410 The next section introduces the main features of GNAT's project facility
10411 through a sequence of examples; subsequent sections will present the syntax
10412 and semantics in more detail. A more formal description of the project
10413 facility appears in the GNAT Reference Manual.
10415 @c *****************************
10416 @c * Examples of Project Files *
10417 @c *****************************
10419 @node Examples of Project Files
10420 @section Examples of Project Files
10422 This section illustrates some of the typical uses of project files and
10423 explains their basic structure and behavior.
10426 * Common Sources with Different ^Switches^Switches^ and Directories::
10427 * Using External Variables::
10428 * Importing Other Projects::
10429 * Extending a Project::
10432 @node Common Sources with Different ^Switches^Switches^ and Directories
10433 @subsection Common Sources with Different ^Switches^Switches^ and Directories
10437 * Specifying the Object Directory::
10438 * Specifying the Exec Directory::
10439 * Project File Packages::
10440 * Specifying ^Switch^Switch^ Settings::
10441 * Main Subprograms::
10442 * Executable File Names::
10443 * Source File Naming Conventions::
10444 * Source Language(s)::
10448 Suppose that the Ada source files @file{pack.ads}, @file{pack.adb}, and
10449 @file{proc.adb} are in the @file{/common} directory. The file
10450 @file{proc.adb} contains an Ada main subprogram @code{Proc} that @code{with}s
10451 package @code{Pack}. We want to compile these source files under two sets
10452 of ^switches^switches^:
10455 When debugging, we want to pass the @option{-g} switch to @command{gnatmake},
10456 and the @option{^-gnata^-gnata^},
10457 @option{^-gnato^-gnato^},
10458 and @option{^-gnatE^-gnatE^} switches to the
10459 compiler; the compiler's output is to appear in @file{/common/debug}
10461 When preparing a release version, we want to pass the @option{^-O2^O2^} switch
10462 to the compiler; the compiler's output is to appear in @file{/common/release}
10466 The GNAT project files shown below, respectively @file{debug.gpr} and
10467 @file{release.gpr} in the @file{/common} directory, achieve these effects.
10480 ^/common/debug^[COMMON.DEBUG]^
10485 ^/common/release^[COMMON.RELEASE]^
10490 Here are the corresponding project files:
10492 @smallexample @c projectfile
10495 for Object_Dir use "debug";
10496 for Main use ("proc");
10499 for ^Default_Switches^Default_Switches^ ("Ada")
10501 for Executable ("proc.adb") use "proc1";
10506 package Compiler is
10507 for ^Default_Switches^Default_Switches^ ("Ada")
10508 use ("-fstack-check",
10511 "^-gnatE^-gnatE^");
10517 @smallexample @c projectfile
10520 for Object_Dir use "release";
10521 for Exec_Dir use ".";
10522 for Main use ("proc");
10524 package Compiler is
10525 for ^Default_Switches^Default_Switches^ ("Ada")
10533 The name of the project defined by @file{debug.gpr} is @code{"Debug"} (case
10534 insensitive), and analogously the project defined by @file{release.gpr} is
10535 @code{"Release"}. For consistency the file should have the same name as the
10536 project, and the project file's extension should be @code{"gpr"}. These
10537 conventions are not required, but a warning is issued if they are not followed.
10539 If the current directory is @file{^/temp^[TEMP]^}, then the command
10541 gnatmake ^-P/common/debug.gpr^/PROJECT_FILE=[COMMON]DEBUG^
10545 generates object and ALI files in @file{^/common/debug^[COMMON.DEBUG]^},
10546 as well as the @code{^proc1^PROC1.EXE^} executable,
10547 using the ^switch^switch^ settings defined in the project file.
10549 Likewise, the command
10551 gnatmake ^-P/common/release.gpr^/PROJECT_FILE=[COMMON]RELEASE^
10555 generates object and ALI files in @file{^/common/release^[COMMON.RELEASE]^},
10556 and the @code{^proc^PROC.EXE^}
10557 executable in @file{^/common^[COMMON]^},
10558 using the ^switch^switch^ settings from the project file.
10561 @unnumberedsubsubsec Source Files
10564 If a project file does not explicitly specify a set of source directories or
10565 a set of source files, then by default the project's source files are the
10566 Ada source files in the project file directory. Thus @file{pack.ads},
10567 @file{pack.adb}, and @file{proc.adb} are the source files for both projects.
10569 @node Specifying the Object Directory
10570 @unnumberedsubsubsec Specifying the Object Directory
10573 Several project properties are modeled by Ada-style @emph{attributes};
10574 a property is defined by supplying the equivalent of an Ada attribute
10575 definition clause in the project file.
10576 A project's object directory is another such a property; the corresponding
10577 attribute is @code{Object_Dir}, and its value is also a string expression,
10578 specified either as absolute or relative. In the later case,
10579 it is relative to the project file directory. Thus the compiler's
10580 output is directed to @file{^/common/debug^[COMMON.DEBUG]^}
10581 (for the @code{Debug} project)
10582 and to @file{^/common/release^[COMMON.RELEASE]^}
10583 (for the @code{Release} project).
10584 If @code{Object_Dir} is not specified, then the default is the project file
10587 @node Specifying the Exec Directory
10588 @unnumberedsubsubsec Specifying the Exec Directory
10591 A project's exec directory is another property; the corresponding
10592 attribute is @code{Exec_Dir}, and its value is also a string expression,
10593 either specified as relative or absolute. If @code{Exec_Dir} is not specified,
10594 then the default is the object directory (which may also be the project file
10595 directory if attribute @code{Object_Dir} is not specified). Thus the executable
10596 is placed in @file{^/common/debug^[COMMON.DEBUG]^}
10597 for the @code{Debug} project (attribute @code{Exec_Dir} not specified)
10598 and in @file{^/common^[COMMON]^} for the @code{Release} project.
10600 @node Project File Packages
10601 @unnumberedsubsubsec Project File Packages
10604 A GNAT tool that is integrated with the Project Manager is modeled by a
10605 corresponding package in the project file. In the example above,
10606 The @code{Debug} project defines the packages @code{Builder}
10607 (for @command{gnatmake}) and @code{Compiler};
10608 the @code{Release} project defines only the @code{Compiler} package.
10610 The Ada-like package syntax is not to be taken literally. Although packages in
10611 project files bear a surface resemblance to packages in Ada source code, the
10612 notation is simply a way to convey a grouping of properties for a named
10613 entity. Indeed, the package names permitted in project files are restricted
10614 to a predefined set, corresponding to the project-aware tools, and the contents
10615 of packages are limited to a small set of constructs.
10616 The packages in the example above contain attribute definitions.
10618 @node Specifying ^Switch^Switch^ Settings
10619 @unnumberedsubsubsec Specifying ^Switch^Switch^ Settings
10622 ^Switch^Switch^ settings for a project-aware tool can be specified through
10623 attributes in the package that corresponds to the tool.
10624 The example above illustrates one of the relevant attributes,
10625 @code{^Default_Switches^Default_Switches^}, which is defined in packages
10626 in both project files.
10627 Unlike simple attributes like @code{Source_Dirs},
10628 @code{^Default_Switches^Default_Switches^} is
10629 known as an @emph{associative array}. When you define this attribute, you must
10630 supply an ``index'' (a literal string), and the effect of the attribute
10631 definition is to set the value of the array at the specified index.
10632 For the @code{^Default_Switches^Default_Switches^} attribute,
10633 the index is a programming language (in our case, Ada),
10634 and the value specified (after @code{use}) must be a list
10635 of string expressions.
10637 The attributes permitted in project files are restricted to a predefined set.
10638 Some may appear at project level, others in packages.
10639 For any attribute that is an associative array, the index must always be a
10640 literal string, but the restrictions on this string (e.g., a file name or a
10641 language name) depend on the individual attribute.
10642 Also depending on the attribute, its specified value will need to be either a
10643 string or a string list.
10645 In the @code{Debug} project, we set the switches for two tools,
10646 @command{gnatmake} and the compiler, and thus we include the two corresponding
10647 packages; each package defines the @code{^Default_Switches^Default_Switches^}
10648 attribute with index @code{"Ada"}.
10649 Note that the package corresponding to
10650 @command{gnatmake} is named @code{Builder}. The @code{Release} project is
10651 similar, but only includes the @code{Compiler} package.
10653 In project @code{Debug} above, the ^switches^switches^ starting with
10654 @option{-gnat} that are specified in package @code{Compiler}
10655 could have been placed in package @code{Builder}, since @command{gnatmake}
10656 transmits all such ^switches^switches^ to the compiler.
10658 @node Main Subprograms
10659 @unnumberedsubsubsec Main Subprograms
10662 One of the specifiable properties of a project is a list of files that contain
10663 main subprograms. This property is captured in the @code{Main} attribute,
10664 whose value is a list of strings. If a project defines the @code{Main}
10665 attribute, it is not necessary to identify the main subprogram(s) when
10666 invoking @command{gnatmake} (see @ref{gnatmake and Project Files}).
10668 @node Executable File Names
10669 @unnumberedsubsubsec Executable File Names
10672 By default, the executable file name corresponding to a main source is
10673 deducted from the main source file name. Through the attributes
10674 @code{Executable} and @code{Executable_Suffix} of package @code{Builder},
10675 it is possible to change this default.
10676 In project @code{Debug} above, the executable file name
10677 for main source @file{^proc.adb^PROC.ADB^} is
10678 @file{^proc1^PROC1.EXE^}.
10679 Attribute @code{Executable_Suffix}, when specified, may change the suffix
10680 of the the executable files, when no attribute @code{Executable} applies:
10681 its value replace the platform-specific executable suffix.
10682 Attributes @code{Executable} and @code{Executable_Suffix} are the only ways to
10683 specify a non default executable file name when several mains are built at once
10684 in a single @command{gnatmake} command.
10686 @node Source File Naming Conventions
10687 @unnumberedsubsubsec Source File Naming Conventions
10690 Since the project files above do not specify any source file naming
10691 conventions, the GNAT defaults are used. The mechanism for defining source
10692 file naming conventions -- a package named @code{Naming} --
10693 is described below (@pxref{Naming Schemes}).
10695 @node Source Language(s)
10696 @unnumberedsubsubsec Source Language(s)
10699 Since the project files do not specify a @code{Languages} attribute, by
10700 default the GNAT tools assume that the language of the project file is Ada.
10701 More generally, a project can comprise source files
10702 in Ada, C, and/or other languages.
10704 @node Using External Variables
10705 @subsection Using External Variables
10708 Instead of supplying different project files for debug and release, we can
10709 define a single project file that queries an external variable (set either
10710 on the command line or via an ^environment variable^logical name^) in order to
10711 conditionally define the appropriate settings. Again, assume that the
10712 source files @file{pack.ads}, @file{pack.adb}, and @file{proc.adb} are
10713 located in directory @file{^/common^[COMMON]^}. The following project file,
10714 @file{build.gpr}, queries the external variable named @code{STYLE} and
10715 defines an object directory and ^switch^switch^ settings based on whether
10716 the value is @code{"deb"} (debug) or @code{"rel"} (release), and where
10717 the default is @code{"deb"}.
10719 @smallexample @c projectfile
10722 for Main use ("proc");
10724 type Style_Type is ("deb", "rel");
10725 Style : Style_Type := external ("STYLE", "deb");
10729 for Object_Dir use "debug";
10732 for Object_Dir use "release";
10733 for Exec_Dir use ".";
10742 for ^Default_Switches^Default_Switches^ ("Ada")
10744 for Executable ("proc") use "proc1";
10751 package Compiler is
10755 for ^Default_Switches^Default_Switches^ ("Ada")
10756 use ("^-gnata^-gnata^",
10758 "^-gnatE^-gnatE^");
10761 for ^Default_Switches^Default_Switches^ ("Ada")
10772 @code{Style_Type} is an example of a @emph{string type}, which is the project
10773 file analog of an Ada enumeration type but whose components are string literals
10774 rather than identifiers. @code{Style} is declared as a variable of this type.
10776 The form @code{external("STYLE", "deb")} is known as an
10777 @emph{external reference}; its first argument is the name of an
10778 @emph{external variable}, and the second argument is a default value to be
10779 used if the external variable doesn't exist. You can define an external
10780 variable on the command line via the @option{^-X^/EXTERNAL_REFERENCE^} switch,
10781 or you can use ^an environment variable^a logical name^
10782 as an external variable.
10784 Each @code{case} construct is expanded by the Project Manager based on the
10785 value of @code{Style}. Thus the command
10788 gnatmake -P/common/build.gpr -XSTYLE=deb
10794 gnatmake /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=deb
10799 is equivalent to the @command{gnatmake} invocation using the project file
10800 @file{debug.gpr} in the earlier example. So is the command
10802 gnatmake ^-P/common/build.gpr^/PROJECT_FILE=[COMMON]BUILD.GPR^
10806 since @code{"deb"} is the default for @code{STYLE}.
10812 gnatmake -P/common/build.gpr -XSTYLE=rel
10818 GNAT MAKE /PROJECT_FILE=[COMMON]BUILD.GPR /EXTERNAL_REFERENCE=STYLE=rel
10823 is equivalent to the @command{gnatmake} invocation using the project file
10824 @file{release.gpr} in the earlier example.
10826 @node Importing Other Projects
10827 @subsection Importing Other Projects
10830 A compilation unit in a source file in one project may depend on compilation
10831 units in source files in other projects. To compile this unit under
10832 control of a project file, the
10833 dependent project must @emph{import} the projects containing the needed source
10835 This effect is obtained using syntax similar to an Ada @code{with} clause,
10836 but where @code{with}ed entities are strings that denote project files.
10838 As an example, suppose that the two projects @code{GUI_Proj} and
10839 @code{Comm_Proj} are defined in the project files @file{gui_proj.gpr} and
10840 @file{comm_proj.gpr} in directories @file{^/gui^[GUI]^}
10841 and @file{^/comm^[COMM]^}, respectively.
10842 Suppose that the source files for @code{GUI_Proj} are
10843 @file{gui.ads} and @file{gui.adb}, and that the source files for
10844 @code{Comm_Proj} are @file{comm.ads} and @file{comm.adb}, where each set of
10845 files is located in its respective project file directory. Schematically:
10864 We want to develop an application in directory @file{^/app^[APP]^} that
10865 @code{with} the packages @code{GUI} and @code{Comm}, using the properties of
10866 the corresponding project files (e.g. the ^switch^switch^ settings
10867 and object directory).
10868 Skeletal code for a main procedure might be something like the following:
10870 @smallexample @c ada
10873 procedure App_Main is
10882 Here is a project file, @file{app_proj.gpr}, that achieves the desired
10885 @smallexample @c projectfile
10887 with "/gui/gui_proj", "/comm/comm_proj";
10888 project App_Proj is
10889 for Main use ("app_main");
10895 Building an executable is achieved through the command:
10897 gnatmake ^-P/app/app_proj^/PROJECT_FILE=[APP]APP_PROJ^
10900 which will generate the @code{^app_main^APP_MAIN.EXE^} executable
10901 in the directory where @file{app_proj.gpr} resides.
10903 If an imported project file uses the standard extension (@code{^gpr^GPR^}) then
10904 (as illustrated above) the @code{with} clause can omit the extension.
10906 Our example specified an absolute path for each imported project file.
10907 Alternatively, the directory name of an imported object can be omitted
10911 The imported project file is in the same directory as the importing project
10914 You have defined ^an environment variable^a logical name^
10915 that includes the directory containing
10916 the needed project file. The syntax of @code{ADA_PROJECT_PATH} is the same as
10917 the syntax of @code{ADA_INCLUDE_PATH} and @code{ADA_OBJECTS_PATH}: a list of
10918 directory names separated by colons (semicolons on Windows).
10922 Thus, if we define @code{ADA_PROJECT_PATH} to include @file{^/gui^[GUI]^} and
10923 @file{^/comm^[COMM]^}, then our project file @file{app_proj.gpr} can be written
10926 @smallexample @c projectfile
10928 with "gui_proj", "comm_proj";
10929 project App_Proj is
10930 for Main use ("app_main");
10936 Importing other projects can create ambiguities.
10937 For example, the same unit might be present in different imported projects, or
10938 it might be present in both the importing project and in an imported project.
10939 Both of these conditions are errors. Note that in the current version of
10940 the Project Manager, it is illegal to have an ambiguous unit even if the
10941 unit is never referenced by the importing project. This restriction may be
10942 relaxed in a future release.
10944 @node Extending a Project
10945 @subsection Extending a Project
10948 In large software systems it is common to have multiple
10949 implementations of a common interface; in Ada terms, multiple versions of a
10950 package body for the same specification. For example, one implementation
10951 might be safe for use in tasking programs, while another might only be used
10952 in sequential applications. This can be modeled in GNAT using the concept
10953 of @emph{project extension}. If one project (the ``child'') @emph{extends}
10954 another project (the ``parent'') then by default all source files of the
10955 parent project are inherited by the child, but the child project can
10956 override any of the parent's source files with new versions, and can also
10957 add new files. This facility is the project analog of a type extension in
10958 Object-Oriented Programming. Project hierarchies are permitted (a child
10959 project may be the parent of yet another project), and a project that
10960 inherits one project can also import other projects.
10962 As an example, suppose that directory @file{^/seq^[SEQ]^} contains the project
10963 file @file{seq_proj.gpr} as well as the source files @file{pack.ads},
10964 @file{pack.adb}, and @file{proc.adb}:
10977 Note that the project file can simply be empty (that is, no attribute or
10978 package is defined):
10980 @smallexample @c projectfile
10982 project Seq_Proj is
10988 implying that its source files are all the Ada source files in the project
10991 Suppose we want to supply an alternate version of @file{pack.adb}, in
10992 directory @file{^/tasking^[TASKING]^}, but use the existing versions of
10993 @file{pack.ads} and @file{proc.adb}. We can define a project
10994 @code{Tasking_Proj} that inherits @code{Seq_Proj}:
10998 ^/tasking^[TASKING]^
11004 project Tasking_Proj extends "/seq/seq_proj" is
11010 The version of @file{pack.adb} used in a build depends on which project file
11013 Note that we could have obtained the desired behavior using project import
11014 rather than project inheritance; a @code{base} project would contain the
11015 sources for @file{pack.ads} and @file{proc.adb}, a sequential project would
11016 import @code{base} and add @file{pack.adb}, and likewise a tasking project
11017 would import @code{base} and add a different version of @file{pack.adb}. The
11018 choice depends on whether other sources in the original project need to be
11019 overridden. If they do, then project extension is necessary, otherwise,
11020 importing is sufficient.
11023 In a project file that extends another project file, it is possible to
11024 indicate that an inherited source is not part of the sources of the extending
11025 project. This is necessary sometimes when a package spec has been overloaded
11026 and no longer requires a body: in this case, it is necessary to indicate that
11027 the inherited body is not part of the sources of the project, otherwise there
11028 will be a compilation error when compiling the spec.
11030 For that purpose, the attribute @code{Locally_Removed_Files} is used.
11031 Its value is a string list: a list of file names.
11033 @smallexample @c @projectfile
11034 project B extends "a" is
11035 for Source_Files use ("pkg.ads");
11036 -- New spec of Pkg does not need a completion
11037 for Locally_Removed_Files use ("pkg.adb");
11041 Attribute @code{Locally_Removed_Files} may also be used to check if a source
11042 is still needed: if it is possible to build using @code{gnatmake} when such
11043 a source is put in attribute @code{Locally_Removed_Files} of a project P, then
11044 it is possible to remove the source completely from a system that includes
11047 @c ***********************
11048 @c * Project File Syntax *
11049 @c ***********************
11051 @node Project File Syntax
11052 @section Project File Syntax
11061 * Associative Array Attributes::
11062 * case Constructions::
11066 This section describes the structure of project files.
11068 A project may be an @emph{independent project}, entirely defined by a single
11069 project file. Any Ada source file in an independent project depends only
11070 on the predefined library and other Ada source files in the same project.
11073 A project may also @dfn{depend on} other projects, in either or both of
11074 the following ways:
11076 @item It may import any number of projects
11077 @item It may extend at most one other project
11081 The dependence relation is a directed acyclic graph (the subgraph reflecting
11082 the ``extends'' relation is a tree).
11084 A project's @dfn{immediate sources} are the source files directly defined by
11085 that project, either implicitly by residing in the project file's directory,
11086 or explicitly through any of the source-related attributes described below.
11087 More generally, a project @var{proj}'s @dfn{sources} are the immediate sources
11088 of @var{proj} together with the immediate sources (unless overridden) of any
11089 project on which @var{proj} depends (either directly or indirectly).
11092 @subsection Basic Syntax
11095 As seen in the earlier examples, project files have an Ada-like syntax.
11096 The minimal project file is:
11097 @smallexample @c projectfile
11106 The identifier @code{Empty} is the name of the project.
11107 This project name must be present after the reserved
11108 word @code{end} at the end of the project file, followed by a semi-colon.
11110 Any name in a project file, such as the project name or a variable name,
11111 has the same syntax as an Ada identifier.
11113 The reserved words of project files are the Ada reserved words plus
11114 @code{extends}, @code{external}, and @code{project}. Note that the only Ada
11115 reserved words currently used in project file syntax are:
11143 Comments in project files have the same syntax as in Ada, two consecutives
11144 hyphens through the end of the line.
11147 @subsection Packages
11150 A project file may contain @emph{packages}. The name of a package must be one
11151 of the identifiers from the following list. A package
11152 with a given name may only appear once in a project file. Package names are
11153 case insensitive. The following package names are legal:
11169 @code{Cross_Reference}
11181 In its simplest form, a package may be empty:
11183 @smallexample @c projectfile
11193 A package may contain @emph{attribute declarations},
11194 @emph{variable declarations} and @emph{case constructions}, as will be
11197 When there is ambiguity between a project name and a package name,
11198 the name always designates the project. To avoid possible confusion, it is
11199 always a good idea to avoid naming a project with one of the
11200 names allowed for packages or any name that starts with @code{gnat}.
11203 @subsection Expressions
11206 An @emph{expression} is either a @emph{string expression} or a
11207 @emph{string list expression}.
11209 A @emph{string expression} is either a @emph{simple string expression} or a
11210 @emph{compound string expression}.
11212 A @emph{simple string expression} is one of the following:
11214 @item A literal string; e.g.@code{"comm/my_proj.gpr"}
11215 @item A string-valued variable reference (see @ref{Variables})
11216 @item A string-valued attribute reference (see @ref{Attributes})
11217 @item An external reference (see @ref{External References in Project Files})
11221 A @emph{compound string expression} is a concatenation of string expressions,
11222 using the operator @code{"&"}
11224 Path & "/" & File_Name & ".ads"
11228 A @emph{string list expression} is either a
11229 @emph{simple string list expression} or a
11230 @emph{compound string list expression}.
11232 A @emph{simple string list expression} is one of the following:
11234 @item A parenthesized list of zero or more string expressions,
11235 separated by commas
11237 File_Names := (File_Name, "gnat.adc", File_Name & ".orig");
11240 @item A string list-valued variable reference
11241 @item A string list-valued attribute reference
11245 A @emph{compound string list expression} is the concatenation (using
11246 @code{"&"}) of a simple string list expression and an expression. Note that
11247 each term in a compound string list expression, except the first, may be
11248 either a string expression or a string list expression.
11250 @smallexample @c projectfile
11252 File_Name_List := () & File_Name; -- One string in this list
11253 Extended_File_Name_List := File_Name_List & (File_Name & ".orig");
11255 Big_List := File_Name_List & Extended_File_Name_List;
11256 -- Concatenation of two string lists: three strings
11257 Illegal_List := "gnat.adc" & Extended_File_Name_List;
11258 -- Illegal: must start with a string list
11263 @subsection String Types
11266 A @emph{string type declaration} introduces a discrete set of string literals.
11267 If a string variable is declared to have this type, its value
11268 is restricted to the given set of literals.
11270 Here is an example of a string type declaration:
11272 @smallexample @c projectfile
11273 type OS is ("NT", "nt", "Unix", "GNU/Linux", "other OS");
11277 Variables of a string type are called @emph{typed variables}; all other
11278 variables are called @emph{untyped variables}. Typed variables are
11279 particularly useful in @code{case} constructions, to support conditional
11280 attribute declarations.
11281 (see @ref{case Constructions}).
11283 The string literals in the list are case sensitive and must all be different.
11284 They may include any graphic characters allowed in Ada, including spaces.
11286 A string type may only be declared at the project level, not inside a package.
11288 A string type may be referenced by its name if it has been declared in the same
11289 project file, or by an expanded name whose prefix is the name of the project
11290 in which it is declared.
11293 @subsection Variables
11296 A variable may be declared at the project file level, or within a package.
11297 Here are some examples of variable declarations:
11299 @smallexample @c projectfile
11301 This_OS : OS := external ("OS"); -- a typed variable declaration
11302 That_OS := "GNU/Linux"; -- an untyped variable declaration
11307 The syntax of a @emph{typed variable declaration} is identical to the Ada
11308 syntax for an object declaration. By contrast, the syntax of an untyped
11309 variable declaration is identical to an Ada assignment statement. In fact,
11310 variable declarations in project files have some of the characteristics of
11311 an assignment, in that successive declarations for the same variable are
11312 allowed. Untyped variable declarations do establish the expected kind of the
11313 variable (string or string list), and successive declarations for it must
11314 respect the initial kind.
11317 A string variable declaration (typed or untyped) declares a variable
11318 whose value is a string. This variable may be used as a string expression.
11319 @smallexample @c projectfile
11320 File_Name := "readme.txt";
11321 Saved_File_Name := File_Name & ".saved";
11325 A string list variable declaration declares a variable whose value is a list
11326 of strings. The list may contain any number (zero or more) of strings.
11328 @smallexample @c projectfile
11330 List_With_One_Element := ("^-gnaty^-gnaty^");
11331 List_With_Two_Elements := List_With_One_Element & "^-gnatg^-gnatg^";
11332 Long_List := ("main.ada", "pack1_.ada", "pack1.ada", "pack2_.ada"
11333 "pack2.ada", "util_.ada", "util.ada");
11337 The same typed variable may not be declared more than once at project level,
11338 and it may not be declared more than once in any package; it is in effect
11341 The same untyped variable may be declared several times. Declarations are
11342 elaborated in the order in which they appear, so the new value replaces
11343 the old one, and any subsequent reference to the variable uses the new value.
11344 However, as noted above, if a variable has been declared as a string, all
11346 declarations must give it a string value. Similarly, if a variable has
11347 been declared as a string list, all subsequent declarations
11348 must give it a string list value.
11350 A @emph{variable reference} may take several forms:
11353 @item The simple variable name, for a variable in the current package (if any)
11354 or in the current project
11355 @item An expanded name, whose prefix is a context name.
11359 A @emph{context} may be one of the following:
11362 @item The name of an existing package in the current project
11363 @item The name of an imported project of the current project
11364 @item The name of an ancestor project (i.e., a project extended by the current
11365 project, either directly or indirectly)
11366 @item An expanded name whose prefix is an imported/parent project name, and
11367 whose selector is a package name in that project.
11371 A variable reference may be used in an expression.
11374 @subsection Attributes
11377 A project (and its packages) may have @emph{attributes} that define
11378 the project's properties. Some attributes have values that are strings;
11379 others have values that are string lists.
11381 There are two categories of attributes: @emph{simple attributes}
11382 and @emph{associative arrays} (see @ref{Associative Array Attributes}).
11384 Legal project attribute names, and attribute names for each legal package are
11385 listed below. Attributes names are case-insensitive.
11387 The following attributes are defined on projects (all are simple attributes):
11389 @multitable @columnfractions .4 .3
11390 @item @emph{Attribute Name}
11392 @item @code{Source_Files}
11394 @item @code{Source_Dirs}
11396 @item @code{Source_List_File}
11398 @item @code{Object_Dir}
11400 @item @code{Exec_Dir}
11402 @item @code{Locally_Removed_Files}
11406 @item @code{Languages}
11408 @item @code{Main_Language}
11410 @item @code{Library_Dir}
11412 @item @code{Library_Name}
11414 @item @code{Library_Kind}
11416 @item @code{Library_Version}
11418 @item @code{Library_Interface}
11420 @item @code{Library_Auto_Init}
11422 @item @code{Library_Options}
11424 @item @code{Library_GCC}
11429 The following attributes are defined for package @code{Naming}
11430 (see @ref{Naming Schemes}):
11432 @multitable @columnfractions .4 .2 .2 .2
11433 @item Attribute Name @tab Category @tab Index @tab Value
11434 @item @code{Spec_Suffix}
11435 @tab associative array
11438 @item @code{Body_Suffix}
11439 @tab associative array
11442 @item @code{Separate_Suffix}
11443 @tab simple attribute
11446 @item @code{Casing}
11447 @tab simple attribute
11450 @item @code{Dot_Replacement}
11451 @tab simple attribute
11455 @tab associative array
11459 @tab associative array
11462 @item @code{Specification_Exceptions}
11463 @tab associative array
11466 @item @code{Implementation_Exceptions}
11467 @tab associative array
11473 The following attributes are defined for packages @code{Builder},
11474 @code{Compiler}, @code{Binder},
11475 @code{Linker}, @code{Cross_Reference}, and @code{Finder}
11476 (see @ref{^Switches^Switches^ and Project Files}).
11478 @multitable @columnfractions .4 .2 .2 .2
11479 @item Attribute Name @tab Category @tab Index @tab Value
11480 @item @code{^Default_Switches^Default_Switches^}
11481 @tab associative array
11484 @item @code{^Switches^Switches^}
11485 @tab associative array
11491 In addition, package @code{Compiler} has a single string attribute
11492 @code{Local_Configuration_Pragmas} and package @code{Builder} has a single
11493 string attribute @code{Global_Configuration_Pragmas}.
11496 Each simple attribute has a default value: the empty string (for string-valued
11497 attributes) and the empty list (for string list-valued attributes).
11499 An attribute declaration defines a new value for an attribute.
11501 Examples of simple attribute declarations:
11503 @smallexample @c projectfile
11504 for Object_Dir use "objects";
11505 for Source_Dirs use ("units", "test/drivers");
11509 The syntax of a @dfn{simple attribute declaration} is similar to that of an
11510 attribute definition clause in Ada.
11512 Attributes references may be appear in expressions.
11513 The general form for such a reference is @code{<entity>'<attribute>}:
11514 Associative array attributes are functions. Associative
11515 array attribute references must have an argument that is a string literal.
11519 @smallexample @c projectfile
11521 Naming'Dot_Replacement
11522 Imported_Project'Source_Dirs
11523 Imported_Project.Naming'Casing
11524 Builder'^Default_Switches^Default_Switches^("Ada")
11528 The prefix of an attribute may be:
11530 @item @code{project} for an attribute of the current project
11531 @item The name of an existing package of the current project
11532 @item The name of an imported project
11533 @item The name of a parent project that is extended by the current project
11534 @item An expanded name whose prefix is imported/parent project name,
11535 and whose selector is a package name
11540 @smallexample @c projectfile
11543 for Source_Dirs use project'Source_Dirs & "units";
11544 for Source_Dirs use project'Source_Dirs & "test/drivers"
11550 In the first attribute declaration, initially the attribute @code{Source_Dirs}
11551 has the default value: an empty string list. After this declaration,
11552 @code{Source_Dirs} is a string list of one element: @code{"units"}.
11553 After the second attribute declaration @code{Source_Dirs} is a string list of
11554 two elements: @code{"units"} and @code{"test/drivers"}.
11556 Note: this example is for illustration only. In practice,
11557 the project file would contain only one attribute declaration:
11559 @smallexample @c projectfile
11560 for Source_Dirs use ("units", "test/drivers");
11563 @node Associative Array Attributes
11564 @subsection Associative Array Attributes
11567 Some attributes are defined as @emph{associative arrays}. An associative
11568 array may be regarded as a function that takes a string as a parameter
11569 and delivers a string or string list value as its result.
11571 Here are some examples of single associative array attribute associations:
11573 @smallexample @c projectfile
11574 for Body ("main") use "Main.ada";
11575 for ^Switches^Switches^ ("main.ada")
11577 "^-gnatv^-gnatv^");
11578 for ^Switches^Switches^ ("main.ada")
11579 use Builder'^Switches^Switches^ ("main.ada")
11584 Like untyped variables and simple attributes, associative array attributes
11585 may be declared several times. Each declaration supplies a new value for the
11586 attribute, and replaces the previous setting.
11589 An associative array attribute may be declared as a full associative array
11590 declaration, with the value of the same attribute in an imported or extended
11593 @smallexample @c projectfile
11595 for Default_Switches use Default.Builder'Default_Switches;
11600 In this example, @code{Default} must be either an project imported by the
11601 current project, or the project that the current project extends. If the
11602 attribute is in a package (in this case, in package @code{Builder}), the same
11603 package needs to be specified.
11606 A full associative array declaration replaces any other declaration for the
11607 attribute, including other full associative array declaration. Single
11608 associative array associations may be declare after a full associative
11609 declaration, modifying the value for a single association of the attribute.
11611 @node case Constructions
11612 @subsection @code{case} Constructions
11615 A @code{case} construction is used in a project file to effect conditional
11617 Here is a typical example:
11619 @smallexample @c projectfile
11622 type OS_Type is ("GNU/Linux", "Unix", "NT", "VMS");
11624 OS : OS_Type := external ("OS", "GNU/Linux");
11628 package Compiler is
11630 when "GNU/Linux" | "Unix" =>
11631 for ^Default_Switches^Default_Switches^ ("Ada")
11632 use ("^-gnath^-gnath^");
11634 for ^Default_Switches^Default_Switches^ ("Ada")
11635 use ("^-gnatP^-gnatP^");
11644 The syntax of a @code{case} construction is based on the Ada case statement
11645 (although there is no @code{null} construction for empty alternatives).
11647 The case expression must a typed string variable.
11648 Each alternative comprises the reserved word @code{when}, either a list of
11649 literal strings separated by the @code{"|"} character or the reserved word
11650 @code{others}, and the @code{"=>"} token.
11651 Each literal string must belong to the string type that is the type of the
11653 An @code{others} alternative, if present, must occur last.
11655 After each @code{=>}, there are zero or more constructions. The only
11656 constructions allowed in a case construction are other case constructions and
11657 attribute declarations. String type declarations, variable declarations and
11658 package declarations are not allowed.
11660 The value of the case variable is often given by an external reference
11661 (see @ref{External References in Project Files}).
11663 @c ****************************************
11664 @c * Objects and Sources in Project Files *
11665 @c ****************************************
11667 @node Objects and Sources in Project Files
11668 @section Objects and Sources in Project Files
11671 * Object Directory::
11673 * Source Directories::
11674 * Source File Names::
11678 Each project has exactly one object directory and one or more source
11679 directories. The source directories must contain at least one source file,
11680 unless the project file explicitly specifies that no source files are present
11681 (see @ref{Source File Names}).
11683 @node Object Directory
11684 @subsection Object Directory
11687 The object directory for a project is the directory containing the compiler's
11688 output (such as @file{ALI} files and object files) for the project's immediate
11691 The object directory is given by the value of the attribute @code{Object_Dir}
11692 in the project file.
11694 @smallexample @c projectfile
11695 for Object_Dir use "objects";
11699 The attribute @var{Object_Dir} has a string value, the path name of the object
11700 directory. The path name may be absolute or relative to the directory of the
11701 project file. This directory must already exist, and be readable and writable.
11703 By default, when the attribute @code{Object_Dir} is not given an explicit value
11704 or when its value is the empty string, the object directory is the same as the
11705 directory containing the project file.
11707 @node Exec Directory
11708 @subsection Exec Directory
11711 The exec directory for a project is the directory containing the executables
11712 for the project's main subprograms.
11714 The exec directory is given by the value of the attribute @code{Exec_Dir}
11715 in the project file.
11717 @smallexample @c projectfile
11718 for Exec_Dir use "executables";
11722 The attribute @var{Exec_Dir} has a string value, the path name of the exec
11723 directory. The path name may be absolute or relative to the directory of the
11724 project file. This directory must already exist, and be writable.
11726 By default, when the attribute @code{Exec_Dir} is not given an explicit value
11727 or when its value is the empty string, the exec directory is the same as the
11728 object directory of the project file.
11730 @node Source Directories
11731 @subsection Source Directories
11734 The source directories of a project are specified by the project file
11735 attribute @code{Source_Dirs}.
11737 This attribute's value is a string list. If the attribute is not given an
11738 explicit value, then there is only one source directory, the one where the
11739 project file resides.
11741 A @code{Source_Dirs} attribute that is explicitly defined to be the empty list,
11744 @smallexample @c projectfile
11745 for Source_Dirs use ();
11749 indicates that the project contains no source files.
11751 Otherwise, each string in the string list designates one or more
11752 source directories.
11754 @smallexample @c projectfile
11755 for Source_Dirs use ("sources", "test/drivers");
11759 If a string in the list ends with @code{"/**"}, then the directory whose path
11760 name precedes the two asterisks, as well as all its subdirectories
11761 (recursively), are source directories.
11763 @smallexample @c projectfile
11764 for Source_Dirs use ("/system/sources/**");
11768 Here the directory @code{/system/sources} and all of its subdirectories
11769 (recursively) are source directories.
11771 To specify that the source directories are the directory of the project file
11772 and all of its subdirectories, you can declare @code{Source_Dirs} as follows:
11773 @smallexample @c projectfile
11774 for Source_Dirs use ("./**");
11778 Each of the source directories must exist and be readable.
11780 @node Source File Names
11781 @subsection Source File Names
11784 In a project that contains source files, their names may be specified by the
11785 attributes @code{Source_Files} (a string list) or @code{Source_List_File}
11786 (a string). Source file names never include any directory information.
11788 If the attribute @code{Source_Files} is given an explicit value, then each
11789 element of the list is a source file name.
11791 @smallexample @c projectfile
11792 for Source_Files use ("main.adb");
11793 for Source_Files use ("main.adb", "pack1.ads", "pack2.adb");
11797 If the attribute @code{Source_Files} is not given an explicit value,
11798 but the attribute @code{Source_List_File} is given a string value,
11799 then the source file names are contained in the text file whose path name
11800 (absolute or relative to the directory of the project file) is the
11801 value of the attribute @code{Source_List_File}.
11803 Each line in the file that is not empty or is not a comment
11804 contains a source file name.
11806 @smallexample @c projectfile
11807 for Source_List_File use "source_list.txt";
11811 By default, if neither the attribute @code{Source_Files} nor the attribute
11812 @code{Source_List_File} is given an explicit value, then each file in the
11813 source directories that conforms to the project's naming scheme
11814 (see @ref{Naming Schemes}) is an immediate source of the project.
11816 A warning is issued if both attributes @code{Source_Files} and
11817 @code{Source_List_File} are given explicit values. In this case, the attribute
11818 @code{Source_Files} prevails.
11820 Each source file name must be the name of one existing source file
11821 in one of the source directories.
11823 A @code{Source_Files} attribute whose value is an empty list
11824 indicates that there are no source files in the project.
11826 If the order of the source directories is known statically, that is if
11827 @code{"/**"} is not used in the string list @code{Source_Dirs}, then there may
11828 be several files with the same source file name. In this case, only the file
11829 in the first directory is considered as an immediate source of the project
11830 file. If the order of the source directories is not known statically, it is
11831 an error to have several files with the same source file name.
11833 Projects can be specified to have no Ada source
11834 files: the value of (@code{Source_Dirs} or @code{Source_Files} may be an empty
11835 list, or the @code{"Ada"} may be absent from @code{Languages}:
11837 @smallexample @c projectfile
11838 for Source_Dirs use ();
11839 for Source_Files use ();
11840 for Languages use ("C", "C++");
11844 Otherwise, a project must contain at least one immediate source.
11846 Projects with no source files are useful as template packages
11847 (see @ref{Packages in Project Files}) for other projects; in particular to
11848 define a package @code{Naming} (see @ref{Naming Schemes}).
11850 @c ****************************
11851 @c * Importing Projects *
11852 @c ****************************
11854 @node Importing Projects
11855 @section Importing Projects
11858 An immediate source of a project P may depend on source files that
11859 are neither immediate sources of P nor in the predefined library.
11860 To get this effect, P must @emph{import} the projects that contain the needed
11863 @smallexample @c projectfile
11865 with "project1", "utilities.gpr";
11866 with "/namings/apex.gpr";
11873 As can be seen in this example, the syntax for importing projects is similar
11874 to the syntax for importing compilation units in Ada. However, project files
11875 use literal strings instead of names, and the @code{with} clause identifies
11876 project files rather than packages.
11878 Each literal string is the file name or path name (absolute or relative) of a
11879 project file. If a string is simply a file name, with no path, then its
11880 location is determined by the @emph{project path}:
11884 If the ^environment variable^logical name^ @env{ADA_PROJECT_PATH} exists,
11885 then the project path includes all the directories in this
11886 ^environment variable^logical name^, plus the directory of the project file.
11889 If the ^environment variable^logical name^ @env{ADA_PROJECT_PATH} does not
11890 exist, then the project path contains only one directory, namely the one where
11891 the project file is located.
11895 If a relative pathname is used, as in
11897 @smallexample @c projectfile
11902 then the path is relative to the directory where the importing project file is
11903 located. Any symbolic link will be fully resolved in the directory
11904 of the importing project file before the imported project file is examined.
11906 If the @code{with}'ed project file name does not have an extension,
11907 the default is @file{^.gpr^.GPR^}. If a file with this extension is not found,
11908 then the file name as specified in the @code{with} clause (no extension) will
11909 be used. In the above example, if a file @code{project1.gpr} is found, then it
11910 will be used; otherwise, if a file @code{^project1^PROJECT1^} exists
11911 then it will be used; if neither file exists, this is an error.
11913 A warning is issued if the name of the project file does not match the
11914 name of the project; this check is case insensitive.
11916 Any source file that is an immediate source of the imported project can be
11917 used by the immediate sources of the importing project, transitively. Thus
11918 if @code{A} imports @code{B}, and @code{B} imports @code{C}, the immediate
11919 sources of @code{A} may depend on the immediate sources of @code{C}, even if
11920 @code{A} does not import @code{C} explicitly. However, this is not recommended,
11921 because if and when @code{B} ceases to import @code{C}, some sources in
11922 @code{A} will no longer compile.
11924 A side effect of this capability is that normally cyclic dependencies are not
11925 permitted: if @code{A} imports @code{B} (directly or indirectly) then @code{B}
11926 is not allowed to import @code{A}. However, there are cases when cyclic
11927 dependencies would be beneficial. For these cases, another form of import
11928 between projects exists, the @code{limited with}: a project @code{A} that
11929 imports a project @code{B} with a straigh @code{with} may also be imported,
11930 directly or indirectly, by @code{B} on the condition that imports from @code{B}
11931 to @code{A} include at least one @code{limited with}.
11933 @smallexample @c 0projectfile
11939 limited with "../a/a.gpr";
11947 limited with "../a/a.gpr";
11953 In the above legal example, there are two project cycles:
11956 @item A -> C -> D -> A
11960 In each of these cycle there is one @code{limited with}: import of @code{A}
11961 from @code{B} and import of @code{A} from @code{D}.
11963 The difference between straight @code{with} and @code{limited with} is that
11964 the name of a project imported with a @code{limited with} cannot be used in the
11965 project that imports it. In particular, its packages cannot be renamed and
11966 its variables cannot be referred to.
11968 An exception to the above rules for @code{limited with} is that for the main
11969 project specified to @command{gnatmake} or to the @command{GNAT} driver a
11970 @code{limited with} is equivalent to a straight @code{with}. For example,
11971 in the example above, projects @code{B} and @code{D} could not be main
11972 projects for @command{gnatmake} or to the @command{GNAT} driver, because they
11973 each have a @code{limited with} that is the only one in a cycle of importing
11976 @c *********************
11977 @c * Project Extension *
11978 @c *********************
11980 @node Project Extension
11981 @section Project Extension
11984 During development of a large system, it is sometimes necessary to use
11985 modified versions of some of the source files, without changing the original
11986 sources. This can be achieved through the @emph{project extension} facility.
11988 @smallexample @c projectfile
11989 project Modified_Utilities extends "/baseline/utilities.gpr" is ...
11993 A project extension declaration introduces an extending project
11994 (the @emph{child}) and a project being extended (the @emph{parent}).
11996 By default, a child project inherits all the sources of its parent.
11997 However, inherited sources can be overridden: a unit in a parent is hidden
11998 by a unit of the same name in the child.
12000 Inherited sources are considered to be sources (but not immediate sources)
12001 of the child project; see @ref{Project File Syntax}.
12003 An inherited source file retains any switches specified in the parent project.
12005 For example if the project @code{Utilities} contains the specification and the
12006 body of an Ada package @code{Util_IO}, then the project
12007 @code{Modified_Utilities} can contain a new body for package @code{Util_IO}.
12008 The original body of @code{Util_IO} will not be considered in program builds.
12009 However, the package specification will still be found in the project
12012 A child project can have only one parent but it may import any number of other
12015 A project is not allowed to import directly or indirectly at the same time a
12016 child project and any of its ancestors.
12018 @c ****************************************
12019 @c * External References in Project Files *
12020 @c ****************************************
12022 @node External References in Project Files
12023 @section External References in Project Files
12026 A project file may contain references to external variables; such references
12027 are called @emph{external references}.
12029 An external variable is either defined as part of the environment (an
12030 environment variable in Unix, for example) or else specified on the command
12031 line via the @option{^-X^/EXTERNAL_REFERENCE=^@emph{vbl}=@emph{value}} switch.
12032 If both, then the command line value is used.
12034 The value of an external reference is obtained by means of the built-in
12035 function @code{external}, which returns a string value.
12036 This function has two forms:
12038 @item @code{external (external_variable_name)}
12039 @item @code{external (external_variable_name, default_value)}
12043 Each parameter must be a string literal. For example:
12045 @smallexample @c projectfile
12047 external ("OS", "GNU/Linux")
12051 In the form with one parameter, the function returns the value of
12052 the external variable given as parameter. If this name is not present in the
12053 environment, the function returns an empty string.
12055 In the form with two string parameters, the second argument is
12056 the value returned when the variable given as the first argument is not
12057 present in the environment. In the example above, if @code{"OS"} is not
12058 the name of ^an environment variable^a logical name^ and is not passed on
12059 the command line, then the returned value is @code{"GNU/Linux"}.
12061 An external reference may be part of a string expression or of a string
12062 list expression, and can therefore appear in a variable declaration or
12063 an attribute declaration.
12065 @smallexample @c projectfile
12067 type Mode_Type is ("Debug", "Release");
12068 Mode : Mode_Type := external ("MODE");
12075 @c *****************************
12076 @c * Packages in Project Files *
12077 @c *****************************
12079 @node Packages in Project Files
12080 @section Packages in Project Files
12083 A @emph{package} defines the settings for project-aware tools within a
12085 For each such tool one can declare a package; the names for these
12086 packages are preset (see @ref{Packages}).
12087 A package may contain variable declarations, attribute declarations, and case
12090 @smallexample @c projectfile
12093 package Builder is -- used by gnatmake
12094 for ^Default_Switches^Default_Switches^ ("Ada")
12103 The syntax of package declarations mimics that of package in Ada.
12105 Most of the packages have an attribute
12106 @code{^Default_Switches^Default_Switches^}.
12107 This attribute is an associative array, and its value is a string list.
12108 The index of the associative array is the name of a programming language (case
12109 insensitive). This attribute indicates the ^switch^switch^
12110 or ^switches^switches^ to be used
12111 with the corresponding tool.
12113 Some packages also have another attribute, @code{^Switches^Switches^},
12114 an associative array whose value is a string list.
12115 The index is the name of a source file.
12116 This attribute indicates the ^switch^switch^
12117 or ^switches^switches^ to be used by the corresponding
12118 tool when dealing with this specific file.
12120 Further information on these ^switch^switch^-related attributes is found in
12121 @ref{^Switches^Switches^ and Project Files}.
12123 A package may be declared as a @emph{renaming} of another package; e.g., from
12124 the project file for an imported project.
12126 @smallexample @c projectfile
12128 with "/global/apex.gpr";
12130 package Naming renames Apex.Naming;
12137 Packages that are renamed in other project files often come from project files
12138 that have no sources: they are just used as templates. Any modification in the
12139 template will be reflected automatically in all the project files that rename
12140 a package from the template.
12142 In addition to the tool-oriented packages, you can also declare a package
12143 named @code{Naming} to establish specialized source file naming conventions
12144 (see @ref{Naming Schemes}).
12146 @c ************************************
12147 @c * Variables from Imported Projects *
12148 @c ************************************
12150 @node Variables from Imported Projects
12151 @section Variables from Imported Projects
12154 An attribute or variable defined in an imported or parent project can
12155 be used in expressions in the importing / extending project.
12156 Such an attribute or variable is denoted by an expanded name whose prefix
12157 is either the name of the project or the expanded name of a package within
12160 @smallexample @c projectfile
12163 project Main extends "base" is
12164 Var1 := Imported.Var;
12165 Var2 := Base.Var & ".new";
12170 for ^Default_Switches^Default_Switches^ ("Ada")
12171 use Imported.Builder.Ada_^Switches^Switches^ &
12172 "^-gnatg^-gnatg^" &
12178 package Compiler is
12179 for ^Default_Switches^Default_Switches^ ("Ada")
12180 use Base.Compiler.Ada_^Switches^Switches^;
12191 The value of @code{Var1} is a copy of the variable @code{Var} defined
12192 in the project file @file{"imported.gpr"}
12194 the value of @code{Var2} is a copy of the value of variable @code{Var}
12195 defined in the project file @file{base.gpr}, concatenated with @code{".new"}
12197 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
12198 @code{Builder} is a string list that includes in its value a copy of the value
12199 of @code{Ada_^Switches^Switches^} defined in the @code{Builder} package
12200 in project file @file{imported.gpr} plus two new elements:
12201 @option{"^-gnatg^-gnatg^"}
12202 and @option{"^-v^-v^"};
12204 attribute @code{^Default_Switches^Default_Switches^ ("Ada")} in package
12205 @code{Compiler} is a copy of the variable @code{Ada_^Switches^Switches^}
12206 defined in the @code{Compiler} package in project file @file{base.gpr},
12207 the project being extended.
12210 @c ******************
12211 @c * Naming Schemes *
12212 @c ******************
12214 @node Naming Schemes
12215 @section Naming Schemes
12218 Sometimes an Ada software system is ported from a foreign compilation
12219 environment to GNAT, and the file names do not use the default GNAT
12220 conventions. Instead of changing all the file names (which for a variety
12221 of reasons might not be possible), you can define the relevant file
12222 naming scheme in the @code{Naming} package in your project file.
12225 Note that the use of pragmas described in @ref{Alternative
12226 File Naming Schemes} by mean of a configuration pragmas file is not
12227 supported when using project files. You must use the features described
12228 in this paragraph. You can however use specify other configuration
12229 pragmas (see @ref{Specifying Configuration Pragmas}).
12232 For example, the following
12233 package models the Apex file naming rules:
12235 @smallexample @c projectfile
12238 for Casing use "lowercase";
12239 for Dot_Replacement use ".";
12240 for Spec_Suffix ("Ada") use ".1.ada";
12241 for Body_Suffix ("Ada") use ".2.ada";
12248 For example, the following package models the DEC Ada file naming rules:
12250 @smallexample @c projectfile
12253 for Casing use "lowercase";
12254 for Dot_Replacement use "__";
12255 for Spec_Suffix ("Ada") use "_.^ada^ada^";
12256 for Body_Suffix ("Ada") use ".^ada^ada^";
12262 (Note that @code{Casing} is @code{"lowercase"} because GNAT gets the file
12263 names in lower case)
12267 You can define the following attributes in package @code{Naming}:
12272 This must be a string with one of the three values @code{"lowercase"},
12273 @code{"uppercase"} or @code{"mixedcase"}; these strings are case insensitive.
12276 If @var{Casing} is not specified, then the default is @code{"lowercase"}.
12278 @item @var{Dot_Replacement}
12279 This must be a string whose value satisfies the following conditions:
12282 @item It must not be empty
12283 @item It cannot start or end with an alphanumeric character
12284 @item It cannot be a single underscore
12285 @item It cannot start with an underscore followed by an alphanumeric
12286 @item It cannot contain a dot @code{'.'} except if the entire string
12291 If @code{Dot_Replacement} is not specified, then the default is @code{"-"}.
12293 @item @var{Spec_Suffix}
12294 This is an associative array (indexed by the programming language name, case
12295 insensitive) whose value is a string that must satisfy the following
12299 @item It must not be empty
12300 @item It must include at least one dot
12303 If @code{Spec_Suffix ("Ada")} is not specified, then the default is
12304 @code{"^.ads^.ADS^"}.
12306 @item @var{Body_Suffix}
12307 This is an associative array (indexed by the programming language name, case
12308 insensitive) whose value is a string that must satisfy the following
12312 @item It must not be empty
12313 @item It must include at least one dot
12314 @item It cannot end with the same string as @code{Spec_Suffix ("Ada")}
12317 If @code{Body_Suffix ("Ada")} is not specified, then the default is
12318 @code{"^.adb^.ADB^"}.
12320 @item @var{Separate_Suffix}
12321 This must be a string whose value satisfies the same conditions as
12322 @code{Body_Suffix}.
12325 If @code{Separate_Suffix ("Ada")} is not specified, then it defaults to same
12326 value as @code{Body_Suffix ("Ada")}.
12330 You can use the associative array attribute @code{Spec} to define
12331 the source file name for an individual Ada compilation unit's spec. The array
12332 index must be a string literal that identifies the Ada unit (case insensitive).
12333 The value of this attribute must be a string that identifies the file that
12334 contains this unit's spec (case sensitive or insensitive depending on the
12337 @smallexample @c projectfile
12338 for Spec ("MyPack.MyChild") use "mypack.mychild.spec";
12343 You can use the associative array attribute @code{Body} to
12344 define the source file name for an individual Ada compilation unit's body
12345 (possibly a subunit). The array index must be a string literal that identifies
12346 the Ada unit (case insensitive). The value of this attribute must be a string
12347 that identifies the file that contains this unit's body or subunit (case
12348 sensitive or insensitive depending on the operating system).
12350 @smallexample @c projectfile
12351 for Body ("MyPack.MyChild") use "mypack.mychild.body";
12355 @c ********************
12356 @c * Library Projects *
12357 @c ********************
12359 @node Library Projects
12360 @section Library Projects
12363 @emph{Library projects} are projects whose object code is placed in a library.
12364 (Note that this facility is not yet supported on all platforms)
12366 To create a library project, you need to define in its project file
12367 two project-level attributes: @code{Library_Name} and @code{Library_Dir}.
12368 Additionally, you may define the library-related attributes
12369 @code{Library_Kind}, @code{Library_Version}, @code{Library_Interface},
12370 @code{Library_Auto_Init}, @code{Library_Options} and @code{Library_GCC}.
12372 The @code{Library_Name} attribute has a string value. There is no restriction
12373 on the name of a library. It is the responsability of the developer to
12374 choose a name that will be accepted by the platform. It is recommanded to
12375 choose names that could be Ada identifiers; such names are almost guaranteed
12376 to be acceptable on all platforms.
12378 The @code{Library_Dir} attribute has a string value that designates the path
12379 (absolute or relative) of the directory where the library will reside.
12380 It must designate an existing directory, and this directory must be
12381 different from the project's object directory. It also needs to be writable.
12383 If both @code{Library_Name} and @code{Library_Dir} are specified and
12384 are legal, then the project file defines a library project. The optional
12385 library-related attributes are checked only for such project files.
12387 The @code{Library_Kind} attribute has a string value that must be one of the
12388 following (case insensitive): @code{"static"}, @code{"dynamic"} or
12389 @code{"relocatable"}. If this attribute is not specified, the library is a
12390 static library, that is an archive of object files that can be potentially
12391 linked into an static executable. Otherwise, the library may be dynamic or
12392 relocatable, that is a library that is loaded only at the start of execution.
12393 Depending on the operating system, there may or may not be a distinction
12394 between dynamic and relocatable libraries. For Unix and VMS Unix there is no
12397 If you need to build both a static and a dynamic library, you should use two
12398 different object directories, since in some cases some extra code needs to
12399 be generated for the latter. For such cases, it is recommended to either use
12400 two different project files, or a single one which uses external variables
12401 to indicate what kind of library should be build.
12403 The @code{Library_Version} attribute has a string value whose interpretation
12404 is platform dependent. It has no effect on VMS and Windows. On Unix, it is
12405 used only for dynamic/relocatable libraries as the internal name of the
12406 library (the @code{"soname"}). If the library file name (built from the
12407 @code{Library_Name}) is different from the @code{Library_Version}, then the
12408 library file will be a symbolic link to the actual file whose name will be
12409 @code{Library_Version}.
12413 @smallexample @c projectfile
12419 for Library_Dir use "lib_dir";
12420 for Library_Name use "dummy";
12421 for Library_Kind use "relocatable";
12422 for Library_Version use "libdummy.so." & Version;
12429 Directory @file{lib_dir} will contain the internal library file whose name
12430 will be @file{libdummy.so.1}, and @file{libdummy.so} will be a symbolic link to
12431 @file{libdummy.so.1}.
12433 When @command{gnatmake} detects that a project file
12434 is a library project file, it will check all immediate sources of the project
12435 and rebuild the library if any of the sources have been recompiled.
12437 When a library is built or rebuilt, an attempt is made to delete all
12438 files in the library directory.
12439 All @file{ALI} files will also be copied from the object directory to the
12440 library directory. To build executables, @command{gnatmake} will use the
12441 library rather than the individual object files. The copy of the @file{ALI}
12442 files are made read-only.
12445 @c **********************************************
12446 @c * Using Third-Party Libraries through Projects
12447 @c **********************************************
12448 @node Using Third-Party Libraries through Projects
12449 @section Using Third-Party Libraries through Projects
12451 Whether you are exporting your own library to make it available to
12452 clients, or you are using a library provided by a third party, it is
12453 convenient to have project files that automatically set the correct
12454 command line switches for the compiler and linker.
12456 Such project files are very similar to the library project files;
12457 @xref{Library Projects}. The only difference is that you set the
12458 @code{Source_Dirs} and @code{Object_Dir} attribute so that they point to the
12459 directories where, respectively, the sources and the read-only ALI files have
12462 If you need to interface with a set of libraries, as opposed to a
12463 single one, you need to create one library project for each of the
12464 libraries. In addition, a top-level project that imports all these
12465 library projects should be provided, so that the user of your library
12466 has a single @code{with} clause to add to his own projects.
12468 For instance, let's assume you are providing two static libraries
12469 @file{liba.a} and @file{libb.a}. The user needs to link with
12470 both of these libraries. Each of these is associated with its
12471 own set of header files. Let's assume furthermore that all the
12472 header files for the two libraries have been installed in the same
12473 directory @file{headers}. The @file{ALI} files are found in the same
12474 @file{headers} directory.
12476 In this case, you should provide the following three projects:
12478 @smallexample @c projectfile
12480 with "liba", "libb";
12481 project My_Library is
12482 for Source_Dirs use ("headers");
12483 for Object_Dir use "headers";
12489 for Source_Dirs use ();
12490 for Library_Dir use "lib";
12491 for Library_Name use "a";
12492 for Library_Kind use "static";
12498 for Source_Dirs use ();
12499 for Library_Dir use "lib";
12500 for Library_Name use "b";
12501 for Library_Kind use "static";
12506 @c *******************************
12507 @c * Stand-alone Library Projects *
12508 @c *******************************
12510 @node Stand-alone Library Projects
12511 @section Stand-alone Library Projects
12514 A Stand-alone Library is a library that contains the necessary code to
12515 elaborate the Ada units that are included in the library. A Stand-alone
12516 Library is suitable to be used in an executable when the main is not
12517 in Ada. However, Stand-alone Libraries may also be used with an Ada main
12520 A Stand-alone Library Project is a Library Project where the library is
12521 a Stand-alone Library.
12523 To be a Stand-alone Library Project, in addition to the two attributes
12524 that make a project a Library Project (@code{Library_Name} and
12525 @code{Library_Dir}, see @ref{Library Projects}), the attribute
12526 @code{Library_Interface} must be defined.
12528 @smallexample @c projectfile
12530 for Library_Dir use "lib_dir";
12531 for Library_Name use "dummy";
12532 for Library_Interface use ("int1", "int1.child");
12536 Attribute @code{Library_Interface} has a non empty string list value,
12537 each string in the list designating a unit contained in an immediate source
12538 of the project file.
12540 When a Stand-alone Library is built, first the binder is invoked to build
12541 a package whose name depends on the library name
12542 (^b~dummy.ads/b^B$DUMMY.ADS/B^ in the example above).
12543 This binder-generated package includes initialization and
12544 finalization procedures whose
12545 names depend on the library name (dummyinit and dummyfinal in the example
12546 above). The object corresponding to this package is included in the library.
12548 A dynamic or relocatable Stand-alone Library is automatically initialized
12549 if automatic initialization of Stand-alone Libraries is supported on the
12550 platform and if attribute @code{Library_Auto_Init} is not specified or
12551 is specified with the value "true". A static Stand-alone Library is never
12552 automatically initialized.
12554 Single string attribute @code{Library_Auto_Init} may be specified with only
12555 two possible values: "false" or "true" (case-insensitive). Specifying
12556 "false" for attribute @code{Library_Auto_Init} will prevent automatic
12557 initialization of dynamic or relocatable libraries.
12559 When a non automatically initialized Stand-alone Library is used
12560 in an executable, its initialization procedure must be called before
12561 any service of the library is used.
12562 When the main subprogram is in Ada, it may mean that the initialization
12563 procedure has to be called during elaboration of another package.
12565 For a Stand-Alone Library, only the @file{ALI} files of the Interface Units
12566 (those that are listed in attribute @code{Library_Interface}) are copied to
12567 the Library Directory. As a consequence, only the Interface Units may be
12568 imported from Ada units outside of the library. If other units are imported,
12569 the binding phase will fail.
12571 When a Stand-Alone Library is bound, the switches that are specified in
12572 the attribute @code{Default_Switches ("Ada")} in package @code{Binder} are
12573 used in the call to @command{gnatbind}.
12575 The string list attribute @code{Library_Options} may be used to specified
12576 additional switches to the call to @command{gcc} to link the library.
12578 The attribute @code{Library_Src_Dir}, may be specified for a
12579 Stand-Alone Library. @code{Library_Src_Dir} is a simple attribute that has a
12580 single string value. Its value must be the path (absolute or relative to the
12581 project directory) of an existing directory. This directory cannot be the
12582 object directory or one of the source directories, but it can be the same as
12583 the library directory. The sources of the Interface
12584 Units of the library, necessary to an Ada client of the library, will be
12585 copied to the designated directory, called Interface Copy directory.
12586 These sources includes the specs of the Interface Units, but they may also
12587 include bodies and subunits, when pragmas @code{Inline} or @code{Inline_Always}
12588 are used, or when there is a generic units in the spec. Before the sources
12589 are copied to the Interface Copy directory, an attempt is made to delete all
12590 files in the Interface Copy directory.
12592 @c *************************************
12593 @c * Switches Related to Project Files *
12594 @c *************************************
12595 @node Switches Related to Project Files
12596 @section Switches Related to Project Files
12599 The following switches are used by GNAT tools that support project files:
12603 @item ^-P^/PROJECT_FILE=^@var{project}
12604 @cindex @option{^-P^/PROJECT_FILE^} (any tool supporting project files)
12605 Indicates the name of a project file. This project file will be parsed with
12606 the verbosity indicated by @option{^-vP^MESSAGE_PROJECT_FILES=^@emph{x}},
12607 if any, and using the external references indicated
12608 by @option{^-X^/EXTERNAL_REFERENCE^} switches, if any.
12610 There may zero, one or more spaces between @option{-P} and @var{project}.
12614 There must be only one @option{^-P^/PROJECT_FILE^} switch on the command line.
12617 Since the Project Manager parses the project file only after all the switches
12618 on the command line are checked, the order of the switches
12619 @option{^-P^/PROJECT_FILE^},
12620 @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}}
12621 or @option{^-X^/EXTERNAL_REFERENCE^} is not significant.
12623 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
12624 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (any tool supporting project files)
12625 Indicates that external variable @var{name} has the value @var{value}.
12626 The Project Manager will use this value for occurrences of
12627 @code{external(name)} when parsing the project file.
12631 If @var{name} or @var{value} includes a space, then @var{name=value} should be
12632 put between quotes.
12640 Several @option{^-X^/EXTERNAL_REFERENCE^} switches can be used simultaneously.
12641 If several @option{^-X^/EXTERNAL_REFERENCE^} switches specify the same
12642 @var{name}, only the last one is used.
12645 An external variable specified with a @option{^-X^/EXTERNAL_REFERENCE^} switch
12646 takes precedence over the value of the same name in the environment.
