1 \input texinfo @c -*-texinfo-*-
5 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
7 @c GNAT DOCUMENTATION o
11 @c Copyright (C) 1995-2004 Free Software Foundation o
14 @c GNAT is maintained by Ada Core Technologies Inc (http://www.gnat.com). o
16 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
18 @setfilename gnat_rm.info
22 @settitle GNAT Reference Manual
24 @setchapternewpage odd
27 @include gcc-common.texi
29 @dircategory GNU Ada tools
31 * GNAT Reference Manual: (gnat_rm). Reference Manual for GNU Ada tools.
35 Copyright @copyright{} 1995-2004, Free Software Foundation
37 Permission is granted to copy, distribute and/or modify this document
38 under the terms of the GNU Free Documentation License, Version 1.2
39 or any later version published by the Free Software Foundation;
40 with the Invariant Sections being ``GNU Free Documentation License'',
41 with the Front-Cover Texts being ``GNAT Reference Manual'', and with
42 no Back-Cover Texts. A copy of the license is included in the section
43 entitled ``GNU Free Documentation License''.
48 @title GNAT Reference Manual
49 @subtitle GNAT, The GNU Ada 95 Compiler
50 @subtitle GCC version @value{version-GCC}
51 @author Ada Core Technologies, Inc.
54 @vskip 0pt plus 1filll
61 @node Top, About This Guide, (dir), (dir)
62 @top GNAT Reference Manual
68 GNAT, The GNU Ada 95 Compiler@*
69 GCC version @value{version-GCC}@*
72 Ada Core Technologies, Inc.
76 * Implementation Defined Pragmas::
77 * Implementation Defined Attributes::
78 * Implementation Advice::
79 * Implementation Defined Characteristics::
80 * Intrinsic Subprograms::
81 * Representation Clauses and Pragmas::
82 * Standard Library Routines::
83 * The Implementation of Standard I/O::
85 * Interfacing to Other Languages::
86 * Specialized Needs Annexes::
87 * Implementation of Specific Ada Features::
88 * Project File Reference::
89 * Obsolescent Features::
90 * GNU Free Documentation License::
93 --- The Detailed Node Listing ---
97 * What This Reference Manual Contains::
98 * Related Information::
100 Implementation Defined Pragmas
102 * Pragma Abort_Defer::
108 * Pragma C_Pass_By_Copy::
110 * Pragma Common_Object::
111 * Pragma Compile_Time_Warning::
112 * Pragma Complex_Representation::
113 * Pragma Component_Alignment::
114 * Pragma Convention_Identifier::
116 * Pragma CPP_Constructor::
117 * Pragma CPP_Virtual::
118 * Pragma CPP_Vtable::
120 * Pragma Detect_Blocking::
121 * Pragma Elaboration_Checks::
123 * Pragma Export_Exception::
124 * Pragma Export_Function::
125 * Pragma Export_Object::
126 * Pragma Export_Procedure::
127 * Pragma Export_Value::
128 * Pragma Export_Valued_Procedure::
129 * Pragma Extend_System::
131 * Pragma External_Name_Casing::
132 * Pragma Finalize_Storage_Only::
133 * Pragma Float_Representation::
135 * Pragma Import_Exception::
136 * Pragma Import_Function::
137 * Pragma Import_Object::
138 * Pragma Import_Procedure::
139 * Pragma Import_Valued_Procedure::
140 * Pragma Initialize_Scalars::
141 * Pragma Inline_Always::
142 * Pragma Inline_Generic::
144 * Pragma Interface_Name::
145 * Pragma Interrupt_Handler::
146 * Pragma Interrupt_State::
147 * Pragma Keep_Names::
150 * Pragma Linker_Alias::
151 * Pragma Linker_Section::
152 * Pragma Long_Float::
153 * Pragma Machine_Attribute::
154 * Pragma Main_Storage::
156 * Pragma Normalize_Scalars::
157 * Pragma Obsolescent::
160 * Pragma Profile (Ravenscar)::
161 * Pragma Profile (Restricted)::
162 * Pragma Propagate_Exceptions::
163 * Pragma Psect_Object::
164 * Pragma Pure_Function::
165 * Pragma Restriction_Warnings::
166 * Pragma Source_File_Name::
167 * Pragma Source_File_Name_Project::
168 * Pragma Source_Reference::
169 * Pragma Stream_Convert::
170 * Pragma Style_Checks::
172 * Pragma Suppress_All::
173 * Pragma Suppress_Exception_Locations::
174 * Pragma Suppress_Initialization::
177 * Pragma Task_Storage::
178 * Pragma Thread_Body::
179 * Pragma Time_Slice::
181 * Pragma Unchecked_Union::
182 * Pragma Unimplemented_Unit::
183 * Pragma Universal_Data::
184 * Pragma Unreferenced::
185 * Pragma Unreserve_All_Interrupts::
186 * Pragma Unsuppress::
187 * Pragma Use_VADS_Size::
188 * Pragma Validity_Checks::
191 * Pragma Weak_External::
193 Implementation Defined Attributes
203 * Default_Bit_Order::
211 * Has_Access_Values::
212 * Has_Discriminants::
218 * Max_Interrupt_Priority::
220 * Maximum_Alignment::
224 * Passed_By_Reference::
235 * Unconstrained_Array::
236 * Universal_Literal_String::
237 * Unrestricted_Access::
243 The Implementation of Standard I/O
245 * Standard I/O Packages::
254 * Operations on C Streams::
255 * Interfacing to C Streams::
259 * Ada.Characters.Latin_9 (a-chlat9.ads)::
260 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
261 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
262 * Ada.Command_Line.Remove (a-colire.ads)::
263 * Ada.Command_Line.Environment (a-colien.ads)::
264 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
265 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
266 * Ada.Exceptions.Traceback (a-exctra.ads)::
267 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
268 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
269 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
270 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
271 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
272 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
273 * GNAT.Array_Split (g-arrspl.ads)::
274 * GNAT.AWK (g-awk.ads)::
275 * GNAT.Bounded_Buffers (g-boubuf.ads)::
276 * GNAT.Bounded_Mailboxes (g-boumai.ads)::
277 * GNAT.Bubble_Sort (g-bubsor.ads)::
278 * GNAT.Bubble_Sort_A (g-busora.ads)::
279 * GNAT.Bubble_Sort_G (g-busorg.ads)::
280 * GNAT.Calendar (g-calend.ads)::
281 * GNAT.Calendar.Time_IO (g-catiio.ads)::
282 * GNAT.Case_Util (g-casuti.ads)::
283 * GNAT.CGI (g-cgi.ads)::
284 * GNAT.CGI.Cookie (g-cgicoo.ads)::
285 * GNAT.CGI.Debug (g-cgideb.ads)::
286 * GNAT.Command_Line (g-comlin.ads)::
287 * GNAT.Compiler_Version (g-comver.ads)::
288 * GNAT.Ctrl_C (g-ctrl_c.ads)::
289 * GNAT.CRC32 (g-crc32.ads)::
290 * GNAT.Current_Exception (g-curexc.ads)::
291 * GNAT.Debug_Pools (g-debpoo.ads)::
292 * GNAT.Debug_Utilities (g-debuti.ads)::
293 * GNAT.Directory_Operations (g-dirope.ads)::
294 * GNAT.Dynamic_HTables (g-dynhta.ads)::
295 * GNAT.Dynamic_Tables (g-dyntab.ads)::
296 * GNAT.Exception_Actions (g-excact.ads)::
297 * GNAT.Exception_Traces (g-exctra.ads)::
298 * GNAT.Exceptions (g-except.ads)::
299 * GNAT.Expect (g-expect.ads)::
300 * GNAT.Float_Control (g-flocon.ads)::
301 * GNAT.Heap_Sort (g-heasor.ads)::
302 * GNAT.Heap_Sort_A (g-hesora.ads)::
303 * GNAT.Heap_Sort_G (g-hesorg.ads)::
304 * GNAT.HTable (g-htable.ads)::
305 * GNAT.IO (g-io.ads)::
306 * GNAT.IO_Aux (g-io_aux.ads)::
307 * GNAT.Lock_Files (g-locfil.ads)::
308 * GNAT.MD5 (g-md5.ads)::
309 * GNAT.Memory_Dump (g-memdum.ads)::
310 * GNAT.Most_Recent_Exception (g-moreex.ads)::
311 * GNAT.OS_Lib (g-os_lib.ads)::
312 * GNAT.Perfect_Hash_Generators (g-pehage.ads)::
313 * GNAT.Regexp (g-regexp.ads)::
314 * GNAT.Registry (g-regist.ads)::
315 * GNAT.Regpat (g-regpat.ads)::
316 * GNAT.Secondary_Stack_Info (g-sestin.ads)::
317 * GNAT.Semaphores (g-semaph.ads)::
318 * GNAT.Signals (g-signal.ads)::
319 * GNAT.Sockets (g-socket.ads)::
320 * GNAT.Source_Info (g-souinf.ads)::
321 * GNAT.Spell_Checker (g-speche.ads)::
322 * GNAT.Spitbol.Patterns (g-spipat.ads)::
323 * GNAT.Spitbol (g-spitbo.ads)::
324 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
325 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
326 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
327 * GNAT.Strings (g-string.ads)::
328 * GNAT.String_Split (g-strspl.ads)::
329 * GNAT.Table (g-table.ads)::
330 * GNAT.Task_Lock (g-tasloc.ads)::
331 * GNAT.Threads (g-thread.ads)::
332 * GNAT.Traceback (g-traceb.ads)::
333 * GNAT.Traceback.Symbolic (g-trasym.ads)::
334 * GNAT.Wide_String_Split (g-wistsp.ads)::
335 * Interfaces.C.Extensions (i-cexten.ads)::
336 * Interfaces.C.Streams (i-cstrea.ads)::
337 * Interfaces.CPP (i-cpp.ads)::
338 * Interfaces.Os2lib (i-os2lib.ads)::
339 * Interfaces.Os2lib.Errors (i-os2err.ads)::
340 * Interfaces.Os2lib.Synchronization (i-os2syn.ads)::
341 * Interfaces.Os2lib.Threads (i-os2thr.ads)::
342 * Interfaces.Packed_Decimal (i-pacdec.ads)::
343 * Interfaces.VxWorks (i-vxwork.ads)::
344 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
345 * System.Address_Image (s-addima.ads)::
346 * System.Assertions (s-assert.ads)::
347 * System.Memory (s-memory.ads)::
348 * System.Partition_Interface (s-parint.ads)::
349 * System.Restrictions (s-restri.ads)::
350 * System.Rident (s-rident.ads)::
351 * System.Task_Info (s-tasinf.ads)::
352 * System.Wch_Cnv (s-wchcnv.ads)::
353 * System.Wch_Con (s-wchcon.ads)::
357 * Text_IO Stream Pointer Positioning::
358 * Text_IO Reading and Writing Non-Regular Files::
360 * Treating Text_IO Files as Streams::
361 * Text_IO Extensions::
362 * Text_IO Facilities for Unbounded Strings::
366 * Wide_Text_IO Stream Pointer Positioning::
367 * Wide_Text_IO Reading and Writing Non-Regular Files::
369 Interfacing to Other Languages
372 * Interfacing to C++::
373 * Interfacing to COBOL::
374 * Interfacing to Fortran::
375 * Interfacing to non-GNAT Ada code::
377 Specialized Needs Annexes
379 Implementation of Specific Ada Features
380 * Machine Code Insertions::
381 * GNAT Implementation of Tasking::
382 * GNAT Implementation of Shared Passive Packages::
383 * Code Generation for Array Aggregates::
384 * The Size of Discriminated Records with Default Discriminants::
386 Project File Reference
390 GNU Free Documentation License
397 @node About This Guide
398 @unnumbered About This Guide
402 This manual contains useful information in writing programs using the
403 GNAT compiler. It includes information on implementation dependent
404 characteristics of GNAT, including all the information required by Annex
410 This manual contains useful information in writing programs using the
411 GNAT Pro compiler. It includes information on implementation dependent
412 characteristics of GNAT Pro, including all the information required by Annex
416 Ada 95 is designed to be highly portable.
417 In general, a program will have the same effect even when compiled by
418 different compilers on different platforms.
419 However, since Ada 95 is designed to be used in a
420 wide variety of applications, it also contains a number of system
421 dependent features to be used in interfacing to the external world.
422 @cindex Implementation-dependent features
425 Note: Any program that makes use of implementation-dependent features
426 may be non-portable. You should follow good programming practice and
427 isolate and clearly document any sections of your program that make use
428 of these features in a non-portable manner.
431 For ease of exposition, ``GNAT Pro'' will be referred to simply as
432 ``GNAT'' in the remainder of this document.
436 * What This Reference Manual Contains::
438 * Related Information::
441 @node What This Reference Manual Contains
442 @unnumberedsec What This Reference Manual Contains
445 This reference manual contains the following chapters:
449 @ref{Implementation Defined Pragmas}, lists GNAT implementation-dependent
450 pragmas, which can be used to extend and enhance the functionality of the
454 @ref{Implementation Defined Attributes}, lists GNAT
455 implementation-dependent attributes which can be used to extend and
456 enhance the functionality of the compiler.
459 @ref{Implementation Advice}, provides information on generally
460 desirable behavior which are not requirements that all compilers must
461 follow since it cannot be provided on all systems, or which may be
462 undesirable on some systems.
465 @ref{Implementation Defined Characteristics}, provides a guide to
466 minimizing implementation dependent features.
469 @ref{Intrinsic Subprograms}, describes the intrinsic subprograms
470 implemented by GNAT, and how they can be imported into user
471 application programs.
474 @ref{Representation Clauses and Pragmas}, describes in detail the
475 way that GNAT represents data, and in particular the exact set
476 of representation clauses and pragmas that is accepted.
479 @ref{Standard Library Routines}, provides a listing of packages and a
480 brief description of the functionality that is provided by Ada's
481 extensive set of standard library routines as implemented by GNAT@.
484 @ref{The Implementation of Standard I/O}, details how the GNAT
485 implementation of the input-output facilities.
488 @ref{The GNAT Library}, is a catalog of packages that complement
489 the Ada predefined library.
492 @ref{Interfacing to Other Languages}, describes how programs
493 written in Ada using GNAT can be interfaced to other programming
496 @ref{Specialized Needs Annexes}, describes the GNAT implementation of all
497 of the specialized needs annexes.
500 @ref{Implementation of Specific Ada Features}, discusses issues related
501 to GNAT's implementation of machine code insertions, tasking, and several
505 @ref{Project File Reference}, presents the syntax and semantics
509 @ref{Obsolescent Features} documents implementation dependent features,
510 including pragmas and attributes, which are considered obsolescent, since
511 there are other preferred ways of achieving the same results. These
512 obsolescent forms are retained for backwards compatibilty.
516 @cindex Ada 95 ISO/ANSI Standard
518 This reference manual assumes that you are familiar with Ada 95
519 language, as described in the International Standard
520 ANSI/ISO/IEC-8652:1995, Jan 1995.
523 @unnumberedsec Conventions
524 @cindex Conventions, typographical
525 @cindex Typographical conventions
528 Following are examples of the typographical and graphic conventions used
533 @code{Functions}, @code{utility program names}, @code{standard names},
540 @file{File Names}, @samp{button names}, and @samp{field names}.
549 [optional information or parameters]
552 Examples are described by text
554 and then shown this way.
559 Commands that are entered by the user are preceded in this manual by the
560 characters @samp{$ } (dollar sign followed by space). If your system uses this
561 sequence as a prompt, then the commands will appear exactly as you see them
562 in the manual. If your system uses some other prompt, then the command will
563 appear with the @samp{$} replaced by whatever prompt character you are using.
565 @node Related Information
566 @unnumberedsec Related Information
568 See the following documents for further information on GNAT:
572 @cite{GNAT User's Guide}, which provides information on how to use
573 the GNAT compiler system.
576 @cite{Ada 95 Reference Manual}, which contains all reference
577 material for the Ada 95 programming language.
580 @cite{Ada 95 Annotated Reference Manual}, which is an annotated version
581 of the standard reference manual cited above. The annotations describe
582 detailed aspects of the design decision, and in particular contain useful
583 sections on Ada 83 compatibility.
586 @cite{DEC Ada, Technical Overview and Comparison on DIGITAL Platforms},
587 which contains specific information on compatibility between GNAT and
591 @cite{DEC Ada, Language Reference Manual, part number AA-PYZAB-TK} which
592 describes in detail the pragmas and attributes provided by the DEC Ada 83
597 @node Implementation Defined Pragmas
598 @chapter Implementation Defined Pragmas
601 Ada 95 defines a set of pragmas that can be used to supply additional
602 information to the compiler. These language defined pragmas are
603 implemented in GNAT and work as described in the Ada 95 Reference
606 In addition, Ada 95 allows implementations to define additional pragmas
607 whose meaning is defined by the implementation. GNAT provides a number
608 of these implementation-dependent pragmas which can be used to extend
609 and enhance the functionality of the compiler. This section of the GNAT
610 Reference Manual describes these additional pragmas.
612 Note that any program using these pragmas may not be portable to other
613 compilers (although GNAT implements this set of pragmas on all
614 platforms). Therefore if portability to other compilers is an important
615 consideration, the use of these pragmas should be minimized.
618 * Pragma Abort_Defer::
624 * Pragma C_Pass_By_Copy::
626 * Pragma Common_Object::
627 * Pragma Compile_Time_Warning::
628 * Pragma Complex_Representation::
629 * Pragma Component_Alignment::
630 * Pragma Convention_Identifier::
632 * Pragma CPP_Constructor::
633 * Pragma CPP_Virtual::
634 * Pragma CPP_Vtable::
636 * Pragma Detect_Blocking::
637 * Pragma Elaboration_Checks::
639 * Pragma Export_Exception::
640 * Pragma Export_Function::
641 * Pragma Export_Object::
642 * Pragma Export_Procedure::
643 * Pragma Export_Value::
644 * Pragma Export_Valued_Procedure::
645 * Pragma Extend_System::
647 * Pragma External_Name_Casing::
648 * Pragma Finalize_Storage_Only::
649 * Pragma Float_Representation::
651 * Pragma Import_Exception::
652 * Pragma Import_Function::
653 * Pragma Import_Object::
654 * Pragma Import_Procedure::
655 * Pragma Import_Valued_Procedure::
656 * Pragma Initialize_Scalars::
657 * Pragma Inline_Always::
658 * Pragma Inline_Generic::
660 * Pragma Interface_Name::
661 * Pragma Interrupt_Handler::
662 * Pragma Interrupt_State::
663 * Pragma Keep_Names::
666 * Pragma Linker_Alias::
667 * Pragma Linker_Section::
668 * Pragma Long_Float::
669 * Pragma Machine_Attribute::
670 * Pragma Main_Storage::
672 * Pragma Normalize_Scalars::
673 * Pragma Obsolescent::
676 * Pragma Profile (Ravenscar)::
677 * Pragma Profile (Restricted)::
678 * Pragma Propagate_Exceptions::
679 * Pragma Psect_Object::
680 * Pragma Pure_Function::
681 * Pragma Restriction_Warnings::
682 * Pragma Source_File_Name::
683 * Pragma Source_File_Name_Project::
684 * Pragma Source_Reference::
685 * Pragma Stream_Convert::
686 * Pragma Style_Checks::
688 * Pragma Suppress_All::
689 * Pragma Suppress_Exception_Locations::
690 * Pragma Suppress_Initialization::
693 * Pragma Task_Storage::
694 * Pragma Thread_Body::
695 * Pragma Time_Slice::
697 * Pragma Unchecked_Union::
698 * Pragma Unimplemented_Unit::
699 * Pragma Universal_Data::
700 * Pragma Unreferenced::
701 * Pragma Unreserve_All_Interrupts::
702 * Pragma Unsuppress::
703 * Pragma Use_VADS_Size::
704 * Pragma Validity_Checks::
707 * Pragma Weak_External::
710 @node Pragma Abort_Defer
711 @unnumberedsec Pragma Abort_Defer
713 @cindex Deferring aborts
721 This pragma must appear at the start of the statement sequence of a
722 handled sequence of statements (right after the @code{begin}). It has
723 the effect of deferring aborts for the sequence of statements (but not
724 for the declarations or handlers, if any, associated with this statement
728 @unnumberedsec Pragma Ada_83
737 A configuration pragma that establishes Ada 83 mode for the unit to
738 which it applies, regardless of the mode set by the command line
739 switches. In Ada 83 mode, GNAT attempts to be as compatible with
740 the syntax and semantics of Ada 83, as defined in the original Ada
741 83 Reference Manual as possible. In particular, the new Ada 95
742 keywords are not recognized, optional package bodies are allowed,
743 and generics may name types with unknown discriminants without using
744 the @code{(<>)} notation. In addition, some but not all of the additional
745 restrictions of Ada 83 are enforced.
747 Ada 83 mode is intended for two purposes. Firstly, it allows existing
748 legacy Ada 83 code to be compiled and adapted to GNAT with less effort.
749 Secondly, it aids in keeping code backwards compatible with Ada 83.
750 However, there is no guarantee that code that is processed correctly
751 by GNAT in Ada 83 mode will in fact compile and execute with an Ada
752 83 compiler, since GNAT does not enforce all the additional checks
756 @unnumberedsec Pragma Ada_95
765 A configuration pragma that establishes Ada 95 mode for the unit to which
766 it applies, regardless of the mode set by the command line switches.
767 This mode is set automatically for the @code{Ada} and @code{System}
768 packages and their children, so you need not specify it in these
769 contexts. This pragma is useful when writing a reusable component that
770 itself uses Ada 95 features, but which is intended to be usable from
771 either Ada 83 or Ada 95 programs.
773 @node Pragma Annotate
774 @unnumberedsec Pragma Annotate
779 pragma Annotate (IDENTIFIER @{, ARG@});
781 ARG ::= NAME | EXPRESSION
785 This pragma is used to annotate programs. @var{identifier} identifies
786 the type of annotation. GNAT verifies this is an identifier, but does
787 not otherwise analyze it. The @var{arg} argument
788 can be either a string literal or an
789 expression. String literals are assumed to be of type
790 @code{Standard.String}. Names of entities are simply analyzed as entity
791 names. All other expressions are analyzed as expressions, and must be
794 The analyzed pragma is retained in the tree, but not otherwise processed
795 by any part of the GNAT compiler. This pragma is intended for use by
796 external tools, including ASIS@.
799 @unnumberedsec Pragma Assert
806 [, static_string_EXPRESSION]);
810 The effect of this pragma depends on whether the corresponding command
811 line switch is set to activate assertions. The pragma expands into code
812 equivalent to the following:
815 if assertions-enabled then
816 if not boolean_EXPRESSION then
817 System.Assertions.Raise_Assert_Failure
824 The string argument, if given, is the message that will be associated
825 with the exception occurrence if the exception is raised. If no second
826 argument is given, the default message is @samp{@var{file}:@var{nnn}},
827 where @var{file} is the name of the source file containing the assert,
828 and @var{nnn} is the line number of the assert. A pragma is not a
829 statement, so if a statement sequence contains nothing but a pragma
830 assert, then a null statement is required in addition, as in:
835 pragma Assert (K > 3, "Bad value for K");
841 Note that, as with the @code{if} statement to which it is equivalent, the
842 type of the expression is either @code{Standard.Boolean}, or any type derived
843 from this standard type.
845 If assertions are disabled (switch @code{-gnata} not used), then there
846 is no effect (and in particular, any side effects from the expression
847 are suppressed). More precisely it is not quite true that the pragma
848 has no effect, since the expression is analyzed, and may cause types
849 to be frozen if they are mentioned here for the first time.
851 If assertions are enabled, then the given expression is tested, and if
852 it is @code{False} then @code{System.Assertions.Raise_Assert_Failure} is called
853 which results in the raising of @code{Assert_Failure} with the given message.
855 If the boolean expression has side effects, these side effects will turn
856 on and off with the setting of the assertions mode, resulting in
857 assertions that have an effect on the program. You should generally
858 avoid side effects in the expression arguments of this pragma. However,
859 the expressions are analyzed for semantic correctness whether or not
860 assertions are enabled, so turning assertions on and off cannot affect
861 the legality of a program.
863 @node Pragma Ast_Entry
864 @unnumberedsec Pragma Ast_Entry
870 pragma AST_Entry (entry_IDENTIFIER);
874 This pragma is implemented only in the OpenVMS implementation of GNAT@. The
875 argument is the simple name of a single entry; at most one @code{AST_Entry}
876 pragma is allowed for any given entry. This pragma must be used in
877 conjunction with the @code{AST_Entry} attribute, and is only allowed after
878 the entry declaration and in the same task type specification or single task
879 as the entry to which it applies. This pragma specifies that the given entry
880 may be used to handle an OpenVMS asynchronous system trap (@code{AST})
881 resulting from an OpenVMS system service call. The pragma does not affect
882 normal use of the entry. For further details on this pragma, see the
883 DEC Ada Language Reference Manual, section 9.12a.
885 @node Pragma C_Pass_By_Copy
886 @unnumberedsec Pragma C_Pass_By_Copy
887 @cindex Passing by copy
888 @findex C_Pass_By_Copy
892 pragma C_Pass_By_Copy
893 ([Max_Size =>] static_integer_EXPRESSION);
897 Normally the default mechanism for passing C convention records to C
898 convention subprograms is to pass them by reference, as suggested by RM
899 B.3(69). Use the configuration pragma @code{C_Pass_By_Copy} to change
900 this default, by requiring that record formal parameters be passed by
901 copy if all of the following conditions are met:
905 The size of the record type does not exceed@*@var{static_integer_expression}.
907 The record type has @code{Convention C}.
909 The formal parameter has this record type, and the subprogram has a
910 foreign (non-Ada) convention.
914 If these conditions are met the argument is passed by copy, i.e.@: in a
915 manner consistent with what C expects if the corresponding formal in the
916 C prototype is a struct (rather than a pointer to a struct).
918 You can also pass records by copy by specifying the convention
919 @code{C_Pass_By_Copy} for the record type, or by using the extended
920 @code{Import} and @code{Export} pragmas, which allow specification of
921 passing mechanisms on a parameter by parameter basis.
924 @unnumberedsec Pragma Comment
930 pragma Comment (static_string_EXPRESSION);
934 This is almost identical in effect to pragma @code{Ident}. It allows the
935 placement of a comment into the object file and hence into the
936 executable file if the operating system permits such usage. The
937 difference is that @code{Comment}, unlike @code{Ident}, has
938 no limitations on placement of the pragma (it can be placed
939 anywhere in the main source unit), and if more than one pragma
940 is used, all comments are retained.
942 @node Pragma Common_Object
943 @unnumberedsec Pragma Common_Object
944 @findex Common_Object
949 pragma Common_Object (
950 [Internal =>] LOCAL_NAME,
951 [, [External =>] EXTERNAL_SYMBOL]
952 [, [Size =>] EXTERNAL_SYMBOL] );
956 | static_string_EXPRESSION
960 This pragma enables the shared use of variables stored in overlaid
961 linker areas corresponding to the use of @code{COMMON}
962 in Fortran. The single
963 object @var{local_name} is assigned to the area designated by
964 the @var{External} argument.
965 You may define a record to correspond to a series
966 of fields. The @var{size} argument
967 is syntax checked in GNAT, but otherwise ignored.
969 @code{Common_Object} is not supported on all platforms. If no
970 support is available, then the code generator will issue a message
971 indicating that the necessary attribute for implementation of this
972 pragma is not available.
974 @node Pragma Compile_Time_Warning
975 @unnumberedsec Pragma Compile_Time_Warning
976 @findex Compile_Time_Warning
981 pragma Compile_Time_Warning
982 (boolean_EXPRESSION, static_string_EXPRESSION);
986 This pragma can be used to generate additional compile time warnings. It
987 is particularly useful in generics, where warnings can be issued for
988 specific problematic instantiations. The first parameter is a boolean
989 expression. The pragma is effective only if the value of this expression
990 is known at compile time, and has the value True. The set of expressions
991 whose values are known at compile time includes all static boolean
992 expressions, and also other values which the compiler can determine
993 at compile time (e.g. the size of a record type set by an explicit
994 size representation clause, or the value of a variable which was
995 initialized to a constant and is known not to have been modified).
996 If these conditions are met, a warning message is generated using
997 the value given as the second argument. This string value may contain
998 embedded ASCII.LF characters to break the message into multiple lines.
1000 @node Pragma Complex_Representation
1001 @unnumberedsec Pragma Complex_Representation
1002 @findex Complex_Representation
1006 @smallexample @c ada
1007 pragma Complex_Representation
1008 ([Entity =>] LOCAL_NAME);
1012 The @var{Entity} argument must be the name of a record type which has
1013 two fields of the same floating-point type. The effect of this pragma is
1014 to force gcc to use the special internal complex representation form for
1015 this record, which may be more efficient. Note that this may result in
1016 the code for this type not conforming to standard ABI (application
1017 binary interface) requirements for the handling of record types. For
1018 example, in some environments, there is a requirement for passing
1019 records by pointer, and the use of this pragma may result in passing
1020 this type in floating-point registers.
1022 @node Pragma Component_Alignment
1023 @unnumberedsec Pragma Component_Alignment
1024 @cindex Alignments of components
1025 @findex Component_Alignment
1029 @smallexample @c ada
1030 pragma Component_Alignment (
1031 [Form =>] ALIGNMENT_CHOICE
1032 [, [Name =>] type_LOCAL_NAME]);
1034 ALIGNMENT_CHOICE ::=
1042 Specifies the alignment of components in array or record types.
1043 The meaning of the @var{Form} argument is as follows:
1046 @findex Component_Size
1047 @item Component_Size
1048 Aligns scalar components and subcomponents of the array or record type
1049 on boundaries appropriate to their inherent size (naturally
1050 aligned). For example, 1-byte components are aligned on byte boundaries,
1051 2-byte integer components are aligned on 2-byte boundaries, 4-byte
1052 integer components are aligned on 4-byte boundaries and so on. These
1053 alignment rules correspond to the normal rules for C compilers on all
1054 machines except the VAX@.
1056 @findex Component_Size_4
1057 @item Component_Size_4
1058 Naturally aligns components with a size of four or fewer
1059 bytes. Components that are larger than 4 bytes are placed on the next
1062 @findex Storage_Unit
1064 Specifies that array or record components are byte aligned, i.e.@:
1065 aligned on boundaries determined by the value of the constant
1066 @code{System.Storage_Unit}.
1070 Specifies that array or record components are aligned on default
1071 boundaries, appropriate to the underlying hardware or operating system or
1072 both. For OpenVMS VAX systems, the @code{Default} choice is the same as
1073 the @code{Storage_Unit} choice (byte alignment). For all other systems,
1074 the @code{Default} choice is the same as @code{Component_Size} (natural
1079 If the @code{Name} parameter is present, @var{type_local_name} must
1080 refer to a local record or array type, and the specified alignment
1081 choice applies to the specified type. The use of
1082 @code{Component_Alignment} together with a pragma @code{Pack} causes the
1083 @code{Component_Alignment} pragma to be ignored. The use of
1084 @code{Component_Alignment} together with a record representation clause
1085 is only effective for fields not specified by the representation clause.
1087 If the @code{Name} parameter is absent, the pragma can be used as either
1088 a configuration pragma, in which case it applies to one or more units in
1089 accordance with the normal rules for configuration pragmas, or it can be
1090 used within a declarative part, in which case it applies to types that
1091 are declared within this declarative part, or within any nested scope
1092 within this declarative part. In either case it specifies the alignment
1093 to be applied to any record or array type which has otherwise standard
1096 If the alignment for a record or array type is not specified (using
1097 pragma @code{Pack}, pragma @code{Component_Alignment}, or a record rep
1098 clause), the GNAT uses the default alignment as described previously.
1100 @node Pragma Convention_Identifier
1101 @unnumberedsec Pragma Convention_Identifier
1102 @findex Convention_Identifier
1103 @cindex Conventions, synonyms
1107 @smallexample @c ada
1108 pragma Convention_Identifier (
1109 [Name =>] IDENTIFIER,
1110 [Convention =>] convention_IDENTIFIER);
1114 This pragma provides a mechanism for supplying synonyms for existing
1115 convention identifiers. The @code{Name} identifier can subsequently
1116 be used as a synonym for the given convention in other pragmas (including
1117 for example pragma @code{Import} or another @code{Convention_Identifier}
1118 pragma). As an example of the use of this, suppose you had legacy code
1119 which used Fortran77 as the identifier for Fortran. Then the pragma:
1121 @smallexample @c ada
1122 pragma Convention_Identifier (Fortran77, Fortran);
1126 would allow the use of the convention identifier @code{Fortran77} in
1127 subsequent code, avoiding the need to modify the sources. As another
1128 example, you could use this to parametrize convention requirements
1129 according to systems. Suppose you needed to use @code{Stdcall} on
1130 windows systems, and @code{C} on some other system, then you could
1131 define a convention identifier @code{Library} and use a single
1132 @code{Convention_Identifier} pragma to specify which convention
1133 would be used system-wide.
1135 @node Pragma CPP_Class
1136 @unnumberedsec Pragma CPP_Class
1138 @cindex Interfacing with C++
1142 @smallexample @c ada
1143 pragma CPP_Class ([Entity =>] LOCAL_NAME);
1147 The argument denotes an entity in the current declarative region
1148 that is declared as a tagged or untagged record type. It indicates that
1149 the type corresponds to an externally declared C++ class type, and is to
1150 be laid out the same way that C++ would lay out the type.
1152 If (and only if) the type is tagged, at least one component in the
1153 record must be of type @code{Interfaces.CPP.Vtable_Ptr}, corresponding
1154 to the C++ Vtable (or Vtables in the case of multiple inheritance) used
1157 Types for which @code{CPP_Class} is specified do not have assignment or
1158 equality operators defined (such operations can be imported or declared
1159 as subprograms as required). Initialization is allowed only by
1160 constructor functions (see pragma @code{CPP_Constructor}).
1162 Pragma @code{CPP_Class} is intended primarily for automatic generation
1163 using an automatic binding generator tool.
1164 See @ref{Interfacing to C++} for related information.
1166 @node Pragma CPP_Constructor
1167 @unnumberedsec Pragma CPP_Constructor
1168 @cindex Interfacing with C++
1169 @findex CPP_Constructor
1173 @smallexample @c ada
1174 pragma CPP_Constructor ([Entity =>] LOCAL_NAME);
1178 This pragma identifies an imported function (imported in the usual way
1179 with pragma @code{Import}) as corresponding to a C++
1180 constructor. The argument is a name that must have been
1181 previously mentioned in a pragma @code{Import}
1182 with @code{Convention} = @code{CPP}, and must be of one of the following
1187 @code{function @var{Fname} return @var{T}'Class}
1190 @code{function @var{Fname} (@dots{}) return @var{T}'Class}
1194 where @var{T} is a tagged type to which the pragma @code{CPP_Class} applies.
1196 The first form is the default constructor, used when an object of type
1197 @var{T} is created on the Ada side with no explicit constructor. Other
1198 constructors (including the copy constructor, which is simply a special
1199 case of the second form in which the one and only argument is of type
1200 @var{T}), can only appear in two contexts:
1204 On the right side of an initialization of an object of type @var{T}.
1206 In an extension aggregate for an object of a type derived from @var{T}.
1210 Although the constructor is described as a function that returns a value
1211 on the Ada side, it is typically a procedure with an extra implicit
1212 argument (the object being initialized) at the implementation
1213 level. GNAT issues the appropriate call, whatever it is, to get the
1214 object properly initialized.
1216 In the case of derived objects, you may use one of two possible forms
1217 for declaring and creating an object:
1220 @item @code{New_Object : Derived_T}
1221 @item @code{New_Object : Derived_T := (@var{constructor-call with} @dots{})}
1225 In the first case the default constructor is called and extension fields
1226 if any are initialized according to the default initialization
1227 expressions in the Ada declaration. In the second case, the given
1228 constructor is called and the extension aggregate indicates the explicit
1229 values of the extension fields.
1231 If no constructors are imported, it is impossible to create any objects
1232 on the Ada side. If no default constructor is imported, only the
1233 initialization forms using an explicit call to a constructor are
1236 Pragma @code{CPP_Constructor} is intended primarily for automatic generation
1237 using an automatic binding generator tool.
1238 See @ref{Interfacing to C++} for more related information.
1240 @node Pragma CPP_Virtual
1241 @unnumberedsec Pragma CPP_Virtual
1242 @cindex Interfacing to C++
1247 @smallexample @c ada
1250 [, [Vtable_Ptr =>] vtable_ENTITY,]
1251 [, [Position =>] static_integer_EXPRESSION]);
1255 This pragma serves the same function as pragma @code{Import} in that
1256 case of a virtual function imported from C++. The @var{Entity} argument
1258 primitive subprogram of a tagged type to which pragma @code{CPP_Class}
1259 applies. The @var{Vtable_Ptr} argument specifies
1260 the Vtable_Ptr component which contains the
1261 entry for this virtual function. The @var{Position} argument
1262 is the sequential number
1263 counting virtual functions for this Vtable starting at 1.
1265 The @code{Vtable_Ptr} and @code{Position} arguments may be omitted if
1266 there is one Vtable_Ptr present (single inheritance case) and all
1267 virtual functions are imported. In that case the compiler can deduce both
1270 No @code{External_Name} or @code{Link_Name} arguments are required for a
1271 virtual function, since it is always accessed indirectly via the
1272 appropriate Vtable entry.
1274 Pragma @code{CPP_Virtual} is intended primarily for automatic generation
1275 using an automatic binding generator tool.
1276 See @ref{Interfacing to C++} for related information.
1278 @node Pragma CPP_Vtable
1279 @unnumberedsec Pragma CPP_Vtable
1280 @cindex Interfacing with C++
1285 @smallexample @c ada
1288 [Vtable_Ptr =>] vtable_ENTITY,
1289 [Entry_Count =>] static_integer_EXPRESSION);
1293 Given a record to which the pragma @code{CPP_Class} applies,
1294 this pragma can be specified for each component of type
1295 @code{CPP.Interfaces.Vtable_Ptr}.
1296 @var{Entity} is the tagged type, @var{Vtable_Ptr}
1297 is the record field of type @code{Vtable_Ptr}, and @var{Entry_Count} is
1298 the number of virtual functions on the C++ side. Not all of these
1299 functions need to be imported on the Ada side.
1301 You may omit the @code{CPP_Vtable} pragma if there is only one
1302 @code{Vtable_Ptr} component in the record and all virtual functions are
1303 imported on the Ada side (the default value for the entry count in this
1304 case is simply the total number of virtual functions).
1306 Pragma @code{CPP_Vtable} is intended primarily for automatic generation
1307 using an automatic binding generator tool.
1308 See @ref{Interfacing to C++} for related information.
1311 @unnumberedsec Pragma Debug
1316 @smallexample @c ada
1317 pragma Debug (PROCEDURE_CALL_WITHOUT_SEMICOLON);
1319 PROCEDURE_CALL_WITHOUT_SEMICOLON ::=
1321 | PROCEDURE_PREFIX ACTUAL_PARAMETER_PART
1325 The argument has the syntactic form of an expression, meeting the
1326 syntactic requirements for pragmas.
1328 If assertions are not enabled on the command line, this pragma has no
1329 effect. If asserts are enabled, the semantics of the pragma is exactly
1330 equivalent to the procedure call statement corresponding to the argument
1331 with a terminating semicolon. Pragmas are permitted in sequences of
1332 declarations, so you can use pragma @code{Debug} to intersperse calls to
1333 debug procedures in the middle of declarations.
1335 @node Pragma Detect_Blocking
1336 @unnumberedsec Pragma Detect_Blocking
1337 @findex Detect_Blocking
1341 @smallexample @c ada
1342 pragma Detect_Blocking;
1346 This is a configuration pragma that forces the detection of potentially
1347 blocking operations within a protected operation, and to raise Program_Error
1350 @node Pragma Elaboration_Checks
1351 @unnumberedsec Pragma Elaboration_Checks
1352 @cindex Elaboration control
1353 @findex Elaboration_Checks
1357 @smallexample @c ada
1358 pragma Elaboration_Checks (Dynamic | Static);
1362 This is a configuration pragma that provides control over the
1363 elaboration model used by the compilation affected by the
1364 pragma. If the parameter is @code{Dynamic},
1365 then the dynamic elaboration
1366 model described in the Ada Reference Manual is used, as though
1367 the @code{-gnatE} switch had been specified on the command
1368 line. If the parameter is @code{Static}, then the default GNAT static
1369 model is used. This configuration pragma overrides the setting
1370 of the command line. For full details on the elaboration models
1371 used by the GNAT compiler, see section ``Elaboration Order
1372 Handling in GNAT'' in the @cite{GNAT User's Guide}.
1374 @node Pragma Eliminate
1375 @unnumberedsec Pragma Eliminate
1376 @cindex Elimination of unused subprograms
1381 @smallexample @c ada
1383 [Unit_Name =>] IDENTIFIER |
1384 SELECTED_COMPONENT);
1387 [Unit_Name =>] IDENTIFIER |
1389 [Entity =>] IDENTIFIER |
1390 SELECTED_COMPONENT |
1392 [,OVERLOADING_RESOLUTION]);
1394 OVERLOADING_RESOLUTION ::= PARAMETER_AND_RESULT_TYPE_PROFILE |
1397 PARAMETER_AND_RESULT_TYPE_PROFILE ::= PROCEDURE_PROFILE |
1400 PROCEDURE_PROFILE ::= Parameter_Types => PARAMETER_TYPES
1402 FUNCTION_PROFILE ::= [Parameter_Types => PARAMETER_TYPES,]
1403 Result_Type => result_SUBTYPE_NAME]
1405 PARAMETER_TYPES ::= (SUBTYPE_NAME @{, SUBTYPE_NAME@})
1406 SUBTYPE_NAME ::= STRING_VALUE
1408 SOURCE_LOCATION ::= Source_Location => SOURCE_TRACE
1409 SOURCE_TRACE ::= STRING_VALUE
1411 STRING_VALUE ::= STRING_LITERAL @{& STRING_LITERAL@}
1415 This pragma indicates that the given entity is not used outside the
1416 compilation unit it is defined in. The entity must be an explicitly declared
1417 subprogram; this includes generic subprogram instances and
1418 subprograms declared in generic package instances.
1420 If the entity to be eliminated is a library level subprogram, then
1421 the first form of pragma @code{Eliminate} is used with only a single argument.
1422 In this form, the @code{Unit_Name} argument specifies the name of the
1423 library level unit to be eliminated.
1425 In all other cases, both @code{Unit_Name} and @code{Entity} arguments
1426 are required. If item is an entity of a library package, then the first
1427 argument specifies the unit name, and the second argument specifies
1428 the particular entity. If the second argument is in string form, it must
1429 correspond to the internal manner in which GNAT stores entity names (see
1430 compilation unit Namet in the compiler sources for details).
1432 The remaining parameters (OVERLOADING_RESOLUTION) are optionally used
1433 to distinguish between overloaded subprograms. If a pragma does not contain
1434 the OVERLOADING_RESOLUTION parameter(s), it is applied to all the overloaded
1435 subprograms denoted by the first two parameters.
1437 Use PARAMETER_AND_RESULT_TYPE_PROFILE to specify the profile of the subprogram
1438 to be eliminated in a manner similar to that used for the extended
1439 @code{Import} and @code{Export} pragmas, except that the subtype names are
1440 always given as strings. At the moment, this form of distinguishing
1441 overloaded subprograms is implemented only partially, so we do not recommend
1442 using it for practical subprogram elimination.
1444 Note, that in case of a parameterless procedure its profile is represented
1445 as @code{Parameter_Types => ("")}
1447 Alternatively, the @code{Source_Location} parameter is used to specify
1448 which overloaded alternative is to be eliminated by pointing to the
1449 location of the DEFINING_PROGRAM_UNIT_NAME of this subprogram in the
1450 source text. The string literal (or concatenation of string literals)
1451 given as SOURCE_TRACE must have the following format:
1453 @smallexample @c ada
1454 SOURCE_TRACE ::= SOURCE_LOCATION@{LBRACKET SOURCE_LOCATION RBRACKET@}
1459 SOURCE_LOCATION ::= FILE_NAME:LINE_NUMBER
1460 FILE_NAME ::= STRING_LITERAL
1461 LINE_NUMBER ::= DIGIT @{DIGIT@}
1464 SOURCE_TRACE should be the short name of the source file (with no directory
1465 information), and LINE_NUMBER is supposed to point to the line where the
1466 defining name of the subprogram is located.
1468 For the subprograms that are not a part of generic instantiations, only one
1469 SOURCE_LOCATION is used. If a subprogram is declared in a package
1470 instantiation, SOURCE_TRACE contains two SOURCE_LOCATIONs, the first one is
1471 the location of the (DEFINING_PROGRAM_UNIT_NAME of the) instantiation, and the
1472 second one denotes the declaration of the corresponding subprogram in the
1473 generic package. This approach is recursively used to create SOURCE_LOCATIONs
1474 in case of nested instantiations.
1476 The effect of the pragma is to allow the compiler to eliminate
1477 the code or data associated with the named entity. Any reference to
1478 an eliminated entity outside the compilation unit it is defined in,
1479 causes a compile time or link time error.
1481 The intention of pragma @code{Eliminate} is to allow a program to be compiled
1482 in a system independent manner, with unused entities eliminated, without
1483 the requirement of modifying the source text. Normally the required set
1484 of @code{Eliminate} pragmas is constructed automatically using the gnatelim
1485 tool. Elimination of unused entities local to a compilation unit is
1486 automatic, without requiring the use of pragma @code{Eliminate}.
1488 Note that the reason this pragma takes string literals where names might
1489 be expected is that a pragma @code{Eliminate} can appear in a context where the
1490 relevant names are not visible.
1492 Note that any change in the source files that includes removing, splitting of
1493 adding lines may make the set of Eliminate pragmas using SOURCE_LOCATION
1496 @node Pragma Export_Exception
1497 @unnumberedsec Pragma Export_Exception
1499 @findex Export_Exception
1503 @smallexample @c ada
1504 pragma Export_Exception (
1505 [Internal =>] LOCAL_NAME,
1506 [, [External =>] EXTERNAL_SYMBOL,]
1507 [, [Form =>] Ada | VMS]
1508 [, [Code =>] static_integer_EXPRESSION]);
1512 | static_string_EXPRESSION
1516 This pragma is implemented only in the OpenVMS implementation of GNAT@. It
1517 causes the specified exception to be propagated outside of the Ada program,
1518 so that it can be handled by programs written in other OpenVMS languages.
1519 This pragma establishes an external name for an Ada exception and makes the
1520 name available to the OpenVMS Linker as a global symbol. For further details
1521 on this pragma, see the
1522 DEC Ada Language Reference Manual, section 13.9a3.2.
1524 @node Pragma Export_Function
1525 @unnumberedsec Pragma Export_Function
1526 @cindex Argument passing mechanisms
1527 @findex Export_Function
1532 @smallexample @c ada
1533 pragma Export_Function (
1534 [Internal =>] LOCAL_NAME,
1535 [, [External =>] EXTERNAL_SYMBOL]
1536 [, [Parameter_Types =>] PARAMETER_TYPES]
1537 [, [Result_Type =>] result_SUBTYPE_MARK]
1538 [, [Mechanism =>] MECHANISM]
1539 [, [Result_Mechanism =>] MECHANISM_NAME]);
1543 | static_string_EXPRESSION
1548 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
1552 | subtype_Name ' Access
1556 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1558 MECHANISM_ASSOCIATION ::=
1559 [formal_parameter_NAME =>] MECHANISM_NAME
1567 Use this pragma to make a function externally callable and optionally
1568 provide information on mechanisms to be used for passing parameter and
1569 result values. We recommend, for the purposes of improving portability,
1570 this pragma always be used in conjunction with a separate pragma
1571 @code{Export}, which must precede the pragma @code{Export_Function}.
1572 GNAT does not require a separate pragma @code{Export}, but if none is
1573 present, @code{Convention Ada} is assumed, which is usually
1574 not what is wanted, so it is usually appropriate to use this
1575 pragma in conjunction with a @code{Export} or @code{Convention}
1576 pragma that specifies the desired foreign convention.
1577 Pragma @code{Export_Function}
1578 (and @code{Export}, if present) must appear in the same declarative
1579 region as the function to which they apply.
1581 @var{internal_name} must uniquely designate the function to which the
1582 pragma applies. If more than one function name exists of this name in
1583 the declarative part you must use the @code{Parameter_Types} and
1584 @code{Result_Type} parameters is mandatory to achieve the required
1585 unique designation. @var{subtype_ mark}s in these parameters must
1586 exactly match the subtypes in the corresponding function specification,
1587 using positional notation to match parameters with subtype marks.
1588 The form with an @code{'Access} attribute can be used to match an
1589 anonymous access parameter.
1592 @cindex Passing by descriptor
1593 Note that passing by descriptor is not supported, even on the OpenVMS
1596 @cindex Suppressing external name
1597 Special treatment is given if the EXTERNAL is an explicit null
1598 string or a static string expressions that evaluates to the null
1599 string. In this case, no external name is generated. This form
1600 still allows the specification of parameter mechanisms.
1602 @node Pragma Export_Object
1603 @unnumberedsec Pragma Export_Object
1604 @findex Export_Object
1608 @smallexample @c ada
1609 pragma Export_Object
1610 [Internal =>] LOCAL_NAME,
1611 [, [External =>] EXTERNAL_SYMBOL]
1612 [, [Size =>] EXTERNAL_SYMBOL]
1616 | static_string_EXPRESSION
1620 This pragma designates an object as exported, and apart from the
1621 extended rules for external symbols, is identical in effect to the use of
1622 the normal @code{Export} pragma applied to an object. You may use a
1623 separate Export pragma (and you probably should from the point of view
1624 of portability), but it is not required. @var{Size} is syntax checked,
1625 but otherwise ignored by GNAT@.
1627 @node Pragma Export_Procedure
1628 @unnumberedsec Pragma Export_Procedure
1629 @findex Export_Procedure
1633 @smallexample @c ada
1634 pragma Export_Procedure (
1635 [Internal =>] LOCAL_NAME
1636 [, [External =>] EXTERNAL_SYMBOL]
1637 [, [Parameter_Types =>] PARAMETER_TYPES]
1638 [, [Mechanism =>] MECHANISM]);
1642 | static_string_EXPRESSION
1647 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
1651 | subtype_Name ' Access
1655 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1657 MECHANISM_ASSOCIATION ::=
1658 [formal_parameter_NAME =>] MECHANISM_NAME
1666 This pragma is identical to @code{Export_Function} except that it
1667 applies to a procedure rather than a function and the parameters
1668 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
1669 GNAT does not require a separate pragma @code{Export}, but if none is
1670 present, @code{Convention Ada} is assumed, which is usually
1671 not what is wanted, so it is usually appropriate to use this
1672 pragma in conjunction with a @code{Export} or @code{Convention}
1673 pragma that specifies the desired foreign convention.
1676 @cindex Passing by descriptor
1677 Note that passing by descriptor is not supported, even on the OpenVMS
1680 @cindex Suppressing external name
1681 Special treatment is given if the EXTERNAL is an explicit null
1682 string or a static string expressions that evaluates to the null
1683 string. In this case, no external name is generated. This form
1684 still allows the specification of parameter mechanisms.