12648 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
12649 @cindex @code{^-vP^/MESSAGES_PROJECT_FILE^} (any tool supporting project files)
12650 @c Previous line uses code vs option command, to stay less than 80 chars
12651 Indicates the verbosity of the parsing of GNAT project files.
12654 @option{-vP0} means Default;
12655 @option{-vP1} means Medium;
12656 @option{-vP2} means High.
12660 There are three possible options for this qualifier: DEFAULT, MEDIUM and
12665 The default is ^Default^DEFAULT^: no output for syntactically correct
12668 If several @option{^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}} switches are present,
12669 only the last one is used.
12673 @c **********************************
12674 @c * Tools Supporting Project Files *
12675 @c **********************************
12677 @node Tools Supporting Project Files
12678 @section Tools Supporting Project Files
12681 * gnatmake and Project Files::
12682 * The GNAT Driver and Project Files::
12684 * Glide and Project Files::
12688 @node gnatmake and Project Files
12689 @subsection gnatmake and Project Files
12692 This section covers several topics related to @command{gnatmake} and
12693 project files: defining ^switches^switches^ for @command{gnatmake}
12694 and for the tools that it invokes; specifying configuration pragmas;
12695 the use of the @code{Main} attribute; building and rebuilding library project
12699 * ^Switches^Switches^ and Project Files::
12700 * Specifying Configuration Pragmas::
12701 * Project Files and Main Subprograms::
12702 * Library Project Files::
12705 @node ^Switches^Switches^ and Project Files
12706 @subsubsection ^Switches^Switches^ and Project Files
12709 It is not currently possible to specify VMS style qualifiers in the project
12710 files; only Unix style ^switches^switches^ may be specified.
12714 For each of the packages @code{Builder}, @code{Compiler}, @code{Binder}, and
12715 @code{Linker}, you can specify a @code{^Default_Switches^Default_Switches^}
12716 attribute, a @code{^Switches^Switches^} attribute, or both;
12717 as their names imply, these ^switch^switch^-related
12718 attributes affect the ^switches^switches^ that are used for each of these GNAT
12720 @command{gnatmake} is invoked. As will be explained below, these
12721 component-specific ^switches^switches^ precede
12722 the ^switches^switches^ provided on the @command{gnatmake} command line.
12724 The @code{^Default_Switches^Default_Switches^} attribute is an associative
12725 array indexed by language name (case insensitive) whose value is a string list.
12728 @smallexample @c projectfile
12730 package Compiler is
12731 for ^Default_Switches^Default_Switches^ ("Ada")
12732 use ("^-gnaty^-gnaty^",
12739 The @code{^Switches^Switches^} attribute is also an associative array,
12740 indexed by a file name (which may or may not be case sensitive, depending
12741 on the operating system) whose value is a string list. For example:
12743 @smallexample @c projectfile
12746 for ^Switches^Switches^ ("main1.adb")
12748 for ^Switches^Switches^ ("main2.adb")
12755 For the @code{Builder} package, the file names must designate source files
12756 for main subprograms. For the @code{Binder} and @code{Linker} packages, the
12757 file names must designate @file{ALI} or source files for main subprograms.
12758 In each case just the file name without an explicit extension is acceptable.
12760 For each tool used in a program build (@command{gnatmake}, the compiler, the
12761 binder, and the linker), the corresponding package @dfn{contributes} a set of
12762 ^switches^switches^ for each file on which the tool is invoked, based on the
12763 ^switch^switch^-related attributes defined in the package.
12764 In particular, the ^switches^switches^
12765 that each of these packages contributes for a given file @var{f} comprise:
12769 the value of attribute @code{^Switches^Switches^ (@var{f})},
12770 if it is specified in the package for the given file,
12772 otherwise, the value of @code{^Default_Switches^Default_Switches^ ("Ada")},
12773 if it is specified in the package.
12777 If neither of these attributes is defined in the package, then the package does
12778 not contribute any ^switches^switches^ for the given file.
12780 When @command{gnatmake} is invoked on a file, the ^switches^switches^ comprise
12781 two sets, in the following order: those contributed for the file
12782 by the @code{Builder} package;
12783 and the switches passed on the command line.
12785 When @command{gnatmake} invokes a tool (compiler, binder, linker) on a file,
12786 the ^switches^switches^ passed to the tool comprise three sets,
12787 in the following order:
12791 the applicable ^switches^switches^ contributed for the file
12792 by the @code{Builder} package in the project file supplied on the command line;
12795 those contributed for the file by the package (in the relevant project file --
12796 see below) corresponding to the tool; and
12799 the applicable switches passed on the command line.
12803 The term @emph{applicable ^switches^switches^} reflects the fact that
12804 @command{gnatmake} ^switches^switches^ may or may not be passed to individual
12805 tools, depending on the individual ^switch^switch^.
12807 @command{gnatmake} may invoke the compiler on source files from different
12808 projects. The Project Manager will use the appropriate project file to
12809 determine the @code{Compiler} package for each source file being compiled.
12810 Likewise for the @code{Binder} and @code{Linker} packages.
12812 As an example, consider the following package in a project file:
12814 @smallexample @c projectfile
12817 package Compiler is
12818 for ^Default_Switches^Default_Switches^ ("Ada")
12820 for ^Switches^Switches^ ("a.adb")
12822 for ^Switches^Switches^ ("b.adb")
12824 "^-gnaty^-gnaty^");
12831 If @command{gnatmake} is invoked with this project file, and it needs to
12832 compile, say, the files @file{a.adb}, @file{b.adb}, and @file{c.adb}, then
12833 @file{a.adb} will be compiled with the ^switch^switch^
12834 @option{^-O1^-O1^},
12835 @file{b.adb} with ^switches^switches^
12837 and @option{^-gnaty^-gnaty^},
12838 and @file{c.adb} with @option{^-g^-g^}.
12840 The following example illustrates the ordering of the ^switches^switches^
12841 contributed by different packages:
12843 @smallexample @c projectfile
12847 for ^Switches^Switches^ ("main.adb")
12855 package Compiler is
12856 for ^Switches^Switches^ ("main.adb")
12864 If you issue the command:
12867 gnatmake ^-Pproj2^/PROJECT_FILE=PROJ2^ -O0 main
12871 then the compiler will be invoked on @file{main.adb} with the following
12872 sequence of ^switches^switches^
12875 ^-g -O1 -O2 -O0^-g -O1 -O2 -O0^
12878 with the last @option{^-O^-O^}
12879 ^switch^switch^ having precedence over the earlier ones;
12880 several other ^switches^switches^
12881 (such as @option{^-c^-c^}) are added implicitly.
12883 The ^switches^switches^
12885 and @option{^-O1^-O1^} are contributed by package
12886 @code{Builder}, @option{^-O2^-O2^} is contributed
12887 by the package @code{Compiler}
12888 and @option{^-O0^-O0^} comes from the command line.
12890 The @option{^-g^-g^}
12891 ^switch^switch^ will also be passed in the invocation of
12892 @command{Gnatlink.}
12894 A final example illustrates switch contributions from packages in different
12897 @smallexample @c projectfile
12900 for Source_Files use ("pack.ads", "pack.adb");
12901 package Compiler is
12902 for ^Default_Switches^Default_Switches^ ("Ada")
12903 use ("^-gnata^-gnata^");
12911 for Source_Files use ("foo_main.adb", "bar_main.adb");
12913 for ^Switches^Switches^ ("foo_main.adb")
12921 -- Ada source file:
12923 procedure Foo_Main is
12931 gnatmake ^-PProj4^/PROJECT_FILE=PROJ4^ foo_main.adb -cargs -gnato
12935 then the ^switches^switches^ passed to the compiler for @file{foo_main.adb} are
12936 @option{^-g^-g^} (contributed by the package @code{Proj4.Builder}) and
12937 @option{^-gnato^-gnato^} (passed on the command line).
12938 When the imported package @code{Pack} is compiled, the ^switches^switches^ used
12939 are @option{^-g^-g^} from @code{Proj4.Builder},
12940 @option{^-gnata^-gnata^} (contributed from package @code{Proj3.Compiler},
12941 and @option{^-gnato^-gnato^} from the command line.
12944 When using @command{gnatmake} with project files, some ^switches^switches^ or
12945 arguments may be expressed as relative paths. As the working directory where
12946 compilation occurs may change, these relative paths are converted to absolute
12947 paths. For the ^switches^switches^ found in a project file, the relative paths
12948 are relative to the project file directory, for the switches on the command
12949 line, they are relative to the directory where @command{gnatmake} is invoked.
12950 The ^switches^switches^ for which this occurs are:
12956 ^-aI^-aI^, as well as all arguments that are not switches (arguments to
12958 ^-o^-o^, object files specified in package @code{Linker} or after
12959 -largs on the command line). The exception to this rule is the ^switch^switch^
12960 ^--RTS=^--RTS=^ for which a relative path argument is never converted.
12962 @node Specifying Configuration Pragmas
12963 @subsubsection Specifying Configuration Pragmas
12965 When using @command{gnatmake} with project files, if there exists a file
12966 @file{gnat.adc} that contains configuration pragmas, this file will be
12969 Configuration pragmas can be defined by means of the following attributes in
12970 project files: @code{Global_Configuration_Pragmas} in package @code{Builder}
12971 and @code{Local_Configuration_Pragmas} in package @code{Compiler}.
12973 Both these attributes are single string attributes. Their values is the path
12974 name of a file containing configuration pragmas. If a path name is relative,
12975 then it is relative to the project directory of the project file where the
12976 attribute is defined.
12978 When compiling a source, the configuration pragmas used are, in order,
12979 those listed in the file designated by attribute
12980 @code{Global_Configuration_Pragmas} in package @code{Builder} of the main
12981 project file, if it is specified, and those listed in the file designated by
12982 attribute @code{Local_Configuration_Pragmas} in package @code{Compiler} of
12983 the project file of the source, if it exists.
12985 @node Project Files and Main Subprograms
12986 @subsubsection Project Files and Main Subprograms
12989 When using a project file, you can invoke @command{gnatmake}
12990 with one or several main subprograms, by specifying their source files on the
12994 gnatmake ^-P^/PROJECT_FILE=^prj main1 main2 main3
12998 Each of these needs to be a source file of the same project, except
12999 when the switch ^-u^/UNIQUE^ is used.
13002 When ^-u^/UNIQUE^ is not used, all the mains need to be sources of the
13003 same project, one of the project in the tree rooted at the project specified
13004 on the command line. The package @code{Builder} of this common project, the
13005 "main project" is the one that is considered by @command{gnatmake}.
13008 When ^-u^/UNIQUE^ is used, the specified source files may be in projects
13009 imported directly or indirectly by the project specified on the command line.
13010 Note that if such a source file is not part of the project specified on the
13011 command line, the ^switches^switches^ found in package @code{Builder} of the
13012 project specified on the command line, if any, that are transmitted
13013 to the compiler will still be used, not those found in the project file of
13017 When using a project file, you can also invoke @command{gnatmake} without
13018 explicitly specifying any main, and the effect depends on whether you have
13019 defined the @code{Main} attribute. This attribute has a string list value,
13020 where each element in the list is the name of a source file (the file
13021 extension is optional) that contains a unit that can be a main subprogram.
13023 If the @code{Main} attribute is defined in a project file as a non-empty
13024 string list and the switch @option{^-u^/UNIQUE^} is not used on the command
13025 line, then invoking @command{gnatmake} with this project file but without any
13026 main on the command line is equivalent to invoking @command{gnatmake} with all
13027 the file names in the @code{Main} attribute on the command line.
13030 @smallexample @c projectfile
13033 for Main use ("main1", "main2", "main3");
13039 With this project file, @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^"}
13041 @code{"gnatmake ^-Pprj^/PROJECT_FILE=PRJ^ main1 main2 main3"}.
13043 When the project attribute @code{Main} is not specified, or is specified
13044 as an empty string list, or when the switch @option{-u} is used on the command
13045 line, then invoking @command{gnatmake} with no main on the command line will
13046 result in all immediate sources of the project file being checked, and
13047 potentially recompiled. Depending on the presence of the switch @option{-u},
13048 sources from other project files on which the immediate sources of the main
13049 project file depend are also checked and potentially recompiled. In other
13050 words, the @option{-u} switch is applied to all of the immediate sources of the
13053 When no main is specified on the command line and attribute @code{Main} exists
13054 and includes several mains, or when several mains are specified on the
13055 command line, the default ^switches^switches^ in package @code{Builder} will
13056 be used for all mains, even if there are specific ^switches^switches^
13057 specified for one or several mains.
13059 But the ^switches^switches^ from package @code{Binder} or @code{Linker} will be
13060 the specific ^switches^switches^ for each main, if they are specified.
13062 @node Library Project Files
13063 @subsubsection Library Project Files
13066 When @command{gnatmake} is invoked with a main project file that is a library
13067 project file, it is not allowed to specify one or more mains on the command
13071 When a library project file is specified, switches ^-b^/ACTION=BIND^ and
13072 ^-l^/ACTION=LINK^ have special meanings.
13075 @item ^-b^/ACTION=BIND^ is only allwed for stand-alone libraries. It indicates
13076 to @command{gnatmake} that @command{gnatbind} should be invoked for the
13079 @item ^-l^/ACTION=LINK^ may be used for all library projects. It indicates
13080 to @command{gnatmake} that the binder generated file should be compiled
13081 (in the case of a stand-alone library) and that the library should be built.
13085 @node The GNAT Driver and Project Files
13086 @subsection The GNAT Driver and Project Files
13089 A number of GNAT tools, other than @command{^gnatmake^gnatmake^}
13091 @command{^gnatbind^gnatbind^},
13092 @command{^gnatfind^gnatfind^},
13093 @command{^gnatlink^gnatlink^},
13094 @command{^gnatls^gnatls^},
13095 @command{^gnatelim^gnatelim^},
13096 @command{^gnatpp^gnatpp^},
13097 and @command{^gnatxref^gnatxref^}. However, none of these tools can be invoked
13098 directly with a project file switch (@option{^-P^/PROJECT_FILE=^}).
13099 They must be invoked through the @command{gnat} driver.
13101 The @command{gnat} driver is a front-end that accepts a number of commands and
13102 call the corresponding tool. It has been designed initially for VMS to convert
13103 VMS style qualifiers to Unix style switches, but it is now available to all
13104 the GNAT supported platforms.
13106 On non VMS platforms, the @command{gnat} driver accepts the following commands
13107 (case insensitive):
13111 BIND to invoke @command{^gnatbind^gnatbind^}
13113 CHOP to invoke @command{^gnatchop^gnatchop^}
13115 CLEAN to invoke @command{^gnatclean^gnatclean^}
13117 COMP or COMPILE to invoke the compiler
13119 ELIM to invoke @command{^gnatelim^gnatelim^}
13121 FIND to invoke @command{^gnatfind^gnatfind^}
13123 KR or KRUNCH to invoke @command{^gnatkr^gnatkr^}
13125 LINK to invoke @command{^gnatlink^gnatlink^}
13127 LS or LIST to invoke @command{^gnatls^gnatls^}
13129 MAKE to invoke @command{^gnatmake^gnatmake^}
13131 NAME to invoke @command{^gnatname^gnatname^}
13133 PREP or PREPROCESS to invoke @command{^gnatprep^gnatprep^}
13135 PP or PRETTY to invoke @command{^gnatpp^gnatpp^}
13137 STUB to invoke @command{^gnatstub^gnatstub^}
13139 XREF to invoke @command{^gnatxref^gnatxref^}
13143 Note that the compiler is invoked using the command
13144 @command{^gnatmake -f -u -c^gnatmake -f -u -c^}.
13147 The command may be followed by switches and arguments for the invoked
13151 gnat bind -C main.ali
13157 Switches may also be put in text files, one switch per line, and the text
13158 files may be specified with their path name preceded by '@@'.
13161 gnat bind @@args.txt main.ali
13165 In addition, for command BIND, COMP or COMPILE, FIND, ELIM, LS or LIST, LINK,
13166 PP or PRETTY and XREF, the project file related switches
13167 (@option{^-P^/PROJECT_FILE^},
13168 @option{^-X^/EXTERNAL_REFERENCE^} and
13169 @option{^-vP^/MESSAGES_PROJECT_FILE=^x}) may be used in addition to
13170 the switches of the invoking tool.
13173 When GNAT PP or GNAT PRETTY is used with a project file, but with no source
13174 specified on the command line, it invokes @command{^gnatpp^gnatpp^} with all
13175 the immediate sources of the specified project file.
13178 For each of these commands, there is optionally a corresponding package
13179 in the main project.
13183 package @code{Binder} for command BIND (invoking @code{^gnatbind^gnatbind^})
13186 package @code{Compiler} for command COMP or COMPILE (invoking the compiler)
13189 package @code{Finder} for command FIND (invoking @code{^gnatfind^gnatfind^})
13192 package @code{Eliminate} for command ELIM (invoking
13193 @code{^gnatelim^gnatelim^})
13196 package @code{Gnatls} for command LS or LIST (invoking @code{^gnatls^gnatls^})
13199 package @code{Linker} for command LINK (invoking @code{^gnatlink^gnatlink^})
13202 package @code{Pretty_Printer} for command PP or PRETTY
13203 (invoking @code{^gnatpp^gnatpp^})
13206 package @code{Cross_Reference} for command XREF (invoking
13207 @code{^gnatxref^gnatxref^})
13212 Package @code{Gnatls} has a unique attribute @code{^Switches^Switches^},
13213 a simple variable with a string list value. It contains ^switches^switches^
13214 for the invocation of @code{^gnatls^gnatls^}.
13216 @smallexample @c projectfile
13220 for ^Switches^Switches^
13229 All other packages have two attribute @code{^Switches^Switches^} and
13230 @code{^Default_Switches^Default_Switches^}.
13233 @code{^Switches^Switches^} is an associated array attribute, indexed by the
13234 source file name, that has a string list value: the ^switches^switches^ to be
13235 used when the tool corresponding to the package is invoked for the specific
13239 @code{^Default_Switches^Default_Switches^} is an associative array attribute,
13240 indexed by the programming language that has a string list value.
13241 @code{^Default_Switches^Default_Switches^ ("Ada")} contains the
13242 ^switches^switches^ for the invocation of the tool corresponding
13243 to the package, except if a specific @code{^Switches^Switches^} attribute
13244 is specified for the source file.
13246 @smallexample @c projectfile
13250 for Source_Dirs use ("./**");
13253 for ^Switches^Switches^ use
13260 package Compiler is
13261 for ^Default_Switches^Default_Switches^ ("Ada")
13262 use ("^-gnatv^-gnatv^",
13263 "^-gnatwa^-gnatwa^");
13269 for ^Default_Switches^Default_Switches^ ("Ada")
13277 for ^Default_Switches^Default_Switches^ ("Ada")
13279 for ^Switches^Switches^ ("main.adb")
13288 for ^Default_Switches^Default_Switches^ ("Ada")
13295 package Cross_Reference is
13296 for ^Default_Switches^Default_Switches^ ("Ada")
13301 end Cross_Reference;
13307 With the above project file, commands such as
13310 ^gnat comp -Pproj main^GNAT COMP /PROJECT_FILE=PROJ MAIN^
13311 ^gnat ls -Pproj main^GNAT LIST /PROJECT_FILE=PROJ MAIN^
13312 ^gnat xref -Pproj main^GNAT XREF /PROJECT_FILE=PROJ MAIN^
13313 ^gnat bind -Pproj main.ali^GNAT BIND /PROJECT_FILE=PROJ MAIN.ALI^
13314 ^gnat link -Pproj main.ali^GNAT LINK /PROJECT_FILE=PROJ MAIN.ALI^
13318 will set up the environment properly and invoke the tool with the switches
13319 found in the package corresponding to the tool:
13320 @code{^Default_Switches^Default_Switches^ ("Ada")} for all tools,
13321 except @code{^Switches^Switches^ ("main.adb")}
13322 for @code{^gnatlink^gnatlink^}.
13325 @node Glide and Project Files
13326 @subsection Glide and Project Files
13329 Glide will automatically recognize the @file{.gpr} extension for
13330 project files, and will
13331 convert them to its own internal format automatically. However, it
13332 doesn't provide a syntax-oriented editor for modifying these
13334 The project file will be loaded as text when you select the menu item
13335 @code{Ada} @result{} @code{Project} @result{} @code{Edit}.
13336 You can edit this text and save the @file{gpr} file;
13337 when you next select this project file in Glide it
13338 will be automatically reloaded.
13341 @c **********************
13342 @node An Extended Example
13343 @section An Extended Example
13346 Suppose that we have two programs, @var{prog1} and @var{prog2},
13347 whose sources are in corresponding directories. We would like
13348 to build them with a single @command{gnatmake} command, and we want to place
13349 their object files into @file{build} subdirectories of the source directories.
13350 Furthermore, we want to have to have two separate subdirectories
13351 in @file{build} -- @file{release} and @file{debug} -- which will contain
13352 the object files compiled with different set of compilation flags.
13354 In other words, we have the following structure:
13371 Here are the project files that we must place in a directory @file{main}
13372 to maintain this structure:
13376 @item We create a @code{Common} project with a package @code{Compiler} that
13377 specifies the compilation ^switches^switches^:
13382 @b{project} Common @b{is}
13384 @b{for} Source_Dirs @b{use} (); -- No source files
13388 @b{type} Build_Type @b{is} ("release", "debug");
13389 Build : Build_Type := External ("BUILD", "debug");
13392 @b{package} Compiler @b{is}
13393 @b{case} Build @b{is}
13394 @b{when} "release" =>
13395 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
13396 @b{use} ("^-O2^-O2^");
13397 @b{when} "debug" =>
13398 @b{for} ^Default_Switches^Default_Switches^ ("Ada")
13399 @b{use} ("^-g^-g^");
13407 @item We create separate projects for the two programs:
13414 @b{project} Prog1 @b{is}
13416 @b{for} Source_Dirs @b{use} ("prog1");
13417 @b{for} Object_Dir @b{use} "prog1/build/" & Common.Build;
13419 @b{package} Compiler @b{renames} Common.Compiler;
13430 @b{project} Prog2 @b{is}
13432 @b{for} Source_Dirs @b{use} ("prog2");
13433 @b{for} Object_Dir @b{use} "prog2/build/" & Common.Build;
13435 @b{package} Compiler @b{renames} Common.Compiler;
13441 @item We create a wrapping project @code{Main}:
13450 @b{project} Main @b{is}
13452 @b{package} Compiler @b{renames} Common.Compiler;
13458 @item Finally we need to create a dummy procedure that @code{with}s (either
13459 explicitly or implicitly) all the sources of our two programs.
13464 Now we can build the programs using the command
13467 gnatmake ^-P^/PROJECT_FILE=^main dummy
13471 for the Debug mode, or
13475 gnatmake -Pmain -XBUILD=release
13481 GNAT MAKE /PROJECT_FILE=main /EXTERNAL_REFERENCE=BUILD=release
13486 for the Release mode.
13488 @c ********************************
13489 @c * Project File Complete Syntax *
13490 @c ********************************
13492 @node Project File Complete Syntax
13493 @section Project File Complete Syntax
13497 context_clause project_declaration
13503 @b{with} path_name @{ , path_name @} ;
13508 project_declaration ::=
13509 simple_project_declaration | project_extension
13511 simple_project_declaration ::=
13512 @b{project} <project_>simple_name @b{is}
13513 @{declarative_item@}
13514 @b{end} <project_>simple_name;
13516 project_extension ::=
13517 @b{project} <project_>simple_name @b{extends} path_name @b{is}
13518 @{declarative_item@}
13519 @b{end} <project_>simple_name;
13521 declarative_item ::=
13522 package_declaration |
13523 typed_string_declaration |
13524 other_declarative_item
13526 package_declaration ::=
13527 package_specification | package_renaming
13529 package_specification ::=
13530 @b{package} package_identifier @b{is}
13531 @{simple_declarative_item@}
13532 @b{end} package_identifier ;
13534 package_identifier ::=
13535 @code{Naming} | @code{Builder} | @code{Compiler} | @code{Binder} |
13536 @code{Linker} | @code{Finder} | @code{Cross_Reference} |
13537 @code{^gnatls^gnatls^} | @code{IDE} | @code{Pretty_Printer}
13539 package_renaming ::==
13540 @b{package} package_identifier @b{renames}
13541 <project_>simple_name.package_identifier ;
13543 typed_string_declaration ::=
13544 @b{type} <typed_string_>_simple_name @b{is}
13545 ( string_literal @{, string_literal@} );
13547 other_declarative_item ::=
13548 attribute_declaration |
13549 typed_variable_declaration |
13550 variable_declaration |
13553 attribute_declaration ::=
13554 full_associative_array_declaration |
13555 @b{for} attribute_designator @b{use} expression ;
13557 full_associative_array_declaration ::=
13558 @b{for} <associative_array_attribute_>simple_name @b{use}
13559 <project_>simple_name [ . <package_>simple_Name ] ' <attribute_>simple_name ;
13561 attribute_designator ::=
13562 <simple_attribute_>simple_name |
13563 <associative_array_attribute_>simple_name ( string_literal )
13565 typed_variable_declaration ::=
13566 <typed_variable_>simple_name : <typed_string_>name := string_expression ;
13568 variable_declaration ::=
13569 <variable_>simple_name := expression;
13579 attribute_reference
13585 ( <string_>expression @{ , <string_>expression @} )
13588 @b{external} ( string_literal [, string_literal] )
13590 attribute_reference ::=
13591 attribute_prefix ' <simple_attribute_>simple_name [ ( literal_string ) ]
13593 attribute_prefix ::=
13595 <project_>simple_name | package_identifier |
13596 <project_>simple_name . package_identifier
13598 case_construction ::=
13599 @b{case} <typed_variable_>name @b{is}
13604 @b{when} discrete_choice_list =>
13605 @{case_construction | attribute_declaration@}
13607 discrete_choice_list ::=
13608 string_literal @{| string_literal@} |
13612 simple_name @{. simple_name@}
13615 identifier (same as Ada)
13620 @node The Cross-Referencing Tools gnatxref and gnatfind
13621 @chapter The Cross-Referencing Tools @code{gnatxref} and @code{gnatfind}
13626 The compiler generates cross-referencing information (unless
13627 you set the @samp{-gnatx} switch), which are saved in the @file{.ali} files.
13628 This information indicates where in the source each entity is declared and
13629 referenced. Note that entities in package Standard are not included, but
13630 entities in all other predefined units are included in the output.
13632 Before using any of these two tools, you need to compile successfully your
13633 application, so that GNAT gets a chance to generate the cross-referencing
13636 The two tools @code{gnatxref} and @code{gnatfind} take advantage of this
13637 information to provide the user with the capability to easily locate the
13638 declaration and references to an entity. These tools are quite similar,
13639 the difference being that @code{gnatfind} is intended for locating
13640 definitions and/or references to a specified entity or entities, whereas
13641 @code{gnatxref} is oriented to generating a full report of all
13644 To use these tools, you must not compile your application using the
13645 @option{-gnatx} switch on the @file{gnatmake} command line
13646 (see @ref{The GNAT Make Program gnatmake}). Otherwise, cross-referencing
13647 information will not be generated.
13650 * gnatxref Switches::
13651 * gnatfind Switches::
13652 * Project Files for gnatxref and gnatfind::
13653 * Regular Expressions in gnatfind and gnatxref::
13654 * Examples of gnatxref Usage::
13655 * Examples of gnatfind Usage::
13658 @node gnatxref Switches
13659 @section @code{gnatxref} Switches
13662 The command invocation for @code{gnatxref} is:
13664 $ gnatxref [switches] sourcefile1 [sourcefile2 ...]
13671 @item sourcefile1, sourcefile2
13672 identifies the source files for which a report is to be generated. The
13673 ``with''ed units will be processed too. You must provide at least one file.
13675 These file names are considered to be regular expressions, so for instance
13676 specifying @file{source*.adb} is the same as giving every file in the current
13677 directory whose name starts with @file{source} and whose extension is
13683 The switches can be :
13686 @item ^-a^/ALL_FILES^
13687 @cindex @option{^-a^/ALL_FILES^} (@command{gnatxref})
13688 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
13689 the read-only files found in the library search path. Otherwise, these files
13690 will be ignored. This option can be used to protect Gnat sources or your own
13691 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
13692 much faster, and their output much smaller. Read-only here refers to access
13693 or permissions status in the file system for the current user.
13696 @cindex @option{-aIDIR} (@command{gnatxref})
13697 When looking for source files also look in directory DIR. The order in which
13698 source file search is undertaken is the same as for @file{gnatmake}.
13701 @cindex @option{-aODIR} (@command{gnatxref})
13702 When searching for library and object files, look in directory
13703 DIR. The order in which library files are searched is the same as for
13707 @cindex @option{-nostdinc} (@command{gnatxref})
13708 Do not look for sources in the system default directory.
13711 @cindex @option{-nostdlib} (@command{gnatxref})
13712 Do not look for library files in the system default directory.
13714 @item --RTS=@var{rts-path}
13715 @cindex @option{--RTS} (@command{gnatxref})
13716 Specifies the default location of the runtime library. Same meaning as the
13717 equivalent @code{gnatmake} flag (see @ref{Switches for gnatmake}).
13719 @item ^-d^/DERIVED_TYPES^
13720 @cindex @option{^-d^/DERIVED_TYPES^} (@command{gnatxref})
13721 If this switch is set @code{gnatxref} will output the parent type
13722 reference for each matching derived types.
13724 @item ^-f^/FULL_PATHNAME^
13725 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatxref})
13726 If this switch is set, the output file names will be preceded by their
13727 directory (if the file was found in the search path). If this switch is
13728 not set, the directory will not be printed.
13730 @item ^-g^/IGNORE_LOCALS^
13731 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatxref})
13732 If this switch is set, information is output only for library-level
13733 entities, ignoring local entities. The use of this switch may accelerate
13734 @code{gnatfind} and @code{gnatxref}.
13737 @cindex @option{-IDIR} (@command{gnatxref})
13738 Equivalent to @samp{-aODIR -aIDIR}.
13741 @cindex @option{-pFILE} (@command{gnatxref})
13742 Specify a project file to use @xref{Project Files}. These project files are
13743 the @file{.adp} files used by Glide. If you need to use the @file{.gpr}
13744 project files, you should use gnatxref through the GNAT driver
13745 (@command{gnat xref -Pproject}).
13747 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
13748 project file in the current directory.
13750 If a project file is either specified or found by the tools, then the content
13751 of the source directory and object directory lines are added as if they
13752 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^}
13753 and @samp{^-aO^OBJECT_SEARCH^}.
13755 Output only unused symbols. This may be really useful if you give your
13756 main compilation unit on the command line, as @code{gnatxref} will then
13757 display every unused entity and 'with'ed package.
13761 Instead of producing the default output, @code{gnatxref} will generate a
13762 @file{tags} file that can be used by vi. For examples how to use this
13763 feature, see @xref{Examples of gnatxref Usage}. The tags file is output
13764 to the standard output, thus you will have to redirect it to a file.
13770 All these switches may be in any order on the command line, and may even
13771 appear after the file names. They need not be separated by spaces, thus
13772 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
13773 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
13775 @node gnatfind Switches
13776 @section @code{gnatfind} Switches
13779 The command line for @code{gnatfind} is:
13782 $ gnatfind [switches] pattern[:sourcefile[:line[:column]]]
13791 An entity will be output only if it matches the regular expression found
13792 in @samp{pattern}, see @xref{Regular Expressions in gnatfind and gnatxref}.
13794 Omitting the pattern is equivalent to specifying @samp{*}, which
13795 will match any entity. Note that if you do not provide a pattern, you
13796 have to provide both a sourcefile and a line.
13798 Entity names are given in Latin-1, with uppercase/lowercase equivalence
13799 for matching purposes. At the current time there is no support for
13800 8-bit codes other than Latin-1, or for wide characters in identifiers.
13803 @code{gnatfind} will look for references, bodies or declarations
13804 of symbols referenced in @file{sourcefile}, at line @samp{line}
13805 and column @samp{column}. See @pxref{Examples of gnatfind Usage}
13806 for syntax examples.
13809 is a decimal integer identifying the line number containing
13810 the reference to the entity (or entities) to be located.
13813 is a decimal integer identifying the exact location on the
13814 line of the first character of the identifier for the
13815 entity reference. Columns are numbered from 1.
13817 @item file1 file2 ...
13818 The search will be restricted to these source files. If none are given, then
13819 the search will be done for every library file in the search path.
13820 These file must appear only after the pattern or sourcefile.
13822 These file names are considered to be regular expressions, so for instance
13823 specifying 'source*.adb' is the same as giving every file in the current
13824 directory whose name starts with 'source' and whose extension is 'adb'.
13826 The location of the spec of the entity will always be displayed, even if it
13827 isn't in one of file1, file2,... The occurrences of the entity in the
13828 separate units of the ones given on the command line will also be displayed.
13830 Note that if you specify at least one file in this part, @code{gnatfind} may
13831 sometimes not be able to find the body of the subprograms...
13836 At least one of 'sourcefile' or 'pattern' has to be present on
13839 The following switches are available:
13843 @item ^-a^/ALL_FILES^
13844 @cindex @option{^-a^/ALL_FILES^} (@command{gnatfind})
13845 If this switch is present, @code{gnatfind} and @code{gnatxref} will parse
13846 the read-only files found in the library search path. Otherwise, these files
13847 will be ignored. This option can be used to protect Gnat sources or your own
13848 libraries from being parsed, thus making @code{gnatfind} and @code{gnatxref}
13849 much faster, and their output much smaller. Read-only here refers to access
13850 or permission status in the file system for the current user.
13853 @cindex @option{-aIDIR} (@command{gnatfind})
13854 When looking for source files also look in directory DIR. The order in which
13855 source file search is undertaken is the same as for @file{gnatmake}.
13858 @cindex @option{-aODIR} (@command{gnatfind})
13859 When searching for library and object files, look in directory
13860 DIR. The order in which library files are searched is the same as for
13864 @cindex @option{-nostdinc} (@command{gnatfind})
13865 Do not look for sources in the system default directory.
13868 @cindex @option{-nostdlib} (@command{gnatfind})
13869 Do not look for library files in the system default directory.
13871 @item --RTS=@var{rts-path}
13872 @cindex @option{--RTS} (@command{gnatfind})
13873 Specifies the default location of the runtime library. Same meaning as the
13874 equivalent @code{gnatmake} flag (see @ref{Switches for gnatmake}).
13876 @item ^-d^/DERIVED_TYPE_INFORMATION^
13877 @cindex @option{^-d^/DERIVED_TYPE_INFORMATION^} (@code{gnatfind})
13878 If this switch is set, then @code{gnatfind} will output the parent type
13879 reference for each matching derived types.
13881 @item ^-e^/EXPRESSIONS^
13882 @cindex @option{^-e^/EXPRESSIONS^} (@command{gnatfind})
13883 By default, @code{gnatfind} accept the simple regular expression set for
13884 @samp{pattern}. If this switch is set, then the pattern will be
13885 considered as full Unix-style regular expression.
13887 @item ^-f^/FULL_PATHNAME^
13888 @cindex @option{^-f^/FULL_PATHNAME^} (@command{gnatfind})
13889 If this switch is set, the output file names will be preceded by their
13890 directory (if the file was found in the search path). If this switch is
13891 not set, the directory will not be printed.
13893 @item ^-g^/IGNORE_LOCALS^
13894 @cindex @option{^-g^/IGNORE_LOCALS^} (@command{gnatfind})
13895 If this switch is set, information is output only for library-level
13896 entities, ignoring local entities. The use of this switch may accelerate
13897 @code{gnatfind} and @code{gnatxref}.
13900 @cindex @option{-IDIR} (@command{gnatfind})
13901 Equivalent to @samp{-aODIR -aIDIR}.
13904 @cindex @option{-pFILE} (@command{gnatfind})
13905 Specify a project file (@pxref{Project Files}) to use.
13906 By default, @code{gnatxref} and @code{gnatfind} will try to locate a
13907 project file in the current directory.
13909 If a project file is either specified or found by the tools, then the content
13910 of the source directory and object directory lines are added as if they
13911 had been specified respectively by @samp{^-aI^/SOURCE_SEARCH^} and
13912 @samp{^-aO^/OBJECT_SEARCH^}.
13914 @item ^-r^/REFERENCES^
13915 @cindex @option{^-r^/REFERENCES^} (@command{gnatfind})
13916 By default, @code{gnatfind} will output only the information about the
13917 declaration, body or type completion of the entities. If this switch is
13918 set, the @code{gnatfind} will locate every reference to the entities in
13919 the files specified on the command line (or in every file in the search
13920 path if no file is given on the command line).
13922 @item ^-s^/PRINT_LINES^
13923 @cindex @option{^-s^/PRINT_LINES^} (@command{gnatfind})
13924 If this switch is set, then @code{gnatfind} will output the content
13925 of the Ada source file lines were the entity was found.
13927 @item ^-t^/TYPE_HIERARCHY^
13928 @cindex @option{^-t^/TYPE_HIERARCHY^} (@command{gnatfind})
13929 If this switch is set, then @code{gnatfind} will output the type hierarchy for
13930 the specified type. It act like -d option but recursively from parent
13931 type to parent type. When this switch is set it is not possible to
13932 specify more than one file.
13937 All these switches may be in any order on the command line, and may even
13938 appear after the file names. They need not be separated by spaces, thus
13939 you can say @samp{gnatxref ^-ag^/ALL_FILES/IGNORE_LOCALS^} instead of
13940 @samp{gnatxref ^-a -g^/ALL_FILES /IGNORE_LOCALS^}.
13942 As stated previously, gnatfind will search in every directory in the
13943 search path. You can force it to look only in the current directory if
13944 you specify @code{*} at the end of the command line.
13946 @node Project Files for gnatxref and gnatfind
13947 @section Project Files for @command{gnatxref} and @command{gnatfind}
13950 Project files allow a programmer to specify how to compile its
13951 application, where to find sources, etc. These files are used
13953 primarily by the Glide Ada mode, but they can also be used
13956 @code{gnatxref} and @code{gnatfind}.
13958 A project file name must end with @file{.gpr}. If a single one is
13959 present in the current directory, then @code{gnatxref} and @code{gnatfind} will
13960 extract the information from it. If multiple project files are found, none of
13961 them is read, and you have to use the @samp{-p} switch to specify the one
13964 The following lines can be included, even though most of them have default
13965 values which can be used in most cases.
13966 The lines can be entered in any order in the file.
13967 Except for @file{src_dir} and @file{obj_dir}, you can only have one instance of
13968 each line. If you have multiple instances, only the last one is taken into
13973 [default: @code{"^./^[]^"}]
13974 specifies a directory where to look for source files. Multiple @code{src_dir}
13975 lines can be specified and they will be searched in the order they
13979 [default: @code{"^./^[]^"}]
13980 specifies a directory where to look for object and library files. Multiple
13981 @code{obj_dir} lines can be specified, and they will be searched in the order
13984 @item comp_opt=SWITCHES
13985 [default: @code{""}]
13986 creates a variable which can be referred to subsequently by using
13987 the @code{$@{comp_opt@}} notation. This is intended to store the default
13988 switches given to @command{gnatmake} and @command{gcc}.
13990 @item bind_opt=SWITCHES
13991 [default: @code{""}]
13992 creates a variable which can be referred to subsequently by using
13993 the @samp{$@{bind_opt@}} notation. This is intended to store the default
13994 switches given to @command{gnatbind}.
13996 @item link_opt=SWITCHES
13997 [default: @code{""}]
13998 creates a variable which can be referred to subsequently by using
13999 the @samp{$@{link_opt@}} notation. This is intended to store the default
14000 switches given to @command{gnatlink}.
14002 @item main=EXECUTABLE
14003 [default: @code{""}]
14004 specifies the name of the executable for the application. This variable can
14005 be referred to in the following lines by using the @samp{$@{main@}} notation.
14008 @item comp_cmd=COMMAND
14009 [default: @code{"GNAT COMPILE /SEARCH=$@{src_dir@} /DEBUG /TRY_SEMANTICS"}]
14012 @item comp_cmd=COMMAND
14013 [default: @code{"gcc -c -I$@{src_dir@} -g -gnatq"}]
14015 specifies the command used to compile a single file in the application.
14018 @item make_cmd=COMMAND
14019 [default: @code{"GNAT MAKE $@{main@}
14020 /SOURCE_SEARCH=$@{src_dir@} /OBJECT_SEARCH=$@{obj_dir@}
14021 /DEBUG /TRY_SEMANTICS /COMPILER_QUALIFIERS $@{comp_opt@}
14022 /BINDER_QUALIFIERS $@{bind_opt@} /LINKER_QUALIFIERS $@{link_opt@}"}]
14025 @item make_cmd=COMMAND
14026 [default: @code{"gnatmake $@{main@} -aI$@{src_dir@}
14027 -aO$@{obj_dir@} -g -gnatq -cargs $@{comp_opt@}
14028 -bargs $@{bind_opt@} -largs $@{link_opt@}"}]
14030 specifies the command used to recompile the whole application.
14032 @item run_cmd=COMMAND
14033 [default: @code{"$@{main@}"}]
14034 specifies the command used to run the application.
14036 @item debug_cmd=COMMAND
14037 [default: @code{"gdb $@{main@}"}]
14038 specifies the command used to debug the application
14043 @command{gnatxref} and @command{gnatfind} only take into account the
14044 @code{src_dir} and @code{obj_dir} lines, and ignore the others.
14046 @node Regular Expressions in gnatfind and gnatxref
14047 @section Regular Expressions in @code{gnatfind} and @code{gnatxref}
14050 As specified in the section about @command{gnatfind}, the pattern can be a
14051 regular expression. Actually, there are to set of regular expressions
14052 which are recognized by the program :
14055 @item globbing patterns
14056 These are the most usual regular expression. They are the same that you
14057 generally used in a Unix shell command line, or in a DOS session.
14059 Here is a more formal grammar :
14066 term ::= elmt -- matches elmt
14067 term ::= elmt elmt -- concatenation (elmt then elmt)
14068 term ::= * -- any string of 0 or more characters
14069 term ::= ? -- matches any character
14070 term ::= [char @{char@}] -- matches any character listed
14071 term ::= [char - char] -- matches any character in range
14075 @item full regular expression
14076 The second set of regular expressions is much more powerful. This is the
14077 type of regular expressions recognized by utilities such a @file{grep}.
14079 The following is the form of a regular expression, expressed in Ada
14080 reference manual style BNF is as follows
14087 regexp ::= term @{| term@} -- alternation (term or term ...)
14089 term ::= item @{item@} -- concatenation (item then item)
14091 item ::= elmt -- match elmt
14092 item ::= elmt * -- zero or more elmt's
14093 item ::= elmt + -- one or more elmt's
14094 item ::= elmt ? -- matches elmt or nothing
14097 elmt ::= nschar -- matches given character
14098 elmt ::= [nschar @{nschar@}] -- matches any character listed
14099 elmt ::= [^^^ nschar @{nschar@}] -- matches any character not listed
14100 elmt ::= [char - char] -- matches chars in given range
14101 elmt ::= \ char -- matches given character
14102 elmt ::= . -- matches any single character
14103 elmt ::= ( regexp ) -- parens used for grouping
14105 char ::= any character, including special characters
14106 nschar ::= any character except ()[].*+?^^^
14110 Following are a few examples :
14114 will match any of the two strings 'abcde' and 'fghi'.
14117 will match any string like 'abd', 'abcd', 'abccd', 'abcccd', and so on
14120 will match any string which has only lowercase characters in it (and at
14121 least one character
14126 @node Examples of gnatxref Usage
14127 @section Examples of @code{gnatxref} Usage
14129 @subsection General Usage
14132 For the following examples, we will consider the following units :
14134 @smallexample @c ada
14140 3: procedure Foo (B : in Integer);
14147 1: package body Main is
14148 2: procedure Foo (B : in Integer) is
14159 2: procedure Print (B : Integer);
14168 The first thing to do is to recompile your application (for instance, in
14169 that case just by doing a @samp{gnatmake main}, so that GNAT generates
14170 the cross-referencing information.
14171 You can then issue any of the following commands:
14173 @item gnatxref main.adb
14174 @code{gnatxref} generates cross-reference information for main.adb
14175 and every unit 'with'ed by main.adb.
14177 The output would be:
14185 Decl: main.ads 3:20
14186 Body: main.adb 2:20
14187 Ref: main.adb 4:13 5:13 6:19
14190 Ref: main.adb 6:8 7:8
14200 Decl: main.ads 3:15
14201 Body: main.adb 2:15
14204 Body: main.adb 1:14
14207 Ref: main.adb 6:12 7:12
14211 that is the entity @code{Main} is declared in main.ads, line 2, column 9,
14212 its body is in main.adb, line 1, column 14 and is not referenced any where.
14214 The entity @code{Print} is declared in bar.ads, line 2, column 15 and it
14215 it referenced in main.adb, line 6 column 12 and line 7 column 12.
14217 @item gnatxref package1.adb package2.ads
14218 @code{gnatxref} will generates cross-reference information for
14219 package1.adb, package2.ads and any other package 'with'ed by any
14225 @subsection Using gnatxref with vi
14227 @code{gnatxref} can generate a tags file output, which can be used
14228 directly from @file{vi}. Note that the standard version of @file{vi}
14229 will not work properly with overloaded symbols. Consider using another
14230 free implementation of @file{vi}, such as @file{vim}.
14233 $ gnatxref -v gnatfind.adb > tags
14237 will generate the tags file for @code{gnatfind} itself (if the sources
14238 are in the search path!).
14240 From @file{vi}, you can then use the command @samp{:tag @i{entity}}
14241 (replacing @i{entity} by whatever you are looking for), and vi will
14242 display a new file with the corresponding declaration of entity.
14245 @node Examples of gnatfind Usage
14246 @section Examples of @code{gnatfind} Usage
14250 @item gnatfind ^-f^/FULL_PATHNAME^ xyz:main.adb
14251 Find declarations for all entities xyz referenced at least once in
14252 main.adb. The references are search in every library file in the search
14255 The directories will be printed as well (as the @samp{^-f^/FULL_PATHNAME^}
14258 The output will look like:
14260 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
14261 ^directory/^[directory]^main.adb:24:10: xyz <= body
14262 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
14266 that is to say, one of the entities xyz found in main.adb is declared at
14267 line 12 of main.ads (and its body is in main.adb), and another one is
14268 declared at line 45 of foo.ads
14270 @item gnatfind ^-fs^/FULL_PATHNAME/SOURCE_LINE^ xyz:main.adb
14271 This is the same command as the previous one, instead @code{gnatfind} will
14272 display the content of the Ada source file lines.
14274 The output will look like:
14277 ^directory/^[directory]^main.ads:106:14: xyz <= declaration
14279 ^directory/^[directory]^main.adb:24:10: xyz <= body
14281 ^directory/^[directory]^foo.ads:45:23: xyz <= declaration
14286 This can make it easier to find exactly the location your are looking
14289 @item gnatfind ^-r^/REFERENCES^ "*x*":main.ads:123 foo.adb
14290 Find references to all entities containing an x that are
14291 referenced on line 123 of main.ads.
14292 The references will be searched only in main.ads and foo.adb.
14294 @item gnatfind main.ads:123
14295 Find declarations and bodies for all entities that are referenced on
14296 line 123 of main.ads.
14298 This is the same as @code{gnatfind "*":main.adb:123}.
14300 @item gnatfind ^mydir/^[mydir]^main.adb:123:45
14301 Find the declaration for the entity referenced at column 45 in
14302 line 123 of file main.adb in directory mydir. Note that it
14303 is usual to omit the identifier name when the column is given,
14304 since the column position identifies a unique reference.
14306 The column has to be the beginning of the identifier, and should not
14307 point to any character in the middle of the identifier.
14312 @c *********************************
14313 @node The GNAT Pretty-Printer gnatpp
14314 @chapter The GNAT Pretty-Printer @command{gnatpp}
14316 @cindex Pretty-Printer
14319 ^The @command{gnatpp} tool^GNAT PRETTY^ is an ASIS-based utility
14320 for source reformatting / pretty-printing.
14321 It takes an Ada source file as input and generates a reformatted
14323 You can specify various style directives via switches; e.g.,
14324 identifier case conventions, rules of indentation, and comment layout.
14326 To produce a reformatted file, @command{gnatpp} generates and uses the ASIS
14327 tree for the input source and thus requires the input to be syntactically and
14328 semantically legal.
14329 If this condition is not met, @command{gnatpp} will terminate with an
14330 error message; no output file will be generated.
14332 If the compilation unit
14333 contained in the input source depends semantically upon units located
14334 outside the current directory, you have to provide the source search path
14335 when invoking @command{gnatpp}; see the description of the @command{gnatpp}
14338 The @command{gnatpp} command has the form
14341 $ gnatpp [@var{switches}] @var{filename}
14348 @var{switches} is an optional sequence of switches defining such properties as
14349 the formatting rules, the source search path, and the destination for the
14353 @var{filename} is the name (including the extension) of the source file to
14354 reformat; ``wildcards'' or several file names on the same gnatpp command are
14355 allowed. The file name may contain path information; it does not have to follow
14356 the GNAT file naming rules
14361 * Switches for gnatpp::
14362 * Formatting Rules::
14365 @node Switches for gnatpp
14366 @section Switches for @command{gnatpp}
14369 The following subsections describe the various switches accepted by
14370 @command{gnatpp}, organized by category.
14373 You specify a switch by supplying a name and generally also a value.
14374 In many cases the values for a switch with a given name are incompatible with
14376 (for example the switch that controls the casing of a reserved word may have
14377 exactly one value: upper case, lower case, or
14378 mixed case) and thus exactly one such switch can be in effect for an
14379 invocation of @command{gnatpp}.
14380 If more than one is supplied, the last one is used.
14381 However, some values for the same switch are mutually compatible.
14382 You may supply several such switches to @command{gnatpp}, but then
14383 each must be specified in full, with both the name and the value.
14384 Abbreviated forms (the name appearing once, followed by each value) are
14386 For example, to set
14387 the alignment of the assignment delimiter both in declarations and in
14388 assignment statements, you must write @option{-A2A3}
14389 (or @option{-A2 -A3}), but not @option{-A23}.
14393 In many cases the set of options for a given qualifier are incompatible with
14394 each other (for example the qualifier that controls the casing of a reserved
14395 word may have exactly one option, which specifies either upper case, lower
14396 case, or mixed case), and thus exactly one such option can be in effect for
14397 an invocation of @command{gnatpp}.
14398 If more than one is supplied, the last one is used.
14399 However, some qualifiers have options that are mutually compatible,
14400 and then you may then supply several such options when invoking
14404 In most cases, it is obvious whether or not the
14405 ^values for a switch with a given name^options for a given qualifier^
14406 are compatible with each other.
14407 When the semantics might not be evident, the summaries below explicitly
14408 indicate the effect.
14411 * Alignment Control::
14413 * Construct Layout Control::
14414 * General Text Layout Control::
14415 * Other Formatting Options::
14416 * Setting the Source Search Path::
14417 * Output File Control::
14418 * Other gnatpp Switches::
14422 @node Alignment Control
14423 @subsection Alignment Control
14424 @cindex Alignment control in @command{gnatpp}
14427 Programs can be easier to read if certain constructs are vertically aligned.
14428 By default all alignments are set ON.
14429 Through the @option{^-A0^/ALIGN=OFF^} switch you may reset the default to
14430 OFF, and then use one or more of the other
14431 ^@option{-A@var{n}} switches^@option{/ALIGN} options^
14432 to activate alignment for specific constructs.
14435 @cindex @option{^-A@var{n}^/ALIGN^} (@command{gnatpp})
14439 Set all alignments to ON
14442 @item ^-A0^/ALIGN=OFF^
14443 Set all alignments to OFF
14445 @item ^-A1^/ALIGN=COLONS^
14446 Align @code{:} in declarations
14448 @item ^-A2^/ALIGN=DECLARATIONS^
14449 Align @code{:=} in initializations in declarations
14451 @item ^-A3^/ALIGN=STATEMENTS^
14452 Align @code{:=} in assignment statements
14454 @item ^-A4^/ALIGN=ARROWS^
14455 Align @code{=>} in associations
14459 The @option{^-A^/ALIGN^} switches are mutually compatible; any combination
14463 @node Casing Control
14464 @subsection Casing Control
14465 @cindex Casing control in @command{gnatpp}
14468 @command{gnatpp} allows you to specify the casing for reserved words,
14469 pragma names, attribute designators and identifiers.
14470 For identifiers you may define a
14471 general rule for name casing but also override this rule
14472 via a set of dictionary files.
14474 Three types of casing are supported: lower case, upper case, and mixed case.
14475 Lower and upper case are self-explanatory (but since some letters in
14476 Latin1 and other GNAT-supported character sets
14477 exist only in lower-case form, an upper case conversion will have no
14479 ``Mixed case'' means that the first letter, and also each letter immediately
14480 following an underscore, are converted to their uppercase forms;
14481 all the other letters are converted to their lowercase forms.
14484 @cindex @option{^-a@var{x}^/ATTRIBUTE^} (@command{gnatpp})
14485 @item ^-aL^/ATTRIBUTE_CASING=LOWER_CASE^
14486 Attribute designators are lower case
14488 @item ^-aU^/ATTRIBUTE_CASING=UPPER_CASE^
14489 Attribute designators are upper case
14491 @item ^-aM^/ATTRIBUTE_CASING=MIXED_CASE^
14492 Attribute designators are mixed case (this is the default)
14494 @cindex @option{^-k@var{x}^/KEYWORD_CASING^} (@command{gnatpp})
14495 @item ^-kL^/KEYWORD_CASING=LOWER_CASE^
14496 Keywords (technically, these are known in Ada as @emph{reserved words}) are
14497 lower case (this is the default)
14499 @item ^-kU^/KEYWORD_CASING=UPPER_CASE^
14500 Keywords are upper case
14502 @cindex @option{^-n@var{x}^/NAME_CASING^} (@command{gnatpp})
14503 @item ^-nD^/NAME_CASING=AS_DECLARED^
14504 Name casing for defining occurrences are as they appear in the source file
14505 (this is the default)
14507 @item ^-nU^/NAME_CASING=UPPER_CASE^
14508 Names are in upper case
14510 @item ^-nL^/NAME_CASING=LOWER_CASE^
14511 Names are in lower case
14513 @item ^-nM^/NAME_CASING=MIXED_CASE^
14514 Names are in mixed case
14516 @cindex @option{^-p@var{x}^/PRAGMA_CASING^} (@command{gnatpp})
14517 @item ^-pL^/PRAGMA_CASING=LOWER_CASE^
14518 Pragma names are lower case
14520 @item ^-pU^/PRAGMA_CASING=UPPER_CASE^
14521 Pragma names are upper case
14523 @item ^-pM^/PRAGMA_CASING=MIXED_CASE^
14524 Pragma names are mixed case (this is the default)
14526 @item ^-D@var{file}^/DICTIONARY=@var{file}^
14527 @cindex @option{^-D^/DICTIONARY^} (@command{gnatpp})
14528 Use @var{file} as a @emph{dictionary file} that defines
14529 the casing for a set of specified names,
14530 thereby overriding the effect on these names by
14531 any explicit or implicit
14532 ^-n^/NAME_CASING^ switch.
14533 To supply more than one dictionary file,
14534 use ^several @option{-D} switches^a list of files as options^.
14537 @option{gnatpp} implicitly uses a @emph{default dictionary file}
14538 to define the casing for the Ada predefined names and
14539 the names declared in the GNAT libraries.
14541 @item ^-D-^/SPECIFIC_CASING^
14542 @cindex @option{^-D-^/SPECIFIC_CASING^} (@command{gnatpp})
14543 Do not use the default dictionary file;
14544 instead, use the casing
14545 defined by a @option{^-n^/NAME_CASING^} switch and any explicit
14550 The structure of a dictionary file, and details on the conventions
14551 used in the default dictionary file, are defined in @ref{Name Casing}.
14553 The @option{^-D-^/SPECIFIC_CASING^} and
14554 @option{^-D@var{file}^/DICTIONARY=@var{file}^} switches are mutually
14558 @node Construct Layout Control
14559 @subsection Construct Layout Control
14560 @cindex Layout control in @command{gnatpp}
14563 This group of @command{gnatpp} switches controls the layout of comments and
14564 complex syntactic constructs. See @ref{Formatting Comments}, for details
14568 @cindex @option{^-c@var{n}^/COMMENTS_LAYOUT^} (@command{gnatpp})
14569 @item ^-c1^/COMMENTS_LAYOUT=DEFAULT^
14570 GNAT-style comment line indentation (this is the default).
14572 @item ^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^
14573 Reference-manual comment line indentation.
14575 @item ^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^
14576 GNAT-style comment beginning
14578 @item ^-c4^/COMMENTS_LAYOUT=REFORMAT^
14579 Reformat comment blocks
14581 @cindex @option{^-l@var{n}^/CONSTRUCT_LAYOUT^} (@command{gnatpp})
14582 @item ^-l1^/CONSTRUCT_LAYOUT=GNAT^
14583 GNAT-style layout (this is the default)
14585 @item ^-l2^/CONSTRUCT_LAYOUT=COMPACT^
14588 @item ^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^
14591 @item ^-notab^/NOTABS^
14592 All the VT characters are removed from the comment text. All the HT characters are
14593 expanded with the sequences of space characters to get to the next tab stops.
14599 The @option{-c1} and @option{-c2} switches are incompatible.
14600 The @option{-c3} and @option{-c4} switches are compatible with each other and
14601 also with @option{-c1} and @option{-c2}.
14603 The @option{-l1}, @option{-l2}, and @option{-l3} switches are incompatible.
14608 For the @option{/COMMENTS_LAYOUT} qualifier:
14611 The @option{DEFAULT} and @option{STANDARD_INDENT} options are incompatible.
14613 The @option{GNAT_BEGINNING} and @option{REFORMAT} options are compatible with
14614 each other and also with @option{DEFAULT} and @option{STANDARD_INDENT}.
14618 The @option{GNAT}, @option{COMPACT}, and @option{UNCOMPACT} options for the
14619 @option{/CONSTRUCT_LAYOUT} qualifier are incompatible.
14622 @node General Text Layout Control
14623 @subsection General Text Layout Control
14626 These switches allow control over line length and indentation.
14629 @item ^-M@i{nnn}^/LINE_LENGTH_MAX=@i{nnn}^
14630 @cindex @option{^-M^/LINE_LENGTH^} (@command{gnatpp})
14631 Maximum line length, @i{nnn} from 32 ..256, the default value is 79
14633 @item ^-i@i{nnn}^/INDENTATION_LEVEL=@i{nnn}^
14634 @cindex @option{^-i^/INDENTATION_LEVEL^} (@command{gnatpp})
14635 Indentation level, @i{nnn} from 1 .. 9, the default value is 3
14637 @item ^-cl@i{nnn}^/CONTINUATION_INDENT=@i{nnn}^
14638 @cindex @option{^-cl^/CONTINUATION_INDENT^} (@command{gnatpp})
14639 Indentation level for continuation lines (relative to the line being
14640 continued), @i{nnn} from 1 .. 9.
14642 value is one less then the (normal) indentation level, unless the
14643 indentation is set to 1 (in which case the default value for continuation
14644 line indentation is also 1)
14648 @node Other Formatting Options
14649 @subsection Other Formatting Options
14652 These switches control the inclusion of missing end/exit labels, and
14653 the indentation level in @b{case} statements.
14656 @item ^-e^/NO_MISSED_LABELS^
14657 @cindex @option{^-e^/NO_MISSED_LABELS^} (@command{gnatpp})
14658 Do not insert missing end/exit labels. An end label is the name of
14659 a construct that may optionally be repeated at the end of the
14660 construct's declaration;
14661 e.g., the names of packages, subprograms, and tasks.
14662 An exit label is the name of a loop that may appear as target
14663 of an exit statement within the loop.
14664 By default, @command{gnatpp} inserts these end/exit labels when
14665 they are absent from the original source. This option suppresses such
14666 insertion, so that the formatted source reflects the original.
14668 @item ^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^
14669 @cindex @option{^-ff^/FORM_FEED_AFTER_PRAGMA_PAGE^} (@command{gnatpp})
14670 Insert a Form Feed character after a pragma Page.
14672 @item ^-T@i{nnn}^/MAX_INDENT=@i{nnn}^
14673 @cindex @option{^-T^/MAX_INDENT^} (@command{gnatpp})
14674 Do not use an additional indentation level for @b{case} alternatives
14675 and variants if there are @i{nnn} or more (the default
14677 If @i{nnn} is 0, an additional indentation level is
14678 used for @b{case} alternatives and variants regardless of their number.
14681 @node Setting the Source Search Path
14682 @subsection Setting the Source Search Path
14685 To define the search path for the input source file, @command{gnatpp}
14686 uses the same switches as the GNAT compiler, with the same effects.
14689 @item ^-I^/SEARCH=^@var{dir}
14690 @cindex @option{^-I^/SEARCH^} (@code{gnatpp})
14691 The same as the corresponding gcc switch
14693 @item ^-I-^/NOCURRENT_DIRECTORY^
14694 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatpp})
14695 The same as the corresponding gcc switch
14697 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE^=@var{path}
14698 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@code{gnatpp})
14699 The same as the corresponding gcc switch
14701 @item ^--RTS^/RUNTIME_SYSTEM^=@var{path}
14702 @cindex @option{^--RTS^/RUNTIME_SYSTEM^} (@code{gnatpp})
14703 The same as the corresponding gcc switch
14708 @node Output File Control
14709 @subsection Output File Control
14712 By default the output is sent to the file whose name is obtained by appending
14713 the ^@file{.pp}^@file{$PP}^ suffix to the name of the input file
14714 (if the file with this name already exists, it is unconditionally overwritten).
14715 Thus if the input file is @file{^my_ada_proc.adb^MY_ADA_PROC.ADB^} then
14716 @command{gnatpp} will produce @file{^my_ada_proc.adb.pp^MY_ADA_PROC.ADB$PP^}
14718 The output may be redirected by the following switches:
14721 @item ^-pipe^/STANDARD_OUTPUT^
14722 @cindex @option{^-pipe^/STANDARD_OUTPUT^} (@code{gnatpp})
14723 Send the output to @code{Standard_Output}
14725 @item ^-o @var{output_file}^/OUTPUT=@var{output_file}^
14726 @cindex @option{^-o^/OUTPUT^} (@code{gnatpp})
14727 Write the output into @var{output_file}.
14728 If @var{output_file} already exists, @command{gnatpp} terminates without
14729 reading or processing the input file.
14731 @item ^-of ^/FORCED_OUTPUT=^@var{output_file}
14732 @cindex @option{^-of^/FORCED_OUTPUT^} (@code{gnatpp})
14733 Write the output into @var{output_file}, overwriting the existing file
14734 (if one is present).
14736 @item ^-r^/REPLACE^
14737 @cindex @option{^-r^/REPLACE^} (@code{gnatpp})
14738 Replace the input source file with the reformatted output, and copy the
14739 original input source into the file whose name is obtained by appending the
14740 ^@file{.npp}^@file{$NPP}^ suffix to the name of the input file.
14741 If a file with this name already exists, @command{gnatpp} terminates without
14742 reading or processing the input file.
14744 @item ^-rf^/OVERRIDING_REPLACE^
14745 @cindex @option{^-rf^/OVERRIDING_REPLACE^} (@code{gnatpp})
14746 Like @option{^-r^/REPLACE^} except that if the file with the specified name
14747 already exists, it is overwritten.
14751 Options @option{^-pipe^/STANDARD_OUTPUT^},
14752 @option{^-o^/OUTPUT^} and
14753 @option{^-of^/FORCED_OUTPUT^} are allowed only if the call to gnatpp
14754 contains only one file to reformat
14756 @node Other gnatpp Switches
14757 @subsection Other @code{gnatpp} Switches
14760 The additional @command{gnatpp} switches are defined in this subsection.
14763 @item ^-v^/VERBOSE^
14764 @cindex @option{^-v^/VERBOSE^} (@code{gnatpp})
14766 @command{gnatpp} generates version information and then
14767 a trace of the actions it takes to produce or obtain the ASIS tree.
14769 @item ^-w^/WARNINGS^
14770 @cindex @option{^-w^/WARNINGS^} (@code{gnatpp})
14772 @command{gnatpp} generates a warning whenever it can not provide
14773 a required layout in the result source.
14777 @node Formatting Rules
14778 @section Formatting Rules
14781 The following subsections show how @command{gnatpp} treats ``white space'',
14782 comments, program layout, and name casing.
14783 They provide the detailed descriptions of the switches shown above.
14786 * White Space and Empty Lines::
14787 * Formatting Comments::
14788 * Construct Layout::
14793 @node White Space and Empty Lines
14794 @subsection White Space and Empty Lines
14797 @command{gnatpp} does not have an option to control space characters.
14798 It will add or remove spaces according to the style illustrated by the
14799 examples in the @cite{Ada Reference Manual}.
14801 The only format effectors
14802 (see @cite{Ada Reference Manual}, paragraph 2.1(13))
14803 that will appear in the output file are platform-specific line breaks,
14804 and also format effectors within (but not at the end of) comments.
14805 In particular, each horizontal tab character that is not inside
14806 a comment will be treated as a space and thus will appear in the
14807 output file as zero or more spaces depending on
14808 the reformatting of the line in which it appears.
14809 The only exception is a Form Feed character, which is inserted after a
14810 pragma @code{Page} when @option{-ff} is set.
14812 The output file will contain no lines with trailing ``white space'' (spaces,
14815 Empty lines in the original source are preserved
14816 only if they separate declarations or statements.
14817 In such contexts, a
14818 sequence of two or more empty lines is replaced by exactly one empty line.
14819 Note that a blank line will be removed if it separates two ``comment blocks''
14820 (a comment block is a sequence of whole-line comments).
14821 In order to preserve a visual separation between comment blocks, use an
14822 ``empty comment'' (a line comprising only hyphens) rather than an empty line.
14823 Likewise, if for some reason you wish to have a sequence of empty lines,
14824 use a sequence of empty comments instead.
14827 @node Formatting Comments
14828 @subsection Formatting Comments
14831 Comments in Ada code are of two kinds:
14834 a @emph{whole-line comment}, which appears by itself (possibly preceded by
14835 ``white space'') on a line
14838 an @emph{end-of-line comment}, which follows some other Ada lexical element
14843 The indentation of a whole-line comment is that of either
14844 the preceding or following line in
14845 the formatted source, depending on switch settings as will be described below.
14847 For an end-of-line comment, @command{gnatpp} leaves the same number of spaces
14848 between the end of the preceding Ada lexical element and the beginning
14849 of the comment as appear in the original source,
14850 unless either the comment has to be split to
14851 satisfy the line length limitation, or else the next line contains a
14852 whole line comment that is considered a continuation of this end-of-line
14853 comment (because it starts at the same position).
14855 cases, the start of the end-of-line comment is moved right to the nearest
14856 multiple of the indentation level.
14857 This may result in a ``line overflow'' (the right-shifted comment extending
14858 beyond the maximum line length), in which case the comment is split as
14861 There is a difference between @option{^-c1^/COMMENTS_LAYOUT=DEFAULT^}
14862 (GNAT-style comment line indentation)
14863 and @option{^-c2^/COMMENTS_LAYOUT=STANDARD_INDENT^}
14864 (reference-manual comment line indentation).
14865 With reference-manual style, a whole-line comment is indented as if it
14866 were a declaration or statement at the same place
14867 (i.e., according to the indentation of the preceding line(s)).
14868 With GNAT style, a whole-line comment that is immediately followed by an
14869 @b{if} or @b{case} statement alternative, a record variant, or the reserved
14870 word @b{begin}, is indented based on the construct that follows it.
14873 @smallexample @c ada
14885 Reference-manual indentation produces:
14887 @smallexample @c ada
14899 while GNAT-style indentation produces:
14901 @smallexample @c ada
14913 The @option{^-c3^/COMMENTS_LAYOUT=GNAT_BEGINNING^} switch
14914 (GNAT style comment beginning) has the following
14919 For each whole-line comment that does not end with two hyphens,
14920 @command{gnatpp} inserts spaces if necessary after the starting two hyphens
14921 to ensure that there are at least two spaces between these hyphens and the
14922 first non-blank character of the comment.
14926 For an end-of-line comment, if in the original source the next line is a
14927 whole-line comment that starts at the same position
14928 as the end-of-line comment,
14929 then the whole-line comment (and all whole-line comments
14930 that follow it and that start at the same position)
14931 will start at this position in the output file.
14934 That is, if in the original source we have:
14936 @smallexample @c ada
14939 A := B + C; -- B must be in the range Low1..High1
14940 -- C must be in the range Low2..High2
14941 --B+C will be in the range Low1+Low2..High1+High2
14947 Then in the formatted source we get
14949 @smallexample @c ada
14952 A := B + C; -- B must be in the range Low1..High1
14953 -- C must be in the range Low2..High2
14954 -- B+C will be in the range Low1+Low2..High1+High2
14960 A comment that exceeds the line length limit will be split.
14962 @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} (reformat comment blocks) is set and
14963 the line belongs to a reformattable block, splitting the line generates a
14964 @command{gnatpp} warning.
14965 The @option{^-c4^/COMMENTS_LAYOUT=REFORMAT^} switch specifies that whole-line
14966 comments may be reformatted in typical
14967 word processor style (that is, moving words between lines and putting as
14968 many words in a line as possible).
14971 @node Construct Layout
14972 @subsection Construct Layout
14975 The difference between GNAT style @option{^-l1^/CONSTRUCT_LAYOUT=GNAT^}
14976 and compact @option{^-l2^/CONSTRUCT_LAYOUT=COMPACT^}
14977 layout on the one hand, and uncompact layout
14978 @option{^-l3^/CONSTRUCT_LAYOUT=UNCOMPACT^} on the other hand,
14979 can be illustrated by the following examples:
14983 @multitable @columnfractions .5 .5
14984 @item @i{GNAT style, compact layout} @tab @i{Uncompact layout}
14987 @smallexample @c ada
14994 @smallexample @c ada
15003 @smallexample @c ada
15011 @smallexample @c ada
15021 @smallexample @c ada
15022 Clear : for J in 1 .. 10 loop
15027 @smallexample @c ada
15029 for J in 1 .. 10 loop
15040 GNAT style, compact layout Uncompact layout
15042 type q is record type q is
15043 a : integer; record
15044 b : integer; a : integer;
15045 end record; b : integer;
15049 Block : declare Block :
15050 A : Integer := 3; declare
15051 begin A : Integer := 3;
15053 end Block; Proc (A, A);
15056 Clear : for J in 1 .. 10 loop Clear :
15057 A (J) := 0; for J in 1 .. 10 loop
15058 end loop Clear; A (J) := 0;
15065 A further difference between GNAT style layout and compact layout is that
15066 GNAT style layout inserts empty lines as separation for
15067 compound statements, return statements and bodies.
15071 @subsection Name Casing
15074 @command{gnatpp} always converts the usage occurrence of a (simple) name to
15075 the same casing as the corresponding defining identifier.
15077 You control the casing for defining occurrences via the
15078 @option{^-n^/NAME_CASING^} switch.
15080 With @option{-nD} (``as declared'', which is the default),
15083 With @option{/NAME_CASING=AS_DECLARED}, which is the default,
15085 defining occurrences appear exactly as in the source file
15086 where they are declared.
15087 The other ^values for this switch^options for this qualifier^ ---
15088 @option{^-nU^UPPER_CASE^},
15089 @option{^-nL^LOWER_CASE^},
15090 @option{^-nM^MIXED_CASE^} ---
15092 ^upper, lower, or mixed case, respectively^the corresponding casing^.
15093 If @command{gnatpp} changes the casing of a defining
15094 occurrence, it analogously changes the casing of all the
15095 usage occurrences of this name.
15097 If the defining occurrence of a name is not in the source compilation unit
15098 currently being processed by @command{gnatpp}, the casing of each reference to
15099 this name is changed according to the value of the @option{^-n^/NAME_CASING^}
15100 switch (subject to the dictionary file mechanism described below).
15101 Thus @command{gnatpp} acts as though the @option{^-n^/NAME_CASING^} switch
15103 casing for the defining occurrence of the name.
15105 Some names may need to be spelled with casing conventions that are not
15106 covered by the upper-, lower-, and mixed-case transformations.
15107 You can arrange correct casing by placing such names in a
15108 @emph{dictionary file},
15109 and then supplying a @option{^-D^/DICTIONARY^} switch.
15110 The casing of names from dictionary files overrides
15111 any @option{^-n^/NAME_CASING^} switch.
15113 To handle the casing of Ada predefined names and the names from GNAT libraries,
15114 @command{gnatpp} assumes a default dictionary file.
15115 The name of each predefined entity is spelled with the same casing as is used
15116 for the entity in the @cite{Ada Reference Manual}.
15117 The name of each entity in the GNAT libraries is spelled with the same casing
15118 as is used in the declaration of that entity.
15120 The @w{@option{^-D-^/SPECIFIC_CASING^}} switch suppresses the use of the
15121 default dictionary file.
15122 Instead, the casing for predefined and GNAT-defined names will be established
15123 by the @option{^-n^/NAME_CASING^} switch or explicit dictionary files.
15124 For example, by default the names @code{Ada.Text_IO} and @code{GNAT.OS_Lib}
15125 will appear as just shown,
15126 even in the presence of a @option{^-nU^/NAME_CASING=UPPER_CASE^} switch.
15127 To ensure that even such names are rendered in uppercase,
15128 additionally supply the @w{@option{^-D-^/SPECIFIC_CASING^}} switch
15129 (or else, less conveniently, place these names in upper case in a dictionary
15132 A dictionary file is
15133 a plain text file; each line in this file can be either a blank line
15134 (containing only space characters and ASCII.HT characters), an Ada comment
15135 line, or the specification of exactly one @emph{casing schema}.
15137 A casing schema is a string that has the following syntax:
15141 @var{casing_schema} ::= @var{identifier} | [*]@var{simple_identifier}[*]
15143 @var{simple_identifier} ::= @var{letter}@{@var{letter_or_digit}@}
15148 (The @code{[]} metanotation stands for an optional part;
15149 see @cite{Ada Reference Manual}, Section 2.3) for the definition of the
15150 @var{identifier} lexical element and the @var{letter_or_digit} category).
15152 The casing schema string can be followed by white space and/or an Ada-style
15153 comment; any amount of white space is allowed before the string.
15155 If a dictionary file is passed as
15157 the value of a @option{-D@var{file}} switch
15160 an option to the @option{/DICTIONARY} qualifier
15163 simple name and every identifier, @command{gnatpp} checks if the dictionary
15164 defines the casing for the name or for some of its parts (the term ``subword''
15165 is used below to denote the part of a name which is delimited by ``_'' or by
15166 the beginning or end of the word and which does not contain any ``_'' inside):
15170 if the whole name is in the dictionary, @command{gnatpp} uses for this name
15171 the casing defined by the dictionary; no subwords are checked for this word
15174 for the first subword (that is, for the subword preceding the leftmost
15175 ``_''), @command{gnatpp} checks if the dictionary contains the corresponding
15176 string of the form @code{@var{simple_identifier}*}, and if it does, the
15177 casing of this @var{simple_identifier} is used for this subword
15180 for the last subword (following the rightmost ``_'') @command{gnatpp}
15181 checks if the dictionary contains the corresponding string of the form
15182 @code{*@var{simple_identifier}}, and if it does, the casing of this
15183 @var{simple_identifier} is used for this subword
15186 for every intermediate subword (surrounded by two'_') @command{gnatpp} checks
15187 if the dictionary contains the corresponding string of the form
15188 @code{*@var{simple_identifier}*}, and if it does, the casing of this
15189 simple_identifier is used for this subword
15192 if more than one dictionary file is passed as @command{gnatpp} switches, each
15193 dictionary adds new casing exceptions and overrides all the existing casing
15194 exceptions set by the previous dictionaries
15197 when @command{gnatpp} checks if the word or subword is in the dictionary,
15198 this check is not case sensitive
15202 For example, suppose we have the following source to reformat:
15204 @smallexample @c ada
15207 name1 : integer := 1;
15208 name4_name3_name2 : integer := 2;
15209 name2_name3_name4 : Boolean;
15212 name2_name3_name4 := name4_name3_name2 > name1;
15218 And suppose we have two dictionaries:
15235 If @command{gnatpp} is called with the following switches:
15239 @command{gnatpp -nM -D dict1 -D dict2 test.adb}
15242 @command{gnatpp test.adb /NAME_CASING=MIXED_CASE /DICTIONARY=(dict1, dict2)}
15247 then we will get the following name casing in the @command{gnatpp} output:
15249 @smallexample @c ada
15252 NAME1 : Integer := 1;
15253 Name4_NAME3_NAME2 : integer := 2;
15254 Name2_NAME3_Name4 : Boolean;
15257 Name2_NAME3_Name4 := Name4_NAME3_NAME2 > NAME1;
15264 @c ***********************************
15265 @node File Name Krunching Using gnatkr
15266 @chapter File Name Krunching Using @code{gnatkr}
15270 This chapter discusses the method used by the compiler to shorten
15271 the default file names chosen for Ada units so that they do not
15272 exceed the maximum length permitted. It also describes the
15273 @code{gnatkr} utility that can be used to determine the result of
15274 applying this shortening.
15278 * Krunching Method::
15279 * Examples of gnatkr Usage::
15283 @section About @code{gnatkr}
15286 The default file naming rule in GNAT
15287 is that the file name must be derived from
15288 the unit name. The exact default rule is as follows:
15291 Take the unit name and replace all dots by hyphens.
15293 If such a replacement occurs in the
15294 second character position of a name, and the first character is
15295 ^a, g, s, or i^A, G, S, or I^ then replace the dot by the character
15296 ^~ (tilde)^$ (dollar sign)^
15297 instead of a minus.
15299 The reason for this exception is to avoid clashes
15300 with the standard names for children of System, Ada, Interfaces,
15301 and GNAT, which use the prefixes ^s- a- i- and g-^S- A- I- and G-^
15304 The @option{^-gnatk^/FILE_NAME_MAX_LENGTH=^@var{nn}}
15305 switch of the compiler activates a ``krunching''
15306 circuit that limits file names to nn characters (where nn is a decimal
15307 integer). For example, using OpenVMS,
15308 where the maximum file name length is
15309 39, the value of nn is usually set to 39, but if you want to generate
15310 a set of files that would be usable if ported to a system with some
15311 different maximum file length, then a different value can be specified.
15312 The default value of 39 for OpenVMS need not be specified.
15314 The @code{gnatkr} utility can be used to determine the krunched name for
15315 a given file, when krunched to a specified maximum length.
15318 @section Using @code{gnatkr}
15321 The @code{gnatkr} command has the form
15325 $ gnatkr @var{name} [@var{length}]
15331 $ gnatkr @var{name} /COUNT=nn
15336 @var{name} is the uncrunched file name, derived from the name of the unit
15337 in the standard manner described in the previous section (i.e. in particular
15338 all dots are replaced by hyphens). The file name may or may not have an
15339 extension (defined as a suffix of the form period followed by arbitrary
15340 characters other than period). If an extension is present then it will
15341 be preserved in the output. For example, when krunching @file{hellofile.ads}
15342 to eight characters, the result will be hellofil.ads.
15344 Note: for compatibility with previous versions of @code{gnatkr} dots may
15345 appear in the name instead of hyphens, but the last dot will always be
15346 taken as the start of an extension. So if @code{gnatkr} is given an argument
15347 such as @file{Hello.World.adb} it will be treated exactly as if the first
15348 period had been a hyphen, and for example krunching to eight characters
15349 gives the result @file{hellworl.adb}.
15351 Note that the result is always all lower case (except on OpenVMS where it is
15352 all upper case). Characters of the other case are folded as required.
15354 @var{length} represents the length of the krunched name. The default
15355 when no argument is given is ^8^39^ characters. A length of zero stands for
15356 unlimited, in other words do not chop except for system files where the
15357 impled crunching length is always eight characters.
15360 The output is the krunched name. The output has an extension only if the
15361 original argument was a file name with an extension.
15363 @node Krunching Method
15364 @section Krunching Method
15367 The initial file name is determined by the name of the unit that the file
15368 contains. The name is formed by taking the full expanded name of the
15369 unit and replacing the separating dots with hyphens and
15370 using ^lowercase^uppercase^
15371 for all letters, except that a hyphen in the second character position is
15372 replaced by a ^tilde^dollar sign^ if the first character is
15373 ^a, i, g, or s^A, I, G, or S^.
15374 The extension is @code{.ads} for a
15375 specification and @code{.adb} for a body.
15376 Krunching does not affect the extension, but the file name is shortened to
15377 the specified length by following these rules:
15381 The name is divided into segments separated by hyphens, tildes or
15382 underscores and all hyphens, tildes, and underscores are
15383 eliminated. If this leaves the name short enough, we are done.
15386 If the name is too long, the longest segment is located (left-most
15387 if there are two of equal length), and shortened by dropping
15388 its last character. This is repeated until the name is short enough.
15390 As an example, consider the krunching of @*@file{our-strings-wide_fixed.adb}
15391 to fit the name into 8 characters as required by some operating systems.
15394 our-strings-wide_fixed 22
15395 our strings wide fixed 19
15396 our string wide fixed 18
15397 our strin wide fixed 17
15398 our stri wide fixed 16
15399 our stri wide fixe 15
15400 our str wide fixe 14
15401 our str wid fixe 13
15407 Final file name: oustwifi.adb
15411 The file names for all predefined units are always krunched to eight
15412 characters. The krunching of these predefined units uses the following
15413 special prefix replacements:
15417 replaced by @file{^a^A^-}
15420 replaced by @file{^g^G^-}
15423 replaced by @file{^i^I^-}
15426 replaced by @file{^s^S^-}
15429 These system files have a hyphen in the second character position. That
15430 is why normal user files replace such a character with a
15431 ^tilde^dollar sign^, to
15432 avoid confusion with system file names.
15434 As an example of this special rule, consider
15435 @*@file{ada-strings-wide_fixed.adb}, which gets krunched as follows:
15438 ada-strings-wide_fixed 22
15439 a- strings wide fixed 18
15440 a- string wide fixed 17
15441 a- strin wide fixed 16
15442 a- stri wide fixed 15
15443 a- stri wide fixe 14
15444 a- str wide fixe 13
15450 Final file name: a-stwifi.adb
15454 Of course no file shortening algorithm can guarantee uniqueness over all
15455 possible unit names, and if file name krunching is used then it is your
15456 responsibility to ensure that no name clashes occur. The utility
15457 program @code{gnatkr} is supplied for conveniently determining the
15458 krunched name of a file.
15460 @node Examples of gnatkr Usage
15461 @section Examples of @code{gnatkr} Usage
15468 $ gnatkr very_long_unit_name.ads --> velounna.ads
15469 $ gnatkr grandparent-parent-child.ads --> grparchi.ads
15470 $ gnatkr Grandparent.Parent.Child.ads --> grparchi.ads
15471 $ gnatkr grandparent-parent-child --> grparchi
15473 $ gnatkr very_long_unit_name.ads/count=6 --> vlunna.ads
15474 $ gnatkr very_long_unit_name.ads/count=0 --> very_long_unit_name.ads
15477 @node Preprocessing Using gnatprep
15478 @chapter Preprocessing Using @code{gnatprep}
15482 The @code{gnatprep} utility provides
15483 a simple preprocessing capability for Ada programs.
15484 It is designed for use with GNAT, but is not dependent on any special
15489 * Switches for gnatprep::
15490 * Form of Definitions File::
15491 * Form of Input Text for gnatprep::
15494 @node Using gnatprep
15495 @section Using @code{gnatprep}
15498 To call @code{gnatprep} use
15501 $ gnatprep [-bcrsu] [-Dsymbol=value] infile outfile [deffile]
15508 is the full name of the input file, which is an Ada source
15509 file containing preprocessor directives.
15512 is the full name of the output file, which is an Ada source
15513 in standard Ada form. When used with GNAT, this file name will
15514 normally have an ads or adb suffix.
15517 is the full name of a text file containing definitions of
15518 symbols to be referenced by the preprocessor. This argument is
15519 optional, and can be replaced by the use of the @option{-D} switch.
15522 is an optional sequence of switches as described in the next section.
15525 @node Switches for gnatprep
15526 @section Switches for @code{gnatprep}
15531 @item ^-b^/BLANK_LINES^
15532 @cindex @option{^-b^/BLANK_LINES^} (@command{gnatprep})
15533 Causes both preprocessor lines and the lines deleted by
15534 preprocessing to be replaced by blank lines in the output source file,
15535 preserving line numbers in the output file.
15537 @item ^-c^/COMMENTS^
15538 @cindex @option{^-c^/COMMENTS^} (@command{gnatprep})
15539 Causes both preprocessor lines and the lines deleted
15540 by preprocessing to be retained in the output source as comments marked
15541 with the special string @code{"--! "}. This option will result in line numbers
15542 being preserved in the output file.
15544 @item ^-Dsymbol=value^/ASSOCIATE="symbol=value"^
15545 @cindex @option{^-D^/ASSOCIATE^} (@command{gnatprep})
15546 Defines a new symbol, associated with value. If no value is given on the
15547 command line, then symbol is considered to be @code{True}. This switch
15548 can be used in place of a definition file.
15552 @cindex @option{/REMOVE} (@command{gnatprep})
15553 This is the default setting which causes lines deleted by preprocessing
15554 to be entirely removed from the output file.
15557 @item ^-r^/REFERENCE^
15558 @cindex @option{^-r^/REFERENCE^} (@command{gnatprep})
15559 Causes a @code{Source_Reference} pragma to be generated that
15560 references the original input file, so that error messages will use
15561 the file name of this original file. The use of this switch implies
15562 that preprocessor lines are not to be removed from the file, so its
15563 use will force @option{^-b^/BLANK_LINES^} mode if
15564 @option{^-c^/COMMENTS^}
15565 has not been specified explicitly.
15567 Note that if the file to be preprocessed contains multiple units, then
15568 it will be necessary to @code{gnatchop} the output file from
15569 @code{gnatprep}. If a @code{Source_Reference} pragma is present
15570 in the preprocessed file, it will be respected by
15571 @code{gnatchop ^-r^/REFERENCE^}
15572 so that the final chopped files will correctly refer to the original
15573 input source file for @code{gnatprep}.
15575 @item ^-s^/SYMBOLS^
15576 @cindex @option{^-s^/SYMBOLS^} (@command{gnatprep})
15577 Causes a sorted list of symbol names and values to be
15578 listed on the standard output file.
15580 @item ^-u^/UNDEFINED^
15581 @cindex @option{^-u^/UNDEFINED^} (@command{gnatprep})
15582 Causes undefined symbols to be treated as having the value FALSE in the context
15583 of a preprocessor test. In the absence of this option, an undefined symbol in
15584 a @code{#if} or @code{#elsif} test will be treated as an error.
15590 Note: if neither @option{-b} nor @option{-c} is present,
15591 then preprocessor lines and
15592 deleted lines are completely removed from the output, unless -r is
15593 specified, in which case -b is assumed.
15596 @node Form of Definitions File
15597 @section Form of Definitions File
15600 The definitions file contains lines of the form
15607 where symbol is an identifier, following normal Ada (case-insensitive)
15608 rules for its syntax, and value is one of the following:
15612 Empty, corresponding to a null substitution
15614 A string literal using normal Ada syntax
15616 Any sequence of characters from the set
15617 (letters, digits, period, underline).
15621 Comment lines may also appear in the definitions file, starting with
15622 the usual @code{--},
15623 and comments may be added to the definitions lines.
15625 @node Form of Input Text for gnatprep
15626 @section Form of Input Text for @code{gnatprep}
15629 The input text may contain preprocessor conditional inclusion lines,
15630 as well as general symbol substitution sequences.
15632 The preprocessor conditional inclusion commands have the form
15637 #if @i{expression} [then]
15639 #elsif @i{expression} [then]
15641 #elsif @i{expression} [then]
15652 In this example, @i{expression} is defined by the following grammar:
15654 @i{expression} ::= <symbol>
15655 @i{expression} ::= <symbol> = "<value>"
15656 @i{expression} ::= <symbol> = <symbol>
15657 @i{expression} ::= <symbol> 'Defined
15658 @i{expression} ::= not @i{expression}
15659 @i{expression} ::= @i{expression} and @i{expression}
15660 @i{expression} ::= @i{expression} or @i{expression}
15661 @i{expression} ::= @i{expression} and then @i{expression}
15662 @i{expression} ::= @i{expression} or else @i{expression}
15663 @i{expression} ::= ( @i{expression} )
15667 For the first test (@i{expression} ::= <symbol>) the symbol must have
15668 either the value true or false, that is to say the right-hand of the
15669 symbol definition must be one of the (case-insensitive) literals
15670 @code{True} or @code{False}. If the value is true, then the
15671 corresponding lines are included, and if the value is false, they are
15674 The test (@i{expression} ::= <symbol> @code{'Defined}) is true only if
15675 the symbol has been defined in the definition file or by a @option{-D}
15676 switch on the command line. Otherwise, the test is false.
15678 The equality tests are case insensitive, as are all the preprocessor lines.
15680 If the symbol referenced is not defined in the symbol definitions file,
15681 then the effect depends on whether or not switch @option{-u}
15682 is specified. If so, then the symbol is treated as if it had the value
15683 false and the test fails. If this switch is not specified, then
15684 it is an error to reference an undefined symbol. It is also an error to
15685 reference a symbol that is defined with a value other than @code{True}
15688 The use of the @code{not} operator inverts the sense of this logical test, so
15689 that the lines are included only if the symbol is not defined.
15690 The @code{then} keyword is optional as shown
15692 The @code{#} must be the first non-blank character on a line, but
15693 otherwise the format is free form. Spaces or tabs may appear between
15694 the @code{#} and the keyword. The keywords and the symbols are case
15695 insensitive as in normal Ada code. Comments may be used on a
15696 preprocessor line, but other than that, no other tokens may appear on a
15697 preprocessor line. Any number of @code{elsif} clauses can be present,
15698 including none at all. The @code{else} is optional, as in Ada.
15700 The @code{#} marking the start of a preprocessor line must be the first
15701 non-blank character on the line, i.e. it must be preceded only by
15702 spaces or horizontal tabs.
15704 Symbol substitution outside of preprocessor lines is obtained by using
15712 anywhere within a source line, except in a comment or within a
15713 string literal. The identifier
15714 following the @code{$} must match one of the symbols defined in the symbol
15715 definition file, and the result is to substitute the value of the
15716 symbol in place of @code{$symbol} in the output file.
15718 Note that although the substitution of strings within a string literal
15719 is not possible, it is possible to have a symbol whose defined value is
15720 a string literal. So instead of setting XYZ to @code{hello} and writing:
15723 Header : String := "$XYZ";
15727 you should set XYZ to @code{"hello"} and write:
15730 Header : String := $XYZ;
15734 and then the substitution will occur as desired.
15737 @node The GNAT Run-Time Library Builder gnatlbr
15738 @chapter The GNAT Run-Time Library Builder @code{gnatlbr}
15740 @cindex Library builder
15743 @code{gnatlbr} is a tool for rebuilding the GNAT run time with user
15744 supplied configuration pragmas.
15747 * Running gnatlbr::
15748 * Switches for gnatlbr::
15749 * Examples of gnatlbr Usage::
15752 @node Running gnatlbr
15753 @section Running @code{gnatlbr}
15756 The @code{gnatlbr} command has the form
15759 $ GNAT LIBRARY /[CREATE | SET | DELETE]=directory [/CONFIG=file]
15762 @node Switches for gnatlbr
15763 @section Switches for @code{gnatlbr}
15766 @code{gnatlbr} recognizes the following switches:
15770 @item /CREATE=directory
15771 @cindex @code{/CREATE} (@code{gnatlbr})
15772 Create the new run-time library in the specified directory.
15774 @item /SET=directory
15775 @cindex @code{/SET} (@code{gnatlbr})
15776 Make the library in the specified directory the current run-time
15779 @item /DELETE=directory
15780 @cindex @code{/DELETE} (@code{gnatlbr})
15781 Delete the run-time library in the specified directory.
15784 @cindex @code{/CONFIG} (@code{gnatlbr})
15786 Use the configuration pragmas in the specified file when building
15790 Use the configuration pragmas in the specified file when compiling.
15794 @node Examples of gnatlbr Usage
15795 @section Example of @code{gnatlbr} Usage
15798 Contents of VAXFLOAT.ADC:
15799 pragma Float_Representation (VAX_Float);
15801 $ GNAT LIBRARY /CREATE=[.VAXFLOAT] /CONFIG=VAXFLOAT.ADC
15803 GNAT LIBRARY rebuilds the run-time library in directory [.VAXFLOAT]
15808 @node The GNAT Library Browser gnatls
15809 @chapter The GNAT Library Browser @code{gnatls}
15811 @cindex Library browser
15814 @code{gnatls} is a tool that outputs information about compiled
15815 units. It gives the relationship between objects, unit names and source
15816 files. It can also be used to check the source dependencies of a unit
15817 as well as various characteristics.
15821 * Switches for gnatls::
15822 * Examples of gnatls Usage::
15825 @node Running gnatls
15826 @section Running @code{gnatls}
15829 The @code{gnatls} command has the form
15832 $ gnatls switches @var{object_or_ali_file}
15836 The main argument is the list of object or @file{ali} files
15837 (@pxref{The Ada Library Information Files})
15838 for which information is requested.
15840 In normal mode, without additional option, @code{gnatls} produces a
15841 four-column listing. Each line represents information for a specific
15842 object. The first column gives the full path of the object, the second
15843 column gives the name of the principal unit in this object, the third
15844 column gives the status of the source and the fourth column gives the
15845 full path of the source representing this unit.
15846 Here is a simple example of use:
15850 ^./^[]^demo1.o demo1 DIF demo1.adb
15851 ^./^[]^demo2.o demo2 OK demo2.adb
15852 ^./^[]^hello.o h1 OK hello.adb
15853 ^./^[]^instr-child.o instr.child MOK instr-child.adb
15854 ^./^[]^instr.o instr OK instr.adb
15855 ^./^[]^tef.o tef DIF tef.adb
15856 ^./^[]^text_io_example.o text_io_example OK text_io_example.adb
15857 ^./^[]^tgef.o tgef DIF tgef.adb
15861 The first line can be interpreted as follows: the main unit which is
15863 object file @file{demo1.o} is demo1, whose main source is in
15864 @file{demo1.adb}. Furthermore, the version of the source used for the
15865 compilation of demo1 has been modified (DIF). Each source file has a status
15866 qualifier which can be:
15869 @item OK (unchanged)
15870 The version of the source file used for the compilation of the
15871 specified unit corresponds exactly to the actual source file.
15873 @item MOK (slightly modified)
15874 The version of the source file used for the compilation of the
15875 specified unit differs from the actual source file but not enough to
15876 require recompilation. If you use gnatmake with the qualifier
15877 @option{^-m (minimal recompilation)^/MINIMAL_RECOMPILATION^}, a file marked
15878 MOK will not be recompiled.
15880 @item DIF (modified)
15881 No version of the source found on the path corresponds to the source
15882 used to build this object.
15884 @item ??? (file not found)
15885 No source file was found for this unit.
15887 @item HID (hidden, unchanged version not first on PATH)
15888 The version of the source that corresponds exactly to the source used
15889 for compilation has been found on the path but it is hidden by another
15890 version of the same source that has been modified.
15894 @node Switches for gnatls
15895 @section Switches for @code{gnatls}
15898 @code{gnatls} recognizes the following switches:
15902 @item ^-a^/ALL_UNITS^
15903 @cindex @option{^-a^/ALL_UNITS^} (@code{gnatls})
15904 Consider all units, including those of the predefined Ada library.
15905 Especially useful with @option{^-d^/DEPENDENCIES^}.
15907 @item ^-d^/DEPENDENCIES^
15908 @cindex @option{^-d^/DEPENDENCIES^} (@code{gnatls})
15909 List sources from which specified units depend on.
15911 @item ^-h^/OUTPUT=OPTIONS^
15912 @cindex @option{^-h^/OUTPUT=OPTIONS^} (@code{gnatls})
15913 Output the list of options.
15915 @item ^-o^/OUTPUT=OBJECTS^
15916 @cindex @option{^-o^/OUTPUT=OBJECTS^} (@code{gnatls})
15917 Only output information about object files.
15919 @item ^-s^/OUTPUT=SOURCES^
15920 @cindex @option{^-s^/OUTPUT=SOURCES^} (@code{gnatls})
15921 Only output information about source files.
15923 @item ^-u^/OUTPUT=UNITS^
15924 @cindex @option{^-u^/OUTPUT=UNITS^} (@code{gnatls})
15925 Only output information about compilation units.
15927 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
15928 @itemx ^-aI^/SOURCE_SEARCH=^@var{dir}
15929 @itemx ^-I^/SEARCH=^@var{dir}
15930 @itemx ^-I-^/NOCURRENT_DIRECTORY^
15932 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatls})
15933 @cindex @option{^-aI^/SOURCE_SEARCH^} (@code{gnatls})
15934 @cindex @option{^-I^/SEARCH^} (@code{gnatls})
15935 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatls})
15936 Source path manipulation. Same meaning as the equivalent @code{gnatmake} flags
15937 (see @ref{Switches for gnatmake}).
15939 @item --RTS=@var{rts-path}
15940 @cindex @option{--RTS} (@code{gnatls})
15941 Specifies the default location of the runtime library. Same meaning as the
15942 equivalent @code{gnatmake} flag (see @ref{Switches for gnatmake}).
15944 @item ^-v^/OUTPUT=VERBOSE^
15945 @cindex @option{^-v^/OUTPUT=VERBOSE^} (@code{gnatls})
15946 Verbose mode. Output the complete source and object paths. Do not use
15947 the default column layout but instead use long format giving as much as
15948 information possible on each requested units, including special
15949 characteristics such as:
15952 @item Preelaborable
15953 The unit is preelaborable in the Ada 95 sense.
15956 No elaboration code has been produced by the compiler for this unit.
15959 The unit is pure in the Ada 95 sense.
15961 @item Elaborate_Body
15962 The unit contains a pragma Elaborate_Body.
15965 The unit contains a pragma Remote_Types.
15967 @item Shared_Passive
15968 The unit contains a pragma Shared_Passive.
15971 This unit is part of the predefined environment and cannot be modified
15974 @item Remote_Call_Interface
15975 The unit contains a pragma Remote_Call_Interface.
15981 @node Examples of gnatls Usage
15982 @section Example of @code{gnatls} Usage
15986 Example of using the verbose switch. Note how the source and
15987 object paths are affected by the -I switch.
15990 $ gnatls -v -I.. demo1.o
15992 GNATLS 3.10w (970212) Copyright 1999 Free Software Foundation, Inc.
15994 Source Search Path:
15995 <Current_Directory>
15997 /home/comar/local/adainclude/
15999 Object Search Path:
16000 <Current_Directory>
16002 /home/comar/local/lib/gcc-lib/mips-sni-sysv4/2.7.2/adalib/
16007 Kind => subprogram body
16008 Flags => No_Elab_Code
16009 Source => demo1.adb modified
16013 The following is an example of use of the dependency list.
16014 Note the use of the -s switch
16015 which gives a straight list of source files. This can be useful for
16016 building specialized scripts.
16019 $ gnatls -d demo2.o
16020 ./demo2.o demo2 OK demo2.adb
16026 $ gnatls -d -s -a demo1.o
16028 /home/comar/local/adainclude/ada.ads
16029 /home/comar/local/adainclude/a-finali.ads
16030 /home/comar/local/adainclude/a-filico.ads
16031 /home/comar/local/adainclude/a-stream.ads
16032 /home/comar/local/adainclude/a-tags.ads
16035 /home/comar/local/adainclude/gnat.ads
16036 /home/comar/local/adainclude/g-io.ads
16038 /home/comar/local/adainclude/system.ads
16039 /home/comar/local/adainclude/s-exctab.ads
16040 /home/comar/local/adainclude/s-finimp.ads
16041 /home/comar/local/adainclude/s-finroo.ads
16042 /home/comar/local/adainclude/s-secsta.ads
16043 /home/comar/local/adainclude/s-stalib.ads
16044 /home/comar/local/adainclude/s-stoele.ads
16045 /home/comar/local/adainclude/s-stratt.ads
16046 /home/comar/local/adainclude/s-tasoli.ads
16047 /home/comar/local/adainclude/s-unstyp.ads
16048 /home/comar/local/adainclude/unchconv.ads
16054 GNAT LIST /DEPENDENCIES /OUTPUT=SOURCES /ALL_UNITS DEMO1.ADB
16056 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]ada.ads
16057 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-finali.ads
16058 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-filico.ads
16059 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-stream.ads
16060 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]a-tags.ads
16064 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]gnat.ads
16065 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]g-io.ads
16067 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]system.ads
16068 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-exctab.ads
16069 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finimp.ads
16070 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-finroo.ads
16071 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-secsta.ads
16072 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stalib.ads
16073 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stoele.ads
16074 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-stratt.ads
16075 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-tasoli.ads
16076 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]s-unstyp.ads
16077 GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB]unchconv.ads
16081 @node Cleaning Up Using gnatclean
16082 @chapter Cleaning Up Using @code{gnatclean}
16084 @cindex Cleaning tool
16087 @code{gnatclean} is a tool that allows the deletion of files produced by the
16088 compiler, binder and linker, including ALI files, object files, tree files,
16089 expanded source files, library files, interface copy source files, binder
16090 generated files and executable files.
16093 * Running gnatclean::
16094 * Switches for gnatclean::
16095 * Examples of gnatclean Usage::
16098 @node Running gnatclean
16099 @section Running @code{gnatclean}
16102 The @code{gnatclean} command has the form:
16105 $ gnatclean switches @var{names}
16109 @var{names} is a list of source file names. Suffixes @code{.^ads^ADS^} and
16110 @code{^adb^ADB^} may be omitted. If a project file is specified using switch
16111 @code{^-P^/PROJECT_FILE=^}, then @var{names} may be completely omitted.
16114 In normal mode, @code{gnatclean} delete the files produced by the compiler and,
16115 if switch @code{^-c^/COMPILER_FILES_ONLY^} is not specified, by the binder and
16116 the linker. In informative-only mode, specified by switch
16117 @code{^-n^/NODELETE^}, the list of files that would have been deleted in
16118 normal mode is listed, but no file is actually deleted.
16120 @node Switches for gnatclean
16121 @section Switches for @code{gnatclean}
16124 @code{gnatclean} recognizes the following switches:
16128 @item ^-c^/COMPILER_FILES_ONLY^
16129 @cindex @option{^-c^/COMPILER_FILES_ONLY^} (@code{gnatclean})
16130 Only attempt to delete the files produced by the compiler, not those produced
16131 by the binder or the linker. The files that are not to be deleted are library
16132 files, interface copy files, binder generated files and executable files.
16134 @item ^-D ^/DIRECTORY_OBJECTS=^@var{dir}
16135 @cindex @option{^-D^/DIRECTORY_OBJECTS^} (@code{gnatclean})
16136 Indicate that ALI and object files should normally be found in directory
16139 @item ^-F^/FULL_PATH_IN_BRIEF_MESSAGES^
16140 @cindex @option{^-F^/FULL_PATH_IN_BRIEF_MESSAGES^} (@code{gnatclean})
16141 When using project files, if some errors or warnings are detected during
16142 parsing and verbose mode is not in effect (no use of switch
16143 ^-v^/VERBOSE^), then error lines start with the full path name of the project
16144 file, rather than its simple file name.
16147 @cindex @option{^-h^/HELP^} (@code{gnatclean})
16148 Output a message explaining the usage of @code{^gnatclean^gnatclean^}.
16150 @item ^-n^/NODELETE^
16151 @cindex @option{^-n^/NODELETE^} (@code{gnatclean})
16152 Informative-only mode. Do not delete any files. Output the list of the files
16153 that would have been deleted if this switch was not specified.
16155 @item ^-P^/PROJECT_FILE=^@var{project}
16156 @cindex @option{^-P^/PROJECT_FILE^} (@code{gnatclean})
16157 Use project file @var{project}. Only one such switch can be used.
16158 When cleaning a project file, the files produced by the compilation of the
16159 immediate sources or inherited sources of the project files are to be
16160 deleted. This is not depending on the presence or not of executable names
16161 on the command line.
16164 @cindex @option{^-q^/QUIET^} (@code{gnatclean})
16165 Quiet output. If there are no error, do not ouuput anything, except in
16166 verbose mode (switch ^-v^/VERBOSE^) or in informative-only mode
16167 (switch ^-n^/NODELETE^).
16169 @item ^-r^/RECURSIVE^
16170 @cindex @option{^-r^/RECURSIVE^} (@code{gnatclean})
16171 When a project file is specified (using switch ^-P^/PROJECT_FILE=^),
16172 clean all imported and extended project files, recursively. If this switch
16173 is not specified, only the files related to the main project file are to be
16174 deleted. This switch has no effect if no project file is specified.
16176 @item ^-v^/VERBOSE^
16177 @cindex @option{^-v^/VERBOSE^} (@code{gnatclean})
16180 @item ^-vP^/MESSAGES_PROJECT_FILE=^@emph{x}
16181 @cindex @option{^-vP^/MESSAGES_PROJECT_FILE^} (@code{gnatclean})
16182 Indicates the verbosity of the parsing of GNAT project files.
16183 See @ref{Switches Related to Project Files}.
16185 @item ^-X^/EXTERNAL_REFERENCE=^@var{name=value}
16186 @cindex @option{^-X^/EXTERNAL_REFERENCE^} (@code{gnatclean})
16187 Indicates that external variable @var{name} has the value @var{value}.
16188 The Project Manager will use this value for occurrences of
16189 @code{external(name)} when parsing the project file.
16190 See @ref{Switches Related to Project Files}.
16192 @item ^-aO^/OBJECT_SEARCH=^@var{dir}
16193 @cindex @option{^-aO^/OBJECT_SEARCH^} (@code{gnatclean})
16194 When searching for ALI and object files, look in directory
16197 @item ^-I^/SEARCH=^@var{dir}
16198 @cindex @option{^-I^/SEARCH^} (@code{gnatclean})
16199 Equivalent to @option{^-aO^/OBJECT_SEARCH=^@var{dir}}.
16201 @item ^-I-^/NOCURRENT_DIRECTORY^
16202 @cindex @option{^-I-^/NOCURRENT_DIRECTORY^} (@code{gnatclean})
16203 @cindex Source files, suppressing search
16204 Do not look for ALI or object files in the directory
16205 where @code{gnatclean} was invoked.
16209 @node Examples of gnatclean Usage
16210 @section Examples of @code{gnatclean} Usage
16213 @node GNAT and Libraries
16214 @chapter GNAT and Libraries
16215 @cindex Library, building, installing
16218 This chapter addresses some of the issues related to building and using
16219 a library with GNAT. It also shows how the GNAT run-time library can be
16223 * Creating an Ada Library::
16224 * Installing an Ada Library::
16225 * Using an Ada Library::
16226 * Creating an Ada Library to be Used in a Non-Ada Context::
16227 * Rebuilding the GNAT Run-Time Library::
16230 @node Creating an Ada Library
16231 @section Creating an Ada Library
16234 In the GNAT environment, a library has two components:
16239 Compiled code and Ali files. See @ref{The Ada Library Information Files}.
16243 In order to use other packages @ref{The GNAT Compilation Model}
16244 requires a certain number of sources to be available to the compiler.
16246 sources required includes the specs of all the packages that make up the
16247 visible part of the library as well as all the sources upon which they
16248 depend. The bodies of all visible generic units must also be provided.
16250 Although it is not strictly mandatory, it is recommended that all sources
16251 needed to recompile the library be provided, so that the user can make
16252 full use of inter-unit inlining and source-level debugging. This can also
16253 make the situation easier for users that need to upgrade their compilation
16254 toolchain and thus need to recompile the library from sources.
16257 The compiled code can be provided in different ways. The simplest way is
16258 to provide directly the set of objects produced by the compiler during
16259 the compilation of the library. It is also possible to group the objects
16260 into an archive using whatever commands are provided by the operating
16261 system. Finally, it is also possible to create a shared library (see
16262 option -shared in the GCC manual).
16265 There are various possibilities for compiling the units that make up the
16266 library: for example with a Makefile @ref{Using the GNU make Utility},
16267 or with a conventional script.
16268 For simple libraries, it is also possible to create a
16269 dummy main program which depends upon all the packages that comprise the
16270 interface of the library. This dummy main program can then be given to
16271 gnatmake, in order to build all the necessary objects. Here is an example
16272 of such a dummy program and the generic commands used to build an
16273 archive or a shared library.
16275 @smallexample @c ada
16279 with My_Lib.Service1;
16280 with My_Lib.Service2;
16281 with My_Lib.Service3;
16282 procedure My_Lib_Dummy is
16289 # compiling the library
16290 $ gnatmake -c my_lib_dummy.adb
16292 # we don't need the dummy object itself
16293 $ rm my_lib_dummy.o my_lib_dummy.ali
16295 # create an archive with the remaining objects
16296 $ ar rc libmy_lib.a *.o
16297 # some systems may require "ranlib" to be run as well
16299 # or create a shared library
16300 $ gcc -shared -o libmy_lib.so *.o
16301 # some systems may require the code to have been compiled with -fPIC
16303 # remove the object files that are now in the library
16306 # Make the ALI files read-only so that gnatmake will not try to
16307 # regenerate the objects that are in the library
16313 When the objects are grouped in an archive or a shared library, the user
16314 needs to specify the desired library at link time, unless a pragma
16315 linker_options has been used in one of the sources:
16316 @smallexample @c ada
16317 pragma Linker_Options ("-lmy_lib");
16321 Please note that the library must have a name of the form libxxx.a or
16322 libxxx.so in order to be accessed by the directive -lxxx at link
16325 @node Installing an Ada Library
16326 @section Installing an Ada Library
16329 In the GNAT model, installing a library consists in copying into a specific
16330 location the files that make up this library. It is possible to install
16331 the sources in a different directory from the other files (ALI, objects,
16332 archives) since the source path and the object path can easily be
16333 specified separately.
16336 For general purpose libraries, it is possible for the system
16337 administrator to put those libraries in the default compiler paths. To
16338 achieve this, he must specify their location in the configuration files
16339 @file{ada_source_path} and @file{ada_object_path} that must be located in
16341 installation tree at the same place as the gcc spec file. The location of
16342 the gcc spec file can be determined as follows:
16348 The configuration files mentioned above have simple format: each line in them
16349 must contain one unique
16350 directory name. Those names are added to the corresponding path
16351 in their order of appearance in the file. The names can be either absolute
16352 or relative, in the latter case, they are relative to where theses files
16356 @file{ada_source_path} and @file{ada_object_path} might actually not be
16358 GNAT installation, in which case, GNAT will look for its run-time library in
16359 he directories @file{adainclude} for the sources and @file{adalib} for the
16360 objects and @file{ALI} files. When the files exist, the compiler does not
16361 look in @file{adainclude} and @file{adalib} at all, and thus the
16362 @file{ada_source_path} file
16363 must contain the location for the GNAT run-time sources (which can simply
16364 be @file{adainclude}). In the same way, the @file{ada_object_path} file must
16365 contain the location for the GNAT run-time objects (which can simply
16369 You can also specify a new default path to the runtime library at compilation
16370 time with the switch @option{--RTS=rts-path}. You can easily choose and change
16371 the runtime you want your program to be compiled with. This switch is
16372 recognized by gcc, gnatmake, gnatbind, gnatls, gnatfind and gnatxref.
16375 It is possible to install a library before or after the standard GNAT
16376 library, by reordering the lines in the configuration files. In general, a
16377 library must be installed before the GNAT library if it redefines
16380 @node Using an Ada Library
16381 @section Using an Ada Library
16384 In order to use a Ada library, you need to make sure that this
16385 library is on both your source and object path
16386 @ref{Search Paths and the Run-Time Library (RTL)}
16387 and @ref{Search Paths for gnatbind}. For
16388 instance, you can use the library @file{mylib} installed in
16389 @file{/dir/my_lib_src} and @file{/dir/my_lib_obj} with the following commands:
16392 $ gnatmake -aI/dir/my_lib_src -aO/dir/my_lib_obj my_appl \
16397 This can be simplified down to the following:
16401 when the following conditions are met:
16404 @file{/dir/my_lib_src} has been added by the user to the environment
16405 variable @code{ADA_INCLUDE_PATH}, or by the administrator to the file
16406 @file{ada_source_path}
16408 @file{/dir/my_lib_obj} has been added by the user to the environment
16409 variable @code{ADA_OBJECTS_PATH}, or by the administrator to the file
16410 @file{ada_object_path}
16412 a pragma @code{Linker_Options}, as mentioned in @ref{Creating an Ada Library},
16413 has been added to the sources.
16417 @node Creating an Ada Library to be Used in a Non-Ada Context
16418 @section Creating an Ada Library to be Used in a Non-Ada Context
16421 The previous sections detailed how to create and install a library that
16422 was usable from an Ada main program. Using this library in a non-Ada
16423 context is not possible, because the elaboration of the library is
16424 automatically done as part of the main program elaboration.
16426 GNAT also provides the ability to build libraries that can be used both
16427 in an Ada and non-Ada context. This section describes how to build such
16428 a library, and then how to use it from a C program. The method for
16429 interfacing with the library from other languages such as Fortran for
16430 instance remains the same.
16432 @subsection Creating the Library
16435 @item Identify the units representing the interface of the library.
16437 Here is an example of simple library interface:
16439 @smallexample @c ada
16440 package Interface is
16442 procedure Do_Something;
16444 procedure Do_Something_Else;
16449 @item Use @code{pragma Export} or @code{pragma Convention} for the
16452 Our package @code{Interface} is then updated as follow:
16453 @smallexample @c ada
16454 package Interface is
16456 procedure Do_Something;
16457 pragma Export (C, Do_Something, "do_something");
16459 procedure Do_Something_Else;
16460 pragma Export (C, Do_Something_Else, "do_something_else");
16465 @item Compile all the units composing the library.
16467 @item Bind the library objects.
16469 This step is performed by invoking gnatbind with the @option{-L<prefix>}
16470 switch. @code{gnatbind} will then generate the library elaboration
16471 procedure (named @code{<prefix>init}) and the run-time finalization
16472 procedure (named @code{<prefix>final}).
16475 # generate the binder file in Ada
16476 $ gnatbind -Lmylib interface
16478 # generate the binder file in C
16479 $ gnatbind -C -Lmylib interface
16482 @item Compile the files generated by the binder
16485 $ gcc -c b~interface.adb
16488 @item Create the library;
16490 The procedure is identical to the procedure explained in
16491 @ref{Creating an Ada Library},
16492 except that @file{b~interface.o} needs to be added to
16493 the list of objects.
16496 # create an archive file
16497 $ ar cr libmylib.a b~interface.o <other object files>
16499 # create a shared library
16500 $ gcc -shared -o libmylib.so b~interface.o <other object files>
16503 @item Provide a ``foreign'' view of the library interface;
16505 The example below shows the content of @code{mylib_interface.h} (note
16506 that there is no rule for the naming of this file, any name can be used)
16508 /* the library elaboration procedure */
16509 extern void mylibinit (void);
16511 /* the library finalization procedure */
16512 extern void mylibfinal (void);
16514 /* the interface exported by the library */
16515 extern void do_something (void);
16516 extern void do_something_else (void);
16520 @subsection Using the Library
16523 Libraries built as explained above can be used from any program, provided
16524 that the elaboration procedures (named @code{mylibinit} in the previous
16525 example) are called before the library services are used. Any number of
16526 libraries can be used simultaneously, as long as the elaboration
16527 procedure of each library is called.
16529 Below is an example of C program that uses our @code{mylib} library.
16532 #include "mylib_interface.h"
16537 /* First, elaborate the library before using it */
16540 /* Main program, using the library exported entities */
16542 do_something_else ();
16544 /* Library finalization at the end of the program */
16551 Note that this same library can be used from an equivalent Ada main
16552 program. In addition, if the libraries are installed as detailed in
16553 @ref{Installing an Ada Library}, it is not necessary to invoke the
16554 library elaboration and finalization routines. The binder will ensure
16555 that this is done as part of the main program elaboration and
16556 finalization phases.
16558 @subsection The Finalization Phase
16561 Invoking any library finalization procedure generated by @code{gnatbind}
16562 shuts down the Ada run time permanently. Consequently, the finalization
16563 of all Ada libraries must be performed at the end of the program. No
16564 call to these libraries nor the Ada run time should be made past the
16565 finalization phase.
16567 @subsection Restrictions in Libraries
16570 The pragmas listed below should be used with caution inside libraries,
16571 as they can create incompatibilities with other Ada libraries:
16573 @item pragma @code{Locking_Policy}
16574 @item pragma @code{Queuing_Policy}
16575 @item pragma @code{Task_Dispatching_Policy}
16576 @item pragma @code{Unreserve_All_Interrupts}
16578 When using a library that contains such pragmas, the user must make sure
16579 that all libraries use the same pragmas with the same values. Otherwise,
16580 a @code{Program_Error} will
16581 be raised during the elaboration of the conflicting
16582 libraries. The usage of these pragmas and its consequences for the user
16583 should therefore be well documented.
16585 Similarly, the traceback in exception occurrences mechanism should be
16586 enabled or disabled in a consistent manner across all libraries.
16587 Otherwise, a Program_Error will be raised during the elaboration of the
16588 conflicting libraries.
16590 If the @code{'Version} and @code{'Body_Version}
16591 attributes are used inside a library, then it is necessary to
16592 perform a @code{gnatbind} step that mentions all @file{ALI} files in all
16593 libraries, so that version identifiers can be properly computed.
16594 In practice these attributes are rarely used, so this is unlikely
16595 to be a consideration.
16597 @node Rebuilding the GNAT Run-Time Library
16598 @section Rebuilding the GNAT Run-Time Library
16601 It may be useful to recompile the GNAT library in various contexts, the
16602 most important one being the use of partition-wide configuration pragmas
16603 such as Normalize_Scalar. A special Makefile called
16604 @code{Makefile.adalib} is provided to that effect and can be found in
16605 the directory containing the GNAT library. The location of this
16606 directory depends on the way the GNAT environment has been installed and can
16607 be determined by means of the command:
16614 The last entry in the object search path usually contains the
16615 gnat library. This Makefile contains its own documentation and in
16616 particular the set of instructions needed to rebuild a new library and
16619 @node Using the GNU make Utility
16620 @chapter Using the GNU @code{make} Utility
16624 This chapter offers some examples of makefiles that solve specific
16625 problems. It does not explain how to write a makefile (see the GNU make
16626 documentation), nor does it try to replace the @code{gnatmake} utility
16627 (@pxref{The GNAT Make Program gnatmake}).
16629 All the examples in this section are specific to the GNU version of
16630 make. Although @code{make} is a standard utility, and the basic language
16631 is the same, these examples use some advanced features found only in
16635 * Using gnatmake in a Makefile::
16636 * Automatically Creating a List of Directories::
16637 * Generating the Command Line Switches::
16638 * Overcoming Command Line Length Limits::
16641 @node Using gnatmake in a Makefile
16642 @section Using gnatmake in a Makefile
16647 Complex project organizations can be handled in a very powerful way by
16648 using GNU make combined with gnatmake. For instance, here is a Makefile
16649 which allows you to build each subsystem of a big project into a separate
16650 shared library. Such a makefile allows you to significantly reduce the link
16651 time of very big applications while maintaining full coherence at
16652 each step of the build process.
16654 The list of dependencies are handled automatically by
16655 @code{gnatmake}. The Makefile is simply used to call gnatmake in each of
16656 the appropriate directories.
16658 Note that you should also read the example on how to automatically
16659 create the list of directories
16660 (@pxref{Automatically Creating a List of Directories})
16661 which might help you in case your project has a lot of subdirectories.
16666 @font@heightrm=cmr8
16669 ## This Makefile is intended to be used with the following directory
16671 ## - The sources are split into a series of csc (computer software components)
16672 ## Each of these csc is put in its own directory.
16673 ## Their name are referenced by the directory names.
16674 ## They will be compiled into shared library (although this would also work
16675 ## with static libraries
16676 ## - The main program (and possibly other packages that do not belong to any
16677 ## csc is put in the top level directory (where the Makefile is).
16678 ## toplevel_dir __ first_csc (sources) __ lib (will contain the library)
16679 ## \_ second_csc (sources) __ lib (will contain the library)
16681 ## Although this Makefile is build for shared library, it is easy to modify
16682 ## to build partial link objects instead (modify the lines with -shared and
16685 ## With this makefile, you can change any file in the system or add any new
16686 ## file, and everything will be recompiled correctly (only the relevant shared
16687 ## objects will be recompiled, and the main program will be re-linked).
16689 # The list of computer software component for your project. This might be
16690 # generated automatically.
16693 # Name of the main program (no extension)
16696 # If we need to build objects with -fPIC, uncomment the following line
16699 # The following variable should give the directory containing libgnat.so
16700 # You can get this directory through 'gnatls -v'. This is usually the last
16701 # directory in the Object_Path.
16704 # The directories for the libraries
16705 # (This macro expands the list of CSC to the list of shared libraries, you
16706 # could simply use the expanded form :
16707 # LIB_DIR=aa/lib/libaa.so bb/lib/libbb.so cc/lib/libcc.so
16708 LIB_DIR=$@{foreach dir,$@{CSC_LIST@},$@{dir@}/lib/lib$@{dir@}.so@}
16710 $@{MAIN@}: objects $@{LIB_DIR@}
16711 gnatbind $@{MAIN@} $@{CSC_LIST:%=-aO%/lib@} -shared
16712 gnatlink $@{MAIN@} $@{CSC_LIST:%=-l%@}
16715 # recompile the sources
16716 gnatmake -c -i $@{MAIN@}.adb $@{NEED_FPIC@} $@{CSC_LIST:%=-I%@}
16718 # Note: In a future version of GNAT, the following commands will be simplified
16719 # by a new tool, gnatmlib
16721 mkdir -p $@{dir $@@ @}
16722 cd $@{dir $@@ @}; gcc -shared -o $@{notdir $@@ @} ../*.o -L$@{GLIB@} -lgnat
16723 cd $@{dir $@@ @}; cp -f ../*.ali .
16725 # The dependencies for the modules
16726 # Note that we have to force the expansion of *.o, since in some cases
16727 # make won't be able to do it itself.
16728 aa/lib/libaa.so: $@{wildcard aa/*.o@}
16729 bb/lib/libbb.so: $@{wildcard bb/*.o@}
16730 cc/lib/libcc.so: $@{wildcard cc/*.o@}
16732 # Make sure all of the shared libraries are in the path before starting the
16735 LD_LIBRARY_PATH=`pwd`/aa/lib:`pwd`/bb/lib:`pwd`/cc/lib ./$@{MAIN@}
16738 $@{RM@} -rf $@{CSC_LIST:%=%/lib@}
16739 $@{RM@} $@{CSC_LIST:%=%/*.ali@}
16740 $@{RM@} $@{CSC_LIST:%=%/*.o@}
16741 $@{RM@} *.o *.ali $@{MAIN@}
16744 @node Automatically Creating a List of Directories
16745 @section Automatically Creating a List of Directories
16748 In most makefiles, you will have to specify a list of directories, and
16749 store it in a variable. For small projects, it is often easier to
16750 specify each of them by hand, since you then have full control over what
16751 is the proper order for these directories, which ones should be
16754 However, in larger projects, which might involve hundreds of
16755 subdirectories, it might be more convenient to generate this list
16758 The example below presents two methods. The first one, although less
16759 general, gives you more control over the list. It involves wildcard
16760 characters, that are automatically expanded by @code{make}. Its
16761 shortcoming is that you need to explicitly specify some of the
16762 organization of your project, such as for instance the directory tree
16763 depth, whether some directories are found in a separate tree,...
16765 The second method is the most general one. It requires an external
16766 program, called @code{find}, which is standard on all Unix systems. All
16767 the directories found under a given root directory will be added to the
16773 @font@heightrm=cmr8
16776 # The examples below are based on the following directory hierarchy:
16777 # All the directories can contain any number of files
16778 # ROOT_DIRECTORY -> a -> aa -> aaa
16781 # -> b -> ba -> baa
16784 # This Makefile creates a variable called DIRS, that can be reused any time
16785 # you need this list (see the other examples in this section)
16787 # The root of your project's directory hierarchy
16791 # First method: specify explicitly the list of directories
16792 # This allows you to specify any subset of all the directories you need.
16795 DIRS := a/aa/ a/ab/ b/ba/
16798 # Second method: use wildcards
16799 # Note that the argument(s) to wildcard below should end with a '/'.
16800 # Since wildcards also return file names, we have to filter them out
16801 # to avoid duplicate directory names.
16802 # We thus use make's @code{dir} and @code{sort} functions.
16803 # It sets DIRs to the following value (note that the directories aaa and baa
16804 # are not given, unless you change the arguments to wildcard).
16805 # DIRS= ./a/a/ ./b/ ./a/aa/ ./a/ab/ ./a/ac/ ./b/ba/ ./b/bb/ ./b/bc/
16808 DIRS := $@{sort $@{dir $@{wildcard $@{ROOT_DIRECTORY@}/*/
16809 $@{ROOT_DIRECTORY@}/*/*/@}@}@}
16812 # Third method: use an external program
16813 # This command is much faster if run on local disks, avoiding NFS slowdowns.
16814 # This is the most complete command: it sets DIRs to the following value:
16815 # DIRS= ./a ./a/aa ./a/aa/aaa ./a/ab ./a/ac ./b ./b/ba ./b/ba/baa ./b/bb ./b/bc
16818 DIRS := $@{shell find $@{ROOT_DIRECTORY@} -type d -print@}
16822 @node Generating the Command Line Switches
16823 @section Generating the Command Line Switches
16826 Once you have created the list of directories as explained in the
16827 previous section (@pxref{Automatically Creating a List of Directories}),
16828 you can easily generate the command line arguments to pass to gnatmake.
16830 For the sake of completeness, this example assumes that the source path
16831 is not the same as the object path, and that you have two separate lists
16835 # see "Automatically creating a list of directories" to create
16840 GNATMAKE_SWITCHES := $@{patsubst %,-aI%,$@{SOURCE_DIRS@}@}
16841 GNATMAKE_SWITCHES += $@{patsubst %,-aO%,$@{OBJECT_DIRS@}@}
16844 gnatmake $@{GNATMAKE_SWITCHES@} main_unit
16847 @node Overcoming Command Line Length Limits
16848 @section Overcoming Command Line Length Limits
16851 One problem that might be encountered on big projects is that many
16852 operating systems limit the length of the command line. It is thus hard to give
16853 gnatmake the list of source and object directories.
16855 This example shows how you can set up environment variables, which will
16856 make @code{gnatmake} behave exactly as if the directories had been
16857 specified on the command line, but have a much higher length limit (or
16858 even none on most systems).
16860 It assumes that you have created a list of directories in your Makefile,
16861 using one of the methods presented in
16862 @ref{Automatically Creating a List of Directories}.
16863 For the sake of completeness, we assume that the object
16864 path (where the ALI files are found) is different from the sources patch.
16866 Note a small trick in the Makefile below: for efficiency reasons, we
16867 create two temporary variables (SOURCE_LIST and OBJECT_LIST), that are
16868 expanded immediately by @code{make}. This way we overcome the standard
16869 make behavior which is to expand the variables only when they are
16872 On Windows, if you are using the standard Windows command shell, you must
16873 replace colons with semicolons in the assignments to these variables.
16878 @font@heightrm=cmr8
16881 # In this example, we create both ADA_INCLUDE_PATH and ADA_OBJECT_PATH.
16882 # This is the same thing as putting the -I arguments on the command line.
16883 # (the equivalent of using -aI on the command line would be to define
16884 # only ADA_INCLUDE_PATH, the equivalent of -aO is ADA_OBJECT_PATH).
16885 # You can of course have different values for these variables.
16887 # Note also that we need to keep the previous values of these variables, since
16888 # they might have been set before running 'make' to specify where the GNAT
16889 # library is installed.
16891 # see "Automatically creating a list of directories" to create these
16897 space:=$@{empty@} $@{empty@}
16898 SOURCE_LIST := $@{subst $@{space@},:,$@{SOURCE_DIRS@}@}
16899 OBJECT_LIST := $@{subst $@{space@},:,$@{OBJECT_DIRS@}@}
16900 ADA_INCLUDE_PATH += $@{SOURCE_LIST@}
16901 ADA_OBJECT_PATH += $@{OBJECT_LIST@}
16902 export ADA_INCLUDE_PATH
16903 export ADA_OBJECT_PATH
16911 @node Finding Memory Problems
16912 @chapter Finding Memory Problems
16915 This chapter describes
16917 the @command{gnatmem} tool, which can be used to track down
16918 ``memory leaks'', and
16920 the GNAT Debug Pool facility, which can be used to detect incorrect uses of
16921 access values (including ``dangling references'').
16925 * The gnatmem Tool::
16927 * The GNAT Debug Pool Facility::
16932 @node The gnatmem Tool
16933 @section The @command{gnatmem} Tool
16937 The @code{gnatmem} utility monitors dynamic allocation and
16938 deallocation activity in a program, and displays information about
16939 incorrect deallocations and possible sources of memory leaks.