1686 @node Pragma Export_Value
1687 @unnumberedsec Pragma Export_Value
1688 @findex Export_Value
1692 @smallexample @c ada
1693 pragma Export_Value (
1694 [Value =>] static_integer_EXPRESSION,
1695 [Link_Name =>] static_string_EXPRESSION);
1699 This pragma serves to export a static integer value for external use.
1700 The first argument specifies the value to be exported. The Link_Name
1701 argument specifies the symbolic name to be associated with the integer
1702 value. This pragma is useful for defining a named static value in Ada
1703 that can be referenced in assembly language units to be linked with
1704 the application. This pragma is currently supported only for the
1705 AAMP target and is ignored for other targets.
1707 @node Pragma Export_Valued_Procedure
1708 @unnumberedsec Pragma Export_Valued_Procedure
1709 @findex Export_Valued_Procedure
1713 @smallexample @c ada
1714 pragma Export_Valued_Procedure (
1715 [Internal =>] LOCAL_NAME
1716 [, [External =>] EXTERNAL_SYMBOL]
1717 [, [Parameter_Types =>] PARAMETER_TYPES]
1718 [, [Mechanism =>] MECHANISM]);
1722 | static_string_EXPRESSION
1727 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
1731 | subtype_Name ' Access
1735 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1737 MECHANISM_ASSOCIATION ::=
1738 [formal_parameter_NAME =>] MECHANISM_NAME
1746 This pragma is identical to @code{Export_Procedure} except that the
1747 first parameter of @var{local_name}, which must be present, must be of
1748 mode @code{OUT}, and externally the subprogram is treated as a function
1749 with this parameter as the result of the function. GNAT provides for
1750 this capability to allow the use of @code{OUT} and @code{IN OUT}
1751 parameters in interfacing to external functions (which are not permitted
1753 GNAT does not require a separate pragma @code{Export}, but if none is
1754 present, @code{Convention Ada} is assumed, which is almost certainly
1755 not what is wanted since the whole point of this pragma is to interface
1756 with foreign language functions, so it is usually appropriate to use this
1757 pragma in conjunction with a @code{Export} or @code{Convention}
1758 pragma that specifies the desired foreign convention.
1761 @cindex Passing by descriptor
1762 Note that passing by descriptor is not supported, even on the OpenVMS
1765 @cindex Suppressing external name
1766 Special treatment is given if the EXTERNAL is an explicit null
1767 string or a static string expressions that evaluates to the null
1768 string. In this case, no external name is generated. This form
1769 still allows the specification of parameter mechanisms.
1771 @node Pragma Extend_System
1772 @unnumberedsec Pragma Extend_System
1773 @cindex @code{system}, extending
1775 @findex Extend_System
1779 @smallexample @c ada
1780 pragma Extend_System ([Name =>] IDENTIFIER);
1784 This pragma is used to provide backwards compatibility with other
1785 implementations that extend the facilities of package @code{System}. In
1786 GNAT, @code{System} contains only the definitions that are present in
1787 the Ada 95 RM@. However, other implementations, notably the DEC Ada 83
1788 implementation, provide many extensions to package @code{System}.
1790 For each such implementation accommodated by this pragma, GNAT provides a
1791 package @code{Aux_@var{xxx}}, e.g.@: @code{Aux_DEC} for the DEC Ada 83
1792 implementation, which provides the required additional definitions. You
1793 can use this package in two ways. You can @code{with} it in the normal
1794 way and access entities either by selection or using a @code{use}
1795 clause. In this case no special processing is required.
1797 However, if existing code contains references such as
1798 @code{System.@var{xxx}} where @var{xxx} is an entity in the extended
1799 definitions provided in package @code{System}, you may use this pragma
1800 to extend visibility in @code{System} in a non-standard way that
1801 provides greater compatibility with the existing code. Pragma
1802 @code{Extend_System} is a configuration pragma whose single argument is
1803 the name of the package containing the extended definition
1804 (e.g.@: @code{Aux_DEC} for the DEC Ada case). A unit compiled under
1805 control of this pragma will be processed using special visibility
1806 processing that looks in package @code{System.Aux_@var{xxx}} where
1807 @code{Aux_@var{xxx}} is the pragma argument for any entity referenced in
1808 package @code{System}, but not found in package @code{System}.
1810 You can use this pragma either to access a predefined @code{System}
1811 extension supplied with the compiler, for example @code{Aux_DEC} or
1812 you can construct your own extension unit following the above
1813 definition. Note that such a package is a child of @code{System}
1814 and thus is considered part of the implementation. To compile
1815 it you will have to use the appropriate switch for compiling
1816 system units. See the GNAT User's Guide for details.
1818 @node Pragma External
1819 @unnumberedsec Pragma External
1824 @smallexample @c ada
1826 [ Convention =>] convention_IDENTIFIER,
1827 [ Entity =>] local_NAME
1828 [, [External_Name =>] static_string_EXPRESSION ]
1829 [, [Link_Name =>] static_string_EXPRESSION ]);
1833 This pragma is identical in syntax and semantics to pragma
1834 @code{Export} as defined in the Ada Reference Manual. It is
1835 provided for compatibility with some Ada 83 compilers that
1836 used this pragma for exactly the same purposes as pragma
1837 @code{Export} before the latter was standardized.
1839 @node Pragma External_Name_Casing
1840 @unnumberedsec Pragma External_Name_Casing
1841 @cindex Dec Ada 83 casing compatibility
1842 @cindex External Names, casing
1843 @cindex Casing of External names
1844 @findex External_Name_Casing
1848 @smallexample @c ada
1849 pragma External_Name_Casing (
1850 Uppercase | Lowercase
1851 [, Uppercase | Lowercase | As_Is]);
1855 This pragma provides control over the casing of external names associated
1856 with Import and Export pragmas. There are two cases to consider:
1859 @item Implicit external names
1860 Implicit external names are derived from identifiers. The most common case
1861 arises when a standard Ada 95 Import or Export pragma is used with only two
1864 @smallexample @c ada
1865 pragma Import (C, C_Routine);
1869 Since Ada is a case insensitive language, the spelling of the identifier in
1870 the Ada source program does not provide any information on the desired
1871 casing of the external name, and so a convention is needed. In GNAT the
1872 default treatment is that such names are converted to all lower case
1873 letters. This corresponds to the normal C style in many environments.
1874 The first argument of pragma @code{External_Name_Casing} can be used to
1875 control this treatment. If @code{Uppercase} is specified, then the name
1876 will be forced to all uppercase letters. If @code{Lowercase} is specified,
1877 then the normal default of all lower case letters will be used.
1879 This same implicit treatment is also used in the case of extended DEC Ada 83
1880 compatible Import and Export pragmas where an external name is explicitly
1881 specified using an identifier rather than a string.
1883 @item Explicit external names
1884 Explicit external names are given as string literals. The most common case
1885 arises when a standard Ada 95 Import or Export pragma is used with three
1888 @smallexample @c ada
1889 pragma Import (C, C_Routine, "C_routine");
1893 In this case, the string literal normally provides the exact casing required
1894 for the external name. The second argument of pragma
1895 @code{External_Name_Casing} may be used to modify this behavior.
1896 If @code{Uppercase} is specified, then the name
1897 will be forced to all uppercase letters. If @code{Lowercase} is specified,
1898 then the name will be forced to all lowercase letters. A specification of
1899 @code{As_Is} provides the normal default behavior in which the casing is
1900 taken from the string provided.
1904 This pragma may appear anywhere that a pragma is valid. In particular, it
1905 can be used as a configuration pragma in the @file{gnat.adc} file, in which
1906 case it applies to all subsequent compilations, or it can be used as a program
1907 unit pragma, in which case it only applies to the current unit, or it can
1908 be used more locally to control individual Import/Export pragmas.
1910 It is primarily intended for use with OpenVMS systems, where many
1911 compilers convert all symbols to upper case by default. For interfacing to
1912 such compilers (e.g.@: the DEC C compiler), it may be convenient to use
1915 @smallexample @c ada
1916 pragma External_Name_Casing (Uppercase, Uppercase);
1920 to enforce the upper casing of all external symbols.
1922 @node Pragma Finalize_Storage_Only
1923 @unnumberedsec Pragma Finalize_Storage_Only
1924 @findex Finalize_Storage_Only
1928 @smallexample @c ada
1929 pragma Finalize_Storage_Only (first_subtype_LOCAL_NAME);
1933 This pragma allows the compiler not to emit a Finalize call for objects
1934 defined at the library level. This is mostly useful for types where
1935 finalization is only used to deal with storage reclamation since in most
1936 environments it is not necessary to reclaim memory just before terminating
1937 execution, hence the name.
1939 @node Pragma Float_Representation
1940 @unnumberedsec Pragma Float_Representation
1942 @findex Float_Representation
1946 @smallexample @c ada
1947 pragma Float_Representation (FLOAT_REP);
1949 FLOAT_REP ::= VAX_Float | IEEE_Float
1954 allows control over the internal representation chosen for the predefined
1955 floating point types declared in the packages @code{Standard} and
1956 @code{System}. On all systems other than OpenVMS, the argument must
1957 be @code{IEEE_Float} and the pragma has no effect. On OpenVMS, the
1958 argument may be @code{VAX_Float} to specify the use of the VAX float
1959 format for the floating-point types in Standard. This requires that
1960 the standard runtime libraries be recompiled. See the
1961 description of the @code{GNAT LIBRARY} command in the OpenVMS version
1962 of the GNAT Users Guide for details on the use of this command.
1965 @unnumberedsec Pragma Ident
1970 @smallexample @c ada
1971 pragma Ident (static_string_EXPRESSION);
1975 This pragma provides a string identification in the generated object file,
1976 if the system supports the concept of this kind of identification string.
1977 This pragma is allowed only in the outermost declarative part or
1978 declarative items of a compilation unit. If more than one @code{Ident}
1979 pragma is given, only the last one processed is effective.
1981 On OpenVMS systems, the effect of the pragma is identical to the effect of
1982 the DEC Ada 83 pragma of the same name. Note that in DEC Ada 83, the
1983 maximum allowed length is 31 characters, so if it is important to
1984 maintain compatibility with this compiler, you should obey this length
1987 @node Pragma Import_Exception
1988 @unnumberedsec Pragma Import_Exception
1990 @findex Import_Exception
1994 @smallexample @c ada
1995 pragma Import_Exception (
1996 [Internal =>] LOCAL_NAME,
1997 [, [External =>] EXTERNAL_SYMBOL,]
1998 [, [Form =>] Ada | VMS]
1999 [, [Code =>] static_integer_EXPRESSION]);
2003 | static_string_EXPRESSION
2007 This pragma is implemented only in the OpenVMS implementation of GNAT@.
2008 It allows OpenVMS conditions (for example, from OpenVMS system services or
2009 other OpenVMS languages) to be propagated to Ada programs as Ada exceptions.
2010 The pragma specifies that the exception associated with an exception
2011 declaration in an Ada program be defined externally (in non-Ada code).
2012 For further details on this pragma, see the
2013 DEC Ada Language Reference Manual, section 13.9a.3.1.
2015 @node Pragma Import_Function
2016 @unnumberedsec Pragma Import_Function
2017 @findex Import_Function
2021 @smallexample @c ada
2022 pragma Import_Function (
2023 [Internal =>] LOCAL_NAME,
2024 [, [External =>] EXTERNAL_SYMBOL]
2025 [, [Parameter_Types =>] PARAMETER_TYPES]
2026 [, [Result_Type =>] SUBTYPE_MARK]
2027 [, [Mechanism =>] MECHANISM]
2028 [, [Result_Mechanism =>] MECHANISM_NAME]
2029 [, [First_Optional_Parameter =>] IDENTIFIER]);
2033 | static_string_EXPRESSION
2037 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
2041 | subtype_Name ' Access
2045 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
2047 MECHANISM_ASSOCIATION ::=
2048 [formal_parameter_NAME =>] MECHANISM_NAME
2053 | Descriptor [([Class =>] CLASS_NAME)]
2055 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
2059 This pragma is used in conjunction with a pragma @code{Import} to
2060 specify additional information for an imported function. The pragma
2061 @code{Import} (or equivalent pragma @code{Interface}) must precede the
2062 @code{Import_Function} pragma and both must appear in the same
2063 declarative part as the function specification.
2065 The @var{Internal} argument must uniquely designate
2066 the function to which the
2067 pragma applies. If more than one function name exists of this name in
2068 the declarative part you must use the @code{Parameter_Types} and
2069 @var{Result_Type} parameters to achieve the required unique
2070 designation. Subtype marks in these parameters must exactly match the
2071 subtypes in the corresponding function specification, using positional
2072 notation to match parameters with subtype marks.
2073 The form with an @code{'Access} attribute can be used to match an
2074 anonymous access parameter.
2076 You may optionally use the @var{Mechanism} and @var{Result_Mechanism}
2077 parameters to specify passing mechanisms for the
2078 parameters and result. If you specify a single mechanism name, it
2079 applies to all parameters. Otherwise you may specify a mechanism on a
2080 parameter by parameter basis using either positional or named
2081 notation. If the mechanism is not specified, the default mechanism
2085 @cindex Passing by descriptor
2086 Passing by descriptor is supported only on the OpenVMS ports of GNAT@.
2088 @code{First_Optional_Parameter} applies only to OpenVMS ports of GNAT@.
2089 It specifies that the designated parameter and all following parameters
2090 are optional, meaning that they are not passed at the generated code
2091 level (this is distinct from the notion of optional parameters in Ada
2092 where the parameters are passed anyway with the designated optional
2093 parameters). All optional parameters must be of mode @code{IN} and have
2094 default parameter values that are either known at compile time
2095 expressions, or uses of the @code{'Null_Parameter} attribute.
2097 @node Pragma Import_Object
2098 @unnumberedsec Pragma Import_Object
2099 @findex Import_Object
2103 @smallexample @c ada
2104 pragma Import_Object
2105 [Internal =>] LOCAL_NAME,
2106 [, [External =>] EXTERNAL_SYMBOL],
2107 [, [Size =>] EXTERNAL_SYMBOL]);
2111 | static_string_EXPRESSION
2115 This pragma designates an object as imported, and apart from the
2116 extended rules for external symbols, is identical in effect to the use of
2117 the normal @code{Import} pragma applied to an object. Unlike the
2118 subprogram case, you need not use a separate @code{Import} pragma,
2119 although you may do so (and probably should do so from a portability
2120 point of view). @var{size} is syntax checked, but otherwise ignored by
2123 @node Pragma Import_Procedure
2124 @unnumberedsec Pragma Import_Procedure
2125 @findex Import_Procedure
2129 @smallexample @c ada
2130 pragma Import_Procedure (
2131 [Internal =>] LOCAL_NAME,
2132 [, [External =>] EXTERNAL_SYMBOL]
2133 [, [Parameter_Types =>] PARAMETER_TYPES]
2134 [, [Mechanism =>] MECHANISM]
2135 [, [First_Optional_Parameter =>] IDENTIFIER]);
2139 | static_string_EXPRESSION
2143 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
2147 | subtype_Name ' Access
2151 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
2153 MECHANISM_ASSOCIATION ::=
2154 [formal_parameter_NAME =>] MECHANISM_NAME
2159 | Descriptor [([Class =>] CLASS_NAME)]
2161 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
2165 This pragma is identical to @code{Import_Function} except that it
2166 applies to a procedure rather than a function and the parameters
2167 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
2169 @node Pragma Import_Valued_Procedure
2170 @unnumberedsec Pragma Import_Valued_Procedure
2171 @findex Import_Valued_Procedure
2175 @smallexample @c ada
2176 pragma Import_Valued_Procedure (
2177 [Internal =>] LOCAL_NAME,
2178 [, [External =>] EXTERNAL_SYMBOL]
2179 [, [Parameter_Types =>] PARAMETER_TYPES]
2180 [, [Mechanism =>] MECHANISM]
2181 [, [First_Optional_Parameter =>] IDENTIFIER]);
2185 | static_string_EXPRESSION
2189 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
2193 | subtype_Name ' Access
2197 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
2199 MECHANISM_ASSOCIATION ::=
2200 [formal_parameter_NAME =>] MECHANISM_NAME
2205 | Descriptor [([Class =>] CLASS_NAME)]
2207 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
2211 This pragma is identical to @code{Import_Procedure} except that the
2212 first parameter of @var{local_name}, which must be present, must be of
2213 mode @code{OUT}, and externally the subprogram is treated as a function
2214 with this parameter as the result of the function. The purpose of this
2215 capability is to allow the use of @code{OUT} and @code{IN OUT}
2216 parameters in interfacing to external functions (which are not permitted
2217 in Ada functions). You may optionally use the @code{Mechanism}
2218 parameters to specify passing mechanisms for the parameters.
2219 If you specify a single mechanism name, it applies to all parameters.
2220 Otherwise you may specify a mechanism on a parameter by parameter
2221 basis using either positional or named notation. If the mechanism is not
2222 specified, the default mechanism is used.
2224 Note that it is important to use this pragma in conjunction with a separate
2225 pragma Import that specifies the desired convention, since otherwise the
2226 default convention is Ada, which is almost certainly not what is required.
2228 @node Pragma Initialize_Scalars
2229 @unnumberedsec Pragma Initialize_Scalars
2230 @findex Initialize_Scalars
2231 @cindex debugging with Initialize_Scalars
2235 @smallexample @c ada
2236 pragma Initialize_Scalars;
2240 This pragma is similar to @code{Normalize_Scalars} conceptually but has
2241 two important differences. First, there is no requirement for the pragma
2242 to be used uniformly in all units of a partition, in particular, it is fine
2243 to use this just for some or all of the application units of a partition,
2244 without needing to recompile the run-time library.
2246 In the case where some units are compiled with the pragma, and some without,
2247 then a declaration of a variable where the type is defined in package
2248 Standard or is locally declared will always be subject to initialization,
2249 as will any declaration of a scalar variable. For composite variables,
2250 whether the variable is initialized may also depend on whether the package
2251 in which the type of the variable is declared is compiled with the pragma.
2253 The other important difference is that there is control over the value used
2254 for initializing scalar objects. At bind time, you can select whether to
2255 initialize with invalid values (like Normalize_Scalars), or with high or
2256 low values, or with a specified bit pattern. See the users guide for binder
2257 options for specifying these cases.
2259 This means that you can compile a program, and then without having to
2260 recompile the program, you can run it with different values being used
2261 for initializing otherwise uninitialized values, to test if your program
2262 behavior depends on the choice. Of course the behavior should not change,
2263 and if it does, then most likely you have an erroneous reference to an
2264 uninitialized value.
2266 Note that pragma @code{Initialize_Scalars} is particularly useful in
2267 conjunction with the enhanced validity checking that is now provided
2268 in GNAT, which checks for invalid values under more conditions.
2269 Using this feature (see description of the @code{-gnatV} flag in the
2270 users guide) in conjunction with pragma @code{Initialize_Scalars}
2271 provides a powerful new tool to assist in the detection of problems
2272 caused by uninitialized variables.
2274 Note: the use of @code{Initialize_Scalars} has a fairly extensive
2275 effect on the generated code. This may cause your code to be
2276 substantially larger. It may also cause an increase in the amount
2277 of stack required, so it is probably a good idea to turn on stack
2278 checking (see description of stack checking in the GNAT users guide)
2279 when using this pragma.
2281 @node Pragma Inline_Always
2282 @unnumberedsec Pragma Inline_Always
2283 @findex Inline_Always
2287 @smallexample @c ada
2288 pragma Inline_Always (NAME [, NAME]);
2292 Similar to pragma @code{Inline} except that inlining is not subject to
2293 the use of option @code{-gnatn} and the inlining happens regardless of
2294 whether this option is used.
2296 @node Pragma Inline_Generic
2297 @unnumberedsec Pragma Inline_Generic
2298 @findex Inline_Generic
2302 @smallexample @c ada
2303 pragma Inline_Generic (generic_package_NAME);
2307 This is implemented for compatibility with DEC Ada 83 and is recognized,
2308 but otherwise ignored, by GNAT@. All generic instantiations are inlined
2309 by default when using GNAT@.
2311 @node Pragma Interface
2312 @unnumberedsec Pragma Interface
2317 @smallexample @c ada
2319 [Convention =>] convention_identifier,
2320 [Entity =>] local_name
2321 [, [External_Name =>] static_string_expression],
2322 [, [Link_Name =>] static_string_expression]);
2326 This pragma is identical in syntax and semantics to
2327 the standard Ada 95 pragma @code{Import}. It is provided for compatibility
2328 with Ada 83. The definition is upwards compatible both with pragma
2329 @code{Interface} as defined in the Ada 83 Reference Manual, and also
2330 with some extended implementations of this pragma in certain Ada 83
2333 @node Pragma Interface_Name
2334 @unnumberedsec Pragma Interface_Name
2335 @findex Interface_Name
2339 @smallexample @c ada
2340 pragma Interface_Name (
2341 [Entity =>] LOCAL_NAME
2342 [, [External_Name =>] static_string_EXPRESSION]
2343 [, [Link_Name =>] static_string_EXPRESSION]);
2347 This pragma provides an alternative way of specifying the interface name
2348 for an interfaced subprogram, and is provided for compatibility with Ada
2349 83 compilers that use the pragma for this purpose. You must provide at
2350 least one of @var{External_Name} or @var{Link_Name}.
2352 @node Pragma Interrupt_Handler
2353 @unnumberedsec Pragma Interrupt_Handler
2354 @findex Interrupt_Handler
2358 @smallexample @c ada
2359 pragma Interrupt_Handler (procedure_LOCAL_NAME);
2363 This program unit pragma is supported for parameterless protected procedures
2364 as described in Annex C of the Ada Reference Manual. On the AAMP target
2365 the pragma can also be specified for nonprotected parameterless procedures
2366 that are declared at the library level (which includes procedures
2367 declared at the top level of a library package). In the case of AAMP,
2368 when this pragma is applied to a nonprotected procedure, the instruction
2369 @code{IERET} is generated for returns from the procedure, enabling
2370 maskable interrupts, in place of the normal return instruction.
2372 @node Pragma Interrupt_State
2373 @unnumberedsec Pragma Interrupt_State
2374 @findex Interrupt_State
2378 @smallexample @c ada
2379 pragma Interrupt_State (Name => value, State => SYSTEM | RUNTIME | USER);
2383 Normally certain interrupts are reserved to the implementation. Any attempt
2384 to attach an interrupt causes Program_Error to be raised, as described in
2385 RM C.3.2(22). A typical example is the @code{SIGINT} interrupt used in
2386 many systems for an @kbd{Ctrl-C} interrupt. Normally this interrupt is
2387 reserved to the implementation, so that @kbd{Ctrl-C} can be used to
2388 interrupt execution. Additionally, signals such as @code{SIGSEGV},
2389 @code{SIGABRT}, @code{SIGFPE} and @code{SIGILL} are often mapped to specific
2390 Ada exceptions, or used to implement run-time functions such as the
2391 @code{abort} statement and stack overflow checking.
2393 Pragma @code{Interrupt_State} provides a general mechanism for overriding
2394 such uses of interrupts. It subsumes the functionality of pragma
2395 @code{Unreserve_All_Interrupts}. Pragma @code{Interrupt_State} is not
2396 available on OS/2, Windows or VMS. On all other platforms than VxWorks,
2397 it applies to signals; on VxWorks, it applies to vectored hardware interrupts
2398 and may be used to mark interrupts required by the board support package
2401 Interrupts can be in one of three states:
2405 The interrupt is reserved (no Ada handler can be installed), and the
2406 Ada run-time may not install a handler. As a result you are guaranteed
2407 standard system default action if this interrupt is raised.
2411 The interrupt is reserved (no Ada handler can be installed). The run time
2412 is allowed to install a handler for internal control purposes, but is
2413 not required to do so.
2417 The interrupt is unreserved. The user may install a handler to provide
2422 These states are the allowed values of the @code{State} parameter of the
2423 pragma. The @code{Name} parameter is a value of the type
2424 @code{Ada.Interrupts.Interrupt_ID}. Typically, it is a name declared in
2425 @code{Ada.Interrupts.Names}.
2427 This is a configuration pragma, and the binder will check that there
2428 are no inconsistencies between different units in a partition in how a
2429 given interrupt is specified. It may appear anywhere a pragma is legal.
2431 The effect is to move the interrupt to the specified state.
2433 By declaring interrupts to be SYSTEM, you guarantee the standard system
2434 action, such as a core dump.
2436 By declaring interrupts to be USER, you guarantee that you can install
2439 Note that certain signals on many operating systems cannot be caught and
2440 handled by applications. In such cases, the pragma is ignored. See the
2441 operating system documentation, or the value of the array @code{Reserved}
2442 declared in the specification of package @code{System.OS_Interface}.
2444 Overriding the default state of signals used by the Ada runtime may interfere
2445 with an application's runtime behavior in the cases of the synchronous signals,
2446 and in the case of the signal used to implement the @code{abort} statement.
2448 @node Pragma Keep_Names
2449 @unnumberedsec Pragma Keep_Names
2454 @smallexample @c ada
2455 pragma Keep_Names ([On =>] enumeration_first_subtype_LOCAL_NAME);
2459 The @var{LOCAL_NAME} argument
2460 must refer to an enumeration first subtype
2461 in the current declarative part. The effect is to retain the enumeration
2462 literal names for use by @code{Image} and @code{Value} even if a global
2463 @code{Discard_Names} pragma applies. This is useful when you want to
2464 generally suppress enumeration literal names and for example you therefore
2465 use a @code{Discard_Names} pragma in the @file{gnat.adc} file, but you
2466 want to retain the names for specific enumeration types.
2468 @node Pragma License
2469 @unnumberedsec Pragma License
2471 @cindex License checking
2475 @smallexample @c ada
2476 pragma License (Unrestricted | GPL | Modified_GPL | Restricted);
2480 This pragma is provided to allow automated checking for appropriate license
2481 conditions with respect to the standard and modified GPL@. A pragma
2482 @code{License}, which is a configuration pragma that typically appears at
2483 the start of a source file or in a separate @file{gnat.adc} file, specifies
2484 the licensing conditions of a unit as follows:
2488 This is used for a unit that can be freely used with no license restrictions.
2489 Examples of such units are public domain units, and units from the Ada
2493 This is used for a unit that is licensed under the unmodified GPL, and which
2494 therefore cannot be @code{with}'ed by a restricted unit.
2497 This is used for a unit licensed under the GNAT modified GPL that includes
2498 a special exception paragraph that specifically permits the inclusion of
2499 the unit in programs without requiring the entire program to be released
2500 under the GPL@. This is the license used for the GNAT run-time which ensures
2501 that the run-time can be used freely in any program without GPL concerns.
2504 This is used for a unit that is restricted in that it is not permitted to
2505 depend on units that are licensed under the GPL@. Typical examples are
2506 proprietary code that is to be released under more restrictive license
2507 conditions. Note that restricted units are permitted to @code{with} units
2508 which are licensed under the modified GPL (this is the whole point of the
2514 Normally a unit with no @code{License} pragma is considered to have an
2515 unknown license, and no checking is done. However, standard GNAT headers
2516 are recognized, and license information is derived from them as follows.
2520 A GNAT license header starts with a line containing 78 hyphens. The following
2521 comment text is searched for the appearance of any of the following strings.
2523 If the string ``GNU General Public License'' is found, then the unit is assumed
2524 to have GPL license, unless the string ``As a special exception'' follows, in
2525 which case the license is assumed to be modified GPL@.
2527 If one of the strings
2528 ``This specification is adapted from the Ada Semantic Interface'' or
2529 ``This specification is derived from the Ada Reference Manual'' is found
2530 then the unit is assumed to be unrestricted.
2534 These default actions means that a program with a restricted license pragma
2535 will automatically get warnings if a GPL unit is inappropriately
2536 @code{with}'ed. For example, the program:
2538 @smallexample @c ada
2541 procedure Secret_Stuff is
2547 if compiled with pragma @code{License} (@code{Restricted}) in a
2548 @file{gnat.adc} file will generate the warning:
2553 >>> license of withed unit "Sem_Ch3" is incompatible
2555 2. with GNAT.Sockets;
2556 3. procedure Secret_Stuff is
2560 Here we get a warning on @code{Sem_Ch3} since it is part of the GNAT
2561 compiler and is licensed under the
2562 GPL, but no warning for @code{GNAT.Sockets} which is part of the GNAT
2563 run time, and is therefore licensed under the modified GPL@.
2565 @node Pragma Link_With
2566 @unnumberedsec Pragma Link_With
2571 @smallexample @c ada
2572 pragma Link_With (static_string_EXPRESSION @{,static_string_EXPRESSION@});
2576 This pragma is provided for compatibility with certain Ada 83 compilers.
2577 It has exactly the same effect as pragma @code{Linker_Options} except
2578 that spaces occurring within one of the string expressions are treated
2579 as separators. For example, in the following case:
2581 @smallexample @c ada
2582 pragma Link_With ("-labc -ldef");
2586 results in passing the strings @code{-labc} and @code{-ldef} as two
2587 separate arguments to the linker. In addition pragma Link_With allows
2588 multiple arguments, with the same effect as successive pragmas.
2590 @node Pragma Linker_Alias
2591 @unnumberedsec Pragma Linker_Alias
2592 @findex Linker_Alias
2596 @smallexample @c ada
2597 pragma Linker_Alias (
2598 [Entity =>] LOCAL_NAME
2599 [Alias =>] static_string_EXPRESSION);
2603 This pragma establishes a linker alias for the given named entity. For
2604 further details on the exact effect, consult the GCC manual.
2606 @node Pragma Linker_Section
2607 @unnumberedsec Pragma Linker_Section
2608 @findex Linker_Section
2612 @smallexample @c ada
2613 pragma Linker_Section (
2614 [Entity =>] LOCAL_NAME
2615 [Section =>] static_string_EXPRESSION);
2619 This pragma specifies the name of the linker section for the given entity.
2620 For further details on the exact effect, consult the GCC manual.
2622 @node Pragma Long_Float
2623 @unnumberedsec Pragma Long_Float
2629 @smallexample @c ada
2630 pragma Long_Float (FLOAT_FORMAT);
2632 FLOAT_FORMAT ::= D_Float | G_Float
2636 This pragma is implemented only in the OpenVMS implementation of GNAT@.
2637 It allows control over the internal representation chosen for the predefined
2638 type @code{Long_Float} and for floating point type representations with
2639 @code{digits} specified in the range 7 through 15.
2640 For further details on this pragma, see the
2641 @cite{DEC Ada Language Reference Manual}, section 3.5.7b. Note that to use
2642 this pragma, the standard runtime libraries must be recompiled. See the
2643 description of the @code{GNAT LIBRARY} command in the OpenVMS version
2644 of the GNAT User's Guide for details on the use of this command.
2646 @node Pragma Machine_Attribute
2647 @unnumberedsec Pragma Machine_Attribute
2648 @findex Machine_Attribute
2652 @smallexample @c ada
2653 pragma Machine_Attribute (
2654 [Attribute_Name =>] string_EXPRESSION,
2655 [Entity =>] LOCAL_NAME);
2659 Machine dependent attributes can be specified for types and/or
2660 declarations. Currently only subprogram entities are supported. This
2661 pragma is semantically equivalent to
2662 @code{__attribute__((@var{string_expression}))} in GNU C,
2663 where @code{@var{string_expression}} is
2664 recognized by the GNU C macros @code{VALID_MACHINE_TYPE_ATTRIBUTE} and
2665 @code{VALID_MACHINE_DECL_ATTRIBUTE} which are defined in the
2666 configuration header file @file{tm.h} for each machine. See the GCC
2667 manual for further information.
2669 @node Pragma Main_Storage
2670 @unnumberedsec Pragma Main_Storage
2672 @findex Main_Storage
2676 @smallexample @c ada
2678 (MAIN_STORAGE_OPTION [, MAIN_STORAGE_OPTION]);
2680 MAIN_STORAGE_OPTION ::=
2681 [WORKING_STORAGE =>] static_SIMPLE_EXPRESSION
2682 | [TOP_GUARD =>] static_SIMPLE_EXPRESSION
2687 This pragma is provided for compatibility with OpenVMS VAX Systems. It has
2688 no effect in GNAT, other than being syntax checked. Note that the pragma
2689 also has no effect in DEC Ada 83 for OpenVMS Alpha Systems.
2691 @node Pragma No_Return
2692 @unnumberedsec Pragma No_Return
2697 @smallexample @c ada
2698 pragma No_Return (procedure_LOCAL_NAME);
2702 @var{procedure_local_NAME} must refer to one or more procedure
2703 declarations in the current declarative part. A procedure to which this
2704 pragma is applied may not contain any explicit @code{return} statements,
2705 and also may not contain any implicit return statements from falling off
2706 the end of a statement sequence. One use of this pragma is to identify
2707 procedures whose only purpose is to raise an exception.
2709 Another use of this pragma is to suppress incorrect warnings about
2710 missing returns in functions, where the last statement of a function
2711 statement sequence is a call to such a procedure.
2713 @node Pragma Normalize_Scalars
2714 @unnumberedsec Pragma Normalize_Scalars
2715 @findex Normalize_Scalars
2719 @smallexample @c ada
2720 pragma Normalize_Scalars;
2724 This is a language defined pragma which is fully implemented in GNAT@. The
2725 effect is to cause all scalar objects that are not otherwise initialized
2726 to be initialized. The initial values are implementation dependent and
2730 @item Standard.Character
2732 Objects whose root type is Standard.Character are initialized to
2733 Character'Last. This will be out of range of the subtype only if
2734 the subtype range excludes this value.
2736 @item Standard.Wide_Character
2738 Objects whose root type is Standard.Wide_Character are initialized to
2739 Wide_Character'Last. This will be out of range of the subtype only if
2740 the subtype range excludes this value.
2744 Objects of an integer type are initialized to base_type'First, where
2745 base_type is the base type of the object type. This will be out of range
2746 of the subtype only if the subtype range excludes this value. For example,
2747 if you declare the subtype:
2749 @smallexample @c ada
2750 subtype Ityp is integer range 1 .. 10;
2754 then objects of type x will be initialized to Integer'First, a negative
2755 number that is certainly outside the range of subtype @code{Ityp}.
2758 Objects of all real types (fixed and floating) are initialized to
2759 base_type'First, where base_Type is the base type of the object type.
2760 This will be out of range of the subtype only if the subtype range
2761 excludes this value.
2764 Objects of a modular type are initialized to typ'Last. This will be out
2765 of range of the subtype only if the subtype excludes this value.
2767 @item Enumeration types
2768 Objects of an enumeration type are initialized to all one-bits, i.e.@: to
2769 the value @code{2 ** typ'Size - 1}. This will be out of range of the
2770 enumeration subtype in all cases except where the subtype contains
2771 exactly 2**8, 2**16, or 2**32 elements.
2775 @node Pragma Obsolescent
2776 @unnumberedsec Pragma Obsolescent
2781 @smallexample @c ada
2782 pragma Obsolescent [(static_string_EXPRESSION)];
2786 This pragma must occur immediately following a subprogram
2787 declaration. It indicates that the associated function or procedure
2788 is considered obsolescent and should not be used. Typically this is
2789 used when an API must be modified by eventually removing or modifying
2790 existing subprograms. The pragma can be used at an intermediate stage
2791 when the subprogram is still present, but will be removed later.
2793 The effect of this pragma is to output a warning message that the
2794 subprogram is obsolescent if the appropriate warning option in the
2795 compiler is activated. If a parameter is present, then a second
2796 warning message is given containing this text.
2798 @node Pragma Passive
2799 @unnumberedsec Pragma Passive
2804 @smallexample @c ada
2805 pragma Passive ([Semaphore | No]);
2809 Syntax checked, but otherwise ignored by GNAT@. This is recognized for
2810 compatibility with DEC Ada 83 implementations, where it is used within a
2811 task definition to request that a task be made passive. If the argument
2812 @code{Semaphore} is present, or the argument is omitted, then DEC Ada 83
2813 treats the pragma as an assertion that the containing task is passive
2814 and that optimization of context switch with this task is permitted and
2815 desired. If the argument @code{No} is present, the task must not be
2816 optimized. GNAT does not attempt to optimize any tasks in this manner
2817 (since protected objects are available in place of passive tasks).
2819 @node Pragma Polling
2820 @unnumberedsec Pragma Polling
2825 @smallexample @c ada
2826 pragma Polling (ON | OFF);
2830 This pragma controls the generation of polling code. This is normally off.
2831 If @code{pragma Polling (ON)} is used then periodic calls are generated to
2832 the routine @code{Ada.Exceptions.Poll}. This routine is a separate unit in the
2833 runtime library, and can be found in file @file{a-excpol.adb}.
2835 Pragma @code{Polling} can appear as a configuration pragma (for example it
2836 can be placed in the @file{gnat.adc} file) to enable polling globally, or it
2837 can be used in the statement or declaration sequence to control polling
2840 A call to the polling routine is generated at the start of every loop and
2841 at the start of every subprogram call. This guarantees that the @code{Poll}
2842 routine is called frequently, and places an upper bound (determined by
2843 the complexity of the code) on the period between two @code{Poll} calls.
2845 The primary purpose of the polling interface is to enable asynchronous
2846 aborts on targets that cannot otherwise support it (for example Windows
2847 NT), but it may be used for any other purpose requiring periodic polling.
2848 The standard version is null, and can be replaced by a user program. This
2849 will require re-compilation of the @code{Ada.Exceptions} package that can
2850 be found in files @file{a-except.ads} and @file{a-except.adb}.
2852 A standard alternative unit (in file @file{4wexcpol.adb} in the standard GNAT
2853 distribution) is used to enable the asynchronous abort capability on
2854 targets that do not normally support the capability. The version of
2855 @code{Poll} in this file makes a call to the appropriate runtime routine
2856 to test for an abort condition.
2858 Note that polling can also be enabled by use of the @code{-gnatP} switch. See
2859 the @cite{GNAT User's Guide} for details.
2861 @node Pragma Profile (Ravenscar)
2862 @unnumberedsec Pragma Profile (Ravenscar)
2867 @smallexample @c ada
2868 pragma Profile (Ravenscar);
2872 A configuration pragma that establishes the following set of configuration
2876 @item Task_Dispatching_Policy (FIFO_Within_Priorities)
2877 [RM D.2.2] Tasks are dispatched following a preemptive
2878 priority-ordered scheduling policy.
2880 @item Locking_Policy (Ceiling_Locking)
2881 [RM D.3] While tasks and interrupts execute a protected action, they inherit
2882 the ceiling priority of the corresponding protected object.
2884 @c @item Detect_Blocking
2885 @c This pragma forces the detection of potentially blocking operations within a
2886 @c protected operation, and to raise Program_Error if that happens.
2890 plus the following set of restrictions:
2893 @item Max_Entry_Queue_Length = 1
2894 Defines the maximum number of calls that are queued on a (protected) entry.
2895 Note that this restrictions is checked at run time. Violation of this
2896 restriction results in the raising of Program_Error exception at the point of
2897 the call. For the Profile (Ravenscar) the value of Max_Entry_Queue_Length is
2898 always 1 and hence no task can be queued on a protected entry.
2900 @item Max_Protected_Entries = 1
2901 [RM D.7] Specifies the maximum number of entries per protected type. The
2902 bounds of every entry family of a protected unit shall be static, or shall be
2903 defined by a discriminant of a subtype whose corresponding bound is static.
2904 For the Profile (Ravenscar) the value of Max_Protected_Entries is always 1.
2906 @item Max_Task_Entries = 0
2907 [RM D.7] Specifies the maximum number of entries
2908 per task. The bounds of every entry family
2909 of a task unit shall be static, or shall be
2910 defined by a discriminant of a subtype whose
2911 corresponding bound is static. A value of zero
2912 indicates that no rendezvous are possible. For
2913 the Profile (Ravenscar), the value of Max_Task_Entries is always
2916 @item No_Abort_Statements
2917 [RM D.7] There are no abort_statements, and there are
2918 no calls to Task_Identification.Abort_Task.
2920 @item No_Asynchronous_Control
2921 [RM D.7] There are no semantic dependences on the package
2922 Asynchronous_Task_Control.
2925 There are no semantic dependencies on the package Ada.Calendar.
2927 @item No_Dynamic_Attachment
2928 There is no call to any of the operations defined in package Ada.Interrupts
2929 (Is_Reserved, Is_Attached, Current_Handler, Attach_Handler, Exchange_Handler,
2930 Detach_Handler, and Reference).
2932 @item No_Dynamic_Priorities
2933 [RM D.7] There are no semantic dependencies on the package Dynamic_Priorities.
2935 @item No_Implicit_Heap_Allocations
2936 [RM D.7] No constructs are allowed to cause implicit heap allocation.
2938 @item No_Local_Protected_Objects
2939 Protected objects and access types that designate
2940 such objects shall be declared only at library level.
2942 @item No_Protected_Type_Allocators
2943 There are no allocators for protected types or
2944 types containing protected subcomponents.
2946 @item No_Relative_Delay
2947 There are no delay_relative statements.
2949 @item No_Requeue_Statements
2950 Requeue statements are not allowed.
2952 @item No_Select_Statements
2953 There are no select_statements.
2955 @item No_Task_Allocators
2956 [RM D.7] There are no allocators for task types
2957 or types containing task subcomponents.
2959 @item No_Task_Attributes_Package
2960 There are no semantic dependencies on the Ada.Task_Attributes package.
2962 @item No_Task_Hierarchy
2963 [RM D.7] All (non-environment) tasks depend
2964 directly on the environment task of the partition.
2966 @item No_Task_Termination
2967 Tasks which terminate are erroneous.
2969 @item Simple_Barriers
2970 Entry barrier condition expressions shall be either static
2971 boolean expressions or boolean objects which are declared in
2972 the protected type which contains the entry.
2976 This set of configuration pragmas and restrictions correspond to the
2977 definition of the ``Ravenscar Profile'' for limited tasking, devised and
2978 published by the @cite{International Real-Time Ada Workshop}, 1997,
2979 and whose most recent description is available at
2980 @url{ftp://ftp.openravenscar.org/openravenscar/ravenscar00.pdf}.
2982 The original definition of the profile was revised at subsequent IRTAW
2983 meetings. It has been included in the ISO
2984 @cite{Guide for the Use of the Ada Programming Language in High
2985 Integrity Systems}, and has been approved by ISO/IEC/SC22/WG9 for inclusion in
2986 the next revision of the standard. The formal definition given by
2987 the Ada Rapporteur Group (ARG) can be found in two Ada Issues (AI-249 and
2988 AI-305) available at
2989 @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs/AI-00249.TXT} and
2990 @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/AIs/AI-00305.TXT}
2993 The above set is a superset of the restrictions provided by pragma
2994 @code{Profile (Restricted)}, it includes six additional restrictions
2995 (@code{Simple_Barriers}, @code{No_Select_Statements},
2996 @code{No_Calendar}, @code{No_Implicit_Heap_Allocations},
2997 @code{No_Relative_Delay} and @code{No_Task_Termination}). This means
2998 that pragma @code{Profile (Ravenscar)}, like the pragma
2999 @code{Profile (Restricted)},
3000 automatically causes the use of a simplified,
3001 more efficient version of the tasking run-time system.
3003 @node Pragma Profile (Restricted)
3004 @unnumberedsec Pragma Profile (Restricted)
3005 @findex Restricted Run Time
3009 @smallexample @c ada
3010 pragma Profile (Restricted);
3014 A configuration pragma that establishes the following set of restrictions:
3017 @item No_Abort_Statements
3018 @item No_Entry_Queue
3019 @item No_Task_Hierarchy
3020 @item No_Task_Allocators
3021 @item No_Dynamic_Priorities
3022 @item No_Terminate_Alternatives
3023 @item No_Dynamic_Attachment
3024 @item No_Protected_Type_Allocators
3025 @item No_Local_Protected_Objects
3026 @item No_Requeue_Statements
3027 @item No_Task_Attributes_Package
3028 @item Max_Asynchronous_Select_Nesting = 0
3029 @item Max_Task_Entries = 0
3030 @item Max_Protected_Entries = 1
3031 @item Max_Select_Alternatives = 0
3035 This set of restrictions causes the automatic selection of a simplified
3036 version of the run time that provides improved performance for the
3037 limited set of tasking functionality permitted by this set of restrictions.
3039 @node Pragma Propagate_Exceptions
3040 @unnumberedsec Pragma Propagate_Exceptions
3041 @findex Propagate_Exceptions
3042 @cindex Zero Cost Exceptions
3046 @smallexample @c ada
3047 pragma Propagate_Exceptions (subprogram_LOCAL_NAME);
3051 This pragma indicates that the given entity, which is the name of an
3052 imported foreign-language subprogram may receive an Ada exception,
3053 and that the exception should be propagated. It is relevant only if
3054 zero cost exception handling is in use, and is thus never needed if
3055 the alternative @code{longjmp} / @code{setjmp} implementation of
3056 exceptions is used (although it is harmless to use it in such cases).
3058 The implementation of fast exceptions always properly propagates
3059 exceptions through Ada code, as described in the Ada Reference Manual.
3060 However, this manual is silent about the propagation of exceptions
3061 through foreign code. For example, consider the
3062 situation where @code{P1} calls
3063 @code{P2}, and @code{P2} calls @code{P3}, where
3064 @code{P1} and @code{P3} are in Ada, but @code{P2} is in C@.
3065 @code{P3} raises an Ada exception. The question is whether or not
3066 it will be propagated through @code{P2} and can be handled in
3069 For the @code{longjmp} / @code{setjmp} implementation of exceptions,
3070 the answer is always yes. For some targets on which zero cost exception
3071 handling is implemented, the answer is also always yes. However, there
3072 are some targets, notably in the current version all x86 architecture
3073 targets, in which the answer is that such propagation does not
3074 happen automatically. If such propagation is required on these
3075 targets, it is mandatory to use @code{Propagate_Exceptions} to
3076 name all foreign language routines through which Ada exceptions
3079 @node Pragma Psect_Object
3080 @unnumberedsec Pragma Psect_Object
3081 @findex Psect_Object
3085 @smallexample @c ada
3086 pragma Psect_Object (
3087 [Internal =>] LOCAL_NAME,
3088 [, [External =>] EXTERNAL_SYMBOL]
3089 [, [Size =>] EXTERNAL_SYMBOL]);
3093 | static_string_EXPRESSION
3097 This pragma is identical in effect to pragma @code{Common_Object}.
3099 @node Pragma Pure_Function
3100 @unnumberedsec Pragma Pure_Function
3101 @findex Pure_Function
3105 @smallexample @c ada
3106 pragma Pure_Function ([Entity =>] function_LOCAL_NAME);
3110 This pragma appears in the same declarative part as a function
3111 declaration (or a set of function declarations if more than one
3112 overloaded declaration exists, in which case the pragma applies
3113 to all entities). It specifies that the function @code{Entity} is
3114 to be considered pure for the purposes of code generation. This means
3115 that the compiler can assume that there are no side effects, and
3116 in particular that two calls with identical arguments produce the
3117 same result. It also means that the function can be used in an
3120 Note that, quite deliberately, there are no static checks to try
3121 to ensure that this promise is met, so @code{Pure_Function} can be used
3122 with functions that are conceptually pure, even if they do modify
3123 global variables. For example, a square root function that is
3124 instrumented to count the number of times it is called is still
3125 conceptually pure, and can still be optimized, even though it
3126 modifies a global variable (the count). Memo functions are another
3127 example (where a table of previous calls is kept and consulted to
3128 avoid re-computation).
3131 Note: Most functions in a @code{Pure} package are automatically pure, and
3132 there is no need to use pragma @code{Pure_Function} for such functions. One
3133 exception is any function that has at least one formal of type
3134 @code{System.Address} or a type derived from it. Such functions are not
3135 considered pure by default, since the compiler assumes that the
3136 @code{Address} parameter may be functioning as a pointer and that the
3137 referenced data may change even if the address value does not.
3138 Similarly, imported functions are not considered to be pure by default,
3139 since there is no way of checking that they are in fact pure. The use
3140 of pragma @code{Pure_Function} for such a function will override these default
3141 assumption, and cause the compiler to treat a designated subprogram as pure
3144 Note: If pragma @code{Pure_Function} is applied to a renamed function, it
3145 applies to the underlying renamed function. This can be used to
3146 disambiguate cases of overloading where some but not all functions
3147 in a set of overloaded functions are to be designated as pure.
3149 @node Pragma Restriction_Warnings
3150 @unnumberedsec Pragma Restriction_Warnings
3151 @findex Restriction_Warnings
3155 @smallexample @c ada
3156 pragma Restriction_Warnings
3157 (restriction_IDENTIFIER @{, restriction_IDENTIFIER@});
3161 This pragma allows a series of restriction identifiers to be
3162 specified (the list of allowed identifiers is the same as for
3163 pragma @code{Restrictions}). For each of these identifiers
3164 the compiler checks for violations of the restriction, but
3165 generates a warning message rather than an error message
3166 if the restriction is violated.
3168 @node Pragma Source_File_Name
3169 @unnumberedsec Pragma Source_File_Name
3170 @findex Source_File_Name
3174 @smallexample @c ada
3175 pragma Source_File_Name (
3176 [Unit_Name =>] unit_NAME,
3177 Spec_File_Name => STRING_LITERAL);
3179 pragma Source_File_Name (
3180 [Unit_Name =>] unit_NAME,
3181 Body_File_Name => STRING_LITERAL);
3185 Use this to override the normal naming convention. It is a configuration
3186 pragma, and so has the usual applicability of configuration pragmas
3187 (i.e.@: it applies to either an entire partition, or to all units in a
3188 compilation, or to a single unit, depending on how it is used.
3189 @var{unit_name} is mapped to @var{file_name_literal}. The identifier for
3190 the second argument is required, and indicates whether this is the file
3191 name for the spec or for the body.
3193 Another form of the @code{Source_File_Name} pragma allows
3194 the specification of patterns defining alternative file naming schemes
3195 to apply to all files.
3197 @smallexample @c ada
3198 pragma Source_File_Name
3199 (Spec_File_Name => STRING_LITERAL
3200 [,Casing => CASING_SPEC]
3201 [,Dot_Replacement => STRING_LITERAL]);
3203 pragma Source_File_Name
3204 (Body_File_Name => STRING_LITERAL
3205 [,Casing => CASING_SPEC]
3206 [,Dot_Replacement => STRING_LITERAL]);
3208 pragma Source_File_Name
3209 (Subunit_File_Name => STRING_LITERAL
3210 [,Casing => CASING_SPEC]
3211 [,Dot_Replacement => STRING_LITERAL]);
3213 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
3217 The first argument is a pattern that contains a single asterisk indicating
3218 the point at which the unit name is to be inserted in the pattern string
3219 to form the file name. The second argument is optional. If present it
3220 specifies the casing of the unit name in the resulting file name string.
3221 The default is lower case. Finally the third argument allows for systematic
3222 replacement of any dots in the unit name by the specified string literal.
3224 A pragma Source_File_Name cannot appear after a
3225 @ref{Pragma Source_File_Name_Project}.
3227 For more details on the use of the @code{Source_File_Name} pragma,
3228 see the sections ``Using Other File Names'' and
3229 ``Alternative File Naming Schemes'' in the @cite{GNAT User's Guide}.
3231 @node Pragma Source_File_Name_Project
3232 @unnumberedsec Pragma Source_File_Name_Project
3233 @findex Source_File_Name_Project
3236 This pragma has the same syntax and semantics as pragma Source_File_Name.
3237 It is only allowed as a stand alone configuration pragma.
3238 It cannot appear after a @ref{Pragma Source_File_Name}, and
3239 most importantly, once pragma Source_File_Name_Project appears,
3240 no further Source_File_Name pragmas are allowed.
3242 The intention is that Source_File_Name_Project pragmas are always
3243 generated by the Project Manager in a manner consistent with the naming
3244 specified in a project file, and when naming is controlled in this manner,
3245 it is not permissible to attempt to modify this naming scheme using
3246 Source_File_Name pragmas (which would not be known to the project manager).
3248 @node Pragma Source_Reference
3249 @unnumberedsec Pragma Source_Reference
3250 @findex Source_Reference
3254 @smallexample @c ada
3255 pragma Source_Reference (INTEGER_LITERAL, STRING_LITERAL);
3259 This pragma must appear as the first line of a source file.
3260 @var{integer_literal} is the logical line number of the line following
3261 the pragma line (for use in error messages and debugging
3262 information). @var{string_literal} is a static string constant that
3263 specifies the file name to be used in error messages and debugging
3264 information. This is most notably used for the output of @code{gnatchop}
3265 with the @code{-r} switch, to make sure that the original unchopped
3266 source file is the one referred to.
3268 The second argument must be a string literal, it cannot be a static
3269 string expression other than a string literal. This is because its value
3270 is needed for error messages issued by all phases of the compiler.
3272 @node Pragma Stream_Convert
3273 @unnumberedsec Pragma Stream_Convert
3274 @findex Stream_Convert
3278 @smallexample @c ada
3279 pragma Stream_Convert (
3280 [Entity =>] type_LOCAL_NAME,
3281 [Read =>] function_NAME,
3282 [Write =>] function_NAME);
3286 This pragma provides an efficient way of providing stream functions for
3287 types defined in packages. Not only is it simpler to use than declaring
3288 the necessary functions with attribute representation clauses, but more
3289 significantly, it allows the declaration to made in such a way that the
3290 stream packages are not loaded unless they are needed. The use of
3291 the Stream_Convert pragma adds no overhead at all, unless the stream
3292 attributes are actually used on the designated type.
3294 The first argument specifies the type for which stream functions are
3295 provided. The second parameter provides a function used to read values
3296 of this type. It must name a function whose argument type may be any
3297 subtype, and whose returned type must be the type given as the first
3298 argument to the pragma.
3300 The meaning of the @var{Read}
3301 parameter is that if a stream attribute directly
3302 or indirectly specifies reading of the type given as the first parameter,
3303 then a value of the type given as the argument to the Read function is
3304 read from the stream, and then the Read function is used to convert this
3305 to the required target type.
3307 Similarly the @var{Write} parameter specifies how to treat write attributes
3308 that directly or indirectly apply to the type given as the first parameter.
3309 It must have an input parameter of the type specified by the first parameter,
3310 and the return type must be the same as the input type of the Read function.
3311 The effect is to first call the Write function to convert to the given stream
3312 type, and then write the result type to the stream.
3314 The Read and Write functions must not be overloaded subprograms. If necessary
3315 renamings can be supplied to meet this requirement.
3316 The usage of this attribute is best illustrated by a simple example, taken
3317 from the GNAT implementation of package Ada.Strings.Unbounded:
3319 @smallexample @c ada
3320 function To_Unbounded (S : String)
3321 return Unbounded_String
3322 renames To_Unbounded_String;
3324 pragma Stream_Convert
3325 (Unbounded_String, To_Unbounded, To_String);
3329 The specifications of the referenced functions, as given in the Ada 95
3330 Reference Manual are:
3332 @smallexample @c ada
3333 function To_Unbounded_String (Source : String)
3334 return Unbounded_String;
3336 function To_String (Source : Unbounded_String)
3341 The effect is that if the value of an unbounded string is written to a
3342 stream, then the representation of the item in the stream is in the same
3343 format used for @code{Standard.String}, and this same representation is
3344 expected when a value of this type is read from the stream.
3346 @node Pragma Style_Checks
3347 @unnumberedsec Pragma Style_Checks
3348 @findex Style_Checks
3352 @smallexample @c ada
3353 pragma Style_Checks (string_LITERAL | ALL_CHECKS |
3354 On | Off [, LOCAL_NAME]);
3358 This pragma is used in conjunction with compiler switches to control the
3359 built in style checking provided by GNAT@. The compiler switches, if set,
3360 provide an initial setting for the switches, and this pragma may be used
3361 to modify these settings, or the settings may be provided entirely by
3362 the use of the pragma. This pragma can be used anywhere that a pragma
3363 is legal, including use as a configuration pragma (including use in
3364 the @file{gnat.adc} file).
3366 The form with a string literal specifies which style options are to be
3367 activated. These are additive, so they apply in addition to any previously
3368 set style check options. The codes for the options are the same as those
3369 used in the @code{-gnaty} switch to @code{gcc} or @code{gnatmake}.
3370 For example the following two methods can be used to enable
3375 @smallexample @c ada
3376 pragma Style_Checks ("l");
3381 gcc -c -gnatyl @dots{}
3386 The form ALL_CHECKS activates all standard checks (its use is equivalent
3387 to the use of the @code{gnaty} switch with no options. See GNAT User's
3390 The forms with @code{Off} and @code{On}
3391 can be used to temporarily disable style checks
3392 as shown in the following example:
3394 @smallexample @c ada
3398 pragma Style_Checks ("k"); -- requires keywords in lower case
3399 pragma Style_Checks (Off); -- turn off style checks
3400 NULL; -- this will not generate an error message
3401 pragma Style_Checks (On); -- turn style checks back on
3402 NULL; -- this will generate an error message
3406 Finally the two argument form is allowed only if the first argument is
3407 @code{On} or @code{Off}. The effect is to turn of semantic style checks
3408 for the specified entity, as shown in the following example:
3410 @smallexample @c ada
3414 pragma Style_Checks ("r"); -- require consistency of identifier casing
3416 Rf1 : Integer := ARG; -- incorrect, wrong case
3417 pragma Style_Checks (Off, Arg);
3418 Rf2 : Integer := ARG; -- OK, no error
3421 @node Pragma Subtitle
3422 @unnumberedsec Pragma Subtitle
3427 @smallexample @c ada
3428 pragma Subtitle ([Subtitle =>] STRING_LITERAL);
3432 This pragma is recognized for compatibility with other Ada compilers
3433 but is ignored by GNAT@.
3435 @node Pragma Suppress_All
3436 @unnumberedsec Pragma Suppress_All
3437 @findex Suppress_All
3441 @smallexample @c ada
3442 pragma Suppress_All;
3446 This pragma can only appear immediately following a compilation
3447 unit. The effect is to apply @code{Suppress (All_Checks)} to the unit
3448 which it follows. This pragma is implemented for compatibility with DEC
3449 Ada 83 usage. The use of pragma @code{Suppress (All_Checks)} as a normal
3450 configuration pragma is the preferred usage in GNAT@.
3452 @node Pragma Suppress_Exception_Locations
3453 @unnumberedsec Pragma Suppress_Exception_Locations
3454 @findex Suppress_Exception_Locations
3458 @smallexample @c ada
3459 pragma Suppress_Exception_Locations;
3463 In normal mode, a raise statement for an exception by default generates
3464 an exception message giving the file name and line number for the location
3465 of the raise. This is useful for debugging and logging purposes, but this
3466 entails extra space for the strings for the messages. The configuration
3467 pragma @code{Suppress_Exception_Locations} can be used to suppress the
3468 generation of these strings, with the result that space is saved, but the
3469 exception message for such raises is null. This configuration pragma may
3470 appear in a global configuration pragma file, or in a specific unit as
3471 usual. It is not required that this pragma be used consistently within
3472 a partition, so it is fine to have some units within a partition compiled
3473 with this pragma and others compiled in normal mode without it.
3475 @node Pragma Suppress_Initialization
3476 @unnumberedsec Pragma Suppress_Initialization
3477 @findex Suppress_Initialization
3478 @cindex Suppressing initialization
3479 @cindex Initialization, suppression of
3483 @smallexample @c ada
3484 pragma Suppress_Initialization ([Entity =>] type_Name);
3488 This pragma suppresses any implicit or explicit initialization
3489 associated with the given type name for all variables of this type.
3491 @node Pragma Task_Info
3492 @unnumberedsec Pragma Task_Info
3497 @smallexample @c ada
3498 pragma Task_Info (EXPRESSION);
3502 This pragma appears within a task definition (like pragma
3503 @code{Priority}) and applies to the task in which it appears. The
3504 argument must be of type @code{System.Task_Info.Task_Info_Type}.
3505 The @code{Task_Info} pragma provides system dependent control over
3506 aspects of tasking implementation, for example, the ability to map
3507 tasks to specific processors. For details on the facilities available
3508 for the version of GNAT that you are using, see the documentation
3509 in the specification of package System.Task_Info in the runtime
3512 @node Pragma Task_Name
3513 @unnumberedsec Pragma Task_Name
3518 @smallexample @c ada
3519 pragma Task_Name (string_EXPRESSION);
3523 This pragma appears within a task definition (like pragma
3524 @code{Priority}) and applies to the task in which it appears. The
3525 argument must be of type String, and provides a name to be used for
3526 the task instance when the task is created. Note that this expression
3527 is not required to be static, and in particular, it can contain
3528 references to task discriminants. This facility can be used to
3529 provide different names for different tasks as they are created,
3530 as illustrated in the example below.
3532 The task name is recorded internally in the run-time structures
3533 and is accessible to tools like the debugger. In addition the
3534 routine @code{Ada.Task_Identification.Image} will return this
3535 string, with a unique task address appended.
3537 @smallexample @c ada
3538 -- Example of the use of pragma Task_Name
3540 with Ada.Task_Identification;
3541 use Ada.Task_Identification;
3542 with Text_IO; use Text_IO;
3545 type Astring is access String;
3547 task type Task_Typ (Name : access String) is
3548 pragma Task_Name (Name.all);
3551 task body Task_Typ is
3552 Nam : constant String := Image (Current_Task);
3554 Put_Line ("-->" & Nam (1 .. 14) & "<--");
3557 type Ptr_Task is access Task_Typ;
3558 Task_Var : Ptr_Task;
3562 new Task_Typ (new String'("This is task 1"));
3564 new Task_Typ (new String'("This is task 2"));
3568 @node Pragma Task_Storage
3569 @unnumberedsec Pragma Task_Storage
3570 @findex Task_Storage
3573 @smallexample @c ada
3574 pragma Task_Storage (
3575 [Task_Type =>] LOCAL_NAME,
3576 [Top_Guard =>] static_integer_EXPRESSION);
3580 This pragma specifies the length of the guard area for tasks. The guard
3581 area is an additional storage area allocated to a task. A value of zero
3582 means that either no guard area is created or a minimal guard area is
3583 created, depending on the target. This pragma can appear anywhere a
3584 @code{Storage_Size} attribute definition clause is allowed for a task
3587 @node Pragma Thread_Body
3588 @unnumberedsec Pragma Thread_Body
3592 @smallexample @c ada
3593 pragma Thread_Body (
3594 [Entity =>] LOCAL_NAME,
3595 [[Secondary_Stack_Size =>] static_integer_EXPRESSION)];
3599 This pragma specifies that the subprogram whose name is given as the
3600 @code{Entity} argument is a thread body, which will be activated
3601 by being called via its Address from foreign code. The purpose is
3602 to allow execution and registration of the foreign thread within the
3603 Ada run-time system.
3605 See the library unit @code{System.Threads} for details on the expansion of
3606 a thread body subprogram, including the calls made to subprograms
3607 within System.Threads to register the task. This unit also lists the
3608 targets and runtime systems for which this pragma is supported.
3610 A thread body subprogram may not be called directly from Ada code, and
3611 it is not permitted to apply the Access (or Unrestricted_Access) attributes
3612 to such a subprogram. The only legitimate way of calling such a subprogram
3613 is to pass its Address to foreign code and then make the call from the
3616 A thread body subprogram may have any parameters, and it may be a function
3617 returning a result. The convention of the thread body subprogram may be
3618 set in the usual manner using @code{pragma Convention}.
3620 The secondary stack size parameter, if given, is used to set the size
3621 of secondary stack for the thread. The secondary stack is allocated as
3622 a local variable of the expanded thread body subprogram, and thus is
3623 allocated out of the main thread stack size. If no secondary stack
3624 size parameter is present, the default size (from the declaration in
3625 @code{System.Secondary_Stack} is used.
3627 @node Pragma Time_Slice
3628 @unnumberedsec Pragma Time_Slice
3633 @smallexample @c ada
3634 pragma Time_Slice (static_duration_EXPRESSION);
3638 For implementations of GNAT on operating systems where it is possible
3639 to supply a time slice value, this pragma may be used for this purpose.
3640 It is ignored if it is used in a system that does not allow this control,
3641 or if it appears in other than the main program unit.
3643 Note that the effect of this pragma is identical to the effect of the
3644 DEC Ada 83 pragma of the same name when operating under OpenVMS systems.
3647 @unnumberedsec Pragma Title
3652 @smallexample @c ada
3653 pragma Title (TITLING_OPTION [, TITLING OPTION]);
3656 [Title =>] STRING_LITERAL,
3657 | [Subtitle =>] STRING_LITERAL
3661 Syntax checked but otherwise ignored by GNAT@. This is a listing control
3662 pragma used in DEC Ada 83 implementations to provide a title and/or
3663 subtitle for the program listing. The program listing generated by GNAT
3664 does not have titles or subtitles.
3666 Unlike other pragmas, the full flexibility of named notation is allowed
3667 for this pragma, i.e.@: the parameters may be given in any order if named
3668 notation is used, and named and positional notation can be mixed
3669 following the normal rules for procedure calls in Ada.
3671 @node Pragma Unchecked_Union
3672 @unnumberedsec Pragma Unchecked_Union
3674 @findex Unchecked_Union
3678 @smallexample @c ada
3679 pragma Unchecked_Union (first_subtype_LOCAL_NAME);
3683 This pragma is used to declare that the specified type should be represented
3685 equivalent to a C union type, and is intended only for use in
3686 interfacing with C code that uses union types. In Ada terms, the named
3687 type must obey the following rules:
3691 It is a non-tagged non-limited record type.
3693 It has a single discrete discriminant with a default value.
3695 The component list consists of a single variant part.
3697 Each variant has a component list with a single component.
3699 No nested variants are allowed.
3701 No component has an explicit default value.
3703 No component has a non-static constraint.
3707 In addition, given a type that meets the above requirements, the
3708 following restrictions apply to its use throughout the program:
3712 The discriminant name can be mentioned only in an aggregate.
3714 No subtypes may be created of this type.
3716 The type may not be constrained by giving a discriminant value.
3718 The type cannot be passed as the actual for a generic formal with a
3723 Equality and inequality operations on @code{unchecked_unions} are not
3724 available, since there is no discriminant to compare and the compiler
3725 does not even know how many bits to compare. It is implementation
3726 dependent whether this is detected at compile time as an illegality or
3727 whether it is undetected and considered to be an erroneous construct. In
3728 GNAT, a direct comparison is illegal, but GNAT does not attempt to catch
3729 the composite case (where two composites are compared that contain an
3730 unchecked union component), so such comparisons are simply considered
3733 The layout of the resulting type corresponds exactly to a C union, where
3734 each branch of the union corresponds to a single variant in the Ada
3735 record. The semantics of the Ada program is not changed in any way by
3736 the pragma, i.e.@: provided the above restrictions are followed, and no
3737 erroneous incorrect references to fields or erroneous comparisons occur,
3738 the semantics is exactly as described by the Ada reference manual.
3739 Pragma @code{Suppress (Discriminant_Check)} applies implicitly to the
3740 type and the default convention is C.
3742 @node Pragma Unimplemented_Unit
3743 @unnumberedsec Pragma Unimplemented_Unit
3744 @findex Unimplemented_Unit
3748 @smallexample @c ada
3749 pragma Unimplemented_Unit;
3753 If this pragma occurs in a unit that is processed by the compiler, GNAT
3754 aborts with the message @samp{@var{xxx} not implemented}, where
3755 @var{xxx} is the name of the current compilation unit. This pragma is
3756 intended to allow the compiler to handle unimplemented library units in
3759 The abort only happens if code is being generated. Thus you can use
3760 specs of unimplemented packages in syntax or semantic checking mode.
3762 @node Pragma Universal_Data
3763 @unnumberedsec Pragma Universal_Data
3764 @findex Universal_Data
3768 @smallexample @c ada
3769 pragma Universal_Data [(library_unit_Name)];
3773 This pragma is supported only for the AAMP target and is ignored for
3774 other targets. The pragma specifies that all library-level objects
3775 (Counter 0 data) associated with the library unit are to be accessed
3776 and updated using universal addressing (24-bit addresses for AAMP5)
3777 rather than the default of 16-bit Data Environment (DENV) addressing.
3778 Use of this pragma will generally result in less efficient code for
3779 references to global data associated with the library unit, but
3780 allows such data to be located anywhere in memory. This pragma is
3781 a library unit pragma, but can also be used as a configuration pragma
3782 (including use in the @file{gnat.adc} file). The functionality
3783 of this pragma is also available by applying the -univ switch on the
3784 compilations of units where universal addressing of the data is desired.
3786 @node Pragma Unreferenced
3787 @unnumberedsec Pragma Unreferenced
3788 @findex Unreferenced
3789 @cindex Warnings, unreferenced
3793 @smallexample @c ada
3794 pragma Unreferenced (local_Name @{, local_Name@});
3798 This pragma signals that the entities whose names are listed are
3799 deliberately not referenced in the current source unit. This
3800 suppresses warnings about the
3801 entities being unreferenced, and in addition a warning will be
3802 generated if one of these entities is in fact referenced in the
3803 same unit as the pragma (or in the corresponding body, or one
3806 This is particularly useful for clearly signaling that a particular
3807 parameter is not referenced in some particular subprogram implementation
3808 and that this is deliberate. It can also be useful in the case of
3809 objects declared only for their initialization or finalization side
3812 If @code{local_Name} identifies more than one matching homonym in the
3813 current scope, then the entity most recently declared is the one to which
3816 The left hand side of an assignment does not count as a reference for the
3817 purpose of this pragma. Thus it is fine to assign to an entity for which
3818 pragma Unreferenced is given.
3820 @node Pragma Unreserve_All_Interrupts
3821 @unnumberedsec Pragma Unreserve_All_Interrupts
3822 @findex Unreserve_All_Interrupts
3826 @smallexample @c ada
3827 pragma Unreserve_All_Interrupts;
3831 Normally certain interrupts are reserved to the implementation. Any attempt
3832 to attach an interrupt causes Program_Error to be raised, as described in
3833 RM C.3.2(22). A typical example is the @code{SIGINT} interrupt used in
3834 many systems for a @kbd{Ctrl-C} interrupt. Normally this interrupt is
3835 reserved to the implementation, so that @kbd{Ctrl-C} can be used to
3836 interrupt execution.
3838 If the pragma @code{Unreserve_All_Interrupts} appears anywhere in any unit in
3839 a program, then all such interrupts are unreserved. This allows the
3840 program to handle these interrupts, but disables their standard
3841 functions. For example, if this pragma is used, then pressing
3842 @kbd{Ctrl-C} will not automatically interrupt execution. However,
3843 a program can then handle the @code{SIGINT} interrupt as it chooses.
3845 For a full list of the interrupts handled in a specific implementation,
3846 see the source code for the specification of @code{Ada.Interrupts.Names} in
3847 file @file{a-intnam.ads}. This is a target dependent file that contains the
3848 list of interrupts recognized for a given target. The documentation in
3849 this file also specifies what interrupts are affected by the use of
3850 the @code{Unreserve_All_Interrupts} pragma.
3852 For a more general facility for controlling what interrupts can be
3853 handled, see pragma @code{Interrupt_State}, which subsumes the functionality
3854 of the @code{Unreserve_All_Interrupts} pragma.
3856 @node Pragma Unsuppress
3857 @unnumberedsec Pragma Unsuppress
3862 @smallexample @c ada
3863 pragma Unsuppress (IDENTIFIER [, [On =>] NAME]);
3867 This pragma undoes the effect of a previous pragma @code{Suppress}. If
3868 there is no corresponding pragma @code{Suppress} in effect, it has no
3869 effect. The range of the effect is the same as for pragma
3870 @code{Suppress}. The meaning of the arguments is identical to that used
3871 in pragma @code{Suppress}.
3873 One important application is to ensure that checks are on in cases where
3874 code depends on the checks for its correct functioning, so that the code
3875 will compile correctly even if the compiler switches are set to suppress
3878 @node Pragma Use_VADS_Size
3879 @unnumberedsec Pragma Use_VADS_Size
3880 @cindex @code{Size}, VADS compatibility
3881 @findex Use_VADS_Size
3885 @smallexample @c ada
3886 pragma Use_VADS_Size;
3890 This is a configuration pragma. In a unit to which it applies, any use
3891 of the 'Size attribute is automatically interpreted as a use of the
3892 'VADS_Size attribute. Note that this may result in incorrect semantic
3893 processing of valid Ada 95 programs. This is intended to aid in the
3894 handling of legacy code which depends on the interpretation of Size
3895 as implemented in the VADS compiler. See description of the VADS_Size
3896 attribute for further details.
3898 @node Pragma Validity_Checks
3899 @unnumberedsec Pragma Validity_Checks
3900 @findex Validity_Checks
3904 @smallexample @c ada
3905 pragma Validity_Checks (string_LITERAL | ALL_CHECKS | On | Off);
3909 This pragma is used in conjunction with compiler switches to control the
3910 built-in validity checking provided by GNAT@. The compiler switches, if set
3911 provide an initial setting for the switches, and this pragma may be used
3912 to modify these settings, or the settings may be provided entirely by
3913 the use of the pragma. This pragma can be used anywhere that a pragma
3914 is legal, including use as a configuration pragma (including use in
3915 the @file{gnat.adc} file).
3917 The form with a string literal specifies which validity options are to be
3918 activated. The validity checks are first set to include only the default
3919 reference manual settings, and then a string of letters in the string
3920 specifies the exact set of options required. The form of this string
3921 is exactly as described for the @code{-gnatVx} compiler switch (see the
3922 GNAT users guide for details). For example the following two methods
3923 can be used to enable validity checking for mode @code{in} and
3924 @code{in out} subprogram parameters:
3928 @smallexample @c ada
3929 pragma Validity_Checks ("im");
3934 gcc -c -gnatVim @dots{}
3939 The form ALL_CHECKS activates all standard checks (its use is equivalent
3940 to the use of the @code{gnatva} switch.
3942 The forms with @code{Off} and @code{On}
3943 can be used to temporarily disable validity checks
3944 as shown in the following example:
3946 @smallexample @c ada
3950 pragma Validity_Checks ("c"); -- validity checks for copies
3951 pragma Validity_Checks (Off); -- turn off validity checks
3952 A := B; -- B will not be validity checked
3953 pragma Validity_Checks (On); -- turn validity checks back on
3954 A := C; -- C will be validity checked
3957 @node Pragma Volatile
3958 @unnumberedsec Pragma Volatile
3963 @smallexample @c ada
3964 pragma Volatile (local_NAME);
3968 This pragma is defined by the Ada 95 Reference Manual, and the GNAT
3969 implementation is fully conformant with this definition. The reason it
3970 is mentioned in this section is that a pragma of the same name was supplied
3971 in some Ada 83 compilers, including DEC Ada 83. The Ada 95 implementation
3972 of pragma Volatile is upwards compatible with the implementation in
3975 @node Pragma Warnings
3976 @unnumberedsec Pragma Warnings
3981 @smallexample @c ada
3982 pragma Warnings (On | Off [, LOCAL_NAME]);
3986 Normally warnings are enabled, with the output being controlled by
3987 the command line switch. Warnings (@code{Off}) turns off generation of
3988 warnings until a Warnings (@code{On}) is encountered or the end of the
3989 current unit. If generation of warnings is turned off using this
3990 pragma, then no warning messages are output, regardless of the
3991 setting of the command line switches.
3993 The form with a single argument is a configuration pragma.
3995 If the @var{local_name} parameter is present, warnings are suppressed for
3996 the specified entity. This suppression is effective from the point where
3997 it occurs till the end of the extended scope of the variable (similar to
3998 the scope of @code{Suppress}).
4000 @node Pragma Weak_External
4001 @unnumberedsec Pragma Weak_External
4002 @findex Weak_External
4006 @smallexample @c ada
4007 pragma Weak_External ([Entity =>] LOCAL_NAME);
4011 This pragma specifies that the given entity should be marked as a weak
4012 external (one that does not have to be resolved) for the linker. For
4013 further details, consult the GCC manual.
4015 @node Implementation Defined Attributes
4016 @chapter Implementation Defined Attributes
4017 Ada 95 defines (throughout the Ada 95 reference manual,
4018 summarized in annex K),
4019 a set of attributes that provide useful additional functionality in all
4020 areas of the language. These language defined attributes are implemented
4021 in GNAT and work as described in the Ada 95 Reference Manual.
4023 In addition, Ada 95 allows implementations to define additional
4024 attributes whose meaning is defined by the implementation. GNAT provides
4025 a number of these implementation-dependent attributes which can be used
4026 to extend and enhance the functionality of the compiler. This section of
4027 the GNAT reference manual describes these additional attributes.
4029 Note that any program using these attributes may not be portable to
4030 other compilers (although GNAT implements this set of attributes on all
4031 platforms). Therefore if portability to other compilers is an important
4032 consideration, you should minimize the use of these attributes.
4043 * Default_Bit_Order::
4051 * Has_Access_Values::
4052 * Has_Discriminants::
4058 * Max_Interrupt_Priority::
4060 * Maximum_Alignment::
4064 * Passed_By_Reference::
4075 * Unconstrained_Array::
4076 * Universal_Literal_String::
4077 * Unrestricted_Access::
4085 @unnumberedsec Abort_Signal
4086 @findex Abort_Signal
4088 @code{Standard'Abort_Signal} (@code{Standard} is the only allowed
4089 prefix) provides the entity for the special exception used to signal
4090 task abort or asynchronous transfer of control. Normally this attribute
4091 should only be used in the tasking runtime (it is highly peculiar, and
4092 completely outside the normal semantics of Ada, for a user program to
4093 intercept the abort exception).
4096 @unnumberedsec Address_Size
4097 @cindex Size of @code{Address}
4098 @findex Address_Size
4100 @code{Standard'Address_Size} (@code{Standard} is the only allowed
4101 prefix) is a static constant giving the number of bits in an
4102 @code{Address}. It is the same value as System.Address'Size,
4103 but has the advantage of being static, while a direct
4104 reference to System.Address'Size is non-static because Address
4108 @unnumberedsec Asm_Input
4111 The @code{Asm_Input} attribute denotes a function that takes two
4112 parameters. The first is a string, the second is an expression of the
4113 type designated by the prefix. The first (string) argument is required
4114 to be a static expression, and is the constraint for the parameter,
4115 (e.g.@: what kind of register is required). The second argument is the
4116 value to be used as the input argument. The possible values for the
4117 constant are the same as those used in the RTL, and are dependent on
4118 the configuration file used to built the GCC back end.
4119 @ref{Machine Code Insertions}
4122 @unnumberedsec Asm_Output
4125 The @code{Asm_Output} attribute denotes a function that takes two
4126 parameters. The first is a string, the second is the name of a variable
4127 of the type designated by the attribute prefix. The first (string)
4128 argument is required to be a static expression and designates the
4129 constraint for the parameter (e.g.@: what kind of register is
4130 required). The second argument is the variable to be updated with the
4131 result. The possible values for constraint are the same as those used in
4132 the RTL, and are dependent on the configuration file used to build the
4133 GCC back end. If there are no output operands, then this argument may
4134 either be omitted, or explicitly given as @code{No_Output_Operands}.
4135 @ref{Machine Code Insertions}
4138 @unnumberedsec AST_Entry
4142 This attribute is implemented only in OpenVMS versions of GNAT@. Applied to
4143 the name of an entry, it yields a value of the predefined type AST_Handler
4144 (declared in the predefined package System, as extended by the use of
4145 pragma @code{Extend_System (Aux_DEC)}). This value enables the given entry to
4146 be called when an AST occurs. For further details, refer to the @cite{DEC Ada
4147 Language Reference Manual}, section 9.12a.
4152 @code{@var{obj}'Bit}, where @var{obj} is any object, yields the bit
4153 offset within the storage unit (byte) that contains the first bit of
4154 storage allocated for the object. The value of this attribute is of the
4155 type @code{Universal_Integer}, and is always a non-negative number not
4156 exceeding the value of @code{System.Storage_Unit}.
4158 For an object that is a variable or a constant allocated in a register,
4159 the value is zero. (The use of this attribute does not force the
4160 allocation of a variable to memory).
4162 For an object that is a formal parameter, this attribute applies
4163 to either the matching actual parameter or to a copy of the
4164 matching actual parameter.
4166 For an access object the value is zero. Note that
4167 @code{@var{obj}.all'Bit} is subject to an @code{Access_Check} for the
4168 designated object. Similarly for a record component
4169 @code{@var{X}.@var{C}'Bit} is subject to a discriminant check and
4170 @code{@var{X}(@var{I}).Bit} and @code{@var{X}(@var{I1}..@var{I2})'Bit}
4171 are subject to index checks.
4173 This attribute is designed to be compatible with the DEC Ada 83 definition
4174 and implementation of the @code{Bit} attribute.
4177 @unnumberedsec Bit_Position
4178 @findex Bit_Position
4180 @code{@var{R.C}'Bit}, where @var{R} is a record object and C is one
4181 of the fields of the record type, yields the bit
4182 offset within the record contains the first bit of
4183 storage allocated for the object. The value of this attribute is of the
4184 type @code{Universal_Integer}. The value depends only on the field
4185 @var{C} and is independent of the alignment of
4186 the containing record @var{R}.
4189 @unnumberedsec Code_Address
4190 @findex Code_Address
4191 @cindex Subprogram address
4192 @cindex Address of subprogram code
4195 attribute may be applied to subprograms in Ada 95, but the
4196 intended effect from the Ada 95 reference manual seems to be to provide
4197 an address value which can be used to call the subprogram by means of
4198 an address clause as in the following example:
4200 @smallexample @c ada
4201 procedure K is @dots{}
4204 for L'Address use K'Address;
4205 pragma Import (Ada, L);
4209 A call to @code{L} is then expected to result in a call to @code{K}@.
4210 In Ada 83, where there were no access-to-subprogram values, this was
4211 a common work around for getting the effect of an indirect call.
4212 GNAT implements the above use of @code{Address} and the technique
4213 illustrated by the example code works correctly.
4215 However, for some purposes, it is useful to have the address of the start
4216 of the generated code for the subprogram. On some architectures, this is
4217 not necessarily the same as the @code{Address} value described above.
4218 For example, the @code{Address} value may reference a subprogram
4219 descriptor rather than the subprogram itself.
4221 The @code{'Code_Address} attribute, which can only be applied to
4222 subprogram entities, always returns the address of the start of the
4223 generated code of the specified subprogram, which may or may not be
4224 the same value as is returned by the corresponding @code{'Address}
4227 @node Default_Bit_Order
4228 @unnumberedsec Default_Bit_Order
4230 @cindex Little endian
4231 @findex Default_Bit_Order
4233 @code{Standard'Default_Bit_Order} (@code{Standard} is the only
4234 permissible prefix), provides the value @code{System.Default_Bit_Order}
4235 as a @code{Pos} value (0 for @code{High_Order_First}, 1 for
4236 @code{Low_Order_First}). This is used to construct the definition of
4237 @code{Default_Bit_Order} in package @code{System}.
4240 @unnumberedsec Elaborated
4243 The prefix of the @code{'Elaborated} attribute must be a unit name. The
4244 value is a Boolean which indicates whether or not the given unit has been
4245 elaborated. This attribute is primarily intended for internal use by the
4246 generated code for dynamic elaboration checking, but it can also be used
4247 in user programs. The value will always be True once elaboration of all
4248 units has been completed. An exception is for units which need no
4249 elaboration, the value is always False for such units.
4252 @unnumberedsec Elab_Body
4255 This attribute can only be applied to a program unit name. It returns
4256 the entity for the corresponding elaboration procedure for elaborating
4257 the body of the referenced unit. This is used in the main generated
4258 elaboration procedure by the binder and is not normally used in any
4259 other context. However, there may be specialized situations in which it
4260 is useful to be able to call this elaboration procedure from Ada code,
4261 e.g.@: if it is necessary to do selective re-elaboration to fix some
4265 @unnumberedsec Elab_Spec
4268 This attribute can only be applied to a program unit name. It returns
4269 the entity for the corresponding elaboration procedure for elaborating
4270 the specification of the referenced unit. This is used in the main
4271 generated elaboration procedure by the binder and is not normally used
4272 in any other context. However, there may be specialized situations in
4273 which it is useful to be able to call this elaboration procedure from
4274 Ada code, e.g.@: if it is necessary to do selective re-elaboration to fix
4279 @cindex Ada 83 attributes
4282 The @code{Emax} attribute is provided for compatibility with Ada 83. See
4283 the Ada 83 reference manual for an exact description of the semantics of
4287 @unnumberedsec Enum_Rep
4288 @cindex Representation of enums
4291 For every enumeration subtype @var{S}, @code{@var{S}'Enum_Rep} denotes a
4292 function with the following spec:
4294 @smallexample @c ada
4295 function @var{S}'Enum_Rep (Arg : @var{S}'Base)
4296 return @i{Universal_Integer};
4300 It is also allowable to apply @code{Enum_Rep} directly to an object of an
4301 enumeration type or to a non-overloaded enumeration
4302 literal. In this case @code{@var{S}'Enum_Rep} is equivalent to
4303 @code{@var{typ}'Enum_Rep(@var{S})} where @var{typ} is the type of the
4304 enumeration literal or object.
4306 The function returns the representation value for the given enumeration
4307 value. This will be equal to value of the @code{Pos} attribute in the
4308 absence of an enumeration representation clause. This is a static
4309 attribute (i.e.@: the result is static if the argument is static).
4311 @code{@var{S}'Enum_Rep} can also be used with integer types and objects,
4312 in which case it simply returns the integer value. The reason for this
4313 is to allow it to be used for @code{(<>)} discrete formal arguments in
4314 a generic unit that can be instantiated with either enumeration types
4315 or integer types. Note that if @code{Enum_Rep} is used on a modular
4316 type whose upper bound exceeds the upper bound of the largest signed
4317 integer type, and the argument is a variable, so that the universal
4318 integer calculation is done at run-time, then the call to @code{Enum_Rep}
4319 may raise @code{Constraint_Error}.
4322 @unnumberedsec Epsilon
4323 @cindex Ada 83 attributes
4326 The @code{Epsilon} attribute is provided for compatibility with Ada 83. See
4327 the Ada 83 reference manual for an exact description of the semantics of
4331 @unnumberedsec Fixed_Value
4334 For every fixed-point type @var{S}, @code{@var{S}'Fixed_Value} denotes a
4335 function with the following specification:
4337 @smallexample @c ada
4338 function @var{S}'Fixed_Value (Arg : @i{Universal_Integer})
4343 The value returned is the fixed-point value @var{V} such that
4345 @smallexample @c ada
4346 @var{V} = Arg * @var{S}'Small
4350 The effect is thus similar to first converting the argument to the
4351 integer type used to represent @var{S}, and then doing an unchecked
4352 conversion to the fixed-point type. The difference is
4353 that there are full range checks, to ensure that the result is in range.
4354 This attribute is primarily intended for use in implementation of the
4355 input-output functions for fixed-point values.
4357 @node Has_Access_Values
4358 @unnumberedsec Has_Access_Values
4359 @cindex Access values, testing for
4360 @findex Has_Access_Values
4362 The prefix of the @code{Has_Access_Values} attribute is a type. The result
4363 is a Boolean value which is True if the is an access type, or is a composite
4364 type with a component (at any nesting depth) that is an access type, and is
4366 The intended use of this attribute is in conjunction with generic
4367 definitions. If the attribute is applied to a generic private type, it
4368 indicates whether or not the corresponding actual type has access values.
4370 @node Has_Discriminants
4371 @unnumberedsec Has_Discriminants
4372 @cindex Discriminants, testing for
4373 @findex Has_Discriminants
4375 The prefix of the @code{Has_Discriminants} attribute is a type. The result
4376 is a Boolean value which is True if the type has discriminants, and False
4377 otherwise. The intended use of this attribute is in conjunction with generic
4378 definitions. If the attribute is applied to a generic private type, it
4379 indicates whether or not the corresponding actual type has discriminants.
4385 The @code{Img} attribute differs from @code{Image} in that it may be
4386 applied to objects as well as types, in which case it gives the
4387 @code{Image} for the subtype of the object. This is convenient for
4390 @smallexample @c ada
4391 Put_Line ("X = " & X'Img);
4395 has the same meaning as the more verbose:
4397 @smallexample @c ada
4398 Put_Line ("X = " & @var{T}'Image (X));
4402 where @var{T} is the (sub)type of the object @code{X}.
4405 @unnumberedsec Integer_Value
4406 @findex Integer_Value
4408 For every integer type @var{S}, @code{@var{S}'Integer_Value} denotes a
4409 function with the following spec:
4411 @smallexample @c ada
4412 function @var{S}'Integer_Value (Arg : @i{Universal_Fixed})
4417 The value returned is the integer value @var{V}, such that
4419 @smallexample @c ada
4420 Arg = @var{V} * @var{T}'Small
4424 where @var{T} is the type of @code{Arg}.
4425 The effect is thus similar to first doing an unchecked conversion from
4426 the fixed-point type to its corresponding implementation type, and then
4427 converting the result to the target integer type. The difference is
4428 that there are full range checks, to ensure that the result is in range.
4429 This attribute is primarily intended for use in implementation of the
4430 standard input-output functions for fixed-point values.
4433 @unnumberedsec Large
4434 @cindex Ada 83 attributes
4437 The @code{Large} attribute is provided for compatibility with Ada 83. See
4438 the Ada 83 reference manual for an exact description of the semantics of
4442 @unnumberedsec Machine_Size
4443 @findex Machine_Size
4445 This attribute is identical to the @code{Object_Size} attribute. It is
4446 provided for compatibility with the DEC Ada 83 attribute of this name.
4449 @unnumberedsec Mantissa
4450 @cindex Ada 83 attributes
4453 The @code{Mantissa} attribute is provided for compatibility with Ada 83. See
4454 the Ada 83 reference manual for an exact description of the semantics of
4457 @node Max_Interrupt_Priority
4458 @unnumberedsec Max_Interrupt_Priority
4459 @cindex Interrupt priority, maximum
4460 @findex Max_Interrupt_Priority
4462 @code{Standard'Max_Interrupt_Priority} (@code{Standard} is the only
4463 permissible prefix), provides the same value as
4464 @code{System.Max_Interrupt_Priority}.
4467 @unnumberedsec Max_Priority
4468 @cindex Priority, maximum
4469 @findex Max_Priority
4471 @code{Standard'Max_Priority} (@code{Standard} is the only permissible
4472 prefix) provides the same value as @code{System.Max_Priority}.
4474 @node Maximum_Alignment
4475 @unnumberedsec Maximum_Alignment
4476 @cindex Alignment, maximum
4477 @findex Maximum_Alignment
4479 @code{Standard'Maximum_Alignment} (@code{Standard} is the only
4480 permissible prefix) provides the maximum useful alignment value for the
4481 target. This is a static value that can be used to specify the alignment
4482 for an object, guaranteeing that it is properly aligned in all
4485 @node Mechanism_Code
4486 @unnumberedsec Mechanism_Code
4487 @cindex Return values, passing mechanism
4488 @cindex Parameters, passing mechanism
4489 @findex Mechanism_Code
4491 @code{@var{function}'Mechanism_Code} yields an integer code for the
4492 mechanism used for the result of function, and
4493 @code{@var{subprogram}'Mechanism_Code (@var{n})} yields the mechanism
4494 used for formal parameter number @var{n} (a static integer value with 1
4495 meaning the first parameter) of @var{subprogram}. The code returned is:
4503 by descriptor (default descriptor class)
4505 by descriptor (UBS: unaligned bit string)
4507 by descriptor (UBSB: aligned bit string with arbitrary bounds)
4509 by descriptor (UBA: unaligned bit array)
4511 by descriptor (S: string, also scalar access type parameter)
4513 by descriptor (SB: string with arbitrary bounds)
4515 by descriptor (A: contiguous array)
4517 by descriptor (NCA: non-contiguous array)
4521 Values from 3 through 10 are only relevant to Digital OpenVMS implementations.
4524 @node Null_Parameter
4525 @unnumberedsec Null_Parameter
4526 @cindex Zero address, passing
4527 @findex Null_Parameter
4529 A reference @code{@var{T}'Null_Parameter} denotes an imaginary object of
4530 type or subtype @var{T} allocated at machine address zero. The attribute
4531 is allowed only as the default expression of a formal parameter, or as
4532 an actual expression of a subprogram call. In either case, the
4533 subprogram must be imported.
4535 The identity of the object is represented by the address zero in the
4536 argument list, independent of the passing mechanism (explicit or
4539 This capability is needed to specify that a zero address should be
4540 passed for a record or other composite object passed by reference.
4541 There is no way of indicating this without the @code{Null_Parameter}
4545 @unnumberedsec Object_Size
4546 @cindex Size, used for objects
4549 The size of an object is not necessarily the same as the size of the type
4550 of an object. This is because by default object sizes are increased to be
4551 a multiple of the alignment of the object. For example,
4552 @code{Natural'Size} is
4553 31, but by default objects of type @code{Natural} will have a size of 32 bits.
4554 Similarly, a record containing an integer and a character:
4556 @smallexample @c ada
4564 will have a size of 40 (that is @code{Rec'Size} will be 40. The
4565 alignment will be 4, because of the
4566 integer field, and so the default size of record objects for this type
4567 will be 64 (8 bytes).
4569 The @code{@var{type}'Object_Size} attribute
4570 has been added to GNAT to allow the
4571 default object size of a type to be easily determined. For example,
4572 @code{Natural'Object_Size} is 32, and
4573 @code{Rec'Object_Size} (for the record type in the above example) will be
4574 64. Note also that, unlike the situation with the
4575 @code{Size} attribute as defined in the Ada RM, the
4576 @code{Object_Size} attribute can be specified individually
4577 for different subtypes. For example:
4579 @smallexample @c ada
4580 type R is new Integer;
4581 subtype R1 is R range 1 .. 10;
4582 subtype R2 is R range 1 .. 10;
4583 for R2'Object_Size use 8;
4587 In this example, @code{R'Object_Size} and @code{R1'Object_Size} are both
4588 32 since the default object size for a subtype is the same as the object size
4589 for the parent subtype. This means that objects of type @code{R}
4591 by default be 32 bits (four bytes). But objects of type
4592 @code{R2} will be only
4593 8 bits (one byte), since @code{R2'Object_Size} has been set to 8.
4595 @node Passed_By_Reference
4596 @unnumberedsec Passed_By_Reference
4597 @cindex Parameters, when passed by reference
4598 @findex Passed_By_Reference
4600 @code{@var{type}'Passed_By_Reference} for any subtype @var{type} returns
4601 a value of type @code{Boolean} value that is @code{True} if the type is
4602 normally passed by reference and @code{False} if the type is normally
4603 passed by copy in calls. For scalar types, the result is always @code{False}
4604 and is static. For non-scalar types, the result is non-static.
4607 @unnumberedsec Range_Length
4608 @findex Range_Length
4610 @code{@var{type}'Range_Length} for any discrete type @var{type} yields
4611 the number of values represented by the subtype (zero for a null
4612 range). The result is static for static subtypes. @code{Range_Length}
4613 applied to the index subtype of a one dimensional array always gives the
4614 same result as @code{Range} applied to the array itself.
4617 @unnumberedsec Safe_Emax
4618 @cindex Ada 83 attributes
4621 The @code{Safe_Emax} attribute is provided for compatibility with Ada 83. See
4622 the Ada 83 reference manual for an exact description of the semantics of
4626 @unnumberedsec Safe_Large
4627 @cindex Ada 83 attributes
4630 The @code{Safe_Large} attribute is provided for compatibility with Ada 83. See
4631 the Ada 83 reference manual for an exact description of the semantics of
4635 @unnumberedsec Small
4636 @cindex Ada 83 attributes
4639 The @code{Small} attribute is defined in Ada 95 only for fixed-point types.
4640 GNAT also allows this attribute to be applied to floating-point types
4641 for compatibility with Ada 83. See
4642 the Ada 83 reference manual for an exact description of the semantics of
4643 this attribute when applied to floating-point types.
4646 @unnumberedsec Storage_Unit
4647 @findex Storage_Unit
4649 @code{Standard'Storage_Unit} (@code{Standard} is the only permissible
4650 prefix) provides the same value as @code{System.Storage_Unit}.
4653 @unnumberedsec Target_Name
4656 @code{Standard'Target_Name} (@code{Standard} is the only permissible
4657 prefix) provides a static string value that identifies the target
4658 for the current compilation. For GCC implementations, this is the
4659 standard gcc target name without the terminating slash (for
4660 example, GNAT 5.0 on windows yields "i586-pc-mingw32msv").
4666 @code{Standard'Tick} (@code{Standard} is the only permissible prefix)
4667 provides the same value as @code{System.Tick},
4670 @unnumberedsec To_Address
4673 The @code{System'To_Address}
4674 (@code{System} is the only permissible prefix)
4675 denotes a function identical to
4676 @code{System.Storage_Elements.To_Address} except that
4677 it is a static attribute. This means that if its argument is
4678 a static expression, then the result of the attribute is a
4679 static expression. The result is that such an expression can be
4680 used in contexts (e.g.@: preelaborable packages) which require a
4681 static expression and where the function call could not be used
4682 (since the function call is always non-static, even if its
4683 argument is static).
4686 @unnumberedsec Type_Class
4689 @code{@var{type}'Type_Class} for any type or subtype @var{type} yields
4690 the value of the type class for the full type of @var{type}. If
4691 @var{type} is a generic formal type, the value is the value for the
4692 corresponding actual subtype. The value of this attribute is of type
4693 @code{System.Aux_DEC.Type_Class}, which has the following definition:
4695 @smallexample @c ada
4697 (Type_Class_Enumeration,
4699 Type_Class_Fixed_Point,
4700 Type_Class_Floating_Point,
4705 Type_Class_Address);
4709 Protected types yield the value @code{Type_Class_Task}, which thus
4710 applies to all concurrent types. This attribute is designed to
4711 be compatible with the DEC Ada 83 attribute of the same name.
4714 @unnumberedsec UET_Address
4717 The @code{UET_Address} attribute can only be used for a prefix which
4718 denotes a library package. It yields the address of the unit exception
4719 table when zero cost exception handling is used. This attribute is
4720 intended only for use within the GNAT implementation. See the unit
4721 @code{Ada.Exceptions} in files @file{a-except.ads} and @file{a-except.adb}
4722 for details on how this attribute is used in the implementation.
4724 @node Unconstrained_Array
4725 @unnumberedsec Unconstrained_Array
4726 @findex Unconstrained_Array
4728 The @code{Unconstrained_Array} attribute can be used with a prefix that
4729 denotes any type or subtype. It is a static attribute that yields
4730 @code{True} if the prefix designates an unconstrained array,
4731 and @code{False} otherwise. In a generic instance, the result is
4732 still static, and yields the result of applying this test to the
4735 @node Universal_Literal_String
4736 @unnumberedsec Universal_Literal_String
4737 @cindex Named numbers, representation of
4738 @findex Universal_Literal_String
4740 The prefix of @code{Universal_Literal_String} must be a named
4741 number. The static result is the string consisting of the characters of
4742 the number as defined in the original source. This allows the user
4743 program to access the actual text of named numbers without intermediate
4744 conversions and without the need to enclose the strings in quotes (which
4745 would preclude their use as numbers). This is used internally for the
4746 construction of values of the floating-point attributes from the file
4747 @file{ttypef.ads}, but may also be used by user programs.
4749 @node Unrestricted_Access
4750 @unnumberedsec Unrestricted_Access
4751 @cindex @code{Access}, unrestricted
4752 @findex Unrestricted_Access
4754 The @code{Unrestricted_Access} attribute is similar to @code{Access}
4755 except that all accessibility and aliased view checks are omitted. This
4756 is a user-beware attribute. It is similar to
4757 @code{Address}, for which it is a desirable replacement where the value
4758 desired is an access type. In other words, its effect is identical to
4759 first applying the @code{Address} attribute and then doing an unchecked
4760 conversion to a desired access type. In GNAT, but not necessarily in
4761 other implementations, the use of static chains for inner level
4762 subprograms means that @code{Unrestricted_Access} applied to a
4763 subprogram yields a value that can be called as long as the subprogram
4764 is in scope (normal Ada 95 accessibility rules restrict this usage).
4766 It is possible to use @code{Unrestricted_Access} for any type, but care
4767 must be excercised if it is used to create pointers to unconstrained
4768 objects. In this case, the resulting pointer has the same scope as the
4769 context of the attribute, and may not be returned to some enclosing
4770 scope. For instance, a function cannot use @code{Unrestricted_Access}
4771 to create a unconstrained pointer and then return that value to the
4775 @unnumberedsec VADS_Size
4776 @cindex @code{Size}, VADS compatibility
4779 The @code{'VADS_Size} attribute is intended to make it easier to port
4780 legacy code which relies on the semantics of @code{'Size} as implemented
4781 by the VADS Ada 83 compiler. GNAT makes a best effort at duplicating the
4782 same semantic interpretation. In particular, @code{'VADS_Size} applied
4783 to a predefined or other primitive type with no Size clause yields the
4784 Object_Size (for example, @code{Natural'Size} is 32 rather than 31 on
4785 typical machines). In addition @code{'VADS_Size} applied to an object
4786 gives the result that would be obtained by applying the attribute to
4787 the corresponding type.
4790 @unnumberedsec Value_Size
4791 @cindex @code{Size}, setting for not-first subtype
4793 @code{@var{type}'Value_Size} is the number of bits required to represent
4794 a value of the given subtype. It is the same as @code{@var{type}'Size},
4795 but, unlike @code{Size}, may be set for non-first subtypes.
4798 @unnumberedsec Wchar_T_Size
4799 @findex Wchar_T_Size
4800 @code{Standard'Wchar_T_Size} (@code{Standard} is the only permissible
4801 prefix) provides the size in bits of the C @code{wchar_t} type
4802 primarily for constructing the definition of this type in
4803 package @code{Interfaces.C}.
4806 @unnumberedsec Word_Size
4808 @code{Standard'Word_Size} (@code{Standard} is the only permissible
4809 prefix) provides the value @code{System.Word_Size}.
4811 @c ------------------------
4812 @node Implementation Advice
4813 @chapter Implementation Advice
4815 The main text of the Ada 95 Reference Manual describes the required
4816 behavior of all Ada 95 compilers, and the GNAT compiler conforms to
4819 In addition, there are sections throughout the Ada 95
4820 reference manual headed
4821 by the phrase ``implementation advice''. These sections are not normative,
4822 i.e.@: they do not specify requirements that all compilers must
4823 follow. Rather they provide advice on generally desirable behavior. You
4824 may wonder why they are not requirements. The most typical answer is
4825 that they describe behavior that seems generally desirable, but cannot
4826 be provided on all systems, or which may be undesirable on some systems.
4828 As far as practical, GNAT follows the implementation advice sections in
4829 the Ada 95 Reference Manual. This chapter contains a table giving the
4830 reference manual section number, paragraph number and several keywords
4831 for each advice. Each entry consists of the text of the advice followed
4832 by the GNAT interpretation of this advice. Most often, this simply says
4833 ``followed'', which means that GNAT follows the advice. However, in a
4834 number of cases, GNAT deliberately deviates from this advice, in which
4835 case the text describes what GNAT does and why.
4837 @cindex Error detection
4838 @unnumberedsec 1.1.3(20): Error Detection
4841 If an implementation detects the use of an unsupported Specialized Needs
4842 Annex feature at run time, it should raise @code{Program_Error} if
4845 Not relevant. All specialized needs annex features are either supported,
4846 or diagnosed at compile time.
4849 @unnumberedsec 1.1.3(31): Child Units
4852 If an implementation wishes to provide implementation-defined
4853 extensions to the functionality of a language-defined library unit, it
4854 should normally do so by adding children to the library unit.