16940 It provides three type of information:
16943 General information concerning memory management, such as the total
16944 number of allocations and deallocations, the amount of allocated
16945 memory and the high water mark, i.e. the largest amount of allocated
16946 memory in the course of program execution.
16949 Backtraces for all incorrect deallocations, that is to say deallocations
16950 which do not correspond to a valid allocation.
16953 Information on each allocation that is potentially the origin of a memory
16958 * Running gnatmem::
16959 * Switches for gnatmem::
16960 * Example of gnatmem Usage::
16963 @node Running gnatmem
16964 @subsection Running @code{gnatmem}
16967 @code{gnatmem} makes use of the output created by the special version of
16968 allocation and deallocation routines that record call information. This
16969 allows to obtain accurate dynamic memory usage history at a minimal cost to
16970 the execution speed. Note however, that @code{gnatmem} is not supported on
16971 all platforms (currently, it is supported on AIX, HP-UX, GNU/Linux x86,
16972 Solaris (sparc and x86) and Windows NT/2000/XP (x86).
16975 The @code{gnatmem} command has the form
16978 $ gnatmem [switches] user_program
16982 The program must have been linked with the instrumented version of the
16983 allocation and deallocation routines. This is done by linking with the
16984 @file{libgmem.a} library. For correct symbolic backtrace information,
16985 the user program should be compiled with debugging options
16986 @ref{Switches for gcc}. For example to build @file{my_program}:
16989 $ gnatmake -g my_program -largs -lgmem
16993 When running @file{my_program} the file @file{gmem.out} is produced. This file
16994 contains information about all allocations and deallocations done by the
16995 program. It is produced by the instrumented allocations and
16996 deallocations routines and will be used by @code{gnatmem}.
16999 Gnatmem must be supplied with the @file{gmem.out} file and the executable to
17000 examine. If the location of @file{gmem.out} file was not explicitly supplied by
17001 @code{-i} switch, gnatmem will assume that this file can be found in the
17002 current directory. For example, after you have executed @file{my_program},
17003 @file{gmem.out} can be analyzed by @code{gnatmem} using the command:
17006 $ gnatmem my_program
17010 This will produce the output with the following format:
17012 *************** debut cc
17014 $ gnatmem my_program
17018 Total number of allocations : 45
17019 Total number of deallocations : 6
17020 Final Water Mark (non freed mem) : 11.29 Kilobytes
17021 High Water Mark : 11.40 Kilobytes
17026 Allocation Root # 2
17027 -------------------
17028 Number of non freed allocations : 11
17029 Final Water Mark (non freed mem) : 1.16 Kilobytes
17030 High Water Mark : 1.27 Kilobytes
17032 my_program.adb:23 my_program.alloc
17038 The first block of output gives general information. In this case, the
17039 Ada construct ``@code{@b{new}}'' was executed 45 times, and only 6 calls to an
17040 Unchecked_Deallocation routine occurred.
17043 Subsequent paragraphs display information on all allocation roots.
17044 An allocation root is a specific point in the execution of the program
17045 that generates some dynamic allocation, such as a ``@code{@b{new}}''
17046 construct. This root is represented by an execution backtrace (or subprogram
17047 call stack). By default the backtrace depth for allocations roots is 1, so
17048 that a root corresponds exactly to a source location. The backtrace can
17049 be made deeper, to make the root more specific.
17051 @node Switches for gnatmem
17052 @subsection Switches for @code{gnatmem}
17055 @code{gnatmem} recognizes the following switches:
17060 @cindex @option{-q} (@code{gnatmem})
17061 Quiet. Gives the minimum output needed to identify the origin of the
17062 memory leaks. Omits statistical information.
17065 @cindex @var{N} (@code{gnatmem})
17066 N is an integer literal (usually between 1 and 10) which controls the
17067 depth of the backtraces defining allocation root. The default value for
17068 N is 1. The deeper the backtrace, the more precise the localization of
17069 the root. Note that the total number of roots can depend on this
17070 parameter. This parameter must be specified @emph{before} the name of the
17071 executable to be analyzed, to avoid ambiguity.
17074 @cindex @option{-b} (@code{gnatmem})
17075 This switch has the same effect as just depth parameter.
17077 @item -i @var{file}
17078 @cindex @option{-i} (@code{gnatmem})
17079 Do the @code{gnatmem} processing starting from @file{file}, rather than
17080 @file{gmem.out} in the current directory.
17083 @cindex @option{-m} (@code{gnatmem})
17084 This switch causes @code{gnatmem} to mask the allocation roots that have less
17085 than n leaks. The default value is 1. Specifying the value of 0 will allow to
17086 examine even the roots that didn't result in leaks.
17089 @cindex @option{-s} (@code{gnatmem})
17090 This switch causes @code{gnatmem} to sort the allocation roots according to the
17091 specified order of sort criteria, each identified by a single letter. The
17092 currently supported criteria are @code{n, h, w} standing respectively for
17093 number of unfreed allocations, high watermark, and final watermark
17094 corresponding to a specific root. The default order is @code{nwh}.
17098 @node Example of gnatmem Usage
17099 @subsection Example of @code{gnatmem} Usage
17102 The following example shows the use of @code{gnatmem}
17103 on a simple memory-leaking program.
17104 Suppose that we have the following Ada program:
17106 @smallexample @c ada
17109 with Unchecked_Deallocation;
17110 procedure Test_Gm is
17112 type T is array (1..1000) of Integer;
17113 type Ptr is access T;
17114 procedure Free is new Unchecked_Deallocation (T, Ptr);
17117 procedure My_Alloc is
17122 procedure My_DeAlloc is
17130 for I in 1 .. 5 loop
17131 for J in I .. 5 loop
17142 The program needs to be compiled with debugging option and linked with
17143 @code{gmem} library:
17146 $ gnatmake -g test_gm -largs -lgmem
17150 Then we execute the program as usual:
17157 Then @code{gnatmem} is invoked simply with
17163 which produces the following output (result may vary on different platforms):
17168 Total number of allocations : 18
17169 Total number of deallocations : 5
17170 Final Water Mark (non freed mem) : 53.00 Kilobytes
17171 High Water Mark : 56.90 Kilobytes
17173 Allocation Root # 1
17174 -------------------
17175 Number of non freed allocations : 11
17176 Final Water Mark (non freed mem) : 42.97 Kilobytes
17177 High Water Mark : 46.88 Kilobytes
17179 test_gm.adb:11 test_gm.my_alloc
17181 Allocation Root # 2
17182 -------------------
17183 Number of non freed allocations : 1
17184 Final Water Mark (non freed mem) : 10.02 Kilobytes
17185 High Water Mark : 10.02 Kilobytes
17187 s-secsta.adb:81 system.secondary_stack.ss_init
17189 Allocation Root # 3
17190 -------------------
17191 Number of non freed allocations : 1
17192 Final Water Mark (non freed mem) : 12 Bytes
17193 High Water Mark : 12 Bytes
17195 s-secsta.adb:181 system.secondary_stack.ss_init
17199 Note that the GNAT run time contains itself a certain number of
17200 allocations that have no corresponding deallocation,
17201 as shown here for root #2 and root
17202 #3. This is a normal behavior when the number of non freed allocations
17203 is one, it allocates dynamic data structures that the run time needs for
17204 the complete lifetime of the program. Note also that there is only one
17205 allocation root in the user program with a single line back trace:
17206 test_gm.adb:11 test_gm.my_alloc, whereas a careful analysis of the
17207 program shows that 'My_Alloc' is called at 2 different points in the
17208 source (line 21 and line 24). If those two allocation roots need to be
17209 distinguished, the backtrace depth parameter can be used:
17212 $ gnatmem 3 test_gm
17216 which will give the following output:
17221 Total number of allocations : 18
17222 Total number of deallocations : 5
17223 Final Water Mark (non freed mem) : 53.00 Kilobytes
17224 High Water Mark : 56.90 Kilobytes
17226 Allocation Root # 1
17227 -------------------
17228 Number of non freed allocations : 10
17229 Final Water Mark (non freed mem) : 39.06 Kilobytes
17230 High Water Mark : 42.97 Kilobytes
17232 test_gm.adb:11 test_gm.my_alloc
17233 test_gm.adb:24 test_gm
17234 b_test_gm.c:52 main
17236 Allocation Root # 2
17237 -------------------
17238 Number of non freed allocations : 1
17239 Final Water Mark (non freed mem) : 10.02 Kilobytes
17240 High Water Mark : 10.02 Kilobytes
17242 s-secsta.adb:81 system.secondary_stack.ss_init
17243 s-secsta.adb:283 <system__secondary_stack___elabb>
17244 b_test_gm.c:33 adainit
17246 Allocation Root # 3
17247 -------------------
17248 Number of non freed allocations : 1
17249 Final Water Mark (non freed mem) : 3.91 Kilobytes
17250 High Water Mark : 3.91 Kilobytes
17252 test_gm.adb:11 test_gm.my_alloc
17253 test_gm.adb:21 test_gm
17254 b_test_gm.c:52 main
17256 Allocation Root # 4
17257 -------------------
17258 Number of non freed allocations : 1
17259 Final Water Mark (non freed mem) : 12 Bytes
17260 High Water Mark : 12 Bytes
17262 s-secsta.adb:181 system.secondary_stack.ss_init
17263 s-secsta.adb:283 <system__secondary_stack___elabb>
17264 b_test_gm.c:33 adainit
17268 The allocation root #1 of the first example has been split in 2 roots #1
17269 and #3 thanks to the more precise associated backtrace.
17274 @node The GNAT Debug Pool Facility
17275 @section The GNAT Debug Pool Facility
17277 @cindex storage, pool, memory corruption
17280 The use of unchecked deallocation and unchecked conversion can easily
17281 lead to incorrect memory references. The problems generated by such
17282 references are usually difficult to tackle because the symptoms can be
17283 very remote from the origin of the problem. In such cases, it is
17284 very helpful to detect the problem as early as possible. This is the
17285 purpose of the Storage Pool provided by @code{GNAT.Debug_Pools}.
17287 In order to use the GNAT specific debugging pool, the user must
17288 associate a debug pool object with each of the access types that may be
17289 related to suspected memory problems. See Ada Reference Manual 13.11.
17290 @smallexample @c ada
17291 type Ptr is access Some_Type;
17292 Pool : GNAT.Debug_Pools.Debug_Pool;
17293 for Ptr'Storage_Pool use Pool;
17297 @code{GNAT.Debug_Pools} is derived from a GNAT-specific kind of
17298 pool: the @code{Checked_Pool}. Such pools, like standard Ada storage pools,
17299 allow the user to redefine allocation and deallocation strategies. They
17300 also provide a checkpoint for each dereference, through the use of
17301 the primitive operation @code{Dereference} which is implicitly called at
17302 each dereference of an access value.
17304 Once an access type has been associated with a debug pool, operations on
17305 values of the type may raise four distinct exceptions,
17306 which correspond to four potential kinds of memory corruption:
17309 @code{GNAT.Debug_Pools.Accessing_Not_Allocated_Storage}
17311 @code{GNAT.Debug_Pools.Accessing_Deallocated_Storage}
17313 @code{GNAT.Debug_Pools.Freeing_Not_Allocated_Storage}
17315 @code{GNAT.Debug_Pools.Freeing_Deallocated_Storage }
17319 For types associated with a Debug_Pool, dynamic allocation is performed using
17321 GNAT allocation routine. References to all allocated chunks of memory
17322 are kept in an internal dictionary.
17323 Several deallocation strategies are provided, whereupon the user can choose
17324 to release the memory to the system, keep it allocated for further invalid
17325 access checks, or fill it with an easily recognizable pattern for debug
17327 The memory pattern is the old IBM hexadecimal convention: @code{16#DEADBEEF#}.
17329 See the documentation in the file g-debpoo.ads for more information on the
17330 various strategies.
17332 Upon each dereference, a check is made that the access value denotes a
17333 properly allocated memory location. Here is a complete example of use of
17334 @code{Debug_Pools}, that includes typical instances of memory corruption:
17335 @smallexample @c ada
17339 with Gnat.Io; use Gnat.Io;
17340 with Unchecked_Deallocation;
17341 with Unchecked_Conversion;
17342 with GNAT.Debug_Pools;
17343 with System.Storage_Elements;
17344 with Ada.Exceptions; use Ada.Exceptions;
17345 procedure Debug_Pool_Test is
17347 type T is access Integer;
17348 type U is access all T;
17350 P : GNAT.Debug_Pools.Debug_Pool;
17351 for T'Storage_Pool use P;
17353 procedure Free is new Unchecked_Deallocation (Integer, T);
17354 function UC is new Unchecked_Conversion (U, T);
17357 procedure Info is new GNAT.Debug_Pools.Print_Info(Put_Line);
17367 Put_Line (Integer'Image(B.all));
17369 when E : others => Put_Line ("raised: " & Exception_Name (E));
17374 when E : others => Put_Line ("raised: " & Exception_Name (E));
17378 Put_Line (Integer'Image(B.all));
17380 when E : others => Put_Line ("raised: " & Exception_Name (E));
17385 when E : others => Put_Line ("raised: " & Exception_Name (E));
17388 end Debug_Pool_Test;
17392 The debug pool mechanism provides the following precise diagnostics on the
17393 execution of this erroneous program:
17396 Total allocated bytes : 0
17397 Total deallocated bytes : 0
17398 Current Water Mark: 0
17402 Total allocated bytes : 8
17403 Total deallocated bytes : 0
17404 Current Water Mark: 8
17407 raised: GNAT.DEBUG_POOLS.ACCESSING_DEALLOCATED_STORAGE
17408 raised: GNAT.DEBUG_POOLS.FREEING_DEALLOCATED_STORAGE
17409 raised: GNAT.DEBUG_POOLS.ACCESSING_NOT_ALLOCATED_STORAGE
17410 raised: GNAT.DEBUG_POOLS.FREEING_NOT_ALLOCATED_STORAGE
17412 Total allocated bytes : 8
17413 Total deallocated bytes : 4
17414 Current Water Mark: 4
17419 @node Creating Sample Bodies Using gnatstub
17420 @chapter Creating Sample Bodies Using @command{gnatstub}
17424 @command{gnatstub} creates body stubs, that is, empty but compilable bodies
17425 for library unit declarations.
17427 To create a body stub, @command{gnatstub} has to compile the library
17428 unit declaration. Therefore, bodies can be created only for legal
17429 library units. Moreover, if a library unit depends semantically upon
17430 units located outside the current directory, you have to provide
17431 the source search path when calling @command{gnatstub}, see the description
17432 of @command{gnatstub} switches below.
17435 * Running gnatstub::
17436 * Switches for gnatstub::
17439 @node Running gnatstub
17440 @section Running @command{gnatstub}
17443 @command{gnatstub} has the command-line interface of the form
17446 $ gnatstub [switches] filename [directory]
17453 is the name of the source file that contains a library unit declaration
17454 for which a body must be created. The file name may contain the path
17456 The file name does not have to follow the GNAT file name conventions. If the
17458 does not follow GNAT file naming conventions, the name of the body file must
17460 explicitly as the value of the @option{^-o^/BODY=^@var{body-name}} option.
17461 If the file name follows the GNAT file naming
17462 conventions and the name of the body file is not provided,
17465 of the body file from the argument file name by replacing the @file{.ads}
17467 with the @file{.adb} suffix.
17470 indicates the directory in which the body stub is to be placed (the default
17475 is an optional sequence of switches as described in the next section
17478 @node Switches for gnatstub
17479 @section Switches for @command{gnatstub}
17485 @cindex @option{^-f^/FULL^} (@command{gnatstub})
17486 If the destination directory already contains a file with the name of the
17488 for the argument spec file, replace it with the generated body stub.
17490 @item ^-hs^/HEADER=SPEC^
17491 @cindex @option{^-hs^/HEADER=SPEC^} (@command{gnatstub})
17492 Put the comment header (i.e., all the comments preceding the
17493 compilation unit) from the source of the library unit declaration
17494 into the body stub.
17496 @item ^-hg^/HEADER=GENERAL^
17497 @cindex @option{^-hg^/HEADER=GENERAL^} (@command{gnatstub})
17498 Put a sample comment header into the body stub.
17502 @cindex @option{-IDIR} (@command{gnatstub})
17504 @cindex @option{-I-} (@command{gnatstub})
17507 @item /NOCURRENT_DIRECTORY
17508 @cindex @option{/NOCURRENT_DIRECTORY} (@command{gnatstub})
17510 ^These switches have ^This switch has^ the same meaning as in calls to
17512 ^They define ^It defines ^ the source search path in the call to
17513 @command{gcc} issued
17514 by @command{gnatstub} to compile an argument source file.
17516 @item ^-gnatec^/CONFIGURATION_PRAGMAS_FILE=^@var{PATH}
17517 @cindex @option{^-gnatec^/CONFIGURATION_PRAGMAS_FILE^} (@command{gnatstub})
17518 This switch has the same meaning as in calls to @command{gcc}.
17519 It defines the additional configuration file to be passed to the call to
17520 @command{gcc} issued
17521 by @command{gnatstub} to compile an argument source file.
17523 @item ^-gnatyM^/MAX_LINE_LENGTH=^@var{n}
17524 @cindex @option{^-gnatyM^/MAX_LINE_LENGTH^} (@command{gnatstub})
17525 (@var{n} is a non-negative integer). Set the maximum line length in the
17526 body stub to @var{n}; the default is 79. The maximum value that can be
17527 specified is 32767.
17529 @item ^-gnaty^/STYLE_CHECKS=^@var{n}
17530 @cindex @option{^-gnaty^/STYLE_CHECKS=^} (@command{gnatstub})
17531 (@var{n} is a non-negative integer from 1 to 9). Set the indentation level in
17532 the generated body sample to @var{n}.
17533 The default indentation is 3.
17535 @item ^-gnatyo^/ORDERED_SUBPROGRAMS^
17536 @cindex @option{^-gnato^/ORDERED_SUBPROGRAMS^} (@command{gnatstub})
17537 Order local bodies alphabetically. (By default local bodies are ordered
17538 in the same way as the corresponding local specs in the argument spec file.)
17540 @item ^-i^/INDENTATION=^@var{n}
17541 @cindex @option{^-i^/INDENTATION^} (@command{gnatstub})
17542 Same as @option{^-gnaty^/STYLE_CHECKS=^@var{n}}
17544 @item ^-k^/TREE_FILE=SAVE^
17545 @cindex @option{^-k^/TREE_FILE=SAVE^} (@command{gnatstub})
17546 Do not remove the tree file (i.e., the snapshot of the compiler internal
17547 structures used by @command{gnatstub}) after creating the body stub.
17549 @item ^-l^/LINE_LENGTH=^@var{n}
17550 @cindex @option{^-l^/LINE_LENGTH^} (@command{gnatstub})
17551 Same as @option{^-gnatyM^/MAX_LINE_LENGTH=^@var{n}}
17553 @item ^-o^/BODY=^@var{body-name}
17554 @cindex @option{^-o^/BODY^} (@command{gnatstub})
17555 Body file name. This should be set if the argument file name does not
17557 the GNAT file naming
17558 conventions. If this switch is omitted the default name for the body will be
17560 from the argument file name according to the GNAT file naming conventions.
17563 @cindex @option{^-q^/QUIET^} (@command{gnatstub})
17564 Quiet mode: do not generate a confirmation when a body is
17565 successfully created, and do not generate a message when a body is not
17569 @item ^-r^/TREE_FILE=REUSE^
17570 @cindex @option{^-r^/TREE_FILE=REUSE^} (@command{gnatstub})
17571 Reuse the tree file (if it exists) instead of creating it. Instead of
17572 creating the tree file for the library unit declaration, @command{gnatstub}
17573 tries to find it in the current directory and use it for creating
17574 a body. If the tree file is not found, no body is created. This option
17575 also implies @option{^-k^/SAVE^}, whether or not
17576 the latter is set explicitly.
17578 @item ^-t^/TREE_FILE=OVERWRITE^
17579 @cindex @option{^-t^/TREE_FILE=OVERWRITE^} (@command{gnatstub})
17580 Overwrite the existing tree file. If the current directory already
17581 contains the file which, according to the GNAT file naming rules should
17582 be considered as a tree file for the argument source file,
17584 will refuse to create the tree file needed to create a sample body
17585 unless this option is set.
17587 @item ^-v^/VERBOSE^
17588 @cindex @option{^-v^/VERBOSE^} (@command{gnatstub})
17589 Verbose mode: generate version information.
17594 @node Other Utility Programs
17595 @chapter Other Utility Programs
17598 This chapter discusses some other utility programs available in the Ada
17602 * Using Other Utility Programs with GNAT::
17603 * The External Symbol Naming Scheme of GNAT::
17605 * Ada Mode for Glide::
17607 * Converting Ada Files to html with gnathtml::
17608 * Installing gnathtml::
17615 @node Using Other Utility Programs with GNAT
17616 @section Using Other Utility Programs with GNAT
17619 The object files generated by GNAT are in standard system format and in
17620 particular the debugging information uses this format. This means
17621 programs generated by GNAT can be used with existing utilities that
17622 depend on these formats.
17625 In general, any utility program that works with C will also often work with
17626 Ada programs generated by GNAT. This includes software utilities such as
17627 gprof (a profiling program), @code{gdb} (the FSF debugger), and utilities such
17631 @node The External Symbol Naming Scheme of GNAT
17632 @section The External Symbol Naming Scheme of GNAT
17635 In order to interpret the output from GNAT, when using tools that are
17636 originally intended for use with other languages, it is useful to
17637 understand the conventions used to generate link names from the Ada
17640 All link names are in all lowercase letters. With the exception of library
17641 procedure names, the mechanism used is simply to use the full expanded
17642 Ada name with dots replaced by double underscores. For example, suppose
17643 we have the following package spec:
17645 @smallexample @c ada
17656 The variable @code{MN} has a full expanded Ada name of @code{QRS.MN}, so
17657 the corresponding link name is @code{qrs__mn}.
17659 Of course if a @code{pragma Export} is used this may be overridden:
17661 @smallexample @c ada
17666 pragma Export (Var1, C, External_Name => "var1_name");
17668 pragma Export (Var2, C, Link_Name => "var2_link_name");
17675 In this case, the link name for @var{Var1} is whatever link name the
17676 C compiler would assign for the C function @var{var1_name}. This typically
17677 would be either @var{var1_name} or @var{_var1_name}, depending on operating
17678 system conventions, but other possibilities exist. The link name for
17679 @var{Var2} is @var{var2_link_name}, and this is not operating system
17683 One exception occurs for library level procedures. A potential ambiguity
17684 arises between the required name @code{_main} for the C main program,
17685 and the name we would otherwise assign to an Ada library level procedure
17686 called @code{Main} (which might well not be the main program).
17688 To avoid this ambiguity, we attach the prefix @code{_ada_} to such
17689 names. So if we have a library level procedure such as
17691 @smallexample @c ada
17694 procedure Hello (S : String);
17700 the external name of this procedure will be @var{_ada_hello}.
17703 @node Ada Mode for Glide
17704 @section Ada Mode for @code{Glide}
17705 @cindex Ada mode (for Glide)
17708 The Glide mode for programming in Ada (both Ada83 and Ada95) helps the
17709 user to understand and navigate existing code, and facilitates writing
17710 new code. It furthermore provides some utility functions for easier
17711 integration of standard Emacs features when programming in Ada.
17713 Its general features include:
17717 An Integrated Development Environment with functionality such as the
17722 ``Project files'' for configuration-specific aspects
17723 (e.g. directories and compilation options)
17726 Compiling and stepping through error messages.
17729 Running and debugging an applications within Glide.
17736 User configurability
17739 Some of the specific Ada mode features are:
17743 Functions for easy and quick stepping through Ada code
17746 Getting cross reference information for identifiers (e.g., finding a
17747 defining occurrence)
17750 Displaying an index menu of types and subprograms, allowing
17751 direct selection for browsing
17754 Automatic color highlighting of the various Ada entities
17757 Glide directly supports writing Ada code, via several facilities:
17761 Switching between spec and body files with possible
17762 autogeneration of body files
17765 Automatic formating of subprogram parameter lists
17768 Automatic indentation according to Ada syntax
17771 Automatic completion of identifiers
17774 Automatic (and configurable) casing of identifiers, keywords, and attributes
17777 Insertion of syntactic templates
17780 Block commenting / uncommenting
17784 For more information, please refer to the online documentation
17785 available in the @code{Glide} @result{} @code{Help} menu.
17789 @node Converting Ada Files to html with gnathtml
17790 @section Converting Ada Files to HTML with @code{gnathtml}
17793 This @code{Perl} script allows Ada source files to be browsed using
17794 standard Web browsers. For installation procedure, see the section
17795 @xref{Installing gnathtml}.
17797 Ada reserved keywords are highlighted in a bold font and Ada comments in
17798 a blue font. Unless your program was compiled with the gcc @option{-gnatx}
17799 switch to suppress the generation of cross-referencing information, user
17800 defined variables and types will appear in a different color; you will
17801 be able to click on any identifier and go to its declaration.
17803 The command line is as follow:
17805 $ perl gnathtml.pl [switches] ada-files
17809 You can pass it as many Ada files as you want. @code{gnathtml} will generate
17810 an html file for every ada file, and a global file called @file{index.htm}.
17811 This file is an index of every identifier defined in the files.
17813 The available switches are the following ones :
17817 @cindex @option{-83} (@code{gnathtml})
17818 Only the subset on the Ada 83 keywords will be highlighted, not the full
17819 Ada 95 keywords set.
17821 @item -cc @var{color}
17822 @cindex @option{-cc} (@code{gnathtml})
17823 This option allows you to change the color used for comments. The default
17824 value is green. The color argument can be any name accepted by html.
17827 @cindex @option{-d} (@code{gnathtml})
17828 If the ada files depend on some other files (using for instance the
17829 @code{with} command, the latter will also be converted to html.
17830 Only the files in the user project will be converted to html, not the files
17831 in the run-time library itself.
17834 @cindex @option{-D} (@code{gnathtml})
17835 This command is the same as @option{-d} above, but @command{gnathtml} will
17836 also look for files in the run-time library, and generate html files for them.
17838 @item -ext @var{extension}
17839 @cindex @option{-ext} (@code{gnathtml})
17840 This option allows you to change the extension of the generated HTML files.
17841 If you do not specify an extension, it will default to @file{htm}.
17844 @cindex @option{-f} (@code{gnathtml})
17845 By default, gnathtml will generate html links only for global entities
17846 ('with'ed units, global variables and types,...). If you specify the
17847 @option{-f} on the command line, then links will be generated for local
17850 @item -l @var{number}
17851 @cindex @option{-l} (@code{gnathtml})
17852 If this switch is provided and @var{number} is not 0, then @code{gnathtml}
17853 will number the html files every @var{number} line.
17856 @cindex @option{-I} (@code{gnathtml})
17857 Specify a directory to search for library files (@file{.ALI} files) and
17858 source files. You can provide several -I switches on the command line,
17859 and the directories will be parsed in the order of the command line.
17862 @cindex @option{-o} (@code{gnathtml})
17863 Specify the output directory for html files. By default, gnathtml will
17864 saved the generated html files in a subdirectory named @file{html/}.
17866 @item -p @var{file}
17867 @cindex @option{-p} (@code{gnathtml})
17868 If you are using Emacs and the most recent Emacs Ada mode, which provides
17869 a full Integrated Development Environment for compiling, checking,
17870 running and debugging applications, you may use @file{.gpr} files
17871 to give the directories where Emacs can find sources and object files.
17873 Using this switch, you can tell gnathtml to use these files. This allows
17874 you to get an html version of your application, even if it is spread
17875 over multiple directories.
17877 @item -sc @var{color}
17878 @cindex @option{-sc} (@code{gnathtml})
17879 This option allows you to change the color used for symbol definitions.
17880 The default value is red. The color argument can be any name accepted by html.
17882 @item -t @var{file}
17883 @cindex @option{-t} (@code{gnathtml})
17884 This switch provides the name of a file. This file contains a list of
17885 file names to be converted, and the effect is exactly as though they had
17886 appeared explicitly on the command line. This
17887 is the recommended way to work around the command line length limit on some
17892 @node Installing gnathtml
17893 @section Installing @code{gnathtml}
17896 @code{Perl} needs to be installed on your machine to run this script.
17897 @code{Perl} is freely available for almost every architecture and
17898 Operating System via the Internet.
17900 On Unix systems, you may want to modify the first line of the script
17901 @code{gnathtml}, to explicitly tell the Operating system where Perl
17902 is. The syntax of this line is :
17904 #!full_path_name_to_perl
17908 Alternatively, you may run the script using the following command line:
17911 $ perl gnathtml.pl [switches] files
17920 The GNAT distribution provides an Ada 95 template for the Digital Language
17921 Sensitive Editor (LSE), a component of DECset. In order to
17922 access it, invoke LSE with the qualifier /ENVIRONMENT=GNU:[LIB]ADA95.ENV.
17929 GNAT supports The Digital Performance Coverage Analyzer (PCA), a component
17930 of DECset. To use it proceed as outlined under ``HELP PCA'', except for running
17931 the collection phase with the /DEBUG qualifier.
17934 $ GNAT MAKE /DEBUG <PROGRAM_NAME>
17935 $ DEFINE LIB$DEBUG PCA$COLLECTOR
17936 $ RUN/DEBUG <PROGRAM_NAME>
17941 @node Running and Debugging Ada Programs
17942 @chapter Running and Debugging Ada Programs
17946 This chapter discusses how to debug Ada programs. An incorrect Ada program
17947 may be handled in three ways by the GNAT compiler:
17951 The illegality may be a violation of the static semantics of Ada. In
17952 that case GNAT diagnoses the constructs in the program that are illegal.
17953 It is then a straightforward matter for the user to modify those parts of
17957 The illegality may be a violation of the dynamic semantics of Ada. In
17958 that case the program compiles and executes, but may generate incorrect
17959 results, or may terminate abnormally with some exception.
17962 When presented with a program that contains convoluted errors, GNAT
17963 itself may terminate abnormally without providing full diagnostics on
17964 the incorrect user program.
17968 * The GNAT Debugger GDB::
17970 * Introduction to GDB Commands::
17971 * Using Ada Expressions::
17972 * Calling User-Defined Subprograms::
17973 * Using the Next Command in a Function::
17976 * Debugging Generic Units::
17977 * GNAT Abnormal Termination or Failure to Terminate::
17978 * Naming Conventions for GNAT Source Files::
17979 * Getting Internal Debugging Information::
17980 * Stack Traceback::
17986 @node The GNAT Debugger GDB
17987 @section The GNAT Debugger GDB
17990 @code{GDB} is a general purpose, platform-independent debugger that
17991 can be used to debug mixed-language programs compiled with @code{GCC},
17992 and in particular is capable of debugging Ada programs compiled with
17993 GNAT. The latest versions of @code{GDB} are Ada-aware and can handle
17994 complex Ada data structures.
17996 The manual @cite{Debugging with GDB}
17998 , located in the GNU:[DOCS] directory,
18000 contains full details on the usage of @code{GDB}, including a section on
18001 its usage on programs. This manual should be consulted for full
18002 details. The section that follows is a brief introduction to the
18003 philosophy and use of @code{GDB}.
18005 When GNAT programs are compiled, the compiler optionally writes debugging
18006 information into the generated object file, including information on
18007 line numbers, and on declared types and variables. This information is
18008 separate from the generated code. It makes the object files considerably
18009 larger, but it does not add to the size of the actual executable that
18010 will be loaded into memory, and has no impact on run-time performance. The
18011 generation of debug information is triggered by the use of the
18012 ^-g^/DEBUG^ switch in the gcc or gnatmake command used to carry out
18013 the compilations. It is important to emphasize that the use of these
18014 options does not change the generated code.
18016 The debugging information is written in standard system formats that
18017 are used by many tools, including debuggers and profilers. The format
18018 of the information is typically designed to describe C types and
18019 semantics, but GNAT implements a translation scheme which allows full
18020 details about Ada types and variables to be encoded into these
18021 standard C formats. Details of this encoding scheme may be found in
18022 the file exp_dbug.ads in the GNAT source distribution. However, the
18023 details of this encoding are, in general, of no interest to a user,
18024 since @code{GDB} automatically performs the necessary decoding.
18026 When a program is bound and linked, the debugging information is
18027 collected from the object files, and stored in the executable image of
18028 the program. Again, this process significantly increases the size of
18029 the generated executable file, but it does not increase the size of
18030 the executable program itself. Furthermore, if this program is run in
18031 the normal manner, it runs exactly as if the debug information were
18032 not present, and takes no more actual memory.
18034 However, if the program is run under control of @code{GDB}, the
18035 debugger is activated. The image of the program is loaded, at which
18036 point it is ready to run. If a run command is given, then the program
18037 will run exactly as it would have if @code{GDB} were not present. This
18038 is a crucial part of the @code{GDB} design philosophy. @code{GDB} is
18039 entirely non-intrusive until a breakpoint is encountered. If no
18040 breakpoint is ever hit, the program will run exactly as it would if no
18041 debugger were present. When a breakpoint is hit, @code{GDB} accesses
18042 the debugging information and can respond to user commands to inspect
18043 variables, and more generally to report on the state of execution.
18047 @section Running GDB
18050 The debugger can be launched directly and simply from @code{glide} or
18051 through its graphical interface: @code{gvd}. It can also be used
18052 directly in text mode. Here is described the basic use of @code{GDB}
18053 in text mode. All the commands described below can be used in the
18054 @code{gvd} console window even though there is usually other more
18055 graphical ways to achieve the same goals.
18059 The command to run the graphical interface of the debugger is
18066 The command to run @code{GDB} in text mode is
18069 $ ^gdb program^$ GDB PROGRAM^
18073 where @code{^program^PROGRAM^} is the name of the executable file. This
18074 activates the debugger and results in a prompt for debugger commands.
18075 The simplest command is simply @code{run}, which causes the program to run
18076 exactly as if the debugger were not present. The following section
18077 describes some of the additional commands that can be given to @code{GDB}.
18080 @c *******************************
18081 @node Introduction to GDB Commands
18082 @section Introduction to GDB Commands
18085 @code{GDB} contains a large repertoire of commands. The manual
18086 @cite{Debugging with GDB}
18088 , located in the GNU:[DOCS] directory,
18090 includes extensive documentation on the use
18091 of these commands, together with examples of their use. Furthermore,
18092 the command @var{help} invoked from within @code{GDB} activates a simple help
18093 facility which summarizes the available commands and their options.
18094 In this section we summarize a few of the most commonly
18095 used commands to give an idea of what @code{GDB} is about. You should create
18096 a simple program with debugging information and experiment with the use of
18097 these @code{GDB} commands on the program as you read through the
18101 @item set args @var{arguments}
18102 The @var{arguments} list above is a list of arguments to be passed to
18103 the program on a subsequent run command, just as though the arguments
18104 had been entered on a normal invocation of the program. The @code{set args}
18105 command is not needed if the program does not require arguments.
18108 The @code{run} command causes execution of the program to start from
18109 the beginning. If the program is already running, that is to say if
18110 you are currently positioned at a breakpoint, then a prompt will ask
18111 for confirmation that you want to abandon the current execution and
18114 @item breakpoint @var{location}
18115 The breakpoint command sets a breakpoint, that is to say a point at which
18116 execution will halt and @code{GDB} will await further
18117 commands. @var{location} is
18118 either a line number within a file, given in the format @code{file:linenumber},
18119 or it is the name of a subprogram. If you request that a breakpoint be set on
18120 a subprogram that is overloaded, a prompt will ask you to specify on which of
18121 those subprograms you want to breakpoint. You can also
18122 specify that all of them should be breakpointed. If the program is run
18123 and execution encounters the breakpoint, then the program
18124 stops and @code{GDB} signals that the breakpoint was encountered by
18125 printing the line of code before which the program is halted.
18127 @item breakpoint exception @var{name}
18128 A special form of the breakpoint command which breakpoints whenever
18129 exception @var{name} is raised.
18130 If @var{name} is omitted,
18131 then a breakpoint will occur when any exception is raised.
18133 @item print @var{expression}
18134 This will print the value of the given expression. Most simple
18135 Ada expression formats are properly handled by @code{GDB}, so the expression
18136 can contain function calls, variables, operators, and attribute references.
18139 Continues execution following a breakpoint, until the next breakpoint or the
18140 termination of the program.
18143 Executes a single line after a breakpoint. If the next statement
18144 is a subprogram call, execution continues into (the first statement of)
18145 the called subprogram.
18148 Executes a single line. If this line is a subprogram call, executes and
18149 returns from the call.
18152 Lists a few lines around the current source location. In practice, it
18153 is usually more convenient to have a separate edit window open with the
18154 relevant source file displayed. Successive applications of this command
18155 print subsequent lines. The command can be given an argument which is a
18156 line number, in which case it displays a few lines around the specified one.
18159 Displays a backtrace of the call chain. This command is typically
18160 used after a breakpoint has occurred, to examine the sequence of calls that
18161 leads to the current breakpoint. The display includes one line for each
18162 activation record (frame) corresponding to an active subprogram.
18165 At a breakpoint, @code{GDB} can display the values of variables local
18166 to the current frame. The command @code{up} can be used to
18167 examine the contents of other active frames, by moving the focus up
18168 the stack, that is to say from callee to caller, one frame at a time.
18171 Moves the focus of @code{GDB} down from the frame currently being
18172 examined to the frame of its callee (the reverse of the previous command),
18174 @item frame @var{n}
18175 Inspect the frame with the given number. The value 0 denotes the frame
18176 of the current breakpoint, that is to say the top of the call stack.
18180 The above list is a very short introduction to the commands that
18181 @code{GDB} provides. Important additional capabilities, including conditional
18182 breakpoints, the ability to execute command sequences on a breakpoint,
18183 the ability to debug at the machine instruction level and many other
18184 features are described in detail in @cite{Debugging with GDB}.
18185 Note that most commands can be abbreviated
18186 (for example, c for continue, bt for backtrace).
18188 @node Using Ada Expressions
18189 @section Using Ada Expressions
18190 @cindex Ada expressions
18193 @code{GDB} supports a fairly large subset of Ada expression syntax, with some
18194 extensions. The philosophy behind the design of this subset is
18198 That @code{GDB} should provide basic literals and access to operations for
18199 arithmetic, dereferencing, field selection, indexing, and subprogram calls,
18200 leaving more sophisticated computations to subprograms written into the
18201 program (which therefore may be called from @code{GDB}).
18204 That type safety and strict adherence to Ada language restrictions
18205 are not particularly important to the @code{GDB} user.
18208 That brevity is important to the @code{GDB} user.
18211 Thus, for brevity, the debugger acts as if there were
18212 implicit @code{with} and @code{use} clauses in effect for all user-written
18213 packages, thus making it unnecessary to fully qualify most names with
18214 their packages, regardless of context. Where this causes ambiguity,
18215 @code{GDB} asks the user's intent.
18217 For details on the supported Ada syntax, see @cite{Debugging with GDB}.
18219 @node Calling User-Defined Subprograms
18220 @section Calling User-Defined Subprograms
18223 An important capability of @code{GDB} is the ability to call user-defined
18224 subprograms while debugging. This is achieved simply by entering
18225 a subprogram call statement in the form:
18228 call subprogram-name (parameters)
18232 The keyword @code{call} can be omitted in the normal case where the
18233 @code{subprogram-name} does not coincide with any of the predefined
18234 @code{GDB} commands.
18236 The effect is to invoke the given subprogram, passing it the
18237 list of parameters that is supplied. The parameters can be expressions and
18238 can include variables from the program being debugged. The
18239 subprogram must be defined
18240 at the library level within your program, and @code{GDB} will call the
18241 subprogram within the environment of your program execution (which
18242 means that the subprogram is free to access or even modify variables
18243 within your program).
18245 The most important use of this facility is in allowing the inclusion of
18246 debugging routines that are tailored to particular data structures
18247 in your program. Such debugging routines can be written to provide a suitably
18248 high-level description of an abstract type, rather than a low-level dump
18249 of its physical layout. After all, the standard
18250 @code{GDB print} command only knows the physical layout of your
18251 types, not their abstract meaning. Debugging routines can provide information
18252 at the desired semantic level and are thus enormously useful.
18254 For example, when debugging GNAT itself, it is crucial to have access to
18255 the contents of the tree nodes used to represent the program internally.
18256 But tree nodes are represented simply by an integer value (which in turn
18257 is an index into a table of nodes).
18258 Using the @code{print} command on a tree node would simply print this integer
18259 value, which is not very useful. But the PN routine (defined in file
18260 treepr.adb in the GNAT sources) takes a tree node as input, and displays
18261 a useful high level representation of the tree node, which includes the
18262 syntactic category of the node, its position in the source, the integers
18263 that denote descendant nodes and parent node, as well as varied
18264 semantic information. To study this example in more detail, you might want to
18265 look at the body of the PN procedure in the stated file.
18267 @node Using the Next Command in a Function
18268 @section Using the Next Command in a Function
18271 When you use the @code{next} command in a function, the current source
18272 location will advance to the next statement as usual. A special case
18273 arises in the case of a @code{return} statement.
18275 Part of the code for a return statement is the ``epilog'' of the function.
18276 This is the code that returns to the caller. There is only one copy of
18277 this epilog code, and it is typically associated with the last return
18278 statement in the function if there is more than one return. In some
18279 implementations, this epilog is associated with the first statement
18282 The result is that if you use the @code{next} command from a return
18283 statement that is not the last return statement of the function you
18284 may see a strange apparent jump to the last return statement or to
18285 the start of the function. You should simply ignore this odd jump.
18286 The value returned is always that from the first return statement
18287 that was stepped through.
18289 @node Ada Exceptions
18290 @section Breaking on Ada Exceptions
18294 You can set breakpoints that trip when your program raises
18295 selected exceptions.
18298 @item break exception
18299 Set a breakpoint that trips whenever (any task in the) program raises
18302 @item break exception @var{name}
18303 Set a breakpoint that trips whenever (any task in the) program raises
18304 the exception @var{name}.
18306 @item break exception unhandled
18307 Set a breakpoint that trips whenever (any task in the) program raises an
18308 exception for which there is no handler.
18310 @item info exceptions
18311 @itemx info exceptions @var{regexp}
18312 The @code{info exceptions} command permits the user to examine all defined
18313 exceptions within Ada programs. With a regular expression, @var{regexp}, as
18314 argument, prints out only those exceptions whose name matches @var{regexp}.
18322 @code{GDB} allows the following task-related commands:
18326 This command shows a list of current Ada tasks, as in the following example:
18333 ID TID P-ID Thread Pri State Name
18334 1 8088000 0 807e000 15 Child Activation Wait main_task
18335 2 80a4000 1 80ae000 15 Accept/Select Wait b
18336 3 809a800 1 80a4800 15 Child Activation Wait a
18337 * 4 80ae800 3 80b8000 15 Running c
18341 In this listing, the asterisk before the first task indicates it to be the
18342 currently running task. The first column lists the task ID that is used
18343 to refer to tasks in the following commands.
18345 @item break @var{linespec} task @var{taskid}
18346 @itemx break @var{linespec} task @var{taskid} if @dots{}
18347 @cindex Breakpoints and tasks
18348 These commands are like the @code{break @dots{} thread @dots{}}.
18349 @var{linespec} specifies source lines.
18351 Use the qualifier @samp{task @var{taskid}} with a breakpoint command
18352 to specify that you only want @code{GDB} to stop the program when a
18353 particular Ada task reaches this breakpoint. @var{taskid} is one of the
18354 numeric task identifiers assigned by @code{GDB}, shown in the first
18355 column of the @samp{info tasks} display.
18357 If you do not specify @samp{task @var{taskid}} when you set a
18358 breakpoint, the breakpoint applies to @emph{all} tasks of your
18361 You can use the @code{task} qualifier on conditional breakpoints as
18362 well; in this case, place @samp{task @var{taskid}} before the
18363 breakpoint condition (before the @code{if}).
18365 @item task @var{taskno}
18366 @cindex Task switching
18368 This command allows to switch to the task referred by @var{taskno}. In
18369 particular, This allows to browse the backtrace of the specified
18370 task. It is advised to switch back to the original task before
18371 continuing execution otherwise the scheduling of the program may be
18376 For more detailed information on the tasking support,
18377 see @cite{Debugging with GDB}.
18379 @node Debugging Generic Units
18380 @section Debugging Generic Units
18381 @cindex Debugging Generic Units
18385 GNAT always uses code expansion for generic instantiation. This means that
18386 each time an instantiation occurs, a complete copy of the original code is
18387 made, with appropriate substitutions of formals by actuals.
18389 It is not possible to refer to the original generic entities in
18390 @code{GDB}, but it is always possible to debug a particular instance of
18391 a generic, by using the appropriate expanded names. For example, if we have
18393 @smallexample @c ada
18398 generic package k is
18399 procedure kp (v1 : in out integer);
18403 procedure kp (v1 : in out integer) is
18409 package k1 is new k;
18410 package k2 is new k;
18412 var : integer := 1;
18425 Then to break on a call to procedure kp in the k2 instance, simply
18429 (gdb) break g.k2.kp
18433 When the breakpoint occurs, you can step through the code of the
18434 instance in the normal manner and examine the values of local variables, as for
18437 @node GNAT Abnormal Termination or Failure to Terminate
18438 @section GNAT Abnormal Termination or Failure to Terminate
18439 @cindex GNAT Abnormal Termination or Failure to Terminate
18442 When presented with programs that contain serious errors in syntax
18444 GNAT may on rare occasions experience problems in operation, such
18446 segmentation fault or illegal memory access, raising an internal
18447 exception, terminating abnormally, or failing to terminate at all.
18448 In such cases, you can activate
18449 various features of GNAT that can help you pinpoint the construct in your
18450 program that is the likely source of the problem.
18452 The following strategies are presented in increasing order of
18453 difficulty, corresponding to your experience in using GNAT and your
18454 familiarity with compiler internals.
18458 Run @code{gcc} with the @option{-gnatf}. This first
18459 switch causes all errors on a given line to be reported. In its absence,
18460 only the first error on a line is displayed.
18462 The @option{-gnatdO} switch causes errors to be displayed as soon as they
18463 are encountered, rather than after compilation is terminated. If GNAT
18464 terminates prematurely or goes into an infinite loop, the last error
18465 message displayed may help to pinpoint the culprit.
18468 Run @code{gcc} with the @option{^-v (verbose)^/VERBOSE^} switch. In this mode,
18469 @code{gcc} produces ongoing information about the progress of the
18470 compilation and provides the name of each procedure as code is
18471 generated. This switch allows you to find which Ada procedure was being
18472 compiled when it encountered a code generation problem.
18475 @cindex @option{-gnatdc} switch
18476 Run @code{gcc} with the @option{-gnatdc} switch. This is a GNAT specific
18477 switch that does for the front-end what @option{^-v^VERBOSE^} does
18478 for the back end. The system prints the name of each unit,
18479 either a compilation unit or nested unit, as it is being analyzed.
18481 Finally, you can start
18482 @code{gdb} directly on the @code{gnat1} executable. @code{gnat1} is the
18483 front-end of GNAT, and can be run independently (normally it is just
18484 called from @code{gcc}). You can use @code{gdb} on @code{gnat1} as you
18485 would on a C program (but @pxref{The GNAT Debugger GDB} for caveats). The
18486 @code{where} command is the first line of attack; the variable
18487 @code{lineno} (seen by @code{print lineno}), used by the second phase of
18488 @code{gnat1} and by the @code{gcc} backend, indicates the source line at
18489 which the execution stopped, and @code{input_file name} indicates the name of
18493 @node Naming Conventions for GNAT Source Files
18494 @section Naming Conventions for GNAT Source Files
18497 In order to examine the workings of the GNAT system, the following
18498 brief description of its organization may be helpful:
18502 Files with prefix @file{^sc^SC^} contain the lexical scanner.
18505 All files prefixed with @file{^par^PAR^} are components of the parser. The
18506 numbers correspond to chapters of the Ada 95 Reference Manual. For example,
18507 parsing of select statements can be found in @file{par-ch9.adb}.
18510 All files prefixed with @file{^sem^SEM^} perform semantic analysis. The
18511 numbers correspond to chapters of the Ada standard. For example, all
18512 issues involving context clauses can be found in @file{sem_ch10.adb}. In
18513 addition, some features of the language require sufficient special processing
18514 to justify their own semantic files: sem_aggr for aggregates, sem_disp for
18515 dynamic dispatching, etc.
18518 All files prefixed with @file{^exp^EXP^} perform normalization and
18519 expansion of the intermediate representation (abstract syntax tree, or AST).
18520 these files use the same numbering scheme as the parser and semantics files.
18521 For example, the construction of record initialization procedures is done in
18522 @file{exp_ch3.adb}.
18525 The files prefixed with @file{^bind^BIND^} implement the binder, which
18526 verifies the consistency of the compilation, determines an order of
18527 elaboration, and generates the bind file.
18530 The files @file{atree.ads} and @file{atree.adb} detail the low-level
18531 data structures used by the front-end.
18534 The files @file{sinfo.ads} and @file{sinfo.adb} detail the structure of
18535 the abstract syntax tree as produced by the parser.
18538 The files @file{einfo.ads} and @file{einfo.adb} detail the attributes of
18539 all entities, computed during semantic analysis.
18542 Library management issues are dealt with in files with prefix
18548 Ada files with the prefix @file{^a-^A-^} are children of @code{Ada}, as
18549 defined in Annex A.
18554 Files with prefix @file{^i-^I-^} are children of @code{Interfaces}, as
18555 defined in Annex B.
18559 Files with prefix @file{^s-^S-^} are children of @code{System}. This includes
18560 both language-defined children and GNAT run-time routines.
18564 Files with prefix @file{^g-^G-^} are children of @code{GNAT}. These are useful
18565 general-purpose packages, fully documented in their specifications. All
18566 the other @file{.c} files are modifications of common @code{gcc} files.
18569 @node Getting Internal Debugging Information
18570 @section Getting Internal Debugging Information
18573 Most compilers have internal debugging switches and modes. GNAT
18574 does also, except GNAT internal debugging switches and modes are not
18575 secret. A summary and full description of all the compiler and binder
18576 debug flags are in the file @file{debug.adb}. You must obtain the
18577 sources of the compiler to see the full detailed effects of these flags.
18579 The switches that print the source of the program (reconstructed from
18580 the internal tree) are of general interest for user programs, as are the
18582 the full internal tree, and the entity table (the symbol table
18583 information). The reconstructed source provides a readable version of the
18584 program after the front-end has completed analysis and expansion,
18585 and is useful when studying the performance of specific constructs.
18586 For example, constraint checks are indicated, complex aggregates
18587 are replaced with loops and assignments, and tasking primitives
18588 are replaced with run-time calls.
18590 @node Stack Traceback
18591 @section Stack Traceback
18593 @cindex stack traceback
18594 @cindex stack unwinding
18597 Traceback is a mechanism to display the sequence of subprogram calls that
18598 leads to a specified execution point in a program. Often (but not always)
18599 the execution point is an instruction at which an exception has been raised.
18600 This mechanism is also known as @i{stack unwinding} because it obtains
18601 its information by scanning the run-time stack and recovering the activation
18602 records of all active subprograms. Stack unwinding is one of the most
18603 important tools for program debugging.
18605 The first entry stored in traceback corresponds to the deepest calling level,
18606 that is to say the subprogram currently executing the instruction
18607 from which we want to obtain the traceback.
18609 Note that there is no runtime performance penalty when stack traceback
18610 is enabled, and no exception is raised during program execution.
18613 * Non-Symbolic Traceback::
18614 * Symbolic Traceback::
18617 @node Non-Symbolic Traceback
18618 @subsection Non-Symbolic Traceback
18619 @cindex traceback, non-symbolic
18622 Note: this feature is not supported on all platforms. See
18623 @file{GNAT.Traceback spec in g-traceb.ads} for a complete list of supported
18627 * Tracebacks From an Unhandled Exception::
18628 * Tracebacks From Exception Occurrences (non-symbolic)::
18629 * Tracebacks From Anywhere in a Program (non-symbolic)::
18632 @node Tracebacks From an Unhandled Exception
18633 @subsubsection Tracebacks From an Unhandled Exception
18636 A runtime non-symbolic traceback is a list of addresses of call instructions.
18637 To enable this feature you must use the @option{-E}
18638 @code{gnatbind}'s option. With this option a stack traceback is stored as part
18639 of exception information. You can retrieve this information using the
18640 @code{addr2line} tool.
18642 Here is a simple example:
18644 @smallexample @c ada
18650 raise Constraint_Error;
18665 $ gnatmake stb -bargs -E
18668 Execution terminated by unhandled exception
18669 Exception name: CONSTRAINT_ERROR
18671 Call stack traceback locations:
18672 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
18676 As we see the traceback lists a sequence of addresses for the unhandled
18677 exception @code{CONSTRAINT_ERROR} raised in procedure P1. It is easy to
18678 guess that this exception come from procedure P1. To translate these
18679 addresses into the source lines where the calls appear, the
18680 @code{addr2line} tool, described below, is invaluable. The use of this tool
18681 requires the program to be compiled with debug information.
18684 $ gnatmake -g stb -bargs -E
18687 Execution terminated by unhandled exception
18688 Exception name: CONSTRAINT_ERROR
18690 Call stack traceback locations:
18691 0x401373 0x40138b 0x40139c 0x401335 0x4011c4 0x4011f1 0x77e892a4
18693 $ addr2line --exe=stb 0x401373 0x40138b 0x40139c 0x401335 0x4011c4
18694 0x4011f1 0x77e892a4
18696 00401373 at d:/stb/stb.adb:5
18697 0040138B at d:/stb/stb.adb:10
18698 0040139C at d:/stb/stb.adb:14
18699 00401335 at d:/stb/b~stb.adb:104
18700 004011C4 at /build/.../crt1.c:200
18701 004011F1 at /build/.../crt1.c:222
18702 77E892A4 in ?? at ??:0
18706 The @code{addr2line} tool has several other useful options:
18710 to get the function name corresponding to any location
18712 @item --demangle=gnat
18713 to use the gnat decoding mode for the function names. Note that
18714 for binutils version 2.9.x the option is simply @option{--demangle}.
18718 $ addr2line --exe=stb --functions --demangle=gnat 0x401373 0x40138b
18719 0x40139c 0x401335 0x4011c4 0x4011f1
18721 00401373 in stb.p1 at d:/stb/stb.adb:5
18722 0040138B in stb.p2 at d:/stb/stb.adb:10
18723 0040139C in stb at d:/stb/stb.adb:14
18724 00401335 in main at d:/stb/b~stb.adb:104
18725 004011C4 in <__mingw_CRTStartup> at /build/.../crt1.c:200
18726 004011F1 in <mainCRTStartup> at /build/.../crt1.c:222
18730 From this traceback we can see that the exception was raised in
18731 @file{stb.adb} at line 5, which was reached from a procedure call in
18732 @file{stb.adb} at line 10, and so on. The @file{b~std.adb} is the binder file,
18733 which contains the call to the main program.
18734 @pxref{Running gnatbind}. The remaining entries are assorted runtime routines,
18735 and the output will vary from platform to platform.
18737 It is also possible to use @code{GDB} with these traceback addresses to debug
18738 the program. For example, we can break at a given code location, as reported
18739 in the stack traceback:
18745 Furthermore, this feature is not implemented inside Windows DLL. Only
18746 the non-symbolic traceback is reported in this case.
18749 (gdb) break *0x401373
18750 Breakpoint 1 at 0x401373: file stb.adb, line 5.
18754 It is important to note that the stack traceback addresses
18755 do not change when debug information is included. This is particularly useful
18756 because it makes it possible to release software without debug information (to
18757 minimize object size), get a field report that includes a stack traceback
18758 whenever an internal bug occurs, and then be able to retrieve the sequence
18759 of calls with the same program compiled with debug information.
18761 @node Tracebacks From Exception Occurrences (non-symbolic)
18762 @subsubsection Tracebacks From Exception Occurrences
18765 Non-symbolic tracebacks are obtained by using the @option{-E} binder argument.
18766 The stack traceback is attached to the exception information string, and can
18767 be retrieved in an exception handler within the Ada program, by means of the
18768 Ada95 facilities defined in @code{Ada.Exceptions}. Here is a simple example:
18770 @smallexample @c ada
18772 with Ada.Exceptions;
18777 use Ada.Exceptions;
18785 Text_IO.Put_Line (Exception_Information (E));
18799 This program will output:
18804 Exception name: CONSTRAINT_ERROR
18805 Message: stb.adb:12
18806 Call stack traceback locations:
18807 0x4015e4 0x401633 0x401644 0x401461 0x4011c4 0x4011f1 0x77e892a4
18810 @node Tracebacks From Anywhere in a Program (non-symbolic)
18811 @subsubsection Tracebacks From Anywhere in a Program
18814 It is also possible to retrieve a stack traceback from anywhere in a
18815 program. For this you need to
18816 use the @code{GNAT.Traceback} API. This package includes a procedure called
18817 @code{Call_Chain} that computes a complete stack traceback, as well as useful
18818 display procedures described below. It is not necessary to use the
18819 @option{-E gnatbind} option in this case, because the stack traceback mechanism
18820 is invoked explicitly.
18823 In the following example we compute a traceback at a specific location in
18824 the program, and we display it using @code{GNAT.Debug_Utilities.Image} to
18825 convert addresses to strings:
18827 @smallexample @c ada
18829 with GNAT.Traceback;
18830 with GNAT.Debug_Utilities;
18836 use GNAT.Traceback;
18839 TB : Tracebacks_Array (1 .. 10);
18840 -- We are asking for a maximum of 10 stack frames.
18842 -- Len will receive the actual number of stack frames returned.
18844 Call_Chain (TB, Len);
18846 Text_IO.Put ("In STB.P1 : ");
18848 for K in 1 .. Len loop
18849 Text_IO.Put (Debug_Utilities.Image (TB (K)));
18870 In STB.P1 : 16#0040_F1E4# 16#0040_14F2# 16#0040_170B# 16#0040_171C#
18871 16#0040_1461# 16#0040_11C4# 16#0040_11F1# 16#77E8_92A4#
18875 You can then get further information by invoking the @code{addr2line}
18876 tool as described earlier (note that the hexadecimal addresses
18877 need to be specified in C format, with a leading ``0x'').
18880 @node Symbolic Traceback
18881 @subsection Symbolic Traceback
18882 @cindex traceback, symbolic
18885 A symbolic traceback is a stack traceback in which procedure names are
18886 associated with each code location.
18889 Note that this feature is not supported on all platforms. See
18890 @file{GNAT.Traceback.Symbolic spec in g-trasym.ads} for a complete
18891 list of currently supported platforms.
18894 Note that the symbolic traceback requires that the program be compiled
18895 with debug information. If it is not compiled with debug information
18896 only the non-symbolic information will be valid.
18899 * Tracebacks From Exception Occurrences (symbolic)::
18900 * Tracebacks From Anywhere in a Program (symbolic)::
18903 @node Tracebacks From Exception Occurrences (symbolic)
18904 @subsubsection Tracebacks From Exception Occurrences
18906 @smallexample @c ada
18908 with GNAT.Traceback.Symbolic;
18914 raise Constraint_Error;
18931 Ada.Text_IO.Put_Line (GNAT.Traceback.Symbolic.Symbolic_Traceback (E));
18936 $ gnatmake -g .\stb -bargs -E -largs -lgnat -laddr2line -lintl
18939 0040149F in stb.p1 at stb.adb:8
18940 004014B7 in stb.p2 at stb.adb:13
18941 004014CF in stb.p3 at stb.adb:18
18942 004015DD in ada.stb at stb.adb:22
18943 00401461 in main at b~stb.adb:168
18944 004011C4 in __mingw_CRTStartup at crt1.c:200
18945 004011F1 in mainCRTStartup at crt1.c:222
18946 77E892A4 in ?? at ??:0
18950 In the above example the ``.\'' syntax in the @command{gnatmake} command
18951 is currently required by @command{addr2line} for files that are in
18952 the current working directory.
18953 Moreover, the exact sequence of linker options may vary from platform
18955 The above @option{-largs} section is for Windows platforms. By contrast,
18956 under Unix there is no need for the @option{-largs} section.
18957 Differences across platforms are due to details of linker implementation.
18959 @node Tracebacks From Anywhere in a Program (symbolic)
18960 @subsubsection Tracebacks From Anywhere in a Program
18963 It is possible to get a symbolic stack traceback
18964 from anywhere in a program, just as for non-symbolic tracebacks.
18965 The first step is to obtain a non-symbolic
18966 traceback, and then call @code{Symbolic_Traceback} to compute the symbolic
18967 information. Here is an example:
18969 @smallexample @c ada
18971 with GNAT.Traceback;
18972 with GNAT.Traceback.Symbolic;
18977 use GNAT.Traceback;
18978 use GNAT.Traceback.Symbolic;
18981 TB : Tracebacks_Array (1 .. 10);
18982 -- We are asking for a maximum of 10 stack frames.
18984 -- Len will receive the actual number of stack frames returned.
18986 Call_Chain (TB, Len);
18987 Text_IO.Put_Line (Symbolic_Traceback (TB (1 .. Len)));
19001 @node Compatibility with DEC Ada
19002 @chapter Compatibility with DEC Ada
19003 @cindex Compatibility
19006 This section of the manual compares DEC Ada for OpenVMS Alpha and GNAT
19007 OpenVMS Alpha. GNAT achieves a high level of compatibility
19008 with DEC Ada, and it should generally be straightforward to port code
19009 from the DEC Ada environment to GNAT. However, there are a few language
19010 and implementation differences of which the user must be aware. These
19011 differences are discussed in this section. In
19012 addition, the operating environment and command structure for the
19013 compiler are different, and these differences are also discussed.
19015 Note that this discussion addresses specifically the implementation
19016 of Ada 83 for DIGITAL OpenVMS Alpha Systems. In cases where the implementation
19017 of DEC Ada differs between OpenVMS Alpha Systems and OpenVMS VAX Systems,
19018 GNAT always follows the Alpha implementation.
19021 * Ada 95 Compatibility::
19022 * Differences in the Definition of Package System::
19023 * Language-Related Features::
19024 * The Package STANDARD::
19025 * The Package SYSTEM::
19026 * Tasking and Task-Related Features::
19027 * Implementation of Tasks in DEC Ada for OpenVMS Alpha Systems::
19028 * Pragmas and Pragma-Related Features::
19029 * Library of Predefined Units::
19031 * Main Program Definition::
19032 * Implementation-Defined Attributes::
19033 * Compiler and Run-Time Interfacing::
19034 * Program Compilation and Library Management::
19036 * Implementation Limits::
19040 @node Ada 95 Compatibility
19041 @section Ada 95 Compatibility
19044 GNAT is an Ada 95 compiler, and DEC Ada is an Ada 83
19045 compiler. Ada 95 is almost completely upwards compatible
19046 with Ada 83, and therefore Ada 83 programs will compile
19047 and run under GNAT with
19048 no changes or only minor changes. The Ada 95 Reference
19049 Manual (ANSI/ISO/IEC-8652:1995) provides details on specific
19052 GNAT provides the switch /83 on the GNAT COMPILE command,
19053 as well as the pragma ADA_83, to force the compiler to
19054 operate in Ada 83 mode. This mode does not guarantee complete
19055 conformance to Ada 83, but in practice is sufficient to
19056 eliminate most sources of incompatibilities.
19057 In particular, it eliminates the recognition of the
19058 additional Ada 95 keywords, so that their use as identifiers
19059 in Ada83 program is legal, and handles the cases of packages
19060 with optional bodies, and generics that instantiate unconstrained
19061 types without the use of @code{(<>)}.
19063 @node Differences in the Definition of Package System
19064 @section Differences in the Definition of Package System
19067 Both the Ada 95 and Ada 83 reference manuals permit a compiler to add
19068 implementation-dependent declarations to package System. In normal mode,
19069 GNAT does not take advantage of this permission, and the version of System
19070 provided by GNAT exactly matches that in the Ada 95 Reference Manual.
19072 However, DEC Ada adds an extensive set of declarations to package System,
19073 as fully documented in the DEC Ada manuals. To minimize changes required
19074 for programs that make use of these extensions, GNAT provides the pragma
19075 Extend_System for extending the definition of package System. By using:
19077 @smallexample @c ada
19080 pragma Extend_System (Aux_DEC);
19086 The set of definitions in System is extended to include those in package
19087 @code{System.Aux_DEC}.
19088 These definitions are incorporated directly into package
19089 System, as though they had been declared there in the first place. For a
19090 list of the declarations added, see the specification of this package,
19091 which can be found in the file @code{s-auxdec.ads} in the GNAT library.