4858 @cindex Bounded errors
4859 @unnumberedsec 1.1.5(12): Bounded Errors
4862 If an implementation detects a bounded error or erroneous
4863 execution, it should raise @code{Program_Error}.
4865 Followed in all cases in which the implementation detects a bounded
4866 error or erroneous execution. Not all such situations are detected at
4870 @unnumberedsec 2.8(16): Pragmas
4873 Normally, implementation-defined pragmas should have no semantic effect
4874 for error-free programs; that is, if the implementation-defined pragmas
4875 are removed from a working program, the program should still be legal,
4876 and should still have the same semantics.
4878 The following implementation defined pragmas are exceptions to this
4890 @item CPP_Constructor
4898 @item Interface_Name
4900 @item Machine_Attribute
4902 @item Unimplemented_Unit
4904 @item Unchecked_Union
4909 In each of the above cases, it is essential to the purpose of the pragma
4910 that this advice not be followed. For details see the separate section
4911 on implementation defined pragmas.
4913 @unnumberedsec 2.8(17-19): Pragmas
4916 Normally, an implementation should not define pragmas that can
4917 make an illegal program legal, except as follows:
4921 A pragma used to complete a declaration, such as a pragma @code{Import};
4925 A pragma used to configure the environment by adding, removing, or
4926 replacing @code{library_items}.
4928 See response to paragraph 16 of this same section.
4930 @cindex Character Sets
4931 @cindex Alternative Character Sets
4932 @unnumberedsec 3.5.2(5): Alternative Character Sets
4935 If an implementation supports a mode with alternative interpretations
4936 for @code{Character} and @code{Wide_Character}, the set of graphic
4937 characters of @code{Character} should nevertheless remain a proper
4938 subset of the set of graphic characters of @code{Wide_Character}. Any
4939 character set ``localizations'' should be reflected in the results of
4940 the subprograms defined in the language-defined package
4941 @code{Characters.Handling} (see A.3) available in such a mode. In a mode with
4942 an alternative interpretation of @code{Character}, the implementation should
4943 also support a corresponding change in what is a legal
4944 @code{identifier_letter}.
4946 Not all wide character modes follow this advice, in particular the JIS
4947 and IEC modes reflect standard usage in Japan, and in these encoding,
4948 the upper half of the Latin-1 set is not part of the wide-character
4949 subset, since the most significant bit is used for wide character
4950 encoding. However, this only applies to the external forms. Internally
4951 there is no such restriction.
4953 @cindex Integer types
4954 @unnumberedsec 3.5.4(28): Integer Types
4958 An implementation should support @code{Long_Integer} in addition to
4959 @code{Integer} if the target machine supports 32-bit (or longer)
4960 arithmetic. No other named integer subtypes are recommended for package
4961 @code{Standard}. Instead, appropriate named integer subtypes should be
4962 provided in the library package @code{Interfaces} (see B.2).
4964 @code{Long_Integer} is supported. Other standard integer types are supported
4965 so this advice is not fully followed. These types
4966 are supported for convenient interface to C, and so that all hardware
4967 types of the machine are easily available.
4968 @unnumberedsec 3.5.4(29): Integer Types
4972 An implementation for a two's complement machine should support
4973 modular types with a binary modulus up to @code{System.Max_Int*2+2}. An
4974 implementation should support a non-binary modules up to @code{Integer'Last}.
4978 @cindex Enumeration values
4979 @unnumberedsec 3.5.5(8): Enumeration Values
4982 For the evaluation of a call on @code{@var{S}'Pos} for an enumeration
4983 subtype, if the value of the operand does not correspond to the internal
4984 code for any enumeration literal of its type (perhaps due to an
4985 un-initialized variable), then the implementation should raise
4986 @code{Program_Error}. This is particularly important for enumeration
4987 types with noncontiguous internal codes specified by an
4988 enumeration_representation_clause.
4993 @unnumberedsec 3.5.7(17): Float Types
4996 An implementation should support @code{Long_Float} in addition to
4997 @code{Float} if the target machine supports 11 or more digits of
4998 precision. No other named floating point subtypes are recommended for
4999 package @code{Standard}. Instead, appropriate named floating point subtypes
5000 should be provided in the library package @code{Interfaces} (see B.2).
5002 @code{Short_Float} and @code{Long_Long_Float} are also provided. The
5003 former provides improved compatibility with other implementations
5004 supporting this type. The latter corresponds to the highest precision
5005 floating-point type supported by the hardware. On most machines, this
5006 will be the same as @code{Long_Float}, but on some machines, it will
5007 correspond to the IEEE extended form. The notable case is all ia32
5008 (x86) implementations, where @code{Long_Long_Float} corresponds to
5009 the 80-bit extended precision format supported in hardware on this
5010 processor. Note that the 128-bit format on SPARC is not supported,
5011 since this is a software rather than a hardware format.
5013 @cindex Multidimensional arrays
5014 @cindex Arrays, multidimensional
5015 @unnumberedsec 3.6.2(11): Multidimensional Arrays
5018 An implementation should normally represent multidimensional arrays in
5019 row-major order, consistent with the notation used for multidimensional
5020 array aggregates (see 4.3.3). However, if a pragma @code{Convention}
5021 (@code{Fortran}, @dots{}) applies to a multidimensional array type, then
5022 column-major order should be used instead (see B.5, ``Interfacing with
5027 @findex Duration'Small
5028 @unnumberedsec 9.6(30-31): Duration'Small
5031 Whenever possible in an implementation, the value of @code{Duration'Small}
5032 should be no greater than 100 microseconds.
5034 Followed. (@code{Duration'Small} = 10**(@minus{}9)).
5038 The time base for @code{delay_relative_statements} should be monotonic;
5039 it need not be the same time base as used for @code{Calendar.Clock}.
5043 @unnumberedsec 10.2.1(12): Consistent Representation
5046 In an implementation, a type declared in a pre-elaborated package should
5047 have the same representation in every elaboration of a given version of
5048 the package, whether the elaborations occur in distinct executions of
5049 the same program, or in executions of distinct programs or partitions
5050 that include the given version.
5052 Followed, except in the case of tagged types. Tagged types involve
5053 implicit pointers to a local copy of a dispatch table, and these pointers
5054 have representations which thus depend on a particular elaboration of the
5055 package. It is not easy to see how it would be possible to follow this
5056 advice without severely impacting efficiency of execution.
5058 @cindex Exception information
5059 @unnumberedsec 11.4.1(19): Exception Information
5062 @code{Exception_Message} by default and @code{Exception_Information}
5063 should produce information useful for
5064 debugging. @code{Exception_Message} should be short, about one
5065 line. @code{Exception_Information} can be long. @code{Exception_Message}
5066 should not include the
5067 @code{Exception_Name}. @code{Exception_Information} should include both
5068 the @code{Exception_Name} and the @code{Exception_Message}.
5070 Followed. For each exception that doesn't have a specified
5071 @code{Exception_Message}, the compiler generates one containing the location
5072 of the raise statement. This location has the form ``file:line'', where
5073 file is the short file name (without path information) and line is the line
5074 number in the file. Note that in the case of the Zero Cost Exception
5075 mechanism, these messages become redundant with the Exception_Information that
5076 contains a full backtrace of the calling sequence, so they are disabled.
5077 To disable explicitly the generation of the source location message, use the
5078 Pragma @code{Discard_Names}.
5080 @cindex Suppression of checks
5081 @cindex Checks, suppression of
5082 @unnumberedsec 11.5(28): Suppression of Checks
5085 The implementation should minimize the code executed for checks that
5086 have been suppressed.
5090 @cindex Representation clauses
5091 @unnumberedsec 13.1 (21-24): Representation Clauses
5094 The recommended level of support for all representation items is
5095 qualified as follows:
5099 An implementation need not support representation items containing
5100 non-static expressions, except that an implementation should support a
5101 representation item for a given entity if each non-static expression in
5102 the representation item is a name that statically denotes a constant
5103 declared before the entity.
5105 Followed. GNAT does not support non-static expressions in representation
5106 clauses unless they are constants declared before the entity. For
5109 @smallexample @c ada
5111 for X'Address use To_address (16#2000#);
5115 will be rejected, since the To_Address expression is non-static. Instead
5118 @smallexample @c ada
5119 X_Address : constant Address : = To_Address (16#2000#);
5121 for X'Address use X_Address;
5126 An implementation need not support a specification for the @code{Size}
5127 for a given composite subtype, nor the size or storage place for an
5128 object (including a component) of a given composite subtype, unless the
5129 constraints on the subtype and its composite subcomponents (if any) are
5130 all static constraints.
5132 Followed. Size Clauses are not permitted on non-static components, as
5137 An aliased component, or a component whose type is by-reference, should
5138 always be allocated at an addressable location.
5142 @cindex Packed types
5143 @unnumberedsec 13.2(6-8): Packed Types
5146 If a type is packed, then the implementation should try to minimize
5147 storage allocated to objects of the type, possibly at the expense of
5148 speed of accessing components, subject to reasonable complexity in
5149 addressing calculations.
5153 The recommended level of support pragma @code{Pack} is:
5155 For a packed record type, the components should be packed as tightly as
5156 possible subject to the Sizes of the component subtypes, and subject to
5157 any @code{record_representation_clause} that applies to the type; the
5158 implementation may, but need not, reorder components or cross aligned
5159 word boundaries to improve the packing. A component whose @code{Size} is
5160 greater than the word size may be allocated an integral number of words.
5162 Followed. Tight packing of arrays is supported for all component sizes
5163 up to 64-bits. If the array component size is 1 (that is to say, if
5164 the component is a boolean type or an enumeration type with two values)
5165 then values of the type are implicitly initialized to zero. This
5166 happens both for objects of the packed type, and for objects that have a
5167 subcomponent of the packed type.
5171 An implementation should support Address clauses for imported
5175 @cindex @code{Address} clauses
5176 @unnumberedsec 13.3(14-19): Address Clauses
5180 For an array @var{X}, @code{@var{X}'Address} should point at the first
5181 component of the array, and not at the array bounds.
5187 The recommended level of support for the @code{Address} attribute is:
5189 @code{@var{X}'Address} should produce a useful result if @var{X} is an
5190 object that is aliased or of a by-reference type, or is an entity whose
5191 @code{Address} has been specified.
5193 Followed. A valid address will be produced even if none of those
5194 conditions have been met. If necessary, the object is forced into
5195 memory to ensure the address is valid.
5199 An implementation should support @code{Address} clauses for imported
5206 Objects (including subcomponents) that are aliased or of a by-reference
5207 type should be allocated on storage element boundaries.
5213 If the @code{Address} of an object is specified, or it is imported or exported,
5214 then the implementation should not perform optimizations based on
5215 assumptions of no aliases.
5219 @cindex @code{Alignment} clauses
5220 @unnumberedsec 13.3(29-35): Alignment Clauses
5223 The recommended level of support for the @code{Alignment} attribute for
5226 An implementation should support specified Alignments that are factors
5227 and multiples of the number of storage elements per word, subject to the
5234 An implementation need not support specified @code{Alignment}s for
5235 combinations of @code{Size}s and @code{Alignment}s that cannot be easily
5236 loaded and stored by available machine instructions.
5242 An implementation need not support specified @code{Alignment}s that are
5243 greater than the maximum @code{Alignment} the implementation ever returns by
5250 The recommended level of support for the @code{Alignment} attribute for
5253 Same as above, for subtypes, but in addition:
5259 For stand-alone library-level objects of statically constrained
5260 subtypes, the implementation should support all @code{Alignment}s
5261 supported by the target linker. For example, page alignment is likely to
5262 be supported for such objects, but not for subtypes.
5266 @cindex @code{Size} clauses
5267 @unnumberedsec 13.3(42-43): Size Clauses
5270 The recommended level of support for the @code{Size} attribute of
5273 A @code{Size} clause should be supported for an object if the specified
5274 @code{Size} is at least as large as its subtype's @code{Size}, and
5275 corresponds to a size in storage elements that is a multiple of the
5276 object's @code{Alignment} (if the @code{Alignment} is nonzero).
5280 @unnumberedsec 13.3(50-56): Size Clauses
5283 If the @code{Size} of a subtype is specified, and allows for efficient
5284 independent addressability (see 9.10) on the target architecture, then
5285 the @code{Size} of the following objects of the subtype should equal the
5286 @code{Size} of the subtype:
5288 Aliased objects (including components).
5294 @code{Size} clause on a composite subtype should not affect the
5295 internal layout of components.
5301 The recommended level of support for the @code{Size} attribute of subtypes is:
5305 The @code{Size} (if not specified) of a static discrete or fixed point
5306 subtype should be the number of bits needed to represent each value
5307 belonging to the subtype using an unbiased representation, leaving space
5308 for a sign bit only if the subtype contains negative values. If such a
5309 subtype is a first subtype, then an implementation should support a
5310 specified @code{Size} for it that reflects this representation.
5316 For a subtype implemented with levels of indirection, the @code{Size}
5317 should include the size of the pointers, but not the size of what they
5322 @cindex @code{Component_Size} clauses
5323 @unnumberedsec 13.3(71-73): Component Size Clauses
5326 The recommended level of support for the @code{Component_Size}
5331 An implementation need not support specified @code{Component_Sizes} that are
5332 less than the @code{Size} of the component subtype.
5338 An implementation should support specified @code{Component_Size}s that
5339 are factors and multiples of the word size. For such
5340 @code{Component_Size}s, the array should contain no gaps between
5341 components. For other @code{Component_Size}s (if supported), the array
5342 should contain no gaps between components when packing is also
5343 specified; the implementation should forbid this combination in cases
5344 where it cannot support a no-gaps representation.
5348 @cindex Enumeration representation clauses
5349 @cindex Representation clauses, enumeration
5350 @unnumberedsec 13.4(9-10): Enumeration Representation Clauses
5353 The recommended level of support for enumeration representation clauses
5356 An implementation need not support enumeration representation clauses
5357 for boolean types, but should at minimum support the internal codes in
5358 the range @code{System.Min_Int.System.Max_Int}.
5362 @cindex Record representation clauses
5363 @cindex Representation clauses, records
5364 @unnumberedsec 13.5.1(17-22): Record Representation Clauses
5367 The recommended level of support for
5368 @*@code{record_representation_clauses} is:
5370 An implementation should support storage places that can be extracted
5371 with a load, mask, shift sequence of machine code, and set with a load,
5372 shift, mask, store sequence, given the available machine instructions
5379 A storage place should be supported if its size is equal to the
5380 @code{Size} of the component subtype, and it starts and ends on a
5381 boundary that obeys the @code{Alignment} of the component subtype.
5387 If the default bit ordering applies to the declaration of a given type,
5388 then for a component whose subtype's @code{Size} is less than the word
5389 size, any storage place that does not cross an aligned word boundary
5390 should be supported.
5396 An implementation may reserve a storage place for the tag field of a
5397 tagged type, and disallow other components from overlapping that place.
5399 Followed. The storage place for the tag field is the beginning of the tagged
5400 record, and its size is Address'Size. GNAT will reject an explicit component
5401 clause for the tag field.
5405 An implementation need not support a @code{component_clause} for a
5406 component of an extension part if the storage place is not after the
5407 storage places of all components of the parent type, whether or not
5408 those storage places had been specified.
5410 Followed. The above advice on record representation clauses is followed,
5411 and all mentioned features are implemented.
5413 @cindex Storage place attributes
5414 @unnumberedsec 13.5.2(5): Storage Place Attributes
5417 If a component is represented using some form of pointer (such as an
5418 offset) to the actual data of the component, and this data is contiguous
5419 with the rest of the object, then the storage place attributes should
5420 reflect the place of the actual data, not the pointer. If a component is
5421 allocated discontinuously from the rest of the object, then a warning
5422 should be generated upon reference to one of its storage place
5425 Followed. There are no such components in GNAT@.
5427 @cindex Bit ordering
5428 @unnumberedsec 13.5.3(7-8): Bit Ordering
5431 The recommended level of support for the non-default bit ordering is:
5435 If @code{Word_Size} = @code{Storage_Unit}, then the implementation
5436 should support the non-default bit ordering in addition to the default
5439 Followed. Word size does not equal storage size in this implementation.
5440 Thus non-default bit ordering is not supported.
5442 @cindex @code{Address}, as private type
5443 @unnumberedsec 13.7(37): Address as Private
5446 @code{Address} should be of a private type.
5450 @cindex Operations, on @code{Address}
5451 @cindex @code{Address}, operations of
5452 @unnumberedsec 13.7.1(16): Address Operations
5455 Operations in @code{System} and its children should reflect the target
5456 environment semantics as closely as is reasonable. For example, on most
5457 machines, it makes sense for address arithmetic to ``wrap around''.
5458 Operations that do not make sense should raise @code{Program_Error}.
5460 Followed. Address arithmetic is modular arithmetic that wraps around. No
5461 operation raises @code{Program_Error}, since all operations make sense.
5463 @cindex Unchecked conversion
5464 @unnumberedsec 13.9(14-17): Unchecked Conversion
5467 The @code{Size} of an array object should not include its bounds; hence,
5468 the bounds should not be part of the converted data.
5474 The implementation should not generate unnecessary run-time checks to
5475 ensure that the representation of @var{S} is a representation of the
5476 target type. It should take advantage of the permission to return by
5477 reference when possible. Restrictions on unchecked conversions should be
5478 avoided unless required by the target environment.
5480 Followed. There are no restrictions on unchecked conversion. A warning is
5481 generated if the source and target types do not have the same size since
5482 the semantics in this case may be target dependent.
5486 The recommended level of support for unchecked conversions is:
5490 Unchecked conversions should be supported and should be reversible in
5491 the cases where this clause defines the result. To enable meaningful use
5492 of unchecked conversion, a contiguous representation should be used for
5493 elementary subtypes, for statically constrained array subtypes whose
5494 component subtype is one of the subtypes described in this paragraph,
5495 and for record subtypes without discriminants whose component subtypes
5496 are described in this paragraph.
5500 @cindex Heap usage, implicit
5501 @unnumberedsec 13.11(23-25): Implicit Heap Usage
5504 An implementation should document any cases in which it dynamically
5505 allocates heap storage for a purpose other than the evaluation of an
5508 Followed, the only other points at which heap storage is dynamically
5509 allocated are as follows:
5513 At initial elaboration time, to allocate dynamically sized global
5517 To allocate space for a task when a task is created.
5520 To extend the secondary stack dynamically when needed. The secondary
5521 stack is used for returning variable length results.
5526 A default (implementation-provided) storage pool for an
5527 access-to-constant type should not have overhead to support deallocation of
5534 A storage pool for an anonymous access type should be created at the
5535 point of an allocator for the type, and be reclaimed when the designated
5536 object becomes inaccessible.
5540 @cindex Unchecked deallocation
5541 @unnumberedsec 13.11.2(17): Unchecked De-allocation
5544 For a standard storage pool, @code{Free} should actually reclaim the
5549 @cindex Stream oriented attributes
5550 @unnumberedsec 13.13.2(17): Stream Oriented Attributes
5553 If a stream element is the same size as a storage element, then the
5554 normal in-memory representation should be used by @code{Read} and
5555 @code{Write} for scalar objects. Otherwise, @code{Read} and @code{Write}
5556 should use the smallest number of stream elements needed to represent
5557 all values in the base range of the scalar type.
5560 Followed. By default, GNAT uses the interpretation suggested by AI-195,
5561 which specifies using the size of the first subtype.
5562 However, such an implementation is based on direct binary
5563 representations and is therefore target- and endianness-dependent.
5564 To address this issue, GNAT also supplies an alternate implementation
5565 of the stream attributes @code{Read} and @code{Write},
5566 which uses the target-independent XDR standard representation
5568 @cindex XDR representation
5569 @cindex @code{Read} attribute
5570 @cindex @code{Write} attribute
5571 @cindex Stream oriented attributes
5572 The XDR implementation is provided as an alternative body of the
5573 @code{System.Stream_Attributes} package, in the file
5574 @file{s-strxdr.adb} in the GNAT library.
5575 There is no @file{s-strxdr.ads} file.
5576 In order to install the XDR implementation, do the following:
5578 @item Replace the default implementation of the
5579 @code{System.Stream_Attributes} package with the XDR implementation.
5580 For example on a Unix platform issue the commands:
5582 $ mv s-stratt.adb s-strold.adb
5583 $ mv s-strxdr.adb s-stratt.adb
5587 Rebuild the GNAT run-time library as documented in the
5588 @cite{GNAT User's Guide}
5591 @unnumberedsec A.1(52): Names of Predefined Numeric Types
5594 If an implementation provides additional named predefined integer types,
5595 then the names should end with @samp{Integer} as in
5596 @samp{Long_Integer}. If an implementation provides additional named
5597 predefined floating point types, then the names should end with
5598 @samp{Float} as in @samp{Long_Float}.
5602 @findex Ada.Characters.Handling
5603 @unnumberedsec A.3.2(49): @code{Ada.Characters.Handling}
5606 If an implementation provides a localized definition of @code{Character}
5607 or @code{Wide_Character}, then the effects of the subprograms in
5608 @code{Characters.Handling} should reflect the localizations. See also
5611 Followed. GNAT provides no such localized definitions.
5613 @cindex Bounded-length strings
5614 @unnumberedsec A.4.4(106): Bounded-Length String Handling
5617 Bounded string objects should not be implemented by implicit pointers
5618 and dynamic allocation.
5620 Followed. No implicit pointers or dynamic allocation are used.
5622 @cindex Random number generation
5623 @unnumberedsec A.5.2(46-47): Random Number Generation
5626 Any storage associated with an object of type @code{Generator} should be
5627 reclaimed on exit from the scope of the object.
5633 If the generator period is sufficiently long in relation to the number
5634 of distinct initiator values, then each possible value of
5635 @code{Initiator} passed to @code{Reset} should initiate a sequence of
5636 random numbers that does not, in a practical sense, overlap the sequence
5637 initiated by any other value. If this is not possible, then the mapping
5638 between initiator values and generator states should be a rapidly
5639 varying function of the initiator value.
5641 Followed. The generator period is sufficiently long for the first
5642 condition here to hold true.
5644 @findex Get_Immediate
5645 @unnumberedsec A.10.7(23): @code{Get_Immediate}
5648 The @code{Get_Immediate} procedures should be implemented with
5649 unbuffered input. For a device such as a keyboard, input should be
5650 @dfn{available} if a key has already been typed, whereas for a disk
5651 file, input should always be available except at end of file. For a file
5652 associated with a keyboard-like device, any line-editing features of the
5653 underlying operating system should be disabled during the execution of
5654 @code{Get_Immediate}.
5656 Followed on all targets except VxWorks. For VxWorks, there is no way to
5657 provide this functionality that does not result in the input buffer being
5658 flushed before the @code{Get_Immediate} call. A special unit
5659 @code{Interfaces.Vxworks.IO} is provided that contains routines to enable
5663 @unnumberedsec B.1(39-41): Pragma @code{Export}
5666 If an implementation supports pragma @code{Export} to a given language,
5667 then it should also allow the main subprogram to be written in that
5668 language. It should support some mechanism for invoking the elaboration
5669 of the Ada library units included in the system, and for invoking the
5670 finalization of the environment task. On typical systems, the
5671 recommended mechanism is to provide two subprograms whose link names are
5672 @code{adainit} and @code{adafinal}. @code{adainit} should contain the
5673 elaboration code for library units. @code{adafinal} should contain the
5674 finalization code. These subprograms should have no effect the second
5675 and subsequent time they are called.
5681 Automatic elaboration of pre-elaborated packages should be
5682 provided when pragma @code{Export} is supported.
5684 Followed when the main program is in Ada. If the main program is in a
5685 foreign language, then
5686 @code{adainit} must be called to elaborate pre-elaborated
5691 For each supported convention @var{L} other than @code{Intrinsic}, an
5692 implementation should support @code{Import} and @code{Export} pragmas
5693 for objects of @var{L}-compatible types and for subprograms, and pragma
5694 @code{Convention} for @var{L}-eligible types and for subprograms,
5695 presuming the other language has corresponding features. Pragma
5696 @code{Convention} need not be supported for scalar types.
5700 @cindex Package @code{Interfaces}
5702 @unnumberedsec B.2(12-13): Package @code{Interfaces}
5705 For each implementation-defined convention identifier, there should be a
5706 child package of package Interfaces with the corresponding name. This
5707 package should contain any declarations that would be useful for
5708 interfacing to the language (implementation) represented by the
5709 convention. Any declarations useful for interfacing to any language on
5710 the given hardware architecture should be provided directly in
5713 Followed. An additional package not defined
5714 in the Ada 95 Reference Manual is @code{Interfaces.CPP}, used
5715 for interfacing to C++.
5719 An implementation supporting an interface to C, COBOL, or Fortran should
5720 provide the corresponding package or packages described in the following
5723 Followed. GNAT provides all the packages described in this section.
5725 @cindex C, interfacing with
5726 @unnumberedsec B.3(63-71): Interfacing with C
5729 An implementation should support the following interface correspondences
5736 An Ada procedure corresponds to a void-returning C function.
5742 An Ada function corresponds to a non-void C function.
5748 An Ada @code{in} scalar parameter is passed as a scalar argument to a C
5755 An Ada @code{in} parameter of an access-to-object type with designated
5756 type @var{T} is passed as a @code{@var{t}*} argument to a C function,
5757 where @var{t} is the C type corresponding to the Ada type @var{T}.
5763 An Ada access @var{T} parameter, or an Ada @code{out} or @code{in out}
5764 parameter of an elementary type @var{T}, is passed as a @code{@var{t}*}
5765 argument to a C function, where @var{t} is the C type corresponding to
5766 the Ada type @var{T}. In the case of an elementary @code{out} or
5767 @code{in out} parameter, a pointer to a temporary copy is used to
5768 preserve by-copy semantics.
5774 An Ada parameter of a record type @var{T}, of any mode, is passed as a
5775 @code{@var{t}*} argument to a C function, where @var{t} is the C
5776 structure corresponding to the Ada type @var{T}.
5778 Followed. This convention may be overridden by the use of the C_Pass_By_Copy
5779 pragma, or Convention, or by explicitly specifying the mechanism for a given
5780 call using an extended import or export pragma.
5784 An Ada parameter of an array type with component type @var{T}, of any
5785 mode, is passed as a @code{@var{t}*} argument to a C function, where
5786 @var{t} is the C type corresponding to the Ada type @var{T}.
5792 An Ada parameter of an access-to-subprogram type is passed as a pointer
5793 to a C function whose prototype corresponds to the designated
5794 subprogram's specification.
5798 @cindex COBOL, interfacing with
5799 @unnumberedsec B.4(95-98): Interfacing with COBOL
5802 An Ada implementation should support the following interface
5803 correspondences between Ada and COBOL@.
5809 An Ada access @var{T} parameter is passed as a @samp{BY REFERENCE} data item of
5810 the COBOL type corresponding to @var{T}.
5816 An Ada in scalar parameter is passed as a @samp{BY CONTENT} data item of
5817 the corresponding COBOL type.
5823 Any other Ada parameter is passed as a @samp{BY REFERENCE} data item of the
5824 COBOL type corresponding to the Ada parameter type; for scalars, a local
5825 copy is used if necessary to ensure by-copy semantics.
5829 @cindex Fortran, interfacing with
5830 @unnumberedsec B.5(22-26): Interfacing with Fortran
5833 An Ada implementation should support the following interface
5834 correspondences between Ada and Fortran:
5840 An Ada procedure corresponds to a Fortran subroutine.
5846 An Ada function corresponds to a Fortran function.
5852 An Ada parameter of an elementary, array, or record type @var{T} is
5853 passed as a @var{T} argument to a Fortran procedure, where @var{T} is
5854 the Fortran type corresponding to the Ada type @var{T}, and where the
5855 INTENT attribute of the corresponding dummy argument matches the Ada
5856 formal parameter mode; the Fortran implementation's parameter passing
5857 conventions are used. For elementary types, a local copy is used if
5858 necessary to ensure by-copy semantics.
5864 An Ada parameter of an access-to-subprogram type is passed as a
5865 reference to a Fortran procedure whose interface corresponds to the
5866 designated subprogram's specification.
5870 @cindex Machine operations
5871 @unnumberedsec C.1(3-5): Access to Machine Operations
5874 The machine code or intrinsic support should allow access to all
5875 operations normally available to assembly language programmers for the
5876 target environment, including privileged instructions, if any.
5882 The interfacing pragmas (see Annex B) should support interface to
5883 assembler; the default assembler should be associated with the
5884 convention identifier @code{Assembler}.
5890 If an entity is exported to assembly language, then the implementation
5891 should allocate it at an addressable location, and should ensure that it
5892 is retained by the linking process, even if not otherwise referenced
5893 from the Ada code. The implementation should assume that any call to a
5894 machine code or assembler subprogram is allowed to read or update every
5895 object that is specified as exported.
5899 @unnumberedsec C.1(10-16): Access to Machine Operations
5902 The implementation should ensure that little or no overhead is
5903 associated with calling intrinsic and machine-code subprograms.
5905 Followed for both intrinsics and machine-code subprograms.
5909 It is recommended that intrinsic subprograms be provided for convenient
5910 access to any machine operations that provide special capabilities or
5911 efficiency and that are not otherwise available through the language
5914 Followed. A full set of machine operation intrinsic subprograms is provided.
5918 Atomic read-modify-write operations---e.g.@:, test and set, compare and
5919 swap, decrement and test, enqueue/dequeue.
5921 Followed on any target supporting such operations.
5925 Standard numeric functions---e.g.@:, sin, log.
5927 Followed on any target supporting such operations.
5931 String manipulation operations---e.g.@:, translate and test.
5933 Followed on any target supporting such operations.
5937 Vector operations---e.g.@:, compare vector against thresholds.
5939 Followed on any target supporting such operations.
5943 Direct operations on I/O ports.
5945 Followed on any target supporting such operations.
5947 @cindex Interrupt support
5948 @unnumberedsec C.3(28): Interrupt Support
5951 If the @code{Ceiling_Locking} policy is not in effect, the
5952 implementation should provide means for the application to specify which
5953 interrupts are to be blocked during protected actions, if the underlying
5954 system allows for a finer-grain control of interrupt blocking.
5956 Followed. The underlying system does not allow for finer-grain control
5957 of interrupt blocking.
5959 @cindex Protected procedure handlers
5960 @unnumberedsec C.3.1(20-21): Protected Procedure Handlers
5963 Whenever possible, the implementation should allow interrupt handlers to
5964 be called directly by the hardware.
5968 This is never possible under IRIX, so this is followed by default.
5970 Followed on any target where the underlying operating system permits
5975 Whenever practical, violations of any
5976 implementation-defined restrictions should be detected before run time.
5978 Followed. Compile time warnings are given when possible.
5980 @cindex Package @code{Interrupts}
5982 @unnumberedsec C.3.2(25): Package @code{Interrupts}
5986 If implementation-defined forms of interrupt handler procedures are
5987 supported, such as protected procedures with parameters, then for each
5988 such form of a handler, a type analogous to @code{Parameterless_Handler}
5989 should be specified in a child package of @code{Interrupts}, with the
5990 same operations as in the predefined package Interrupts.
5994 @cindex Pre-elaboration requirements
5995 @unnumberedsec C.4(14): Pre-elaboration Requirements
5998 It is recommended that pre-elaborated packages be implemented in such a
5999 way that there should be little or no code executed at run time for the
6000 elaboration of entities not already covered by the Implementation
6003 Followed. Executable code is generated in some cases, e.g.@: loops
6004 to initialize large arrays.
6006 @unnumberedsec C.5(8): Pragma @code{Discard_Names}
6010 If the pragma applies to an entity, then the implementation should
6011 reduce the amount of storage used for storing names associated with that
6016 @cindex Package @code{Task_Attributes}
6017 @findex Task_Attributes
6018 @unnumberedsec C.7.2(30): The Package Task_Attributes
6021 Some implementations are targeted to domains in which memory use at run
6022 time must be completely deterministic. For such implementations, it is
6023 recommended that the storage for task attributes will be pre-allocated
6024 statically and not from the heap. This can be accomplished by either
6025 placing restrictions on the number and the size of the task's
6026 attributes, or by using the pre-allocated storage for the first @var{N}
6027 attribute objects, and the heap for the others. In the latter case,
6028 @var{N} should be documented.
6030 Not followed. This implementation is not targeted to such a domain.
6032 @cindex Locking Policies
6033 @unnumberedsec D.3(17): Locking Policies
6037 The implementation should use names that end with @samp{_Locking} for
6038 locking policies defined by the implementation.
6040 Followed. A single implementation-defined locking policy is defined,
6041 whose name (@code{Inheritance_Locking}) follows this suggestion.
6043 @cindex Entry queuing policies
6044 @unnumberedsec D.4(16): Entry Queuing Policies
6047 Names that end with @samp{_Queuing} should be used
6048 for all implementation-defined queuing policies.
6050 Followed. No such implementation-defined queuing policies exist.
6052 @cindex Preemptive abort
6053 @unnumberedsec D.6(9-10): Preemptive Abort
6056 Even though the @code{abort_statement} is included in the list of
6057 potentially blocking operations (see 9.5.1), it is recommended that this
6058 statement be implemented in a way that never requires the task executing
6059 the @code{abort_statement} to block.
6065 On a multi-processor, the delay associated with aborting a task on
6066 another processor should be bounded; the implementation should use
6067 periodic polling, if necessary, to achieve this.
6071 @cindex Tasking restrictions
6072 @unnumberedsec D.7(21): Tasking Restrictions
6075 When feasible, the implementation should take advantage of the specified
6076 restrictions to produce a more efficient implementation.
6078 GNAT currently takes advantage of these restrictions by providing an optimized
6079 run time when the Ravenscar profile and the GNAT restricted run time set
6080 of restrictions are specified. See pragma @code{Profile (Ravenscar)} and
6081 pragma @code{Profile (Restricted)} for more details.
6083 @cindex Time, monotonic
6084 @unnumberedsec D.8(47-49): Monotonic Time
6087 When appropriate, implementations should provide configuration
6088 mechanisms to change the value of @code{Tick}.
6090 Such configuration mechanisms are not appropriate to this implementation
6091 and are thus not supported.
6095 It is recommended that @code{Calendar.Clock} and @code{Real_Time.Clock}
6096 be implemented as transformations of the same time base.
6102 It is recommended that the @dfn{best} time base which exists in
6103 the underlying system be available to the application through
6104 @code{Clock}. @dfn{Best} may mean highest accuracy or largest range.
6108 @cindex Partition communication subsystem
6110 @unnumberedsec E.5(28-29): Partition Communication Subsystem
6113 Whenever possible, the PCS on the called partition should allow for
6114 multiple tasks to call the RPC-receiver with different messages and
6115 should allow them to block until the corresponding subprogram body
6118 Followed by GLADE, a separately supplied PCS that can be used with
6123 The @code{Write} operation on a stream of type @code{Params_Stream_Type}
6124 should raise @code{Storage_Error} if it runs out of space trying to
6125 write the @code{Item} into the stream.
6127 Followed by GLADE, a separately supplied PCS that can be used with
6130 @cindex COBOL support
6131 @unnumberedsec F(7): COBOL Support
6134 If COBOL (respectively, C) is widely supported in the target
6135 environment, implementations supporting the Information Systems Annex
6136 should provide the child package @code{Interfaces.COBOL} (respectively,
6137 @code{Interfaces.C}) specified in Annex B and should support a
6138 @code{convention_identifier} of COBOL (respectively, C) in the interfacing
6139 pragmas (see Annex B), thus allowing Ada programs to interface with
6140 programs written in that language.
6144 @cindex Decimal radix support
6145 @unnumberedsec F.1(2): Decimal Radix Support
6148 Packed decimal should be used as the internal representation for objects
6149 of subtype @var{S} when @var{S}'Machine_Radix = 10.
6151 Not followed. GNAT ignores @var{S}'Machine_Radix and always uses binary
6155 @unnumberedsec G: Numerics
6158 If Fortran (respectively, C) is widely supported in the target
6159 environment, implementations supporting the Numerics Annex
6160 should provide the child package @code{Interfaces.Fortran} (respectively,
6161 @code{Interfaces.C}) specified in Annex B and should support a
6162 @code{convention_identifier} of Fortran (respectively, C) in the interfacing
6163 pragmas (see Annex B), thus allowing Ada programs to interface with
6164 programs written in that language.
6168 @cindex Complex types
6169 @unnumberedsec G.1.1(56-58): Complex Types
6172 Because the usual mathematical meaning of multiplication of a complex
6173 operand and a real operand is that of the scaling of both components of
6174 the former by the latter, an implementation should not perform this
6175 operation by first promoting the real operand to complex type and then
6176 performing a full complex multiplication. In systems that, in the
6177 future, support an Ada binding to IEC 559:1989, the latter technique
6178 will not generate the required result when one of the components of the
6179 complex operand is infinite. (Explicit multiplication of the infinite
6180 component by the zero component obtained during promotion yields a NaN
6181 that propagates into the final result.) Analogous advice applies in the
6182 case of multiplication of a complex operand and a pure-imaginary
6183 operand, and in the case of division of a complex operand by a real or
6184 pure-imaginary operand.
6190 Similarly, because the usual mathematical meaning of addition of a
6191 complex operand and a real operand is that the imaginary operand remains
6192 unchanged, an implementation should not perform this operation by first
6193 promoting the real operand to complex type and then performing a full
6194 complex addition. In implementations in which the @code{Signed_Zeros}
6195 attribute of the component type is @code{True} (and which therefore
6196 conform to IEC 559:1989 in regard to the handling of the sign of zero in
6197 predefined arithmetic operations), the latter technique will not
6198 generate the required result when the imaginary component of the complex
6199 operand is a negatively signed zero. (Explicit addition of the negative
6200 zero to the zero obtained during promotion yields a positive zero.)
6201 Analogous advice applies in the case of addition of a complex operand
6202 and a pure-imaginary operand, and in the case of subtraction of a
6203 complex operand and a real or pure-imaginary operand.
6209 Implementations in which @code{Real'Signed_Zeros} is @code{True} should
6210 attempt to provide a rational treatment of the signs of zero results and
6211 result components. As one example, the result of the @code{Argument}
6212 function should have the sign of the imaginary component of the
6213 parameter @code{X} when the point represented by that parameter lies on
6214 the positive real axis; as another, the sign of the imaginary component
6215 of the @code{Compose_From_Polar} function should be the same as
6216 (respectively, the opposite of) that of the @code{Argument} parameter when that
6217 parameter has a value of zero and the @code{Modulus} parameter has a
6218 nonnegative (respectively, negative) value.
6222 @cindex Complex elementary functions
6223 @unnumberedsec G.1.2(49): Complex Elementary Functions
6226 Implementations in which @code{Complex_Types.Real'Signed_Zeros} is
6227 @code{True} should attempt to provide a rational treatment of the signs
6228 of zero results and result components. For example, many of the complex
6229 elementary functions have components that are odd functions of one of
6230 the parameter components; in these cases, the result component should
6231 have the sign of the parameter component at the origin. Other complex
6232 elementary functions have zero components whose sign is opposite that of
6233 a parameter component at the origin, or is always positive or always
6238 @cindex Accuracy requirements
6239 @unnumberedsec G.2.4(19): Accuracy Requirements
6242 The versions of the forward trigonometric functions without a
6243 @code{Cycle} parameter should not be implemented by calling the
6244 corresponding version with a @code{Cycle} parameter of
6245 @code{2.0*Numerics.Pi}, since this will not provide the required
6246 accuracy in some portions of the domain. For the same reason, the
6247 version of @code{Log} without a @code{Base} parameter should not be
6248 implemented by calling the corresponding version with a @code{Base}
6249 parameter of @code{Numerics.e}.
6253 @cindex Complex arithmetic accuracy
6254 @cindex Accuracy, complex arithmetic
6255 @unnumberedsec G.2.6(15): Complex Arithmetic Accuracy
6259 The version of the @code{Compose_From_Polar} function without a
6260 @code{Cycle} parameter should not be implemented by calling the
6261 corresponding version with a @code{Cycle} parameter of
6262 @code{2.0*Numerics.Pi}, since this will not provide the required
6263 accuracy in some portions of the domain.
6267 @c -----------------------------------------
6268 @node Implementation Defined Characteristics
6269 @chapter Implementation Defined Characteristics
6272 In addition to the implementation dependent pragmas and attributes, and
6273 the implementation advice, there are a number of other features of Ada
6274 95 that are potentially implementation dependent. These are mentioned
6275 throughout the Ada 95 Reference Manual, and are summarized in annex M@.
6277 A requirement for conforming Ada compilers is that they provide
6278 documentation describing how the implementation deals with each of these
6279 issues. In this chapter, you will find each point in annex M listed
6280 followed by a description in italic font of how GNAT
6284 implementation on IRIX 5.3 operating system or greater
6286 handles the implementation dependence.
6288 You can use this chapter as a guide to minimizing implementation
6289 dependent features in your programs if portability to other compilers
6290 and other operating systems is an important consideration. The numbers
6291 in each section below correspond to the paragraph number in the Ada 95
6297 @strong{2}. Whether or not each recommendation given in Implementation
6298 Advice is followed. See 1.1.2(37).
6301 @xref{Implementation Advice}.
6306 @strong{3}. Capacity limitations of the implementation. See 1.1.3(3).
6309 The complexity of programs that can be processed is limited only by the
6310 total amount of available virtual memory, and disk space for the
6311 generated object files.
6316 @strong{4}. Variations from the standard that are impractical to avoid
6317 given the implementation's execution environment. See 1.1.3(6).
6320 There are no variations from the standard.
6325 @strong{5}. Which @code{code_statement}s cause external
6326 interactions. See 1.1.3(10).
6329 Any @code{code_statement} can potentially cause external interactions.
6334 @strong{6}. The coded representation for the text of an Ada
6335 program. See 2.1(4).
6338 See separate section on source representation.
6343 @strong{7}. The control functions allowed in comments. See 2.1(14).
6346 See separate section on source representation.
6351 @strong{8}. The representation for an end of line. See 2.2(2).
6354 See separate section on source representation.
6359 @strong{9}. Maximum supported line length and lexical element
6360 length. See 2.2(15).
6363 The maximum line length is 255 characters an the maximum length of a
6364 lexical element is also 255 characters.
6369 @strong{10}. Implementation defined pragmas. See 2.8(14).
6373 @xref{Implementation Defined Pragmas}.
6378 @strong{11}. Effect of pragma @code{Optimize}. See 2.8(27).
6381 Pragma @code{Optimize}, if given with a @code{Time} or @code{Space}
6382 parameter, checks that the optimization flag is set, and aborts if it is
6388 @strong{12}. The sequence of characters of the value returned by
6389 @code{@var{S}'Image} when some of the graphic characters of
6390 @code{@var{S}'Wide_Image} are not defined in @code{Character}. See
6394 The sequence of characters is as defined by the wide character encoding
6395 method used for the source. See section on source representation for
6401 @strong{13}. The predefined integer types declared in
6402 @code{Standard}. See 3.5.4(25).
6406 @item Short_Short_Integer
6409 (Short) 16 bit signed
6413 64 bit signed (Alpha OpenVMS only)
6414 32 bit signed (all other targets)
6415 @item Long_Long_Integer
6422 @strong{14}. Any nonstandard integer types and the operators defined
6423 for them. See 3.5.4(26).
6426 There are no nonstandard integer types.
6431 @strong{15}. Any nonstandard real types and the operators defined for
6435 There are no nonstandard real types.
6440 @strong{16}. What combinations of requested decimal precision and range
6441 are supported for floating point types. See 3.5.7(7).
6444 The precision and range is as defined by the IEEE standard.
6449 @strong{17}. The predefined floating point types declared in
6450 @code{Standard}. See 3.5.7(16).
6457 (Short) 32 bit IEEE short
6460 @item Long_Long_Float
6461 64 bit IEEE long (80 bit IEEE long on x86 processors)
6467 @strong{18}. The small of an ordinary fixed point type. See 3.5.9(8).
6470 @code{Fine_Delta} is 2**(@minus{}63)
6475 @strong{19}. What combinations of small, range, and digits are
6476 supported for fixed point types. See 3.5.9(10).
6479 Any combinations are permitted that do not result in a small less than
6480 @code{Fine_Delta} and do not result in a mantissa larger than 63 bits.
6481 If the mantissa is larger than 53 bits on machines where Long_Long_Float
6482 is 64 bits (true of all architectures except ia32), then the output from
6483 Text_IO is accurate to only 53 bits, rather than the full mantissa. This
6484 is because floating-point conversions are used to convert fixed point.
6489 @strong{20}. The result of @code{Tags.Expanded_Name} for types declared
6490 within an unnamed @code{block_statement}. See 3.9(10).
6493 Block numbers of the form @code{B@var{nnn}}, where @var{nnn} is a
6494 decimal integer are allocated.
6499 @strong{21}. Implementation-defined attributes. See 4.1.4(12).
6502 @xref{Implementation Defined Attributes}.
6507 @strong{22}. Any implementation-defined time types. See 9.6(6).
6510 There are no implementation-defined time types.
6515 @strong{23}. The time base associated with relative delays.
6518 See 9.6(20). The time base used is that provided by the C library
6519 function @code{gettimeofday}.
6524 @strong{24}. The time base of the type @code{Calendar.Time}. See
6528 The time base used is that provided by the C library function
6529 @code{gettimeofday}.
6534 @strong{25}. The time zone used for package @code{Calendar}
6535 operations. See 9.6(24).
6538 The time zone used by package @code{Calendar} is the current system time zone
6539 setting for local time, as accessed by the C library function
6545 @strong{26}. Any limit on @code{delay_until_statements} of
6546 @code{select_statements}. See 9.6(29).
6549 There are no such limits.
6554 @strong{27}. Whether or not two non overlapping parts of a composite
6555 object are independently addressable, in the case where packing, record
6556 layout, or @code{Component_Size} is specified for the object. See
6560 Separate components are independently addressable if they do not share
6561 overlapping storage units.
6566 @strong{28}. The representation for a compilation. See 10.1(2).
6569 A compilation is represented by a sequence of files presented to the
6570 compiler in a single invocation of the @code{gcc} command.
6575 @strong{29}. Any restrictions on compilations that contain multiple
6576 compilation_units. See 10.1(4).
6579 No single file can contain more than one compilation unit, but any
6580 sequence of files can be presented to the compiler as a single
6586 @strong{30}. The mechanisms for creating an environment and for adding
6587 and replacing compilation units. See 10.1.4(3).
6590 See separate section on compilation model.
6595 @strong{31}. The manner of explicitly assigning library units to a
6596 partition. See 10.2(2).
6599 If a unit contains an Ada main program, then the Ada units for the partition
6600 are determined by recursive application of the rules in the Ada Reference
6601 Manual section 10.2(2-6). In other words, the Ada units will be those that
6602 are needed by the main program, and then this definition of need is applied
6603 recursively to those units, and the partition contains the transitive
6604 closure determined by this relationship. In short, all the necessary units
6605 are included, with no need to explicitly specify the list. If additional
6606 units are required, e.g.@: by foreign language units, then all units must be
6607 mentioned in the context clause of one of the needed Ada units.
6609 If the partition contains no main program, or if the main program is in
6610 a language other than Ada, then GNAT
6611 provides the binder options @code{-z} and @code{-n} respectively, and in
6612 this case a list of units can be explicitly supplied to the binder for
6613 inclusion in the partition (all units needed by these units will also
6614 be included automatically). For full details on the use of these
6615 options, refer to the @cite{GNAT User's Guide} sections on Binding
6621 @strong{32}. The implementation-defined means, if any, of specifying
6622 which compilation units are needed by a given compilation unit. See
6626 The units needed by a given compilation unit are as defined in
6627 the Ada Reference Manual section 10.2(2-6). There are no
6628 implementation-defined pragmas or other implementation-defined
6629 means for specifying needed units.
6634 @strong{33}. The manner of designating the main subprogram of a
6635 partition. See 10.2(7).
6638 The main program is designated by providing the name of the
6639 corresponding @file{ALI} file as the input parameter to the binder.
6644 @strong{34}. The order of elaboration of @code{library_items}. See
6648 The first constraint on ordering is that it meets the requirements of
6649 chapter 10 of the Ada 95 Reference Manual. This still leaves some
6650 implementation dependent choices, which are resolved by first
6651 elaborating bodies as early as possible (i.e.@: in preference to specs
6652 where there is a choice), and second by evaluating the immediate with
6653 clauses of a unit to determine the probably best choice, and
6654 third by elaborating in alphabetical order of unit names
6655 where a choice still remains.
6660 @strong{35}. Parameter passing and function return for the main
6661 subprogram. See 10.2(21).
6664 The main program has no parameters. It may be a procedure, or a function
6665 returning an integer type. In the latter case, the returned integer
6666 value is the return code of the program (overriding any value that
6667 may have been set by a call to @code{Ada.Command_Line.Set_Exit_Status}).
6672 @strong{36}. The mechanisms for building and running partitions. See
6676 GNAT itself supports programs with only a single partition. The GNATDIST
6677 tool provided with the GLADE package (which also includes an implementation
6678 of the PCS) provides a completely flexible method for building and running
6679 programs consisting of multiple partitions. See the separate GLADE manual
6685 @strong{37}. The details of program execution, including program
6686 termination. See 10.2(25).
6689 See separate section on compilation model.
6694 @strong{38}. The semantics of any non-active partitions supported by the
6695 implementation. See 10.2(28).
6698 Passive partitions are supported on targets where shared memory is
6699 provided by the operating system. See the GLADE reference manual for
6705 @strong{39}. The information returned by @code{Exception_Message}. See
6709 Exception message returns the null string unless a specific message has
6710 been passed by the program.
6715 @strong{40}. The result of @code{Exceptions.Exception_Name} for types
6716 declared within an unnamed @code{block_statement}. See 11.4.1(12).
6719 Blocks have implementation defined names of the form @code{B@var{nnn}}
6720 where @var{nnn} is an integer.
6725 @strong{41}. The information returned by
6726 @code{Exception_Information}. See 11.4.1(13).
6729 @code{Exception_Information} returns a string in the following format:
6732 @emph{Exception_Name:} nnnnn
6733 @emph{Message:} mmmmm
6735 @emph{Call stack traceback locations:}
6736 0xhhhh 0xhhhh 0xhhhh ... 0xhhh
6744 @code{nnnn} is the fully qualified name of the exception in all upper
6745 case letters. This line is always present.
6748 @code{mmmm} is the message (this line present only if message is non-null)
6751 @code{ppp} is the Process Id value as a decimal integer (this line is
6752 present only if the Process Id is non-zero). Currently we are
6753 not making use of this field.
6756 The Call stack traceback locations line and the following values
6757 are present only if at least one traceback location was recorded.
6758 The values are given in C style format, with lower case letters
6759 for a-f, and only as many digits present as are necessary.
6763 The line terminator sequence at the end of each line, including
6764 the last line is a single @code{LF} character (@code{16#0A#}).
6769 @strong{42}. Implementation-defined check names. See 11.5(27).
6772 No implementation-defined check names are supported.
6777 @strong{43}. The interpretation of each aspect of representation. See
6781 See separate section on data representations.
6786 @strong{44}. Any restrictions placed upon representation items. See
6790 See separate section on data representations.
6795 @strong{45}. The meaning of @code{Size} for indefinite subtypes. See
6799 Size for an indefinite subtype is the maximum possible size, except that
6800 for the case of a subprogram parameter, the size of the parameter object
6806 @strong{46}. The default external representation for a type tag. See
6810 The default external representation for a type tag is the fully expanded
6811 name of the type in upper case letters.
6816 @strong{47}. What determines whether a compilation unit is the same in
6817 two different partitions. See 13.3(76).
6820 A compilation unit is the same in two different partitions if and only
6821 if it derives from the same source file.
6826 @strong{48}. Implementation-defined components. See 13.5.1(15).
6829 The only implementation defined component is the tag for a tagged type,
6830 which contains a pointer to the dispatching table.
6835 @strong{49}. If @code{Word_Size} = @code{Storage_Unit}, the default bit
6836 ordering. See 13.5.3(5).
6839 @code{Word_Size} (32) is not the same as @code{Storage_Unit} (8) for this
6840 implementation, so no non-default bit ordering is supported. The default
6841 bit ordering corresponds to the natural endianness of the target architecture.
6846 @strong{50}. The contents of the visible part of package @code{System}
6847 and its language-defined children. See 13.7(2).
6850 See the definition of these packages in files @file{system.ads} and
6851 @file{s-stoele.ads}.
6856 @strong{51}. The contents of the visible part of package
6857 @code{System.Machine_Code}, and the meaning of
6858 @code{code_statements}. See 13.8(7).
6861 See the definition and documentation in file @file{s-maccod.ads}.
6866 @strong{52}. The effect of unchecked conversion. See 13.9(11).
6869 Unchecked conversion between types of the same size
6870 and results in an uninterpreted transmission of the bits from one type
6871 to the other. If the types are of unequal sizes, then in the case of
6872 discrete types, a shorter source is first zero or sign extended as
6873 necessary, and a shorter target is simply truncated on the left.
6874 For all non-discrete types, the source is first copied if necessary
6875 to ensure that the alignment requirements of the target are met, then
6876 a pointer is constructed to the source value, and the result is obtained
6877 by dereferencing this pointer after converting it to be a pointer to the
6883 @strong{53}. The manner of choosing a storage pool for an access type
6884 when @code{Storage_Pool} is not specified for the type. See 13.11(17).
6887 There are 3 different standard pools used by the compiler when
6888 @code{Storage_Pool} is not specified depending whether the type is local
6889 to a subprogram or defined at the library level and whether
6890 @code{Storage_Size}is specified or not. See documentation in the runtime
6891 library units @code{System.Pool_Global}, @code{System.Pool_Size} and
6892 @code{System.Pool_Local} in files @file{s-poosiz.ads},
6893 @file{s-pooglo.ads} and @file{s-pooloc.ads} for full details on the
6899 @strong{54}. Whether or not the implementation provides user-accessible
6900 names for the standard pool type(s). See 13.11(17).
6904 See documentation in the sources of the run time mentioned in paragraph
6905 @strong{53} . All these pools are accessible by means of @code{with}'ing
6911 @strong{55}. The meaning of @code{Storage_Size}. See 13.11(18).
6914 @code{Storage_Size} is measured in storage units, and refers to the
6915 total space available for an access type collection, or to the primary
6916 stack space for a task.
6921 @strong{56}. Implementation-defined aspects of storage pools. See
6925 See documentation in the sources of the run time mentioned in paragraph
6926 @strong{53} for details on GNAT-defined aspects of storage pools.
6931 @strong{57}. The set of restrictions allowed in a pragma
6932 @code{Restrictions}. See 13.12(7).
6935 All RM defined Restriction identifiers are implemented. The following
6936 additional restriction identifiers are provided. There are two separate
6937 lists of implementation dependent restriction identifiers. The first
6938 set requires consistency throughout a partition (in other words, if the
6939 restriction identifier is used for any compilation unit in the partition,
6940 then all compilation units in the partition must obey the restriction.
6944 @item Simple_Barriers
6945 @findex Simple_Barriers
6946 This restriction ensures at compile time that barriers in entry declarations
6947 for protected types are restricted to either static boolean expressions or
6948 references to simple boolean variables defined in the private part of the
6949 protected type. No other form of entry barriers is permitted. This is one
6950 of the restrictions of the Ravenscar profile for limited tasking (see also
6951 pragma @code{Profile (Ravenscar)}).
6953 @item Max_Entry_Queue_Length => Expr
6954 @findex Max_Entry_Queue_Length
6955 This restriction is a declaration that any protected entry compiled in
6956 the scope of the restriction has at most the specified number of
6957 tasks waiting on the entry
6958 at any one time, and so no queue is required. This restriction is not
6959 checked at compile time. A program execution is erroneous if an attempt
6960 is made to queue more than the specified number of tasks on such an entry.
6964 This restriction ensures at compile time that there is no implicit or
6965 explicit dependence on the package @code{Ada.Calendar}.
6967 @item No_Direct_Boolean_Operators
6968 @findex No_Direct_Boolean_Operators
6969 This restriction ensures that no logical (and/or/xor) or comparison
6970 operators are used on operands of type Boolean (or any type derived
6971 from Boolean). This is intended for use in safety critical programs
6972 where the certification protocol requires the use of short-circuit
6973 (and then, or else) forms for all composite boolean operations.
6975 @item No_Dynamic_Attachment
6976 @findex No_Dynamic_Attachment
6977 This restriction ensures that there is no call to any of the operations
6978 defined in package Ada.Interrupts.
6980 @item No_Enumeration_Maps
6981 @findex No_Enumeration_Maps
6982 This restriction ensures at compile time that no operations requiring
6983 enumeration maps are used (that is Image and Value attributes applied
6984 to enumeration types).
6986 @item No_Entry_Calls_In_Elaboration_Code
6987 @findex No_Entry_Calls_In_Elaboration_Code
6988 This restriction ensures at compile time that no task or protected entry
6989 calls are made during elaboration code. As a result of the use of this
6990 restriction, the compiler can assume that no code past an accept statement
6991 in a task can be executed at elaboration time.
6993 @item No_Exception_Handlers
6994 @findex No_Exception_Handlers
6995 This restriction ensures at compile time that there are no explicit
6996 exception handlers. It also indicates that no exception propagation will
6997 be provided. In this mode, exceptions may be raised but will result in
6998 an immediate call to the last chance handler, a routine that the user
6999 must define with the following profile:
7001 procedure Last_Chance_Handler
7002 (Source_Location : System.Address; Line : Integer);
7003 pragma Export (C, Last_Chance_Handler,
7004 "__gnat_last_chance_handler");
7006 The parameter is a C null-terminated string representing a message to be
7007 associated with the exception (typically the source location of the raise
7008 statement generated by the compiler). The Line parameter when non-zero
7009 represents the line number in the source program where the raise occurs.
7011 @item No_Exception_Streams
7012 @findex No_Exception_Streams
7013 This restriction ensures at compile time that no stream operations for
7014 types Exception_Id or Exception_Occurrence are used. This also makes it
7015 impossible to pass exceptions to or from a partition with this restriction
7016 in a distributed environment. If this exception is active, then the generated
7017 code is simplified by omitting the otherwise-required global registration
7018 of exceptions when they are declared.
7020 @item No_Implicit_Conditionals
7021 @findex No_Implicit_Conditionals
7022 This restriction ensures that the generated code does not contain any
7023 implicit conditionals, either by modifying the generated code where possible,
7024 or by rejecting any construct that would otherwise generate an implicit
7025 conditional. Note that this check does not include run time constraint
7026 checks, which on some targets may generate implicit conditionals as
7027 well. To control the latter, constraint checks can be suppressed in the
7030 @item No_Implicit_Dynamic_Code
7031 @findex No_Implicit_Dynamic_Code
7032 This restriction prevents the compiler from building ``trampolines''.
7033 This is a structure that is built on the stack and contains dynamic
7034 code to be executed at run time. A trampoline is needed to indirectly
7035 address a nested subprogram (that is a subprogram that is not at the
7036 library level). The restriction prevents the use of any of the
7037 attributes @code{Address}, @code{Access} or @code{Unrestricted_Access}
7038 being applied to a subprogram that is not at the library level.
7040 @item No_Implicit_Loops
7041 @findex No_Implicit_Loops
7042 This restriction ensures that the generated code does not contain any
7043 implicit @code{for} loops, either by modifying
7044 the generated code where possible,
7045 or by rejecting any construct that would otherwise generate an implicit
7048 @item No_Initialize_Scalars
7049 @findex No_Initialize_Scalars
7050 This restriction ensures that no unit in the partition is compiled with
7051 pragma Initialize_Scalars. This allows the generation of more efficient
7052 code, and in particular eliminates dummy null initialization routines that
7053 are otherwise generated for some record and array types.
7055 @item No_Local_Protected_Objects
7056 @findex No_Local_Protected_Objects
7057 This restriction ensures at compile time that protected objects are
7058 only declared at the library level.
7060 @item No_Protected_Type_Allocators
7061 @findex No_Protected_Type_Allocators
7062 This restriction ensures at compile time that there are no allocator
7063 expressions that attempt to allocate protected objects.
7065 @item No_Secondary_Stack
7066 @findex No_Secondary_Stack
7067 This restriction ensures at compile time that the generated code does not
7068 contain any reference to the secondary stack. The secondary stack is used
7069 to implement functions returning unconstrained objects (arrays or records)
7072 @item No_Select_Statements
7073 @findex No_Select_Statements
7074 This restriction ensures at compile time no select statements of any kind
7075 are permitted, that is the keyword @code{select} may not appear.
7076 This is one of the restrictions of the Ravenscar
7077 profile for limited tasking (see also pragma @code{Profile (Ravenscar)}).
7079 @item No_Standard_Storage_Pools
7080 @findex No_Standard_Storage_Pools
7081 This restriction ensures at compile time that no access types
7082 use the standard default storage pool. Any access type declared must
7083 have an explicit Storage_Pool attribute defined specifying a
7084 user-defined storage pool.
7088 This restriction ensures at compile/bind time that there are no
7089 stream objects created (and therefore no actual stream operations).
7090 This restriction does not forbid dependences on the package
7091 @code{Ada.Streams}. So it is permissible to with
7092 @code{Ada.Streams} (or another package that does so itself)
7093 as long as no actual stream objects are created.
7095 @item No_Task_Attributes_Package
7096 @findex No_Task_Attributes_Package
7097 This restriction ensures at compile time that there are no implicit or
7098 explicit dependencies on the package @code{Ada.Task_Attributes}.
7100 @item No_Task_Termination
7101 @findex No_Task_Termination
7102 This restriction ensures at compile time that no terminate alternatives
7103 appear in any task body.
7107 This restriction prevents the declaration of tasks or task types throughout
7108 the partition. It is similar in effect to the use of @code{Max_Tasks => 0}
7109 except that violations are caught at compile time and cause an error message
7110 to be output either by the compiler or binder.
7112 @item No_Wide_Characters
7113 @findex No_Wide_Characters
7114 This restriction ensures at compile time that no uses of the types
7115 @code{Wide_Character} or @code{Wide_String}
7116 appear, and that no wide character literals
7117 appear in the program (that is literals representing characters not in
7118 type @code{Character}.
7120 @item Static_Priorities
7121 @findex Static_Priorities
7122 This restriction ensures at compile time that all priority expressions
7123 are static, and that there are no dependencies on the package
7124 @code{Ada.Dynamic_Priorities}.
7126 @item Static_Storage_Size
7127 @findex Static_Storage_Size
7128 This restriction ensures at compile time that any expression appearing
7129 in a Storage_Size pragma or attribute definition clause is static.
7134 The second set of implementation dependent restriction identifiers
7135 does not require partition-wide consistency.
7136 The restriction may be enforced for a single
7137 compilation unit without any effect on any of the
7138 other compilation units in the partition.
7142 @item No_Elaboration_Code
7143 @findex No_Elaboration_Code
7144 This restriction ensures at compile time that no elaboration code is
7145 generated. Note that this is not the same condition as is enforced
7146 by pragma @code{Preelaborate}. There are cases in which pragma
7147 @code{Preelaborate} still permits code to be generated (e.g.@: code
7148 to initialize a large array to all zeroes), and there are cases of units
7149 which do not meet the requirements for pragma @code{Preelaborate},
7150 but for which no elaboration code is generated. Generally, it is
7151 the case that preelaborable units will meet the restrictions, with
7152 the exception of large aggregates initialized with an others_clause,
7153 and exception declarations (which generate calls to a run-time
7154 registry procedure). Note that this restriction is enforced on
7155 a unit by unit basis, it need not be obeyed consistently
7156 throughout a partition.
7158 @item No_Entry_Queue
7159 @findex No_Entry_Queue
7160 This restriction is a declaration that any protected entry compiled in
7161 the scope of the restriction has at most one task waiting on the entry
7162 at any one time, and so no queue is required. This restriction is not
7163 checked at compile time. A program execution is erroneous if an attempt
7164 is made to queue a second task on such an entry.
7166 @item No_Implementation_Attributes
7167 @findex No_Implementation_Attributes
7168 This restriction checks at compile time that no GNAT-defined attributes
7169 are present. With this restriction, the only attributes that can be used
7170 are those defined in the Ada 95 Reference Manual.
7172 @item No_Implementation_Pragmas
7173 @findex No_Implementation_Pragmas
7174 This restriction checks at compile time that no GNAT-defined pragmas
7175 are present. With this restriction, the only pragmas that can be used
7176 are those defined in the Ada 95 Reference Manual.
7178 @item No_Implementation_Restrictions
7179 @findex No_Implementation_Restrictions
7180 This restriction checks at compile time that no GNAT-defined restriction
7181 identifiers (other than @code{No_Implementation_Restrictions} itself)
7182 are present. With this restriction, the only other restriction identifiers
7183 that can be used are those defined in the Ada 95 Reference Manual.
7190 @strong{58}. The consequences of violating limitations on
7191 @code{Restrictions} pragmas. See 13.12(9).
7194 Restrictions that can be checked at compile time result in illegalities
7195 if violated. Currently there are no other consequences of violating
7201 @strong{59}. The representation used by the @code{Read} and
7202 @code{Write} attributes of elementary types in terms of stream
7203 elements. See 13.13.2(9).
7206 The representation is the in-memory representation of the base type of
7207 the type, using the number of bits corresponding to the
7208 @code{@var{type}'Size} value, and the natural ordering of the machine.
7213 @strong{60}. The names and characteristics of the numeric subtypes
7214 declared in the visible part of package @code{Standard}. See A.1(3).
7217 See items describing the integer and floating-point types supported.
7222 @strong{61}. The accuracy actually achieved by the elementary
7223 functions. See A.5.1(1).
7226 The elementary functions correspond to the functions available in the C
7227 library. Only fast math mode is implemented.
7232 @strong{62}. The sign of a zero result from some of the operators or
7233 functions in @code{Numerics.Generic_Elementary_Functions}, when
7234 @code{Float_Type'Signed_Zeros} is @code{True}. See A.5.1(46).
7237 The sign of zeroes follows the requirements of the IEEE 754 standard on
7243 @strong{63}. The value of
7244 @code{Numerics.Float_Random.Max_Image_Width}. See A.5.2(27).
7247 Maximum image width is 649, see library file @file{a-numran.ads}.
7252 @strong{64}. The value of
7253 @code{Numerics.Discrete_Random.Max_Image_Width}. See A.5.2(27).
7256 Maximum image width is 80, see library file @file{a-nudira.ads}.
7261 @strong{65}. The algorithms for random number generation. See
7265 The algorithm is documented in the source files @file{a-numran.ads} and
7266 @file{a-numran.adb}.
7271 @strong{66}. The string representation of a random number generator's
7272 state. See A.5.2(38).
7275 See the documentation contained in the file @file{a-numran.adb}.
7280 @strong{67}. The minimum time interval between calls to the
7281 time-dependent Reset procedure that are guaranteed to initiate different
7282 random number sequences. See A.5.2(45).
7285 The minimum period between reset calls to guarantee distinct series of
7286 random numbers is one microsecond.
7291 @strong{68}. The values of the @code{Model_Mantissa},
7292 @code{Model_Emin}, @code{Model_Epsilon}, @code{Model},
7293 @code{Safe_First}, and @code{Safe_Last} attributes, if the Numerics
7294 Annex is not supported. See A.5.3(72).
7297 See the source file @file{ttypef.ads} for the values of all numeric
7303 @strong{69}. Any implementation-defined characteristics of the
7304 input-output packages. See A.7(14).
7307 There are no special implementation defined characteristics for these
7313 @strong{70}. The value of @code{Buffer_Size} in @code{Storage_IO}. See
7317 All type representations are contiguous, and the @code{Buffer_Size} is
7318 the value of @code{@var{type}'Size} rounded up to the next storage unit
7324 @strong{71}. External files for standard input, standard output, and
7325 standard error See A.10(5).
7328 These files are mapped onto the files provided by the C streams
7329 libraries. See source file @file{i-cstrea.ads} for further details.
7334 @strong{72}. The accuracy of the value produced by @code{Put}. See
7338 If more digits are requested in the output than are represented by the
7339 precision of the value, zeroes are output in the corresponding least
7340 significant digit positions.
7345 @strong{73}. The meaning of @code{Argument_Count}, @code{Argument}, and
7346 @code{Command_Name}. See A.15(1).
7349 These are mapped onto the @code{argv} and @code{argc} parameters of the
7350 main program in the natural manner.
7355 @strong{74}. Implementation-defined convention names. See B.1(11).
7358 The following convention names are supported
7366 Synonym for Assembler
7368 Synonym for Assembler
7371 @item C_Pass_By_Copy
7372 Allowed only for record types, like C, but also notes that record
7373 is to be passed by copy rather than reference.
7379 Treated the same as C
7381 Treated the same as C
7385 For support of pragma @code{Import} with convention Intrinsic, see
7386 separate section on Intrinsic Subprograms.
7388 Stdcall (used for Windows implementations only). This convention correspond
7389 to the WINAPI (previously called Pascal convention) C/C++ convention under
7390 Windows. A function with this convention cleans the stack before exit.
7396 Stubbed is a special convention used to indicate that the body of the
7397 subprogram will be entirely ignored. Any call to the subprogram
7398 is converted into a raise of the @code{Program_Error} exception. If a
7399 pragma @code{Import} specifies convention @code{stubbed} then no body need
7400 be present at all. This convention is useful during development for the
7401 inclusion of subprograms whose body has not yet been written.
7405 In addition, all otherwise unrecognized convention names are also
7406 treated as being synonymous with convention C@. In all implementations
7407 except for VMS, use of such other names results in a warning. In VMS
7408 implementations, these names are accepted silently.
7413 @strong{75}. The meaning of link names. See B.1(36).
7416 Link names are the actual names used by the linker.
7421 @strong{76}. The manner of choosing link names when neither the link
7422 name nor the address of an imported or exported entity is specified. See
7426 The default linker name is that which would be assigned by the relevant
7427 external language, interpreting the Ada name as being in all lower case
7433 @strong{77}. The effect of pragma @code{Linker_Options}. See B.1(37).
7436 The string passed to @code{Linker_Options} is presented uninterpreted as
7437 an argument to the link command, unless it contains Ascii.NUL characters.
7438 NUL characters if they appear act as argument separators, so for example
7440 @smallexample @c ada
7441 pragma Linker_Options ("-labc" & ASCII.Nul & "-ldef");
7445 causes two separate arguments @code{-labc} and @code{-ldef} to be passed to the
7446 linker. The order of linker options is preserved for a given unit. The final
7447 list of options passed to the linker is in reverse order of the elaboration
7448 order. For example, linker options fo a body always appear before the options
7449 from the corresponding package spec.
7454 @strong{78}. The contents of the visible part of package
7455 @code{Interfaces} and its language-defined descendants. See B.2(1).
7458 See files with prefix @file{i-} in the distributed library.
7463 @strong{79}. Implementation-defined children of package
7464 @code{Interfaces}. The contents of the visible part of package
7465 @code{Interfaces}. See B.2(11).
7468 See files with prefix @file{i-} in the distributed library.
7473 @strong{80}. The types @code{Floating}, @code{Long_Floating},
7474 @code{Binary}, @code{Long_Binary}, @code{Decimal_ Element}, and
7475 @code{COBOL_Character}; and the initialization of the variables
7476 @code{Ada_To_COBOL} and @code{COBOL_To_Ada}, in
7477 @code{Interfaces.COBOL}. See B.4(50).
7484 (Floating) Long_Float
7489 @item Decimal_Element
7491 @item COBOL_Character
7496 For initialization, see the file @file{i-cobol.ads} in the distributed library.
7501 @strong{81}. Support for access to machine instructions. See C.1(1).
7504 See documentation in file @file{s-maccod.ads} in the distributed library.
7509 @strong{82}. Implementation-defined aspects of access to machine
7510 operations. See C.1(9).
7513 See documentation in file @file{s-maccod.ads} in the distributed library.
7518 @strong{83}. Implementation-defined aspects of interrupts. See C.3(2).
7521 Interrupts are mapped to signals or conditions as appropriate. See
7523 @code{Ada.Interrupt_Names} in source file @file{a-intnam.ads} for details
7524 on the interrupts supported on a particular target.
7529 @strong{84}. Implementation-defined aspects of pre-elaboration. See
7533 GNAT does not permit a partition to be restarted without reloading,
7534 except under control of the debugger.
7539 @strong{85}. The semantics of pragma @code{Discard_Names}. See C.5(7).
7542 Pragma @code{Discard_Names} causes names of enumeration literals to
7543 be suppressed. In the presence of this pragma, the Image attribute
7544 provides the image of the Pos of the literal, and Value accepts
7550 @strong{86}. The result of the @code{Task_Identification.Image}
7551 attribute. See C.7.1(7).
7554 The result of this attribute is an 8-digit hexadecimal string
7555 representing the virtual address of the task control block.
7560 @strong{87}. The value of @code{Current_Task} when in a protected entry
7561 or interrupt handler. See C.7.1(17).
7564 Protected entries or interrupt handlers can be executed by any
7565 convenient thread, so the value of @code{Current_Task} is undefined.
7570 @strong{88}. The effect of calling @code{Current_Task} from an entry
7571 body or interrupt handler. See C.7.1(19).
7574 The effect of calling @code{Current_Task} from an entry body or
7575 interrupt handler is to return the identification of the task currently
7581 @strong{89}. Implementation-defined aspects of
7582 @code{Task_Attributes}. See C.7.2(19).
7585 There are no implementation-defined aspects of @code{Task_Attributes}.
7590 @strong{90}. Values of all @code{Metrics}. See D(2).
7593 The metrics information for GNAT depends on the performance of the
7594 underlying operating system. The sources of the run-time for tasking
7595 implementation, together with the output from @code{-gnatG} can be
7596 used to determine the exact sequence of operating systems calls made
7597 to implement various tasking constructs. Together with appropriate
7598 information on the performance of the underlying operating system,
7599 on the exact target in use, this information can be used to determine
7600 the required metrics.
7605 @strong{91}. The declarations of @code{Any_Priority} and
7606 @code{Priority}. See D.1(11).
7609 See declarations in file @file{system.ads}.
7614 @strong{92}. Implementation-defined execution resources. See D.1(15).
7617 There are no implementation-defined execution resources.
7622 @strong{93}. Whether, on a multiprocessor, a task that is waiting for
7623 access to a protected object keeps its processor busy. See D.2.1(3).
7626 On a multi-processor, a task that is waiting for access to a protected
7627 object does not keep its processor busy.
7632 @strong{94}. The affect of implementation defined execution resources
7633 on task dispatching. See D.2.1(9).
7638 Tasks map to IRIX threads, and the dispatching policy is as defined by
7639 the IRIX implementation of threads.
7641 Tasks map to threads in the threads package used by GNAT@. Where possible
7642 and appropriate, these threads correspond to native threads of the
7643 underlying operating system.
7648 @strong{95}. Implementation-defined @code{policy_identifiers} allowed
7649 in a pragma @code{Task_Dispatching_Policy}. See D.2.2(3).
7652 There are no implementation-defined policy-identifiers allowed in this
7658 @strong{96}. Implementation-defined aspects of priority inversion. See
7662 Execution of a task cannot be preempted by the implementation processing
7663 of delay expirations for lower priority tasks.
7668 @strong{97}. Implementation defined task dispatching. See D.2.2(18).
7673 Tasks map to IRIX threads, and the dispatching policy is as defied by
7674 the IRIX implementation of threads.
7676 The policy is the same as that of the underlying threads implementation.
7681 @strong{98}. Implementation-defined @code{policy_identifiers} allowed
7682 in a pragma @code{Locking_Policy}. See D.3(4).
7685 The only implementation defined policy permitted in GNAT is
7686 @code{Inheritance_Locking}. On targets that support this policy, locking
7687 is implemented by inheritance, i.e.@: the task owning the lock operates
7688 at a priority equal to the highest priority of any task currently
7689 requesting the lock.
7694 @strong{99}. Default ceiling priorities. See D.3(10).
7697 The ceiling priority of protected objects of the type
7698 @code{System.Interrupt_Priority'Last} as described in the Ada 95
7699 Reference Manual D.3(10),
7704 @strong{100}. The ceiling of any protected object used internally by
7705 the implementation. See D.3(16).
7708 The ceiling priority of internal protected objects is
7709 @code{System.Priority'Last}.
7714 @strong{101}. Implementation-defined queuing policies. See D.4(1).
7717 There are no implementation-defined queueing policies.
7722 @strong{102}. On a multiprocessor, any conditions that cause the
7723 completion of an aborted construct to be delayed later than what is
7724 specified for a single processor. See D.6(3).
7727 The semantics for abort on a multi-processor is the same as on a single
7728 processor, there are no further delays.
7733 @strong{103}. Any operations that implicitly require heap storage
7734 allocation. See D.7(8).
7737 The only operation that implicitly requires heap storage allocation is
7743 @strong{104}. Implementation-defined aspects of pragma
7744 @code{Restrictions}. See D.7(20).
7747 There are no such implementation-defined aspects.
7752 @strong{105}. Implementation-defined aspects of package
7753 @code{Real_Time}. See D.8(17).
7756 There are no implementation defined aspects of package @code{Real_Time}.
7761 @strong{106}. Implementation-defined aspects of
7762 @code{delay_statements}. See D.9(8).
7765 Any difference greater than one microsecond will cause the task to be
7766 delayed (see D.9(7)).
7771 @strong{107}. The upper bound on the duration of interrupt blocking
7772 caused by the implementation. See D.12(5).
7775 The upper bound is determined by the underlying operating system. In
7776 no cases is it more than 10 milliseconds.
7781 @strong{108}. The means for creating and executing distributed
7785 The GLADE package provides a utility GNATDIST for creating and executing
7786 distributed programs. See the GLADE reference manual for further details.
7791 @strong{109}. Any events that can result in a partition becoming
7792 inaccessible. See E.1(7).
7795 See the GLADE reference manual for full details on such events.
7800 @strong{110}. The scheduling policies, treatment of priorities, and
7801 management of shared resources between partitions in certain cases. See
7805 See the GLADE reference manual for full details on these aspects of
7806 multi-partition execution.
7811 @strong{111}. Events that cause the version of a compilation unit to
7815 Editing the source file of a compilation unit, or the source files of
7816 any units on which it is dependent in a significant way cause the version
7817 to change. No other actions cause the version number to change. All changes
7818 are significant except those which affect only layout, capitalization or
7824 @strong{112}. Whether the execution of the remote subprogram is
7825 immediately aborted as a result of cancellation. See E.4(13).
7828 See the GLADE reference manual for details on the effect of abort in
7829 a distributed application.
7834 @strong{113}. Implementation-defined aspects of the PCS@. See E.5(25).
7837 See the GLADE reference manual for a full description of all implementation
7838 defined aspects of the PCS@.
7843 @strong{114}. Implementation-defined interfaces in the PCS@. See
7847 See the GLADE reference manual for a full description of all
7848 implementation defined interfaces.
7853 @strong{115}. The values of named numbers in the package
7854 @code{Decimal}. See F.2(7).
7866 @item Max_Decimal_Digits
7873 @strong{116}. The value of @code{Max_Picture_Length} in the package
7874 @code{Text_IO.Editing}. See F.3.3(16).
7882 @strong{117}. The value of @code{Max_Picture_Length} in the package
7883 @code{Wide_Text_IO.Editing}. See F.3.4(5).
7891 @strong{118}. The accuracy actually achieved by the complex elementary
7892 functions and by other complex arithmetic operations. See G.1(1).
7895 Standard library functions are used for the complex arithmetic
7896 operations. Only fast math mode is currently supported.
7901 @strong{119}. The sign of a zero result (or a component thereof) from
7902 any operator or function in @code{Numerics.Generic_Complex_Types}, when
7903 @code{Real'Signed_Zeros} is True. See G.1.1(53).
7906 The signs of zero values are as recommended by the relevant
7907 implementation advice.
7912 @strong{120}. The sign of a zero result (or a component thereof) from
7913 any operator or function in
7914 @code{Numerics.Generic_Complex_Elementary_Functions}, when
7915 @code{Real'Signed_Zeros} is @code{True}. See G.1.2(45).
7918 The signs of zero values are as recommended by the relevant
7919 implementation advice.
7924 @strong{121}. Whether the strict mode or the relaxed mode is the
7925 default. See G.2(2).
7928 The strict mode is the default. There is no separate relaxed mode. GNAT
7929 provides a highly efficient implementation of strict mode.
7934 @strong{122}. The result interval in certain cases of fixed-to-float
7935 conversion. See G.2.1(10).
7938 For cases where the result interval is implementation dependent, the
7939 accuracy is that provided by performing all operations in 64-bit IEEE
7940 floating-point format.
7945 @strong{123}. The result of a floating point arithmetic operation in
7946 overflow situations, when the @code{Machine_Overflows} attribute of the
7947 result type is @code{False}. See G.2.1(13).
7950 Infinite and Nan values are produced as dictated by the IEEE
7951 floating-point standard.
7956 @strong{124}. The result interval for division (or exponentiation by a
7957 negative exponent), when the floating point hardware implements division
7958 as multiplication by a reciprocal. See G.2.1(16).
7961 Not relevant, division is IEEE exact.
7966 @strong{125}. The definition of close result set, which determines the
7967 accuracy of certain fixed point multiplications and divisions. See
7971 Operations in the close result set are performed using IEEE long format
7972 floating-point arithmetic. The input operands are converted to
7973 floating-point, the operation is done in floating-point, and the result
7974 is converted to the target type.
7979 @strong{126}. Conditions on a @code{universal_real} operand of a fixed
7980 point multiplication or division for which the result shall be in the
7981 perfect result set. See G.2.3(22).
7984 The result is only defined to be in the perfect result set if the result
7985 can be computed by a single scaling operation involving a scale factor
7986 representable in 64-bits.
7991 @strong{127}. The result of a fixed point arithmetic operation in
7992 overflow situations, when the @code{Machine_Overflows} attribute of the
7993 result type is @code{False}. See G.2.3(27).
7996 Not relevant, @code{Machine_Overflows} is @code{True} for fixed-point
8002 @strong{128}. The result of an elementary function reference in
8003 overflow situations, when the @code{Machine_Overflows} attribute of the
8004 result type is @code{False}. See G.2.4(4).
8007 IEEE infinite and Nan values are produced as appropriate.
8012 @strong{129}. The value of the angle threshold, within which certain
8013 elementary functions, complex arithmetic operations, and complex
8014 elementary functions yield results conforming to a maximum relative
8015 error bound. See G.2.4(10).
8018 Information on this subject is not yet available.
8023 @strong{130}. The accuracy of certain elementary functions for
8024 parameters beyond the angle threshold. See G.2.4(10).
8027 Information on this subject is not yet available.
8032 @strong{131}. The result of a complex arithmetic operation or complex
8033 elementary function reference in overflow situations, when the
8034 @code{Machine_Overflows} attribute of the corresponding real type is
8035 @code{False}. See G.2.6(5).
8038 IEEE infinite and Nan values are produced as appropriate.
8043 @strong{132}. The accuracy of certain complex arithmetic operations and
8044 certain complex elementary functions for parameters (or components
8045 thereof) beyond the angle threshold. See G.2.6(8).
8048 Information on those subjects is not yet available.
8053 @strong{133}. Information regarding bounded errors and erroneous
8054 execution. See H.2(1).
8057 Information on this subject is not yet available.
8062 @strong{134}. Implementation-defined aspects of pragma
8063 @code{Inspection_Point}. See H.3.2(8).
8066 Pragma @code{Inspection_Point} ensures that the variable is live and can
8067 be examined by the debugger at the inspection point.
8072 @strong{135}. Implementation-defined aspects of pragma
8073 @code{Restrictions}. See H.4(25).
8076 There are no implementation-defined aspects of pragma @code{Restrictions}. The
8077 use of pragma @code{Restrictions [No_Exceptions]} has no effect on the
8078 generated code. Checks must suppressed by use of pragma @code{Suppress}.
8083 @strong{136}. Any restrictions on pragma @code{Restrictions}. See
8087 There are no restrictions on pragma @code{Restrictions}.
8089 @node Intrinsic Subprograms
8090 @chapter Intrinsic Subprograms
8091 @cindex Intrinsic Subprograms
8094 * Intrinsic Operators::
8095 * Enclosing_Entity::
8096 * Exception_Information::
8097 * Exception_Message::
8105 * Shift_Right_Arithmetic::
8110 GNAT allows a user application program to write the declaration:
8112 @smallexample @c ada
8113 pragma Import (Intrinsic, name);
8117 providing that the name corresponds to one of the implemented intrinsic
8118 subprograms in GNAT, and that the parameter profile of the referenced
8119 subprogram meets the requirements. This chapter describes the set of
8120 implemented intrinsic subprograms, and the requirements on parameter profiles.
8121 Note that no body is supplied; as with other uses of pragma Import, the
8122 body is supplied elsewhere (in this case by the compiler itself). Note
8123 that any use of this feature is potentially non-portable, since the
8124 Ada standard does not require Ada compilers to implement this feature.
8126 @node Intrinsic Operators
8127 @section Intrinsic Operators
8128 @cindex Intrinsic operator
8131 All the predefined numeric operators in package Standard
8132 in @code{pragma Import (Intrinsic,..)}
8133 declarations. In the binary operator case, the operands must have the same
8134 size. The operand or operands must also be appropriate for
8135 the operator. For example, for addition, the operands must
8136 both be floating-point or both be fixed-point, and the
8137 right operand for @code{"**"} must have a root type of
8138 @code{Standard.Integer'Base}.
8139 You can use an intrinsic operator declaration as in the following example:
8141 @smallexample @c ada
8142 type Int1 is new Integer;
8143 type Int2 is new Integer;
8145 function "+" (X1 : Int1; X2 : Int2) return Int1;
8146 function "+" (X1 : Int1; X2 : Int2) return Int2;
8147 pragma Import (Intrinsic, "+");
8151 This declaration would permit ``mixed mode'' arithmetic on items
8152 of the differing types @code{Int1} and @code{Int2}.
8153 It is also possible to specify such operators for private types, if the
8154 full views are appropriate arithmetic types.
8156 @node Enclosing_Entity
8157 @section Enclosing_Entity
8158 @cindex Enclosing_Entity
8160 This intrinsic subprogram is used in the implementation of the
8161 library routine @code{GNAT.Source_Info}. The only useful use of the
8162 intrinsic import in this case is the one in this unit, so an
8163 application program should simply call the function
8164 @code{GNAT.Source_Info.Enclosing_Entity} to obtain the name of
8165 the current subprogram, package, task, entry, or protected subprogram.
8167 @node Exception_Information
8168 @section Exception_Information
8169 @cindex Exception_Information'
8171 This intrinsic subprogram is used in the implementation of the
8172 library routine @code{GNAT.Current_Exception}. The only useful
8173 use of the intrinsic import in this case is the one in this unit,
8174 so an application program should simply call the function
8175 @code{GNAT.Current_Exception.Exception_Information} to obtain
8176 the exception information associated with the current exception.
8178 @node Exception_Message
8179 @section Exception_Message
8180 @cindex Exception_Message
8182 This intrinsic subprogram is used in the implementation of the
8183 library routine @code{GNAT.Current_Exception}. The only useful
8184 use of the intrinsic import in this case is the one in this unit,
8185 so an application program should simply call the function
8186 @code{GNAT.Current_Exception.Exception_Message} to obtain
8187 the message associated with the current exception.
8189 @node Exception_Name
8190 @section Exception_Name
8191 @cindex Exception_Name
8193 This intrinsic subprogram is used in the implementation of the
8194 library routine @code{GNAT.Current_Exception}. The only useful
8195 use of the intrinsic import in this case is the one in this unit,
8196 so an application program should simply call the function
8197 @code{GNAT.Current_Exception.Exception_Name} to obtain
8198 the name of the current exception.
8204 This intrinsic subprogram is used in the implementation of the
8205 library routine @code{GNAT.Source_Info}. The only useful use of the
8206 intrinsic import in this case is the one in this unit, so an
8207 application program should simply call the function
8208 @code{GNAT.Source_Info.File} to obtain the name of the current
8215 This intrinsic subprogram is used in the implementation of the
8216 library routine @code{GNAT.Source_Info}. The only useful use of the
8217 intrinsic import in this case is the one in this unit, so an
8218 application program should simply call the function
8219 @code{GNAT.Source_Info.Line} to obtain the number of the current
8223 @section Rotate_Left
8226 In standard Ada 95, the @code{Rotate_Left} function is available only
8227 for the predefined modular types in package @code{Interfaces}. However, in
8228 GNAT it is possible to define a Rotate_Left function for a user
8229 defined modular type or any signed integer type as in this example:
8231 @smallexample @c ada
8233 (Value : My_Modular_Type;
8235 return My_Modular_Type;
8239 The requirements are that the profile be exactly as in the example
8240 above. The only modifications allowed are in the formal parameter
8241 names, and in the type of @code{Value} and the return type, which
8242 must be the same, and must be either a signed integer type, or
8243 a modular integer type with a binary modulus, and the size must
8244 be 8. 16, 32 or 64 bits.
8247 @section Rotate_Right
8248 @cindex Rotate_Right
8250 A @code{Rotate_Right} function can be defined for any user defined
8251 binary modular integer type, or signed integer type, as described
8252 above for @code{Rotate_Left}.
8258 A @code{Shift_Left} function can be defined for any user defined
8259 binary modular integer type, or signed integer type, as described
8260 above for @code{Rotate_Left}.
8263 @section Shift_Right
8266 A @code{Shift_Right} function can be defined for any user defined
8267 binary modular integer type, or signed integer type, as described
8268 above for @code{Rotate_Left}.
8270 @node Shift_Right_Arithmetic
8271 @section Shift_Right_Arithmetic
8272 @cindex Shift_Right_Arithmetic
8274 A @code{Shift_Right_Arithmetic} function can be defined for any user
8275 defined binary modular integer type, or signed integer type, as described
8276 above for @code{Rotate_Left}.
8278 @node Source_Location
8279 @section Source_Location
8280 @cindex Source_Location
8282 This intrinsic subprogram is used in the implementation of the
8283 library routine @code{GNAT.Source_Info}. The only useful use of the
8284 intrinsic import in this case is the one in this unit, so an
8285 application program should simply call the function
8286 @code{GNAT.Source_Info.Source_Location} to obtain the current
8287 source file location.
8289 @node Representation Clauses and Pragmas
8290 @chapter Representation Clauses and Pragmas
8291 @cindex Representation Clauses
8294 * Alignment Clauses::
8296 * Storage_Size Clauses::
8297 * Size of Variant Record Objects::
8298 * Biased Representation ::
8299 * Value_Size and Object_Size Clauses::
8300 * Component_Size Clauses::
8301 * Bit_Order Clauses::
8302 * Effect of Bit_Order on Byte Ordering::
8303 * Pragma Pack for Arrays::
8304 * Pragma Pack for Records::
8305 * Record Representation Clauses::
8306 * Enumeration Clauses::
8308 * Effect of Convention on Representation::
8309 * Determining the Representations chosen by GNAT::
8313 @cindex Representation Clause
8314 @cindex Representation Pragma
8315 @cindex Pragma, representation
8316 This section describes the representation clauses accepted by GNAT, and
8317 their effect on the representation of corresponding data objects.
8319 GNAT fully implements Annex C (Systems Programming). This means that all
8320 the implementation advice sections in chapter 13 are fully implemented.
8321 However, these sections only require a minimal level of support for
8322 representation clauses. GNAT provides much more extensive capabilities,
8323 and this section describes the additional capabilities provided.
8325 @node Alignment Clauses
8326 @section Alignment Clauses
8327 @cindex Alignment Clause
8330 GNAT requires that all alignment clauses specify a power of 2, and all
8331 default alignments are always a power of 2. The default alignment
8332 values are as follows:
8335 @item @emph{Primitive Types}.
8336 For primitive types, the alignment is the minimum of the actual size of
8337 objects of the type divided by @code{Storage_Unit},
8338 and the maximum alignment supported by the target.
8339 (This maximum alignment is given by the GNAT-specific attribute
8340 @code{Standard'Maximum_Alignment}; see @ref{Maximum_Alignment}.)
8341 @cindex @code{Maximum_Alignment} attribute
8342 For example, for type @code{Long_Float}, the object size is 8 bytes, and the
8343 default alignment will be 8 on any target that supports alignments
8344 this large, but on some targets, the maximum alignment may be smaller
8345 than 8, in which case objects of type @code{Long_Float} will be maximally
8348 @item @emph{Arrays}.
8349 For arrays, the alignment is equal to the alignment of the component type
8350 for the normal case where no packing or component size is given. If the
8351 array is packed, and the packing is effective (see separate section on
8352 packed arrays), then the alignment will be one for long packed arrays,
8353 or arrays whose length is not known at compile time. For short packed
8354 arrays, which are handled internally as modular types, the alignment
8355 will be as described for primitive types, e.g.@: a packed array of length
8356 31 bits will have an object size of four bytes, and an alignment of 4.
8358 @item @emph{Records}.
8359 For the normal non-packed case, the alignment of a record is equal to
8360 the maximum alignment of any of its components. For tagged records, this
8361 includes the implicit access type used for the tag. If a pragma @code{Pack} is
8362 used and all fields are packable (see separate section on pragma @code{Pack}),
8363 then the resulting alignment is 1.
8365 A special case is when:
8368 the size of the record is given explicitly, or a
8369 full record representation clause is given, and
8371 the size of the record is 2, 4, or 8 bytes.
8374 In this case, an alignment is chosen to match the
8375 size of the record. For example, if we have:
8377 @smallexample @c ada
8378 type Small is record
8381 for Small'Size use 16;
8385 then the default alignment of the record type @code{Small} is 2, not 1. This
8386 leads to more efficient code when the record is treated as a unit, and also
8387 allows the type to specified as @code{Atomic} on architectures requiring
8393 An alignment clause may
8394 always specify a larger alignment than the default value, up to some
8395 maximum value dependent on the target (obtainable by using the
8396 attribute reference @code{Standard'Maximum_Alignment}).
8398 it is permissible to specify a smaller alignment than the default value
8399 is for a record with a record representation clause.
8400 In this case, packable fields for which a component clause is
8401 given still result in a default alignment corresponding to the original
8402 type, but this may be overridden, since these components in fact only
8403 require an alignment of one byte. For example, given
8405 @smallexample @c ada
8411 A at 0 range 0 .. 31;
8414 for V'alignment use 1;
8418 @cindex Alignment, default
8419 The default alignment for the type @code{V} is 4, as a result of the
8420 Integer field in the record, but since this field is placed with a
8421 component clause, it is permissible, as shown, to override the default
8422 alignment of the record with a smaller value.
8425 @section Size Clauses
8429 The default size for a type @code{T} is obtainable through the
8430 language-defined attribute @code{T'Size} and also through the
8431 equivalent GNAT-defined attribute @code{T'Value_Size}.
8432 For objects of type @code{T}, GNAT will generally increase the type size
8433 so that the object size (obtainable through the GNAT-defined attribute
8434 @code{T'Object_Size})
8435 is a multiple of @code{T'Alignment * Storage_Unit}.
8438 @smallexample @c ada
8439 type Smallint is range 1 .. 6;
8448 In this example, @code{Smallint'Size} = @code{Smallint'Value_Size} = 3,
8449 as specified by the RM rules,
8450 but objects of this type will have a size of 8
8451 (@code{Smallint'Object_Size} = 8),
8452 since objects by default occupy an integral number
8453 of storage units. On some targets, notably older
8454 versions of the Digital Alpha, the size of stand
8455 alone objects of this type may be 32, reflecting
8456 the inability of the hardware to do byte load/stores.
8458 Similarly, the size of type @code{Rec} is 40 bits
8459 (@code{Rec'Size} = @code{Rec'Value_Size} = 40), but
8460 the alignment is 4, so objects of this type will have
8461 their size increased to 64 bits so that it is a multiple
8462 of the alignment (in bits). This decision is
8463 in accordance with the specific Implementation Advice in RM 13.3(43):
8466 A @code{Size} clause should be supported for an object if the specified
8467 @code{Size} is at least as large as its subtype's @code{Size}, and corresponds
8468 to a size in storage elements that is a multiple of the object's
8469 @code{Alignment} (if the @code{Alignment} is nonzero).
8473 An explicit size clause may be used to override the default size by
8474 increasing it. For example, if we have:
8476 @smallexample @c ada
8477 type My_Boolean is new Boolean;
8478 for My_Boolean'Size use 32;
8482 then values of this type will always be 32 bits long. In the case of
8483 discrete types, the size can be increased up to 64 bits, with the effect
8484 that the entire specified field is used to hold the value, sign- or
8485 zero-extended as appropriate. If more than 64 bits is specified, then
8486 padding space is allocated after the value, and a warning is issued that
8487 there are unused bits.
8489 Similarly the size of records and arrays may be increased, and the effect
8490 is to add padding bits after the value. This also causes a warning message
8493 The largest Size value permitted in GNAT is 2**31@minus{}1. Since this is a
8494 Size in bits, this corresponds to an object of size 256 megabytes (minus
8495 one). This limitation is true on all targets. The reason for this
8496 limitation is that it improves the quality of the code in many cases
8497 if it is known that a Size value can be accommodated in an object of
8500 @node Storage_Size Clauses
8501 @section Storage_Size Clauses
8502 @cindex Storage_Size Clause
8505 For tasks, the @code{Storage_Size} clause specifies the amount of space
8506 to be allocated for the task stack. This cannot be extended, and if the
8507 stack is exhausted, then @code{Storage_Error} will be raised (if stack
8508 checking is enabled). Use a @code{Storage_Size} attribute definition clause,
8509 or a @code{Storage_Size} pragma in the task definition to set the
8510 appropriate required size. A useful technique is to include in every
8511 task definition a pragma of the form:
8513 @smallexample @c ada
8514 pragma Storage_Size (Default_Stack_Size);
8518 Then @code{Default_Stack_Size} can be defined in a global package, and
8519 modified as required. Any tasks requiring stack sizes different from the
8520 default can have an appropriate alternative reference in the pragma.
8522 For access types, the @code{Storage_Size} clause specifies the maximum
8523 space available for allocation of objects of the type. If this space is
8524 exceeded then @code{Storage_Error} will be raised by an allocation attempt.
8525 In the case where the access type is declared local to a subprogram, the
8526 use of a @code{Storage_Size} clause triggers automatic use of a special
8527 predefined storage pool (@code{System.Pool_Size}) that ensures that all
8528 space for the pool is automatically reclaimed on exit from the scope in
8529 which the type is declared.
8531 A special case recognized by the compiler is the specification of a
8532 @code{Storage_Size} of zero for an access type. This means that no
8533 items can be allocated from the pool, and this is recognized at compile
8534 time, and all the overhead normally associated with maintaining a fixed
8535 size storage pool is eliminated. Consider the following example:
8537 @smallexample @c ada
8539 type R is array (Natural) of Character;
8540 type P is access all R;
8541 for P'Storage_Size use 0;
8542 -- Above access type intended only for interfacing purposes
8546 procedure g (m : P);
8547 pragma Import (C, g);
8558 As indicated in this example, these dummy storage pools are often useful in
8559 connection with interfacing where no object will ever be allocated. If you
8560 compile the above example, you get the warning:
8563 p.adb:16:09: warning: allocation from empty storage pool
8564 p.adb:16:09: warning: Storage_Error will be raised at run time
8568 Of course in practice, there will not be any explicit allocators in the
8569 case of such an access declaration.
8571 @node Size of Variant Record Objects
8572 @section Size of Variant Record Objects
8573 @cindex Size, variant record objects
8574 @cindex Variant record objects, size
8577 In the case of variant record objects, there is a question whether Size gives
8578 information about a particular variant, or the maximum size required
8579 for any variant. Consider the following program
8581 @smallexample @c ada
8582 with Text_IO; use Text_IO;
8584 type R1 (A : Boolean := False) is record
8586 when True => X : Character;
8595 Put_Line (Integer'Image (V1'Size));
8596 Put_Line (Integer'Image (V2'Size));
8601 Here we are dealing with a variant record, where the True variant
8602 requires 16 bits, and the False variant requires 8 bits.
8603 In the above example, both V1 and V2 contain the False variant,
8604 which is only 8 bits long. However, the result of running the
8613 The reason for the difference here is that the discriminant value of
8614 V1 is fixed, and will always be False. It is not possible to assign
8615 a True variant value to V1, therefore 8 bits is sufficient. On the
8616 other hand, in the case of V2, the initial discriminant value is
8617 False (from the default), but it is possible to assign a True
8618 variant value to V2, therefore 16 bits must be allocated for V2
8619 in the general case, even fewer bits may be needed at any particular
8620 point during the program execution.
8622 As can be seen from the output of this program, the @code{'Size}
8623 attribute applied to such an object in GNAT gives the actual allocated
8624 size of the variable, which is the largest size of any of the variants.
8625 The Ada Reference Manual is not completely clear on what choice should
8626 be made here, but the GNAT behavior seems most consistent with the
8627 language in the RM@.
8629 In some cases, it may be desirable to obtain the size of the current
8630 variant, rather than the size of the largest variant. This can be
8631 achieved in GNAT by making use of the fact that in the case of a
8632 subprogram parameter, GNAT does indeed return the size of the current
8633 variant (because a subprogram has no way of knowing how much space
8634 is actually allocated for the actual).
8636 Consider the following modified version of the above program:
8638 @smallexample @c ada
8639 with Text_IO; use Text_IO;
8641 type R1 (A : Boolean := False) is record
8643 when True => X : Character;
8650 function Size (V : R1) return Integer is
8656 Put_Line (Integer'Image (V2'Size));
8657 Put_Line (Integer'IMage (Size (V2)));
8659 Put_Line (Integer'Image (V2'Size));
8660 Put_Line (Integer'IMage (Size (V2)));
8665 The output from this program is
8675 Here we see that while the @code{'Size} attribute always returns
8676 the maximum size, regardless of the current variant value, the
8677 @code{Size} function does indeed return the size of the current
8680 @node Biased Representation
8681 @section Biased Representation
8682 @cindex Size for biased representation
8683 @cindex Biased representation
8686 In the case of scalars with a range starting at other than zero, it is
8687 possible in some cases to specify a size smaller than the default minimum
8688 value, and in such cases, GNAT uses an unsigned biased representation,
8689 in which zero is used to represent the lower bound, and successive values
8690 represent successive values of the type.
8692 For example, suppose we have the declaration:
8694 @smallexample @c ada
8695 type Small is range -7 .. -4;
8696 for Small'Size use 2;
8700 Although the default size of type @code{Small} is 4, the @code{Size}
8701 clause is accepted by GNAT and results in the following representation
8705 -7 is represented as 2#00#
8706 -6 is represented as 2#01#
8707 -5 is represented as 2#10#
8708 -4 is represented as 2#11#
8712 Biased representation is only used if the specified @code{Size} clause
8713 cannot be accepted in any other manner. These reduced sizes that force
8714 biased representation can be used for all discrete types except for
8715 enumeration types for which a representation clause is given.