19092 The pragma Extend_System is a configuration pragma, which means that
19093 it can be placed in the file @file{gnat.adc}, so that it will automatically
19094 apply to all subsequent compilations. See the section on Configuration
19095 Pragmas for further details.
19097 An alternative approach that avoids the use of the non-standard
19098 Extend_System pragma is to add a context clause to the unit that
19099 references these facilities:
19101 @smallexample @c ada
19104 with System.Aux_DEC;
19105 use System.Aux_DEC;
19111 The effect is not quite semantically identical to incorporating
19112 the declarations directly into package @code{System},
19113 but most programs will not notice a difference
19114 unless they use prefix notation (e.g. @code{System.Integer_8})
19116 entities directly in package @code{System}.
19117 For units containing such references,
19118 the prefixes must either be removed, or the pragma @code{Extend_System}
19121 @node Language-Related Features
19122 @section Language-Related Features
19125 The following sections highlight differences in types,
19126 representations of types, operations, alignment, and
19130 * Integer Types and Representations::
19131 * Floating-Point Types and Representations::
19132 * Pragmas Float_Representation and Long_Float::
19133 * Fixed-Point Types and Representations::
19134 * Record and Array Component Alignment::
19135 * Address Clauses::
19136 * Other Representation Clauses::
19139 @node Integer Types and Representations
19140 @subsection Integer Types and Representations
19143 The set of predefined integer types is identical in DEC Ada and GNAT.
19144 Furthermore the representation of these integer types is also identical,
19145 including the capability of size clauses forcing biased representation.
19148 DEC Ada for OpenVMS Alpha systems has defined the
19149 following additional integer types in package System:
19170 When using GNAT, the first four of these types may be obtained from the
19171 standard Ada 95 package @code{Interfaces}.
19172 Alternatively, by use of the pragma
19173 @code{Extend_System}, identical
19174 declarations can be referenced directly in package @code{System}.
19175 On both GNAT and DEC Ada, the maximum integer size is 64 bits.
19177 @node Floating-Point Types and Representations
19178 @subsection Floating-Point Types and Representations
19179 @cindex Floating-Point types
19182 The set of predefined floating-point types is identical in DEC Ada and GNAT.
19183 Furthermore the representation of these floating-point
19184 types is also identical. One important difference is that the default
19185 representation for DEC Ada is VAX_Float, but the default representation
19188 Specific types may be declared to be VAX_Float or IEEE, using the pragma
19189 @code{Float_Representation} as described in the DEC Ada documentation.
19190 For example, the declarations:
19192 @smallexample @c ada
19195 type F_Float is digits 6;
19196 pragma Float_Representation (VAX_Float, F_Float);
19202 declare a type F_Float that will be represented in VAX_Float format.
19203 This set of declarations actually appears in System.Aux_DEC, which provides
19204 the full set of additional floating-point declarations provided in
19205 the DEC Ada version of package
19206 System. This and similar declarations may be accessed in a user program
19207 by using pragma @code{Extend_System}. The use of this
19208 pragma, and the related pragma @code{Long_Float} is described in further
19209 detail in the following section.
19211 @node Pragmas Float_Representation and Long_Float
19212 @subsection Pragmas Float_Representation and Long_Float
19215 DEC Ada provides the pragma @code{Float_Representation}, which
19216 acts as a program library switch to allow control over
19217 the internal representation chosen for the predefined
19218 floating-point types declared in the package @code{Standard}.
19219 The format of this pragma is as follows:
19224 @b{pragma} @code{Float_Representation}(VAX_Float | IEEE_Float);
19230 This pragma controls the representation of floating-point
19235 @code{VAX_Float} specifies that floating-point
19236 types are represented by default with the VAX hardware types
19237 F-floating, D-floating, G-floating. Note that the H-floating
19238 type is available only on DIGITAL Vax systems, and is not available
19239 in either DEC Ada or GNAT for Alpha systems.
19242 @code{IEEE_Float} specifies that floating-point
19243 types are represented by default with the IEEE single and
19244 double floating-point types.
19248 GNAT provides an identical implementation of the pragma
19249 @code{Float_Representation}, except that it functions as a
19250 configuration pragma, as defined by Ada 95. Note that the
19251 notion of configuration pragma corresponds closely to the
19252 DEC Ada notion of a program library switch.
19254 When no pragma is used in GNAT, the default is IEEE_Float, which is different
19255 from DEC Ada 83, where the default is VAX_Float. In addition, the
19256 predefined libraries in GNAT are built using IEEE_Float, so it is not
19257 advisable to change the format of numbers passed to standard library
19258 routines, and if necessary explicit type conversions may be needed.
19260 The use of IEEE_Float is recommended in GNAT since it is more efficient,
19261 and (given that it conforms to an international standard) potentially more
19262 portable. The situation in which VAX_Float may be useful is in interfacing
19263 to existing code and data that expects the use of VAX_Float. There are
19264 two possibilities here. If the requirement for the use of VAX_Float is
19265 localized, then the best approach is to use the predefined VAX_Float
19266 types in package @code{System}, as extended by
19267 @code{Extend_System}. For example, use @code{System.F_Float}
19268 to specify the 32-bit @code{F-Float} format.
19270 Alternatively, if an entire program depends heavily on the use of
19271 the @code{VAX_Float} and in particular assumes that the types in
19272 package @code{Standard} are in @code{Vax_Float} format, then it
19273 may be desirable to reconfigure GNAT to assume Vax_Float by default.
19274 This is done by using the GNAT LIBRARY command to rebuild the library, and
19275 then using the general form of the @code{Float_Representation}
19276 pragma to ensure that this default format is used throughout.
19277 The form of the GNAT LIBRARY command is:
19280 GNAT LIBRARY /CONFIG=@i{file} /CREATE=@i{directory}
19284 where @i{file} contains the new configuration pragmas
19285 and @i{directory} is the directory to be created to contain
19289 On OpenVMS systems, DEC Ada provides the pragma @code{Long_Float}
19290 to allow control over the internal representation chosen
19291 for the predefined type @code{Long_Float} and for floating-point
19292 type declarations with digits specified in the range 7 .. 15.
19293 The format of this pragma is as follows:
19295 @smallexample @c ada
19297 pragma Long_Float (D_FLOAT | G_FLOAT);
19301 @node Fixed-Point Types and Representations
19302 @subsection Fixed-Point Types and Representations
19305 On DEC Ada for OpenVMS Alpha systems, rounding is
19306 away from zero for both positive and negative numbers.
19307 Therefore, +0.5 rounds to 1 and -0.5 rounds to -1.
19309 On GNAT for OpenVMS Alpha, the results of operations
19310 on fixed-point types are in accordance with the Ada 95
19311 rules. In particular, results of operations on decimal
19312 fixed-point types are truncated.
19314 @node Record and Array Component Alignment
19315 @subsection Record and Array Component Alignment
19318 On DEC Ada for OpenVMS Alpha, all non composite components
19319 are aligned on natural boundaries. For example, 1-byte
19320 components are aligned on byte boundaries, 2-byte
19321 components on 2-byte boundaries, 4-byte components on 4-byte
19322 byte boundaries, and so on. The OpenVMS Alpha hardware
19323 runs more efficiently with naturally aligned data.
19325 ON GNAT for OpenVMS Alpha, alignment rules are compatible
19326 with DEC Ada for OpenVMS Alpha.
19328 @node Address Clauses
19329 @subsection Address Clauses
19332 In DEC Ada and GNAT, address clauses are supported for
19333 objects and imported subprograms.
19334 The predefined type @code{System.Address} is a private type
19335 in both compilers, with the same representation (it is simply
19336 a machine pointer). Addition, subtraction, and comparison
19337 operations are available in the standard Ada 95 package
19338 @code{System.Storage_Elements}, or in package @code{System}
19339 if it is extended to include @code{System.Aux_DEC} using a
19340 pragma @code{Extend_System} as previously described.
19342 Note that code that with's both this extended package @code{System}
19343 and the package @code{System.Storage_Elements} should not @code{use}
19344 both packages, or ambiguities will result. In general it is better
19345 not to mix these two sets of facilities. The Ada 95 package was
19346 designed specifically to provide the kind of features that DEC Ada
19347 adds directly to package @code{System}.
19349 GNAT is compatible with DEC Ada in its handling of address
19350 clauses, except for some limitations in
19351 the form of address clauses for composite objects with
19352 initialization. Such address clauses are easily replaced
19353 by the use of an explicitly-defined constant as described
19354 in the Ada 95 Reference Manual (13.1(22)). For example, the sequence
19357 @smallexample @c ada
19359 X, Y : Integer := Init_Func;
19360 Q : String (X .. Y) := "abc";
19362 for Q'Address use Compute_Address;
19367 will be rejected by GNAT, since the address cannot be computed at the time
19368 that Q is declared. To achieve the intended effect, write instead:
19370 @smallexample @c ada
19373 X, Y : Integer := Init_Func;
19374 Q_Address : constant Address := Compute_Address;
19375 Q : String (X .. Y) := "abc";
19377 for Q'Address use Q_Address;
19383 which will be accepted by GNAT (and other Ada 95 compilers), and is also
19384 backwards compatible with Ada 83. A fuller description of the restrictions
19385 on address specifications is found in the GNAT Reference Manual.
19387 @node Other Representation Clauses
19388 @subsection Other Representation Clauses
19391 GNAT supports in a compatible manner all the representation
19392 clauses supported by DEC Ada. In addition, it
19393 supports representation clause forms that are new in Ada 95
19394 including COMPONENT_SIZE and SIZE clauses for objects.
19396 @node The Package STANDARD
19397 @section The Package STANDARD
19400 The package STANDARD, as implemented by DEC Ada, is fully
19401 described in the Reference Manual for the Ada Programming
19402 Language (ANSI/MIL-STD-1815A-1983) and in the DEC Ada
19403 Language Reference Manual. As implemented by GNAT, the
19404 package STANDARD is described in the Ada 95 Reference
19407 In addition, DEC Ada supports the Latin-1 character set in
19408 the type CHARACTER. GNAT supports the Latin-1 character set
19409 in the type CHARACTER and also Unicode (ISO 10646 BMP) in
19410 the type WIDE_CHARACTER.
19412 The floating-point types supported by GNAT are those
19413 supported by DEC Ada, but defaults are different, and are controlled by
19414 pragmas. See @pxref{Floating-Point Types and Representations} for details.
19416 @node The Package SYSTEM
19417 @section The Package SYSTEM
19420 DEC Ada provides a system-specific version of the package
19421 SYSTEM for each platform on which the language ships.
19422 For the complete specification of the package SYSTEM, see
19423 Appendix F of the DEC Ada Language Reference Manual.
19425 On DEC Ada, the package SYSTEM includes the following conversion functions:
19427 @item TO_ADDRESS(INTEGER)
19429 @item TO_ADDRESS(UNSIGNED_LONGWORD)
19431 @item TO_ADDRESS(universal_integer)
19433 @item TO_INTEGER(ADDRESS)
19435 @item TO_UNSIGNED_LONGWORD(ADDRESS)
19437 @item Function IMPORT_VALUE return UNSIGNED_LONGWORD and the
19438 functions IMPORT_ADDRESS and IMPORT_LARGEST_VALUE
19442 By default, GNAT supplies a version of SYSTEM that matches
19443 the definition given in the Ada 95 Reference Manual.
19445 is a subset of the DIGITAL system definitions, which is as
19446 close as possible to the original definitions. The only difference
19447 is that the definition of SYSTEM_NAME is different:
19449 @smallexample @c ada
19452 type Name is (SYSTEM_NAME_GNAT);
19453 System_Name : constant Name := SYSTEM_NAME_GNAT;
19459 Also, GNAT adds the new Ada 95 declarations for
19460 BIT_ORDER and DEFAULT_BIT_ORDER.
19462 However, the use of the following pragma causes GNAT
19463 to extend the definition of package SYSTEM so that it
19464 encompasses the full set of DIGITAL-specific extensions,
19465 including the functions listed above:
19467 @smallexample @c ada
19469 pragma Extend_System (Aux_DEC);
19474 The pragma Extend_System is a configuration pragma that
19475 is most conveniently placed in the @file{gnat.adc} file. See the
19476 GNAT Reference Manual for further details.
19478 DEC Ada does not allow the recompilation of the package
19479 SYSTEM. Instead DEC Ada provides several pragmas (SYSTEM_
19480 NAME, STORAGE_UNIT, and MEMORY_SIZE) to modify values in
19481 the package SYSTEM. On OpenVMS Alpha systems, the pragma
19482 SYSTEM_NAME takes the enumeration literal OPENVMS_AXP as
19483 its single argument.
19485 GNAT does permit the recompilation of package SYSTEM using
19486 a special switch (@option{-gnatg}) and this switch can be used if
19487 it is necessary to modify the definitions in SYSTEM. GNAT does
19488 not permit the specification of SYSTEM_NAME, STORAGE_UNIT
19489 or MEMORY_SIZE by any other means.
19491 On GNAT systems, the pragma SYSTEM_NAME takes the
19492 enumeration literal SYSTEM_NAME_GNAT.
19494 The definitions provided by the use of
19496 @smallexample @c ada
19497 pragma Extend_System (AUX_Dec);
19501 are virtually identical to those provided by the DEC Ada 83 package
19502 System. One important difference is that the name of the TO_ADDRESS
19503 function for type UNSIGNED_LONGWORD is changed to TO_ADDRESS_LONG.
19504 See the GNAT Reference manual for a discussion of why this change was
19508 The version of TO_ADDRESS taking a universal integer argument is in fact
19509 an extension to Ada 83 not strictly compatible with the reference manual.
19510 In GNAT, we are constrained to be exactly compatible with the standard,
19511 and this means we cannot provide this capability. In DEC Ada 83, the
19512 point of this definition is to deal with a call like:
19514 @smallexample @c ada
19515 TO_ADDRESS (16#12777#);
19519 Normally, according to the Ada 83 standard, one would expect this to be
19520 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
19521 of TO_ADDRESS. However, in DEC Ada 83, there is no ambiguity, since the
19522 definition using universal_integer takes precedence.
19524 In GNAT, since the version with universal_integer cannot be supplied, it is
19525 not possible to be 100% compatible. Since there are many programs using
19526 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
19527 to change the name of the function in the UNSIGNED_LONGWORD case, so the
19528 declarations provided in the GNAT version of AUX_Dec are:
19530 @smallexample @c ada
19531 function To_Address (X : Integer) return Address;
19532 pragma Pure_Function (To_Address);
19534 function To_Address_Long (X : Unsigned_Longword) return Address;
19535 pragma Pure_Function (To_Address_Long);
19539 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
19540 change the name to TO_ADDRESS_LONG.
19542 @node Tasking and Task-Related Features
19543 @section Tasking and Task-Related Features
19546 The concepts relevant to a comparison of tasking on GNAT
19547 and on DEC Ada for OpenVMS Alpha systems are discussed in
19548 the following sections.
19550 For detailed information on concepts related to tasking in
19551 DEC Ada, see the DEC Ada Language Reference Manual and the
19552 relevant run-time reference manual.
19554 @node Implementation of Tasks in DEC Ada for OpenVMS Alpha Systems
19555 @section Implementation of Tasks in DEC Ada for OpenVMS Alpha Systems
19558 On OpenVMS Alpha systems, each Ada task (except a passive
19559 task) is implemented as a single stream of execution
19560 that is created and managed by the kernel. On these
19561 systems, DEC Ada tasking support is based on DECthreads,
19562 an implementation of the POSIX standard for threads.
19564 Although tasks are implemented as threads, all tasks in
19565 an Ada program are part of the same process. As a result,
19566 resources such as open files and virtual memory can be
19567 shared easily among tasks. Having all tasks in one process
19568 allows better integration with the programming environment
19569 (the shell and the debugger, for example).
19571 Also, on OpenVMS Alpha systems, DEC Ada tasks and foreign
19572 code that calls DECthreads routines can be used together.
19573 The interaction between Ada tasks and DECthreads routines
19574 can have some benefits. For example when on OpenVMS Alpha,
19575 DEC Ada can call C code that is already threaded.
19576 GNAT on OpenVMS Alpha uses the facilities of DECthreads,
19577 and Ada tasks are mapped to threads.
19580 * Assigning Task IDs::
19581 * Task IDs and Delays::
19582 * Task-Related Pragmas::
19583 * Scheduling and Task Priority::
19585 * External Interrupts::
19588 @node Assigning Task IDs
19589 @subsection Assigning Task IDs
19592 The DEC Ada Run-Time Library always assigns %TASK 1 to
19593 the environment task that executes the main program. On
19594 OpenVMS Alpha systems, %TASK 0 is often used for tasks
19595 that have been created but are not yet activated.
19597 On OpenVMS Alpha systems, task IDs are assigned at
19598 activation. On GNAT systems, task IDs are also assigned at
19599 task creation but do not have the same form or values as
19600 task ID values in DEC Ada. There is no null task, and the
19601 environment task does not have a specific task ID value.
19603 @node Task IDs and Delays
19604 @subsection Task IDs and Delays
19607 On OpenVMS Alpha systems, tasking delays are implemented
19608 using Timer System Services. The Task ID is used for the
19609 identification of the timer request (the REQIDT parameter).
19610 If Timers are used in the application take care not to use
19611 0 for the identification, because cancelling such a timer
19612 will cancel all timers and may lead to unpredictable results.
19614 @node Task-Related Pragmas
19615 @subsection Task-Related Pragmas
19618 Ada supplies the pragma TASK_STORAGE, which allows
19619 specification of the size of the guard area for a task
19620 stack. (The guard area forms an area of memory that has no
19621 read or write access and thus helps in the detection of
19622 stack overflow.) On OpenVMS Alpha systems, if the pragma
19623 TASK_STORAGE specifies a value of zero, a minimal guard
19624 area is created. In the absence of a pragma TASK_STORAGE, a default guard
19627 GNAT supplies the following task-related pragmas:
19632 This pragma appears within a task definition and
19633 applies to the task in which it appears. The argument
19634 must be of type SYSTEM.TASK_INFO.TASK_INFO_TYPE.
19638 GNAT implements pragma TASK_STORAGE in the same way as
19640 Both DEC Ada and GNAT supply the pragmas PASSIVE,
19641 SUPPRESS, and VOLATILE.
19643 @node Scheduling and Task Priority
19644 @subsection Scheduling and Task Priority
19647 DEC Ada implements the Ada language requirement that
19648 when two tasks are eligible for execution and they have
19649 different priorities, the lower priority task does not
19650 execute while the higher priority task is waiting. The DEC
19651 Ada Run-Time Library keeps a task running until either the
19652 task is suspended or a higher priority task becomes ready.
19654 On OpenVMS Alpha systems, the default strategy is round-
19655 robin with preemption. Tasks of equal priority take turns
19656 at the processor. A task is run for a certain period of
19657 time and then placed at the rear of the ready queue for
19658 its priority level.
19660 DEC Ada provides the implementation-defined pragma TIME_SLICE,
19661 which can be used to enable or disable round-robin
19662 scheduling of tasks with the same priority.
19663 See the relevant DEC Ada run-time reference manual for
19664 information on using the pragmas to control DEC Ada task
19667 GNAT follows the scheduling rules of Annex D (real-time
19668 Annex) of the Ada 95 Reference Manual. In general, this
19669 scheduling strategy is fully compatible with DEC Ada
19670 although it provides some additional constraints (as
19671 fully documented in Annex D).
19672 GNAT implements time slicing control in a manner compatible with
19673 DEC Ada 83, by means of the pragma Time_Slice, whose semantics are identical
19674 to the DEC Ada 83 pragma of the same name.
19675 Note that it is not possible to mix GNAT tasking and
19676 DEC Ada 83 tasking in the same program, since the two run times are
19679 @node The Task Stack
19680 @subsection The Task Stack
19683 In DEC Ada, a task stack is allocated each time a
19684 non passive task is activated. As soon as the task is
19685 terminated, the storage for the task stack is deallocated.
19686 If you specify a size of zero (bytes) with T'STORAGE_SIZE,
19687 a default stack size is used. Also, regardless of the size
19688 specified, some additional space is allocated for task
19689 management purposes. On OpenVMS Alpha systems, at least
19690 one page is allocated.
19692 GNAT handles task stacks in a similar manner. According to
19693 the Ada 95 rules, it provides the pragma STORAGE_SIZE as
19694 an alternative method for controlling the task stack size.
19695 The specification of the attribute T'STORAGE_SIZE is also
19696 supported in a manner compatible with DEC Ada.
19698 @node External Interrupts
19699 @subsection External Interrupts
19702 On DEC Ada, external interrupts can be associated with task entries.
19703 GNAT is compatible with DEC Ada in its handling of external interrupts.
19705 @node Pragmas and Pragma-Related Features
19706 @section Pragmas and Pragma-Related Features
19709 Both DEC Ada and GNAT supply all language-defined pragmas
19710 as specified by the Ada 83 standard. GNAT also supplies all
19711 language-defined pragmas specified in the Ada 95 Reference Manual.
19712 In addition, GNAT implements the implementation-defined pragmas
19718 @item COMMON_OBJECT
19720 @item COMPONENT_ALIGNMENT
19722 @item EXPORT_EXCEPTION
19724 @item EXPORT_FUNCTION
19726 @item EXPORT_OBJECT
19728 @item EXPORT_PROCEDURE
19730 @item EXPORT_VALUED_PROCEDURE
19732 @item FLOAT_REPRESENTATION
19736 @item IMPORT_EXCEPTION
19738 @item IMPORT_FUNCTION
19740 @item IMPORT_OBJECT
19742 @item IMPORT_PROCEDURE
19744 @item IMPORT_VALUED_PROCEDURE
19746 @item INLINE_GENERIC
19748 @item INTERFACE_NAME
19758 @item SHARE_GENERIC
19770 These pragmas are all fully implemented, with the exception of @code{Title},
19771 @code{Passive}, and @code{Share_Generic}, which are
19772 recognized, but which have no
19773 effect in GNAT. The effect of @code{Passive} may be obtained by the
19774 use of protected objects in Ada 95. In GNAT, all generics are inlined.
19776 Unlike DEC Ada, the GNAT 'EXPORT_@i{subprogram}' pragmas require
19777 a separate subprogram specification which must appear before the
19780 GNAT also supplies a number of implementation-defined pragmas as follows:
19782 @item C_PASS_BY_COPY
19784 @item EXTEND_SYSTEM
19786 @item SOURCE_FILE_NAME
19804 @item CPP_CONSTRUCTOR
19806 @item CPP_DESTRUCTOR
19816 @item LINKER_SECTION
19818 @item MACHINE_ATTRIBUTE
19822 @item PURE_FUNCTION
19824 @item SOURCE_REFERENCE
19828 @item UNCHECKED_UNION
19830 @item UNIMPLEMENTED_UNIT
19832 @item UNIVERSAL_DATA
19834 @item WEAK_EXTERNAL
19838 For full details on these GNAT implementation-defined pragmas, see
19839 the GNAT Reference Manual.
19842 * Restrictions on the Pragma INLINE::
19843 * Restrictions on the Pragma INTERFACE::
19844 * Restrictions on the Pragma SYSTEM_NAME::
19847 @node Restrictions on the Pragma INLINE
19848 @subsection Restrictions on the Pragma INLINE
19851 DEC Ada applies the following restrictions to the pragma INLINE:
19853 @item Parameters cannot be a task type.
19855 @item Function results cannot be task types, unconstrained
19856 array types, or unconstrained types with discriminants.
19858 @item Bodies cannot declare the following:
19860 @item Subprogram body or stub (imported subprogram is allowed)
19864 @item Generic declarations
19866 @item Instantiations
19870 @item Access types (types derived from access types allowed)
19872 @item Array or record types
19874 @item Dependent tasks
19876 @item Direct recursive calls of subprogram or containing
19877 subprogram, directly or via a renaming
19883 In GNAT, the only restriction on pragma INLINE is that the
19884 body must occur before the call if both are in the same
19885 unit, and the size must be appropriately small. There are
19886 no other specific restrictions which cause subprograms to
19887 be incapable of being inlined.
19889 @node Restrictions on the Pragma INTERFACE
19890 @subsection Restrictions on the Pragma INTERFACE
19893 The following lists and describes the restrictions on the
19894 pragma INTERFACE on DEC Ada and GNAT:
19896 @item Languages accepted: Ada, Bliss, C, Fortran, Default.
19897 Default is the default on OpenVMS Alpha systems.
19899 @item Parameter passing: Language specifies default
19900 mechanisms but can be overridden with an EXPORT pragma.
19903 @item Ada: Use internal Ada rules.
19905 @item Bliss, C: Parameters must be mode @code{in}; cannot be
19906 record or task type. Result cannot be a string, an
19907 array, or a record.
19909 @item Fortran: Parameters cannot be a task. Result cannot
19910 be a string, an array, or a record.
19915 GNAT is entirely upwards compatible with DEC Ada, and in addition allows
19916 record parameters for all languages.
19918 @node Restrictions on the Pragma SYSTEM_NAME
19919 @subsection Restrictions on the Pragma SYSTEM_NAME
19922 For DEC Ada for OpenVMS Alpha, the enumeration literal
19923 for the type NAME is OPENVMS_AXP. In GNAT, the enumeration
19924 literal for the type NAME is SYSTEM_NAME_GNAT.
19926 @node Library of Predefined Units
19927 @section Library of Predefined Units
19930 A library of predefined units is provided as part of the
19931 DEC Ada and GNAT implementations. DEC Ada does not provide
19932 the package MACHINE_CODE but instead recommends importing
19935 The GNAT versions of the DEC Ada Run-Time Library (ADA$PREDEFINED:)
19936 units are taken from the OpenVMS Alpha version, not the OpenVMS VAX
19937 version. During GNAT installation, the DEC Ada Predefined
19938 Library units are copied into the GNU:[LIB.OPENVMS7_x.2_8_x.DECLIB]
19939 (aka DECLIB) directory and patched to remove Ada 95 incompatibilities
19940 and to make them interoperable with GNAT, @pxref{Changes to DECLIB}
19943 The GNAT RTL is contained in
19944 the GNU:[LIB.OPENVMS7_x.2_8_x.ADALIB] (aka ADALIB) directory and
19945 the default search path is set up to find DECLIB units in preference
19946 to ADALIB units with the same name (TEXT_IO, SEQUENTIAL_IO, and DIRECT_IO,
19949 However, it is possible to change the default so that the
19950 reverse is true, or even to mix them using child package
19951 notation. The DEC Ada 83 units are available as DEC.xxx where xxx
19952 is the package name, and the Ada units are available in the
19953 standard manner defined for Ada 95, that is to say as Ada.xxx. To
19954 change the default, set ADA_INCLUDE_PATH and ADA_OBJECTS_PATH
19955 appropriately. For example, to change the default to use the Ada95
19959 $ DEFINE ADA_INCLUDE_PATH GNU:[LIB.OPENVMS7_1.2_8_1.ADAINCLUDE],-
19960 GNU:[LIB.OPENVMS7_1.2_8_1.DECLIB]
19961 $ DEFINE ADA_OBJECTS_PATH GNU:[LIB.OPENVMS7_1.2_8_1.ADALIB],-
19962 GNU:[LIB.OPENVMS7_1.2_8_1.DECLIB]
19966 * Changes to DECLIB::
19969 @node Changes to DECLIB
19970 @subsection Changes to DECLIB
19973 The changes made to the DEC Ada predefined library for GNAT and Ada 95
19974 compatibility are minor and include the following:
19977 @item Adjusting the location of pragmas and record representation
19978 clauses to obey Ada 95 rules
19980 @item Adding the proper notation to generic formal parameters
19981 that take unconstrained types in instantiation
19983 @item Adding pragma ELABORATE_BODY to package specifications
19984 that have package bodies not otherwise allowed
19986 @item Occurrences of the identifier @code{"PROTECTED"} are renamed to
19988 Currently these are found only in the STARLET package spec.
19992 None of the above changes is visible to users.
19998 On OpenVMS Alpha, DEC Ada provides the following strongly-typed bindings:
20001 @item Command Language Interpreter (CLI interface)
20003 @item DECtalk Run-Time Library (DTK interface)
20005 @item Librarian utility routines (LBR interface)
20007 @item General Purpose Run-Time Library (LIB interface)
20009 @item Math Run-Time Library (MTH interface)
20011 @item National Character Set Run-Time Library (NCS interface)
20013 @item Compiled Code Support Run-Time Library (OTS interface)
20015 @item Parallel Processing Run-Time Library (PPL interface)
20017 @item Screen Management Run-Time Library (SMG interface)
20019 @item Sort Run-Time Library (SOR interface)
20021 @item String Run-Time Library (STR interface)
20023 @item STARLET System Library
20026 @item X Window System Version 11R4 and 11R5 (X, XLIB interface)
20028 @item X Windows Toolkit (XT interface)
20030 @item X/Motif Version 1.1.3 and 1.2 (XM interface)
20034 GNAT provides implementations of these DEC bindings in the DECLIB directory.
20036 The X/Motif bindings used to build DECLIB are whatever versions are in the
20037 DEC Ada @file{ADA$PREDEFINED} directory with extension @file{.ADC}.
20038 The build script will
20039 automatically add a pragma Linker_Options to packages @code{Xm}, @code{Xt},
20041 causing the default X/Motif sharable image libraries to be linked in. This
20042 is done via options files named @file{xm.opt}, @file{xt.opt}, and
20043 @file{x_lib.opt} (also located in the @file{DECLIB} directory).
20045 It may be necessary to edit these options files to update or correct the
20046 library names if, for example, the newer X/Motif bindings from
20047 @file{ADA$EXAMPLES}
20048 had been (previous to installing GNAT) copied and renamed to supersede the
20049 default @file{ADA$PREDEFINED} versions.
20052 * Shared Libraries and Options Files::
20053 * Interfaces to C::
20056 @node Shared Libraries and Options Files
20057 @subsection Shared Libraries and Options Files
20060 When using the DEC Ada
20061 predefined X and Motif bindings, the linking with their sharable images is
20062 done automatically by @command{GNAT LINK}.
20063 When using other X and Motif bindings, you need
20064 to add the corresponding sharable images to the command line for
20065 @code{GNAT LINK}. When linking with shared libraries, or with
20066 @file{.OPT} files, you must
20067 also add them to the command line for @command{GNAT LINK}.
20069 A shared library to be used with GNAT is built in the same way as other
20070 libraries under VMS. The VMS Link command can be used in standard fashion.
20072 @node Interfaces to C
20073 @subsection Interfaces to C
20077 provides the following Ada types and operations:
20080 @item C types package (C_TYPES)
20082 @item C strings (C_TYPES.NULL_TERMINATED)
20084 @item Other_types (SHORT_INT)
20088 Interfacing to C with GNAT, one can use the above approach
20089 described for DEC Ada or the facilities of Annex B of
20090 the Ada 95 Reference Manual (packages INTERFACES.C,
20091 INTERFACES.C.STRINGS and INTERFACES.C.POINTERS). For more
20092 information, see the section ``Interfacing to C'' in the
20093 @cite{GNAT Reference Manual}.
20095 The @option{-gnatF} qualifier forces default and explicit
20096 @code{External_Name} parameters in pragmas Import and Export
20097 to be uppercased for compatibility with the default behavior
20098 of Compaq C. The qualifier has no effect on @code{Link_Name} parameters.
20100 @node Main Program Definition
20101 @section Main Program Definition
20104 The following section discusses differences in the
20105 definition of main programs on DEC Ada and GNAT.
20106 On DEC Ada, main programs are defined to meet the
20107 following conditions:
20109 @item Procedure with no formal parameters (returns 0 upon
20112 @item Procedure with no formal parameters (returns 42 when
20113 unhandled exceptions are raised)
20115 @item Function with no formal parameters whose returned value
20116 is of a discrete type
20118 @item Procedure with one OUT formal of a discrete type for
20119 which a specification of pragma EXPORT_VALUED_PROCEDURE is given.
20124 When declared with the pragma EXPORT_VALUED_PROCEDURE,
20125 a main function or main procedure returns a discrete
20126 value whose size is less than 64 bits (32 on VAX systems),
20127 the value is zero- or sign-extended as appropriate.
20128 On GNAT, main programs are defined as follows:
20130 @item Must be a non-generic, parameter-less subprogram that
20131 is either a procedure or function returning an Ada
20132 STANDARD.INTEGER (the predefined type)
20134 @item Cannot be a generic subprogram or an instantiation of a
20138 @node Implementation-Defined Attributes
20139 @section Implementation-Defined Attributes
20142 GNAT provides all DEC Ada implementation-defined
20145 @node Compiler and Run-Time Interfacing
20146 @section Compiler and Run-Time Interfacing
20149 DEC Ada provides the following ways to pass options to the linker
20152 @item /WAIT and /SUBMIT qualifiers
20154 @item /COMMAND qualifier
20156 @item /[NO]MAP qualifier
20158 @item /OUTPUT=file-spec
20160 @item /[NO]DEBUG and /[NO]TRACEBACK qualifiers
20164 To pass options to the linker, GNAT provides the following
20168 @item @option{/EXECUTABLE=exec-name}
20170 @item @option{/VERBOSE qualifier}
20172 @item @option{/[NO]DEBUG} and @option{/[NO]TRACEBACK} qualifiers
20176 For more information on these switches, see
20177 @ref{Switches for gnatlink}.
20178 In DEC Ada, the command-line switch @option{/OPTIMIZE} is available
20179 to control optimization. DEC Ada also supplies the
20182 @item @code{OPTIMIZE}
20184 @item @code{INLINE}
20186 @item @code{INLINE_GENERIC}
20188 @item @code{SUPPRESS_ALL}
20190 @item @code{PASSIVE}
20194 In GNAT, optimization is controlled strictly by command
20195 line parameters, as described in the corresponding section of this guide.
20196 The DIGITAL pragmas for control of optimization are
20197 recognized but ignored.
20199 Note that in GNAT, the default is optimization off, whereas in DEC Ada 83,
20200 the default is that optimization is turned on.
20202 @node Program Compilation and Library Management
20203 @section Program Compilation and Library Management
20206 DEC Ada and GNAT provide a comparable set of commands to
20207 build programs. DEC Ada also provides a program library,
20208 which is a concept that does not exist on GNAT. Instead,
20209 GNAT provides directories of sources that are compiled as
20212 The following table summarizes
20213 the DEC Ada commands and provides
20214 equivalent GNAT commands. In this table, some GNAT
20215 equivalents reflect the fact that GNAT does not use the
20216 concept of a program library. Instead, it uses a model
20217 in which collections of source and object files are used
20218 in a manner consistent with other languages like C and
20219 Fortran. Therefore, standard system file commands are used
20220 to manipulate these elements. Those GNAT commands are marked with
20222 Note that, unlike DEC Ada, none of the GNAT commands accepts wild cards.
20225 @multitable @columnfractions .35 .65
20227 @item @emph{DEC Ada Command}
20228 @tab @emph{GNAT Equivalent / Description}
20230 @item @command{ADA}
20231 @tab @command{GNAT COMPILE}@*
20232 Invokes the compiler to compile one or more Ada source files.
20234 @item @command{ACS ATTACH}@*
20235 @tab [No equivalent]@*
20236 Switches control of terminal from current process running the program
20239 @item @command{ACS CHECK}
20240 @tab @command{GNAT MAKE /DEPENDENCY_LIST}@*
20241 Forms the execution closure of one
20242 or more compiled units and checks completeness and currency.
20244 @item @command{ACS COMPILE}
20245 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
20246 Forms the execution closure of one or
20247 more specified units, checks completeness and currency,
20248 identifies units that have revised source files, compiles same,
20249 and recompiles units that are or will become obsolete.
20250 Also completes incomplete generic instantiations.
20252 @item @command{ACS COPY FOREIGN}
20254 Copies a foreign object file into the program library as a
20257 @item @command{ACS COPY UNIT}
20259 Copies a compiled unit from one program library to another.
20261 @item @command{ACS CREATE LIBRARY}
20262 @tab Create /directory (*)@*
20263 Creates a program library.
20265 @item @command{ACS CREATE SUBLIBRARY}
20266 @tab Create /directory (*)@*
20267 Creates a program sublibrary.
20269 @item @command{ACS DELETE LIBRARY}
20271 Deletes a program library and its contents.
20273 @item @command{ACS DELETE SUBLIBRARY}
20275 Deletes a program sublibrary and its contents.
20277 @item @command{ACS DELETE UNIT}
20278 @tab Delete file (*)@*
20279 On OpenVMS systems, deletes one or more compiled units from
20280 the current program library.
20282 @item @command{ACS DIRECTORY}
20283 @tab Directory (*)@*
20284 On OpenVMS systems, lists units contained in the current
20287 @item @command{ACS ENTER FOREIGN}
20289 Allows the import of a foreign body as an Ada library
20290 specification and enters a reference to a pointer.
20292 @item @command{ACS ENTER UNIT}
20294 Enters a reference (pointer) from the current program library to
20295 a unit compiled into another program library.
20297 @item @command{ACS EXIT}
20298 @tab [No equivalent]@*
20299 Exits from the program library manager.
20301 @item @command{ACS EXPORT}
20303 Creates an object file that contains system-specific object code
20304 for one or more units. With GNAT, object files can simply be copied
20305 into the desired directory.
20307 @item @command{ACS EXTRACT SOURCE}
20309 Allows access to the copied source file for each Ada compilation unit
20311 @item @command{ACS HELP}
20312 @tab @command{HELP GNAT}@*
20313 Provides online help.
20315 @item @command{ACS LINK}
20316 @tab @command{GNAT LINK}@*
20317 Links an object file containing Ada units into an executable file.
20319 @item @command{ACS LOAD}
20321 Loads (partially compiles) Ada units into the program library.
20322 Allows loading a program from a collection of files into a library
20323 without knowing the relationship among units.
20325 @item @command{ACS MERGE}
20327 Merges into the current program library, one or more units from
20328 another library where they were modified.
20330 @item @command{ACS RECOMPILE}
20331 @tab @command{GNAT MAKE /ACTIONS=COMPILE}@*
20332 Recompiles from external or copied source files any obsolete
20333 unit in the closure. Also, completes any incomplete generic
20336 @item @command{ACS REENTER}
20337 @tab @command{GNAT MAKE}@*
20338 Reenters current references to units compiled after last entered
20339 with the @command{ACS ENTER UNIT} command.
20341 @item @command{ACS SET LIBRARY}
20342 @tab Set default (*)@*
20343 Defines a program library to be the compilation context as well
20344 as the target library for compiler output and commands in general.
20346 @item @command{ACS SET PRAGMA}
20347 @tab Edit @file{gnat.adc} (*)@*
20348 Redefines specified values of the library characteristics
20349 @code{LONG_ FLOAT}, @code{MEMORY_SIZE}, @code{SYSTEM_NAME},
20350 and @code{Float_Representation}.
20352 @item @command{ACS SET SOURCE}
20353 @tab Define @code{ADA_INCLUDE_PATH} path (*)@*
20354 Defines the source file search list for the @command{ACS COMPILE} command.
20356 @item @command{ACS SHOW LIBRARY}
20357 @tab Directory (*)@*
20358 Lists information about one or more program libraries.
20360 @item @command{ACS SHOW PROGRAM}
20361 @tab [No equivalent]@*
20362 Lists information about the execution closure of one or
20363 more units in the program library.
20365 @item @command{ACS SHOW SOURCE}
20366 @tab Show logical @code{ADA_INCLUDE_PATH}@*
20367 Shows the source file search used when compiling units.
20369 @item @command{ACS SHOW VERSION}
20370 @tab Compile with @option{VERBOSE} option
20371 Displays the version number of the compiler and program library
20374 @item @command{ACS SPAWN}
20375 @tab [No equivalent]@*
20376 Creates a subprocess of the current process (same as @command{DCL SPAWN}
20379 @item @command{ACS VERIFY}
20380 @tab [No equivalent]@*
20381 Performs a series of consistency checks on a program library to
20382 determine whether the library structure and library files are in
20389 @section Input-Output
20392 On OpenVMS Alpha systems, DEC Ada uses OpenVMS Record
20393 Management Services (RMS) to perform operations on
20397 DEC Ada and GNAT predefine an identical set of input-
20398 output packages. To make the use of the
20399 generic TEXT_IO operations more convenient, DEC Ada
20400 provides predefined library packages that instantiate the
20401 integer and floating-point operations for the predefined
20402 integer and floating-point types as shown in the following table.
20404 @multitable @columnfractions .45 .55
20405 @item @emph{Package Name} @tab Instantiation
20407 @item @code{INTEGER_TEXT_IO}
20408 @tab @code{INTEGER_IO(INTEGER)}
20410 @item @code{SHORT_INTEGER_TEXT_IO}
20411 @tab @code{INTEGER_IO(SHORT_INTEGER)}
20413 @item @code{SHORT_SHORT_INTEGER_TEXT_IO}
20414 @tab @code{INTEGER_IO(SHORT_SHORT_INTEGER)}
20416 @item @code{FLOAT_TEXT_IO}
20417 @tab @code{FLOAT_IO(FLOAT)}
20419 @item @code{LONG_FLOAT_TEXT_IO}
20420 @tab @code{FLOAT_IO(LONG_FLOAT)}
20424 The DEC Ada predefined packages and their operations
20425 are implemented using OpenVMS Alpha files and input-
20426 output facilities. DEC Ada supports asynchronous input-
20427 output on OpenVMS Alpha. Familiarity with the following is
20430 @item RMS file organizations and access methods
20432 @item OpenVMS file specifications and directories
20434 @item OpenVMS File Definition Language (FDL)
20438 GNAT provides I/O facilities that are completely
20439 compatible with DEC Ada. The distribution includes the
20440 standard DEC Ada versions of all I/O packages, operating
20441 in a manner compatible with DEC Ada. In particular, the
20442 following packages are by default the DEC Ada (Ada 83)
20443 versions of these packages rather than the renamings
20444 suggested in annex J of the Ada 95 Reference Manual:
20446 @item @code{TEXT_IO}
20448 @item @code{SEQUENTIAL_IO}
20450 @item @code{DIRECT_IO}
20454 The use of the standard Ada 95 syntax for child packages (for
20455 example, @code{ADA.TEXT_IO}) retrieves the Ada 95 versions of these
20456 packages, as defined in the Ada 95 Reference Manual.
20457 GNAT provides DIGITAL-compatible predefined instantiations
20458 of the @code{TEXT_IO} packages, and also
20459 provides the standard predefined instantiations required
20460 by the Ada 95 Reference Manual.
20462 For further information on how GNAT interfaces to the file
20463 system or how I/O is implemented in programs written in
20464 mixed languages, see the chapter ``Implementation of the
20465 Standard I/O'' in the @cite{GNAT Reference Manual}.
20466 This chapter covers the following:
20468 @item Standard I/O packages
20470 @item @code{FORM} strings
20472 @item @code{ADA.DIRECT_IO}
20474 @item @code{ADA.SEQUENTIAL_IO}
20476 @item @code{ADA.TEXT_IO}
20478 @item Stream pointer positioning
20480 @item Reading and writing non-regular files
20482 @item @code{GET_IMMEDIATE}
20484 @item Treating @code{TEXT_IO} files as streams
20491 @node Implementation Limits
20492 @section Implementation Limits
20495 The following table lists implementation limits for DEC Ada
20497 @multitable @columnfractions .60 .20 .20
20499 @item @emph{Compilation Parameter}
20500 @tab @emph{DEC Ada}
20504 @item In a subprogram or entry declaration, maximum number of
20505 formal parameters that are of an unconstrained record type
20510 @item Maximum identifier length (number of characters)
20515 @item Maximum number of characters in a source line
20520 @item Maximum collection size (number of bytes)
20525 @item Maximum number of discriminants for a record type
20530 @item Maximum number of formal parameters in an entry or
20531 subprogram declaration
20536 @item Maximum number of dimensions in an array type
20541 @item Maximum number of library units and subunits in a compilation.
20546 @item Maximum number of library units and subunits in an execution.
20551 @item Maximum number of objects declared with the pragma @code{COMMON_OBJECT}
20552 or @code{PSECT_OBJECT}
20557 @item Maximum number of enumeration literals in an enumeration type
20563 @item Maximum number of lines in a source file
20568 @item Maximum number of bits in any object
20573 @item Maximum size of the static portion of a stack frame (approximate)
20584 @c **************************************
20585 @node Platform-Specific Information for the Run-Time Libraries
20586 @appendix Platform-Specific Information for the Run-Time Libraries
20587 @cindex Tasking and threads libraries
20588 @cindex Threads libraries and tasking
20589 @cindex Run-time libraries (platform-specific information)
20592 The GNAT run-time implementation
20593 may vary with respect to both the underlying threads library and
20594 the exception handling scheme.
20595 For threads support, one or more of the following are supplied:
20597 @item @b{native threads library}, a binding to the thread package from
20598 the underlying operating system
20600 @item @b{FSU threads library}, a binding to the Florida State University
20601 threads implementation, which complies fully with the requirements of Annex D
20603 @item @b{pthreads library} (Sparc Solaris only), a binding to the Solaris
20604 POSIX thread package
20608 For exception handling, either or both of two models are supplied:
20610 @item @b{Zero-Cost Exceptions} (``ZCX''),@footnote{
20611 Most programs should experience a substantial speed improvement by
20612 being compiled with a ZCX run-time.
20613 This is especially true for
20614 tasking applications or applications with many exception handlers.}
20615 @cindex Zero-Cost Exceptions
20616 @cindex ZCX (Zero-Cost Exceptions)
20617 which uses binder-generated tables that
20618 are interrogated at run time to locate a handler
20620 @item @b{setjmp / longjmp} (``SJLJ''),
20621 @cindex setjmp/longjmp Exception Model
20622 @cindex SJLJ (setjmp/longjmp Exception Model)
20623 which uses dynamically-set data to establish
20624 the set of handlers
20628 This appendix summarizes which combinations of threads and exception support
20629 are supplied on various GNAT platforms.
20630 It then shows how to select a particular library either
20631 permanently or temporarily,
20632 explains the properties of (and tradeoffs among) the various threads
20633 libraries, and provides some additional
20634 information about several specific platforms.
20637 * Summary of Run-Time Configurations::
20638 * Specifying a Run-Time Library::
20639 * Choosing between Native and FSU Threads Libraries::
20640 * Choosing the Scheduling Policy::
20641 * Solaris-Specific Considerations::
20642 * IRIX-Specific Considerations::
20643 * Linux-Specific Considerations::
20647 @node Summary of Run-Time Configurations
20648 @section Summary of Run-Time Configurations
20651 @multitable @columnfractions .30 .70
20652 @item @b{alpha-openvms}
20653 @item @code{@ @ }@i{rts-native (default)}
20654 @item @code{@ @ @ @ }Tasking @tab native VMS threads
20655 @item @code{@ @ @ @ }Exceptions @tab ZCX
20658 @item @code{@ @ }@i{rts-native (default)}
20659 @item @code{@ @ @ @ }Tasking @tab native HP threads library
20660 @item @code{@ @ @ @ }Exceptions @tab ZCX
20662 @item @code{@ @ }@i{rts-sjlj}
20663 @item @code{@ @ @ @ }Tasking @tab native HP threads library
20664 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20666 @item @b{sparc-solaris} @tab
20667 @item @code{@ @ }@i{rts-native (default)}
20668 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
20669 @item @code{@ @ @ @ }Exceptions @tab ZCX
20671 @item @code{@ @ }@i{rts-fsu} @tab
20672 @item @code{@ @ @ @ }Tasking @tab FSU threads library
20673 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20675 @item @code{@ @ }@i{rts-m64}
20676 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
20677 @item @code{@ @ @ @ }Exceptions @tab ZCX
20678 @item @code{@ @ @ @ }Constraints @tab Use only when compiling in 64-bit mode;
20679 @item @tab Use only on Solaris 8 or later.
20680 @item @tab @xref{Building and Debugging 64-bit Applications}, for details.
20682 @item @code{@ @ }@i{rts-pthread}
20683 @item @code{@ @ @ @ }Tasking @tab pthreads library
20684 @item @code{@ @ @ @ }Exceptions @tab ZCX
20686 @item @code{@ @ }@i{rts-sjlj}
20687 @item @code{@ @ @ @ }Tasking @tab native Solaris threads library
20688 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20690 @item @b{x86-linux}
20691 @item @code{@ @ }@i{rts-native (default)}
20692 @item @code{@ @ @ @ }Tasking @tab LinuxThread library
20693 @item @code{@ @ @ @ }Exceptions @tab ZCX
20695 @item @code{@ @ }@i{rts-fsu}
20696 @item @code{@ @ @ @ }Tasking @tab FSU threads library
20697 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20699 @item @code{@ @ }@i{rts-sjlj}
20700 @item @code{@ @ @ @ }Tasking @tab LinuxThread library
20701 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20703 @item @b{x86-windows}
20704 @item @code{@ @ }@i{rts-native (default)}
20705 @item @code{@ @ @ @ }Tasking @tab native Win32 threads
20706 @item @code{@ @ @ @ }Exceptions @tab SJLJ
20712 @node Specifying a Run-Time Library
20713 @section Specifying a Run-Time Library
20716 The @file{adainclude} subdirectory containing the sources of the GNAT
20717 run-time library, and the @file{adalib} subdirectory containing the
20718 @file{ALI} files and the static and/or shared GNAT library, are located
20719 in the gcc target-dependent area:
20722 target=$prefix/lib/gcc-lib/gcc-@i{dumpmachine}/gcc-@i{dumpversion}/
20726 As indicated above, on some platforms several run-time libraries are supplied.
20727 These libraries are installed in the target dependent area and
20728 contain a complete source and binary subdirectory. The detailed description
20729 below explains the differences between the different libraries in terms of
20730 their thread support.
20732 The default run-time library (when GNAT is installed) is @emph{rts-native}.
20733 This default run time is selected by the means of soft links.
20734 For example on x86-linux:
20740 +--- adainclude----------+
20742 +--- adalib-----------+ |
20744 +--- rts-native | |
20746 | +--- adainclude <---+
20748 | +--- adalib <----+
20765 If the @i{rts-fsu} library is to be selected on a permanent basis,
20766 these soft links can be modified with the following commands:
20770 $ rm -f adainclude adalib
20771 $ ln -s rts-fsu/adainclude adainclude
20772 $ ln -s rts-fsu/adalib adalib
20776 Alternatively, you can specify @file{rts-fsu/adainclude} in the file
20777 @file{$target/ada_source_path} and @file{rts-fsu/adalib} in
20778 @file{$target/ada_object_path}.
20780 Selecting another run-time library temporarily can be
20781 achieved by the regular mechanism for GNAT object or source path selection:
20785 Set the environment variables:
20788 $ ADA_INCLUDE_PATH=$target/rts-fsu/adainclude:$ADA_INCLUDE_PATH
20789 $ ADA_OBJECTS_PATH=$target/rts-fsu/adalib:$ADA_OBJECTS_PATH
20790 $ export ADA_INCLUDE_PATH ADA_OBJECTS_PATH
20794 Use @option{-aI$target/rts-fsu/adainclude}
20795 and @option{-aO$target/rts-fsu/adalib}
20796 on the @command{gnatmake} command line
20799 Use the switch @option{--RTS}; e.g., @option{--RTS=fsu}
20800 @cindex @option{--RTS} option
20804 You can similarly switch to @emph{rts-sjlj}.
20806 @node Choosing between Native and FSU Threads Libraries
20807 @section Choosing between Native and FSU Threads Libraries
20808 @cindex Native threads library
20809 @cindex FSU threads library
20812 Some GNAT implementations offer a choice between
20813 native threads and FSU threads.
20817 The @emph{native threads} library correspond to the standard system threads
20818 implementation (e.g. LinuxThreads on GNU/Linux,
20819 @cindex LinuxThreads library
20820 POSIX threads on AIX, or
20821 Solaris threads on Solaris). When this option is chosen, GNAT provides
20822 a full and accurate implementation of the core language tasking model
20823 as described in Chapter 9 of the Ada Reference Manual,
20824 but might not (and probably does not) implement
20825 the exact semantics as specified in @w{Annex D} (the Real-Time Systems Annex).
20826 @cindex Annex D (Real-Time Systems Annex) compliance
20827 @cindex Real-Time Systems Annex compliance
20828 Indeed, the reason that a choice of libraries is offered
20829 on a given target is because some of the
20830 ACATS tests for @w{Annex D} fail using the native threads library.
20831 As far as possible, this library is implemented
20832 in accordance with Ada semantics (e.g., modifying priorities as required
20833 to simulate ceiling locking),
20834 but there are often slight inaccuracies, most often in the area of
20835 absolutely respecting the priority rules on a single
20837 Moreover, it is not possible in general to define the exact behavior,
20838 because the native threads implementations
20839 are not well enough documented.
20841 On systems where the @code{SCHED_FIFO} POSIX scheduling policy is supported,
20842 @cindex POSIX scheduling policies
20843 @cindex @code{SCHED_FIFO} scheduling policy
20844 native threads will provide a behavior very close to the @w{Annex D}
20845 requirements (i.e., a run-till-blocked scheduler with fixed priorities), but
20846 on some systems (in particular GNU/Linux and Solaris), you need to have root
20847 privileges to use the @code{SCHED_FIFO} policy.
20850 The @emph{FSU threads} library provides a completely accurate implementation
20852 Thus, operating with this library, GNAT is 100% compliant with both the core
20853 and all @w{Annex D}
20855 The formal validations for implementations offering
20856 a choice of threads packages are always carried out using the FSU
20861 From these considerations, it might seem that FSU threads are the
20863 but that is by no means always the case. The FSU threads package
20864 operates with all Ada tasks appearing to the system to be a single
20865 thread. This is often considerably more efficient than operating
20866 with separate threads, since for example, switching between tasks
20867 can be accomplished without the (in some cases considerable)
20868 overhead of a context switch between two system threads. However,
20869 it means that you may well lose concurrency at the system
20870 level. Notably, some system operations (such as I/O) may block all
20871 tasks in a program and not just the calling task. More
20872 significantly, the FSU threads approach likely means you cannot
20873 take advantage of multiple processors, since for this you need
20874 separate threads (or even separate processes) to operate on
20875 different processors.
20877 For most programs, the native threads library is
20878 usually the better choice. Use the FSU threads if absolute
20879 conformance to @w{Annex D} is important for your application, or if
20880 you find that the improved efficiency of FSU threads is significant to you.
20882 Note also that to take full advantage of Florist and Glade, it is highly
20883 recommended that you use native threads.
20886 @node Choosing the Scheduling Policy
20887 @section Choosing the Scheduling Policy
20890 When using a POSIX threads implementation, you have a choice of several
20891 scheduling policies: @code{SCHED_FIFO},
20892 @cindex @code{SCHED_FIFO} scheduling policy
20894 @cindex @code{SCHED_RR} scheduling policy
20895 and @code{SCHED_OTHER}.
20896 @cindex @code{SCHED_OTHER} scheduling policy
20897 Typically, the default is @code{SCHED_OTHER}, while using @code{SCHED_FIFO}
20898 or @code{SCHED_RR} requires special (e.g., root) privileges.
20900 By default, GNAT uses the @code{SCHED_OTHER} policy. To specify
20902 @cindex @code{SCHED_FIFO} scheduling policy
20903 you can use one of the following:
20907 @code{pragma Time_Slice (0.0)}
20908 @cindex pragma Time_Slice
20910 the corresponding binder option @option{-T0}
20911 @cindex @option{-T0} option
20913 @code{pragma Task_Dispatching_Policy (FIFO_Within_Priorities)}
20914 @cindex pragma Task_Dispatching_Policy
20918 To specify @code{SCHED_RR},
20919 @cindex @code{SCHED_RR} scheduling policy
20920 you should use @code{pragma Time_Slice} with a
20921 value greater than @code{0.0}, or else use the corresponding @option{-T}
20926 @node Solaris-Specific Considerations
20927 @section Solaris-Specific Considerations
20928 @cindex Solaris Sparc threads libraries
20931 This section addresses some topics related to the various threads libraries
20932 on Sparc Solaris and then provides some information on building and
20933 debugging 64-bit applications.
20936 * Solaris Threads Issues::
20937 * Building and Debugging 64-bit Applications::
20941 @node Solaris Threads Issues
20942 @subsection Solaris Threads Issues
20945 Starting with version 3.14, GNAT under Solaris comes with a new tasking
20946 run-time library based on POSIX threads --- @emph{rts-pthread}.
20947 @cindex rts-pthread threads library
20948 This run-time library has the advantage of being mostly shared across all
20949 POSIX-compliant thread implementations, and it also provides under
20950 @w{Solaris 8} the @code{PTHREAD_PRIO_INHERIT}
20951 @cindex @code{PTHREAD_PRIO_INHERIT} policy (under rts-pthread)
20952 and @code{PTHREAD_PRIO_PROTECT}
20953 @cindex @code{PTHREAD_PRIO_PROTECT} policy (under rts-pthread)
20954 semantics that can be selected using the predefined pragma
20955 @code{Locking_Policy}
20956 @cindex pragma Locking_Policy (under rts-pthread)
20958 @code{Inheritance_Locking} and @code{Ceiling_Locking} as the policy.
20959 @cindex @code{Inheritance_Locking} (under rts-pthread)
20960 @cindex @code{Ceiling_Locking} (under rts-pthread)
20962 As explained above, the native run-time library is based on the Solaris thread
20963 library (@code{libthread}) and is the default library.
20964 The FSU run-time library is based on the FSU threads.
20965 @cindex FSU threads library
20967 Starting with Solaris 2.5.1, when the Solaris threads library is used
20968 (this is the default), programs
20969 compiled with GNAT can automatically take advantage of
20970 and can thus execute on multiple processors.
20971 The user can alternatively specify a processor on which the program should run
20972 to emulate a single-processor system. The multiprocessor / uniprocessor choice
20974 setting the environment variable @code{GNAT_PROCESSOR}
20975 @cindex @code{GNAT_PROCESSOR} environment variable (on Sparc Solaris)
20976 to one of the following:
20980 Use the default configuration (run the program on all
20981 available processors) - this is the same as having
20982 @code{GNAT_PROCESSOR} unset
20985 Let the run-time implementation choose one processor and run the program on
20988 @item 0 .. Last_Proc
20989 Run the program on the specified processor.
20990 @code{Last_Proc} is equal to @code{_SC_NPROCESSORS_CONF - 1}
20991 (where @code{_SC_NPROCESSORS_CONF} is a system variable).
20995 @node Building and Debugging 64-bit Applications
20996 @subsection Building and Debugging 64-bit Applications
20999 In a 64-bit application, all the sources involved must be compiled with the
21000 @option{-m64} command-line option, and a specific GNAT library (compiled with
21001 this option) is required.
21002 The easiest way to build a 64bit application is to add
21003 @option{-m64 --RTS=m64} to the @command{gnatmake} flags.
21005 To debug these applications, dwarf-2 debug information is required, so you
21006 have to add @option{-gdwarf-2} to your gnatmake arguments.
21007 In addition, a special
21008 version of gdb, called @command{gdb64}, needs to be used.
21010 To summarize, building and debugging a ``Hello World'' program in 64-bit mode
21014 $ gnatmake -m64 -gdwarf-2 --RTS=m64 hello.adb
21020 @node IRIX-Specific Considerations
21021 @section IRIX-Specific Considerations
21022 @cindex IRIX thread library
21025 On SGI IRIX, the thread library depends on which compiler is used.
21026 The @emph{o32 ABI} compiler comes with a run-time library based on the
21027 user-level @code{athread}
21028 library. Thus kernel-level capabilities such as nonblocking system
21029 calls or time slicing can only be achieved reliably by specifying different
21030 @code{sprocs} via the pragma @code{Task_Info}
21031 @cindex pragma Task_Info (and IRIX threads)
21033 @code{System.Task_Info} package.
21034 @cindex @code{System.Task_Info} package (and IRIX threads)
21035 See the @cite{GNAT Reference Manual} for further information.
21037 The @emph{n32 ABI} compiler comes with a run-time library based on the
21038 kernel POSIX threads and thus does not have the limitations mentioned above.
21041 @node Linux-Specific Considerations
21042 @section Linux-Specific Considerations
21043 @cindex Linux threads libraries
21046 The default thread library under GNU/Linux has the following disadvantages
21047 compared to other native thread libraries:
21050 @item The size of the task's stack is limited to 2 megabytes.
21051 @item The signal model is not POSIX compliant, which means that to send a
21052 signal to the process, you need to send the signal to all threads,
21053 e.g. by using @code{killpg()}.
21058 @c *******************************
21059 @node Example of Binder Output File
21060 @appendix Example of Binder Output File
21063 This Appendix displays the source code for @command{gnatbind}'s output
21064 file generated for a simple ``Hello World'' program.
21065 Comments have been added for clarification purposes.
21068 @smallexample @c adanocomment
21072 -- The package is called Ada_Main unless this name is actually used
21073 -- as a unit name in the partition, in which case some other unique
21077 package ada_main is
21079 Elab_Final_Code : Integer;
21080 pragma Import (C, Elab_Final_Code, "__gnat_inside_elab_final_code");
21082 -- The main program saves the parameters (argument count,
21083 -- argument values, environment pointer) in global variables
21084 -- for later access by other units including
21085 -- Ada.Command_Line.
21087 gnat_argc : Integer;
21088 gnat_argv : System.Address;
21089 gnat_envp : System.Address;
21091 -- The actual variables are stored in a library routine. This
21092 -- is useful for some shared library situations, where there
21093 -- are problems if variables are not in the library.
21095 pragma Import (C, gnat_argc);
21096 pragma Import (C, gnat_argv);
21097 pragma Import (C, gnat_envp);
21099 -- The exit status is similarly an external location
21101 gnat_exit_status : Integer;
21102 pragma Import (C, gnat_exit_status);
21104 GNAT_Version : constant String :=
21105 "GNAT Version: 3.15w (20010315)";
21106 pragma Export (C, GNAT_Version, "__gnat_version");
21108 -- This is the generated adafinal routine that performs
21109 -- finalization at the end of execution. In the case where
21110 -- Ada is the main program, this main program makes a call
21111 -- to adafinal at program termination.
21113 procedure adafinal;
21114 pragma Export (C, adafinal, "adafinal");
21116 -- This is the generated adainit routine that performs
21117 -- initialization at the start of execution. In the case
21118 -- where Ada is the main program, this main program makes
21119 -- a call to adainit at program startup.
21122 pragma Export (C, adainit, "adainit");
21124 -- This routine is called at the start of execution. It is
21125 -- a dummy routine that is used by the debugger to breakpoint
21126 -- at the start of execution.
21128 procedure Break_Start;
21129 pragma Import (C, Break_Start, "__gnat_break_start");
21131 -- This is the actual generated main program (it would be
21132 -- suppressed if the no main program switch were used). As
21133 -- required by standard system conventions, this program has
21134 -- the external name main.
21138 argv : System.Address;
21139 envp : System.Address)
21141 pragma Export (C, main, "main");
21143 -- The following set of constants give the version
21144 -- identification values for every unit in the bound
21145 -- partition. This identification is computed from all
21146 -- dependent semantic units, and corresponds to the
21147 -- string that would be returned by use of the
21148 -- Body_Version or Version attributes.