8717 @node Value_Size and Object_Size Clauses
8718 @section Value_Size and Object_Size Clauses
8721 @cindex Size, of objects
8724 In Ada 95, @code{T'Size} for a type @code{T} is the minimum number of bits
8725 required to hold values of type @code{T}. Although this interpretation was
8726 allowed in Ada 83, it was not required, and this requirement in practice
8727 can cause some significant difficulties. For example, in most Ada 83
8728 compilers, @code{Natural'Size} was 32. However, in Ada 95,
8729 @code{Natural'Size} is
8730 typically 31. This means that code may change in behavior when moving
8731 from Ada 83 to Ada 95. For example, consider:
8733 @smallexample @c ada
8740 at 0 range 0 .. Natural'Size - 1;
8741 at 0 range Natural'Size .. 2 * Natural'Size - 1;
8746 In the above code, since the typical size of @code{Natural} objects
8747 is 32 bits and @code{Natural'Size} is 31, the above code can cause
8748 unexpected inefficient packing in Ada 95, and in general there are
8749 cases where the fact that the object size can exceed the
8750 size of the type causes surprises.
8752 To help get around this problem GNAT provides two implementation
8753 defined attributes, @code{Value_Size} and @code{Object_Size}. When
8754 applied to a type, these attributes yield the size of the type
8755 (corresponding to the RM defined size attribute), and the size of
8756 objects of the type respectively.
8758 The @code{Object_Size} is used for determining the default size of
8759 objects and components. This size value can be referred to using the
8760 @code{Object_Size} attribute. The phrase ``is used'' here means that it is
8761 the basis of the determination of the size. The backend is free to
8762 pad this up if necessary for efficiency, e.g.@: an 8-bit stand-alone
8763 character might be stored in 32 bits on a machine with no efficient
8764 byte access instructions such as the Alpha.
8766 The default rules for the value of @code{Object_Size} for
8767 discrete types are as follows:
8771 The @code{Object_Size} for base subtypes reflect the natural hardware
8772 size in bits (run the compiler with @option{-gnatS} to find those values
8773 for numeric types). Enumeration types and fixed-point base subtypes have
8774 8, 16, 32 or 64 bits for this size, depending on the range of values
8778 The @code{Object_Size} of a subtype is the same as the
8779 @code{Object_Size} of
8780 the type from which it is obtained.
8783 The @code{Object_Size} of a derived base type is copied from the parent
8784 base type, and the @code{Object_Size} of a derived first subtype is copied
8785 from the parent first subtype.
8789 The @code{Value_Size} attribute
8790 is the (minimum) number of bits required to store a value
8792 This value is used to determine how tightly to pack
8793 records or arrays with components of this type, and also affects
8794 the semantics of unchecked conversion (unchecked conversions where
8795 the @code{Value_Size} values differ generate a warning, and are potentially
8798 The default rules for the value of @code{Value_Size} are as follows:
8802 The @code{Value_Size} for a base subtype is the minimum number of bits
8803 required to store all values of the type (including the sign bit
8804 only if negative values are possible).
8807 If a subtype statically matches the first subtype of a given type, then it has
8808 by default the same @code{Value_Size} as the first subtype. This is a
8809 consequence of RM 13.1(14) (``if two subtypes statically match,
8810 then their subtype-specific aspects are the same''.)
8813 All other subtypes have a @code{Value_Size} corresponding to the minimum
8814 number of bits required to store all values of the subtype. For
8815 dynamic bounds, it is assumed that the value can range down or up
8816 to the corresponding bound of the ancestor
8820 The RM defined attribute @code{Size} corresponds to the
8821 @code{Value_Size} attribute.
8823 The @code{Size} attribute may be defined for a first-named subtype. This sets
8824 the @code{Value_Size} of
8825 the first-named subtype to the given value, and the
8826 @code{Object_Size} of this first-named subtype to the given value padded up
8827 to an appropriate boundary. It is a consequence of the default rules
8828 above that this @code{Object_Size} will apply to all further subtypes. On the
8829 other hand, @code{Value_Size} is affected only for the first subtype, any
8830 dynamic subtypes obtained from it directly, and any statically matching
8831 subtypes. The @code{Value_Size} of any other static subtypes is not affected.
8833 @code{Value_Size} and
8834 @code{Object_Size} may be explicitly set for any subtype using
8835 an attribute definition clause. Note that the use of these attributes
8836 can cause the RM 13.1(14) rule to be violated. If two access types
8837 reference aliased objects whose subtypes have differing @code{Object_Size}
8838 values as a result of explicit attribute definition clauses, then it
8839 is erroneous to convert from one access subtype to the other.
8841 At the implementation level, Esize stores the Object_Size and the
8842 RM_Size field stores the @code{Value_Size} (and hence the value of the
8843 @code{Size} attribute,
8844 which, as noted above, is equivalent to @code{Value_Size}).
8846 To get a feel for the difference, consider the following examples (note
8847 that in each case the base is @code{Short_Short_Integer} with a size of 8):
8850 Object_Size Value_Size
8852 type x1 is range 0 .. 5; 8 3
8854 type x2 is range 0 .. 5;
8855 for x2'size use 12; 16 12
8857 subtype x3 is x2 range 0 .. 3; 16 2
8859 subtype x4 is x2'base range 0 .. 10; 8 4
8861 subtype x5 is x2 range 0 .. dynamic; 16 3*
8863 subtype x6 is x2'base range 0 .. dynamic; 8 3*
8868 Note: the entries marked ``3*'' are not actually specified by the Ada 95 RM,
8869 but it seems in the spirit of the RM rules to allocate the minimum number
8870 of bits (here 3, given the range for @code{x2})
8871 known to be large enough to hold the given range of values.
8873 So far, so good, but GNAT has to obey the RM rules, so the question is
8874 under what conditions must the RM @code{Size} be used.
8875 The following is a list
8876 of the occasions on which the RM @code{Size} must be used:
8880 Component size for packed arrays or records
8883 Value of the attribute @code{Size} for a type
8886 Warning about sizes not matching for unchecked conversion
8890 For record types, the @code{Object_Size} is always a multiple of the
8891 alignment of the type (this is true for all types). In some cases the
8892 @code{Value_Size} can be smaller. Consider:
8902 On a typical 32-bit architecture, the X component will be four bytes, and
8903 require four-byte alignment, and the Y component will be one byte. In this
8904 case @code{R'Value_Size} will be 40 (bits) since this is the minimum size
8905 required to store a value of this type, and for example, it is permissible
8906 to have a component of type R in an outer record whose component size is
8907 specified to be 48 bits. However, @code{R'Object_Size} will be 64 (bits),
8908 since it must be rounded up so that this value is a multiple of the
8909 alignment (4 bytes = 32 bits).
8912 For all other types, the @code{Object_Size}
8913 and Value_Size are the same (and equivalent to the RM attribute @code{Size}).
8914 Only @code{Size} may be specified for such types.
8916 @node Component_Size Clauses
8917 @section Component_Size Clauses
8918 @cindex Component_Size Clause
8921 Normally, the value specified in a component clause must be consistent
8922 with the subtype of the array component with regard to size and alignment.
8923 In other words, the value specified must be at least equal to the size
8924 of this subtype, and must be a multiple of the alignment value.
8926 In addition, component size clauses are allowed which cause the array
8927 to be packed, by specifying a smaller value. The cases in which this
8928 is allowed are for component size values in the range 1 through 63. The value
8929 specified must not be smaller than the Size of the subtype. GNAT will
8930 accurately honor all packing requests in this range. For example, if
8933 @smallexample @c ada
8934 type r is array (1 .. 8) of Natural;
8935 for r'Component_Size use 31;
8939 then the resulting array has a length of 31 bytes (248 bits = 8 * 31).
8940 Of course access to the components of such an array is considerably
8941 less efficient than if the natural component size of 32 is used.
8943 @node Bit_Order Clauses
8944 @section Bit_Order Clauses
8945 @cindex Bit_Order Clause
8946 @cindex bit ordering
8947 @cindex ordering, of bits
8950 For record subtypes, GNAT permits the specification of the @code{Bit_Order}
8951 attribute. The specification may either correspond to the default bit
8952 order for the target, in which case the specification has no effect and
8953 places no additional restrictions, or it may be for the non-standard
8954 setting (that is the opposite of the default).
8956 In the case where the non-standard value is specified, the effect is
8957 to renumber bits within each byte, but the ordering of bytes is not
8958 affected. There are certain
8959 restrictions placed on component clauses as follows:
8963 @item Components fitting within a single storage unit.
8965 These are unrestricted, and the effect is merely to renumber bits. For
8966 example if we are on a little-endian machine with @code{Low_Order_First}
8967 being the default, then the following two declarations have exactly
8970 @smallexample @c ada
8973 B : Integer range 1 .. 120;
8977 A at 0 range 0 .. 0;
8978 B at 0 range 1 .. 7;
8983 B : Integer range 1 .. 120;
8986 for R2'Bit_Order use High_Order_First;
8989 A at 0 range 7 .. 7;
8990 B at 0 range 0 .. 6;
8995 The useful application here is to write the second declaration with the
8996 @code{Bit_Order} attribute definition clause, and know that it will be treated
8997 the same, regardless of whether the target is little-endian or big-endian.
8999 @item Components occupying an integral number of bytes.
9001 These are components that exactly fit in two or more bytes. Such component
9002 declarations are allowed, but have no effect, since it is important to realize
9003 that the @code{Bit_Order} specification does not affect the ordering of bytes.
9004 In particular, the following attempt at getting an endian-independent integer
9007 @smallexample @c ada
9012 for R2'Bit_Order use High_Order_First;
9015 A at 0 range 0 .. 31;
9020 This declaration will result in a little-endian integer on a
9021 little-endian machine, and a big-endian integer on a big-endian machine.
9022 If byte flipping is required for interoperability between big- and
9023 little-endian machines, this must be explicitly programmed. This capability
9024 is not provided by @code{Bit_Order}.
9026 @item Components that are positioned across byte boundaries
9028 but do not occupy an integral number of bytes. Given that bytes are not
9029 reordered, such fields would occupy a non-contiguous sequence of bits
9030 in memory, requiring non-trivial code to reassemble. They are for this
9031 reason not permitted, and any component clause specifying such a layout
9032 will be flagged as illegal by GNAT@.
9037 Since the misconception that Bit_Order automatically deals with all
9038 endian-related incompatibilities is a common one, the specification of
9039 a component field that is an integral number of bytes will always
9040 generate a warning. This warning may be suppressed using
9041 @code{pragma Suppress} if desired. The following section contains additional
9042 details regarding the issue of byte ordering.
9044 @node Effect of Bit_Order on Byte Ordering
9045 @section Effect of Bit_Order on Byte Ordering
9046 @cindex byte ordering
9047 @cindex ordering, of bytes
9050 In this section we will review the effect of the @code{Bit_Order} attribute
9051 definition clause on byte ordering. Briefly, it has no effect at all, but
9052 a detailed example will be helpful. Before giving this
9053 example, let us review the precise
9054 definition of the effect of defining @code{Bit_Order}. The effect of a
9055 non-standard bit order is described in section 15.5.3 of the Ada
9059 2 A bit ordering is a method of interpreting the meaning of
9060 the storage place attributes.
9064 To understand the precise definition of storage place attributes in
9065 this context, we visit section 13.5.1 of the manual:
9068 13 A record_representation_clause (without the mod_clause)
9069 specifies the layout. The storage place attributes (see 13.5.2)
9070 are taken from the values of the position, first_bit, and last_bit
9071 expressions after normalizing those values so that first_bit is
9072 less than Storage_Unit.
9076 The critical point here is that storage places are taken from
9077 the values after normalization, not before. So the @code{Bit_Order}
9078 interpretation applies to normalized values. The interpretation
9079 is described in the later part of the 15.5.3 paragraph:
9082 2 A bit ordering is a method of interpreting the meaning of
9083 the storage place attributes. High_Order_First (known in the
9084 vernacular as ``big endian'') means that the first bit of a
9085 storage element (bit 0) is the most significant bit (interpreting
9086 the sequence of bits that represent a component as an unsigned
9087 integer value). Low_Order_First (known in the vernacular as
9088 ``little endian'') means the opposite: the first bit is the
9093 Note that the numbering is with respect to the bits of a storage
9094 unit. In other words, the specification affects only the numbering
9095 of bits within a single storage unit.
9097 We can make the effect clearer by giving an example.
9099 Suppose that we have an external device which presents two bytes, the first
9100 byte presented, which is the first (low addressed byte) of the two byte
9101 record is called Master, and the second byte is called Slave.
9103 The left most (most significant bit is called Control for each byte, and
9104 the remaining 7 bits are called V1, V2, @dots{} V7, where V7 is the rightmost
9105 (least significant) bit.
9107 On a big-endian machine, we can write the following representation clause
9109 @smallexample @c ada
9111 Master_Control : Bit;
9119 Slave_Control : Bit;
9130 Master_Control at 0 range 0 .. 0;
9131 Master_V1 at 0 range 1 .. 1;
9132 Master_V2 at 0 range 2 .. 2;
9133 Master_V3 at 0 range 3 .. 3;
9134 Master_V4 at 0 range 4 .. 4;
9135 Master_V5 at 0 range 5 .. 5;
9136 Master_V6 at 0 range 6 .. 6;
9137 Master_V7 at 0 range 7 .. 7;
9138 Slave_Control at 1 range 0 .. 0;
9139 Slave_V1 at 1 range 1 .. 1;
9140 Slave_V2 at 1 range 2 .. 2;
9141 Slave_V3 at 1 range 3 .. 3;
9142 Slave_V4 at 1 range 4 .. 4;
9143 Slave_V5 at 1 range 5 .. 5;
9144 Slave_V6 at 1 range 6 .. 6;
9145 Slave_V7 at 1 range 7 .. 7;
9150 Now if we move this to a little endian machine, then the bit ordering within
9151 the byte is backwards, so we have to rewrite the record rep clause as:
9153 @smallexample @c ada
9155 Master_Control at 0 range 7 .. 7;
9156 Master_V1 at 0 range 6 .. 6;
9157 Master_V2 at 0 range 5 .. 5;
9158 Master_V3 at 0 range 4 .. 4;
9159 Master_V4 at 0 range 3 .. 3;
9160 Master_V5 at 0 range 2 .. 2;
9161 Master_V6 at 0 range 1 .. 1;
9162 Master_V7 at 0 range 0 .. 0;
9163 Slave_Control at 1 range 7 .. 7;
9164 Slave_V1 at 1 range 6 .. 6;
9165 Slave_V2 at 1 range 5 .. 5;
9166 Slave_V3 at 1 range 4 .. 4;
9167 Slave_V4 at 1 range 3 .. 3;
9168 Slave_V5 at 1 range 2 .. 2;
9169 Slave_V6 at 1 range 1 .. 1;
9170 Slave_V7 at 1 range 0 .. 0;
9175 It is a nuisance to have to rewrite the clause, especially if
9176 the code has to be maintained on both machines. However,
9177 this is a case that we can handle with the
9178 @code{Bit_Order} attribute if it is implemented.
9179 Note that the implementation is not required on byte addressed
9180 machines, but it is indeed implemented in GNAT.
9181 This means that we can simply use the
9182 first record clause, together with the declaration
9184 @smallexample @c ada
9185 for Data'Bit_Order use High_Order_First;
9189 and the effect is what is desired, namely the layout is exactly the same,
9190 independent of whether the code is compiled on a big-endian or little-endian
9193 The important point to understand is that byte ordering is not affected.
9194 A @code{Bit_Order} attribute definition never affects which byte a field
9195 ends up in, only where it ends up in that byte.
9196 To make this clear, let us rewrite the record rep clause of the previous
9199 @smallexample @c ada
9200 for Data'Bit_Order use High_Order_First;
9202 Master_Control at 0 range 0 .. 0;
9203 Master_V1 at 0 range 1 .. 1;
9204 Master_V2 at 0 range 2 .. 2;
9205 Master_V3 at 0 range 3 .. 3;
9206 Master_V4 at 0 range 4 .. 4;
9207 Master_V5 at 0 range 5 .. 5;
9208 Master_V6 at 0 range 6 .. 6;
9209 Master_V7 at 0 range 7 .. 7;
9210 Slave_Control at 0 range 8 .. 8;
9211 Slave_V1 at 0 range 9 .. 9;
9212 Slave_V2 at 0 range 10 .. 10;
9213 Slave_V3 at 0 range 11 .. 11;
9214 Slave_V4 at 0 range 12 .. 12;
9215 Slave_V5 at 0 range 13 .. 13;
9216 Slave_V6 at 0 range 14 .. 14;
9217 Slave_V7 at 0 range 15 .. 15;
9222 This is exactly equivalent to saying (a repeat of the first example):
9224 @smallexample @c ada
9225 for Data'Bit_Order use High_Order_First;
9227 Master_Control at 0 range 0 .. 0;
9228 Master_V1 at 0 range 1 .. 1;
9229 Master_V2 at 0 range 2 .. 2;
9230 Master_V3 at 0 range 3 .. 3;
9231 Master_V4 at 0 range 4 .. 4;
9232 Master_V5 at 0 range 5 .. 5;
9233 Master_V6 at 0 range 6 .. 6;
9234 Master_V7 at 0 range 7 .. 7;
9235 Slave_Control at 1 range 0 .. 0;
9236 Slave_V1 at 1 range 1 .. 1;
9237 Slave_V2 at 1 range 2 .. 2;
9238 Slave_V3 at 1 range 3 .. 3;
9239 Slave_V4 at 1 range 4 .. 4;
9240 Slave_V5 at 1 range 5 .. 5;
9241 Slave_V6 at 1 range 6 .. 6;
9242 Slave_V7 at 1 range 7 .. 7;
9247 Why are they equivalent? Well take a specific field, the @code{Slave_V2}
9248 field. The storage place attributes are obtained by normalizing the
9249 values given so that the @code{First_Bit} value is less than 8. After
9250 normalizing the values (0,10,10) we get (1,2,2) which is exactly what
9251 we specified in the other case.
9253 Now one might expect that the @code{Bit_Order} attribute might affect
9254 bit numbering within the entire record component (two bytes in this
9255 case, thus affecting which byte fields end up in), but that is not
9256 the way this feature is defined, it only affects numbering of bits,
9257 not which byte they end up in.
9259 Consequently it never makes sense to specify a starting bit number
9260 greater than 7 (for a byte addressable field) if an attribute
9261 definition for @code{Bit_Order} has been given, and indeed it
9262 may be actively confusing to specify such a value, so the compiler
9263 generates a warning for such usage.
9265 If you do need to control byte ordering then appropriate conditional
9266 values must be used. If in our example, the slave byte came first on
9267 some machines we might write:
9269 @smallexample @c ada
9270 Master_Byte_First constant Boolean := @dots{};
9272 Master_Byte : constant Natural :=
9273 1 - Boolean'Pos (Master_Byte_First);
9274 Slave_Byte : constant Natural :=
9275 Boolean'Pos (Master_Byte_First);
9277 for Data'Bit_Order use High_Order_First;
9279 Master_Control at Master_Byte range 0 .. 0;
9280 Master_V1 at Master_Byte range 1 .. 1;
9281 Master_V2 at Master_Byte range 2 .. 2;
9282 Master_V3 at Master_Byte range 3 .. 3;
9283 Master_V4 at Master_Byte range 4 .. 4;
9284 Master_V5 at Master_Byte range 5 .. 5;
9285 Master_V6 at Master_Byte range 6 .. 6;
9286 Master_V7 at Master_Byte range 7 .. 7;
9287 Slave_Control at Slave_Byte range 0 .. 0;
9288 Slave_V1 at Slave_Byte range 1 .. 1;
9289 Slave_V2 at Slave_Byte range 2 .. 2;
9290 Slave_V3 at Slave_Byte range 3 .. 3;
9291 Slave_V4 at Slave_Byte range 4 .. 4;
9292 Slave_V5 at Slave_Byte range 5 .. 5;
9293 Slave_V6 at Slave_Byte range 6 .. 6;
9294 Slave_V7 at Slave_Byte range 7 .. 7;
9299 Now to switch between machines, all that is necessary is
9300 to set the boolean constant @code{Master_Byte_First} in
9301 an appropriate manner.
9303 @node Pragma Pack for Arrays
9304 @section Pragma Pack for Arrays
9305 @cindex Pragma Pack (for arrays)
9308 Pragma @code{Pack} applied to an array has no effect unless the component type
9309 is packable. For a component type to be packable, it must be one of the
9316 Any type whose size is specified with a size clause
9318 Any packed array type with a static size
9322 For all these cases, if the component subtype size is in the range
9323 1 through 63, then the effect of the pragma @code{Pack} is exactly as though a
9324 component size were specified giving the component subtype size.
9325 For example if we have:
9327 @smallexample @c ada
9328 type r is range 0 .. 17;
9330 type ar is array (1 .. 8) of r;
9335 Then the component size of @code{ar} will be set to 5 (i.e.@: to @code{r'size},
9336 and the size of the array @code{ar} will be exactly 40 bits.
9338 Note that in some cases this rather fierce approach to packing can produce
9339 unexpected effects. For example, in Ada 95, type Natural typically has a
9340 size of 31, meaning that if you pack an array of Natural, you get 31-bit
9341 close packing, which saves a few bits, but results in far less efficient
9342 access. Since many other Ada compilers will ignore such a packing request,
9343 GNAT will generate a warning on some uses of pragma @code{Pack} that it guesses
9344 might not be what is intended. You can easily remove this warning by
9345 using an explicit @code{Component_Size} setting instead, which never generates
9346 a warning, since the intention of the programmer is clear in this case.
9348 GNAT treats packed arrays in one of two ways. If the size of the array is
9349 known at compile time and is less than 64 bits, then internally the array
9350 is represented as a single modular type, of exactly the appropriate number
9351 of bits. If the length is greater than 63 bits, or is not known at compile
9352 time, then the packed array is represented as an array of bytes, and the
9353 length is always a multiple of 8 bits.
9355 Note that to represent a packed array as a modular type, the alignment must
9356 be suitable for the modular type involved. For example, on typical machines
9357 a 32-bit packed array will be represented by a 32-bit modular integer with
9358 an alignment of four bytes. If you explicitly override the default alignment
9359 with an alignment clause that is too small, the modular representation
9360 cannot be used. For example, consider the following set of declarations:
9362 @smallexample @c ada
9363 type R is range 1 .. 3;
9364 type S is array (1 .. 31) of R;
9365 for S'Component_Size use 2;
9367 for S'Alignment use 1;
9371 If the alignment clause were not present, then a 62-bit modular
9372 representation would be chosen (typically with an alignment of 4 or 8
9373 bytes depending on the target). But the default alignment is overridden
9374 with the explicit alignment clause. This means that the modular
9375 representation cannot be used, and instead the array of bytes
9376 representation must be used, meaning that the length must be a multiple
9377 of 8. Thus the above set of declarations will result in a diagnostic
9378 rejecting the size clause and noting that the minimum size allowed is 64.
9380 @cindex Pragma Pack (for type Natural)
9381 @cindex Pragma Pack warning
9383 One special case that is worth noting occurs when the base type of the
9384 component size is 8/16/32 and the subtype is one bit less. Notably this
9385 occurs with subtype @code{Natural}. Consider:
9387 @smallexample @c ada
9388 type Arr is array (1 .. 32) of Natural;
9393 In all commonly used Ada 83 compilers, this pragma Pack would be ignored,
9394 since typically @code{Natural'Size} is 32 in Ada 83, and in any case most
9395 Ada 83 compilers did not attempt 31 bit packing.
9397 In Ada 95, @code{Natural'Size} is required to be 31. Furthermore, GNAT really
9398 does pack 31-bit subtype to 31 bits. This may result in a substantial
9399 unintended performance penalty when porting legacy Ada 83 code. To help
9400 prevent this, GNAT generates a warning in such cases. If you really want 31
9401 bit packing in a case like this, you can set the component size explicitly:
9403 @smallexample @c ada
9404 type Arr is array (1 .. 32) of Natural;
9405 for Arr'Component_Size use 31;
9409 Here 31-bit packing is achieved as required, and no warning is generated,
9410 since in this case the programmer intention is clear.
9412 @node Pragma Pack for Records
9413 @section Pragma Pack for Records
9414 @cindex Pragma Pack (for records)
9417 Pragma @code{Pack} applied to a record will pack the components to reduce
9418 wasted space from alignment gaps and by reducing the amount of space
9419 taken by components. We distinguish between @emph{packable} components and
9420 @emph{non-packable} components.
9421 Components of the following types are considered packable:
9424 All primitive types are packable.
9427 Small packed arrays, whose size does not exceed 64 bits, and where the
9428 size is statically known at compile time, are represented internally
9429 as modular integers, and so they are also packable.
9434 All packable components occupy the exact number of bits corresponding to
9435 their @code{Size} value, and are packed with no padding bits, i.e.@: they
9436 can start on an arbitrary bit boundary.
9438 All other types are non-packable, they occupy an integral number of
9440 are placed at a boundary corresponding to their alignment requirements.
9442 For example, consider the record
9444 @smallexample @c ada
9445 type Rb1 is array (1 .. 13) of Boolean;
9448 type Rb2 is array (1 .. 65) of Boolean;
9463 The representation for the record x2 is as follows:
9465 @smallexample @c ada
9466 for x2'Size use 224;
9468 l1 at 0 range 0 .. 0;
9469 l2 at 0 range 1 .. 64;
9470 l3 at 12 range 0 .. 31;
9471 l4 at 16 range 0 .. 0;
9472 l5 at 16 range 1 .. 13;
9473 l6 at 18 range 0 .. 71;
9478 Studying this example, we see that the packable fields @code{l1}
9480 of length equal to their sizes, and placed at specific bit boundaries (and
9481 not byte boundaries) to
9482 eliminate padding. But @code{l3} is of a non-packable float type, so
9483 it is on the next appropriate alignment boundary.
9485 The next two fields are fully packable, so @code{l4} and @code{l5} are
9486 minimally packed with no gaps. However, type @code{Rb2} is a packed
9487 array that is longer than 64 bits, so it is itself non-packable. Thus
9488 the @code{l6} field is aligned to the next byte boundary, and takes an
9489 integral number of bytes, i.e.@: 72 bits.
9491 @node Record Representation Clauses
9492 @section Record Representation Clauses
9493 @cindex Record Representation Clause
9496 Record representation clauses may be given for all record types, including
9497 types obtained by record extension. Component clauses are allowed for any
9498 static component. The restrictions on component clauses depend on the type
9501 @cindex Component Clause
9502 For all components of an elementary type, the only restriction on component
9503 clauses is that the size must be at least the 'Size value of the type
9504 (actually the Value_Size). There are no restrictions due to alignment,
9505 and such components may freely cross storage boundaries.
9507 Packed arrays with a size up to and including 64 bits are represented
9508 internally using a modular type with the appropriate number of bits, and
9509 thus the same lack of restriction applies. For example, if you declare:
9511 @smallexample @c ada
9512 type R is array (1 .. 49) of Boolean;
9518 then a component clause for a component of type R may start on any
9519 specified bit boundary, and may specify a value of 49 bits or greater.
9521 Packed bit arrays that are longer than 64 bits must always be placed
9522 on a storage unit (byte) boundary. Any component clause that does not
9523 meet this requirement will be rejected.
9525 The rules for other types are different for GNAT 3 and GNAT 5 versions
9526 (based on GCC 2 and GCC 3 respectively). In GNAT 5, larger components
9527 (other than packed arrays)
9528 may also be placed on arbitrary boundaries, so for example, the following
9531 @smallexample @c ada
9532 type R is array (1 .. 10) of Boolean;
9541 G at 0 range 0 .. 0;
9542 H at 0 range 1 .. 1;
9543 L at 0 range 2 .. 81;
9544 R at 0 range 82 .. 161;
9549 In GNAT 3, there are more severe restrictions on larger components.
9550 For non-primitive types, including packed arrays with a size greater than
9551 64 bits, component clauses must respect the alignment requirement of the
9552 type, in particular, always starting on a byte boundary, and the length
9553 must be a multiple of the storage unit.
9555 The following rules regarding tagged types are enforced in both GNAT 3 and
9558 The tag field of a tagged type always occupies an address sized field at
9559 the start of the record. No component clause may attempt to overlay this
9562 In the case of a record extension T1, of a type T, no component clause applied
9563 to the type T1 can specify a storage location that would overlap the first
9564 T'Size bytes of the record.
9566 @node Enumeration Clauses
9567 @section Enumeration Clauses
9569 The only restriction on enumeration clauses is that the range of values
9570 must be representable. For the signed case, if one or more of the
9571 representation values are negative, all values must be in the range:
9573 @smallexample @c ada
9574 System.Min_Int .. System.Max_Int
9578 For the unsigned case, where all values are non negative, the values must
9581 @smallexample @c ada
9582 0 .. System.Max_Binary_Modulus;
9586 A @emph{confirming} representation clause is one in which the values range
9587 from 0 in sequence, i.e.@: a clause that confirms the default representation
9588 for an enumeration type.
9589 Such a confirming representation
9590 is permitted by these rules, and is specially recognized by the compiler so
9591 that no extra overhead results from the use of such a clause.
9593 If an array has an index type which is an enumeration type to which an
9594 enumeration clause has been applied, then the array is stored in a compact
9595 manner. Consider the declarations:
9597 @smallexample @c ada
9598 type r is (A, B, C);
9599 for r use (A => 1, B => 5, C => 10);
9600 type t is array (r) of Character;
9604 The array type t corresponds to a vector with exactly three elements and
9605 has a default size equal to @code{3*Character'Size}. This ensures efficient
9606 use of space, but means that accesses to elements of the array will incur
9607 the overhead of converting representation values to the corresponding
9608 positional values, (i.e.@: the value delivered by the @code{Pos} attribute).
9610 @node Address Clauses
9611 @section Address Clauses
9612 @cindex Address Clause
9614 The reference manual allows a general restriction on representation clauses,
9615 as found in RM 13.1(22):
9618 An implementation need not support representation
9619 items containing nonstatic expressions, except that
9620 an implementation should support a representation item
9621 for a given entity if each nonstatic expression in the
9622 representation item is a name that statically denotes
9623 a constant declared before the entity.
9627 In practice this is applicable only to address clauses, since this is the
9628 only case in which a non-static expression is permitted by the syntax. As
9629 the AARM notes in sections 13.1 (22.a-22.h):
9632 22.a Reason: This is to avoid the following sort of thing:
9634 22.b X : Integer := F(@dots{});
9635 Y : Address := G(@dots{});
9636 for X'Address use Y;
9638 22.c In the above, we have to evaluate the
9639 initialization expression for X before we
9640 know where to put the result. This seems
9641 like an unreasonable implementation burden.
9643 22.d The above code should instead be written
9646 22.e Y : constant Address := G(@dots{});
9647 X : Integer := F(@dots{});
9648 for X'Address use Y;
9650 22.f This allows the expression ``Y'' to be safely
9651 evaluated before X is created.
9653 22.g The constant could be a formal parameter of mode in.
9655 22.h An implementation can support other nonstatic
9656 expressions if it wants to. Expressions of type
9657 Address are hardly ever static, but their value
9658 might be known at compile time anyway in many
9663 GNAT does indeed permit many additional cases of non-static expressions. In
9664 particular, if the type involved is elementary there are no restrictions
9665 (since in this case, holding a temporary copy of the initialization value,
9666 if one is present, is inexpensive). In addition, if there is no implicit or
9667 explicit initialization, then there are no restrictions. GNAT will reject
9668 only the case where all three of these conditions hold:
9673 The type of the item is non-elementary (e.g.@: a record or array).
9676 There is explicit or implicit initialization required for the object.
9677 Note that access values are always implicitly initialized, and also
9678 in GNAT, certain bit-packed arrays (those having a dynamic length or
9679 a length greater than 64) will also be implicitly initialized to zero.
9682 The address value is non-static. Here GNAT is more permissive than the
9683 RM, and allows the address value to be the address of a previously declared
9684 stand-alone variable, as long as it does not itself have an address clause.
9686 @smallexample @c ada
9687 Anchor : Some_Initialized_Type;
9688 Overlay : Some_Initialized_Type;
9689 for Overlay'Address use Anchor'Address;
9693 However, the prefix of the address clause cannot be an array component, or
9694 a component of a discriminated record.
9699 As noted above in section 22.h, address values are typically non-static. In
9700 particular the To_Address function, even if applied to a literal value, is
9701 a non-static function call. To avoid this minor annoyance, GNAT provides
9702 the implementation defined attribute 'To_Address. The following two
9703 expressions have identical values:
9707 @smallexample @c ada
9708 To_Address (16#1234_0000#)
9709 System'To_Address (16#1234_0000#);
9713 except that the second form is considered to be a static expression, and
9714 thus when used as an address clause value is always permitted.
9717 Additionally, GNAT treats as static an address clause that is an
9718 unchecked_conversion of a static integer value. This simplifies the porting
9719 of legacy code, and provides a portable equivalent to the GNAT attribute
9722 Another issue with address clauses is the interaction with alignment
9723 requirements. When an address clause is given for an object, the address
9724 value must be consistent with the alignment of the object (which is usually
9725 the same as the alignment of the type of the object). If an address clause
9726 is given that specifies an inappropriately aligned address value, then the
9727 program execution is erroneous.
9729 Since this source of erroneous behavior can have unfortunate effects, GNAT
9730 checks (at compile time if possible, generating a warning, or at execution
9731 time with a run-time check) that the alignment is appropriate. If the
9732 run-time check fails, then @code{Program_Error} is raised. This run-time
9733 check is suppressed if range checks are suppressed, or if
9734 @code{pragma Restrictions (No_Elaboration_Code)} is in effect.
9737 An address clause cannot be given for an exported object. More
9738 understandably the real restriction is that objects with an address
9739 clause cannot be exported. This is because such variables are not
9740 defined by the Ada program, so there is no external object to export.
9743 It is permissible to give an address clause and a pragma Import for the
9744 same object. In this case, the variable is not really defined by the
9745 Ada program, so there is no external symbol to be linked. The link name
9746 and the external name are ignored in this case. The reason that we allow this
9747 combination is that it provides a useful idiom to avoid unwanted
9748 initializations on objects with address clauses.
9750 When an address clause is given for an object that has implicit or
9751 explicit initialization, then by default initialization takes place. This
9752 means that the effect of the object declaration is to overwrite the
9753 memory at the specified address. This is almost always not what the
9754 programmer wants, so GNAT will output a warning:
9764 for Ext'Address use System'To_Address (16#1234_1234#);
9766 >>> warning: implicit initialization of "Ext" may
9767 modify overlaid storage
9768 >>> warning: use pragma Import for "Ext" to suppress
9769 initialization (RM B(24))
9775 As indicated by the warning message, the solution is to use a (dummy) pragma
9776 Import to suppress this initialization. The pragma tell the compiler that the
9777 object is declared and initialized elsewhere. The following package compiles
9778 without warnings (and the initialization is suppressed):
9780 @smallexample @c ada
9788 for Ext'Address use System'To_Address (16#1234_1234#);
9789 pragma Import (Ada, Ext);
9794 A final issue with address clauses involves their use for overlaying
9795 variables, as in the following example:
9796 @cindex Overlaying of objects
9798 @smallexample @c ada
9801 for B'Address use A'Address;
9805 or alternatively, using the form recommended by the RM:
9807 @smallexample @c ada
9809 Addr : constant Address := A'Address;
9811 for B'Address use Addr;
9815 In both of these cases, @code{A}
9816 and @code{B} become aliased to one another via the
9817 address clause. This use of address clauses to overlay
9818 variables, achieving an effect similar to unchecked
9819 conversion was erroneous in Ada 83, but in Ada 95
9820 the effect is implementation defined. Furthermore, the
9821 Ada 95 RM specifically recommends that in a situation
9822 like this, @code{B} should be subject to the following
9823 implementation advice (RM 13.3(19)):
9826 19 If the Address of an object is specified, or it is imported
9827 or exported, then the implementation should not perform
9828 optimizations based on assumptions of no aliases.
9832 GNAT follows this recommendation, and goes further by also applying
9833 this recommendation to the overlaid variable (@code{A}
9834 in the above example) in this case. This means that the overlay
9835 works "as expected", in that a modification to one of the variables
9836 will affect the value of the other.
9838 @node Effect of Convention on Representation
9839 @section Effect of Convention on Representation
9840 @cindex Convention, effect on representation
9843 Normally the specification of a foreign language convention for a type or
9844 an object has no effect on the chosen representation. In particular, the
9845 representation chosen for data in GNAT generally meets the standard system
9846 conventions, and for example records are laid out in a manner that is
9847 consistent with C@. This means that specifying convention C (for example)
9850 There are three exceptions to this general rule:
9854 @item Convention Fortran and array subtypes
9855 If pragma Convention Fortran is specified for an array subtype, then in
9856 accordance with the implementation advice in section 3.6.2(11) of the
9857 Ada Reference Manual, the array will be stored in a Fortran-compatible
9858 column-major manner, instead of the normal default row-major order.
9860 @item Convention C and enumeration types
9861 GNAT normally stores enumeration types in 8, 16, or 32 bits as required
9862 to accommodate all values of the type. For example, for the enumeration
9865 @smallexample @c ada
9866 type Color is (Red, Green, Blue);
9870 8 bits is sufficient to store all values of the type, so by default, objects
9871 of type @code{Color} will be represented using 8 bits. However, normal C
9872 convention is to use 32 bits for all enum values in C, since enum values
9873 are essentially of type int. If pragma @code{Convention C} is specified for an
9874 Ada enumeration type, then the size is modified as necessary (usually to
9875 32 bits) to be consistent with the C convention for enum values.
9877 @item Convention C/Fortran and Boolean types
9878 In C, the usual convention for boolean values, that is values used for
9879 conditions, is that zero represents false, and nonzero values represent
9880 true. In Ada, the normal convention is that two specific values, typically
9881 0/1, are used to represent false/true respectively.
9883 Fortran has a similar convention for @code{LOGICAL} values (any nonzero
9884 value represents true).
9886 To accommodate the Fortran and C conventions, if a pragma Convention specifies
9887 C or Fortran convention for a derived Boolean, as in the following example:
9889 @smallexample @c ada
9890 type C_Switch is new Boolean;
9891 pragma Convention (C, C_Switch);
9895 then the GNAT generated code will treat any nonzero value as true. For truth
9896 values generated by GNAT, the conventional value 1 will be used for True, but
9897 when one of these values is read, any nonzero value is treated as True.
9901 @node Determining the Representations chosen by GNAT
9902 @section Determining the Representations chosen by GNAT
9903 @cindex Representation, determination of
9904 @cindex @code{-gnatR} switch
9907 Although the descriptions in this section are intended to be complete, it is
9908 often easier to simply experiment to see what GNAT accepts and what the
9909 effect is on the layout of types and objects.
9911 As required by the Ada RM, if a representation clause is not accepted, then
9912 it must be rejected as illegal by the compiler. However, when a
9913 representation clause or pragma is accepted, there can still be questions
9914 of what the compiler actually does. For example, if a partial record
9915 representation clause specifies the location of some components and not
9916 others, then where are the non-specified components placed? Or if pragma
9917 @code{Pack} is used on a record, then exactly where are the resulting
9918 fields placed? The section on pragma @code{Pack} in this chapter can be
9919 used to answer the second question, but it is often easier to just see
9920 what the compiler does.
9922 For this purpose, GNAT provides the option @code{-gnatR}. If you compile
9923 with this option, then the compiler will output information on the actual
9924 representations chosen, in a format similar to source representation
9925 clauses. For example, if we compile the package:
9927 @smallexample @c ada
9929 type r (x : boolean) is tagged record
9931 when True => S : String (1 .. 100);
9936 type r2 is new r (false) with record
9941 y2 at 16 range 0 .. 31;
9948 type x1 is array (1 .. 10) of x;
9949 for x1'component_size use 11;
9951 type ia is access integer;
9953 type Rb1 is array (1 .. 13) of Boolean;
9956 type Rb2 is array (1 .. 65) of Boolean;
9972 using the switch @code{-gnatR} we obtain the following output:
9975 Representation information for unit q
9976 -------------------------------------
9979 for r'Alignment use 4;
9981 x at 4 range 0 .. 7;
9982 _tag at 0 range 0 .. 31;
9983 s at 5 range 0 .. 799;
9986 for r2'Size use 160;
9987 for r2'Alignment use 4;
9989 x at 4 range 0 .. 7;
9990 _tag at 0 range 0 .. 31;
9991 _parent at 0 range 0 .. 63;
9992 y2 at 16 range 0 .. 31;
9996 for x'Alignment use 1;
9998 y at 0 range 0 .. 7;
10001 for x1'Size use 112;
10002 for x1'Alignment use 1;
10003 for x1'Component_Size use 11;
10005 for rb1'Size use 13;
10006 for rb1'Alignment use 2;
10007 for rb1'Component_Size use 1;
10009 for rb2'Size use 72;
10010 for rb2'Alignment use 1;
10011 for rb2'Component_Size use 1;
10013 for x2'Size use 224;
10014 for x2'Alignment use 4;
10016 l1 at 0 range 0 .. 0;
10017 l2 at 0 range 1 .. 64;
10018 l3 at 12 range 0 .. 31;
10019 l4 at 16 range 0 .. 0;
10020 l5 at 16 range 1 .. 13;
10021 l6 at 18 range 0 .. 71;
10026 The Size values are actually the Object_Size, i.e.@: the default size that
10027 will be allocated for objects of the type.
10028 The ?? size for type r indicates that we have a variant record, and the
10029 actual size of objects will depend on the discriminant value.
10031 The Alignment values show the actual alignment chosen by the compiler
10032 for each record or array type.
10034 The record representation clause for type r shows where all fields
10035 are placed, including the compiler generated tag field (whose location
10036 cannot be controlled by the programmer).
10038 The record representation clause for the type extension r2 shows all the
10039 fields present, including the parent field, which is a copy of the fields
10040 of the parent type of r2, i.e.@: r1.
10042 The component size and size clauses for types rb1 and rb2 show
10043 the exact effect of pragma @code{Pack} on these arrays, and the record
10044 representation clause for type x2 shows how pragma @code{Pack} affects
10047 In some cases, it may be useful to cut and paste the representation clauses
10048 generated by the compiler into the original source to fix and guarantee
10049 the actual representation to be used.
10051 @node Standard Library Routines
10052 @chapter Standard Library Routines
10055 The Ada 95 Reference Manual contains in Annex A a full description of an
10056 extensive set of standard library routines that can be used in any Ada
10057 program, and which must be provided by all Ada compilers. They are
10058 analogous to the standard C library used by C programs.
10060 GNAT implements all of the facilities described in annex A, and for most
10061 purposes the description in the Ada 95
10062 reference manual, or appropriate Ada
10063 text book, will be sufficient for making use of these facilities.
10065 In the case of the input-output facilities, @xref{The Implementation of
10066 Standard I/O}, gives details on exactly how GNAT interfaces to the
10067 file system. For the remaining packages, the Ada 95 reference manual
10068 should be sufficient. The following is a list of the packages included,
10069 together with a brief description of the functionality that is provided.
10071 For completeness, references are included to other predefined library
10072 routines defined in other sections of the Ada 95 reference manual (these are
10073 cross-indexed from annex A).
10077 This is a parent package for all the standard library packages. It is
10078 usually included implicitly in your program, and itself contains no
10079 useful data or routines.
10081 @item Ada.Calendar (9.6)
10082 @code{Calendar} provides time of day access, and routines for
10083 manipulating times and durations.
10085 @item Ada.Characters (A.3.1)
10086 This is a dummy parent package that contains no useful entities
10088 @item Ada.Characters.Handling (A.3.2)
10089 This package provides some basic character handling capabilities,
10090 including classification functions for classes of characters (e.g.@: test
10091 for letters, or digits).
10093 @item Ada.Characters.Latin_1 (A.3.3)
10094 This package includes a complete set of definitions of the characters
10095 that appear in type CHARACTER@. It is useful for writing programs that
10096 will run in international environments. For example, if you want an
10097 upper case E with an acute accent in a string, it is often better to use
10098 the definition of @code{UC_E_Acute} in this package. Then your program
10099 will print in an understandable manner even if your environment does not
10100 support these extended characters.
10102 @item Ada.Command_Line (A.15)
10103 This package provides access to the command line parameters and the name
10104 of the current program (analogous to the use of @code{argc} and @code{argv}
10105 in C), and also allows the exit status for the program to be set in a
10106 system-independent manner.
10108 @item Ada.Decimal (F.2)
10109 This package provides constants describing the range of decimal numbers
10110 implemented, and also a decimal divide routine (analogous to the COBOL
10111 verb DIVIDE .. GIVING .. REMAINDER ..)
10113 @item Ada.Direct_IO (A.8.4)
10114 This package provides input-output using a model of a set of records of
10115 fixed-length, containing an arbitrary definite Ada type, indexed by an
10116 integer record number.
10118 @item Ada.Dynamic_Priorities (D.5)
10119 This package allows the priorities of a task to be adjusted dynamically
10120 as the task is running.
10122 @item Ada.Exceptions (11.4.1)
10123 This package provides additional information on exceptions, and also
10124 contains facilities for treating exceptions as data objects, and raising
10125 exceptions with associated messages.
10127 @item Ada.Finalization (7.6)
10128 This package contains the declarations and subprograms to support the
10129 use of controlled types, providing for automatic initialization and
10130 finalization (analogous to the constructors and destructors of C++)
10132 @item Ada.Interrupts (C.3.2)
10133 This package provides facilities for interfacing to interrupts, which
10134 includes the set of signals or conditions that can be raised and
10135 recognized as interrupts.
10137 @item Ada.Interrupts.Names (C.3.2)
10138 This package provides the set of interrupt names (actually signal
10139 or condition names) that can be handled by GNAT@.
10141 @item Ada.IO_Exceptions (A.13)
10142 This package defines the set of exceptions that can be raised by use of
10143 the standard IO packages.
10146 This package contains some standard constants and exceptions used
10147 throughout the numerics packages. Note that the constants pi and e are
10148 defined here, and it is better to use these definitions than rolling
10151 @item Ada.Numerics.Complex_Elementary_Functions
10152 Provides the implementation of standard elementary functions (such as
10153 log and trigonometric functions) operating on complex numbers using the
10154 standard @code{Float} and the @code{Complex} and @code{Imaginary} types
10155 created by the package @code{Numerics.Complex_Types}.
10157 @item Ada.Numerics.Complex_Types
10158 This is a predefined instantiation of
10159 @code{Numerics.Generic_Complex_Types} using @code{Standard.Float} to
10160 build the type @code{Complex} and @code{Imaginary}.
10162 @item Ada.Numerics.Discrete_Random
10163 This package provides a random number generator suitable for generating
10164 random integer values from a specified range.
10166 @item Ada.Numerics.Float_Random
10167 This package provides a random number generator suitable for generating
10168 uniformly distributed floating point values.
10170 @item Ada.Numerics.Generic_Complex_Elementary_Functions
10171 This is a generic version of the package that provides the
10172 implementation of standard elementary functions (such as log and
10173 trigonometric functions) for an arbitrary complex type.
10175 The following predefined instantiations of this package are provided:
10179 @code{Ada.Numerics.Short_Complex_Elementary_Functions}
10181 @code{Ada.Numerics.Complex_Elementary_Functions}
10183 @code{Ada.Numerics.
10184 Long_Complex_Elementary_Functions}
10187 @item Ada.Numerics.Generic_Complex_Types
10188 This is a generic package that allows the creation of complex types,
10189 with associated complex arithmetic operations.
10191 The following predefined instantiations of this package exist
10194 @code{Ada.Numerics.Short_Complex_Complex_Types}
10196 @code{Ada.Numerics.Complex_Complex_Types}
10198 @code{Ada.Numerics.Long_Complex_Complex_Types}
10201 @item Ada.Numerics.Generic_Elementary_Functions
10202 This is a generic package that provides the implementation of standard
10203 elementary functions (such as log an trigonometric functions) for an
10204 arbitrary float type.
10206 The following predefined instantiations of this package exist
10210 @code{Ada.Numerics.Short_Elementary_Functions}
10212 @code{Ada.Numerics.Elementary_Functions}
10214 @code{Ada.Numerics.Long_Elementary_Functions}
10217 @item Ada.Real_Time (D.8)
10218 This package provides facilities similar to those of @code{Calendar}, but
10219 operating with a finer clock suitable for real time control. Note that
10220 annex D requires that there be no backward clock jumps, and GNAT generally
10221 guarantees this behavior, but of course if the external clock on which
10222 the GNAT runtime depends is deliberately reset by some external event,
10223 then such a backward jump may occur.
10225 @item Ada.Sequential_IO (A.8.1)
10226 This package provides input-output facilities for sequential files,
10227 which can contain a sequence of values of a single type, which can be
10228 any Ada type, including indefinite (unconstrained) types.
10230 @item Ada.Storage_IO (A.9)
10231 This package provides a facility for mapping arbitrary Ada types to and
10232 from a storage buffer. It is primarily intended for the creation of new
10235 @item Ada.Streams (13.13.1)
10236 This is a generic package that provides the basic support for the
10237 concept of streams as used by the stream attributes (@code{Input},
10238 @code{Output}, @code{Read} and @code{Write}).
10240 @item Ada.Streams.Stream_IO (A.12.1)
10241 This package is a specialization of the type @code{Streams} defined in
10242 package @code{Streams} together with a set of operations providing
10243 Stream_IO capability. The Stream_IO model permits both random and
10244 sequential access to a file which can contain an arbitrary set of values
10245 of one or more Ada types.
10247 @item Ada.Strings (A.4.1)
10248 This package provides some basic constants used by the string handling
10251 @item Ada.Strings.Bounded (A.4.4)
10252 This package provides facilities for handling variable length
10253 strings. The bounded model requires a maximum length. It is thus
10254 somewhat more limited than the unbounded model, but avoids the use of
10255 dynamic allocation or finalization.
10257 @item Ada.Strings.Fixed (A.4.3)
10258 This package provides facilities for handling fixed length strings.
10260 @item Ada.Strings.Maps (A.4.2)
10261 This package provides facilities for handling character mappings and
10262 arbitrarily defined subsets of characters. For instance it is useful in
10263 defining specialized translation tables.
10265 @item Ada.Strings.Maps.Constants (A.4.6)
10266 This package provides a standard set of predefined mappings and
10267 predefined character sets. For example, the standard upper to lower case
10268 conversion table is found in this package. Note that upper to lower case
10269 conversion is non-trivial if you want to take the entire set of
10270 characters, including extended characters like E with an acute accent,
10271 into account. You should use the mappings in this package (rather than
10272 adding 32 yourself) to do case mappings.
10274 @item Ada.Strings.Unbounded (A.4.5)
10275 This package provides facilities for handling variable length
10276 strings. The unbounded model allows arbitrary length strings, but
10277 requires the use of dynamic allocation and finalization.
10279 @item Ada.Strings.Wide_Bounded (A.4.7)
10280 @itemx Ada.Strings.Wide_Fixed (A.4.7)
10281 @itemx Ada.Strings.Wide_Maps (A.4.7)
10282 @itemx Ada.Strings.Wide_Maps.Constants (A.4.7)
10283 @itemx Ada.Strings.Wide_Unbounded (A.4.7)
10284 These packages provide analogous capabilities to the corresponding
10285 packages without @samp{Wide_} in the name, but operate with the types
10286 @code{Wide_String} and @code{Wide_Character} instead of @code{String}
10287 and @code{Character}.
10289 @item Ada.Synchronous_Task_Control (D.10)
10290 This package provides some standard facilities for controlling task
10291 communication in a synchronous manner.
10294 This package contains definitions for manipulation of the tags of tagged
10297 @item Ada.Task_Attributes
10298 This package provides the capability of associating arbitrary
10299 task-specific data with separate tasks.