21150 type Version_32 is mod 2 ** 32;
21151 u00001 : constant Version_32 := 16#7880BEB3#;
21152 u00002 : constant Version_32 := 16#0D24CBD0#;
21153 u00003 : constant Version_32 := 16#3283DBEB#;
21154 u00004 : constant Version_32 := 16#2359F9ED#;
21155 u00005 : constant Version_32 := 16#664FB847#;
21156 u00006 : constant Version_32 := 16#68E803DF#;
21157 u00007 : constant Version_32 := 16#5572E604#;
21158 u00008 : constant Version_32 := 16#46B173D8#;
21159 u00009 : constant Version_32 := 16#156A40CF#;
21160 u00010 : constant Version_32 := 16#033DABE0#;
21161 u00011 : constant Version_32 := 16#6AB38FEA#;
21162 u00012 : constant Version_32 := 16#22B6217D#;
21163 u00013 : constant Version_32 := 16#68A22947#;
21164 u00014 : constant Version_32 := 16#18CC4A56#;
21165 u00015 : constant Version_32 := 16#08258E1B#;
21166 u00016 : constant Version_32 := 16#367D5222#;
21167 u00017 : constant Version_32 := 16#20C9ECA4#;
21168 u00018 : constant Version_32 := 16#50D32CB6#;
21169 u00019 : constant Version_32 := 16#39A8BB77#;
21170 u00020 : constant Version_32 := 16#5CF8FA2B#;
21171 u00021 : constant Version_32 := 16#2F1EB794#;
21172 u00022 : constant Version_32 := 16#31AB6444#;
21173 u00023 : constant Version_32 := 16#1574B6E9#;
21174 u00024 : constant Version_32 := 16#5109C189#;
21175 u00025 : constant Version_32 := 16#56D770CD#;
21176 u00026 : constant Version_32 := 16#02F9DE3D#;
21177 u00027 : constant Version_32 := 16#08AB6B2C#;
21178 u00028 : constant Version_32 := 16#3FA37670#;
21179 u00029 : constant Version_32 := 16#476457A0#;
21180 u00030 : constant Version_32 := 16#731E1B6E#;
21181 u00031 : constant Version_32 := 16#23C2E789#;
21182 u00032 : constant Version_32 := 16#0F1BD6A1#;
21183 u00033 : constant Version_32 := 16#7C25DE96#;
21184 u00034 : constant Version_32 := 16#39ADFFA2#;
21185 u00035 : constant Version_32 := 16#571DE3E7#;
21186 u00036 : constant Version_32 := 16#5EB646AB#;
21187 u00037 : constant Version_32 := 16#4249379B#;
21188 u00038 : constant Version_32 := 16#0357E00A#;
21189 u00039 : constant Version_32 := 16#3784FB72#;
21190 u00040 : constant Version_32 := 16#2E723019#;
21191 u00041 : constant Version_32 := 16#623358EA#;
21192 u00042 : constant Version_32 := 16#107F9465#;
21193 u00043 : constant Version_32 := 16#6843F68A#;
21194 u00044 : constant Version_32 := 16#63305874#;
21195 u00045 : constant Version_32 := 16#31E56CE1#;
21196 u00046 : constant Version_32 := 16#02917970#;
21197 u00047 : constant Version_32 := 16#6CCBA70E#;
21198 u00048 : constant Version_32 := 16#41CD4204#;
21199 u00049 : constant Version_32 := 16#572E3F58#;
21200 u00050 : constant Version_32 := 16#20729FF5#;
21201 u00051 : constant Version_32 := 16#1D4F93E8#;
21202 u00052 : constant Version_32 := 16#30B2EC3D#;
21203 u00053 : constant Version_32 := 16#34054F96#;
21204 u00054 : constant Version_32 := 16#5A199860#;
21205 u00055 : constant Version_32 := 16#0E7F912B#;
21206 u00056 : constant Version_32 := 16#5760634A#;
21207 u00057 : constant Version_32 := 16#5D851835#;
21209 -- The following Export pragmas export the version numbers
21210 -- with symbolic names ending in B (for body) or S
21211 -- (for spec) so that they can be located in a link. The
21212 -- information provided here is sufficient to track down
21213 -- the exact versions of units used in a given build.
21215 pragma Export (C, u00001, "helloB");
21216 pragma Export (C, u00002, "system__standard_libraryB");
21217 pragma Export (C, u00003, "system__standard_libraryS");
21218 pragma Export (C, u00004, "adaS");
21219 pragma Export (C, u00005, "ada__text_ioB");
21220 pragma Export (C, u00006, "ada__text_ioS");
21221 pragma Export (C, u00007, "ada__exceptionsB");
21222 pragma Export (C, u00008, "ada__exceptionsS");
21223 pragma Export (C, u00009, "gnatS");
21224 pragma Export (C, u00010, "gnat__heap_sort_aB");
21225 pragma Export (C, u00011, "gnat__heap_sort_aS");
21226 pragma Export (C, u00012, "systemS");
21227 pragma Export (C, u00013, "system__exception_tableB");
21228 pragma Export (C, u00014, "system__exception_tableS");
21229 pragma Export (C, u00015, "gnat__htableB");
21230 pragma Export (C, u00016, "gnat__htableS");
21231 pragma Export (C, u00017, "system__exceptionsS");
21232 pragma Export (C, u00018, "system__machine_state_operationsB");
21233 pragma Export (C, u00019, "system__machine_state_operationsS");
21234 pragma Export (C, u00020, "system__machine_codeS");
21235 pragma Export (C, u00021, "system__storage_elementsB");
21236 pragma Export (C, u00022, "system__storage_elementsS");
21237 pragma Export (C, u00023, "system__secondary_stackB");
21238 pragma Export (C, u00024, "system__secondary_stackS");
21239 pragma Export (C, u00025, "system__parametersB");
21240 pragma Export (C, u00026, "system__parametersS");
21241 pragma Export (C, u00027, "system__soft_linksB");
21242 pragma Export (C, u00028, "system__soft_linksS");
21243 pragma Export (C, u00029, "system__stack_checkingB");
21244 pragma Export (C, u00030, "system__stack_checkingS");
21245 pragma Export (C, u00031, "system__tracebackB");
21246 pragma Export (C, u00032, "system__tracebackS");
21247 pragma Export (C, u00033, "ada__streamsS");
21248 pragma Export (C, u00034, "ada__tagsB");
21249 pragma Export (C, u00035, "ada__tagsS");
21250 pragma Export (C, u00036, "system__string_opsB");
21251 pragma Export (C, u00037, "system__string_opsS");
21252 pragma Export (C, u00038, "interfacesS");
21253 pragma Export (C, u00039, "interfaces__c_streamsB");
21254 pragma Export (C, u00040, "interfaces__c_streamsS");
21255 pragma Export (C, u00041, "system__file_ioB");
21256 pragma Export (C, u00042, "system__file_ioS");
21257 pragma Export (C, u00043, "ada__finalizationB");
21258 pragma Export (C, u00044, "ada__finalizationS");
21259 pragma Export (C, u00045, "system__finalization_rootB");
21260 pragma Export (C, u00046, "system__finalization_rootS");
21261 pragma Export (C, u00047, "system__finalization_implementationB");
21262 pragma Export (C, u00048, "system__finalization_implementationS");
21263 pragma Export (C, u00049, "system__string_ops_concat_3B");
21264 pragma Export (C, u00050, "system__string_ops_concat_3S");
21265 pragma Export (C, u00051, "system__stream_attributesB");
21266 pragma Export (C, u00052, "system__stream_attributesS");
21267 pragma Export (C, u00053, "ada__io_exceptionsS");
21268 pragma Export (C, u00054, "system__unsigned_typesS");
21269 pragma Export (C, u00055, "system__file_control_blockS");
21270 pragma Export (C, u00056, "ada__finalization__list_controllerB");
21271 pragma Export (C, u00057, "ada__finalization__list_controllerS");
21273 -- BEGIN ELABORATION ORDER
21276 -- gnat.heap_sort_a (spec)
21277 -- gnat.heap_sort_a (body)
21278 -- gnat.htable (spec)
21279 -- gnat.htable (body)
21280 -- interfaces (spec)
21282 -- system.machine_code (spec)
21283 -- system.parameters (spec)
21284 -- system.parameters (body)
21285 -- interfaces.c_streams (spec)
21286 -- interfaces.c_streams (body)
21287 -- system.standard_library (spec)
21288 -- ada.exceptions (spec)
21289 -- system.exception_table (spec)
21290 -- system.exception_table (body)
21291 -- ada.io_exceptions (spec)
21292 -- system.exceptions (spec)
21293 -- system.storage_elements (spec)
21294 -- system.storage_elements (body)
21295 -- system.machine_state_operations (spec)
21296 -- system.machine_state_operations (body)
21297 -- system.secondary_stack (spec)
21298 -- system.stack_checking (spec)
21299 -- system.soft_links (spec)
21300 -- system.soft_links (body)
21301 -- system.stack_checking (body)
21302 -- system.secondary_stack (body)
21303 -- system.standard_library (body)
21304 -- system.string_ops (spec)
21305 -- system.string_ops (body)
21308 -- ada.streams (spec)
21309 -- system.finalization_root (spec)
21310 -- system.finalization_root (body)
21311 -- system.string_ops_concat_3 (spec)
21312 -- system.string_ops_concat_3 (body)
21313 -- system.traceback (spec)
21314 -- system.traceback (body)
21315 -- ada.exceptions (body)
21316 -- system.unsigned_types (spec)
21317 -- system.stream_attributes (spec)
21318 -- system.stream_attributes (body)
21319 -- system.finalization_implementation (spec)
21320 -- system.finalization_implementation (body)
21321 -- ada.finalization (spec)
21322 -- ada.finalization (body)
21323 -- ada.finalization.list_controller (spec)
21324 -- ada.finalization.list_controller (body)
21325 -- system.file_control_block (spec)
21326 -- system.file_io (spec)
21327 -- system.file_io (body)
21328 -- ada.text_io (spec)
21329 -- ada.text_io (body)
21331 -- END ELABORATION ORDER
21335 -- The following source file name pragmas allow the generated file
21336 -- names to be unique for different main programs. They are needed
21337 -- since the package name will always be Ada_Main.
21339 pragma Source_File_Name (ada_main, Spec_File_Name => "b~hello.ads");
21340 pragma Source_File_Name (ada_main, Body_File_Name => "b~hello.adb");
21342 -- Generated package body for Ada_Main starts here
21344 package body ada_main is
21346 -- The actual finalization is performed by calling the
21347 -- library routine in System.Standard_Library.Adafinal
21349 procedure Do_Finalize;
21350 pragma Import (C, Do_Finalize, "system__standard_library__adafinal");
21357 procedure adainit is
21359 -- These booleans are set to True once the associated unit has
21360 -- been elaborated. It is also used to avoid elaborating the
21361 -- same unit twice.
21364 pragma Import (Ada, E040, "interfaces__c_streams_E");
21367 pragma Import (Ada, E008, "ada__exceptions_E");
21370 pragma Import (Ada, E014, "system__exception_table_E");
21373 pragma Import (Ada, E053, "ada__io_exceptions_E");
21376 pragma Import (Ada, E017, "system__exceptions_E");
21379 pragma Import (Ada, E024, "system__secondary_stack_E");
21382 pragma Import (Ada, E030, "system__stack_checking_E");
21385 pragma Import (Ada, E028, "system__soft_links_E");
21388 pragma Import (Ada, E035, "ada__tags_E");
21391 pragma Import (Ada, E033, "ada__streams_E");
21394 pragma Import (Ada, E046, "system__finalization_root_E");
21397 pragma Import (Ada, E048, "system__finalization_implementation_E");
21400 pragma Import (Ada, E044, "ada__finalization_E");
21403 pragma Import (Ada, E057, "ada__finalization__list_controller_E");
21406 pragma Import (Ada, E055, "system__file_control_block_E");
21409 pragma Import (Ada, E042, "system__file_io_E");
21412 pragma Import (Ada, E006, "ada__text_io_E");
21414 -- Set_Globals is a library routine that stores away the
21415 -- value of the indicated set of global values in global
21416 -- variables within the library.
21418 procedure Set_Globals
21419 (Main_Priority : Integer;
21420 Time_Slice_Value : Integer;
21421 WC_Encoding : Character;
21422 Locking_Policy : Character;
21423 Queuing_Policy : Character;
21424 Task_Dispatching_Policy : Character;
21425 Adafinal : System.Address;
21426 Unreserve_All_Interrupts : Integer;
21427 Exception_Tracebacks : Integer);
21428 @findex __gnat_set_globals
21429 pragma Import (C, Set_Globals, "__gnat_set_globals");
21431 -- SDP_Table_Build is a library routine used to build the
21432 -- exception tables. See unit Ada.Exceptions in files
21433 -- a-except.ads/adb for full details of how zero cost
21434 -- exception handling works. This procedure, the call to
21435 -- it, and the two following tables are all omitted if the
21436 -- build is in longjmp/setjump exception mode.
21438 @findex SDP_Table_Build
21439 @findex Zero Cost Exceptions
21440 procedure SDP_Table_Build
21441 (SDP_Addresses : System.Address;
21442 SDP_Count : Natural;
21443 Elab_Addresses : System.Address;
21444 Elab_Addr_Count : Natural);
21445 pragma Import (C, SDP_Table_Build, "__gnat_SDP_Table_Build");
21447 -- Table of Unit_Exception_Table addresses. Used for zero
21448 -- cost exception handling to build the top level table.
21450 ST : aliased constant array (1 .. 23) of System.Address := (
21452 Ada.Text_Io'UET_Address,
21453 Ada.Exceptions'UET_Address,
21454 Gnat.Heap_Sort_A'UET_Address,
21455 System.Exception_Table'UET_Address,
21456 System.Machine_State_Operations'UET_Address,
21457 System.Secondary_Stack'UET_Address,
21458 System.Parameters'UET_Address,
21459 System.Soft_Links'UET_Address,
21460 System.Stack_Checking'UET_Address,
21461 System.Traceback'UET_Address,
21462 Ada.Streams'UET_Address,
21463 Ada.Tags'UET_Address,
21464 System.String_Ops'UET_Address,
21465 Interfaces.C_Streams'UET_Address,
21466 System.File_Io'UET_Address,
21467 Ada.Finalization'UET_Address,
21468 System.Finalization_Root'UET_Address,
21469 System.Finalization_Implementation'UET_Address,
21470 System.String_Ops_Concat_3'UET_Address,
21471 System.Stream_Attributes'UET_Address,
21472 System.File_Control_Block'UET_Address,
21473 Ada.Finalization.List_Controller'UET_Address);
21475 -- Table of addresses of elaboration routines. Used for
21476 -- zero cost exception handling to make sure these
21477 -- addresses are included in the top level procedure
21480 EA : aliased constant array (1 .. 23) of System.Address := (
21481 adainit'Code_Address,
21482 Do_Finalize'Code_Address,
21483 Ada.Exceptions'Elab_Spec'Address,
21484 System.Exceptions'Elab_Spec'Address,
21485 Interfaces.C_Streams'Elab_Spec'Address,
21486 System.Exception_Table'Elab_Body'Address,
21487 Ada.Io_Exceptions'Elab_Spec'Address,
21488 System.Stack_Checking'Elab_Spec'Address,
21489 System.Soft_Links'Elab_Body'Address,
21490 System.Secondary_Stack'Elab_Body'Address,
21491 Ada.Tags'Elab_Spec'Address,
21492 Ada.Tags'Elab_Body'Address,
21493 Ada.Streams'Elab_Spec'Address,
21494 System.Finalization_Root'Elab_Spec'Address,
21495 Ada.Exceptions'Elab_Body'Address,
21496 System.Finalization_Implementation'Elab_Spec'Address,
21497 System.Finalization_Implementation'Elab_Body'Address,
21498 Ada.Finalization'Elab_Spec'Address,
21499 Ada.Finalization.List_Controller'Elab_Spec'Address,
21500 System.File_Control_Block'Elab_Spec'Address,
21501 System.File_Io'Elab_Body'Address,
21502 Ada.Text_Io'Elab_Spec'Address,
21503 Ada.Text_Io'Elab_Body'Address);
21505 -- Start of processing for adainit
21509 -- Call SDP_Table_Build to build the top level procedure
21510 -- table for zero cost exception handling (omitted in
21511 -- longjmp/setjump mode).
21513 SDP_Table_Build (ST'Address, 23, EA'Address, 23);
21515 -- Call Set_Globals to record various information for
21516 -- this partition. The values are derived by the binder
21517 -- from information stored in the ali files by the compiler.
21519 @findex __gnat_set_globals
21521 (Main_Priority => -1,
21522 -- Priority of main program, -1 if no pragma Priority used
21524 Time_Slice_Value => -1,
21525 -- Time slice from Time_Slice pragma, -1 if none used
21527 WC_Encoding => 'b',
21528 -- Wide_Character encoding used, default is brackets
21530 Locking_Policy => ' ',
21531 -- Locking_Policy used, default of space means not
21532 -- specified, otherwise it is the first character of
21533 -- the policy name.
21535 Queuing_Policy => ' ',
21536 -- Queuing_Policy used, default of space means not
21537 -- specified, otherwise it is the first character of
21538 -- the policy name.
21540 Task_Dispatching_Policy => ' ',
21541 -- Task_Dispatching_Policy used, default of space means
21542 -- not specified, otherwise first character of the
21545 Adafinal => System.Null_Address,
21546 -- Address of Adafinal routine, not used anymore
21548 Unreserve_All_Interrupts => 0,
21549 -- Set true if pragma Unreserve_All_Interrupts was used
21551 Exception_Tracebacks => 0);
21552 -- Indicates if exception tracebacks are enabled
21554 Elab_Final_Code := 1;
21556 -- Now we have the elaboration calls for all units in the partition.
21557 -- The Elab_Spec and Elab_Body attributes generate references to the
21558 -- implicit elaboration procedures generated by the compiler for
21559 -- each unit that requires elaboration.
21562 Interfaces.C_Streams'Elab_Spec;
21566 Ada.Exceptions'Elab_Spec;
21569 System.Exception_Table'Elab_Body;
21573 Ada.Io_Exceptions'Elab_Spec;
21577 System.Exceptions'Elab_Spec;
21581 System.Stack_Checking'Elab_Spec;
21584 System.Soft_Links'Elab_Body;
21589 System.Secondary_Stack'Elab_Body;
21593 Ada.Tags'Elab_Spec;
21596 Ada.Tags'Elab_Body;
21600 Ada.Streams'Elab_Spec;
21604 System.Finalization_Root'Elab_Spec;
21608 Ada.Exceptions'Elab_Body;
21612 System.Finalization_Implementation'Elab_Spec;
21615 System.Finalization_Implementation'Elab_Body;
21619 Ada.Finalization'Elab_Spec;
21623 Ada.Finalization.List_Controller'Elab_Spec;
21627 System.File_Control_Block'Elab_Spec;
21631 System.File_Io'Elab_Body;
21635 Ada.Text_Io'Elab_Spec;
21638 Ada.Text_Io'Elab_Body;
21642 Elab_Final_Code := 0;
21650 procedure adafinal is
21659 -- main is actually a function, as in the ANSI C standard,
21660 -- defined to return the exit status. The three parameters
21661 -- are the argument count, argument values and environment
21664 @findex Main Program
21667 argv : System.Address;
21668 envp : System.Address)
21671 -- The initialize routine performs low level system
21672 -- initialization using a standard library routine which
21673 -- sets up signal handling and performs any other
21674 -- required setup. The routine can be found in file
21677 @findex __gnat_initialize
21678 procedure initialize;
21679 pragma Import (C, initialize, "__gnat_initialize");
21681 -- The finalize routine performs low level system
21682 -- finalization using a standard library routine. The
21683 -- routine is found in file a-final.c and in the standard
21684 -- distribution is a dummy routine that does nothing, so
21685 -- really this is a hook for special user finalization.
21687 @findex __gnat_finalize
21688 procedure finalize;
21689 pragma Import (C, finalize, "__gnat_finalize");
21691 -- We get to the main program of the partition by using
21692 -- pragma Import because if we try to with the unit and
21693 -- call it Ada style, then not only do we waste time
21694 -- recompiling it, but also, we don't really know the right
21695 -- switches (e.g. identifier character set) to be used
21698 procedure Ada_Main_Program;
21699 pragma Import (Ada, Ada_Main_Program, "_ada_hello");
21701 -- Start of processing for main
21704 -- Save global variables
21710 -- Call low level system initialization
21714 -- Call our generated Ada initialization routine
21718 -- This is the point at which we want the debugger to get
21723 -- Now we call the main program of the partition
21727 -- Perform Ada finalization
21731 -- Perform low level system finalization
21735 -- Return the proper exit status
21736 return (gnat_exit_status);
21739 -- This section is entirely comments, so it has no effect on the
21740 -- compilation of the Ada_Main package. It provides the list of
21741 -- object files and linker options, as well as some standard
21742 -- libraries needed for the link. The gnatlink utility parses
21743 -- this b~hello.adb file to read these comment lines to generate
21744 -- the appropriate command line arguments for the call to the
21745 -- system linker. The BEGIN/END lines are used for sentinels for
21746 -- this parsing operation.
21748 -- The exact file names will of course depend on the environment,
21749 -- host/target and location of files on the host system.
21751 @findex Object file list
21752 -- BEGIN Object file/option list
21755 -- -L/usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/
21756 -- /usr/local/gnat/lib/gcc-lib/i686-pc-linux-gnu/2.8.1/adalib/libgnat.a
21757 -- END Object file/option list
21763 The Ada code in the above example is exactly what is generated by the
21764 binder. We have added comments to more clearly indicate the function
21765 of each part of the generated @code{Ada_Main} package.
21767 The code is standard Ada in all respects, and can be processed by any
21768 tools that handle Ada. In particular, it is possible to use the debugger
21769 in Ada mode to debug the generated @code{Ada_Main} package. For example,
21770 suppose that for reasons that you do not understand, your program is crashing
21771 during elaboration of the body of @code{Ada.Text_IO}. To locate this bug,
21772 you can place a breakpoint on the call:
21774 @smallexample @c ada
21775 Ada.Text_Io'Elab_Body;
21779 and trace the elaboration routine for this package to find out where
21780 the problem might be (more usually of course you would be debugging
21781 elaboration code in your own application).
21784 @node Elaboration Order Handling in GNAT
21785 @appendix Elaboration Order Handling in GNAT
21786 @cindex Order of elaboration
21787 @cindex Elaboration control
21790 * Elaboration Code in Ada 95::
21791 * Checking the Elaboration Order in Ada 95::
21792 * Controlling the Elaboration Order in Ada 95::
21793 * Controlling Elaboration in GNAT - Internal Calls::
21794 * Controlling Elaboration in GNAT - External Calls::
21795 * Default Behavior in GNAT - Ensuring Safety::
21796 * Treatment of Pragma Elaborate::
21797 * Elaboration Issues for Library Tasks::
21798 * Mixing Elaboration Models::
21799 * What to Do If the Default Elaboration Behavior Fails::
21800 * Elaboration for Access-to-Subprogram Values::
21801 * Summary of Procedures for Elaboration Control::
21802 * Other Elaboration Order Considerations::
21806 This chapter describes the handling of elaboration code in Ada 95 and
21807 in GNAT, and discusses how the order of elaboration of program units can
21808 be controlled in GNAT, either automatically or with explicit programming
21811 @node Elaboration Code in Ada 95
21812 @section Elaboration Code in Ada 95
21815 Ada 95 provides rather general mechanisms for executing code at elaboration
21816 time, that is to say before the main program starts executing. Such code arises
21820 @item Initializers for variables.
21821 Variables declared at the library level, in package specs or bodies, can
21822 require initialization that is performed at elaboration time, as in:
21823 @smallexample @c ada
21825 Sqrt_Half : Float := Sqrt (0.5);
21829 @item Package initialization code
21830 Code in a @code{BEGIN-END} section at the outer level of a package body is
21831 executed as part of the package body elaboration code.
21833 @item Library level task allocators
21834 Tasks that are declared using task allocators at the library level
21835 start executing immediately and hence can execute at elaboration time.
21839 Subprogram calls are possible in any of these contexts, which means that
21840 any arbitrary part of the program may be executed as part of the elaboration
21841 code. It is even possible to write a program which does all its work at
21842 elaboration time, with a null main program, although stylistically this
21843 would usually be considered an inappropriate way to structure
21846 An important concern arises in the context of elaboration code:
21847 we have to be sure that it is executed in an appropriate order. What we
21848 have is a series of elaboration code sections, potentially one section
21849 for each unit in the program. It is important that these execute
21850 in the correct order. Correctness here means that, taking the above
21851 example of the declaration of @code{Sqrt_Half},
21852 if some other piece of
21853 elaboration code references @code{Sqrt_Half},
21854 then it must run after the
21855 section of elaboration code that contains the declaration of
21858 There would never be any order of elaboration problem if we made a rule
21859 that whenever you @code{with} a unit, you must elaborate both the spec and body
21860 of that unit before elaborating the unit doing the @code{with}'ing:
21862 @smallexample @c ada
21866 package Unit_2 is ...
21872 would require that both the body and spec of @code{Unit_1} be elaborated
21873 before the spec of @code{Unit_2}. However, a rule like that would be far too
21874 restrictive. In particular, it would make it impossible to have routines
21875 in separate packages that were mutually recursive.
21877 You might think that a clever enough compiler could look at the actual
21878 elaboration code and determine an appropriate correct order of elaboration,
21879 but in the general case, this is not possible. Consider the following
21882 In the body of @code{Unit_1}, we have a procedure @code{Func_1}
21884 the variable @code{Sqrt_1}, which is declared in the elaboration code
21885 of the body of @code{Unit_1}:
21887 @smallexample @c ada
21889 Sqrt_1 : Float := Sqrt (0.1);
21894 The elaboration code of the body of @code{Unit_1} also contains:
21896 @smallexample @c ada
21899 if expression_1 = 1 then
21900 Q := Unit_2.Func_2;
21907 @code{Unit_2} is exactly parallel,
21908 it has a procedure @code{Func_2} that references
21909 the variable @code{Sqrt_2}, which is declared in the elaboration code of
21910 the body @code{Unit_2}:
21912 @smallexample @c ada
21914 Sqrt_2 : Float := Sqrt (0.1);
21919 The elaboration code of the body of @code{Unit_2} also contains:
21921 @smallexample @c ada
21924 if expression_2 = 2 then
21925 Q := Unit_1.Func_1;
21932 Now the question is, which of the following orders of elaboration is
21957 If you carefully analyze the flow here, you will see that you cannot tell
21958 at compile time the answer to this question.
21959 If @code{expression_1} is not equal to 1,
21960 and @code{expression_2} is not equal to 2,
21961 then either order is acceptable, because neither of the function calls is
21962 executed. If both tests evaluate to true, then neither order is acceptable
21963 and in fact there is no correct order.
21965 If one of the two expressions is true, and the other is false, then one
21966 of the above orders is correct, and the other is incorrect. For example,
21967 if @code{expression_1} = 1 and @code{expression_2} /= 2,
21968 then the call to @code{Func_2}
21969 will occur, but not the call to @code{Func_1.}
21970 This means that it is essential
21971 to elaborate the body of @code{Unit_1} before
21972 the body of @code{Unit_2}, so the first
21973 order of elaboration is correct and the second is wrong.
21975 By making @code{expression_1} and @code{expression_2}
21976 depend on input data, or perhaps
21977 the time of day, we can make it impossible for the compiler or binder
21978 to figure out which of these expressions will be true, and hence it
21979 is impossible to guarantee a safe order of elaboration at run time.
21981 @node Checking the Elaboration Order in Ada 95
21982 @section Checking the Elaboration Order in Ada 95
21985 In some languages that involve the same kind of elaboration problems,
21986 e.g. Java and C++, the programmer is expected to worry about these
21987 ordering problems himself, and it is common to
21988 write a program in which an incorrect elaboration order gives
21989 surprising results, because it references variables before they
21991 Ada 95 is designed to be a safe language, and a programmer-beware approach is
21992 clearly not sufficient. Consequently, the language provides three lines
21996 @item Standard rules
21997 Some standard rules restrict the possible choice of elaboration
21998 order. In particular, if you @code{with} a unit, then its spec is always
21999 elaborated before the unit doing the @code{with}. Similarly, a parent
22000 spec is always elaborated before the child spec, and finally
22001 a spec is always elaborated before its corresponding body.
22003 @item Dynamic elaboration checks
22004 @cindex Elaboration checks
22005 @cindex Checks, elaboration
22006 Dynamic checks are made at run time, so that if some entity is accessed
22007 before it is elaborated (typically by means of a subprogram call)
22008 then the exception (@code{Program_Error}) is raised.
22010 @item Elaboration control
22011 Facilities are provided for the programmer to specify the desired order
22015 Let's look at these facilities in more detail. First, the rules for
22016 dynamic checking. One possible rule would be simply to say that the
22017 exception is raised if you access a variable which has not yet been
22018 elaborated. The trouble with this approach is that it could require
22019 expensive checks on every variable reference. Instead Ada 95 has two
22020 rules which are a little more restrictive, but easier to check, and
22024 @item Restrictions on calls
22025 A subprogram can only be called at elaboration time if its body
22026 has been elaborated. The rules for elaboration given above guarantee
22027 that the spec of the subprogram has been elaborated before the
22028 call, but not the body. If this rule is violated, then the
22029 exception @code{Program_Error} is raised.
22031 @item Restrictions on instantiations
22032 A generic unit can only be instantiated if the body of the generic
22033 unit has been elaborated. Again, the rules for elaboration given above
22034 guarantee that the spec of the generic unit has been elaborated
22035 before the instantiation, but not the body. If this rule is
22036 violated, then the exception @code{Program_Error} is raised.
22040 The idea is that if the body has been elaborated, then any variables
22041 it references must have been elaborated; by checking for the body being
22042 elaborated we guarantee that none of its references causes any
22043 trouble. As we noted above, this is a little too restrictive, because a
22044 subprogram that has no non-local references in its body may in fact be safe
22045 to call. However, it really would be unsafe to rely on this, because
22046 it would mean that the caller was aware of details of the implementation
22047 in the body. This goes against the basic tenets of Ada.
22049 A plausible implementation can be described as follows.
22050 A Boolean variable is associated with each subprogram
22051 and each generic unit. This variable is initialized to False, and is set to
22052 True at the point body is elaborated. Every call or instantiation checks the
22053 variable, and raises @code{Program_Error} if the variable is False.
22055 Note that one might think that it would be good enough to have one Boolean
22056 variable for each package, but that would not deal with cases of trying
22057 to call a body in the same package as the call
22058 that has not been elaborated yet.
22059 Of course a compiler may be able to do enough analysis to optimize away
22060 some of the Boolean variables as unnecessary, and @code{GNAT} indeed
22061 does such optimizations, but still the easiest conceptual model is to
22062 think of there being one variable per subprogram.
22064 @node Controlling the Elaboration Order in Ada 95
22065 @section Controlling the Elaboration Order in Ada 95
22068 In the previous section we discussed the rules in Ada 95 which ensure
22069 that @code{Program_Error} is raised if an incorrect elaboration order is
22070 chosen. This prevents erroneous executions, but we need mechanisms to
22071 specify a correct execution and avoid the exception altogether.
22072 To achieve this, Ada 95 provides a number of features for controlling
22073 the order of elaboration. We discuss these features in this section.
22075 First, there are several ways of indicating to the compiler that a given
22076 unit has no elaboration problems:
22079 @item packages that do not require a body
22080 In Ada 95, a library package that does not require a body does not permit
22081 a body. This means that if we have a such a package, as in:
22083 @smallexample @c ada
22086 package Definitions is
22088 type m is new integer;
22090 type a is array (1 .. 10) of m;
22091 type b is array (1 .. 20) of m;
22099 A package that @code{with}'s @code{Definitions} may safely instantiate
22100 @code{Definitions.Subp} because the compiler can determine that there
22101 definitely is no package body to worry about in this case
22104 @cindex pragma Pure
22106 Places sufficient restrictions on a unit to guarantee that
22107 no call to any subprogram in the unit can result in an
22108 elaboration problem. This means that the compiler does not need
22109 to worry about the point of elaboration of such units, and in
22110 particular, does not need to check any calls to any subprograms
22113 @item pragma Preelaborate
22114 @findex Preelaborate
22115 @cindex pragma Preelaborate
22116 This pragma places slightly less stringent restrictions on a unit than
22118 but these restrictions are still sufficient to ensure that there
22119 are no elaboration problems with any calls to the unit.
22121 @item pragma Elaborate_Body
22122 @findex Elaborate_Body
22123 @cindex pragma Elaborate_Body
22124 This pragma requires that the body of a unit be elaborated immediately
22125 after its spec. Suppose a unit @code{A} has such a pragma,
22126 and unit @code{B} does
22127 a @code{with} of unit @code{A}. Recall that the standard rules require
22128 the spec of unit @code{A}
22129 to be elaborated before the @code{with}'ing unit; given the pragma in
22130 @code{A}, we also know that the body of @code{A}
22131 will be elaborated before @code{B}, so
22132 that calls to @code{A} are safe and do not need a check.
22137 unlike pragma @code{Pure} and pragma @code{Preelaborate},
22139 @code{Elaborate_Body} does not guarantee that the program is
22140 free of elaboration problems, because it may not be possible
22141 to satisfy the requested elaboration order.
22142 Let's go back to the example with @code{Unit_1} and @code{Unit_2}.
22144 marks @code{Unit_1} as @code{Elaborate_Body},
22145 and not @code{Unit_2,} then the order of
22146 elaboration will be:
22158 Now that means that the call to @code{Func_1} in @code{Unit_2}
22159 need not be checked,
22160 it must be safe. But the call to @code{Func_2} in
22161 @code{Unit_1} may still fail if
22162 @code{Expression_1} is equal to 1,
22163 and the programmer must still take
22164 responsibility for this not being the case.
22166 If all units carry a pragma @code{Elaborate_Body}, then all problems are
22167 eliminated, except for calls entirely within a body, which are
22168 in any case fully under programmer control. However, using the pragma
22169 everywhere is not always possible.
22170 In particular, for our @code{Unit_1}/@code{Unit_2} example, if
22171 we marked both of them as having pragma @code{Elaborate_Body}, then
22172 clearly there would be no possible elaboration order.
22174 The above pragmas allow a server to guarantee safe use by clients, and
22175 clearly this is the preferable approach. Consequently a good rule in
22176 Ada 95 is to mark units as @code{Pure} or @code{Preelaborate} if possible,
22177 and if this is not possible,
22178 mark them as @code{Elaborate_Body} if possible.
22179 As we have seen, there are situations where neither of these
22180 three pragmas can be used.
22181 So we also provide methods for clients to control the
22182 order of elaboration of the servers on which they depend:
22185 @item pragma Elaborate (unit)
22187 @cindex pragma Elaborate
22188 This pragma is placed in the context clause, after a @code{with} clause,
22189 and it requires that the body of the named unit be elaborated before
22190 the unit in which the pragma occurs. The idea is to use this pragma
22191 if the current unit calls at elaboration time, directly or indirectly,
22192 some subprogram in the named unit.
22194 @item pragma Elaborate_All (unit)
22195 @findex Elaborate_All
22196 @cindex pragma Elaborate_All
22197 This is a stronger version of the Elaborate pragma. Consider the
22201 Unit A @code{with}'s unit B and calls B.Func in elab code
22202 Unit B @code{with}'s unit C, and B.Func calls C.Func
22206 Now if we put a pragma @code{Elaborate (B)}
22207 in unit @code{A}, this ensures that the
22208 body of @code{B} is elaborated before the call, but not the
22209 body of @code{C}, so
22210 the call to @code{C.Func} could still cause @code{Program_Error} to
22213 The effect of a pragma @code{Elaborate_All} is stronger, it requires
22214 not only that the body of the named unit be elaborated before the
22215 unit doing the @code{with}, but also the bodies of all units that the
22216 named unit uses, following @code{with} links transitively. For example,
22217 if we put a pragma @code{Elaborate_All (B)} in unit @code{A},
22219 not only that the body of @code{B} be elaborated before @code{A},
22221 body of @code{C}, because @code{B} @code{with}'s @code{C}.
22225 We are now in a position to give a usage rule in Ada 95 for avoiding
22226 elaboration problems, at least if dynamic dispatching and access to
22227 subprogram values are not used. We will handle these cases separately
22230 The rule is simple. If a unit has elaboration code that can directly or
22231 indirectly make a call to a subprogram in a @code{with}'ed unit, or instantiate
22232 a generic unit in a @code{with}'ed unit,
22233 then if the @code{with}'ed unit does not have
22234 pragma @code{Pure} or @code{Preelaborate}, then the client should have
22235 a pragma @code{Elaborate_All}
22236 for the @code{with}'ed unit. By following this rule a client is
22237 assured that calls can be made without risk of an exception.
22238 If this rule is not followed, then a program may be in one of four
22242 @item No order exists
22243 No order of elaboration exists which follows the rules, taking into
22244 account any @code{Elaborate}, @code{Elaborate_All},
22245 or @code{Elaborate_Body} pragmas. In
22246 this case, an Ada 95 compiler must diagnose the situation at bind
22247 time, and refuse to build an executable program.
22249 @item One or more orders exist, all incorrect
22250 One or more acceptable elaboration orders exists, and all of them
22251 generate an elaboration order problem. In this case, the binder
22252 can build an executable program, but @code{Program_Error} will be raised
22253 when the program is run.
22255 @item Several orders exist, some right, some incorrect
22256 One or more acceptable elaboration orders exists, and some of them
22257 work, and some do not. The programmer has not controlled
22258 the order of elaboration, so the binder may or may not pick one of
22259 the correct orders, and the program may or may not raise an
22260 exception when it is run. This is the worst case, because it means
22261 that the program may fail when moved to another compiler, or even
22262 another version of the same compiler.
22264 @item One or more orders exists, all correct
22265 One ore more acceptable elaboration orders exist, and all of them
22266 work. In this case the program runs successfully. This state of
22267 affairs can be guaranteed by following the rule we gave above, but
22268 may be true even if the rule is not followed.
22272 Note that one additional advantage of following our Elaborate_All rule
22273 is that the program continues to stay in the ideal (all orders OK) state
22274 even if maintenance
22275 changes some bodies of some subprograms. Conversely, if a program that does
22276 not follow this rule happens to be safe at some point, this state of affairs
22277 may deteriorate silently as a result of maintenance changes.
22279 You may have noticed that the above discussion did not mention
22280 the use of @code{Elaborate_Body}. This was a deliberate omission. If you
22281 @code{with} an @code{Elaborate_Body} unit, it still may be the case that
22282 code in the body makes calls to some other unit, so it is still necessary
22283 to use @code{Elaborate_All} on such units.
22285 @node Controlling Elaboration in GNAT - Internal Calls
22286 @section Controlling Elaboration in GNAT - Internal Calls
22289 In the case of internal calls, i.e. calls within a single package, the
22290 programmer has full control over the order of elaboration, and it is up
22291 to the programmer to elaborate declarations in an appropriate order. For
22294 @smallexample @c ada
22297 function One return Float;
22301 function One return Float is
22310 will obviously raise @code{Program_Error} at run time, because function
22311 One will be called before its body is elaborated. In this case GNAT will
22312 generate a warning that the call will raise @code{Program_Error}:
22318 2. function One return Float;
22320 4. Q : Float := One;
22322 >>> warning: cannot call "One" before body is elaborated
22323 >>> warning: Program_Error will be raised at run time
22326 6. function One return Float is
22339 Note that in this particular case, it is likely that the call is safe, because
22340 the function @code{One} does not access any global variables.
22341 Nevertheless in Ada 95, we do not want the validity of the check to depend on
22342 the contents of the body (think about the separate compilation case), so this
22343 is still wrong, as we discussed in the previous sections.
22345 The error is easily corrected by rearranging the declarations so that the
22346 body of One appears before the declaration containing the call
22347 (note that in Ada 95,
22348 declarations can appear in any order, so there is no restriction that
22349 would prevent this reordering, and if we write:
22351 @smallexample @c ada
22354 function One return Float;
22356 function One return Float is
22367 then all is well, no warning is generated, and no
22368 @code{Program_Error} exception
22370 Things are more complicated when a chain of subprograms is executed:
22372 @smallexample @c ada
22375 function A return Integer;
22376 function B return Integer;
22377 function C return Integer;
22379 function B return Integer is begin return A; end;
22380 function C return Integer is begin return B; end;
22384 function A return Integer is begin return 1; end;
22390 Now the call to @code{C}
22391 at elaboration time in the declaration of @code{X} is correct, because
22392 the body of @code{C} is already elaborated,
22393 and the call to @code{B} within the body of
22394 @code{C} is correct, but the call
22395 to @code{A} within the body of @code{B} is incorrect, because the body
22396 of @code{A} has not been elaborated, so @code{Program_Error}
22397 will be raised on the call to @code{A}.
22398 In this case GNAT will generate a
22399 warning that @code{Program_Error} may be
22400 raised at the point of the call. Let's look at the warning:
22406 2. function A return Integer;
22407 3. function B return Integer;
22408 4. function C return Integer;
22410 6. function B return Integer is begin return A; end;
22412 >>> warning: call to "A" before body is elaborated may
22413 raise Program_Error
22414 >>> warning: "B" called at line 7
22415 >>> warning: "C" called at line 9
22417 7. function C return Integer is begin return B; end;
22419 9. X : Integer := C;
22421 11. function A return Integer is begin return 1; end;
22431 Note that the message here says ``may raise'', instead of the direct case,
22432 where the message says ``will be raised''. That's because whether
22434 actually called depends in general on run-time flow of control.
22435 For example, if the body of @code{B} said
22437 @smallexample @c ada
22440 function B return Integer is
22442 if some-condition-depending-on-input-data then
22453 then we could not know until run time whether the incorrect call to A would
22454 actually occur, so @code{Program_Error} might
22455 or might not be raised. It is possible for a compiler to
22456 do a better job of analyzing bodies, to
22457 determine whether or not @code{Program_Error}
22458 might be raised, but it certainly
22459 couldn't do a perfect job (that would require solving the halting problem
22460 and is provably impossible), and because this is a warning anyway, it does
22461 not seem worth the effort to do the analysis. Cases in which it
22462 would be relevant are rare.
22464 In practice, warnings of either of the forms given
22465 above will usually correspond to
22466 real errors, and should be examined carefully and eliminated.
22467 In the rare case where a warning is bogus, it can be suppressed by any of
22468 the following methods:
22472 Compile with the @option{-gnatws} switch set
22475 Suppress @code{Elaboration_Check} for the called subprogram
22478 Use pragma @code{Warnings_Off} to turn warnings off for the call
22482 For the internal elaboration check case,
22483 GNAT by default generates the
22484 necessary run-time checks to ensure
22485 that @code{Program_Error} is raised if any
22486 call fails an elaboration check. Of course this can only happen if a
22487 warning has been issued as described above. The use of pragma
22488 @code{Suppress (Elaboration_Check)} may (but is not guaranteed to) suppress
22489 some of these checks, meaning that it may be possible (but is not
22490 guaranteed) for a program to be able to call a subprogram whose body
22491 is not yet elaborated, without raising a @code{Program_Error} exception.
22493 @node Controlling Elaboration in GNAT - External Calls
22494 @section Controlling Elaboration in GNAT - External Calls
22497 The previous section discussed the case in which the execution of a
22498 particular thread of elaboration code occurred entirely within a
22499 single unit. This is the easy case to handle, because a programmer
22500 has direct and total control over the order of elaboration, and
22501 furthermore, checks need only be generated in cases which are rare
22502 and which the compiler can easily detect.
22503 The situation is more complex when separate compilation is taken into account.
22504 Consider the following:
22506 @smallexample @c ada
22510 function Sqrt (Arg : Float) return Float;
22513 package body Math is
22514 function Sqrt (Arg : Float) return Float is
22523 X : Float := Math.Sqrt (0.5);
22536 where @code{Main} is the main program. When this program is executed, the
22537 elaboration code must first be executed, and one of the jobs of the
22538 binder is to determine the order in which the units of a program are
22539 to be elaborated. In this case we have four units: the spec and body
22541 the spec of @code{Stuff} and the body of @code{Main}).
22542 In what order should the four separate sections of elaboration code
22545 There are some restrictions in the order of elaboration that the binder
22546 can choose. In particular, if unit U has a @code{with}
22547 for a package @code{X}, then you
22548 are assured that the spec of @code{X}
22549 is elaborated before U , but you are
22550 not assured that the body of @code{X}
22551 is elaborated before U.
22552 This means that in the above case, the binder is allowed to choose the
22563 but that's not good, because now the call to @code{Math.Sqrt}
22564 that happens during
22565 the elaboration of the @code{Stuff}
22566 spec happens before the body of @code{Math.Sqrt} is
22567 elaborated, and hence causes @code{Program_Error} exception to be raised.
22568 At first glance, one might say that the binder is misbehaving, because
22569 obviously you want to elaborate the body of something you @code{with}
22571 that is not a general rule that can be followed in all cases. Consider
22573 @smallexample @c ada
22581 package body Y is ...
22584 package body X is ...
22590 This is a common arrangement, and, apart from the order of elaboration
22591 problems that might arise in connection with elaboration code, this works fine.
22592 A rule that says that you must first elaborate the body of anything you
22593 @code{with} cannot work in this case:
22594 the body of @code{X} @code{with}'s @code{Y},
22595 which means you would have to
22596 elaborate the body of @code{Y} first, but that @code{with}'s @code{X},
22598 you have to elaborate the body of @code{X} first, but ... and we have a
22599 loop that cannot be broken.
22601 It is true that the binder can in many cases guess an order of elaboration
22602 that is unlikely to cause a @code{Program_Error}
22603 exception to be raised, and it tries to do so (in the
22604 above example of @code{Math/Stuff/Spec}, the GNAT binder will
22606 elaborate the body of @code{Math} right after its spec, so all will be well).
22608 However, a program that blindly relies on the binder to be helpful can
22609 get into trouble, as we discussed in the previous sections, so
22611 provides a number of facilities for assisting the programmer in
22612 developing programs that are robust with respect to elaboration order.
22614 @node Default Behavior in GNAT - Ensuring Safety
22615 @section Default Behavior in GNAT - Ensuring Safety
22618 The default behavior in GNAT ensures elaboration safety. In its
22619 default mode GNAT implements the
22620 rule we previously described as the right approach. Let's restate it:
22624 @emph{If a unit has elaboration code that can directly or indirectly make a
22625 call to a subprogram in a @code{with}'ed unit, or instantiate a generic unit
22626 in a @code{with}'ed unit, then if the @code{with}'ed unit
22627 does not have pragma @code{Pure} or
22628 @code{Preelaborate}, then the client should have an
22629 @code{Elaborate_All} for the @code{with}'ed unit.}
22633 By following this rule a client is assured that calls and instantiations
22634 can be made without risk of an exception.
22636 In this mode GNAT traces all calls that are potentially made from
22637 elaboration code, and puts in any missing implicit @code{Elaborate_All}
22639 The advantage of this approach is that no elaboration problems
22640 are possible if the binder can find an elaboration order that is
22641 consistent with these implicit @code{Elaborate_All} pragmas. The
22642 disadvantage of this approach is that no such order may exist.
22644 If the binder does not generate any diagnostics, then it means that it
22645 has found an elaboration order that is guaranteed to be safe. However,
22646 the binder may still be relying on implicitly generated
22647 @code{Elaborate_All} pragmas so portability to other compilers than
22648 GNAT is not guaranteed.
22650 If it is important to guarantee portability, then the compilations should
22653 (warn on elaboration problems) switch. This will cause warning messages
22654 to be generated indicating the missing @code{Elaborate_All} pragmas.
22655 Consider the following source program:
22657 @smallexample @c ada
22662 m : integer := k.r;
22669 where it is clear that there
22670 should be a pragma @code{Elaborate_All}
22671 for unit @code{k}. An implicit pragma will be generated, and it is
22672 likely that the binder will be able to honor it. However, if you want
22673 to port this program to some other Ada compiler than GNAT.
22674 it is safer to include the pragma explicitly in the source. If this
22675 unit is compiled with the
22677 switch, then the compiler outputs a warning:
22684 3. m : integer := k.r;
22686 >>> warning: call to "r" may raise Program_Error
22687 >>> warning: missing pragma Elaborate_All for "k"
22695 and these warnings can be used as a guide for supplying manually
22696 the missing pragmas. It is usually a bad idea to use this warning
22697 option during development. That's because it will warn you when
22698 you need to put in a pragma, but cannot warn you when it is time
22699 to take it out. So the use of pragma Elaborate_All may lead to
22700 unnecessary dependencies and even false circularities.
22702 This default mode is more restrictive than the Ada Reference
22703 Manual, and it is possible to construct programs which will compile
22704 using the dynamic model described there, but will run into a
22705 circularity using the safer static model we have described.
22707 Of course any Ada compiler must be able to operate in a mode
22708 consistent with the requirements of the Ada Reference Manual,
22709 and in particular must have the capability of implementing the
22710 standard dynamic model of elaboration with run-time checks.
22712 In GNAT, this standard mode can be achieved either by the use of
22713 the @option{-gnatE} switch on the compiler (@code{gcc} or @code{gnatmake})
22714 command, or by the use of the configuration pragma:
22716 @smallexample @c ada
22717 pragma Elaboration_Checks (RM);
22721 Either approach will cause the unit affected to be compiled using the
22722 standard dynamic run-time elaboration checks described in the Ada
22723 Reference Manual. The static model is generally preferable, since it
22724 is clearly safer to rely on compile and link time checks rather than
22725 run-time checks. However, in the case of legacy code, it may be
22726 difficult to meet the requirements of the static model. This
22727 issue is further discussed in
22728 @ref{What to Do If the Default Elaboration Behavior Fails}.
22730 Note that the static model provides a strict subset of the allowed
22731 behavior and programs of the Ada Reference Manual, so if you do
22732 adhere to the static model and no circularities exist,
22733 then you are assured that your program will
22734 work using the dynamic model, providing that you remove any
22735 pragma Elaborate statements from the source.
22737 @node Treatment of Pragma Elaborate
22738 @section Treatment of Pragma Elaborate
22739 @cindex Pragma Elaborate
22742 The use of @code{pragma Elaborate}
22743 should generally be avoided in Ada 95 programs.
22744 The reason for this is that there is no guarantee that transitive calls
22745 will be properly handled. Indeed at one point, this pragma was placed
22746 in Annex J (Obsolescent Features), on the grounds that it is never useful.
22748 Now that's a bit restrictive. In practice, the case in which
22749 @code{pragma Elaborate} is useful is when the caller knows that there
22750 are no transitive calls, or that the called unit contains all necessary
22751 transitive @code{pragma Elaborate} statements, and legacy code often
22752 contains such uses.
22754 Strictly speaking the static mode in GNAT should ignore such pragmas,
22755 since there is no assurance at compile time that the necessary safety
22756 conditions are met. In practice, this would cause GNAT to be incompatible
22757 with correctly written Ada 83 code that had all necessary
22758 @code{pragma Elaborate} statements in place. Consequently, we made the
22759 decision that GNAT in its default mode will believe that if it encounters
22760 a @code{pragma Elaborate} then the programmer knows what they are doing,
22761 and it will trust that no elaboration errors can occur.
22763 The result of this decision is two-fold. First to be safe using the
22764 static mode, you should remove all @code{pragma Elaborate} statements.
22765 Second, when fixing circularities in existing code, you can selectively
22766 use @code{pragma Elaborate} statements to convince the static mode of
22767 GNAT that it need not generate an implicit @code{pragma Elaborate_All}
22770 When using the static mode with @option{-gnatwl}, any use of
22771 @code{pragma Elaborate} will generate a warning about possible
22774 @node Elaboration Issues for Library Tasks
22775 @section Elaboration Issues for Library Tasks
22776 @cindex Library tasks, elaboration issues
22777 @cindex Elaboration of library tasks
22780 In this section we examine special elaboration issues that arise for
22781 programs that declare library level tasks.
22783 Generally the model of execution of an Ada program is that all units are
22784 elaborated, and then execution of the program starts. However, the
22785 declaration of library tasks definitely does not fit this model. The
22786 reason for this is that library tasks start as soon as they are declared
22787 (more precisely, as soon as the statement part of the enclosing package
22788 body is reached), that is to say before elaboration
22789 of the program is complete. This means that if such a task calls a
22790 subprogram, or an entry in another task, the callee may or may not be
22791 elaborated yet, and in the standard
22792 Reference Manual model of dynamic elaboration checks, you can even
22793 get timing dependent Program_Error exceptions, since there can be
22794 a race between the elaboration code and the task code.
22796 The static model of elaboration in GNAT seeks to avoid all such
22797 dynamic behavior, by being conservative, and the conservative
22798 approach in this particular case is to assume that all the code
22799 in a task body is potentially executed at elaboration time if
22800 a task is declared at the library level.
22802 This can definitely result in unexpected circularities. Consider
22803 the following example
22805 @smallexample @c ada
22811 type My_Int is new Integer;
22813 function Ident (M : My_Int) return My_Int;
22817 package body Decls is
22818 task body Lib_Task is
22824 function Ident (M : My_Int) return My_Int is
22832 procedure Put_Val (Arg : Decls.My_Int);
22836 package body Utils is
22837 procedure Put_Val (Arg : Decls.My_Int) is
22839 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
22846 Decls.Lib_Task.Start;
22851 If the above example is compiled in the default static elaboration
22852 mode, then a circularity occurs. The circularity comes from the call
22853 @code{Utils.Put_Val} in the task body of @code{Decls.Lib_Task}. Since
22854 this call occurs in elaboration code, we need an implicit pragma
22855 @code{Elaborate_All} for @code{Utils}. This means that not only must
22856 the spec and body of @code{Utils} be elaborated before the body
22857 of @code{Decls}, but also the spec and body of any unit that is
22858 @code{with'ed} by the body of @code{Utils} must also be elaborated before
22859 the body of @code{Decls}. This is the transitive implication of
22860 pragma @code{Elaborate_All} and it makes sense, because in general
22861 the body of @code{Put_Val} might have a call to something in a
22862 @code{with'ed} unit.
22864 In this case, the body of Utils (actually its spec) @code{with's}
22865 @code{Decls}. Unfortunately this means that the body of @code{Decls}
22866 must be elaborated before itself, in case there is a call from the
22867 body of @code{Utils}.
22869 Here is the exact chain of events we are worrying about:
22873 In the body of @code{Decls} a call is made from within the body of a library
22874 task to a subprogram in the package @code{Utils}. Since this call may
22875 occur at elaboration time (given that the task is activated at elaboration
22876 time), we have to assume the worst, i.e. that the
22877 call does happen at elaboration time.
22880 This means that the body and spec of @code{Util} must be elaborated before
22881 the body of @code{Decls} so that this call does not cause an access before
22885 Within the body of @code{Util}, specifically within the body of
22886 @code{Util.Put_Val} there may be calls to any unit @code{with}'ed
22890 One such @code{with}'ed package is package @code{Decls}, so there
22891 might be a call to a subprogram in @code{Decls} in @code{Put_Val}.
22892 In fact there is such a call in this example, but we would have to
22893 assume that there was such a call even if it were not there, since
22894 we are not supposed to write the body of @code{Decls} knowing what
22895 is in the body of @code{Utils}; certainly in the case of the
22896 static elaboration model, the compiler does not know what is in
22897 other bodies and must assume the worst.
22900 This means that the spec and body of @code{Decls} must also be
22901 elaborated before we elaborate the unit containing the call, but
22902 that unit is @code{Decls}! This means that the body of @code{Decls}
22903 must be elaborated before itself, and that's a circularity.
22907 Indeed, if you add an explicit pragma Elaborate_All for @code{Utils} in
22908 the body of @code{Decls} you will get a true Ada Reference Manual
22909 circularity that makes the program illegal.
22911 In practice, we have found that problems with the static model of
22912 elaboration in existing code often arise from library tasks, so
22913 we must address this particular situation.
22915 Note that if we compile and run the program above, using the dynamic model of
22916 elaboration (that is to say use the @option{-gnatE} switch),
22917 then it compiles, binds,
22918 links, and runs, printing the expected result of 2. Therefore in some sense
22919 the circularity here is only apparent, and we need to capture
22920 the properties of this program that distinguish it from other library-level
22921 tasks that have real elaboration problems.
22923 We have four possible answers to this question:
22928 Use the dynamic model of elaboration.
22930 If we use the @option{-gnatE} switch, then as noted above, the program works.
22931 Why is this? If we examine the task body, it is apparent that the task cannot
22933 @code{accept} statement until after elaboration has been completed, because
22934 the corresponding entry call comes from the main program, not earlier.
22935 This is why the dynamic model works here. But that's really giving
22936 up on a precise analysis, and we prefer to take this approach only if we cannot
22938 problem in any other manner. So let us examine two ways to reorganize
22939 the program to avoid the potential elaboration problem.
22942 Split library tasks into separate packages.
22944 Write separate packages, so that library tasks are isolated from
22945 other declarations as much as possible. Let us look at a variation on
22948 @smallexample @c ada
22956 package body Decls1 is
22957 task body Lib_Task is
22965 type My_Int is new Integer;
22966 function Ident (M : My_Int) return My_Int;
22970 package body Decls2 is
22971 function Ident (M : My_Int) return My_Int is
22979 procedure Put_Val (Arg : Decls2.My_Int);
22983 package body Utils is
22984 procedure Put_Val (Arg : Decls2.My_Int) is
22986 Text_IO.Put_Line (Decls2.My_Int'Image (Decls2.Ident (Arg)));
22993 Decls1.Lib_Task.Start;
22998 All we have done is to split @code{Decls} into two packages, one
22999 containing the library task, and one containing everything else. Now
23000 there is no cycle, and the program compiles, binds, links and executes
23001 using the default static model of elaboration.
23004 Declare separate task types.
23006 A significant part of the problem arises because of the use of the
23007 single task declaration form. This means that the elaboration of
23008 the task type, and the elaboration of the task itself (i.e. the
23009 creation of the task) happen at the same time. A good rule
23010 of style in Ada 95 is to always create explicit task types. By
23011 following the additional step of placing task objects in separate
23012 packages from the task type declaration, many elaboration problems
23013 are avoided. Here is another modified example of the example program:
23015 @smallexample @c ada
23017 task type Lib_Task_Type is
23021 type My_Int is new Integer;
23023 function Ident (M : My_Int) return My_Int;
23027 package body Decls is
23028 task body Lib_Task_Type is
23034 function Ident (M : My_Int) return My_Int is
23042 procedure Put_Val (Arg : Decls.My_Int);
23046 package body Utils is
23047 procedure Put_Val (Arg : Decls.My_Int) is
23049 Text_IO.Put_Line (Decls.My_Int'Image (Decls.Ident (Arg)));
23055 Lib_Task : Decls.Lib_Task_Type;
23061 Declst.Lib_Task.Start;
23066 What we have done here is to replace the @code{task} declaration in
23067 package @code{Decls} with a @code{task type} declaration. Then we
23068 introduce a separate package @code{Declst} to contain the actual
23069 task object. This separates the elaboration issues for
23070 the @code{task type}
23071 declaration, which causes no trouble, from the elaboration issues
23072 of the task object, which is also unproblematic, since it is now independent
23073 of the elaboration of @code{Utils}.
23074 This separation of concerns also corresponds to
23075 a generally sound engineering principle of separating declarations
23076 from instances. This version of the program also compiles, binds, links,
23077 and executes, generating the expected output.
23080 Use No_Entry_Calls_In_Elaboration_Code restriction.
23081 @cindex No_Entry_Calls_In_Elaboration_Code
23083 The previous two approaches described how a program can be restructured
23084 to avoid the special problems caused by library task bodies. in practice,
23085 however, such restructuring may be difficult to apply to existing legacy code,
23086 so we must consider solutions that do not require massive rewriting.
23088 Let us consider more carefully why our original sample program works
23089 under the dynamic model of elaboration. The reason is that the code
23090 in the task body blocks immediately on the @code{accept}
23091 statement. Now of course there is nothing to prohibit elaboration
23092 code from making entry calls (for example from another library level task),
23093 so we cannot tell in isolation that
23094 the task will not execute the accept statement during elaboration.
23096 However, in practice it is very unusual to see elaboration code
23097 make any entry calls, and the pattern of tasks starting
23098 at elaboration time and then immediately blocking on @code{accept} or
23099 @code{select} statements is very common. What this means is that
23100 the compiler is being too pessimistic when it analyzes the
23101 whole package body as though it might be executed at elaboration
23104 If we know that the elaboration code contains no entry calls, (a very safe
23105 assumption most of the time, that could almost be made the default
23106 behavior), then we can compile all units of the program under control
23107 of the following configuration pragma:
23110 pragma Restrictions (No_Entry_Calls_In_Elaboration_Code);
23114 This pragma can be placed in the @file{gnat.adc} file in the usual
23115 manner. If we take our original unmodified program and compile it
23116 in the presence of a @file{gnat.adc} containing the above pragma,
23117 then once again, we can compile, bind, link, and execute, obtaining
23118 the expected result. In the presence of this pragma, the compiler does
23119 not trace calls in a task body, that appear after the first @code{accept}
23120 or @code{select} statement, and therefore does not report a potential
23121 circularity in the original program.
23123 The compiler will check to the extent it can that the above
23124 restriction is not violated, but it is not always possible to do a
23125 complete check at compile time, so it is important to use this
23126 pragma only if the stated restriction is in fact met, that is to say
23127 no task receives an entry call before elaboration of all units is completed.
23131 @node Mixing Elaboration Models
23132 @section Mixing Elaboration Models
23134 So far, we have assumed that the entire program is either compiled
23135 using the dynamic model or static model, ensuring consistency. It
23136 is possible to mix the two models, but rules have to be followed
23137 if this mixing is done to ensure that elaboration checks are not
23140 The basic rule is that @emph{a unit compiled with the static model cannot
23141 be @code{with'ed} by a unit compiled with the dynamic model}. The
23142 reason for this is that in the static model, a unit assumes that
23143 its clients guarantee to use (the equivalent of) pragma
23144 @code{Elaborate_All} so that no elaboration checks are required
23145 in inner subprograms, and this assumption is violated if the
23146 client is compiled with dynamic checks.
23148 The precise rule is as follows. A unit that is compiled with dynamic
23149 checks can only @code{with} a unit that meets at least one of the
23150 following criteria:
23155 The @code{with'ed} unit is itself compiled with dynamic elaboration
23156 checks (that is with the @option{-gnatE} switch.
23159 The @code{with'ed} unit is an internal GNAT implementation unit from
23160 the System, Interfaces, Ada, or GNAT hierarchies.
23163 The @code{with'ed} unit has pragma Preelaborate or pragma Pure.
23166 The @code{with'ing} unit (that is the client) has an explicit pragma
23167 @code{Elaborate_All} for the @code{with'ed} unit.
23172 If this rule is violated, that is if a unit with dynamic elaboration
23173 checks @code{with's} a unit that does not meet one of the above four
23174 criteria, then the binder (@code{gnatbind}) will issue a warning
23175 similar to that in the following example:
23178 warning: "x.ads" has dynamic elaboration checks and with's
23179 warning: "y.ads" which has static elaboration checks
23183 These warnings indicate that the rule has been violated, and that as a result
23184 elaboration checks may be missed in the resulting executable file.
23185 This warning may be suppressed using the @option{-ws} binder switch
23186 in the usual manner.
23188 One useful application of this mixing rule is in the case of a subsystem
23189 which does not itself @code{with} units from the remainder of the
23190 application. In this case, the entire subsystem can be compiled with
23191 dynamic checks to resolve a circularity in the subsystem, while
23192 allowing the main application that uses this subsystem to be compiled
23193 using the more reliable default static model.
23195 @node What to Do If the Default Elaboration Behavior Fails
23196 @section What to Do If the Default Elaboration Behavior Fails
23199 If the binder cannot find an acceptable order, it outputs detailed
23200 diagnostics. For example:
23206 error: elaboration circularity detected
23207 info: "proc (body)" must be elaborated before "pack (body)"
23208 info: reason: Elaborate_All probably needed in unit "pack (body)"
23209 info: recompile "pack (body)" with -gnatwl
23210 info: for full details
23211 info: "proc (body)"
23212 info: is needed by its spec:
23213 info: "proc (spec)"
23214 info: which is withed by:
23215 info: "pack (body)"
23216 info: "pack (body)" must be elaborated before "proc (body)"
23217 info: reason: pragma Elaborate in unit "proc (body)"
23223 In this case we have a cycle that the binder cannot break. On the one
23224 hand, there is an explicit pragma Elaborate in @code{proc} for
23225 @code{pack}. This means that the body of @code{pack} must be elaborated
23226 before the body of @code{proc}. On the other hand, there is elaboration
23227 code in @code{pack} that calls a subprogram in @code{proc}. This means
23228 that for maximum safety, there should really be a pragma
23229 Elaborate_All in @code{pack} for @code{proc} which would require that
23230 the body of @code{proc} be elaborated before the body of
23231 @code{pack}. Clearly both requirements cannot be satisfied.
23232 Faced with a circularity of this kind, you have three different options.
23235 @item Fix the program
23236 The most desirable option from the point of view of long-term maintenance
23237 is to rearrange the program so that the elaboration problems are avoided.
23238 One useful technique is to place the elaboration code into separate
23239 child packages. Another is to move some of the initialization code to
23240 explicitly called subprograms, where the program controls the order
23241 of initialization explicitly. Although this is the most desirable option,
23242 it may be impractical and involve too much modification, especially in
23243 the case of complex legacy code.
23245 @item Perform dynamic checks
23246 If the compilations are done using the
23248 (dynamic elaboration check) switch, then GNAT behaves in
23249 a quite different manner. Dynamic checks are generated for all calls
23250 that could possibly result in raising an exception. With this switch,
23251 the compiler does not generate implicit @code{Elaborate_All} pragmas.
23252 The behavior then is exactly as specified in the Ada 95 Reference Manual.
23253 The binder will generate an executable program that may or may not
23254 raise @code{Program_Error}, and then it is the programmer's job to ensure
23255 that it does not raise an exception. Note that it is important to
23256 compile all units with the switch, it cannot be used selectively.
23258 @item Suppress checks
23259 The drawback of dynamic checks is that they generate a
23260 significant overhead at run time, both in space and time. If you
23261 are absolutely sure that your program cannot raise any elaboration
23262 exceptions, and you still want to use the dynamic elaboration model,
23263 then you can use the configuration pragma
23264 @code{Suppress (Elaboration_Check)} to suppress all such checks. For
23265 example this pragma could be placed in the @file{gnat.adc} file.
23267 @item Suppress checks selectively
23268 When you know that certain calls in elaboration code cannot possibly
23269 lead to an elaboration error, and the binder nevertheless generates warnings
23270 on those calls and inserts Elaborate_All pragmas that lead to elaboration
23271 circularities, it is possible to remove those warnings locally and obtain
23272 a program that will bind. Clearly this can be unsafe, and it is the
23273 responsibility of the programmer to make sure that the resulting program has
23274 no elaboration anomalies. The pragma @code{Suppress (Elaboration_Check)} can
23275 be used with different granularity to suppress warnings and break
23276 elaboration circularities:
23280 Place the pragma that names the called subprogram in the declarative part
23281 that contains the call.
23284 Place the pragma in the declarative part, without naming an entity. This
23285 disables warnings on all calls in the corresponding declarative region.
23288 Place the pragma in the package spec that declares the called subprogram,
23289 and name the subprogram. This disables warnings on all elaboration calls to
23293 Place the pragma in the package spec that declares the called subprogram,
23294 without naming any entity. This disables warnings on all elaboration calls to
23295 all subprograms declared in this spec.
23297 @item Use Pragma Elaborate
23298 As previously described in section @xref{Treatment of Pragma Elaborate},
23299 GNAT in static mode assumes that a @code{pragma} Elaborate indicates correctly
23300 that no elaboration checks are required on calls to the designated unit.
23301 There may be cases in which the caller knows that no transitive calls
23302 can occur, so that a @code{pragma Elaborate} will be sufficient in a
23303 case where @code{pragma Elaborate_All} would cause a circularity.
23307 These five cases are listed in order of decreasing safety, and therefore
23308 require increasing programmer care in their application. Consider the
23311 @smallexample @c adanocomment
23313 function F1 return Integer;
23318 function F2 return Integer;
23319 function Pure (x : integer) return integer;
23320 -- pragma Suppress (Elaboration_Check, On => Pure); -- (3)
23321 -- pragma Suppress (Elaboration_Check); -- (4)
23325 package body Pack1 is
23326 function F1 return Integer is
23330 Val : integer := Pack2.Pure (11); -- Elab. call (1)
23333 -- pragma Suppress(Elaboration_Check, Pack2.F2); -- (1)
23334 -- pragma Suppress(Elaboration_Check); -- (2)
23336 X1 := Pack2.F2 + 1; -- Elab. call (2)
23341 package body Pack2 is
23342 function F2 return Integer is
23346 function Pure (x : integer) return integer is
23348 return x ** 3 - 3 * x;
23352 with Pack1, Ada.Text_IO;
23355 Ada.Text_IO.Put_Line(Pack1.X1'Img); -- 101
23358 In the absence of any pragmas, an attempt to bind this program produces
23359 the following diagnostics:
23365 error: elaboration circularity detected
23366 info: "pack1 (body)" must be elaborated before "pack1 (body)"
23367 info: reason: Elaborate_All probably needed in unit "pack1 (body)"
23368 info: recompile "pack1 (body)" with -gnatwl for full details
23369 info: "pack1 (body)"
23370 info: must be elaborated along with its spec:
23371 info: "pack1 (spec)"
23372 info: which is withed by:
23373 info: "pack2 (body)"
23374 info: which must be elaborated along with its spec:
23375 info: "pack2 (spec)"
23376 info: which is withed by:
23377 info: "pack1 (body)"
23380 The sources of the circularity are the two calls to @code{Pack2.Pure} and
23381 @code{Pack2.F2} in the body of @code{Pack1}. We can see that the call to
23382 F2 is safe, even though F2 calls F1, because the call appears after the
23383 elaboration of the body of F1. Therefore the pragma (1) is safe, and will
23384 remove the warning on the call. It is also possible to use pragma (2)
23385 because there are no other potentially unsafe calls in the block.
23388 The call to @code{Pure} is safe because this function does not depend on the
23389 state of @code{Pack2}. Therefore any call to this function is safe, and it
23390 is correct to place pragma (3) in the corresponding package spec.
23393 Finally, we could place pragma (4) in the spec of @code{Pack2} to disable
23394 warnings on all calls to functions declared therein. Note that this is not
23395 necessarily safe, and requires more detailed examination of the subprogram
23396 bodies involved. In particular, a call to @code{F2} requires that @code{F1}
23397 be already elaborated.
23401 It is hard to generalize on which of these four approaches should be
23402 taken. Obviously if it is possible to fix the program so that the default
23403 treatment works, this is preferable, but this may not always be practical.
23404 It is certainly simple enough to use
23406 but the danger in this case is that, even if the GNAT binder
23407 finds a correct elaboration order, it may not always do so,
23408 and certainly a binder from another Ada compiler might not. A
23409 combination of testing and analysis (for which the warnings generated
23412 switch can be useful) must be used to ensure that the program is free
23413 of errors. One switch that is useful in this testing is the
23414 @option{^-p (pessimistic elaboration order)^/PESSIMISTIC_ELABORATION_ORDER^}
23417 Normally the binder tries to find an order that has the best chance of
23418 of avoiding elaboration problems. With this switch, the binder
23419 plays a devil's advocate role, and tries to choose the order that
23420 has the best chance of failing. If your program works even with this
23421 switch, then it has a better chance of being error free, but this is still
23424 For an example of this approach in action, consider the C-tests (executable
23425 tests) from the ACVC suite. If these are compiled and run with the default
23426 treatment, then all but one of them succeed without generating any error
23427 diagnostics from the binder. However, there is one test that fails, and
23428 this is not surprising, because the whole point of this test is to ensure
23429 that the compiler can handle cases where it is impossible to determine
23430 a correct order statically, and it checks that an exception is indeed
23431 raised at run time.
23433 This one test must be compiled and run using the
23435 switch, and then it passes. Alternatively, the entire suite can
23436 be run using this switch. It is never wrong to run with the dynamic
23437 elaboration switch if your code is correct, and we assume that the
23438 C-tests are indeed correct (it is less efficient, but efficiency is
23439 not a factor in running the ACVC tests.)
23441 @node Elaboration for Access-to-Subprogram Values
23442 @section Elaboration for Access-to-Subprogram Values
23443 @cindex Access-to-subprogram
23446 The introduction of access-to-subprogram types in Ada 95 complicates
23447 the handling of elaboration. The trouble is that it becomes
23448 impossible to tell at compile time which procedure
23449 is being called. This means that it is not possible for the binder
23450 to analyze the elaboration requirements in this case.
23452 If at the point at which the access value is created
23453 (i.e., the evaluation of @code{P'Access} for a subprogram @code{P}),
23454 the body of the subprogram is
23455 known to have been elaborated, then the access value is safe, and its use
23456 does not require a check. This may be achieved by appropriate arrangement
23457 of the order of declarations if the subprogram is in the current unit,
23458 or, if the subprogram is in another unit, by using pragma
23459 @code{Pure}, @code{Preelaborate}, or @code{Elaborate_Body}
23460 on the referenced unit.
23462 If the referenced body is not known to have been elaborated at the point
23463 the access value is created, then any use of the access value must do a
23464 dynamic check, and this dynamic check will fail and raise a
23465 @code{Program_Error} exception if the body has not been elaborated yet.
23466 GNAT will generate the necessary checks, and in addition, if the
23468 switch is set, will generate warnings that such checks are required.
23470 The use of dynamic dispatching for tagged types similarly generates
23471 a requirement for dynamic checks, and premature calls to any primitive
23472 operation of a tagged type before the body of the operation has been
23473 elaborated, will result in the raising of @code{Program_Error}.
23475 @node Summary of Procedures for Elaboration Control
23476 @section Summary of Procedures for Elaboration Control
23477 @cindex Elaboration control
23480 First, compile your program with the default options, using none of
23481 the special elaboration control switches. If the binder successfully
23482 binds your program, then you can be confident that, apart from issues
23483 raised by the use of access-to-subprogram types and dynamic dispatching,
23484 the program is free of elaboration errors. If it is important that the
23485 program be portable, then use the
23487 switch to generate warnings about missing @code{Elaborate_All}
23488 pragmas, and supply the missing pragmas.
23490 If the program fails to bind using the default static elaboration
23491 handling, then you can fix the program to eliminate the binder
23492 message, or recompile the entire program with the
23493 @option{-gnatE} switch to generate dynamic elaboration checks,
23494 and, if you are sure there really are no elaboration problems,
23495 use a global pragma @code{Suppress (Elaboration_Check)}.
23497 @node Other Elaboration Order Considerations
23498 @section Other Elaboration Order Considerations
23500 This section has been entirely concerned with the issue of finding a valid
23501 elaboration order, as defined by the Ada Reference Manual. In a case
23502 where several elaboration orders are valid, the task is to find one
23503 of the possible valid elaboration orders (and the static model in GNAT
23504 will ensure that this is achieved).
23506 The purpose of the elaboration rules in the Ada Reference Manual is to
23507 make sure that no entity is accessed before it has been elaborated. For
23508 a subprogram, this means that the spec and body must have been elaborated
23509 before the subprogram is called. For an object, this means that the object
23510 must have been elaborated before its value is read or written. A violation
23511 of either of these two requirements is an access before elaboration order,
23512 and this section has been all about avoiding such errors.
23514 In the case where more than one order of elaboration is possible, in the
23515 sense that access before elaboration errors are avoided, then any one of
23516 the orders is ``correct'' in the sense that it meets the requirements of
23517 the Ada Reference Manual, and no such error occurs.
23519 However, it may be the case for a given program, that there are
23520 constraints on the order of elaboration that come not from consideration
23521 of avoiding elaboration errors, but rather from extra-lingual logic
23522 requirements. Consider this example:
23524 @smallexample @c ada
23525 with Init_Constants;
23526 package Constants is
23531 package Init_Constants is
23532 procedure P; -- require a body
23533 end Init_Constants;
23536 package body Init_Constants is
23537 procedure P is begin null; end;
23541 end Init_Constants;
23545 Z : Integer := Constants.X + Constants.Y;
23549 with Text_IO; use Text_IO;
23552 Put_Line (Calc.Z'Img);
23557 In this example, there is more than one valid order of elaboration. For
23558 example both the following are correct orders:
23561 Init_Constants spec
23564 Init_Constants body
23569 Init_Constants spec
23570 Init_Constants body
23577 There is no language rule to prefer one or the other, both are correct
23578 from an order of elaboration point of view. But the programmatic effects
23579 of the two orders are very different. In the first, the elaboration routine
23580 of @code{Calc} initializes @code{Z} to zero, and then the main program
23581 runs with this value of zero. But in the second order, the elaboration
23582 routine of @code{Calc} runs after the body of Init_Constants has set
23583 @code{X} and @code{Y} and thus @code{Z} is set to 7 before @code{Main}
23586 One could perhaps by applying pretty clever non-artificial intelligence
23587 to the situation guess that it is more likely that the second order of
23588 elaboration is the one desired, but there is no formal linguistic reason
23589 to prefer one over the other. In fact in this particular case, GNAT will
23590 prefer the second order, because of the rule that bodies are elaborated
23591 as soon as possible, but it's just luck that this is what was wanted
23592 (if indeed the second order was preferred).
23594 If the program cares about the order of elaboration routines in a case like
23595 this, it is important to specify the order required. In this particular
23596 case, that could have been achieved by adding to the spec of Calc:
23598 @smallexample @c ada
23599 pragma Elaborate_All (Constants);
23603 which requires that the body (if any) and spec of @code{Constants},
23604 as well as the body and spec of any unit @code{with}'ed by
23605 @code{Constants} be elaborated before @code{Calc} is elaborated.
23607 Clearly no automatic method can always guess which alternative you require,
23608 and if you are working with legacy code that had constraints of this kind
23609 which were not properly specified by adding @code{Elaborate} or
23610 @code{Elaborate_All} pragmas, then indeed it is possible that two different
23611 compilers can choose different orders.
23613 The @code{gnatbind}
23614 @option{^-p^/PESSIMISTIC_ELABORATION^} switch may be useful in smoking
23615 out problems. This switch causes bodies to be elaborated as late as possible
23616 instead of as early as possible. In the example above, it would have forced
23617 the choice of the first elaboration order. If you get different results
23618 when using this switch, and particularly if one set of results is right,
23619 and one is wrong as far as you are concerned, it shows that you have some
23620 missing @code{Elaborate} pragmas. For the example above, we have the
23624 gnatmake -f -q main
23627 gnatmake -f -q main -bargs -p
23633 It is of course quite unlikely that both these results are correct, so
23634 it is up to you in a case like this to investigate the source of the
23635 difference, by looking at the two elaboration orders that are chosen,
23636 and figuring out which is correct, and then adding the necessary
23637 @code{Elaborate_All} pragmas to ensure the desired order.
23640 @node Inline Assembler
23641 @appendix Inline Assembler
23644 If you need to write low-level software that interacts directly
23645 with the hardware, Ada provides two ways to incorporate assembly
23646 language code into your program. First, you can import and invoke
23647 external routines written in assembly language, an Ada feature fully
23648 supported by GNAT. However, for small sections of code it may be simpler
23649 or more efficient to include assembly language statements directly
23650 in your Ada source program, using the facilities of the implementation-defined
23651 package @code{System.Machine_Code}, which incorporates the gcc
23652 Inline Assembler. The Inline Assembler approach offers a number of advantages,
23653 including the following:
23656 @item No need to use non-Ada tools
23657 @item Consistent interface over different targets
23658 @item Automatic usage of the proper calling conventions
23659 @item Access to Ada constants and variables
23660 @item Definition of intrinsic routines
23661 @item Possibility of inlining a subprogram comprising assembler code
23662 @item Code optimizer can take Inline Assembler code into account
23665 This chapter presents a series of examples to show you how to use
23666 the Inline Assembler. Although it focuses on the Intel x86,
23667 the general approach applies also to other processors.
23668 It is assumed that you are familiar with Ada
23669 and with assembly language programming.
23672 * Basic Assembler Syntax::
23673 * A Simple Example of Inline Assembler::
23674 * Output Variables in Inline Assembler::
23675 * Input Variables in Inline Assembler::
23676 * Inlining Inline Assembler Code::
23677 * Other Asm Functionality::
23678 * A Complete Example::
23681 @c ---------------------------------------------------------------------------
23682 @node Basic Assembler Syntax
23683 @section Basic Assembler Syntax
23686 The assembler used by GNAT and gcc is based not on the Intel assembly
23687 language, but rather on a language that descends from the AT&T Unix
23688 assembler @emph{as} (and which is often referred to as ``AT&T syntax'').
23689 The following table summarizes the main features of @emph{as} syntax
23690 and points out the differences from the Intel conventions.
23691 See the gcc @emph{as} and @emph{gas} (an @emph{as} macro
23692 pre-processor) documentation for further information.
23695 @item Register names
23696 gcc / @emph{as}: Prefix with ``%''; for example @code{%eax}
23698 Intel: No extra punctuation; for example @code{eax}
23700 @item Immediate operand
23701 gcc / @emph{as}: Prefix with ``$''; for example @code{$4}
23703 Intel: No extra punctuation; for example @code{4}
23706 gcc / @emph{as}: Prefix with ``$''; for example @code{$loc}
23708 Intel: No extra punctuation; for example @code{loc}
23710 @item Memory contents
23711 gcc / @emph{as}: No extra punctuation; for example @code{loc}
23713 Intel: Square brackets; for example @code{[loc]}
23715 @item Register contents
23716 gcc / @emph{as}: Parentheses; for example @code{(%eax)}
23718 Intel: Square brackets; for example @code{[eax]}
23720 @item Hexadecimal numbers
23721 gcc / @emph{as}: Leading ``0x'' (C language syntax); for example @code{0xA0}
23723 Intel: Trailing ``h''; for example @code{A0h}
23726 gcc / @emph{as}: Explicit in op code; for example @code{movw} to move
23729 Intel: Implicit, deduced by assembler; for example @code{mov}
23731 @item Instruction repetition
23732 gcc / @emph{as}: Split into two lines; for example
23738 Intel: Keep on one line; for example @code{rep stosl}
23740 @item Order of operands
23741 gcc / @emph{as}: Source first; for example @code{movw $4, %eax}
23743 Intel: Destination first; for example @code{mov eax, 4}
23746 @c ---------------------------------------------------------------------------
23747 @node A Simple Example of Inline Assembler
23748 @section A Simple Example of Inline Assembler
23751 The following example will generate a single assembly language statement,
23752 @code{nop}, which does nothing. Despite its lack of run-time effect,
23753 the example will be useful in illustrating the basics of
23754 the Inline Assembler facility.
23756 @smallexample @c ada
23758 with System.Machine_Code; use System.Machine_Code;
23759 procedure Nothing is
23766 @code{Asm} is a procedure declared in package @code{System.Machine_Code};
23767 here it takes one parameter, a @emph{template string} that must be a static
23768 expression and that will form the generated instruction.
23769 @code{Asm} may be regarded as a compile-time procedure that parses
23770 the template string and additional parameters (none here),
23771 from which it generates a sequence of assembly language instructions.
23773 The examples in this chapter will illustrate several of the forms
23774 for invoking @code{Asm}; a complete specification of the syntax
23775 is found in the @cite{GNAT Reference Manual}.
23777 Under the standard GNAT conventions, the @code{Nothing} procedure
23778 should be in a file named @file{nothing.adb}.
23779 You can build the executable in the usual way:
23783 However, the interesting aspect of this example is not its run-time behavior
23784 but rather the generated assembly code.
23785 To see this output, invoke the compiler as follows:
23787 gcc -c -S -fomit-frame-pointer -gnatp @file{nothing.adb}
23789 where the options are:
23793 compile only (no bind or link)
23795 generate assembler listing
23796 @item -fomit-frame-pointer
23797 do not set up separate stack frames
23799 do not add runtime checks
23802 This gives a human-readable assembler version of the code. The resulting
23803 file will have the same name as the Ada source file, but with a @code{.s}
23804 extension. In our example, the file @file{nothing.s} has the following
23809 .file "nothing.adb"
23811 ___gnu_compiled_ada:
23814 .globl __ada_nothing
23826 The assembly code you included is clearly indicated by
23827 the compiler, between the @code{#APP} and @code{#NO_APP}
23828 delimiters. The character before the 'APP' and 'NOAPP'
23829 can differ on different targets. For example, GNU/Linux uses '#APP' while
23830 on NT you will see '/APP'.
23832 If you make a mistake in your assembler code (such as using the
23833 wrong size modifier, or using a wrong operand for the instruction) GNAT
23834 will report this error in a temporary file, which will be deleted when
23835 the compilation is finished. Generating an assembler file will help
23836 in such cases, since you can assemble this file separately using the
23837 @emph{as} assembler that comes with gcc.
23839 Assembling the file using the command
23842 as @file{nothing.s}
23845 will give you error messages whose lines correspond to the assembler
23846 input file, so you can easily find and correct any mistakes you made.
23847 If there are no errors, @emph{as} will generate an object file
23848 @file{nothing.out}.
23850 @c ---------------------------------------------------------------------------
23851 @node Output Variables in Inline Assembler
23852 @section Output Variables in Inline Assembler
23855 The examples in this section, showing how to access the processor flags,
23856 illustrate how to specify the destination operands for assembly language
23859 @smallexample @c ada
23861 with Interfaces; use Interfaces;
23862 with Ada.Text_IO; use Ada.Text_IO;
23863 with System.Machine_Code; use System.Machine_Code;
23864 procedure Get_Flags is
23865 Flags : Unsigned_32;
23868 Asm ("pushfl" & LF & HT & -- push flags on stack
23869 "popl %%eax" & LF & HT & -- load eax with flags
23870 "movl %%eax, %0", -- store flags in variable
23871 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
23872 Put_Line ("Flags register:" & Flags'Img);
23877 In order to have a nicely aligned assembly listing, we have separated
23878 multiple assembler statements in the Asm template string with linefeed
23879 (ASCII.LF) and horizontal tab (ASCII.HT) characters.
23880 The resulting section of the assembly output file is:
23887 movl %eax, -40(%ebp)
23892 It would have been legal to write the Asm invocation as:
23895 Asm ("pushfl popl %%eax movl %%eax, %0")
23898 but in the generated assembler file, this would come out as:
23902 pushfl popl %eax movl %eax, -40(%ebp)
23906 which is not so convenient for the human reader.
23908 We use Ada comments
23909 at the end of each line to explain what the assembler instructions
23910 actually do. This is a useful convention.
23912 When writing Inline Assembler instructions, you need to precede each register
23913 and variable name with a percent sign. Since the assembler already requires
23914 a percent sign at the beginning of a register name, you need two consecutive
23915 percent signs for such names in the Asm template string, thus @code{%%eax}.
23916 In the generated assembly code, one of the percent signs will be stripped off.
23918 Names such as @code{%0}, @code{%1}, @code{%2}, etc., denote input or output
23919 variables: operands you later define using @code{Input} or @code{Output}
23920 parameters to @code{Asm}.
23921 An output variable is illustrated in
23922 the third statement in the Asm template string:
23926 The intent is to store the contents of the eax register in a variable that can
23927 be accessed in Ada. Simply writing @code{movl %%eax, Flags} would not
23928 necessarily work, since the compiler might optimize by using a register
23929 to hold Flags, and the expansion of the @code{movl} instruction would not be
23930 aware of this optimization. The solution is not to store the result directly
23931 but rather to advise the compiler to choose the correct operand form;
23932 that is the purpose of the @code{%0} output variable.
23934 Information about the output variable is supplied in the @code{Outputs}
23935 parameter to @code{Asm}:
23937 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
23940 The output is defined by the @code{Asm_Output} attribute of the target type;
23941 the general format is
23943 Type'Asm_Output (constraint_string, variable_name)
23946 The constraint string directs the compiler how
23947 to store/access the associated variable. In the example
23949 Unsigned_32'Asm_Output ("=m", Flags);
23951 the @code{"m"} (memory) constraint tells the compiler that the variable
23952 @code{Flags} should be stored in a memory variable, thus preventing
23953 the optimizer from keeping it in a register. In contrast,
23955 Unsigned_32'Asm_Output ("=r", Flags);
23957 uses the @code{"r"} (register) constraint, telling the compiler to
23958 store the variable in a register.
23960 If the constraint is preceded by the equal character (@strong{=}), it tells
23961 the compiler that the variable will be used to store data into it.
23963 In the @code{Get_Flags} example, we used the @code{"g"} (global) constraint,
23964 allowing the optimizer to choose whatever it deems best.
23966 There are a fairly large number of constraints, but the ones that are
23967 most useful (for the Intel x86 processor) are the following:
23973 global (i.e. can be stored anywhere)
23991 use one of eax, ebx, ecx or edx
23993 use one of eax, ebx, ecx, edx, esi or edi
23996 The full set of constraints is described in the gcc and @emph{as}
23997 documentation; note that it is possible to combine certain constraints
23998 in one constraint string.
24000 You specify the association of an output variable with an assembler operand
24001 through the @code{%}@emph{n} notation, where @emph{n} is a non-negative
24003 @smallexample @c ada
24005 Asm ("pushfl" & LF & HT & -- push flags on stack
24006 "popl %%eax" & LF & HT & -- load eax with flags
24007 "movl %%eax, %0", -- store flags in variable
24008 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
24012 @code{%0} will be replaced in the expanded code by the appropriate operand,
24014 the compiler decided for the @code{Flags} variable.
24016 In general, you may have any number of output variables:
24019 Count the operands starting at 0; thus @code{%0}, @code{%1}, etc.
24021 Specify the @code{Outputs} parameter as a parenthesized comma-separated list
24022 of @code{Asm_Output} attributes
24026 @smallexample @c ada
24028 Asm ("movl %%eax, %0" & LF & HT &
24029 "movl %%ebx, %1" & LF & HT &
24031 Outputs => (Unsigned_32'Asm_Output ("=g", Var_A), -- %0 = Var_A
24032 Unsigned_32'Asm_Output ("=g", Var_B), -- %1 = Var_B
24033 Unsigned_32'Asm_Output ("=g", Var_C))); -- %2 = Var_C
24037 where @code{Var_A}, @code{Var_B}, and @code{Var_C} are variables
24038 in the Ada program.
24040 As a variation on the @code{Get_Flags} example, we can use the constraints
24041 string to direct the compiler to store the eax register into the @code{Flags}
24042 variable, instead of including the store instruction explicitly in the
24043 @code{Asm} template string:
24045 @smallexample @c ada
24047 with Interfaces; use Interfaces;
24048 with Ada.Text_IO; use Ada.Text_IO;
24049 with System.Machine_Code; use System.Machine_Code;
24050 procedure Get_Flags_2 is
24051 Flags : Unsigned_32;
24054 Asm ("pushfl" & LF & HT & -- push flags on stack
24055 "popl %%eax", -- save flags in eax
24056 Outputs => Unsigned_32'Asm_Output ("=a", Flags));
24057 Put_Line ("Flags register:" & Flags'Img);
24063 The @code{"a"} constraint tells the compiler that the @code{Flags}
24064 variable will come from the eax register. Here is the resulting code:
24072 movl %eax,-40(%ebp)
24077 The compiler generated the store of eax into Flags after
24078 expanding the assembler code.
24080 Actually, there was no need to pop the flags into the eax register;
24081 more simply, we could just pop the flags directly into the program variable:
24083 @smallexample @c ada
24085 with Interfaces; use Interfaces;
24086 with Ada.Text_IO; use Ada.Text_IO;
24087 with System.Machine_Code; use System.Machine_Code;
24088 procedure Get_Flags_3 is
24089 Flags : Unsigned_32;
24092 Asm ("pushfl" & LF & HT & -- push flags on stack
24093 "pop %0", -- save flags in Flags
24094 Outputs => Unsigned_32'Asm_Output ("=g", Flags));
24095 Put_Line ("Flags register:" & Flags'Img);
24100 @c ---------------------------------------------------------------------------
24101 @node Input Variables in Inline Assembler
24102 @section Input Variables in Inline Assembler
24105 The example in this section illustrates how to specify the source operands
24106 for assembly language statements.
24107 The program simply increments its input value by 1:
24109 @smallexample @c ada
24111 with Interfaces; use Interfaces;
24112 with Ada.Text_IO; use Ada.Text_IO;
24113 with System.Machine_Code; use System.Machine_Code;
24114 procedure Increment is
24116 function Incr (Value : Unsigned_32) return Unsigned_32 is
24117 Result : Unsigned_32;
24120 Inputs => Unsigned_32'Asm_Input ("a", Value),
24121 Outputs => Unsigned_32'Asm_Output ("=a", Result));
24125 Value : Unsigned_32;
24129 Put_Line ("Value before is" & Value'Img);
24130 Value := Incr (Value);
24131 Put_Line ("Value after is" & Value'Img);
24136 The @code{Outputs} parameter to @code{Asm} specifies
24137 that the result will be in the eax register and that it is to be stored
24138 in the @code{Result} variable.
24140 The @code{Inputs} parameter looks much like the @code{Outputs} parameter,
24141 but with an @code{Asm_Input} attribute.
24142 The @code{"="} constraint, indicating an output value, is not present.
24144 You can have multiple input variables, in the same way that you can have more
24145 than one output variable.
24147 The parameter count (%0, %1) etc, now starts at the first input
24148 statement, and continues with the output statements.
24149 When both parameters use the same variable, the
24150 compiler will treat them as the same %n operand, which is the case here.
24152 Just as the @code{Outputs} parameter causes the register to be stored into the
24153 target variable after execution of the assembler statements, so does the
24154 @code{Inputs} parameter cause its variable to be loaded into the register
24155 before execution of the assembler statements.
24157 Thus the effect of the @code{Asm} invocation is:
24159 @item load the 32-bit value of @code{Value} into eax
24160 @item execute the @code{incl %eax} instruction
24161 @item store the contents of eax into the @code{Result} variable
24164 The resulting assembler file (with @option{-O2} optimization) contains:
24167 _increment__incr.1:
24180 @c ---------------------------------------------------------------------------
24181 @node Inlining Inline Assembler Code
24182 @section Inlining Inline Assembler Code
24185 For a short subprogram such as the @code{Incr} function in the previous
24186 section, the overhead of the call and return (creating / deleting the stack
24187 frame) can be significant, compared to the amount of code in the subprogram
24188 body. A solution is to apply Ada's @code{Inline} pragma to the subprogram,
24189 which directs the compiler to expand invocations of the subprogram at the
24190 point(s) of call, instead of setting up a stack frame for out-of-line calls.
24191 Here is the resulting program:
24193 @smallexample @c ada
24195 with Interfaces; use Interfaces;
24196 with Ada.Text_IO; use Ada.Text_IO;
24197 with System.Machine_Code; use System.Machine_Code;
24198 procedure Increment_2 is
24200 function Incr (Value : Unsigned_32) return Unsigned_32 is
24201 Result : Unsigned_32;
24204 Inputs => Unsigned_32'Asm_Input ("a", Value),
24205 Outputs => Unsigned_32'Asm_Output ("=a", Result));
24208 pragma Inline (Increment);
24210 Value : Unsigned_32;
24214 Put_Line ("Value before is" & Value'Img);
24215 Value := Increment (Value);
24216 Put_Line ("Value after is" & Value'Img);
24221 Compile the program with both optimization (@option{-O2}) and inlining
24222 enabled (@option{-gnatpn} instead of @option{-gnatp}).
24224 The @code{Incr} function is still compiled as usual, but at the
24225 point in @code{Increment} where our function used to be called:
24230 call _increment__incr.1
24235 the code for the function body directly appears:
24248 thus saving the overhead of stack frame setup and an out-of-line call.
24250 @c ---------------------------------------------------------------------------
24251 @node Other Asm Functionality
24252 @section Other @code{Asm} Functionality
24255 This section describes two important parameters to the @code{Asm}
24256 procedure: @code{Clobber}, which identifies register usage;
24257 and @code{Volatile}, which inhibits unwanted optimizations.
24260 * The Clobber Parameter::
24261 * The Volatile Parameter::
24264 @c ---------------------------------------------------------------------------
24265 @node The Clobber Parameter
24266 @subsection The @code{Clobber} Parameter
24269 One of the dangers of intermixing assembly language and a compiled language
24270 such as Ada is that the compiler needs to be aware of which registers are
24271 being used by the assembly code. In some cases, such as the earlier examples,
24272 the constraint string is sufficient to indicate register usage (e.g.,
24274 the eax register). But more generally, the compiler needs an explicit
24275 identification of the registers that are used by the Inline Assembly
24278 Using a register that the compiler doesn't know about
24279 could be a side effect of an instruction (like @code{mull}
24280 storing its result in both eax and edx).
24281 It can also arise from explicit register usage in your
24282 assembly code; for example:
24285 Asm ("movl %0, %%ebx" & LF & HT &
24287 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
24288 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out));
24292 where the compiler (since it does not analyze the @code{Asm} template string)
24293 does not know you are using the ebx register.
24295 In such cases you need to supply the @code{Clobber} parameter to @code{Asm},
24296 to identify the registers that will be used by your assembly code:
24300 Asm ("movl %0, %%ebx" & LF & HT &
24302 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
24303 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
24308 The Clobber parameter is a static string expression specifying the
24309 register(s) you are using. Note that register names are @emph{not} prefixed
24310 by a percent sign. Also, if more than one register is used then their names
24311 are separated by commas; e.g., @code{"eax, ebx"}
24313 The @code{Clobber} parameter has several additional uses:
24315 @item Use ``register'' name @code{cc} to indicate that flags might have changed
24316 @item Use ``register'' name @code{memory} if you changed a memory location
24319 @c ---------------------------------------------------------------------------
24320 @node The Volatile Parameter
24321 @subsection The @code{Volatile} Parameter
24322 @cindex Volatile parameter
24325 Compiler optimizations in the presence of Inline Assembler may sometimes have
24326 unwanted effects. For example, when an @code{Asm} invocation with an input
24327 variable is inside a loop, the compiler might move the loading of the input
24328 variable outside the loop, regarding it as a one-time initialization.
24330 If this effect is not desired, you can disable such optimizations by setting
24331 the @code{Volatile} parameter to @code{True}; for example:
24333 @smallexample @c ada
24335 Asm ("movl %0, %%ebx" & LF & HT &
24337 Inputs => Unsigned_32'Asm_Input ("g", Var_In),
24338 Outputs => Unsigned_32'Asm_Output ("=g", Var_Out),
24344 By default, @code{Volatile} is set to @code{False} unless there is no
24345 @code{Outputs} parameter.
24347 Although setting @code{Volatile} to @code{True} prevents unwanted
24348 optimizations, it will also disable other optimizations that might be
24349 important for efficiency. In general, you should set @code{Volatile}
24350 to @code{True} only if the compiler's optimizations have created
24353 @c ---------------------------------------------------------------------------
24354 @node A Complete Example
24355 @section A Complete Example
24358 This section contains a complete program illustrating a realistic usage
24359 of GNAT's Inline Assembler capabilities. It comprises a main procedure
24360 @code{Check_CPU} and a package @code{Intel_CPU}.
24361 The package declares a collection of functions that detect the properties
24362 of the 32-bit x86 processor that is running the program.
24363 The main procedure invokes these functions and displays the information.
24365 The Intel_CPU package could be enhanced by adding functions to
24366 detect the type of x386 co-processor, the processor caching options and
24367 special operations such as the SIMD extensions.
24369 Although the Intel_CPU package has been written for 32-bit Intel
24370 compatible CPUs, it is OS neutral. It has been tested on DOS,
24371 Windows/NT and GNU/Linux.
24374 * Check_CPU Procedure::
24375 * Intel_CPU Package Specification::
24376 * Intel_CPU Package Body::
24379 @c ---------------------------------------------------------------------------
24380 @node Check_CPU Procedure
24381 @subsection @code{Check_CPU} Procedure
24382 @cindex Check_CPU procedure
24384 @smallexample @c adanocomment
24385 ---------------------------------------------------------------------
24387 -- Uses the Intel_CPU package to identify the CPU the program is --
24388 -- running on, and some of the features it supports. --
24390 ---------------------------------------------------------------------
24392 with Intel_CPU; -- Intel CPU detection functions
24393 with Ada.Text_IO; -- Standard text I/O
24394 with Ada.Command_Line; -- To set the exit status
24396 procedure Check_CPU is
24398 Type_Found : Boolean := False;
24399 -- Flag to indicate that processor was identified
24401 Features : Intel_CPU.Processor_Features;
24402 -- The processor features
24404 Signature : Intel_CPU.Processor_Signature;
24405 -- The processor type signature
24409 -----------------------------------
24410 -- Display the program banner. --
24411 -----------------------------------
24413 Ada.Text_IO.Put_Line (Ada.Command_Line.Command_Name &
24414 ": check Intel CPU version and features, v1.0");
24415 Ada.Text_IO.Put_Line ("distribute freely, but no warranty whatsoever");
24416 Ada.Text_IO.New_Line;
24418 -----------------------------------------------------------------------
24419 -- We can safely start with the assumption that we are on at least --
24420 -- a x386 processor. If the CPUID instruction is present, then we --
24421 -- have a later processor type. --
24422 -----------------------------------------------------------------------
24424 if Intel_CPU.Has_CPUID = False then
24426 -- No CPUID instruction, so we assume this is indeed a x386
24427 -- processor. We can still check if it has a FP co-processor.
24428 if Intel_CPU.Has_FPU then
24429 Ada.Text_IO.Put_Line
24430 ("x386-type processor with a FP co-processor");
24432 Ada.Text_IO.Put_Line
24433 ("x386-type processor without a FP co-processor");
24434 end if; -- check for FPU
24437 Ada.Command_Line.Set_Exit_Status (Ada.Command_Line.Success);
24440 end if; -- check for CPUID
24442 -----------------------------------------------------------------------
24443 -- If CPUID is supported, check if this is a true Intel processor, --
24444 -- if it is not, display a warning. --
24445 -----------------------------------------------------------------------
24447 if Intel_CPU.Vendor_ID /= Intel_CPU.Intel_Processor then
24448 Ada.Text_IO.Put_Line ("*** This is a Intel compatible processor");
24449 Ada.Text_IO.Put_Line ("*** Some information may be incorrect");
24450 end if; -- check if Intel
24452 ----------------------------------------------------------------------
24453 -- With the CPUID instruction present, we can assume at least a --
24454 -- x486 processor. If the CPUID support level is < 1 then we have --
24455 -- to leave it at that. --
24456 ----------------------------------------------------------------------
24458 if Intel_CPU.CPUID_Level < 1 then
24460 -- Ok, this is a x486 processor. we still can get the Vendor ID
24461 Ada.Text_IO.Put_Line ("x486-type processor");
24462 Ada.Text_IO.Put_Line ("Vendor ID is " & Intel_CPU.Vendor_ID);
24464 -- We can also check if there is a FPU present
24465 if Intel_CPU.Has_FPU then
24466 Ada.Text_IO.Put_Line ("Floating-Point support");
24468 Ada.Text_IO.Put_Line ("No Floating-Point support");
24469 end if; -- check for FPU
24472 Ada.Command_Line.Set_Exit_Status (Ada.Command_Line.Success);
24475 end if; -- check CPUID level
24477 ---------------------------------------------------------------------
24478 -- With a CPUID level of 1 we can use the processor signature to --
24479 -- determine it's exact type. --
24480 ---------------------------------------------------------------------
24482 Signature := Intel_CPU.Signature;
24484 ----------------------------------------------------------------------
24485 -- Ok, now we go into a lot of messy comparisons to get the --
24486 -- processor type. For clarity, no attememt to try to optimize the --
24487 -- comparisons has been made. Note that since Intel_CPU does not --
24488 -- support getting cache info, we cannot distinguish between P5 --
24489 -- and Celeron types yet. --
24490 ----------------------------------------------------------------------
24493 if Signature.Processor_Type = 2#00# and
24494 Signature.Family = 2#0100# and
24495 Signature.Model = 2#0100# then
24496 Type_Found := True;
24497 Ada.Text_IO.Put_Line ("x486SL processor");
24500 -- x486DX2 Write-Back
24501 if Signature.Processor_Type = 2#00# and
24502 Signature.Family = 2#0100# and
24503 Signature.Model = 2#0111# then
24504 Type_Found := True;
24505 Ada.Text_IO.Put_Line ("Write-Back Enhanced x486DX2 processor");
24509 if Signature.Processor_Type = 2#00# and
24510 Signature.Family = 2#0100# and
24511 Signature.Model = 2#1000# then
24512 Type_Found := True;
24513 Ada.Text_IO.Put_Line ("x486DX4 processor");
24516 -- x486DX4 Overdrive
24517 if Signature.Processor_Type = 2#01# and
24518 Signature.Family = 2#0100# and
24519 Signature.Model = 2#1000# then
24520 Type_Found := True;
24521 Ada.Text_IO.Put_Line ("x486DX4 OverDrive processor");
24524 -- Pentium (60, 66)
24525 if Signature.Processor_Type = 2#00# and
24526 Signature.Family = 2#0101# and
24527 Signature.Model = 2#0001# then
24528 Type_Found := True;
24529 Ada.Text_IO.Put_Line ("Pentium processor (60, 66)");
24532 -- Pentium (75, 90, 100, 120, 133, 150, 166, 200)
24533 if Signature.Processor_Type = 2#00# and
24534 Signature.Family = 2#0101# and
24535 Signature.Model = 2#0010# then
24536 Type_Found := True;
24537 Ada.Text_IO.Put_Line
24538 ("Pentium processor (75, 90, 100, 120, 133, 150, 166, 200)");
24541 -- Pentium OverDrive (60, 66)
24542 if Signature.Processor_Type = 2#01# and
24543 Signature.Family = 2#0101# and
24544 Signature.Model = 2#0001# then
24545 Type_Found := True;
24546 Ada.Text_IO.Put_Line ("Pentium OverDrive processor (60, 66)");
24549 -- Pentium OverDrive (75, 90, 100, 120, 133, 150, 166, 200)
24550 if Signature.Processor_Type = 2#01# and
24551 Signature.Family = 2#0101# and
24552 Signature.Model = 2#0010# then
24553 Type_Found := True;
24554 Ada.Text_IO.Put_Line
24555 ("Pentium OverDrive cpu (75, 90, 100, 120, 133, 150, 166, 200)");
24558 -- Pentium OverDrive processor for x486 processor-based systems
24559 if Signature.Processor_Type = 2#01# and
24560 Signature.Family = 2#0101# and
24561 Signature.Model = 2#0011# then
24562 Type_Found := True;
24563 Ada.Text_IO.Put_Line
24564 ("Pentium OverDrive processor for x486 processor-based systems");
24567 -- Pentium processor with MMX technology (166, 200)
24568 if Signature.Processor_Type = 2#00# and
24569 Signature.Family = 2#0101# and
24570 Signature.Model = 2#0100# then
24571 Type_Found := True;
24572 Ada.Text_IO.Put_Line
24573 ("Pentium processor with MMX technology (166, 200)");
24576 -- Pentium OverDrive with MMX for Pentium (75, 90, 100, 120, 133)
24577 if Signature.Processor_Type = 2#01# and
24578 Signature.Family = 2#0101# and
24579 Signature.Model = 2#0100# then
24580 Type_Found := True;
24581 Ada.Text_IO.Put_Line
24582 ("Pentium OverDrive processor with MMX " &
24583 "technology for Pentium processor (75, 90, 100, 120, 133)");
24586 -- Pentium Pro processor
24587 if Signature.Processor_Type = 2#00# and
24588 Signature.Family = 2#0110# and
24589 Signature.Model = 2#0001# then
24590 Type_Found := True;
24591 Ada.Text_IO.Put_Line ("Pentium Pro processor");
24594 -- Pentium II processor, model 3
24595 if Signature.Processor_Type = 2#00# and
24596 Signature.Family = 2#0110# and
24597 Signature.Model = 2#0011# then
24598 Type_Found := True;
24599 Ada.Text_IO.Put_Line ("Pentium II processor, model 3");
24602 -- Pentium II processor, model 5 or Celeron processor
24603 if Signature.Processor_Type = 2#00# and
24604 Signature.Family = 2#0110# and
24605 Signature.Model = 2#0101# then
24606 Type_Found := True;
24607 Ada.Text_IO.Put_Line
24608 ("Pentium II processor, model 5 or Celeron processor");
24611 -- Pentium Pro OverDrive processor
24612 if Signature.Processor_Type = 2#01# and
24613 Signature.Family = 2#0110# and
24614 Signature.Model = 2#0011# then
24615 Type_Found := True;
24616 Ada.Text_IO.Put_Line ("Pentium Pro OverDrive processor");
24619 -- If no type recognized, we have an unknown. Display what
24621 if Type_Found = False then
24622 Ada.Text_IO.Put_Line ("Unknown processor");
24625 -----------------------------------------
24626 -- Display processor stepping level. --
24627 -----------------------------------------
24629 Ada.Text_IO.Put_Line ("Stepping level:" & Signature.Stepping'Img);
24631 ---------------------------------
24632 -- Display vendor ID string. --
24633 ---------------------------------
24635 Ada.Text_IO.Put_Line ("Vendor ID: " & Intel_CPU.Vendor_ID);
24637 ------------------------------------
24638 -- Get the processors features. --
24639 ------------------------------------
24641 Features := Intel_CPU.Features;
24643 -----------------------------
24644 -- Check for a FPU unit. --
24645 -----------------------------
24647 if Features.FPU = True then
24648 Ada.Text_IO.Put_Line ("Floating-Point unit available");
24650 Ada.Text_IO.Put_Line ("no Floating-Point unit");
24651 end if; -- check for FPU
24653 --------------------------------
24654 -- List processor features. --
24655 --------------------------------
24657 Ada.Text_IO.Put_Line ("Supported features: ");
24659 -- Virtual Mode Extension
24660 if Features.VME = True then
24661 Ada.Text_IO.Put_Line (" VME - Virtual Mode Extension");
24664 -- Debugging Extension
24665 if Features.DE = True then
24666 Ada.Text_IO.Put_Line (" DE - Debugging Extension");
24669 -- Page Size Extension
24670 if Features.PSE = True then
24671 Ada.Text_IO.Put_Line (" PSE - Page Size Extension");
24674 -- Time Stamp Counter
24675 if Features.TSC = True then
24676 Ada.Text_IO.Put_Line (" TSC - Time Stamp Counter");
24679 -- Model Specific Registers
24680 if Features.MSR = True then
24681 Ada.Text_IO.Put_Line (" MSR - Model Specific Registers");
24684 -- Physical Address Extension
24685 if Features.PAE = True then
24686 Ada.Text_IO.Put_Line (" PAE - Physical Address Extension");
24689 -- Machine Check Extension
24690 if Features.MCE = True then
24691 Ada.Text_IO.Put_Line (" MCE - Machine Check Extension");
24694 -- CMPXCHG8 instruction supported
24695 if Features.CX8 = True then
24696 Ada.Text_IO.Put_Line (" CX8 - CMPXCHG8 instruction");
24699 -- on-chip APIC hardware support
24700 if Features.APIC = True then
24701 Ada.Text_IO.Put_Line (" APIC - on-chip APIC hardware support");
24704 -- Fast System Call
24705 if Features.SEP = True then
24706 Ada.Text_IO.Put_Line (" SEP - Fast System Call");
24709 -- Memory Type Range Registers
24710 if Features.MTRR = True then
24711 Ada.Text_IO.Put_Line (" MTTR - Memory Type Range Registers");
24714 -- Page Global Enable
24715 if Features.PGE = True then
24716 Ada.Text_IO.Put_Line (" PGE - Page Global Enable");
24719 -- Machine Check Architecture
24720 if Features.MCA = True then
24721 Ada.Text_IO.Put_Line (" MCA - Machine Check Architecture");
24724 -- Conditional Move Instruction Supported
24725 if Features.CMOV = True then
24726 Ada.Text_IO.Put_Line
24727 (" CMOV - Conditional Move Instruction Supported");
24730 -- Page Attribute Table
24731 if Features.PAT = True then
24732 Ada.Text_IO.Put_Line (" PAT - Page Attribute Table");
24735 -- 36-bit Page Size Extension
24736 if Features.PSE_36 = True then
24737 Ada.Text_IO.Put_Line (" PSE_36 - 36-bit Page Size Extension");
24740 -- MMX technology supported
24741 if Features.MMX = True then
24742 Ada.Text_IO.Put_Line (" MMX - MMX technology supported");
24745 -- Fast FP Save and Restore
24746 if Features.FXSR = True then
24747 Ada.Text_IO.Put_Line (" FXSR - Fast FP Save and Restore");
24750 ---------------------
24751 -- Program done. --
24752 ---------------------
24754 Ada.Command_Line.Set_Exit_Status (Ada.Command_Line.Success);
24759 Ada.Command_Line.Set_Exit_Status (Ada.Command_Line.Failure);
24765 @c ---------------------------------------------------------------------------
24766 @node Intel_CPU Package Specification
24767 @subsection @code{Intel_CPU} Package Specification
24768 @cindex Intel_CPU package specification
24770 @smallexample @c adanocomment
24771 -------------------------------------------------------------------------
24773 -- file: intel_cpu.ads --
24775 -- ********************************************* --
24776 -- * WARNING: for 32-bit Intel processors only * --
24777 -- ********************************************* --
24779 -- This package contains a number of subprograms that are useful in --
24780 -- determining the Intel x86 CPU (and the features it supports) on --
24781 -- which the program is running. --
24783 -- The package is based upon the information given in the Intel --
24784 -- Application Note AP-485: "Intel Processor Identification and the --
24785 -- CPUID Instruction" as of April 1998. This application note can be --
24786 -- found on www.intel.com. --
24788 -- It currently deals with 32-bit processors only, will not detect --
24789 -- features added after april 1998, and does not guarantee proper --
24790 -- results on Intel-compatible processors. --
24792 -- Cache info and x386 fpu type detection are not supported. --
24794 -- This package does not use any privileged instructions, so should --
24795 -- work on any OS running on a 32-bit Intel processor. --
24797 -------------------------------------------------------------------------
24799 with Interfaces; use Interfaces;
24800 -- for using unsigned types
24802 with System.Machine_Code; use System.Machine_Code;
24803 -- for using inline assembler code
24805 with Ada.Characters.Latin_1; use Ada.Characters.Latin_1;
24806 -- for inserting control characters
24808 package Intel_CPU is
24810 ----------------------
24811 -- Processor bits --
24812 ----------------------
24814 subtype Num_Bits is Natural range 0 .. 31;
24815 -- the number of processor bits (32)
24817 --------------------------
24818 -- Processor register --
24819 --------------------------
24821 -- define a processor register type for easy access to
24822 -- the individual bits
24824 type Processor_Register is array (Num_Bits) of Boolean;
24825 pragma Pack (Processor_Register);
24826 for Processor_Register'Size use 32;
24828 -------------------------
24829 -- Unsigned register --
24830 -------------------------
24832 -- define a processor register type for easy access to
24833 -- the individual bytes
24835 type Unsigned_Register is
24843 for Unsigned_Register use
24845 L1 at 0 range 0 .. 7;
24846 H1 at 0 range 8 .. 15;
24847 L2 at 0 range 16 .. 23;
24848 H2 at 0 range 24 .. 31;
24851 for Unsigned_Register'Size use 32;
24853 ---------------------------------
24854 -- Intel processor vendor ID --
24855 ---------------------------------
24857 Intel_Processor : constant String (1 .. 12) := "GenuineIntel";
24858 -- indicates an Intel manufactured processor
24860 ------------------------------------
24861 -- Processor signature register --
24862 ------------------------------------
24864 -- a register type to hold the processor signature
24866 type Processor_Signature is
24868 Stepping : Natural range 0 .. 15;
24869 Model : Natural range 0 .. 15;
24870 Family : Natural range 0 .. 15;
24871 Processor_Type : Natural range 0 .. 3;
24872 Reserved : Natural range 0 .. 262143;
24875 for Processor_Signature use
24877 Stepping at 0 range 0 .. 3;
24878 Model at 0 range 4 .. 7;
24879 Family at 0 range 8 .. 11;
24880 Processor_Type at 0 range 12 .. 13;
24881 Reserved at 0 range 14 .. 31;
24884 for Processor_Signature'Size use 32;
24886 -----------------------------------
24887 -- Processor features register --
24888 -----------------------------------
24890 -- a processor register to hold the processor feature flags
24892 type Processor_Features is
24894 FPU : Boolean; -- floating point unit on chip
24895 VME : Boolean; -- virtual mode extension
24896 DE : Boolean; -- debugging extension
24897 PSE : Boolean; -- page size extension
24898 TSC : Boolean; -- time stamp counter
24899 MSR : Boolean; -- model specific registers
24900 PAE : Boolean; -- physical address extension
24901 MCE : Boolean; -- machine check extension
24902 CX8 : Boolean; -- cmpxchg8 instruction
24903 APIC : Boolean; -- on-chip apic hardware
24904 Res_1 : Boolean; -- reserved for extensions
24905 SEP : Boolean; -- fast system call
24906 MTRR : Boolean; -- memory type range registers
24907 PGE : Boolean; -- page global enable
24908 MCA : Boolean; -- machine check architecture
24909 CMOV : Boolean; -- conditional move supported
24910 PAT : Boolean; -- page attribute table
24911 PSE_36 : Boolean; -- 36-bit page size extension
24912 Res_2 : Natural range 0 .. 31; -- reserved for extensions
24913 MMX : Boolean; -- MMX technology supported
24914 FXSR : Boolean; -- fast FP save and restore
24915 Res_3 : Natural range 0 .. 127; -- reserved for extensions
24918 for Processor_Features use
24920 FPU at 0 range 0 .. 0;
24921 VME at 0 range 1 .. 1;
24922 DE at 0 range 2 .. 2;
24923 PSE at 0 range 3 .. 3;
24924 TSC at 0 range 4 .. 4;
24925 MSR at 0 range 5 .. 5;
24926 PAE at 0 range 6 .. 6;
24927 MCE at 0 range 7 .. 7;
24928 CX8 at 0 range 8 .. 8;
24929 APIC at 0 range 9 .. 9;
24930 Res_1 at 0 range 10 .. 10;
24931 SEP at 0 range 11 .. 11;
24932 MTRR at 0 range 12 .. 12;
24933 PGE at 0 range 13 .. 13;
24934 MCA at 0 range 14 .. 14;
24935 CMOV at 0 range 15 .. 15;
24936 PAT at 0 range 16 .. 16;
24937 PSE_36 at 0 range 17 .. 17;
24938 Res_2 at 0 range 18 .. 22;
24939 MMX at 0 range 23 .. 23;
24940 FXSR at 0 range 24 .. 24;
24941 Res_3 at 0 range 25 .. 31;
24944 for Processor_Features'Size use 32;
24946 -------------------
24948 -------------------
24950 function Has_FPU return Boolean;
24951 -- return True if a FPU is found
24952 -- use only if CPUID is not supported
24954 function Has_CPUID return Boolean;
24955 -- return True if the processor supports the CPUID instruction
24957 function CPUID_Level return Natural;
24958 -- return the CPUID support level (0, 1 or 2)
24959 -- can only be called if the CPUID instruction is supported
24961 function Vendor_ID return String;
24962 -- return the processor vendor identification string
24963 -- can only be called if the CPUID instruction is supported
24965 function Signature return Processor_Signature;
24966 -- return the processor signature
24967 -- can only be called if the CPUID instruction is supported
24969 function Features return Processor_Features;
24970 -- return the processors features
24971 -- can only be called if the CPUID instruction is supported
24975 ------------------------
24976 -- EFLAGS bit names --
24977 ------------------------
24979 ID_Flag : constant Num_Bits := 21;
24985 @c ---------------------------------------------------------------------------
24986 @node Intel_CPU Package Body
24987 @subsection @code{Intel_CPU} Package Body
24988 @cindex Intel_CPU package body
24990 @smallexample @c adanocomment
24991 package body Intel_CPU is
24993 ---------------------------
24994 -- Detect FPU presence --
24995 ---------------------------
24997 -- There is a FPU present if we can set values to the FPU Status
24998 -- and Control Words.
25000 function Has_FPU return Boolean is
25002 Register : Unsigned_16;
25003 -- processor register to store a word
25007 -- check if we can change the status word
25010 -- the assembler code
25011 "finit" & LF & HT & -- reset status word
25012 "movw $0x5A5A, %%ax" & LF & HT & -- set value status word
25013 "fnstsw %0" & LF & HT & -- save status word
25014 "movw %%ax, %0", -- store status word
25016 -- output stored in Register
25017 -- register must be a memory location
25018 Outputs => Unsigned_16'Asm_output ("=m", Register),
25020 -- tell compiler that we used eax
25023 -- if the status word is zero, there is no FPU
25024 if Register = 0 then
25025 return False; -- no status word
25026 end if; -- check status word value
25028 -- check if we can get the control word
25031 -- the assembler code
25032 "fnstcw %0", -- save the control word
25034 -- output into Register
25035 -- register must be a memory location
25036 Outputs => Unsigned_16'Asm_output ("=m", Register));
25038 -- check the relevant bits
25039 if (Register and 16#103F#) /= 16#003F# then
25040 return False; -- no control word
25041 end if; -- check control word value
25048 --------------------------------
25049 -- Detect CPUID instruction --
25050 --------------------------------
25052 -- The processor supports the CPUID instruction if it is possible
25053 -- to change the value of ID flag bit in the EFLAGS register.
25055 function Has_CPUID return Boolean is
25057 Original_Flags, Modified_Flags : Processor_Register;
25058 -- EFLAG contents before and after changing the ID flag
25062 -- try flipping the ID flag in the EFLAGS register
25065 -- the assembler code
25066 "pushfl" & LF & HT & -- push EFLAGS on stack
25067 "pop %%eax" & LF & HT & -- pop EFLAGS into eax
25068 "movl %%eax, %0" & LF & HT & -- save EFLAGS content
25069 "xor $0x200000, %%eax" & LF & HT & -- flip ID flag
25070 "push %%eax" & LF & HT & -- push EFLAGS on stack
25071 "popfl" & LF & HT & -- load EFLAGS register
25072 "pushfl" & LF & HT & -- push EFLAGS on stack
25073 "pop %1", -- save EFLAGS content
25075 -- output values, may be anything
25076 -- Original_Flags is %0
25077 -- Modified_Flags is %1
25079 (Processor_Register'Asm_output ("=g", Original_Flags),
25080 Processor_Register'Asm_output ("=g", Modified_Flags)),
25082 -- tell compiler eax is destroyed
25085 -- check if CPUID is supported
25086 if Original_Flags(ID_Flag) /= Modified_Flags(ID_Flag) then
25087 return True; -- ID flag was modified
25089 return False; -- ID flag unchanged
25090 end if; -- check for CPUID
25094 -------------------------------
25095 -- Get CPUID support level --
25096 -------------------------------
25098 function CPUID_Level return Natural is
25100 Level : Unsigned_32;
25101 -- returned support level
25105 -- execute CPUID, storing the results in the Level register
25108 -- the assembler code
25109 "cpuid", -- execute CPUID
25111 -- zero is stored in eax
25112 -- returning the support level in eax
25113 Inputs => Unsigned_32'Asm_input ("a", 0),
25115 -- eax is stored in Level
25116 Outputs => Unsigned_32'Asm_output ("=a", Level),
25118 -- tell compiler ebx, ecx and edx registers are destroyed
25119 Clobber => "ebx, ecx, edx");
25121 -- return the support level
25122 return Natural (Level);
25126 --------------------------------
25127 -- Get CPU Vendor ID String --
25128 --------------------------------
25130 -- The vendor ID string is returned in the ebx, ecx and edx register
25131 -- after executing the CPUID instruction with eax set to zero.
25132 -- In case of a true Intel processor the string returned is
25135 function Vendor_ID return String is
25137 Ebx, Ecx, Edx : Unsigned_Register;
25138 -- registers containing the vendor ID string
25140 Vendor_ID : String (1 .. 12);
25141 -- the vendor ID string
25145 -- execute CPUID, storing the results in the processor registers
25148 -- the assembler code
25149 "cpuid", -- execute CPUID
25151 -- zero stored in eax
25152 -- vendor ID string returned in ebx, ecx and edx
25153 Inputs => Unsigned_32'Asm_input ("a", 0),
25155 -- ebx is stored in Ebx
25156 -- ecx is stored in Ecx
25157 -- edx is stored in Edx
25158 Outputs => (Unsigned_Register'Asm_output ("=b", Ebx),
25159 Unsigned_Register'Asm_output ("=c", Ecx),
25160 Unsigned_Register'Asm_output ("=d", Edx)));
25162 -- now build the vendor ID string
25163 Vendor_ID( 1) := Character'Val (Ebx.L1);
25164 Vendor_ID( 2) := Character'Val (Ebx.H1);
25165 Vendor_ID( 3) := Character'Val (Ebx.L2);
25166 Vendor_ID( 4) := Character'Val (Ebx.H2);
25167 Vendor_ID( 5) := Character'Val (Edx.L1);
25168 Vendor_ID( 6) := Character'Val (Edx.H1);
25169 Vendor_ID( 7) := Character'Val (Edx.L2);
25170 Vendor_ID( 8) := Character'Val (Edx.H2);
25171 Vendor_ID( 9) := Character'Val (Ecx.L1);
25172 Vendor_ID(10) := Character'Val (Ecx.H1);
25173 Vendor_ID(11) := Character'Val (Ecx.L2);
25174 Vendor_ID(12) := Character'Val (Ecx.H2);
25181 -------------------------------
25182 -- Get processor signature --
25183 -------------------------------
25185 function Signature return Processor_Signature is
25187 Result : Processor_Signature;
25188 -- processor signature returned
25192 -- execute CPUID, storing the results in the Result variable
25195 -- the assembler code
25196 "cpuid", -- execute CPUID
25198 -- one is stored in eax
25199 -- processor signature returned in eax
25200 Inputs => Unsigned_32'Asm_input ("a", 1),
25202 -- eax is stored in Result
25203 Outputs => Processor_Signature'Asm_output ("=a", Result),
25205 -- tell compiler that ebx, ecx and edx are also destroyed
25206 Clobber => "ebx, ecx, edx");
25208 -- return processor signature
25213 ------------------------------
25214 -- Get processor features --
25215 ------------------------------
25217 function Features return Processor_Features is
25219 Result : Processor_Features;
25220 -- processor features returned
25224 -- execute CPUID, storing the results in the Result variable
25227 -- the assembler code
25228 "cpuid", -- execute CPUID
25230 -- one stored in eax
25231 -- processor features returned in edx
25232 Inputs => Unsigned_32'Asm_input ("a", 1),
25234 -- edx is stored in Result
25235 Outputs => Processor_Features'Asm_output ("=d", Result),
25237 -- tell compiler that ebx and ecx are also destroyed
25238 Clobber => "ebx, ecx");
25240 -- return processor signature
25247 @c END OF INLINE ASSEMBLER CHAPTER
25248 @c ===============================
25252 @c ***********************************
25253 @c * Compatibility and Porting Guide *
25254 @c ***********************************
25255 @node Compatibility and Porting Guide
25256 @appendix Compatibility and Porting Guide
25259 This chapter describes the compatibility issues that may arise between
25260 GNAT and other Ada 83 and Ada 95 compilation systems, and shows how GNAT
25261 can expedite porting
25262 applications developed in other Ada environments.
25265 * Compatibility with Ada 83::
25266 * Implementation-dependent characteristics::
25267 * Compatibility with DEC Ada 83::
25268 * Compatibility with Other Ada 95 Systems::
25269 * Representation Clauses::
25272 @node Compatibility with Ada 83
25273 @section Compatibility with Ada 83
25274 @cindex Compatibility (between Ada 83 and Ada 95)
25277 Ada 95 is designed to be highly upwards compatible with Ada 83. In
25278 particular, the design intention is that the difficulties associated
25279 with moving from Ada 83 to Ada 95 should be no greater than those
25280 that occur when moving from one Ada 83 system to another.
25282 However, there are a number of points at which there are minor
25283 incompatibilities. The @cite{Ada 95 Annotated Reference Manual} contains
25284 full details of these issues,
25285 and should be consulted for a complete treatment.
25287 following subsections treat the most likely issues to be encountered.
25290 * Legal Ada 83 programs that are illegal in Ada 95::
25291 * More deterministic semantics::
25292 * Changed semantics::
25293 * Other language compatibility issues::
25296 @node Legal Ada 83 programs that are illegal in Ada 95
25297 @subsection Legal Ada 83 programs that are illegal in Ada 95
25300 @item Character literals
25301 Some uses of character literals are ambiguous. Since Ada 95 has introduced
25302 @code{Wide_Character} as a new predefined character type, some uses of
25303 character literals that were legal in Ada 83 are illegal in Ada 95.
25305 @smallexample @c ada
25306 for Char in 'A' .. 'Z' loop ... end loop;
25309 The problem is that @code{'A'} and @code{'Z'} could be from either
25310 @code{Character} or @code{Wide_Character}. The simplest correction
25311 is to make the type explicit; e.g.:
25312 @smallexample @c ada
25313 for Char in Character range 'A' .. 'Z' loop ... end loop;
25316 @item New reserved words
25317 The identifiers @code{abstract}, @code{aliased}, @code{protected},
25318 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
25319 Existing Ada 83 code using any of these identifiers must be edited to
25320 use some alternative name.
25322 @item Freezing rules
25323 The rules in Ada 95 are slightly different with regard to the point at
25324 which entities are frozen, and representation pragmas and clauses are
25325 not permitted past the freeze point. This shows up most typically in
25326 the form of an error message complaining that a representation item
25327 appears too late, and the appropriate corrective action is to move
25328 the item nearer to the declaration of the entity to which it refers.
25330 A particular case is that representation pragmas
25333 extended DEC Ada 83 compatibility pragmas such as @code{Export_Procedure})
25335 cannot be applied to a subprogram body. If necessary, a separate subprogram
25336 declaration must be introduced to which the pragma can be applied.
25338 @item Optional bodies for library packages
25339 In Ada 83, a package that did not require a package body was nevertheless
25340 allowed to have one. This lead to certain surprises in compiling large
25341 systems (situations in which the body could be unexpectedly ignored by the
25342 binder). In Ada 95, if a package does not require a body then it is not
25343 permitted to have a body. To fix this problem, simply remove a redundant
25344 body if it is empty, or, if it is non-empty, introduce a dummy declaration
25345 into the spec that makes the body required. One approach is to add a private
25346 part to the package declaration (if necessary), and define a parameterless
25347 procedure called @code{Requires_Body}, which must then be given a dummy
25348 procedure body in the package body, which then becomes required.
25349 Another approach (assuming that this does not introduce elaboration
25350 circularities) is to add an @code{Elaborate_Body} pragma to the package spec,
25351 since one effect of this pragma is to require the presence of a package body.
25353 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
25354 In Ada 95, the exception @code{Numeric_Error} is a renaming of
25355 @code{Constraint_Error}.