10302 This package provides basic text input-output capabilities for
10303 character, string and numeric data. The subpackages of this
10304 package are listed next.
10306 @item Ada.Text_IO.Decimal_IO
10307 Provides input-output facilities for decimal fixed-point types
10309 @item Ada.Text_IO.Enumeration_IO
10310 Provides input-output facilities for enumeration types.
10312 @item Ada.Text_IO.Fixed_IO
10313 Provides input-output facilities for ordinary fixed-point types.
10315 @item Ada.Text_IO.Float_IO
10316 Provides input-output facilities for float types. The following
10317 predefined instantiations of this generic package are available:
10321 @code{Short_Float_Text_IO}
10323 @code{Float_Text_IO}
10325 @code{Long_Float_Text_IO}
10328 @item Ada.Text_IO.Integer_IO
10329 Provides input-output facilities for integer types. The following
10330 predefined instantiations of this generic package are available:
10333 @item Short_Short_Integer
10334 @code{Ada.Short_Short_Integer_Text_IO}
10335 @item Short_Integer
10336 @code{Ada.Short_Integer_Text_IO}
10338 @code{Ada.Integer_Text_IO}
10340 @code{Ada.Long_Integer_Text_IO}
10341 @item Long_Long_Integer
10342 @code{Ada.Long_Long_Integer_Text_IO}
10345 @item Ada.Text_IO.Modular_IO
10346 Provides input-output facilities for modular (unsigned) types
10348 @item Ada.Text_IO.Complex_IO (G.1.3)
10349 This package provides basic text input-output capabilities for complex
10352 @item Ada.Text_IO.Editing (F.3.3)
10353 This package contains routines for edited output, analogous to the use
10354 of pictures in COBOL@. The picture formats used by this package are a
10355 close copy of the facility in COBOL@.
10357 @item Ada.Text_IO.Text_Streams (A.12.2)
10358 This package provides a facility that allows Text_IO files to be treated
10359 as streams, so that the stream attributes can be used for writing
10360 arbitrary data, including binary data, to Text_IO files.
10362 @item Ada.Unchecked_Conversion (13.9)
10363 This generic package allows arbitrary conversion from one type to
10364 another of the same size, providing for breaking the type safety in
10365 special circumstances.
10367 If the types have the same Size (more accurately the same Value_Size),
10368 then the effect is simply to transfer the bits from the source to the
10369 target type without any modification. This usage is well defined, and
10370 for simple types whose representation is typically the same across
10371 all implementations, gives a portable method of performing such
10374 If the types do not have the same size, then the result is implementation
10375 defined, and thus may be non-portable. The following describes how GNAT
10376 handles such unchecked conversion cases.
10378 If the types are of different sizes, and are both discrete types, then
10379 the effect is of a normal type conversion without any constraint checking.
10380 In particular if the result type has a larger size, the result will be
10381 zero or sign extended. If the result type has a smaller size, the result
10382 will be truncated by ignoring high order bits.
10384 If the types are of different sizes, and are not both discrete types,
10385 then the conversion works as though pointers were created to the source
10386 and target, and the pointer value is converted. The effect is that bits
10387 are copied from successive low order storage units and bits of the source
10388 up to the length of the target type.
10390 A warning is issued if the lengths differ, since the effect in this
10391 case is implementation dependent, and the above behavior may not match
10392 that of some other compiler.
10394 A pointer to one type may be converted to a pointer to another type using
10395 unchecked conversion. The only case in which the effect is undefined is
10396 when one or both pointers are pointers to unconstrained array types. In
10397 this case, the bounds information may get incorrectly transferred, and in
10398 particular, GNAT uses double size pointers for such types, and it is
10399 meaningless to convert between such pointer types. GNAT will issue a
10400 warning if the alignment of the target designated type is more strict
10401 than the alignment of the source designated type (since the result may
10402 be unaligned in this case).
10404 A pointer other than a pointer to an unconstrained array type may be
10405 converted to and from System.Address. Such usage is common in Ada 83
10406 programs, but note that Ada.Address_To_Access_Conversions is the
10407 preferred method of performing such conversions in Ada 95. Neither
10408 unchecked conversion nor Ada.Address_To_Access_Conversions should be
10409 used in conjunction with pointers to unconstrained objects, since
10410 the bounds information cannot be handled correctly in this case.
10412 @item Ada.Unchecked_Deallocation (13.11.2)
10413 This generic package allows explicit freeing of storage previously
10414 allocated by use of an allocator.
10416 @item Ada.Wide_Text_IO (A.11)
10417 This package is similar to @code{Ada.Text_IO}, except that the external
10418 file supports wide character representations, and the internal types are
10419 @code{Wide_Character} and @code{Wide_String} instead of @code{Character}
10420 and @code{String}. It contains generic subpackages listed next.
10422 @item Ada.Wide_Text_IO.Decimal_IO
10423 Provides input-output facilities for decimal fixed-point types
10425 @item Ada.Wide_Text_IO.Enumeration_IO
10426 Provides input-output facilities for enumeration types.
10428 @item Ada.Wide_Text_IO.Fixed_IO
10429 Provides input-output facilities for ordinary fixed-point types.
10431 @item Ada.Wide_Text_IO.Float_IO
10432 Provides input-output facilities for float types. The following
10433 predefined instantiations of this generic package are available:
10437 @code{Short_Float_Wide_Text_IO}
10439 @code{Float_Wide_Text_IO}
10441 @code{Long_Float_Wide_Text_IO}
10444 @item Ada.Wide_Text_IO.Integer_IO
10445 Provides input-output facilities for integer types. The following
10446 predefined instantiations of this generic package are available:
10449 @item Short_Short_Integer
10450 @code{Ada.Short_Short_Integer_Wide_Text_IO}
10451 @item Short_Integer
10452 @code{Ada.Short_Integer_Wide_Text_IO}
10454 @code{Ada.Integer_Wide_Text_IO}
10456 @code{Ada.Long_Integer_Wide_Text_IO}
10457 @item Long_Long_Integer
10458 @code{Ada.Long_Long_Integer_Wide_Text_IO}
10461 @item Ada.Wide_Text_IO.Modular_IO
10462 Provides input-output facilities for modular (unsigned) types
10464 @item Ada.Wide_Text_IO.Complex_IO (G.1.3)
10465 This package is similar to @code{Ada.Text_IO.Complex_IO}, except that the
10466 external file supports wide character representations.
10468 @item Ada.Wide_Text_IO.Editing (F.3.4)
10469 This package is similar to @code{Ada.Text_IO.Editing}, except that the
10470 types are @code{Wide_Character} and @code{Wide_String} instead of
10471 @code{Character} and @code{String}.
10473 @item Ada.Wide_Text_IO.Streams (A.12.3)
10474 This package is similar to @code{Ada.Text_IO.Streams}, except that the
10475 types are @code{Wide_Character} and @code{Wide_String} instead of
10476 @code{Character} and @code{String}.
10479 @node The Implementation of Standard I/O
10480 @chapter The Implementation of Standard I/O
10483 GNAT implements all the required input-output facilities described in
10484 A.6 through A.14. These sections of the Ada 95 reference manual describe the
10485 required behavior of these packages from the Ada point of view, and if
10486 you are writing a portable Ada program that does not need to know the
10487 exact manner in which Ada maps to the outside world when it comes to
10488 reading or writing external files, then you do not need to read this
10489 chapter. As long as your files are all regular files (not pipes or
10490 devices), and as long as you write and read the files only from Ada, the
10491 description in the Ada 95 reference manual is sufficient.
10493 However, if you want to do input-output to pipes or other devices, such
10494 as the keyboard or screen, or if the files you are dealing with are
10495 either generated by some other language, or to be read by some other
10496 language, then you need to know more about the details of how the GNAT
10497 implementation of these input-output facilities behaves.
10499 In this chapter we give a detailed description of exactly how GNAT
10500 interfaces to the file system. As always, the sources of the system are
10501 available to you for answering questions at an even more detailed level,
10502 but for most purposes the information in this chapter will suffice.
10504 Another reason that you may need to know more about how input-output is
10505 implemented arises when you have a program written in mixed languages
10506 where, for example, files are shared between the C and Ada sections of
10507 the same program. GNAT provides some additional facilities, in the form
10508 of additional child library packages, that facilitate this sharing, and
10509 these additional facilities are also described in this chapter.
10512 * Standard I/O Packages::
10521 * Operations on C Streams::
10522 * Interfacing to C Streams::
10525 @node Standard I/O Packages
10526 @section Standard I/O Packages
10529 The Standard I/O packages described in Annex A for
10535 Ada.Text_IO.Complex_IO
10537 Ada.Text_IO.Text_Streams,
10541 Ada.Wide_Text_IO.Complex_IO,
10543 Ada.Wide_Text_IO.Text_Streams
10553 are implemented using the C
10554 library streams facility; where
10558 All files are opened using @code{fopen}.
10560 All input/output operations use @code{fread}/@code{fwrite}.
10564 There is no internal buffering of any kind at the Ada library level. The
10565 only buffering is that provided at the system level in the
10566 implementation of the C library routines that support streams. This
10567 facilitates shared use of these streams by mixed language programs.
10570 @section FORM Strings
10573 The format of a FORM string in GNAT is:
10576 "keyword=value,keyword=value,@dots{},keyword=value"
10580 where letters may be in upper or lower case, and there are no spaces
10581 between values. The order of the entries is not important. Currently
10582 there are two keywords defined.
10590 The use of these parameters is described later in this section.
10596 Direct_IO can only be instantiated for definite types. This is a
10597 restriction of the Ada language, which means that the records are fixed
10598 length (the length being determined by @code{@var{type}'Size}, rounded
10599 up to the next storage unit boundary if necessary).
10601 The records of a Direct_IO file are simply written to the file in index
10602 sequence, with the first record starting at offset zero, and subsequent
10603 records following. There is no control information of any kind. For
10604 example, if 32-bit integers are being written, each record takes
10605 4-bytes, so the record at index @var{K} starts at offset
10606 (@var{K}@minus{}1)*4.
10608 There is no limit on the size of Direct_IO files, they are expanded as
10609 necessary to accommodate whatever records are written to the file.
10611 @node Sequential_IO
10612 @section Sequential_IO
10615 Sequential_IO may be instantiated with either a definite (constrained)
10616 or indefinite (unconstrained) type.
10618 For the definite type case, the elements written to the file are simply
10619 the memory images of the data values with no control information of any
10620 kind. The resulting file should be read using the same type, no validity
10621 checking is performed on input.
10623 For the indefinite type case, the elements written consist of two
10624 parts. First is the size of the data item, written as the memory image
10625 of a @code{Interfaces.C.size_t} value, followed by the memory image of
10626 the data value. The resulting file can only be read using the same
10627 (unconstrained) type. Normal assignment checks are performed on these
10628 read operations, and if these checks fail, @code{Data_Error} is
10629 raised. In particular, in the array case, the lengths must match, and in
10630 the variant record case, if the variable for a particular read operation
10631 is constrained, the discriminants must match.
10633 Note that it is not possible to use Sequential_IO to write variable
10634 length array items, and then read the data back into different length
10635 arrays. For example, the following will raise @code{Data_Error}:
10637 @smallexample @c ada
10638 package IO is new Sequential_IO (String);
10643 IO.Write (F, "hello!")
10644 IO.Reset (F, Mode=>In_File);
10651 On some Ada implementations, this will print @code{hell}, but the program is
10652 clearly incorrect, since there is only one element in the file, and that
10653 element is the string @code{hello!}.
10655 In Ada 95, this kind of behavior can be legitimately achieved using
10656 Stream_IO, and this is the preferred mechanism. In particular, the above
10657 program fragment rewritten to use Stream_IO will work correctly.
10663 Text_IO files consist of a stream of characters containing the following
10664 special control characters:
10667 LF (line feed, 16#0A#) Line Mark
10668 FF (form feed, 16#0C#) Page Mark
10672 A canonical Text_IO file is defined as one in which the following
10673 conditions are met:
10677 The character @code{LF} is used only as a line mark, i.e.@: to mark the end
10681 The character @code{FF} is used only as a page mark, i.e.@: to mark the
10682 end of a page and consequently can appear only immediately following a
10683 @code{LF} (line mark) character.
10686 The file ends with either @code{LF} (line mark) or @code{LF}-@code{FF}
10687 (line mark, page mark). In the former case, the page mark is implicitly
10688 assumed to be present.
10692 A file written using Text_IO will be in canonical form provided that no
10693 explicit @code{LF} or @code{FF} characters are written using @code{Put}
10694 or @code{Put_Line}. There will be no @code{FF} character at the end of
10695 the file unless an explicit @code{New_Page} operation was performed
10696 before closing the file.
10698 A canonical Text_IO file that is a regular file, i.e.@: not a device or a
10699 pipe, can be read using any of the routines in Text_IO@. The
10700 semantics in this case will be exactly as defined in the Ada 95 reference
10701 manual and all the routines in Text_IO are fully implemented.
10703 A text file that does not meet the requirements for a canonical Text_IO
10704 file has one of the following:
10708 The file contains @code{FF} characters not immediately following a
10709 @code{LF} character.
10712 The file contains @code{LF} or @code{FF} characters written by
10713 @code{Put} or @code{Put_Line}, which are not logically considered to be
10714 line marks or page marks.
10717 The file ends in a character other than @code{LF} or @code{FF},
10718 i.e.@: there is no explicit line mark or page mark at the end of the file.
10722 Text_IO can be used to read such non-standard text files but subprograms
10723 to do with line or page numbers do not have defined meanings. In
10724 particular, a @code{FF} character that does not follow a @code{LF}
10725 character may or may not be treated as a page mark from the point of
10726 view of page and line numbering. Every @code{LF} character is considered
10727 to end a line, and there is an implied @code{LF} character at the end of
10731 * Text_IO Stream Pointer Positioning::
10732 * Text_IO Reading and Writing Non-Regular Files::
10734 * Treating Text_IO Files as Streams::
10735 * Text_IO Extensions::
10736 * Text_IO Facilities for Unbounded Strings::
10739 @node Text_IO Stream Pointer Positioning
10740 @subsection Stream Pointer Positioning
10743 @code{Ada.Text_IO} has a definition of current position for a file that
10744 is being read. No internal buffering occurs in Text_IO, and usually the
10745 physical position in the stream used to implement the file corresponds
10746 to this logical position defined by Text_IO@. There are two exceptions:
10750 After a call to @code{End_Of_Page} that returns @code{True}, the stream
10751 is positioned past the @code{LF} (line mark) that precedes the page
10752 mark. Text_IO maintains an internal flag so that subsequent read
10753 operations properly handle the logical position which is unchanged by
10754 the @code{End_Of_Page} call.
10757 After a call to @code{End_Of_File} that returns @code{True}, if the
10758 Text_IO file was positioned before the line mark at the end of file
10759 before the call, then the logical position is unchanged, but the stream
10760 is physically positioned right at the end of file (past the line mark,
10761 and past a possible page mark following the line mark. Again Text_IO
10762 maintains internal flags so that subsequent read operations properly
10763 handle the logical position.
10767 These discrepancies have no effect on the observable behavior of
10768 Text_IO, but if a single Ada stream is shared between a C program and
10769 Ada program, or shared (using @samp{shared=yes} in the form string)
10770 between two Ada files, then the difference may be observable in some
10773 @node Text_IO Reading and Writing Non-Regular Files
10774 @subsection Reading and Writing Non-Regular Files
10777 A non-regular file is a device (such as a keyboard), or a pipe. Text_IO
10778 can be used for reading and writing. Writing is not affected and the
10779 sequence of characters output is identical to the normal file case, but
10780 for reading, the behavior of Text_IO is modified to avoid undesirable
10781 look-ahead as follows:
10783 An input file that is not a regular file is considered to have no page
10784 marks. Any @code{Ascii.FF} characters (the character normally used for a
10785 page mark) appearing in the file are considered to be data
10786 characters. In particular:
10790 @code{Get_Line} and @code{Skip_Line} do not test for a page mark
10791 following a line mark. If a page mark appears, it will be treated as a
10795 This avoids the need to wait for an extra character to be typed or
10796 entered from the pipe to complete one of these operations.
10799 @code{End_Of_Page} always returns @code{False}
10802 @code{End_Of_File} will return @code{False} if there is a page mark at
10803 the end of the file.
10807 Output to non-regular files is the same as for regular files. Page marks
10808 may be written to non-regular files using @code{New_Page}, but as noted
10809 above they will not be treated as page marks on input if the output is
10810 piped to another Ada program.
10812 Another important discrepancy when reading non-regular files is that the end
10813 of file indication is not ``sticky''. If an end of file is entered, e.g.@: by
10814 pressing the @key{EOT} key,
10816 is signaled once (i.e.@: the test @code{End_Of_File}
10817 will yield @code{True}, or a read will
10818 raise @code{End_Error}), but then reading can resume
10819 to read data past that end of
10820 file indication, until another end of file indication is entered.
10822 @node Get_Immediate
10823 @subsection Get_Immediate
10824 @cindex Get_Immediate
10827 Get_Immediate returns the next character (including control characters)
10828 from the input file. In particular, Get_Immediate will return LF or FF
10829 characters used as line marks or page marks. Such operations leave the
10830 file positioned past the control character, and it is thus not treated
10831 as having its normal function. This means that page, line and column
10832 counts after this kind of Get_Immediate call are set as though the mark
10833 did not occur. In the case where a Get_Immediate leaves the file
10834 positioned between the line mark and page mark (which is not normally
10835 possible), it is undefined whether the FF character will be treated as a
10838 @node Treating Text_IO Files as Streams
10839 @subsection Treating Text_IO Files as Streams
10840 @cindex Stream files
10843 The package @code{Text_IO.Streams} allows a Text_IO file to be treated
10844 as a stream. Data written to a Text_IO file in this stream mode is
10845 binary data. If this binary data contains bytes 16#0A# (@code{LF}) or
10846 16#0C# (@code{FF}), the resulting file may have non-standard
10847 format. Similarly if read operations are used to read from a Text_IO
10848 file treated as a stream, then @code{LF} and @code{FF} characters may be
10849 skipped and the effect is similar to that described above for
10850 @code{Get_Immediate}.
10852 @node Text_IO Extensions
10853 @subsection Text_IO Extensions
10854 @cindex Text_IO extensions
10857 A package GNAT.IO_Aux in the GNAT library provides some useful extensions
10858 to the standard @code{Text_IO} package:
10861 @item function File_Exists (Name : String) return Boolean;
10862 Determines if a file of the given name exists.
10864 @item function Get_Line return String;
10865 Reads a string from the standard input file. The value returned is exactly
10866 the length of the line that was read.
10868 @item function Get_Line (File : Ada.Text_IO.File_Type) return String;
10869 Similar, except that the parameter File specifies the file from which
10870 the string is to be read.
10874 @node Text_IO Facilities for Unbounded Strings
10875 @subsection Text_IO Facilities for Unbounded Strings
10876 @cindex Text_IO for unbounded strings
10877 @cindex Unbounded_String, Text_IO operations
10880 The package @code{Ada.Strings.Unbounded.Text_IO}
10881 in library files @code{a-suteio.ads/adb} contains some GNAT-specific
10882 subprograms useful for Text_IO operations on unbounded strings:
10886 @item function Get_Line (File : File_Type) return Unbounded_String;
10887 Reads a line from the specified file
10888 and returns the result as an unbounded string.
10890 @item procedure Put (File : File_Type; U : Unbounded_String);
10891 Writes the value of the given unbounded string to the specified file
10892 Similar to the effect of
10893 @code{Put (To_String (U))} except that an extra copy is avoided.
10895 @item procedure Put_Line (File : File_Type; U : Unbounded_String);
10896 Writes the value of the given unbounded string to the specified file,
10897 followed by a @code{New_Line}.
10898 Similar to the effect of @code{Put_Line (To_String (U))} except
10899 that an extra copy is avoided.
10903 In the above procedures, @code{File} is of type @code{Ada.Text_IO.File_Type}
10904 and is optional. If the parameter is omitted, then the standard input or
10905 output file is referenced as appropriate.
10907 The package @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} in library
10908 files @file{a-swuwti.ads} and @file{a-swuwti.adb} provides similar extended
10909 @code{Wide_Text_IO} functionality for unbounded wide strings.
10912 @section Wide_Text_IO
10915 @code{Wide_Text_IO} is similar in most respects to Text_IO, except that
10916 both input and output files may contain special sequences that represent
10917 wide character values. The encoding scheme for a given file may be
10918 specified using a FORM parameter:
10925 as part of the FORM string (WCEM = wide character encoding method),
10926 where @var{x} is one of the following characters
10932 Upper half encoding
10944 The encoding methods match those that
10945 can be used in a source
10946 program, but there is no requirement that the encoding method used for
10947 the source program be the same as the encoding method used for files,
10948 and different files may use different encoding methods.
10950 The default encoding method for the standard files, and for opened files
10951 for which no WCEM parameter is given in the FORM string matches the
10952 wide character encoding specified for the main program (the default
10953 being brackets encoding if no coding method was specified with -gnatW).
10957 In this encoding, a wide character is represented by a five character
10965 where @var{a}, @var{b}, @var{c}, @var{d} are the four hexadecimal
10966 characters (using upper case letters) of the wide character code. For
10967 example, ESC A345 is used to represent the wide character with code
10968 16#A345#. This scheme is compatible with use of the full
10969 @code{Wide_Character} set.
10971 @item Upper Half Coding
10972 The wide character with encoding 16#abcd#, where the upper bit is on
10973 (i.e.@: a is in the range 8-F) is represented as two bytes 16#ab# and
10974 16#cd#. The second byte may never be a format control character, but is
10975 not required to be in the upper half. This method can be also used for
10976 shift-JIS or EUC where the internal coding matches the external coding.
10978 @item Shift JIS Coding
10979 A wide character is represented by a two character sequence 16#ab# and
10980 16#cd#, with the restrictions described for upper half encoding as
10981 described above. The internal character code is the corresponding JIS
10982 character according to the standard algorithm for Shift-JIS
10983 conversion. Only characters defined in the JIS code set table can be
10984 used with this encoding method.
10987 A wide character is represented by a two character sequence 16#ab# and
10988 16#cd#, with both characters being in the upper half. The internal
10989 character code is the corresponding JIS character according to the EUC
10990 encoding algorithm. Only characters defined in the JIS code set table
10991 can be used with this encoding method.
10994 A wide character is represented using
10995 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
10996 10646-1/Am.2. Depending on the character value, the representation
10997 is a one, two, or three byte sequence:
11000 16#0000#-16#007f#: 2#0xxxxxxx#
11001 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
11002 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
11006 where the xxx bits correspond to the left-padded bits of the
11007 16-bit character value. Note that all lower half ASCII characters
11008 are represented as ASCII bytes and all upper half characters and
11009 other wide characters are represented as sequences of upper-half
11010 (The full UTF-8 scheme allows for encoding 31-bit characters as
11011 6-byte sequences, but in this implementation, all UTF-8 sequences
11012 of four or more bytes length will raise a Constraint_Error, as
11013 will all invalid UTF-8 sequences.)
11015 @item Brackets Coding
11016 In this encoding, a wide character is represented by the following eight
11017 character sequence:
11024 where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
11025 characters (using uppercase letters) of the wide character code. For
11026 example, @code{["A345"]} is used to represent the wide character with code
11028 This scheme is compatible with use of the full Wide_Character set.
11029 On input, brackets coding can also be used for upper half characters,
11030 e.g.@: @code{["C1"]} for lower case a. However, on output, brackets notation
11031 is only used for wide characters with a code greater than @code{16#FF#}.
11036 For the coding schemes other than Hex and Brackets encoding,
11037 not all wide character
11038 values can be represented. An attempt to output a character that cannot
11039 be represented using the encoding scheme for the file causes
11040 Constraint_Error to be raised. An invalid wide character sequence on
11041 input also causes Constraint_Error to be raised.
11044 * Wide_Text_IO Stream Pointer Positioning::
11045 * Wide_Text_IO Reading and Writing Non-Regular Files::
11048 @node Wide_Text_IO Stream Pointer Positioning
11049 @subsection Stream Pointer Positioning
11052 @code{Ada.Wide_Text_IO} is similar to @code{Ada.Text_IO} in its handling
11053 of stream pointer positioning (@pxref{Text_IO}). There is one additional
11056 If @code{Ada.Wide_Text_IO.Look_Ahead} reads a character outside the
11057 normal lower ASCII set (i.e.@: a character in the range:
11059 @smallexample @c ada
11060 Wide_Character'Val (16#0080#) .. Wide_Character'Val (16#FFFF#)
11064 then although the logical position of the file pointer is unchanged by
11065 the @code{Look_Ahead} call, the stream is physically positioned past the
11066 wide character sequence. Again this is to avoid the need for buffering
11067 or backup, and all @code{Wide_Text_IO} routines check the internal
11068 indication that this situation has occurred so that this is not visible
11069 to a normal program using @code{Wide_Text_IO}. However, this discrepancy
11070 can be observed if the wide text file shares a stream with another file.
11072 @node Wide_Text_IO Reading and Writing Non-Regular Files
11073 @subsection Reading and Writing Non-Regular Files
11076 As in the case of Text_IO, when a non-regular file is read, it is
11077 assumed that the file contains no page marks (any form characters are
11078 treated as data characters), and @code{End_Of_Page} always returns
11079 @code{False}. Similarly, the end of file indication is not sticky, so
11080 it is possible to read beyond an end of file.
11086 A stream file is a sequence of bytes, where individual elements are
11087 written to the file as described in the Ada 95 reference manual. The type
11088 @code{Stream_Element} is simply a byte. There are two ways to read or
11089 write a stream file.
11093 The operations @code{Read} and @code{Write} directly read or write a
11094 sequence of stream elements with no control information.
11097 The stream attributes applied to a stream file transfer data in the
11098 manner described for stream attributes.
11102 @section Shared Files
11105 Section A.14 of the Ada 95 Reference Manual allows implementations to
11106 provide a wide variety of behavior if an attempt is made to access the
11107 same external file with two or more internal files.
11109 To provide a full range of functionality, while at the same time
11110 minimizing the problems of portability caused by this implementation
11111 dependence, GNAT handles file sharing as follows:
11115 In the absence of a @samp{shared=@var{xxx}} form parameter, an attempt
11116 to open two or more files with the same full name is considered an error
11117 and is not supported. The exception @code{Use_Error} will be
11118 raised. Note that a file that is not explicitly closed by the program
11119 remains open until the program terminates.
11122 If the form parameter @samp{shared=no} appears in the form string, the
11123 file can be opened or created with its own separate stream identifier,
11124 regardless of whether other files sharing the same external file are
11125 opened. The exact effect depends on how the C stream routines handle
11126 multiple accesses to the same external files using separate streams.
11129 If the form parameter @samp{shared=yes} appears in the form string for
11130 each of two or more files opened using the same full name, the same
11131 stream is shared between these files, and the semantics are as described
11132 in Ada 95 Reference Manual, Section A.14.
11136 When a program that opens multiple files with the same name is ported
11137 from another Ada compiler to GNAT, the effect will be that
11138 @code{Use_Error} is raised.
11140 The documentation of the original compiler and the documentation of the
11141 program should then be examined to determine if file sharing was
11142 expected, and @samp{shared=@var{xxx}} parameters added to @code{Open}
11143 and @code{Create} calls as required.
11145 When a program is ported from GNAT to some other Ada compiler, no
11146 special attention is required unless the @samp{shared=@var{xxx}} form
11147 parameter is used in the program. In this case, you must examine the
11148 documentation of the new compiler to see if it supports the required
11149 file sharing semantics, and form strings modified appropriately. Of
11150 course it may be the case that the program cannot be ported if the
11151 target compiler does not support the required functionality. The best
11152 approach in writing portable code is to avoid file sharing (and hence
11153 the use of the @samp{shared=@var{xxx}} parameter in the form string)
11156 One common use of file sharing in Ada 83 is the use of instantiations of
11157 Sequential_IO on the same file with different types, to achieve
11158 heterogeneous input-output. Although this approach will work in GNAT if
11159 @samp{shared=yes} is specified, it is preferable in Ada 95 to use Stream_IO
11160 for this purpose (using the stream attributes)
11163 @section Open Modes
11166 @code{Open} and @code{Create} calls result in a call to @code{fopen}
11167 using the mode shown in the following table:
11170 @center @code{Open} and @code{Create} Call Modes
11172 @b{OPEN } @b{CREATE}
11173 Append_File "r+" "w+"
11175 Out_File (Direct_IO) "r+" "w"
11176 Out_File (all other cases) "w" "w"
11177 Inout_File "r+" "w+"
11181 If text file translation is required, then either @samp{b} or @samp{t}
11182 is added to the mode, depending on the setting of Text. Text file
11183 translation refers to the mapping of CR/LF sequences in an external file
11184 to LF characters internally. This mapping only occurs in DOS and
11185 DOS-like systems, and is not relevant to other systems.
11187 A special case occurs with Stream_IO@. As shown in the above table, the
11188 file is initially opened in @samp{r} or @samp{w} mode for the
11189 @code{In_File} and @code{Out_File} cases. If a @code{Set_Mode} operation
11190 subsequently requires switching from reading to writing or vice-versa,
11191 then the file is reopened in @samp{r+} mode to permit the required operation.
11193 @node Operations on C Streams
11194 @section Operations on C Streams
11195 The package @code{Interfaces.C_Streams} provides an Ada program with direct
11196 access to the C library functions for operations on C streams:
11198 @smallexample @c adanocomment
11199 package Interfaces.C_Streams is
11200 -- Note: the reason we do not use the types that are in
11201 -- Interfaces.C is that we want to avoid dragging in the
11202 -- code in this unit if possible.
11203 subtype chars is System.Address;
11204 -- Pointer to null-terminated array of characters
11205 subtype FILEs is System.Address;
11206 -- Corresponds to the C type FILE*
11207 subtype voids is System.Address;
11208 -- Corresponds to the C type void*
11209 subtype int is Integer;
11210 subtype long is Long_Integer;
11211 -- Note: the above types are subtypes deliberately, and it
11212 -- is part of this spec that the above correspondences are
11213 -- guaranteed. This means that it is legitimate to, for
11214 -- example, use Integer instead of int. We provide these
11215 -- synonyms for clarity, but in some cases it may be
11216 -- convenient to use the underlying types (for example to
11217 -- avoid an unnecessary dependency of a spec on the spec
11219 type size_t is mod 2 ** Standard'Address_Size;
11220 NULL_Stream : constant FILEs;
11221 -- Value returned (NULL in C) to indicate an
11222 -- fdopen/fopen/tmpfile error
11223 ----------------------------------
11224 -- Constants Defined in stdio.h --
11225 ----------------------------------
11226 EOF : constant int;
11227 -- Used by a number of routines to indicate error or
11229 IOFBF : constant int;
11230 IOLBF : constant int;
11231 IONBF : constant int;
11232 -- Used to indicate buffering mode for setvbuf call
11233 SEEK_CUR : constant int;
11234 SEEK_END : constant int;
11235 SEEK_SET : constant int;
11236 -- Used to indicate origin for fseek call
11237 function stdin return FILEs;
11238 function stdout return FILEs;
11239 function stderr return FILEs;
11240 -- Streams associated with standard files
11241 --------------------------
11242 -- Standard C functions --
11243 --------------------------
11244 -- The functions selected below are ones that are
11245 -- available in DOS, OS/2, UNIX and Xenix (but not
11246 -- necessarily in ANSI C). These are very thin interfaces
11247 -- which copy exactly the C headers. For more
11248 -- documentation on these functions, see the Microsoft C
11249 -- "Run-Time Library Reference" (Microsoft Press, 1990,
11250 -- ISBN 1-55615-225-6), which includes useful information
11251 -- on system compatibility.
11252 procedure clearerr (stream : FILEs);
11253 function fclose (stream : FILEs) return int;
11254 function fdopen (handle : int; mode : chars) return FILEs;
11255 function feof (stream : FILEs) return int;
11256 function ferror (stream : FILEs) return int;
11257 function fflush (stream : FILEs) return int;
11258 function fgetc (stream : FILEs) return int;
11259 function fgets (strng : chars; n : int; stream : FILEs)
11261 function fileno (stream : FILEs) return int;
11262 function fopen (filename : chars; Mode : chars)
11264 -- Note: to maintain target independence, use
11265 -- text_translation_required, a boolean variable defined in
11266 -- a-sysdep.c to deal with the target dependent text
11267 -- translation requirement. If this variable is set,
11268 -- then b/t should be appended to the standard mode
11269 -- argument to set the text translation mode off or on
11271 function fputc (C : int; stream : FILEs) return int;
11272 function fputs (Strng : chars; Stream : FILEs) return int;
11289 function ftell (stream : FILEs) return long;
11296 function isatty (handle : int) return int;
11297 procedure mktemp (template : chars);
11298 -- The return value (which is just a pointer to template)
11300 procedure rewind (stream : FILEs);
11301 function rmtmp return int;
11309 function tmpfile return FILEs;
11310 function ungetc (c : int; stream : FILEs) return int;
11311 function unlink (filename : chars) return int;
11312 ---------------------
11313 -- Extra functions --
11314 ---------------------
11315 -- These functions supply slightly thicker bindings than
11316 -- those above. They are derived from functions in the
11317 -- C Run-Time Library, but may do a bit more work than
11318 -- just directly calling one of the Library functions.
11319 function is_regular_file (handle : int) return int;
11320 -- Tests if given handle is for a regular file (result 1)
11321 -- or for a non-regular file (pipe or device, result 0).
11322 ---------------------------------
11323 -- Control of Text/Binary Mode --
11324 ---------------------------------
11325 -- If text_translation_required is true, then the following
11326 -- functions may be used to dynamically switch a file from
11327 -- binary to text mode or vice versa. These functions have
11328 -- no effect if text_translation_required is false (i.e. in
11329 -- normal UNIX mode). Use fileno to get a stream handle.
11330 procedure set_binary_mode (handle : int);
11331 procedure set_text_mode (handle : int);
11332 ----------------------------
11333 -- Full Path Name support --
11334 ----------------------------
11335 procedure full_name (nam : chars; buffer : chars);
11336 -- Given a NUL terminated string representing a file
11337 -- name, returns in buffer a NUL terminated string
11338 -- representing the full path name for the file name.
11339 -- On systems where it is relevant the drive is also
11340 -- part of the full path name. It is the responsibility
11341 -- of the caller to pass an actual parameter for buffer
11342 -- that is big enough for any full path name. Use
11343 -- max_path_len given below as the size of buffer.
11344 max_path_len : integer;
11345 -- Maximum length of an allowable full path name on the
11346 -- system, including a terminating NUL character.
11347 end Interfaces.C_Streams;
11350 @node Interfacing to C Streams
11351 @section Interfacing to C Streams
11354 The packages in this section permit interfacing Ada files to C Stream
11357 @smallexample @c ada
11358 with Interfaces.C_Streams;
11359 package Ada.Sequential_IO.C_Streams is
11360 function C_Stream (F : File_Type)
11361 return Interfaces.C_Streams.FILEs;
11363 (File : in out File_Type;
11364 Mode : in File_Mode;
11365 C_Stream : in Interfaces.C_Streams.FILEs;
11366 Form : in String := "");
11367 end Ada.Sequential_IO.C_Streams;
11369 with Interfaces.C_Streams;
11370 package Ada.Direct_IO.C_Streams is
11371 function C_Stream (F : File_Type)
11372 return Interfaces.C_Streams.FILEs;
11374 (File : in out File_Type;
11375 Mode : in File_Mode;
11376 C_Stream : in Interfaces.C_Streams.FILEs;
11377 Form : in String := "");
11378 end Ada.Direct_IO.C_Streams;
11380 with Interfaces.C_Streams;
11381 package Ada.Text_IO.C_Streams is
11382 function C_Stream (F : File_Type)
11383 return Interfaces.C_Streams.FILEs;
11385 (File : in out File_Type;
11386 Mode : in File_Mode;
11387 C_Stream : in Interfaces.C_Streams.FILEs;
11388 Form : in String := "");
11389 end Ada.Text_IO.C_Streams;
11391 with Interfaces.C_Streams;
11392 package Ada.Wide_Text_IO.C_Streams is
11393 function C_Stream (F : File_Type)
11394 return Interfaces.C_Streams.FILEs;
11396 (File : in out File_Type;
11397 Mode : in File_Mode;
11398 C_Stream : in Interfaces.C_Streams.FILEs;
11399 Form : in String := "");
11400 end Ada.Wide_Text_IO.C_Streams;
11402 with Interfaces.C_Streams;
11403 package Ada.Stream_IO.C_Streams is
11404 function C_Stream (F : File_Type)
11405 return Interfaces.C_Streams.FILEs;
11407 (File : in out File_Type;
11408 Mode : in File_Mode;
11409 C_Stream : in Interfaces.C_Streams.FILEs;
11410 Form : in String := "");
11411 end Ada.Stream_IO.C_Streams;
11415 In each of these five packages, the @code{C_Stream} function obtains the
11416 @code{FILE} pointer from a currently opened Ada file. It is then
11417 possible to use the @code{Interfaces.C_Streams} package to operate on
11418 this stream, or the stream can be passed to a C program which can
11419 operate on it directly. Of course the program is responsible for
11420 ensuring that only appropriate sequences of operations are executed.
11422 One particular use of relevance to an Ada program is that the
11423 @code{setvbuf} function can be used to control the buffering of the
11424 stream used by an Ada file. In the absence of such a call the standard
11425 default buffering is used.
11427 The @code{Open} procedures in these packages open a file giving an
11428 existing C Stream instead of a file name. Typically this stream is
11429 imported from a C program, allowing an Ada file to operate on an
11432 @node The GNAT Library
11433 @chapter The GNAT Library
11436 The GNAT library contains a number of general and special purpose packages.
11437 It represents functionality that the GNAT developers have found useful, and
11438 which is made available to GNAT users. The packages described here are fully
11439 supported, and upwards compatibility will be maintained in future releases,
11440 so you can use these facilities with the confidence that the same functionality
11441 will be available in future releases.
11443 The chapter here simply gives a brief summary of the facilities available.
11444 The full documentation is found in the spec file for the package. The full
11445 sources of these library packages, including both spec and body, are provided
11446 with all GNAT releases. For example, to find out the full specifications of
11447 the SPITBOL pattern matching capability, including a full tutorial and
11448 extensive examples, look in the @file{g-spipat.ads} file in the library.
11450 For each entry here, the package name (as it would appear in a @code{with}
11451 clause) is given, followed by the name of the corresponding spec file in
11452 parentheses. The packages are children in four hierarchies, @code{Ada},
11453 @code{Interfaces}, @code{System}, and @code{GNAT}, the latter being a
11454 GNAT-specific hierarchy.
11456 Note that an application program should only use packages in one of these
11457 four hierarchies if the package is defined in the Ada Reference Manual,
11458 or is listed in this section of the GNAT Programmers Reference Manual.
11459 All other units should be considered internal implementation units and
11460 should not be directly @code{with}'ed by application code. The use of
11461 a @code{with} statement that references one of these internal implementation
11462 units makes an application potentially dependent on changes in versions
11463 of GNAT, and will generate a warning message.
11466 * Ada.Characters.Latin_9 (a-chlat9.ads)::
11467 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
11468 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
11469 * Ada.Command_Line.Remove (a-colire.ads)::
11470 * Ada.Command_Line.Environment (a-colien.ads)::
11471 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
11472 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
11473 * Ada.Exceptions.Traceback (a-exctra.ads)::
11474 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
11475 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
11476 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
11477 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
11478 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
11479 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
11480 * GNAT.Array_Split (g-arrspl.ads)::
11481 * GNAT.AWK (g-awk.ads)::
11482 * GNAT.Bounded_Buffers (g-boubuf.ads)::
11483 * GNAT.Bounded_Mailboxes (g-boumai.ads)::
11484 * GNAT.Bubble_Sort (g-bubsor.ads)::
11485 * GNAT.Bubble_Sort_A (g-busora.ads)::
11486 * GNAT.Bubble_Sort_G (g-busorg.ads)::
11487 * GNAT.Calendar (g-calend.ads)::
11488 * GNAT.Calendar.Time_IO (g-catiio.ads)::
11489 * GNAT.CRC32 (g-crc32.ads)::
11490 * GNAT.Case_Util (g-casuti.ads)::
11491 * GNAT.CGI (g-cgi.ads)::
11492 * GNAT.CGI.Cookie (g-cgicoo.ads)::
11493 * GNAT.CGI.Debug (g-cgideb.ads)::
11494 * GNAT.Command_Line (g-comlin.ads)::
11495 * GNAT.Compiler_Version (g-comver.ads)::
11496 * GNAT.Ctrl_C (g-ctrl_c.ads)::
11497 * GNAT.Current_Exception (g-curexc.ads)::
11498 * GNAT.Debug_Pools (g-debpoo.ads)::
11499 * GNAT.Debug_Utilities (g-debuti.ads)::
11500 * GNAT.Directory_Operations (g-dirope.ads)::
11501 * GNAT.Dynamic_HTables (g-dynhta.ads)::
11502 * GNAT.Dynamic_Tables (g-dyntab.ads)::
11503 * GNAT.Exception_Actions (g-excact.ads)::
11504 * GNAT.Exception_Traces (g-exctra.ads)::
11505 * GNAT.Exceptions (g-except.ads)::
11506 * GNAT.Expect (g-expect.ads)::
11507 * GNAT.Float_Control (g-flocon.ads)::
11508 * GNAT.Heap_Sort (g-heasor.ads)::
11509 * GNAT.Heap_Sort_A (g-hesora.ads)::
11510 * GNAT.Heap_Sort_G (g-hesorg.ads)::
11511 * GNAT.HTable (g-htable.ads)::
11512 * GNAT.IO (g-io.ads)::
11513 * GNAT.IO_Aux (g-io_aux.ads)::
11514 * GNAT.Lock_Files (g-locfil.ads)::
11515 * GNAT.MD5 (g-md5.ads)::
11516 * GNAT.Memory_Dump (g-memdum.ads)::
11517 * GNAT.Most_Recent_Exception (g-moreex.ads)::
11518 * GNAT.OS_Lib (g-os_lib.ads)::
11519 * GNAT.Perfect_Hash_Generators (g-pehage.ads)::
11520 * GNAT.Regexp (g-regexp.ads)::
11521 * GNAT.Registry (g-regist.ads)::
11522 * GNAT.Regpat (g-regpat.ads)::
11523 * GNAT.Secondary_Stack_Info (g-sestin.ads)::
11524 * GNAT.Semaphores (g-semaph.ads)::
11525 * GNAT.Signals (g-signal.ads)::
11526 * GNAT.Sockets (g-socket.ads)::
11527 * GNAT.Source_Info (g-souinf.ads)::
11528 * GNAT.Spell_Checker (g-speche.ads)::
11529 * GNAT.Spitbol.Patterns (g-spipat.ads)::
11530 * GNAT.Spitbol (g-spitbo.ads)::
11531 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
11532 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
11533 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
11534 * GNAT.Strings (g-string.ads)::
11535 * GNAT.String_Split (g-strspl.ads)::
11536 * GNAT.Table (g-table.ads)::
11537 * GNAT.Task_Lock (g-tasloc.ads)::
11538 * GNAT.Threads (g-thread.ads)::
11539 * GNAT.Traceback (g-traceb.ads)::
11540 * GNAT.Traceback.Symbolic (g-trasym.ads)::
11541 * GNAT.Wide_String_Split (g-wistsp.ads)::
11542 * Interfaces.C.Extensions (i-cexten.ads)::
11543 * Interfaces.C.Streams (i-cstrea.ads)::
11544 * Interfaces.CPP (i-cpp.ads)::
11545 * Interfaces.Os2lib (i-os2lib.ads)::
11546 * Interfaces.Os2lib.Errors (i-os2err.ads)::
11547 * Interfaces.Os2lib.Synchronization (i-os2syn.ads)::
11548 * Interfaces.Os2lib.Threads (i-os2thr.ads)::
11549 * Interfaces.Packed_Decimal (i-pacdec.ads)::
11550 * Interfaces.VxWorks (i-vxwork.ads)::
11551 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
11552 * System.Address_Image (s-addima.ads)::
11553 * System.Assertions (s-assert.ads)::
11554 * System.Memory (s-memory.ads)::
11555 * System.Partition_Interface (s-parint.ads)::
11556 * System.Restrictions (s-restri.ads)::
11557 * System.Rident (s-rident.ads)::
11558 * System.Task_Info (s-tasinf.ads)::
11559 * System.Wch_Cnv (s-wchcnv.ads)::
11560 * System.Wch_Con (s-wchcon.ads)::
11563 @node Ada.Characters.Latin_9 (a-chlat9.ads)
11564 @section @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
11565 @cindex @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
11566 @cindex Latin_9 constants for Character
11569 This child of @code{Ada.Characters}
11570 provides a set of definitions corresponding to those in the
11571 RM-defined package @code{Ada.Characters.Latin_1} but with the
11572 few modifications required for @code{Latin-9}
11573 The provision of such a package
11574 is specifically authorized by the Ada Reference Manual
11577 @node Ada.Characters.Wide_Latin_1 (a-cwila1.ads)
11578 @section @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
11579 @cindex @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
11580 @cindex Latin_1 constants for Wide_Character
11583 This child of @code{Ada.Characters}
11584 provides a set of definitions corresponding to those in the
11585 RM-defined package @code{Ada.Characters.Latin_1} but with the
11586 types of the constants being @code{Wide_Character}
11587 instead of @code{Character}. The provision of such a package
11588 is specifically authorized by the Ada Reference Manual
11591 @node Ada.Characters.Wide_Latin_9 (a-cwila9.ads)
11592 @section @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
11593 @cindex @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
11594 @cindex Latin_9 constants for Wide_Character
11597 This child of @code{Ada.Characters}
11598 provides a set of definitions corresponding to those in the
11599 GNAT defined package @code{Ada.Characters.Latin_9} but with the
11600 types of the constants being @code{Wide_Character}
11601 instead of @code{Character}. The provision of such a package
11602 is specifically authorized by the Ada Reference Manual
11605 @node Ada.Command_Line.Remove (a-colire.ads)
11606 @section @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
11607 @cindex @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
11608 @cindex Removing command line arguments
11609 @cindex Command line, argument removal
11612 This child of @code{Ada.Command_Line}
11613 provides a mechanism for logically removing
11614 arguments from the argument list. Once removed, an argument is not visible
11615 to further calls on the subprograms in @code{Ada.Command_Line} will not
11616 see the removed argument.
11618 @node Ada.Command_Line.Environment (a-colien.ads)
11619 @section @code{Ada.Command_Line.Environment} (@file{a-colien.ads})
11620 @cindex @code{Ada.Command_Line.Environment} (@file{a-colien.ads})
11621 @cindex Environment entries
11624 This child of @code{Ada.Command_Line}
11625 provides a mechanism for obtaining environment values on systems
11626 where this concept makes sense.
11628 @node Ada.Direct_IO.C_Streams (a-diocst.ads)
11629 @section @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
11630 @cindex @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
11631 @cindex C Streams, Interfacing with Direct_IO
11634 This package provides subprograms that allow interfacing between
11635 C streams and @code{Direct_IO}. The stream identifier can be
11636 extracted from a file opened on the Ada side, and an Ada file
11637 can be constructed from a stream opened on the C side.
11639 @node Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)
11640 @section @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
11641 @cindex @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
11642 @cindex Null_Occurrence, testing for
11645 This child subprogram provides a way of testing for the null
11646 exception occurrence (@code{Null_Occurrence}) without raising
11649 @node Ada.Exceptions.Traceback (a-exctra.ads)
11650 @section @code{Ada.Exceptions.Traceback} (@file{a-exctra.ads})
11651 @cindex @code{Ada.Exceptions.Traceback} (@file{a-exctra.ads})
11652 @cindex Traceback for Exception Occurrence
11655 This child package provides the subprogram (@code{Tracebacks}) to
11656 give a traceback array of addresses based on an exception
11659 @node Ada.Sequential_IO.C_Streams (a-siocst.ads)
11660 @section @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
11661 @cindex @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
11662 @cindex C Streams, Interfacing with Sequential_IO
11665 This package provides subprograms that allow interfacing between
11666 C streams and @code{Sequential_IO}. The stream identifier can be
11667 extracted from a file opened on the Ada side, and an Ada file
11668 can be constructed from a stream opened on the C side.
11670 @node Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)
11671 @section @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
11672 @cindex @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
11673 @cindex C Streams, Interfacing with Stream_IO
11676 This package provides subprograms that allow interfacing between
11677 C streams and @code{Stream_IO}. The stream identifier can be
11678 extracted from a file opened on the Ada side, and an Ada file
11679 can be constructed from a stream opened on the C side.
11681 @node Ada.Strings.Unbounded.Text_IO (a-suteio.ads)
11682 @section @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
11683 @cindex @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
11684 @cindex @code{Unbounded_String}, IO support
11685 @cindex @code{Text_IO}, extensions for unbounded strings
11688 This package provides subprograms for Text_IO for unbounded
11689 strings, avoiding the necessity for an intermediate operation
11690 with ordinary strings.
11692 @node Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)
11693 @section @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
11694 @cindex @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
11695 @cindex @code{Unbounded_Wide_String}, IO support
11696 @cindex @code{Text_IO}, extensions for unbounded wide strings
11699 This package provides subprograms for Text_IO for unbounded
11700 wide strings, avoiding the necessity for an intermediate operation
11701 with ordinary wide strings.
11703 @node Ada.Text_IO.C_Streams (a-tiocst.ads)
11704 @section @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
11705 @cindex @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
11706 @cindex C Streams, Interfacing with @code{Text_IO}
11709 This package provides subprograms that allow interfacing between
11710 C streams and @code{Text_IO}. The stream identifier can be
11711 extracted from a file opened on the Ada side, and an Ada file
11712 can be constructed from a stream opened on the C side.
11714 @node Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)
11715 @section @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
11716 @cindex @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
11717 @cindex C Streams, Interfacing with @code{Wide_Text_IO}
11720 This package provides subprograms that allow interfacing between
11721 C streams and @code{Wide_Text_IO}. The stream identifier can be
11722 extracted from a file opened on the Ada side, and an Ada file
11723 can be constructed from a stream opened on the C side.
11725 @node GNAT.Array_Split (g-arrspl.ads)
11726 @section @code{GNAT.Array_Split} (@file{g-arrspl.ads})
11727 @cindex @code{GNAT.Array_Split} (@file{g-arrspl.ads})
11728 @cindex Array splitter
11731 Useful array-manipulation routines: given a set of separators, split
11732 an array wherever the separators appear, and provide direct access
11733 to the resulting slices.
11735 @node GNAT.AWK (g-awk.ads)
11736 @section @code{GNAT.AWK} (@file{g-awk.ads})
11737 @cindex @code{GNAT.AWK} (@file{g-awk.ads})
11742 Provides AWK-like parsing functions, with an easy interface for parsing one
11743 or more files containing formatted data. The file is viewed as a database
11744 where each record is a line and a field is a data element in this line.
11746 @node GNAT.Bounded_Buffers (g-boubuf.ads)
11747 @section @code{GNAT.Bounded_Buffers} (@file{g-boubuf.ads})
11748 @cindex @code{GNAT.Bounded_Buffers} (@file{g-boubuf.ads})
11750 @cindex Bounded Buffers
11753 Provides a concurrent generic bounded buffer abstraction. Instances are
11754 useful directly or as parts of the implementations of other abstractions,
11757 @node GNAT.Bounded_Mailboxes (g-boumai.ads)
11758 @section @code{GNAT.Bounded_Mailboxes} (@file{g-boumai.ads})
11759 @cindex @code{GNAT.Bounded_Mailboxes} (@file{g-boumai.ads})
11764 Provides a thread-safe asynchronous intertask mailbox communication facility.
11766 @node GNAT.Bubble_Sort (g-bubsor.ads)
11767 @section @code{GNAT.Bubble_Sort} (@file{g-bubsor.ads})
11768 @cindex @code{GNAT.Bubble_Sort} (@file{g-bubsor.ads})
11770 @cindex Bubble sort
11773 Provides a general implementation of bubble sort usable for sorting arbitrary
11774 data items. Exchange and comparison procedures are provided by passing
11775 access-to-procedure values.