25356 This means that it is illegal to have separate exception handlers for
25357 the two exceptions. The fix is simply to remove the handler for the
25358 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
25359 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
25361 @item Indefinite subtypes in generics
25362 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String})
25363 as the actual for a generic formal private type, but then the instantiation
25364 would be illegal if there were any instances of declarations of variables
25365 of this type in the generic body. In Ada 95, to avoid this clear violation
25366 of the methodological principle known as the ``contract model'',
25367 the generic declaration explicitly indicates whether
25368 or not such instantiations are permitted. If a generic formal parameter
25369 has explicit unknown discriminants, indicated by using @code{(<>)} after the
25370 type name, then it can be instantiated with indefinite types, but no
25371 stand-alone variables can be declared of this type. Any attempt to declare
25372 such a variable will result in an illegality at the time the generic is
25373 declared. If the @code{(<>)} notation is not used, then it is illegal
25374 to instantiate the generic with an indefinite type.
25375 This is the potential incompatibility issue when porting Ada 83 code to Ada 95.
25376 It will show up as a compile time error, and
25377 the fix is usually simply to add the @code{(<>)} to the generic declaration.
25380 @node More deterministic semantics
25381 @subsection More deterministic semantics
25385 Conversions from real types to integer types round away from 0. In Ada 83
25386 the conversion Integer(2.5) could deliver either 2 or 3 as its value. This
25387 implementation freedom was intended to support unbiased rounding in
25388 statistical applications, but in practice it interfered with portability.
25389 In Ada 95 the conversion semantics are unambiguous, and rounding away from 0
25390 is required. Numeric code may be affected by this change in semantics.
25391 Note, though, that this issue is no worse than already existed in Ada 83
25392 when porting code from one vendor to another.
25395 The Real-Time Annex introduces a set of policies that define the behavior of
25396 features that were implementation dependent in Ada 83, such as the order in
25397 which open select branches are executed.
25400 @node Changed semantics
25401 @subsection Changed semantics
25404 The worst kind of incompatibility is one where a program that is legal in
25405 Ada 83 is also legal in Ada 95 but can have an effect in Ada 95 that was not
25406 possible in Ada 83. Fortunately this is extremely rare, but the one
25407 situation that you should be alert to is the change in the predefined type
25408 @code{Character} from 7-bit ASCII to 8-bit Latin-1.
25411 @item range of @code{Character}
25412 The range of @code{Standard.Character} is now the full 256 characters
25413 of Latin-1, whereas in most Ada 83 implementations it was restricted
25414 to 128 characters. Although some of the effects of
25415 this change will be manifest in compile-time rejection of legal
25416 Ada 83 programs it is possible for a working Ada 83 program to have
25417 a different effect in Ada 95, one that was not permitted in Ada 83.
25418 As an example, the expression
25419 @code{Character'Pos(Character'Last)} returned @code{127} in Ada 83 and now
25420 delivers @code{255} as its value.
25421 In general, you should look at the logic of any
25422 character-processing Ada 83 program and see whether it needs to be adapted
25423 to work correctly with Latin-1. Note that the predefined Ada 95 API has a
25424 character handling package that may be relevant if code needs to be adapted
25425 to account for the additional Latin-1 elements.
25426 The desirable fix is to
25427 modify the program to accommodate the full character set, but in some cases
25428 it may be convenient to define a subtype or derived type of Character that
25429 covers only the restricted range.
25433 @node Other language compatibility issues
25434 @subsection Other language compatibility issues
25436 @item @option{-gnat83 switch}
25437 All implementations of GNAT provide a switch that causes GNAT to operate
25438 in Ada 83 mode. In this mode, some but not all compatibility problems
25439 of the type described above are handled automatically. For example, the
25440 new Ada 95 reserved words are treated simply as identifiers as in Ada 83.
25442 in practice, it is usually advisable to make the necessary modifications
25443 to the program to remove the need for using this switch.
25444 See @ref{Compiling Ada 83 Programs}.
25446 @item Support for removed Ada 83 pragmas and attributes
25447 A number of pragmas and attributes from Ada 83 have been removed from Ada 95,
25448 generally because they have been replaced by other mechanisms. Ada 95
25449 compilers are allowed, but not required, to implement these missing
25450 elements. In contrast with some other Ada 95 compilers, GNAT implements all
25451 such pragmas and attributes, eliminating this compatibility concern. These
25452 include @code{pragma Interface} and the floating point type attributes
25453 (@code{Emax}, @code{Mantissa}, etc.), among other items.
25457 @node Implementation-dependent characteristics
25458 @section Implementation-dependent characteristics
25460 Although the Ada language defines the semantics of each construct as
25461 precisely as practical, in some situations (for example for reasons of
25462 efficiency, or where the effect is heavily dependent on the host or target
25463 platform) the implementation is allowed some freedom. In porting Ada 83
25464 code to GNAT, you need to be aware of whether / how the existing code
25465 exercised such implementation dependencies. Such characteristics fall into
25466 several categories, and GNAT offers specific support in assisting the
25467 transition from certain Ada 83 compilers.
25470 * Implementation-defined pragmas::
25471 * Implementation-defined attributes::
25473 * Elaboration order::
25474 * Target-specific aspects::
25478 @node Implementation-defined pragmas
25479 @subsection Implementation-defined pragmas
25482 Ada compilers are allowed to supplement the language-defined pragmas, and
25483 these are a potential source of non-portability. All GNAT-defined pragmas
25484 are described in the GNAT Reference Manual, and these include several that
25485 are specifically intended to correspond to other vendors' Ada 83 pragmas.
25486 For migrating from VADS, the pragma @code{Use_VADS_Size} may be useful.
25488 compatibility with DEC Ada 83, GNAT supplies the pragmas
25489 @code{Extend_System}, @code{Ident}, @code{Inline_Generic},
25490 @code{Interface_Name}, @code{Passive}, @code{Suppress_All},
25491 and @code{Volatile}.
25492 Other relevant pragmas include @code{External} and @code{Link_With}.
25493 Some vendor-specific
25494 Ada 83 pragmas (@code{Share_Generic}, @code{Subtitle}, and @code{Title}) are
25496 avoiding compiler rejection of units that contain such pragmas; they are not
25497 relevant in a GNAT context and hence are not otherwise implemented.
25499 @node Implementation-defined attributes
25500 @subsection Implementation-defined attributes
25502 Analogous to pragmas, the set of attributes may be extended by an
25503 implementation. All GNAT-defined attributes are described in the
25504 @cite{GNAT Reference Manual}, and these include several that are specifically
25506 to correspond to other vendors' Ada 83 attributes. For migrating from VADS,
25507 the attribute @code{VADS_Size} may be useful. For compatibility with DEC
25508 Ada 83, GNAT supplies the attributes @code{Bit}, @code{Machine_Size} and
25512 @subsection Libraries
25514 Vendors may supply libraries to supplement the standard Ada API. If Ada 83
25515 code uses vendor-specific libraries then there are several ways to manage
25519 If the source code for the libraries (specifications and bodies) are
25520 available, then the libraries can be migrated in the same way as the
25523 If the source code for the specifications but not the bodies are
25524 available, then you can reimplement the bodies.
25526 Some new Ada 95 features obviate the need for library support. For
25527 example most Ada 83 vendors supplied a package for unsigned integers. The
25528 Ada 95 modular type feature is the preferred way to handle this need, so
25529 instead of migrating or reimplementing the unsigned integer package it may
25530 be preferable to retrofit the application using modular types.
25533 @node Elaboration order
25534 @subsection Elaboration order
25536 The implementation can choose any elaboration order consistent with the unit
25537 dependency relationship. This freedom means that some orders can result in
25538 Program_Error being raised due to an ``Access Before Elaboration'': an attempt
25539 to invoke a subprogram its body has been elaborated, or to instantiate a
25540 generic before the generic body has been elaborated. By default GNAT
25541 attempts to choose a safe order (one that will not encounter access before
25542 elaboration problems) by implicitly inserting Elaborate_All pragmas where
25543 needed. However, this can lead to the creation of elaboration circularities
25544 and a resulting rejection of the program by gnatbind. This issue is
25545 thoroughly described in @ref{Elaboration Order Handling in GNAT}.
25546 In brief, there are several
25547 ways to deal with this situation:
25551 Modify the program to eliminate the circularities, e.g. by moving
25552 elaboration-time code into explicitly-invoked procedures
25554 Constrain the elaboration order by including explicit @code{Elaborate_Body} or
25555 @code{Elaborate} pragmas, and then inhibit the generation of implicit
25556 @code{Elaborate_All}
25557 pragmas either globally (as an effect of the @option{-gnatE} switch) or locally
25558 (by selectively suppressing elaboration checks via pragma
25559 @code{Suppress(Elaboration_Check)} when it is safe to do so).
25562 @node Target-specific aspects
25563 @subsection Target-specific aspects
25565 Low-level applications need to deal with machine addresses, data
25566 representations, interfacing with assembler code, and similar issues. If
25567 such an Ada 83 application is being ported to different target hardware (for
25568 example where the byte endianness has changed) then you will need to
25569 carefully examine the program logic; the porting effort will heavily depend
25570 on the robustness of the original design. Moreover, Ada 95 is sometimes
25571 incompatible with typical Ada 83 compiler practices regarding implicit
25572 packing, the meaning of the Size attribute, and the size of access values.
25573 GNAT's approach to these issues is described in @ref{Representation Clauses}.
25576 @node Compatibility with Other Ada 95 Systems
25577 @section Compatibility with Other Ada 95 Systems
25580 Providing that programs avoid the use of implementation dependent and
25581 implementation defined features of Ada 95, as documented in the Ada 95
25582 reference manual, there should be a high degree of portability between
25583 GNAT and other Ada 95 systems. The following are specific items which
25584 have proved troublesome in moving GNAT programs to other Ada 95
25585 compilers, but do not affect porting code to GNAT@.
25588 @item Ada 83 Pragmas and Attributes
25589 Ada 95 compilers are allowed, but not required, to implement the missing
25590 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
25591 GNAT implements all such pragmas and attributes, eliminating this as
25592 a compatibility concern, but some other Ada 95 compilers reject these
25593 pragmas and attributes.
25595 @item Special-needs Annexes
25596 GNAT implements the full set of special needs annexes. At the
25597 current time, it is the only Ada 95 compiler to do so. This means that
25598 programs making use of these features may not be portable to other Ada
25599 95 compilation systems.
25601 @item Representation Clauses
25602 Some other Ada 95 compilers implement only the minimal set of
25603 representation clauses required by the Ada 95 reference manual. GNAT goes
25604 far beyond this minimal set, as described in the next section.
25607 @node Representation Clauses
25608 @section Representation Clauses
25611 The Ada 83 reference manual was quite vague in describing both the minimal
25612 required implementation of representation clauses, and also their precise
25613 effects. The Ada 95 reference manual is much more explicit, but the minimal
25614 set of capabilities required in Ada 95 is quite limited.
25616 GNAT implements the full required set of capabilities described in the
25617 Ada 95 reference manual, but also goes much beyond this, and in particular
25618 an effort has been made to be compatible with existing Ada 83 usage to the
25619 greatest extent possible.
25621 A few cases exist in which Ada 83 compiler behavior is incompatible with
25622 requirements in the Ada 95 reference manual. These are instances of
25623 intentional or accidental dependence on specific implementation dependent
25624 characteristics of these Ada 83 compilers. The following is a list of
25625 the cases most likely to arise in existing legacy Ada 83 code.
25628 @item Implicit Packing
25629 Some Ada 83 compilers allowed a Size specification to cause implicit
25630 packing of an array or record. This could cause expensive implicit
25631 conversions for change of representation in the presence of derived
25632 types, and the Ada design intends to avoid this possibility.
25633 Subsequent AI's were issued to make it clear that such implicit
25634 change of representation in response to a Size clause is inadvisable,
25635 and this recommendation is represented explicitly in the Ada 95 RM
25636 as implementation advice that is followed by GNAT@.
25637 The problem will show up as an error
25638 message rejecting the size clause. The fix is simply to provide
25639 the explicit pragma @code{Pack}, or for more fine tuned control, provide
25640 a Component_Size clause.
25642 @item Meaning of Size Attribute
25643 The Size attribute in Ada 95 for discrete types is defined as being the
25644 minimal number of bits required to hold values of the type. For example,
25645 on a 32-bit machine, the size of Natural will typically be 31 and not
25646 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
25647 some 32 in this situation. This problem will usually show up as a compile
25648 time error, but not always. It is a good idea to check all uses of the
25649 'Size attribute when porting Ada 83 code. The GNAT specific attribute
25650 Object_Size can provide a useful way of duplicating the behavior of
25651 some Ada 83 compiler systems.
25653 @item Size of Access Types
25654 A common assumption in Ada 83 code is that an access type is in fact a pointer,
25655 and that therefore it will be the same size as a System.Address value. This
25656 assumption is true for GNAT in most cases with one exception. For the case of
25657 a pointer to an unconstrained array type (where the bounds may vary from one
25658 value of the access type to another), the default is to use a ``fat pointer'',
25659 which is represented as two separate pointers, one to the bounds, and one to
25660 the array. This representation has a number of advantages, including improved
25661 efficiency. However, it may cause some difficulties in porting existing Ada 83
25662 code which makes the assumption that, for example, pointers fit in 32 bits on
25663 a machine with 32-bit addressing.
25665 To get around this problem, GNAT also permits the use of ``thin pointers'' for
25666 access types in this case (where the designated type is an unconstrained array
25667 type). These thin pointers are indeed the same size as a System.Address value.
25668 To specify a thin pointer, use a size clause for the type, for example:
25670 @smallexample @c ada
25671 type X is access all String;
25672 for X'Size use Standard'Address_Size;
25676 which will cause the type X to be represented using a single pointer.
25677 When using this representation, the bounds are right behind the array.
25678 This representation is slightly less efficient, and does not allow quite
25679 such flexibility in the use of foreign pointers or in using the
25680 Unrestricted_Access attribute to create pointers to non-aliased objects.
25681 But for any standard portable use of the access type it will work in
25682 a functionally correct manner and allow porting of existing code.
25683 Note that another way of forcing a thin pointer representation
25684 is to use a component size clause for the element size in an array,
25685 or a record representation clause for an access field in a record.
25688 @node Compatibility with DEC Ada 83
25689 @section Compatibility with DEC Ada 83
25692 The VMS version of GNAT fully implements all the pragmas and attributes
25693 provided by DEC Ada 83, as well as providing the standard DEC Ada 83
25694 libraries, including Starlet. In addition, data layouts and parameter
25695 passing conventions are highly compatible. This means that porting
25696 existing DEC Ada 83 code to GNAT in VMS systems should be easier than
25697 most other porting efforts. The following are some of the most
25698 significant differences between GNAT and DEC Ada 83.
25701 @item Default floating-point representation
25702 In GNAT, the default floating-point format is IEEE, whereas in DEC Ada 83,
25703 it is VMS format. GNAT does implement the necessary pragmas
25704 (Long_Float, Float_Representation) for changing this default.
25707 The package System in GNAT exactly corresponds to the definition in the
25708 Ada 95 reference manual, which means that it excludes many of the
25709 DEC Ada 83 extensions. However, a separate package Aux_DEC is provided
25710 that contains the additional definitions, and a special pragma,
25711 Extend_System allows this package to be treated transparently as an
25712 extension of package System.
25715 The definitions provided by Aux_DEC are exactly compatible with those
25716 in the DEC Ada 83 version of System, with one exception.
25717 DEC Ada provides the following declarations:
25719 @smallexample @c ada
25720 TO_ADDRESS (INTEGER)
25721 TO_ADDRESS (UNSIGNED_LONGWORD)
25722 TO_ADDRESS (universal_integer)
25726 The version of TO_ADDRESS taking a universal integer argument is in fact
25727 an extension to Ada 83 not strictly compatible with the reference manual.
25728 In GNAT, we are constrained to be exactly compatible with the standard,
25729 and this means we cannot provide this capability. In DEC Ada 83, the
25730 point of this definition is to deal with a call like:
25732 @smallexample @c ada
25733 TO_ADDRESS (16#12777#);
25737 Normally, according to the Ada 83 standard, one would expect this to be
25738 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
25739 of TO_ADDRESS@. However, in DEC Ada 83, there is no ambiguity, since the
25740 definition using universal_integer takes precedence.
25742 In GNAT, since the version with universal_integer cannot be supplied, it is
25743 not possible to be 100% compatible. Since there are many programs using
25744 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
25745 to change the name of the function in the UNSIGNED_LONGWORD case, so the
25746 declarations provided in the GNAT version of AUX_Dec are:
25748 @smallexample @c ada
25749 function To_Address (X : Integer) return Address;
25750 pragma Pure_Function (To_Address);
25752 function To_Address_Long (X : Unsigned_Longword)
25754 pragma Pure_Function (To_Address_Long);
25758 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
25759 change the name to TO_ADDRESS_LONG@.
25761 @item Task_Id values
25762 The Task_Id values assigned will be different in the two systems, and GNAT
25763 does not provide a specified value for the Task_Id of the environment task,
25764 which in GNAT is treated like any other declared task.
25767 For full details on these and other less significant compatibility issues,
25768 see appendix E of the Digital publication entitled @cite{DEC Ada, Technical
25769 Overview and Comparison on DIGITAL Platforms}.
25771 For GNAT running on other than VMS systems, all the DEC Ada 83 pragmas and
25772 attributes are recognized, although only a subset of them can sensibly
25773 be implemented. The description of pragmas in this reference manual
25774 indicates whether or not they are applicable to non-VMS systems.
25779 @node Microsoft Windows Topics
25780 @appendix Microsoft Windows Topics
25786 This chapter describes topics that are specific to the Microsoft Windows
25787 platforms (NT, 2000, and XP Professional).
25790 * Using GNAT on Windows::
25791 * Using a network installation of GNAT::
25792 * CONSOLE and WINDOWS subsystems::
25793 * Temporary Files::
25794 * Mixed-Language Programming on Windows::
25795 * Windows Calling Conventions::
25796 * Introduction to Dynamic Link Libraries (DLLs)::
25797 * Using DLLs with GNAT::
25798 * Building DLLs with GNAT::
25799 * GNAT and Windows Resources::
25800 * Debugging a DLL::
25801 * GNAT and COM/DCOM Objects::
25804 @node Using GNAT on Windows
25805 @section Using GNAT on Windows
25808 One of the strengths of the GNAT technology is that its tool set
25809 (@code{gcc}, @code{gnatbind}, @code{gnatlink}, @code{gnatmake}, the
25810 @code{gdb} debugger, etc.) is used in the same way regardless of the
25813 On Windows this tool set is complemented by a number of Microsoft-specific
25814 tools that have been provided to facilitate interoperability with Windows
25815 when this is required. With these tools:
25820 You can build applications using the @code{CONSOLE} or @code{WINDOWS}
25824 You can use any Dynamically Linked Library (DLL) in your Ada code (both
25825 relocatable and non-relocatable DLLs are supported).
25828 You can build Ada DLLs for use in other applications. These applications
25829 can be written in a language other than Ada (e.g., C, C++, etc). Again both
25830 relocatable and non-relocatable Ada DLLs are supported.
25833 You can include Windows resources in your Ada application.
25836 You can use or create COM/DCOM objects.
25840 Immediately below are listed all known general GNAT-for-Windows restrictions.
25841 Other restrictions about specific features like Windows Resources and DLLs
25842 are listed in separate sections below.
25847 It is not possible to use @code{GetLastError} and @code{SetLastError}
25848 when tasking, protected records, or exceptions are used. In these
25849 cases, in order to implement Ada semantics, the GNAT run-time system
25850 calls certain Win32 routines that set the last error variable to 0 upon
25851 success. It should be possible to use @code{GetLastError} and
25852 @code{SetLastError} when tasking, protected record, and exception
25853 features are not used, but it is not guaranteed to work.
25856 It is not possible to link against Microsoft libraries except for
25857 import libraries. The library must be built to be compatible with
25858 @file{MSVCRT.LIB} (/MD Microsoft compiler option), @file{LIBC.LIB} and
25859 @file{LIBCMT.LIB} (/ML or /MT Microsoft compiler options) are known to
25860 not be compatible with the GNAT runtime. Even if the library is
25861 compatible with @file{MSVCRT.LIB} it is not guaranteed to work.
25864 When the compilation environment is located on FAT32 drives, users may
25865 experience recompilations of the source files that have not changed if
25866 Daylight Saving Time (DST) state has changed since the last time files
25867 were compiled. NTFS drives do not have this problem.
25870 No components of the GNAT toolset use any entries in the Windows
25871 registry. The only entries that can be created are file associations and
25872 PATH settings, provided the user has chosen to create them at installation
25873 time, as well as some minimal book-keeping information needed to correctly
25874 uninstall or integrate different GNAT products.
25877 @node Using a network installation of GNAT
25878 @section Using a network installation of GNAT
25881 Make sure the system on which GNAT is installed is accessible from the
25882 current machine, i.e. the install location is shared over the network.
25883 Shared resources are accessed on Windows by means of UNC paths, which
25884 have the format @code{\\server\sharename\path}
25886 In order to use such a network installation, simply add the UNC path of the
25887 @file{bin} directory of your GNAT installation in front of your PATH. For
25888 example, if GNAT is installed in @file{\GNAT} directory of a share location
25889 called @file{c-drive} on a machine @file{LOKI}, the following command will
25892 @code{@ @ @ path \\loki\c-drive\gnat\bin;%path%}
25894 Be aware that every compilation using the network installation results in the
25895 transfer of large amounts of data across the network and will likely cause
25896 serious performance penalty.
25898 @node CONSOLE and WINDOWS subsystems
25899 @section CONSOLE and WINDOWS subsystems
25900 @cindex CONSOLE Subsystem
25901 @cindex WINDOWS Subsystem
25905 There are two main subsystems under Windows. The @code{CONSOLE} subsystem
25906 (which is the default subsystem) will always create a console when
25907 launching the application. This is not something desirable when the
25908 application has a Windows GUI. To get rid of this console the
25909 application must be using the @code{WINDOWS} subsystem. To do so
25910 the @option{-mwindows} linker option must be specified.
25913 $ gnatmake winprog -largs -mwindows
25916 @node Temporary Files
25917 @section Temporary Files
25918 @cindex Temporary files
25921 It is possible to control where temporary files gets created by setting
25922 the TMP environment variable. The file will be created:
25925 @item Under the directory pointed to by the TMP environment variable if
25926 this directory exists.
25928 @item Under c:\temp, if the TMP environment variable is not set (or not
25929 pointing to a directory) and if this directory exists.
25931 @item Under the current working directory otherwise.
25935 This allows you to determine exactly where the temporary
25936 file will be created. This is particularly useful in networked
25937 environments where you may not have write access to some
25940 @node Mixed-Language Programming on Windows
25941 @section Mixed-Language Programming on Windows
25944 Developing pure Ada applications on Windows is no different than on
25945 other GNAT-supported platforms. However, when developing or porting an
25946 application that contains a mix of Ada and C/C++, the choice of your
25947 Windows C/C++ development environment conditions your overall
25948 interoperability strategy.
25950 If you use @code{gcc} to compile the non-Ada part of your application,
25951 there are no Windows-specific restrictions that affect the overall
25952 interoperability with your Ada code. If you plan to use
25953 Microsoft tools (e.g. Microsoft Visual C/C++), you should be aware of
25954 the following limitations:
25958 You cannot link your Ada code with an object or library generated with
25959 Microsoft tools if these use the @code{.tls} section (Thread Local
25960 Storage section) since the GNAT linker does not yet support this section.
25963 You cannot link your Ada code with an object or library generated with
25964 Microsoft tools if these use I/O routines other than those provided in
25965 the Microsoft DLL: @code{msvcrt.dll}. This is because the GNAT run time
25966 uses the services of @code{msvcrt.dll} for its I/Os. Use of other I/O
25967 libraries can cause a conflict with @code{msvcrt.dll} services. For
25968 instance Visual C++ I/O stream routines conflict with those in
25973 If you do want to use the Microsoft tools for your non-Ada code and hit one
25974 of the above limitations, you have two choices:
25978 Encapsulate your non Ada code in a DLL to be linked with your Ada
25979 application. In this case, use the Microsoft or whatever environment to
25980 build the DLL and use GNAT to build your executable
25981 (@pxref{Using DLLs with GNAT}).
25984 Or you can encapsulate your Ada code in a DLL to be linked with the
25985 other part of your application. In this case, use GNAT to build the DLL
25986 (@pxref{Building DLLs with GNAT}) and use the Microsoft or whatever
25987 environment to build your executable.
25990 @node Windows Calling Conventions
25991 @section Windows Calling Conventions
25996 * C Calling Convention::
25997 * Stdcall Calling Convention::
25998 * DLL Calling Convention::
26002 When a subprogram @code{F} (caller) calls a subprogram @code{G}
26003 (callee), there are several ways to push @code{G}'s parameters on the
26004 stack and there are several possible scenarios to clean up the stack
26005 upon @code{G}'s return. A calling convention is an agreed upon software
26006 protocol whereby the responsibilities between the caller (@code{F}) and
26007 the callee (@code{G}) are clearly defined. Several calling conventions
26008 are available for Windows:
26012 @code{C} (Microsoft defined)
26015 @code{Stdcall} (Microsoft defined)
26018 @code{DLL} (GNAT specific)
26021 @node C Calling Convention
26022 @subsection @code{C} Calling Convention
26025 This is the default calling convention used when interfacing to C/C++
26026 routines compiled with either @code{gcc} or Microsoft Visual C++.
26028 In the @code{C} calling convention subprogram parameters are pushed on the
26029 stack by the caller from right to left. The caller itself is in charge of
26030 cleaning up the stack after the call. In addition, the name of a routine
26031 with @code{C} calling convention is mangled by adding a leading underscore.
26033 The name to use on the Ada side when importing (or exporting) a routine
26034 with @code{C} calling convention is the name of the routine. For
26035 instance the C function:
26038 int get_val (long);
26042 should be imported from Ada as follows:
26044 @smallexample @c ada
26046 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
26047 pragma Import (C, Get_Val, External_Name => "get_val");
26052 Note that in this particular case the @code{External_Name} parameter could
26053 have been omitted since, when missing, this parameter is taken to be the
26054 name of the Ada entity in lower case. When the @code{Link_Name} parameter
26055 is missing, as in the above example, this parameter is set to be the
26056 @code{External_Name} with a leading underscore.
26058 When importing a variable defined in C, you should always use the @code{C}
26059 calling convention unless the object containing the variable is part of a
26060 DLL (in which case you should use the @code{DLL} calling convention,
26061 @pxref{DLL Calling Convention}).
26063 @node Stdcall Calling Convention
26064 @subsection @code{Stdcall} Calling Convention
26067 This convention, which was the calling convention used for Pascal
26068 programs, is used by Microsoft for all the routines in the Win32 API for
26069 efficiency reasons. It must be used to import any routine for which this
26070 convention was specified.
26072 In the @code{Stdcall} calling convention subprogram parameters are pushed
26073 on the stack by the caller from right to left. The callee (and not the
26074 caller) is in charge of cleaning the stack on routine exit. In addition,
26075 the name of a routine with @code{Stdcall} calling convention is mangled by
26076 adding a leading underscore (as for the @code{C} calling convention) and a
26077 trailing @code{@@}@code{@i{nn}}, where @i{nn} is the overall size (in
26078 bytes) of the parameters passed to the routine.
26080 The name to use on the Ada side when importing a C routine with a
26081 @code{Stdcall} calling convention is the name of the C routine. The leading
26082 underscore and trailing @code{@@}@code{@i{nn}} are added automatically by
26083 the compiler. For instance the Win32 function:
26086 @b{APIENTRY} int get_val (long);
26090 should be imported from Ada as follows:
26092 @smallexample @c ada
26094 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
26095 pragma Import (Stdcall, Get_Val);
26096 -- On the x86 a long is 4 bytes, so the Link_Name is "_get_val@@4"
26101 As for the @code{C} calling convention, when the @code{External_Name}
26102 parameter is missing, it is taken to be the name of the Ada entity in lower
26103 case. If instead of writing the above import pragma you write:
26105 @smallexample @c ada
26107 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
26108 pragma Import (Stdcall, Get_Val, External_Name => "retrieve_val");
26113 then the imported routine is @code{_retrieve_val@@4}. However, if instead
26114 of specifying the @code{External_Name} parameter you specify the
26115 @code{Link_Name} as in the following example:
26117 @smallexample @c ada
26119 function Get_Val (V : Interfaces.C.long) return Interfaces.C.int;
26120 pragma Import (Stdcall, Get_Val, Link_Name => "retrieve_val");
26125 then the imported routine is @code{retrieve_val@@4}, that is, there is no
26126 trailing underscore but the appropriate @code{@@}@code{@i{nn}} is always
26127 added at the end of the @code{Link_Name} by the compiler.
26130 Note, that in some special cases a DLL's entry point name lacks a trailing
26131 @code{@@}@code{@i{nn}} while the exported name generated for a call has it.
26132 The @code{gnatdll} tool, which creates the import library for the DLL, is able
26133 to handle those cases (see the description of the switches in
26134 @pxref{Using gnatdll} section).
26136 @node DLL Calling Convention
26137 @subsection @code{DLL} Calling Convention
26140 This convention, which is GNAT-specific, must be used when you want to
26141 import in Ada a variables defined in a DLL. For functions and procedures
26142 this convention is equivalent to the @code{Stdcall} convention. As an
26143 example, if a DLL contains a variable defined as:
26150 then, to access this variable from Ada you should write:
26152 @smallexample @c ada
26154 My_Var : Interfaces.C.int;
26155 pragma Import (DLL, My_Var);
26159 The remarks concerning the @code{External_Name} and @code{Link_Name}
26160 parameters given in the previous sections equally apply to the @code{DLL}
26161 calling convention.
26163 @node Introduction to Dynamic Link Libraries (DLLs)
26164 @section Introduction to Dynamic Link Libraries (DLLs)
26168 A Dynamically Linked Library (DLL) is a library that can be shared by
26169 several applications running under Windows. A DLL can contain any number of
26170 routines and variables.
26172 One advantage of DLLs is that you can change and enhance them without
26173 forcing all the applications that depend on them to be relinked or
26174 recompiled. However, you should be aware than all calls to DLL routines are
26175 slower since, as you will understand below, such calls are indirect.
26177 To illustrate the remainder of this section, suppose that an application
26178 wants to use the services of a DLL @file{API.dll}. To use the services
26179 provided by @file{API.dll} you must statically link against an import
26180 library which contains a jump table with an entry for each routine and
26181 variable exported by the DLL. In the Microsoft world this import library is
26182 called @file{API.lib}. When using GNAT this import library is called either
26183 @file{libAPI.a} or @file{libapi.a} (names are case insensitive).
26185 After you have statically linked your application with the import library
26186 and you run your application, here is what happens:
26190 Your application is loaded into memory.
26193 The DLL @file{API.dll} is mapped into the address space of your
26194 application. This means that:
26198 The DLL will use the stack of the calling thread.
26201 The DLL will use the virtual address space of the calling process.
26204 The DLL will allocate memory from the virtual address space of the calling
26208 Handles (pointers) can be safely exchanged between routines in the DLL
26209 routines and routines in the application using the DLL.
26213 The entries in the @file{libAPI.a} or @file{API.lib} jump table which is
26214 part of your application are initialized with the addresses of the routines
26215 and variables in @file{API.dll}.
26218 If present in @file{API.dll}, routines @code{DllMain} or
26219 @code{DllMainCRTStartup} are invoked. These routines typically contain
26220 the initialization code needed for the well-being of the routines and
26221 variables exported by the DLL.
26225 There is an additional point which is worth mentioning. In the Windows
26226 world there are two kind of DLLs: relocatable and non-relocatable
26227 DLLs. Non-relocatable DLLs can only be loaded at a very specific address
26228 in the target application address space. If the addresses of two
26229 non-relocatable DLLs overlap and these happen to be used by the same
26230 application, a conflict will occur and the application will run
26231 incorrectly. Hence, when possible, it is always preferable to use and
26232 build relocatable DLLs. Both relocatable and non-relocatable DLLs are
26233 supported by GNAT. Note that the @option{-s} linker option (see GNU Linker
26234 User's Guide) removes the debugging symbols from the DLL but the DLL can
26235 still be relocated.
26237 As a side note, an interesting difference between Microsoft DLLs and
26238 Unix shared libraries, is the fact that on most Unix systems all public
26239 routines are exported by default in a Unix shared library, while under
26240 Windows the exported routines must be listed explicitly in a definition
26241 file (@pxref{The Definition File}).
26243 @node Using DLLs with GNAT
26244 @section Using DLLs with GNAT
26247 * Creating an Ada Spec for the DLL Services::
26248 * Creating an Import Library::
26252 To use the services of a DLL, say @file{API.dll}, in your Ada application
26257 The Ada spec for the routines and/or variables you want to access in
26258 @file{API.dll}. If not available this Ada spec must be built from the C/C++
26259 header files provided with the DLL.
26262 The import library (@file{libAPI.a} or @file{API.lib}). As previously
26263 mentioned an import library is a statically linked library containing the
26264 import table which will be filled at load time to point to the actual
26265 @file{API.dll} routines. Sometimes you don't have an import library for the
26266 DLL you want to use. The following sections will explain how to build one.
26269 The actual DLL, @file{API.dll}.
26273 Once you have all the above, to compile an Ada application that uses the
26274 services of @file{API.dll} and whose main subprogram is @code{My_Ada_App},
26275 you simply issue the command
26278 $ gnatmake my_ada_app -largs -lAPI
26282 The argument @option{-largs -lAPI} at the end of the @code{gnatmake} command
26283 tells the GNAT linker to look first for a library named @file{API.lib}
26284 (Microsoft-style name) and if not found for a library named @file{libAPI.a}
26285 (GNAT-style name). Note that if the Ada package spec for @file{API.dll}
26286 contains the following pragma
26288 @smallexample @c ada
26289 pragma Linker_Options ("-lAPI");
26293 you do not have to add @option{-largs -lAPI} at the end of the @code{gnatmake}
26296 If any one of the items above is missing you will have to create it
26297 yourself. The following sections explain how to do so using as an
26298 example a fictitious DLL called @file{API.dll}.
26300 @node Creating an Ada Spec for the DLL Services
26301 @subsection Creating an Ada Spec for the DLL Services
26304 A DLL typically comes with a C/C++ header file which provides the
26305 definitions of the routines and variables exported by the DLL. The Ada
26306 equivalent of this header file is a package spec that contains definitions
26307 for the imported entities. If the DLL you intend to use does not come with
26308 an Ada spec you have to generate one such spec yourself. For example if
26309 the header file of @file{API.dll} is a file @file{api.h} containing the
26310 following two definitions:
26322 then the equivalent Ada spec could be:
26324 @smallexample @c ada
26327 with Interfaces.C.Strings;
26332 function Get (Str : C.Strings.Chars_Ptr) return C.int;
26335 pragma Import (C, Get);
26336 pragma Import (DLL, Some_Var);
26343 Note that a variable is @strong{always imported with a DLL convention}. A
26344 function can have @code{C}, @code{Stdcall} or @code{DLL} convention. For
26345 subprograms, the @code{DLL} convention is a synonym of @code{Stdcall}
26346 (@pxref{Windows Calling Conventions}).
26348 @node Creating an Import Library
26349 @subsection Creating an Import Library
26350 @cindex Import library
26353 * The Definition File::
26354 * GNAT-Style Import Library::
26355 * Microsoft-Style Import Library::
26359 If a Microsoft-style import library @file{API.lib} or a GNAT-style
26360 import library @file{libAPI.a} is available with @file{API.dll} you
26361 can skip this section. Otherwise read on.
26363 @node The Definition File
26364 @subsubsection The Definition File
26365 @cindex Definition file
26369 As previously mentioned, and unlike Unix systems, the list of symbols
26370 that are exported from a DLL must be provided explicitly in Windows.
26371 The main goal of a definition file is precisely that: list the symbols
26372 exported by a DLL. A definition file (usually a file with a @code{.def}
26373 suffix) has the following structure:
26379 [DESCRIPTION @i{string}]
26389 @item LIBRARY @i{name}
26390 This section, which is optional, gives the name of the DLL.
26392 @item DESCRIPTION @i{string}
26393 This section, which is optional, gives a description string that will be
26394 embedded in the import library.
26397 This section gives the list of exported symbols (procedures, functions or
26398 variables). For instance in the case of @file{API.dll} the @code{EXPORTS}
26399 section of @file{API.def} looks like:
26413 Note that you must specify the correct suffix (@code{@@}@code{@i{nn}})
26414 (@pxref{Windows Calling Conventions}) for a Stdcall
26415 calling convention function in the exported symbols list.
26418 There can actually be other sections in a definition file, but these
26419 sections are not relevant to the discussion at hand.
26421 @node GNAT-Style Import Library
26422 @subsubsection GNAT-Style Import Library
26425 To create a static import library from @file{API.dll} with the GNAT tools
26426 you should proceed as follows:
26430 Create the definition file @file{API.def} (@pxref{The Definition File}).
26431 For that use the @code{dll2def} tool as follows:
26434 $ dll2def API.dll > API.def
26438 @code{dll2def} is a very simple tool: it takes as input a DLL and prints
26439 to standard output the list of entry points in the DLL. Note that if
26440 some routines in the DLL have the @code{Stdcall} convention
26441 (@pxref{Windows Calling Conventions}) with stripped @code{@@}@i{nn}
26442 suffix then you'll have to edit @file{api.def} to add it.
26445 Here are some hints to find the right @code{@@}@i{nn} suffix.
26449 If you have the Microsoft import library (.lib), it is possible to get
26450 the right symbols by using Microsoft @code{dumpbin} tool (see the
26451 corresponding Microsoft documentation for further details).
26454 $ dumpbin /exports api.lib
26458 If you have a message about a missing symbol at link time the compiler
26459 tells you what symbol is expected. You just have to go back to the
26460 definition file and add the right suffix.
26464 Build the import library @code{libAPI.a}, using @code{gnatdll}
26465 (@pxref{Using gnatdll}) as follows:
26468 $ gnatdll -e API.def -d API.dll
26472 @code{gnatdll} takes as input a definition file @file{API.def} and the
26473 name of the DLL containing the services listed in the definition file
26474 @file{API.dll}. The name of the static import library generated is
26475 computed from the name of the definition file as follows: if the
26476 definition file name is @i{xyz}@code{.def}, the import library name will
26477 be @code{lib}@i{xyz}@code{.a}. Note that in the previous example option
26478 @option{-e} could have been removed because the name of the definition
26479 file (before the ``@code{.def}'' suffix) is the same as the name of the
26480 DLL (@pxref{Using gnatdll} for more information about @code{gnatdll}).
26483 @node Microsoft-Style Import Library
26484 @subsubsection Microsoft-Style Import Library
26487 With GNAT you can either use a GNAT-style or Microsoft-style import
26488 library. A Microsoft import library is needed only if you plan to make an
26489 Ada DLL available to applications developed with Microsoft
26490 tools (@pxref{Mixed-Language Programming on Windows}).
26492 To create a Microsoft-style import library for @file{API.dll} you
26493 should proceed as follows:
26497 Create the definition file @file{API.def} from the DLL. For this use either
26498 the @code{dll2def} tool as described above or the Microsoft @code{dumpbin}
26499 tool (see the corresponding Microsoft documentation for further details).
26502 Build the actual import library using Microsoft's @code{lib} utility:
26505 $ lib -machine:IX86 -def:API.def -out:API.lib
26509 If you use the above command the definition file @file{API.def} must
26510 contain a line giving the name of the DLL:
26517 See the Microsoft documentation for further details about the usage of
26521 @node Building DLLs with GNAT
26522 @section Building DLLs with GNAT
26523 @cindex DLLs, building
26526 * Limitations When Using Ada DLLs from Ada::
26527 * Exporting Ada Entities::
26528 * Ada DLLs and Elaboration::
26529 * Ada DLLs and Finalization::
26530 * Creating a Spec for Ada DLLs::
26531 * Creating the Definition File::
26536 This section explains how to build DLLs containing Ada code. These DLLs
26537 will be referred to as Ada DLLs in the remainder of this section.
26539 The steps required to build an Ada DLL that is to be used by Ada as well as
26540 non-Ada applications are as follows:
26544 You need to mark each Ada @i{entity} exported by the DLL with a @code{C} or
26545 @code{Stdcall} calling convention to avoid any Ada name mangling for the
26546 entities exported by the DLL (@pxref{Exporting Ada Entities}). You can
26547 skip this step if you plan to use the Ada DLL only from Ada applications.
26550 Your Ada code must export an initialization routine which calls the routine
26551 @code{adainit} generated by @code{gnatbind} to perform the elaboration of
26552 the Ada code in the DLL (@pxref{Ada DLLs and Elaboration}). The initialization
26553 routine exported by the Ada DLL must be invoked by the clients of the DLL
26554 to initialize the DLL.
26557 When useful, the DLL should also export a finalization routine which calls
26558 routine @code{adafinal} generated by @code{gnatbind} to perform the
26559 finalization of the Ada code in the DLL (@pxref{Ada DLLs and Finalization}).
26560 The finalization routine exported by the Ada DLL must be invoked by the
26561 clients of the DLL when the DLL services are no further needed.
26564 You must provide a spec for the services exported by the Ada DLL in each
26565 of the programming languages to which you plan to make the DLL available.
26568 You must provide a definition file listing the exported entities
26569 (@pxref{The Definition File}).
26572 Finally you must use @code{gnatdll} to produce the DLL and the import
26573 library (@pxref{Using gnatdll}).
26577 Note that a relocatable DLL stripped using the @code{strip} binutils
26578 tool will not be relocatable anymore. To build a DLL without debug
26579 information pass @code{-largs -s} to @code{gnatdll}.
26581 @node Limitations When Using Ada DLLs from Ada
26582 @subsection Limitations When Using Ada DLLs from Ada
26585 When using Ada DLLs from Ada applications there is a limitation users
26586 should be aware of. Because on Windows the GNAT run time is not in a DLL of
26587 its own, each Ada DLL includes a part of the GNAT run time. Specifically,
26588 each Ada DLL includes the services of the GNAT run time that are necessary
26589 to the Ada code inside the DLL. As a result, when an Ada program uses an
26590 Ada DLL there are two independent GNAT run times: one in the Ada DLL and
26591 one in the main program.
26593 It is therefore not possible to exchange GNAT run-time objects between the
26594 Ada DLL and the main Ada program. Example of GNAT run-time objects are file
26595 handles (e.g. @code{Text_IO.File_Type}), tasks types, protected objects
26598 It is completely safe to exchange plain elementary, array or record types,
26599 Windows object handles, etc.
26601 @node Exporting Ada Entities
26602 @subsection Exporting Ada Entities
26603 @cindex Export table
26606 Building a DLL is a way to encapsulate a set of services usable from any
26607 application. As a result, the Ada entities exported by a DLL should be
26608 exported with the @code{C} or @code{Stdcall} calling conventions to avoid
26609 any Ada name mangling. Please note that the @code{Stdcall} convention
26610 should only be used for subprograms, not for variables. As an example here
26611 is an Ada package @code{API}, spec and body, exporting two procedures, a
26612 function, and a variable:
26614 @smallexample @c ada
26617 with Interfaces.C; use Interfaces;
26619 Count : C.int := 0;
26620 function Factorial (Val : C.int) return C.int;
26622 procedure Initialize_API;
26623 procedure Finalize_API;
26624 -- Initialization & Finalization routines. More in the next section.
26626 pragma Export (C, Initialize_API);
26627 pragma Export (C, Finalize_API);
26628 pragma Export (C, Count);
26629 pragma Export (C, Factorial);
26635 @smallexample @c ada
26638 package body API is
26639 function Factorial (Val : C.int) return C.int is
26642 Count := Count + 1;
26643 for K in 1 .. Val loop
26649 procedure Initialize_API is
26651 pragma Import (C, Adainit);
26654 end Initialize_API;
26656 procedure Finalize_API is
26657 procedure Adafinal;
26658 pragma Import (C, Adafinal);
26668 If the Ada DLL you are building will only be used by Ada applications
26669 you do not have to export Ada entities with a @code{C} or @code{Stdcall}
26670 convention. As an example, the previous package could be written as
26673 @smallexample @c ada
26677 Count : Integer := 0;
26678 function Factorial (Val : Integer) return Integer;
26680 procedure Initialize_API;
26681 procedure Finalize_API;
26682 -- Initialization and Finalization routines.
26688 @smallexample @c ada
26691 package body API is
26692 function Factorial (Val : Integer) return Integer is
26693 Fact : Integer := 1;
26695 Count := Count + 1;
26696 for K in 1 .. Val loop
26703 -- The remainder of this package body is unchanged.
26710 Note that if you do not export the Ada entities with a @code{C} or
26711 @code{Stdcall} convention you will have to provide the mangled Ada names
26712 in the definition file of the Ada DLL
26713 (@pxref{Creating the Definition File}).
26715 @node Ada DLLs and Elaboration
26716 @subsection Ada DLLs and Elaboration
26717 @cindex DLLs and elaboration
26720 The DLL that you are building contains your Ada code as well as all the
26721 routines in the Ada library that are needed by it. The first thing a
26722 user of your DLL must do is elaborate the Ada code
26723 (@pxref{Elaboration Order Handling in GNAT}).
26725 To achieve this you must export an initialization routine
26726 (@code{Initialize_API} in the previous example), which must be invoked
26727 before using any of the DLL services. This elaboration routine must call
26728 the Ada elaboration routine @code{adainit} generated by the GNAT binder
26729 (@pxref{Binding with Non-Ada Main Programs}). See the body of
26730 @code{Initialize_Api} for an example. Note that the GNAT binder is
26731 automatically invoked during the DLL build process by the @code{gnatdll}
26732 tool (@pxref{Using gnatdll}).
26734 When a DLL is loaded, Windows systematically invokes a routine called
26735 @code{DllMain}. It would therefore be possible to call @code{adainit}
26736 directly from @code{DllMain} without having to provide an explicit
26737 initialization routine. Unfortunately, it is not possible to call
26738 @code{adainit} from the @code{DllMain} if your program has library level
26739 tasks because access to the @code{DllMain} entry point is serialized by
26740 the system (that is, only a single thread can execute ``through'' it at a
26741 time), which means that the GNAT run time will deadlock waiting for the
26742 newly created task to complete its initialization.
26744 @node Ada DLLs and Finalization
26745 @subsection Ada DLLs and Finalization
26746 @cindex DLLs and finalization
26749 When the services of an Ada DLL are no longer needed, the client code should
26750 invoke the DLL finalization routine, if available. The DLL finalization
26751 routine is in charge of releasing all resources acquired by the DLL. In the
26752 case of the Ada code contained in the DLL, this is achieved by calling
26753 routine @code{adafinal} generated by the GNAT binder
26754 (@pxref{Binding with Non-Ada Main Programs}).
26755 See the body of @code{Finalize_Api} for an
26756 example. As already pointed out the GNAT binder is automatically invoked
26757 during the DLL build process by the @code{gnatdll} tool
26758 (@pxref{Using gnatdll}).
26760 @node Creating a Spec for Ada DLLs
26761 @subsection Creating a Spec for Ada DLLs
26764 To use the services exported by the Ada DLL from another programming
26765 language (e.g. C), you have to translate the specs of the exported Ada
26766 entities in that language. For instance in the case of @code{API.dll},
26767 the corresponding C header file could look like:
26772 extern int *_imp__count;
26773 #define count (*_imp__count)
26774 int factorial (int);
26780 It is important to understand that when building an Ada DLL to be used by
26781 other Ada applications, you need two different specs for the packages
26782 contained in the DLL: one for building the DLL and the other for using
26783 the DLL. This is because the @code{DLL} calling convention is needed to
26784 use a variable defined in a DLL, but when building the DLL, the variable
26785 must have either the @code{Ada} or @code{C} calling convention. As an
26786 example consider a DLL comprising the following package @code{API}:
26788 @smallexample @c ada
26792 Count : Integer := 0;
26794 -- Remainder of the package omitted.
26801 After producing a DLL containing package @code{API}, the spec that
26802 must be used to import @code{API.Count} from Ada code outside of the
26805 @smallexample @c ada
26810 pragma Import (DLL, Count);
26816 @node Creating the Definition File
26817 @subsection Creating the Definition File
26820 The definition file is the last file needed to build the DLL. It lists
26821 the exported symbols. As an example, the definition file for a DLL
26822 containing only package @code{API} (where all the entities are exported
26823 with a @code{C} calling convention) is:
26838 If the @code{C} calling convention is missing from package @code{API},
26839 then the definition file contains the mangled Ada names of the above
26840 entities, which in this case are:
26849 api__initialize_api
26854 @node Using gnatdll
26855 @subsection Using @code{gnatdll}
26859 * gnatdll Example::
26860 * gnatdll behind the Scenes::
26865 @code{gnatdll} is a tool to automate the DLL build process once all the Ada
26866 and non-Ada sources that make up your DLL have been compiled.
26867 @code{gnatdll} is actually in charge of two distinct tasks: build the
26868 static import library for the DLL and the actual DLL. The form of the
26869 @code{gnatdll} command is
26873 $ gnatdll [@var{switches}] @var{list-of-files} [-largs @var{opts}]
26878 where @i{list-of-files} is a list of ALI and object files. The object
26879 file list must be the exact list of objects corresponding to the non-Ada
26880 sources whose services are to be included in the DLL. The ALI file list
26881 must be the exact list of ALI files for the corresponding Ada sources
26882 whose services are to be included in the DLL. If @i{list-of-files} is
26883 missing, only the static import library is generated.
26886 You may specify any of the following switches to @code{gnatdll}:
26889 @item -a[@var{address}]
26890 @cindex @option{-a} (@code{gnatdll})
26891 Build a non-relocatable DLL at @var{address}. If @var{address} is not
26892 specified the default address @var{0x11000000} will be used. By default,
26893 when this switch is missing, @code{gnatdll} builds relocatable DLL. We
26894 advise the reader to build relocatable DLL.
26896 @item -b @var{address}
26897 @cindex @option{-b} (@code{gnatdll})
26898 Set the relocatable DLL base address. By default the address is
26901 @item -bargs @var{opts}
26902 @cindex @option{-bargs} (@code{gnatdll})
26903 Binder options. Pass @var{opts} to the binder.
26905 @item -d @var{dllfile}
26906 @cindex @option{-d} (@code{gnatdll})
26907 @var{dllfile} is the name of the DLL. This switch must be present for
26908 @code{gnatdll} to do anything. The name of the generated import library is
26909 obtained algorithmically from @var{dllfile} as shown in the following
26910 example: if @var{dllfile} is @code{xyz.dll}, the import library name is
26911 @code{libxyz.a}. The name of the definition file to use (if not specified
26912 by option @option{-e}) is obtained algorithmically from @var{dllfile}
26913 as shown in the following example:
26914 if @var{dllfile} is @code{xyz.dll}, the definition
26915 file used is @code{xyz.def}.
26917 @item -e @var{deffile}
26918 @cindex @option{-e} (@code{gnatdll})
26919 @var{deffile} is the name of the definition file.
26922 @cindex @option{-g} (@code{gnatdll})
26923 Generate debugging information. This information is stored in the object
26924 file and copied from there to the final DLL file by the linker,
26925 where it can be read by the debugger. You must use the
26926 @option{-g} switch if you plan on using the debugger or the symbolic
26930 @cindex @option{-h} (@code{gnatdll})
26931 Help mode. Displays @code{gnatdll} switch usage information.
26934 @cindex @option{-I} (@code{gnatdll})
26935 Direct @code{gnatdll} to search the @var{dir} directory for source and
26936 object files needed to build the DLL.
26937 (@pxref{Search Paths and the Run-Time Library (RTL)}).
26940 @cindex @option{-k} (@code{gnatdll})
26941 Removes the @code{@@}@i{nn} suffix from the import library's exported
26942 names. You must specified this option if you want to use a
26943 @code{Stdcall} function in a DLL for which the @code{@@}@i{nn} suffix
26944 has been removed. This is the case for most of the Windows NT DLL for
26945 example. This option has no effect when @option{-n} option is specified.
26947 @item -l @var{file}
26948 @cindex @option{-l} (@code{gnatdll})
26949 The list of ALI and object files used to build the DLL are listed in
26950 @var{file}, instead of being given in the command line. Each line in
26951 @var{file} contains the name of an ALI or object file.
26954 @cindex @option{-n} (@code{gnatdll})
26955 No Import. Do not create the import library.
26958 @cindex @option{-q} (@code{gnatdll})
26959 Quiet mode. Do not display unnecessary messages.
26962 @cindex @option{-v} (@code{gnatdll})
26963 Verbose mode. Display extra information.
26965 @item -largs @var{opts}
26966 @cindex @option{-largs} (@code{gnatdll})
26967 Linker options. Pass @var{opts} to the linker.
26970 @node gnatdll Example
26971 @subsubsection @code{gnatdll} Example
26974 As an example the command to build a relocatable DLL from @file{api.adb}
26975 once @file{api.adb} has been compiled and @file{api.def} created is
26978 $ gnatdll -d api.dll api.ali
26982 The above command creates two files: @file{libapi.a} (the import
26983 library) and @file{api.dll} (the actual DLL). If you want to create
26984 only the DLL, just type:
26987 $ gnatdll -d api.dll -n api.ali
26991 Alternatively if you want to create just the import library, type:
26994 $ gnatdll -d api.dll
26997 @node gnatdll behind the Scenes
26998 @subsubsection @code{gnatdll} behind the Scenes
27001 This section details the steps involved in creating a DLL. @code{gnatdll}
27002 does these steps for you. Unless you are interested in understanding what
27003 goes on behind the scenes, you should skip this section.
27005 We use the previous example of a DLL containing the Ada package @code{API},
27006 to illustrate the steps necessary to build a DLL. The starting point is a
27007 set of objects that will make up the DLL and the corresponding ALI
27008 files. In the case of this example this means that @file{api.o} and
27009 @file{api.ali} are available. To build a relocatable DLL, @code{gnatdll} does
27014 @code{gnatdll} builds the base file (@file{api.base}). A base file gives
27015 the information necessary to generate relocation information for the
27021 $ gnatlink api -o api.jnk -mdll -Wl,--base-file,api.base
27026 In addition to the base file, the @code{gnatlink} command generates an
27027 output file @file{api.jnk} which can be discarded. The @option{-mdll} switch
27028 asks @code{gnatlink} to generate the routines @code{DllMain} and
27029 @code{DllMainCRTStartup} that are called by the Windows loader when the DLL
27030 is loaded into memory.
27033 @code{gnatdll} uses @code{dlltool} (@pxref{Using dlltool}) to build the
27034 export table (@file{api.exp}). The export table contains the relocation
27035 information in a form which can be used during the final link to ensure
27036 that the Windows loader is able to place the DLL anywhere in memory.
27040 $ dlltool --dllname api.dll --def api.def --base-file api.base \
27041 --output-exp api.exp
27046 @code{gnatdll} builds the base file using the new export table. Note that
27047 @code{gnatbind} must be called once again since the binder generated file
27048 has been deleted during the previous call to @code{gnatlink}.
27053 $ gnatlink api -o api.jnk api.exp -mdll
27054 -Wl,--base-file,api.base
27059 @code{gnatdll} builds the new export table using the new base file and
27060 generates the DLL import library @file{libAPI.a}.
27064 $ dlltool --dllname api.dll --def api.def --base-file api.base \
27065 --output-exp api.exp --output-lib libAPI.a
27070 Finally @code{gnatdll} builds the relocatable DLL using the final export
27076 $ gnatlink api api.exp -o api.dll -mdll
27081 @node Using dlltool
27082 @subsubsection Using @code{dlltool}
27085 @code{dlltool} is the low-level tool used by @code{gnatdll} to build
27086 DLLs and static import libraries. This section summarizes the most
27087 common @code{dlltool} switches. The form of the @code{dlltool} command
27091 $ dlltool [@var{switches}]
27095 @code{dlltool} switches include:
27098 @item --base-file @var{basefile}
27099 @cindex @option{--base-file} (@command{dlltool})
27100 Read the base file @var{basefile} generated by the linker. This switch
27101 is used to create a relocatable DLL.
27103 @item --def @var{deffile}
27104 @cindex @option{--def} (@command{dlltool})
27105 Read the definition file.
27107 @item --dllname @var{name}
27108 @cindex @option{--dllname} (@command{dlltool})
27109 Gives the name of the DLL. This switch is used to embed the name of the
27110 DLL in the static import library generated by @code{dlltool} with switch
27111 @option{--output-lib}.
27114 @cindex @option{-k} (@command{dlltool})
27115 Kill @code{@@}@i{nn} from exported names
27116 (@pxref{Windows Calling Conventions}
27117 for a discussion about @code{Stdcall}-style symbols.
27120 @cindex @option{--help} (@command{dlltool})
27121 Prints the @code{dlltool} switches with a concise description.
27123 @item --output-exp @var{exportfile}
27124 @cindex @option{--output-exp} (@command{dlltool})
27125 Generate an export file @var{exportfile}. The export file contains the
27126 export table (list of symbols in the DLL) and is used to create the DLL.
27128 @item --output-lib @i{libfile}
27129 @cindex @option{--output-lib} (@command{dlltool})
27130 Generate a static import library @var{libfile}.
27133 @cindex @option{-v} (@command{dlltool})
27136 @item --as @i{assembler-name}
27137 @cindex @option{--as} (@command{dlltool})
27138 Use @i{assembler-name} as the assembler. The default is @code{as}.
27141 @node GNAT and Windows Resources
27142 @section GNAT and Windows Resources
27143 @cindex Resources, windows
27146 * Building Resources::
27147 * Compiling Resources::
27148 * Using Resources::
27152 Resources are an easy way to add Windows specific objects to your
27153 application. The objects that can be added as resources include:
27182 This section explains how to build, compile and use resources.
27184 @node Building Resources
27185 @subsection Building Resources
27186 @cindex Resources, building
27189 A resource file is an ASCII file. By convention resource files have an
27190 @file{.rc} extension.
27191 The easiest way to build a resource file is to use Microsoft tools
27192 such as @code{imagedit.exe} to build bitmaps, icons and cursors and
27193 @code{dlgedit.exe} to build dialogs.
27194 It is always possible to build an @file{.rc} file yourself by writing a
27197 It is not our objective to explain how to write a resource file. A
27198 complete description of the resource script language can be found in the
27199 Microsoft documentation.
27201 @node Compiling Resources
27202 @subsection Compiling Resources
27205 @cindex Resources, compiling
27208 This section describes how to build a GNAT-compatible (COFF) object file
27209 containing the resources. This is done using the Resource Compiler
27210 @code{windres} as follows:
27213 $ windres -i myres.rc -o myres.o
27217 By default @code{windres} will run @code{gcc} to preprocess the @file{.rc}
27218 file. You can specify an alternate preprocessor (usually named
27219 @file{cpp.exe}) using the @code{windres} @option{--preprocessor}
27220 parameter. A list of all possible options may be obtained by entering
27221 the command @code{windres} @option{--help}.
27223 It is also possible to use the Microsoft resource compiler @code{rc.exe}
27224 to produce a @file{.res} file (binary resource file). See the
27225 corresponding Microsoft documentation for further details. In this case
27226 you need to use @code{windres} to translate the @file{.res} file to a
27227 GNAT-compatible object file as follows:
27230 $ windres -i myres.res -o myres.o
27233 @node Using Resources
27234 @subsection Using Resources
27235 @cindex Resources, using
27238 To include the resource file in your program just add the
27239 GNAT-compatible object file for the resource(s) to the linker
27240 arguments. With @code{gnatmake} this is done by using the @option{-largs}
27244 $ gnatmake myprog -largs myres.o
27247 @node Debugging a DLL
27248 @section Debugging a DLL
27249 @cindex DLL debugging
27252 * Program and DLL Both Built with GCC/GNAT::
27253 * Program Built with Foreign Tools and DLL Built with GCC/GNAT::
27257 Debugging a DLL is similar to debugging a standard program. But
27258 we have to deal with two different executable parts: the DLL and the
27259 program that uses it. We have the following four possibilities:
27263 The program and the DLL are built with @code{GCC/GNAT}.
27265 The program is built with foreign tools and the DLL is built with
27268 The program is built with @code{GCC/GNAT} and the DLL is built with
27274 In this section we address only cases one and two above.
27275 There is no point in trying to debug
27276 a DLL with @code{GNU/GDB}, if there is no GDB-compatible debugging
27277 information in it. To do so you must use a debugger compatible with the
27278 tools suite used to build the DLL.
27280 @node Program and DLL Both Built with GCC/GNAT
27281 @subsection Program and DLL Both Built with GCC/GNAT
27284 This is the simplest case. Both the DLL and the program have @code{GDB}
27285 compatible debugging information. It is then possible to break anywhere in
27286 the process. Let's suppose here that the main procedure is named
27287 @code{ada_main} and that in the DLL there is an entry point named
27291 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) and
27292 program must have been built with the debugging information (see GNAT -g
27293 switch). Here are the step-by-step instructions for debugging it:
27296 @item Launch @code{GDB} on the main program.
27302 @item Break on the main procedure and run the program.
27305 (gdb) break ada_main
27310 This step is required to be able to set a breakpoint inside the DLL. As long
27311 as the program is not run, the DLL is not loaded. This has the
27312 consequence that the DLL debugging information is also not loaded, so it is not
27313 possible to set a breakpoint in the DLL.
27315 @item Set a breakpoint inside the DLL
27318 (gdb) break ada_dll
27325 At this stage a breakpoint is set inside the DLL. From there on
27326 you can use the standard approach to debug the whole program
27327 (@pxref{Running and Debugging Ada Programs}).
27329 @node Program Built with Foreign Tools and DLL Built with GCC/GNAT
27330 @subsection Program Built with Foreign Tools and DLL Built with GCC/GNAT
27333 * Debugging the DLL Directly::
27334 * Attaching to a Running Process::
27338 In this case things are slightly more complex because it is not possible to
27339 start the main program and then break at the beginning to load the DLL and the
27340 associated DLL debugging information. It is not possible to break at the
27341 beginning of the program because there is no @code{GDB} debugging information,
27342 and therefore there is no direct way of getting initial control. This
27343 section addresses this issue by describing some methods that can be used
27344 to break somewhere in the DLL to debug it.
27347 First suppose that the main procedure is named @code{main} (this is for
27348 example some C code built with Microsoft Visual C) and that there is a
27349 DLL named @code{test.dll} containing an Ada entry point named
27353 The DLL (@pxref{Introduction to Dynamic Link Libraries (DLLs)}) must have
27354 been built with debugging information (see GNAT -g option).
27356 @node Debugging the DLL Directly
27357 @subsubsection Debugging the DLL Directly
27361 Launch the debugger on the DLL.
27367 @item Set a breakpoint on a DLL subroutine.
27370 (gdb) break ada_dll
27374 Specify the executable file to @code{GDB}.
27377 (gdb) exec-file main.exe
27388 This will run the program until it reaches the breakpoint that has been
27389 set. From that point you can use the standard way to debug a program
27390 as described in (@pxref{Running and Debugging Ada Programs}).
27395 It is also possible to debug the DLL by attaching to a running process.
27397 @node Attaching to a Running Process
27398 @subsubsection Attaching to a Running Process
27399 @cindex DLL debugging, attach to process
27402 With @code{GDB} it is always possible to debug a running process by
27403 attaching to it. It is possible to debug a DLL this way. The limitation
27404 of this approach is that the DLL must run long enough to perform the
27405 attach operation. It may be useful for instance to insert a time wasting
27406 loop in the code of the DLL to meet this criterion.
27410 @item Launch the main program @file{main.exe}.
27416 @item Use the Windows @i{Task Manager} to find the process ID. Let's say
27417 that the process PID for @file{main.exe} is 208.
27425 @item Attach to the running process to be debugged.
27431 @item Load the process debugging information.
27434 (gdb) symbol-file main.exe
27437 @item Break somewhere in the DLL.
27440 (gdb) break ada_dll
27443 @item Continue process execution.
27452 This last step will resume the process execution, and stop at
27453 the breakpoint we have set. From there you can use the standard
27454 approach to debug a program as described in
27455 (@pxref{Running and Debugging Ada Programs}).
27457 @node GNAT and COM/DCOM Objects
27458 @section GNAT and COM/DCOM Objects
27463 This section is temporarily left blank.
27468 @c **********************************
27469 @c * GNU Free Documentation License *
27470 @c **********************************
27472 @c GNU Free Documentation License
27474 @node Index,,GNU Free Documentation License, Top
27480 @c Put table of contents at end, otherwise it precedes the "title page" in
27481 @c the .txt version
27482 @c Edit the pdf file to move the contents to the beginning, after the title