11777 @node GNAT.Bubble_Sort_A (g-busora.ads)
11778 @section @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
11779 @cindex @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
11781 @cindex Bubble sort
11784 Provides a general implementation of bubble sort usable for sorting arbitrary
11785 data items. Move and comparison procedures are provided by passing
11786 access-to-procedure values. This is an older version, retained for
11787 compatibility. Usually @code{GNAT.Bubble_Sort} will be preferable.
11789 @node GNAT.Bubble_Sort_G (g-busorg.ads)
11790 @section @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
11791 @cindex @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
11793 @cindex Bubble sort
11796 Similar to @code{Bubble_Sort_A} except that the move and sorting procedures
11797 are provided as generic parameters, this improves efficiency, especially
11798 if the procedures can be inlined, at the expense of duplicating code for
11799 multiple instantiations.
11801 @node GNAT.Calendar (g-calend.ads)
11802 @section @code{GNAT.Calendar} (@file{g-calend.ads})
11803 @cindex @code{GNAT.Calendar} (@file{g-calend.ads})
11804 @cindex @code{Calendar}
11807 Extends the facilities provided by @code{Ada.Calendar} to include handling
11808 of days of the week, an extended @code{Split} and @code{Time_Of} capability.
11809 Also provides conversion of @code{Ada.Calendar.Time} values to and from the
11810 C @code{timeval} format.
11812 @node GNAT.Calendar.Time_IO (g-catiio.ads)
11813 @section @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
11814 @cindex @code{Calendar}
11816 @cindex @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
11818 @node GNAT.CRC32 (g-crc32.ads)
11819 @section @code{GNAT.CRC32} (@file{g-crc32.ads})
11820 @cindex @code{GNAT.CRC32} (@file{g-crc32.ads})
11822 @cindex Cyclic Redundancy Check
11825 This package implements the CRC-32 algorithm. For a full description
11826 of this algorithm see
11827 ``Computation of Cyclic Redundancy Checks via Table Look-Up'',
11828 @cite{Communications of the ACM}, Vol.@: 31 No.@: 8, pp.@: 1008-1013,
11829 Aug.@: 1988. Sarwate, D.V@.
11832 Provides an extended capability for formatted output of time values with
11833 full user control over the format. Modeled on the GNU Date specification.
11835 @node GNAT.Case_Util (g-casuti.ads)
11836 @section @code{GNAT.Case_Util} (@file{g-casuti.ads})
11837 @cindex @code{GNAT.Case_Util} (@file{g-casuti.ads})
11838 @cindex Casing utilities
11839 @cindex Character handling (@code{GNAT.Case_Util})
11842 A set of simple routines for handling upper and lower casing of strings
11843 without the overhead of the full casing tables
11844 in @code{Ada.Characters.Handling}.
11846 @node GNAT.CGI (g-cgi.ads)
11847 @section @code{GNAT.CGI} (@file{g-cgi.ads})
11848 @cindex @code{GNAT.CGI} (@file{g-cgi.ads})
11849 @cindex CGI (Common Gateway Interface)
11852 This is a package for interfacing a GNAT program with a Web server via the
11853 Common Gateway Interface (CGI)@. Basically this package parses the CGI
11854 parameters, which are a set of key/value pairs sent by the Web server. It
11855 builds a table whose index is the key and provides some services to deal
11858 @node GNAT.CGI.Cookie (g-cgicoo.ads)
11859 @section @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
11860 @cindex @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
11861 @cindex CGI (Common Gateway Interface) cookie support
11862 @cindex Cookie support in CGI
11865 This is a package to interface a GNAT program with a Web server via the
11866 Common Gateway Interface (CGI). It exports services to deal with Web
11867 cookies (piece of information kept in the Web client software).
11869 @node GNAT.CGI.Debug (g-cgideb.ads)
11870 @section @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
11871 @cindex @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
11872 @cindex CGI (Common Gateway Interface) debugging
11875 This is a package to help debugging CGI (Common Gateway Interface)
11876 programs written in Ada.
11878 @node GNAT.Command_Line (g-comlin.ads)
11879 @section @code{GNAT.Command_Line} (@file{g-comlin.ads})
11880 @cindex @code{GNAT.Command_Line} (@file{g-comlin.ads})
11881 @cindex Command line
11884 Provides a high level interface to @code{Ada.Command_Line} facilities,
11885 including the ability to scan for named switches with optional parameters
11886 and expand file names using wild card notations.
11888 @node GNAT.Compiler_Version (g-comver.ads)
11889 @section @code{GNAT.Compiler_Version} (@file{g-comver.ads})
11890 @cindex @code{GNAT.Compiler_Version} (@file{g-comver.ads})
11891 @cindex Compiler Version
11892 @cindex Version, of compiler
11895 Provides a routine for obtaining the version of the compiler used to
11896 compile the program. More accurately this is the version of the binder
11897 used to bind the program (this will normally be the same as the version
11898 of the compiler if a consistent tool set is used to compile all units
11901 @node GNAT.Ctrl_C (g-ctrl_c.ads)
11902 @section @code{GNAT.Ctrl_C} (@file{g-ctrl_c.ads})
11903 @cindex @code{GNAT.Ctrl_C} (@file{g-ctrl_c.ads})
11907 Provides a simple interface to handle Ctrl-C keyboard events.
11909 @node GNAT.Current_Exception (g-curexc.ads)
11910 @section @code{GNAT.Current_Exception} (@file{g-curexc.ads})
11911 @cindex @code{GNAT.Current_Exception} (@file{g-curexc.ads})
11912 @cindex Current exception
11913 @cindex Exception retrieval
11916 Provides access to information on the current exception that has been raised
11917 without the need for using the Ada-95 exception choice parameter specification
11918 syntax. This is particularly useful in simulating typical facilities for
11919 obtaining information about exceptions provided by Ada 83 compilers.
11921 @node GNAT.Debug_Pools (g-debpoo.ads)
11922 @section @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
11923 @cindex @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
11925 @cindex Debug pools
11926 @cindex Memory corruption debugging
11929 Provide a debugging storage pools that helps tracking memory corruption
11930 problems. See section ``Finding memory problems with GNAT Debug Pool'' in
11931 the @cite{GNAT User's Guide}.
11933 @node GNAT.Debug_Utilities (g-debuti.ads)
11934 @section @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
11935 @cindex @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
11939 Provides a few useful utilities for debugging purposes, including conversion
11940 to and from string images of address values. Supports both C and Ada formats
11941 for hexadecimal literals.
11943 @node GNAT.Directory_Operations (g-dirope.ads)
11944 @section @code{GNAT.Directory_Operations} (g-dirope.ads)
11945 @cindex @code{GNAT.Directory_Operations} (g-dirope.ads)
11946 @cindex Directory operations
11949 Provides a set of routines for manipulating directories, including changing
11950 the current directory, making new directories, and scanning the files in a
11953 @node GNAT.Dynamic_HTables (g-dynhta.ads)
11954 @section @code{GNAT.Dynamic_HTables} (@file{g-dynhta.ads})
11955 @cindex @code{GNAT.Dynamic_HTables} (@file{g-dynhta.ads})
11956 @cindex Hash tables
11959 A generic implementation of hash tables that can be used to hash arbitrary
11960 data. Provided in two forms, a simple form with built in hash functions,
11961 and a more complex form in which the hash function is supplied.
11964 This package provides a facility similar to that of @code{GNAT.HTable},
11965 except that this package declares a type that can be used to define
11966 dynamic instances of the hash table, while an instantiation of
11967 @code{GNAT.HTable} creates a single instance of the hash table.
11969 @node GNAT.Dynamic_Tables (g-dyntab.ads)
11970 @section @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
11971 @cindex @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
11972 @cindex Table implementation
11973 @cindex Arrays, extendable
11976 A generic package providing a single dimension array abstraction where the
11977 length of the array can be dynamically modified.
11980 This package provides a facility similar to that of @code{GNAT.Table},
11981 except that this package declares a type that can be used to define
11982 dynamic instances of the table, while an instantiation of
11983 @code{GNAT.Table} creates a single instance of the table type.
11985 @node GNAT.Exception_Actions (g-excact.ads)
11986 @section @code{GNAT.Exception_Actions} (@file{g-excact.ads})
11987 @cindex @code{GNAT.Exception_Actions} (@file{g-excact.ads})
11988 @cindex Exception actions
11991 Provides callbacks when an exception is raised. Callbacks can be registered
11992 for specific exceptions, or when any exception is raised. This
11993 can be used for instance to force a core dump to ease debugging.
11995 @node GNAT.Exception_Traces (g-exctra.ads)
11996 @section @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
11997 @cindex @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
11998 @cindex Exception traces
12002 Provides an interface allowing to control automatic output upon exception
12005 @node GNAT.Exceptions (g-except.ads)
12006 @section @code{GNAT.Exceptions} (@file{g-expect.ads})
12007 @cindex @code{GNAT.Exceptions} (@file{g-expect.ads})
12008 @cindex Exceptions, Pure
12009 @cindex Pure packages, exceptions
12012 Normally it is not possible to raise an exception with
12013 a message from a subprogram in a pure package, since the
12014 necessary types and subprograms are in @code{Ada.Exceptions}
12015 which is not a pure unit. @code{GNAT.Exceptions} provides a
12016 facility for getting around this limitation for a few
12017 predefined exceptions, and for example allow raising
12018 @code{Constraint_Error} with a message from a pure subprogram.
12020 @node GNAT.Expect (g-expect.ads)
12021 @section @code{GNAT.Expect} (@file{g-expect.ads})
12022 @cindex @code{GNAT.Expect} (@file{g-expect.ads})
12025 Provides a set of subprograms similar to what is available
12026 with the standard Tcl Expect tool.
12027 It allows you to easily spawn and communicate with an external process.
12028 You can send commands or inputs to the process, and compare the output
12029 with some expected regular expression. Currently @code{GNAT.Expect}
12030 is implemented on all native GNAT ports except for OpenVMS@.
12031 It is not implemented for cross ports, and in particular is not
12032 implemented for VxWorks or LynxOS@.
12034 @node GNAT.Float_Control (g-flocon.ads)
12035 @section @code{GNAT.Float_Control} (@file{g-flocon.ads})
12036 @cindex @code{GNAT.Float_Control} (@file{g-flocon.ads})
12037 @cindex Floating-Point Processor
12040 Provides an interface for resetting the floating-point processor into the
12041 mode required for correct semantic operation in Ada. Some third party
12042 library calls may cause this mode to be modified, and the Reset procedure
12043 in this package can be used to reestablish the required mode.
12045 @node GNAT.Heap_Sort (g-heasor.ads)
12046 @section @code{GNAT.Heap_Sort} (@file{g-heasor.ads})
12047 @cindex @code{GNAT.Heap_Sort} (@file{g-heasor.ads})
12051 Provides a general implementation of heap sort usable for sorting arbitrary
12052 data items. Exchange and comparison procedures are provided by passing
12053 access-to-procedure values. The algorithm used is a modified heap sort
12054 that performs approximately N*log(N) comparisons in the worst case.
12056 @node GNAT.Heap_Sort_A (g-hesora.ads)
12057 @section @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
12058 @cindex @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
12062 Provides a general implementation of heap sort usable for sorting arbitrary
12063 data items. Move and comparison procedures are provided by passing
12064 access-to-procedure values. The algorithm used is a modified heap sort
12065 that performs approximately N*log(N) comparisons in the worst case.
12066 This differs from @code{GNAT.Heap_Sort} in having a less convenient
12067 interface, but may be slightly more efficient.
12069 @node GNAT.Heap_Sort_G (g-hesorg.ads)
12070 @section @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
12071 @cindex @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
12075 Similar to @code{Heap_Sort_A} except that the move and sorting procedures
12076 are provided as generic parameters, this improves efficiency, especially
12077 if the procedures can be inlined, at the expense of duplicating code for
12078 multiple instantiations.
12080 @node GNAT.HTable (g-htable.ads)
12081 @section @code{GNAT.HTable} (@file{g-htable.ads})
12082 @cindex @code{GNAT.HTable} (@file{g-htable.ads})
12083 @cindex Hash tables
12086 A generic implementation of hash tables that can be used to hash arbitrary
12087 data. Provides two approaches, one a simple static approach, and the other
12088 allowing arbitrary dynamic hash tables.
12090 @node GNAT.IO (g-io.ads)
12091 @section @code{GNAT.IO} (@file{g-io.ads})
12092 @cindex @code{GNAT.IO} (@file{g-io.ads})
12094 @cindex Input/Output facilities
12097 A simple preelaborable input-output package that provides a subset of
12098 simple Text_IO functions for reading characters and strings from
12099 Standard_Input, and writing characters, strings and integers to either
12100 Standard_Output or Standard_Error.
12102 @node GNAT.IO_Aux (g-io_aux.ads)
12103 @section @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
12104 @cindex @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
12106 @cindex Input/Output facilities
12108 Provides some auxiliary functions for use with Text_IO, including a test
12109 for whether a file exists, and functions for reading a line of text.
12111 @node GNAT.Lock_Files (g-locfil.ads)
12112 @section @code{GNAT.Lock_Files} (@file{g-locfil.ads})
12113 @cindex @code{GNAT.Lock_Files} (@file{g-locfil.ads})
12114 @cindex File locking
12115 @cindex Locking using files
12118 Provides a general interface for using files as locks. Can be used for
12119 providing program level synchronization.
12121 @node GNAT.MD5 (g-md5.ads)
12122 @section @code{GNAT.MD5} (@file{g-md5.ads})
12123 @cindex @code{GNAT.MD5} (@file{g-md5.ads})
12124 @cindex Message Digest MD5
12127 Implements the MD5 Message-Digest Algorithm as described in RFC 1321.
12129 @node GNAT.Memory_Dump (g-memdum.ads)
12130 @section @code{GNAT.Memory_Dump} (@file{g-memdum.ads})
12131 @cindex @code{GNAT.Memory_Dump} (@file{g-memdum.ads})
12132 @cindex Dump Memory
12135 Provides a convenient routine for dumping raw memory to either the
12136 standard output or standard error files. Uses GNAT.IO for actual
12139 @node GNAT.Most_Recent_Exception (g-moreex.ads)
12140 @section @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
12141 @cindex @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
12142 @cindex Exception, obtaining most recent
12145 Provides access to the most recently raised exception. Can be used for
12146 various logging purposes, including duplicating functionality of some
12147 Ada 83 implementation dependent extensions.
12149 @node GNAT.OS_Lib (g-os_lib.ads)
12150 @section @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
12151 @cindex @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
12152 @cindex Operating System interface
12153 @cindex Spawn capability
12156 Provides a range of target independent operating system interface functions,
12157 including time/date management, file operations, subprocess management,
12158 including a portable spawn procedure, and access to environment variables
12159 and error return codes.
12161 @node GNAT.Perfect_Hash_Generators (g-pehage.ads)
12162 @section @code{GNAT.Perfect_Hash_Generators} (@file{g-pehage.ads})
12163 @cindex @code{GNAT.Perfect_Hash_Generators} (@file{g-pehage.ads})
12164 @cindex Hash functions
12167 Provides a generator of static minimal perfect hash functions. No
12168 collisions occur and each item can be retrieved from the table in one
12169 probe (perfect property). The hash table size corresponds to the exact
12170 size of the key set and no larger (minimal property). The key set has to
12171 be know in advance (static property). The hash functions are also order
12172 preservering. If w2 is inserted after w1 in the generator, their
12173 hashcode are in the same order. These hashing functions are very
12174 convenient for use with realtime applications.
12176 @node GNAT.Regexp (g-regexp.ads)
12177 @section @code{GNAT.Regexp} (@file{g-regexp.ads})
12178 @cindex @code{GNAT.Regexp} (@file{g-regexp.ads})
12179 @cindex Regular expressions
12180 @cindex Pattern matching
12183 A simple implementation of regular expressions, using a subset of regular
12184 expression syntax copied from familiar Unix style utilities. This is the
12185 simples of the three pattern matching packages provided, and is particularly
12186 suitable for ``file globbing'' applications.
12188 @node GNAT.Registry (g-regist.ads)
12189 @section @code{GNAT.Registry} (@file{g-regist.ads})
12190 @cindex @code{GNAT.Registry} (@file{g-regist.ads})
12191 @cindex Windows Registry
12194 This is a high level binding to the Windows registry. It is possible to
12195 do simple things like reading a key value, creating a new key. For full
12196 registry API, but at a lower level of abstraction, refer to the Win32.Winreg
12197 package provided with the Win32Ada binding
12199 @node GNAT.Regpat (g-regpat.ads)
12200 @section @code{GNAT.Regpat} (@file{g-regpat.ads})
12201 @cindex @code{GNAT.Regpat} (@file{g-regpat.ads})
12202 @cindex Regular expressions
12203 @cindex Pattern matching
12206 A complete implementation of Unix-style regular expression matching, copied
12207 from the original V7 style regular expression library written in C by
12208 Henry Spencer (and binary compatible with this C library).
12210 @node GNAT.Secondary_Stack_Info (g-sestin.ads)
12211 @section @code{GNAT.Secondary_Stack_Info} (@file{g-sestin.ads})
12212 @cindex @code{GNAT.Secondary_Stack_Info} (@file{g-sestin.ads})
12213 @cindex Secondary Stack Info
12216 Provide the capability to query the high water mark of the current task's
12219 @node GNAT.Semaphores (g-semaph.ads)
12220 @section @code{GNAT.Semaphores} (@file{g-semaph.ads})
12221 @cindex @code{GNAT.Semaphores} (@file{g-semaph.ads})
12225 Provides classic counting and binary semaphores using protected types.
12227 @node GNAT.Signals (g-signal.ads)
12228 @section @code{GNAT.Signals} (@file{g-signal.ads})
12229 @cindex @code{GNAT.Signals} (@file{g-signal.ads})
12233 Provides the ability to manipulate the blocked status of signals on supported
12236 @node GNAT.Sockets (g-socket.ads)
12237 @section @code{GNAT.Sockets} (@file{g-socket.ads})
12238 @cindex @code{GNAT.Sockets} (@file{g-socket.ads})
12242 A high level and portable interface to develop sockets based applications.
12243 This package is based on the sockets thin binding found in
12244 @code{GNAT.Sockets.Thin}. Currently @code{GNAT.Sockets} is implemented
12245 on all native GNAT ports except for OpenVMS@. It is not implemented
12246 for the LynxOS@ cross port.
12248 @node GNAT.Source_Info (g-souinf.ads)
12249 @section @code{GNAT.Source_Info} (@file{g-souinf.ads})
12250 @cindex @code{GNAT.Source_Info} (@file{g-souinf.ads})
12251 @cindex Source Information
12254 Provides subprograms that give access to source code information known at
12255 compile time, such as the current file name and line number.
12257 @node GNAT.Spell_Checker (g-speche.ads)
12258 @section @code{GNAT.Spell_Checker} (@file{g-speche.ads})
12259 @cindex @code{GNAT.Spell_Checker} (@file{g-speche.ads})
12260 @cindex Spell checking
12263 Provides a function for determining whether one string is a plausible
12264 near misspelling of another string.
12266 @node GNAT.Spitbol.Patterns (g-spipat.ads)
12267 @section @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
12268 @cindex @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
12269 @cindex SPITBOL pattern matching
12270 @cindex Pattern matching
12273 A complete implementation of SNOBOL4 style pattern matching. This is the
12274 most elaborate of the pattern matching packages provided. It fully duplicates
12275 the SNOBOL4 dynamic pattern construction and matching capabilities, using the
12276 efficient algorithm developed by Robert Dewar for the SPITBOL system.
12278 @node GNAT.Spitbol (g-spitbo.ads)
12279 @section @code{GNAT.Spitbol} (@file{g-spitbo.ads})
12280 @cindex @code{GNAT.Spitbol} (@file{g-spitbo.ads})
12281 @cindex SPITBOL interface
12284 The top level package of the collection of SPITBOL-style functionality, this
12285 package provides basic SNOBOL4 string manipulation functions, such as
12286 Pad, Reverse, Trim, Substr capability, as well as a generic table function
12287 useful for constructing arbitrary mappings from strings in the style of
12288 the SNOBOL4 TABLE function.
12290 @node GNAT.Spitbol.Table_Boolean (g-sptabo.ads)
12291 @section @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
12292 @cindex @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
12293 @cindex Sets of strings
12294 @cindex SPITBOL Tables
12297 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
12298 for type @code{Standard.Boolean}, giving an implementation of sets of
12301 @node GNAT.Spitbol.Table_Integer (g-sptain.ads)
12302 @section @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
12303 @cindex @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
12304 @cindex Integer maps
12306 @cindex SPITBOL Tables
12309 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
12310 for type @code{Standard.Integer}, giving an implementation of maps
12311 from string to integer values.
12313 @node GNAT.Spitbol.Table_VString (g-sptavs.ads)
12314 @section @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
12315 @cindex @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
12316 @cindex String maps
12318 @cindex SPITBOL Tables
12321 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table} for
12322 a variable length string type, giving an implementation of general
12323 maps from strings to strings.
12325 @node GNAT.Strings (g-string.ads)
12326 @section @code{GNAT.Strings} (@file{g-string.ads})
12327 @cindex @code{GNAT.Strings} (@file{g-string.ads})
12330 Common String access types and related subprograms. Basically it
12331 defines a string access and an array of string access types.
12333 @node GNAT.String_Split (g-strspl.ads)
12334 @section @code{GNAT.String_Split} (@file{g-strspl.ads})
12335 @cindex @code{GNAT.String_Split} (@file{g-strspl.ads})
12336 @cindex String splitter
12339 Useful string-manipulation routines: given a set of separators, split
12340 a string wherever the separators appear, and provide direct access
12341 to the resulting slices. This package is instantiated from
12342 @code{GNAT.Array_Split}.
12344 @node GNAT.Table (g-table.ads)
12345 @section @code{GNAT.Table} (@file{g-table.ads})
12346 @cindex @code{GNAT.Table} (@file{g-table.ads})
12347 @cindex Table implementation
12348 @cindex Arrays, extendable
12351 A generic package providing a single dimension array abstraction where the
12352 length of the array can be dynamically modified.
12355 This package provides a facility similar to that of @code{GNAT.Dynamic_Tables},
12356 except that this package declares a single instance of the table type,
12357 while an instantiation of @code{GNAT.Dynamic_Tables} creates a type that can be
12358 used to define dynamic instances of the table.
12360 @node GNAT.Task_Lock (g-tasloc.ads)
12361 @section @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
12362 @cindex @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
12363 @cindex Task synchronization
12364 @cindex Task locking
12368 A very simple facility for locking and unlocking sections of code using a
12369 single global task lock. Appropriate for use in situations where contention
12370 between tasks is very rarely expected.
12372 @node GNAT.Threads (g-thread.ads)
12373 @section @code{GNAT.Threads} (@file{g-thread.ads})
12374 @cindex @code{GNAT.Threads} (@file{g-thread.ads})
12375 @cindex Foreign threads
12376 @cindex Threads, foreign
12379 Provides facilities for creating and destroying threads with explicit calls.
12380 These threads are known to the GNAT run-time system. These subprograms are
12381 exported C-convention procedures intended to be called from foreign code.
12382 By using these primitives rather than directly calling operating systems
12383 routines, compatibility with the Ada tasking runt-time is provided.
12385 @node GNAT.Traceback (g-traceb.ads)
12386 @section @code{GNAT.Traceback} (@file{g-traceb.ads})
12387 @cindex @code{GNAT.Traceback} (@file{g-traceb.ads})
12388 @cindex Trace back facilities
12391 Provides a facility for obtaining non-symbolic traceback information, useful
12392 in various debugging situations.
12394 @node GNAT.Traceback.Symbolic (g-trasym.ads)
12395 @section @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
12396 @cindex @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
12397 @cindex Trace back facilities
12400 Provides symbolic traceback information that includes the subprogram
12401 name and line number information.
12403 @node GNAT.Wide_String_Split (g-wistsp.ads)
12404 @section @code{GNAT.Wide_String_Split} (@file{g-wistsp.ads})
12405 @cindex @code{GNAT.Wide_String_Split} (@file{g-wistsp.ads})
12406 @cindex Wide_String splitter
12409 Useful wide_string-manipulation routines: given a set of separators, split
12410 a wide_string wherever the separators appear, and provide direct access
12411 to the resulting slices. This package is instantiated from
12412 @code{GNAT.Array_Split}.
12414 @node Interfaces.C.Extensions (i-cexten.ads)
12415 @section @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
12416 @cindex @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
12419 This package contains additional C-related definitions, intended
12420 for use with either manually or automatically generated bindings
12423 @node Interfaces.C.Streams (i-cstrea.ads)
12424 @section @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
12425 @cindex @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
12426 @cindex C streams, interfacing
12429 This package is a binding for the most commonly used operations
12432 @node Interfaces.CPP (i-cpp.ads)
12433 @section @code{Interfaces.CPP} (@file{i-cpp.ads})
12434 @cindex @code{Interfaces.CPP} (@file{i-cpp.ads})
12435 @cindex C++ interfacing
12436 @cindex Interfacing, to C++
12439 This package provides facilities for use in interfacing to C++. It
12440 is primarily intended to be used in connection with automated tools
12441 for the generation of C++ interfaces.
12443 @node Interfaces.Os2lib (i-os2lib.ads)
12444 @section @code{Interfaces.Os2lib} (@file{i-os2lib.ads})
12445 @cindex @code{Interfaces.Os2lib} (@file{i-os2lib.ads})
12446 @cindex Interfacing, to OS/2
12447 @cindex OS/2 interfacing
12450 This package provides interface definitions to the OS/2 library.
12451 It is a thin binding which is a direct translation of the
12452 various @file{<bse@.h>} files.
12454 @node Interfaces.Os2lib.Errors (i-os2err.ads)
12455 @section @code{Interfaces.Os2lib.Errors} (@file{i-os2err.ads})
12456 @cindex @code{Interfaces.Os2lib.Errors} (@file{i-os2err.ads})
12457 @cindex OS/2 Error codes
12458 @cindex Interfacing, to OS/2
12459 @cindex OS/2 interfacing
12462 This package provides definitions of the OS/2 error codes.
12464 @node Interfaces.Os2lib.Synchronization (i-os2syn.ads)
12465 @section @code{Interfaces.Os2lib.Synchronization} (@file{i-os2syn.ads})
12466 @cindex @code{Interfaces.Os2lib.Synchronization} (@file{i-os2syn.ads})
12467 @cindex Interfacing, to OS/2
12468 @cindex Synchronization, OS/2
12469 @cindex OS/2 synchronization primitives
12472 This is a child package that provides definitions for interfacing
12473 to the @code{OS/2} synchronization primitives.
12475 @node Interfaces.Os2lib.Threads (i-os2thr.ads)
12476 @section @code{Interfaces.Os2lib.Threads} (@file{i-os2thr.ads})
12477 @cindex @code{Interfaces.Os2lib.Threads} (@file{i-os2thr.ads})
12478 @cindex Interfacing, to OS/2
12479 @cindex Thread control, OS/2
12480 @cindex OS/2 thread interfacing
12483 This is a child package that provides definitions for interfacing
12484 to the @code{OS/2} thread primitives.
12486 @node Interfaces.Packed_Decimal (i-pacdec.ads)
12487 @section @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
12488 @cindex @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
12489 @cindex IBM Packed Format
12490 @cindex Packed Decimal
12493 This package provides a set of routines for conversions to and
12494 from a packed decimal format compatible with that used on IBM
12497 @node Interfaces.VxWorks (i-vxwork.ads)
12498 @section @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
12499 @cindex @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
12500 @cindex Interfacing to VxWorks
12501 @cindex VxWorks, interfacing
12504 This package provides a limited binding to the VxWorks API.
12505 In particular, it interfaces with the
12506 VxWorks hardware interrupt facilities.
12508 @node Interfaces.VxWorks.IO (i-vxwoio.ads)
12509 @section @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
12510 @cindex @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
12511 @cindex Interfacing to VxWorks' I/O
12512 @cindex VxWorks, I/O interfacing
12513 @cindex VxWorks, Get_Immediate
12514 @cindex Get_Immediate, VxWorks
12517 This package provides a binding to the ioctl (IO/Control)
12518 function of VxWorks, defining a set of option values and
12519 function codes. A particular use of this package is
12520 to enable the use of Get_Immediate under VxWorks.
12522 @node System.Address_Image (s-addima.ads)
12523 @section @code{System.Address_Image} (@file{s-addima.ads})
12524 @cindex @code{System.Address_Image} (@file{s-addima.ads})
12525 @cindex Address image
12526 @cindex Image, of an address
12529 This function provides a useful debugging
12530 function that gives an (implementation dependent)
12531 string which identifies an address.
12533 @node System.Assertions (s-assert.ads)
12534 @section @code{System.Assertions} (@file{s-assert.ads})
12535 @cindex @code{System.Assertions} (@file{s-assert.ads})
12537 @cindex Assert_Failure, exception
12540 This package provides the declaration of the exception raised
12541 by an run-time assertion failure, as well as the routine that
12542 is used internally to raise this assertion.
12544 @node System.Memory (s-memory.ads)
12545 @section @code{System.Memory} (@file{s-memory.ads})
12546 @cindex @code{System.Memory} (@file{s-memory.ads})
12547 @cindex Memory allocation
12550 This package provides the interface to the low level routines used
12551 by the generated code for allocation and freeing storage for the
12552 default storage pool (analogous to the C routines malloc and free.
12553 It also provides a reallocation interface analogous to the C routine
12554 realloc. The body of this unit may be modified to provide alternative
12555 allocation mechanisms for the default pool, and in addition, direct
12556 calls to this unit may be made for low level allocation uses (for
12557 example see the body of @code{GNAT.Tables}).
12559 @node System.Partition_Interface (s-parint.ads)
12560 @section @code{System.Partition_Interface} (@file{s-parint.ads})
12561 @cindex @code{System.Partition_Interface} (@file{s-parint.ads})
12562 @cindex Partition intefacing functions
12565 This package provides facilities for partition interfacing. It
12566 is used primarily in a distribution context when using Annex E
12569 @node System.Restrictions (s-restri.ads)
12570 @section @code{System.Restrictions} (@file{s-restri.ads})
12571 @cindex @code{System.Restrictions} (@file{s-restri.ads})
12572 @cindex Run-time restrictions access
12575 This package provides facilities for accessing at run-time
12576 the status of restrictions specified at compile time for
12577 the partition. Information is available both with regard
12578 to actual restrictions specified, and with regard to
12579 compiler determined information on which restrictions
12580 are violated by one or more packages in the partition.
12582 @node System.Rident (s-rident.ads)
12583 @section @code{System.Rident} (@file{s-rident.ads})
12584 @cindex @code{System.Rident} (@file{s-rident.ads})
12585 @cindex Restrictions definitions
12588 This package provides definitions of the restrictions
12589 identifiers supported by GNAT, and also the format of
12590 the restrictions provided in package System.Restrictions.
12591 It is not normally necessary to @code{with} this generic package
12592 since the necessary instantiation is included in
12593 package System.Restrictions.
12595 @node System.Task_Info (s-tasinf.ads)
12596 @section @code{System.Task_Info} (@file{s-tasinf.ads})
12597 @cindex @code{System.Task_Info} (@file{s-tasinf.ads})
12598 @cindex Task_Info pragma
12601 This package provides target dependent functionality that is used
12602 to support the @code{Task_Info} pragma
12604 @node System.Wch_Cnv (s-wchcnv.ads)
12605 @section @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
12606 @cindex @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
12607 @cindex Wide Character, Representation
12608 @cindex Wide String, Conversion
12609 @cindex Representation of wide characters
12612 This package provides routines for converting between
12613 wide characters and a representation as a value of type
12614 @code{Standard.String}, using a specified wide character
12615 encoding method. It uses definitions in
12616 package @code{System.Wch_Con}.
12618 @node System.Wch_Con (s-wchcon.ads)
12619 @section @code{System.Wch_Con} (@file{s-wchcon.ads})
12620 @cindex @code{System.Wch_Con} (@file{s-wchcon.ads})
12623 This package provides definitions and descriptions of
12624 the various methods used for encoding wide characters
12625 in ordinary strings. These definitions are used by
12626 the package @code{System.Wch_Cnv}.
12628 @node Interfacing to Other Languages
12629 @chapter Interfacing to Other Languages
12631 The facilities in annex B of the Ada 95 Reference Manual are fully
12632 implemented in GNAT, and in addition, a full interface to C++ is
12636 * Interfacing to C::
12637 * Interfacing to C++::
12638 * Interfacing to COBOL::
12639 * Interfacing to Fortran::
12640 * Interfacing to non-GNAT Ada code::
12643 @node Interfacing to C
12644 @section Interfacing to C
12647 Interfacing to C with GNAT can use one of two approaches:
12651 The types in the package @code{Interfaces.C} may be used.
12653 Standard Ada types may be used directly. This may be less portable to
12654 other compilers, but will work on all GNAT compilers, which guarantee
12655 correspondence between the C and Ada types.
12659 Pragma @code{Convention C} may be applied to Ada types, but mostly has no
12660 effect, since this is the default. The following table shows the
12661 correspondence between Ada scalar types and the corresponding C types.
12666 @item Short_Integer
12668 @item Short_Short_Integer
12672 @item Long_Long_Integer
12680 @item Long_Long_Float
12681 This is the longest floating-point type supported by the hardware.
12685 Additionally, there are the following general correspondences between Ada
12689 Ada enumeration types map to C enumeration types directly if pragma
12690 @code{Convention C} is specified, which causes them to have int
12691 length. Without pragma @code{Convention C}, Ada enumeration types map to
12692 8, 16, or 32 bits (i.e.@: C types @code{signed char}, @code{short},
12693 @code{int}, respectively) depending on the number of values passed.
12694 This is the only case in which pragma @code{Convention C} affects the
12695 representation of an Ada type.
12698 Ada access types map to C pointers, except for the case of pointers to
12699 unconstrained types in Ada, which have no direct C equivalent.
12702 Ada arrays map directly to C arrays.
12705 Ada records map directly to C structures.
12708 Packed Ada records map to C structures where all members are bit fields
12709 of the length corresponding to the @code{@var{type}'Size} value in Ada.
12712 @node Interfacing to C++
12713 @section Interfacing to C++
12716 The interface to C++ makes use of the following pragmas, which are
12717 primarily intended to be constructed automatically using a binding generator
12718 tool, although it is possible to construct them by hand. No suitable binding
12719 generator tool is supplied with GNAT though.
12721 Using these pragmas it is possible to achieve complete
12722 inter-operability between Ada tagged types and C class definitions.
12723 See @ref{Implementation Defined Pragmas}, for more details.
12726 @item pragma CPP_Class ([Entity =>] @var{local_name})
12727 The argument denotes an entity in the current declarative region that is
12728 declared as a tagged or untagged record type. It indicates that the type
12729 corresponds to an externally declared C++ class type, and is to be laid
12730 out the same way that C++ would lay out the type.
12732 @item pragma CPP_Constructor ([Entity =>] @var{local_name})
12733 This pragma identifies an imported function (imported in the usual way
12734 with pragma @code{Import}) as corresponding to a C++ constructor.
12736 @item pragma CPP_Vtable @dots{}
12737 One @code{CPP_Vtable} pragma can be present for each component of type
12738 @code{CPP.Interfaces.Vtable_Ptr} in a record to which pragma @code{CPP_Class}
12742 @node Interfacing to COBOL
12743 @section Interfacing to COBOL
12746 Interfacing to COBOL is achieved as described in section B.4 of
12747 the Ada 95 reference manual.
12749 @node Interfacing to Fortran
12750 @section Interfacing to Fortran
12753 Interfacing to Fortran is achieved as described in section B.5 of the
12754 reference manual. The pragma @code{Convention Fortran}, applied to a
12755 multi-dimensional array causes the array to be stored in column-major
12756 order as required for convenient interface to Fortran.
12758 @node Interfacing to non-GNAT Ada code
12759 @section Interfacing to non-GNAT Ada code
12761 It is possible to specify the convention @code{Ada} in a pragma
12762 @code{Import} or pragma @code{Export}. However this refers to
12763 the calling conventions used by GNAT, which may or may not be
12764 similar enough to those used by some other Ada 83 or Ada 95
12765 compiler to allow interoperation.
12767 If arguments types are kept simple, and if the foreign compiler generally
12768 follows system calling conventions, then it may be possible to integrate
12769 files compiled by other Ada compilers, provided that the elaboration
12770 issues are adequately addressed (for example by eliminating the
12771 need for any load time elaboration).
12773 In particular, GNAT running on VMS is designed to
12774 be highly compatible with the DEC Ada 83 compiler, so this is one
12775 case in which it is possible to import foreign units of this type,
12776 provided that the data items passed are restricted to simple scalar
12777 values or simple record types without variants, or simple array
12778 types with fixed bounds.
12780 @node Specialized Needs Annexes
12781 @chapter Specialized Needs Annexes
12784 Ada 95 defines a number of specialized needs annexes, which are not
12785 required in all implementations. However, as described in this chapter,
12786 GNAT implements all of these special needs annexes:
12789 @item Systems Programming (Annex C)
12790 The Systems Programming Annex is fully implemented.
12792 @item Real-Time Systems (Annex D)
12793 The Real-Time Systems Annex is fully implemented.
12795 @item Distributed Systems (Annex E)
12796 Stub generation is fully implemented in the GNAT compiler. In addition,
12797 a complete compatible PCS is available as part of the GLADE system,
12798 a separate product. When the two
12799 products are used in conjunction, this annex is fully implemented.
12801 @item Information Systems (Annex F)
12802 The Information Systems annex is fully implemented.
12804 @item Numerics (Annex G)
12805 The Numerics Annex is fully implemented.
12807 @item Safety and Security (Annex H)
12808 The Safety and Security annex is fully implemented.
12811 @node Implementation of Specific Ada Features
12812 @chapter Implementation of Specific Ada Features
12815 This chapter describes the GNAT implementation of several Ada language
12819 * Machine Code Insertions::
12820 * GNAT Implementation of Tasking::
12821 * GNAT Implementation of Shared Passive Packages::
12822 * Code Generation for Array Aggregates::
12823 * The Size of Discriminated Records with Default Discriminants::
12826 @node Machine Code Insertions
12827 @section Machine Code Insertions
12830 Package @code{Machine_Code} provides machine code support as described
12831 in the Ada 95 Reference Manual in two separate forms:
12834 Machine code statements, consisting of qualified expressions that
12835 fit the requirements of RM section 13.8.
12837 An intrinsic callable procedure, providing an alternative mechanism of
12838 including machine instructions in a subprogram.
12842 The two features are similar, and both are closely related to the mechanism
12843 provided by the asm instruction in the GNU C compiler. Full understanding
12844 and use of the facilities in this package requires understanding the asm
12845 instruction as described in @cite{Using the GNU Compiler Collection (GCC)}
12846 by Richard Stallman. The relevant section is titled ``Extensions to the C
12847 Language Family'' -> ``Assembler Instructions with C Expression Operands''.
12849 Calls to the function @code{Asm} and the procedure @code{Asm} have identical
12850 semantic restrictions and effects as described below. Both are provided so
12851 that the procedure call can be used as a statement, and the function call
12852 can be used to form a code_statement.
12854 The first example given in the GCC documentation is the C @code{asm}
12857 asm ("fsinx %1 %0" : "=f" (result) : "f" (angle));
12861 The equivalent can be written for GNAT as:
12863 @smallexample @c ada
12864 Asm ("fsinx %1 %0",
12865 My_Float'Asm_Output ("=f", result),
12866 My_Float'Asm_Input ("f", angle));
12870 The first argument to @code{Asm} is the assembler template, and is
12871 identical to what is used in GNU C@. This string must be a static
12872 expression. The second argument is the output operand list. It is
12873 either a single @code{Asm_Output} attribute reference, or a list of such
12874 references enclosed in parentheses (technically an array aggregate of
12877 The @code{Asm_Output} attribute denotes a function that takes two
12878 parameters. The first is a string, the second is the name of a variable
12879 of the type designated by the attribute prefix. The first (string)
12880 argument is required to be a static expression and designates the
12881 constraint for the parameter (e.g.@: what kind of register is
12882 required). The second argument is the variable to be updated with the
12883 result. The possible values for constraint are the same as those used in
12884 the RTL, and are dependent on the configuration file used to build the
12885 GCC back end. If there are no output operands, then this argument may
12886 either be omitted, or explicitly given as @code{No_Output_Operands}.
12888 The second argument of @code{@var{my_float}'Asm_Output} functions as
12889 though it were an @code{out} parameter, which is a little curious, but
12890 all names have the form of expressions, so there is no syntactic
12891 irregularity, even though normally functions would not be permitted
12892 @code{out} parameters. The third argument is the list of input
12893 operands. It is either a single @code{Asm_Input} attribute reference, or
12894 a list of such references enclosed in parentheses (technically an array
12895 aggregate of such references).
12897 The @code{Asm_Input} attribute denotes a function that takes two
12898 parameters. The first is a string, the second is an expression of the
12899 type designated by the prefix. The first (string) argument is required
12900 to be a static expression, and is the constraint for the parameter,
12901 (e.g.@: what kind of register is required). The second argument is the
12902 value to be used as the input argument. The possible values for the
12903 constant are the same as those used in the RTL, and are dependent on
12904 the configuration file used to built the GCC back end.
12906 If there are no input operands, this argument may either be omitted, or
12907 explicitly given as @code{No_Input_Operands}. The fourth argument, not
12908 present in the above example, is a list of register names, called the
12909 @dfn{clobber} argument. This argument, if given, must be a static string
12910 expression, and is a space or comma separated list of names of registers
12911 that must be considered destroyed as a result of the @code{Asm} call. If
12912 this argument is the null string (the default value), then the code
12913 generator assumes that no additional registers are destroyed.
12915 The fifth argument, not present in the above example, called the
12916 @dfn{volatile} argument, is by default @code{False}. It can be set to
12917 the literal value @code{True} to indicate to the code generator that all
12918 optimizations with respect to the instruction specified should be
12919 suppressed, and that in particular, for an instruction that has outputs,
12920 the instruction will still be generated, even if none of the outputs are
12921 used. See the full description in the GCC manual for further details.
12923 The @code{Asm} subprograms may be used in two ways. First the procedure
12924 forms can be used anywhere a procedure call would be valid, and
12925 correspond to what the RM calls ``intrinsic'' routines. Such calls can
12926 be used to intersperse machine instructions with other Ada statements.
12927 Second, the function forms, which return a dummy value of the limited
12928 private type @code{Asm_Insn}, can be used in code statements, and indeed
12929 this is the only context where such calls are allowed. Code statements
12930 appear as aggregates of the form:
12932 @smallexample @c ada
12933 Asm_Insn'(Asm (@dots{}));
12934 Asm_Insn'(Asm_Volatile (@dots{}));
12938 In accordance with RM rules, such code statements are allowed only
12939 within subprograms whose entire body consists of such statements. It is
12940 not permissible to intermix such statements with other Ada statements.
12942 Typically the form using intrinsic procedure calls is more convenient
12943 and more flexible. The code statement form is provided to meet the RM
12944 suggestion that such a facility should be made available. The following
12945 is the exact syntax of the call to @code{Asm}. As usual, if named notation
12946 is used, the arguments may be given in arbitrary order, following the
12947 normal rules for use of positional and named arguments)
12951 [Template =>] static_string_EXPRESSION
12952 [,[Outputs =>] OUTPUT_OPERAND_LIST ]
12953 [,[Inputs =>] INPUT_OPERAND_LIST ]
12954 [,[Clobber =>] static_string_EXPRESSION ]
12955 [,[Volatile =>] static_boolean_EXPRESSION] )
12957 OUTPUT_OPERAND_LIST ::=
12958 [PREFIX.]No_Output_Operands
12959 | OUTPUT_OPERAND_ATTRIBUTE
12960 | (OUTPUT_OPERAND_ATTRIBUTE @{,OUTPUT_OPERAND_ATTRIBUTE@})
12962 OUTPUT_OPERAND_ATTRIBUTE ::=
12963 SUBTYPE_MARK'Asm_Output (static_string_EXPRESSION, NAME)
12965 INPUT_OPERAND_LIST ::=
12966 [PREFIX.]No_Input_Operands
12967 | INPUT_OPERAND_ATTRIBUTE
12968 | (INPUT_OPERAND_ATTRIBUTE @{,INPUT_OPERAND_ATTRIBUTE@})
12970 INPUT_OPERAND_ATTRIBUTE ::=
12971 SUBTYPE_MARK'Asm_Input (static_string_EXPRESSION, EXPRESSION)
12975 The identifiers @code{No_Input_Operands} and @code{No_Output_Operands}
12976 are declared in the package @code{Machine_Code} and must be referenced
12977 according to normal visibility rules. In particular if there is no
12978 @code{use} clause for this package, then appropriate package name
12979 qualification is required.
12981 @node GNAT Implementation of Tasking
12982 @section GNAT Implementation of Tasking
12985 This chapter outlines the basic GNAT approach to tasking (in particular,
12986 a multi-layered library for portability) and discusses issues related
12987 to compliance with the Real-Time Systems Annex.
12990 * Mapping Ada Tasks onto the Underlying Kernel Threads::
12991 * Ensuring Compliance with the Real-Time Annex::
12994 @node Mapping Ada Tasks onto the Underlying Kernel Threads
12995 @subsection Mapping Ada Tasks onto the Underlying Kernel Threads
12998 GNAT's run-time support comprises two layers:
13001 @item GNARL (GNAT Run-time Layer)
13002 @item GNULL (GNAT Low-level Library)
13006 In GNAT, Ada's tasking services rely on a platform and OS independent
13007 layer known as GNARL@. This code is responsible for implementing the
13008 correct semantics of Ada's task creation, rendezvous, protected
13011 GNARL decomposes Ada's tasking semantics into simpler lower level
13012 operations such as create a thread, set the priority of a thread,
13013 yield, create a lock, lock/unlock, etc. The spec for these low-level
13014 operations constitutes GNULLI, the GNULL Interface. This interface is
13015 directly inspired from the POSIX real-time API@.
13017 If the underlying executive or OS implements the POSIX standard
13018 faithfully, the GNULL Interface maps as is to the services offered by
13019 the underlying kernel. Otherwise, some target dependent glue code maps
13020 the services offered by the underlying kernel to the semantics expected
13023 Whatever the underlying OS (VxWorks, UNIX, OS/2, Windows NT, etc.) the
13024 key point is that each Ada task is mapped on a thread in the underlying
13025 kernel. For example, in the case of VxWorks, one Ada task = one VxWorks task.
13027 In addition Ada task priorities map onto the underlying thread priorities.
13028 Mapping Ada tasks onto the underlying kernel threads has several advantages:
13032 The underlying scheduler is used to schedule the Ada tasks. This
13033 makes Ada tasks as efficient as kernel threads from a scheduling
13037 Interaction with code written in C containing threads is eased
13038 since at the lowest level Ada tasks and C threads map onto the same
13039 underlying kernel concept.
13042 When an Ada task is blocked during I/O the remaining Ada tasks are
13046 On multiprocessor systems Ada tasks can execute in parallel.
13050 Some threads libraries offer a mechanism to fork a new process, with the
13051 child process duplicating the threads from the parent.
13053 support this functionality when the parent contains more than one task.
13054 @cindex Forking a new process
13056 @node Ensuring Compliance with the Real-Time Annex
13057 @subsection Ensuring Compliance with the Real-Time Annex
13058 @cindex Real-Time Systems Annex compliance
13061 Although mapping Ada tasks onto
13062 the underlying threads has significant advantages, it does create some
13063 complications when it comes to respecting the scheduling semantics
13064 specified in the real-time annex (Annex D).
13066 For instance the Annex D requirement for the @code{FIFO_Within_Priorities}
13067 scheduling policy states:
13070 @emph{When the active priority of a ready task that is not running
13071 changes, or the setting of its base priority takes effect, the
13072 task is removed from the ready queue for its old active priority
13073 and is added at the tail of the ready queue for its new active
13074 priority, except in the case where the active priority is lowered
13075 due to the loss of inherited priority, in which case the task is
13076 added at the head of the ready queue for its new active priority.}
13080 While most kernels do put tasks at the end of the priority queue when
13081 a task changes its priority, (which respects the main
13082 FIFO_Within_Priorities requirement), almost none keep a thread at the
13083 beginning of its priority queue when its priority drops from the loss
13084 of inherited priority.
13086 As a result most vendors have provided incomplete Annex D implementations.
13088 The GNAT run-time, has a nice cooperative solution to this problem
13089 which ensures that accurate FIFO_Within_Priorities semantics are
13092 The principle is as follows. When an Ada task T is about to start
13093 running, it checks whether some other Ada task R with the same
13094 priority as T has been suspended due to the loss of priority
13095 inheritance. If this is the case, T yields and is placed at the end of
13096 its priority queue. When R arrives at the front of the queue it
13099 Note that this simple scheme preserves the relative order of the tasks
13100 that were ready to execute in the priority queue where R has been
13103 @node GNAT Implementation of Shared Passive Packages
13104 @section GNAT Implementation of Shared Passive Packages
13105 @cindex Shared passive packages
13108 GNAT fully implements the pragma @code{Shared_Passive} for
13109 @cindex pragma @code{Shared_Passive}
13110 the purpose of designating shared passive packages.
13111 This allows the use of passive partitions in the
13112 context described in the Ada Reference Manual; i.e. for communication
13113 between separate partitions of a distributed application using the
13114 features in Annex E.
13116 @cindex Distribution Systems Annex
13118 However, the implementation approach used by GNAT provides for more
13119 extensive usage as follows:
13122 @item Communication between separate programs
13124 This allows separate programs to access the data in passive
13125 partitions, using protected objects for synchronization where
13126 needed. The only requirement is that the two programs have a
13127 common shared file system. It is even possible for programs
13128 running on different machines with different architectures
13129 (e.g. different endianness) to communicate via the data in
13130 a passive partition.
13132 @item Persistence between program runs
13134 The data in a passive package can persist from one run of a
13135 program to another, so that a later program sees the final
13136 values stored by a previous run of the same program.
13141 The implementation approach used is to store the data in files. A
13142 separate stream file is created for each object in the package, and
13143 an access to an object causes the corresponding file to be read or
13146 The environment variable @code{SHARED_MEMORY_DIRECTORY} should be
13147 @cindex @code{SHARED_MEMORY_DIRECTORY} environment variable
13148 set to the directory to be used for these files.
13149 The files in this directory
13150 have names that correspond to their fully qualified names. For
13151 example, if we have the package
13153 @smallexample @c ada
13155 pragma Shared_Passive (X);
13162 and the environment variable is set to @code{/stemp/}, then the files created
13163 will have the names:
13171 These files are created when a value is initially written to the object, and
13172 the files are retained until manually deleted. This provides the persistence
13173 semantics. If no file exists, it means that no partition has assigned a value
13174 to the variable; in this case the initial value declared in the package
13175 will be used. This model ensures that there are no issues in synchronizing
13176 the elaboration process, since elaboration of passive packages elaborates the
13177 initial values, but does not create the files.
13179 The files are written using normal @code{Stream_IO} access.
13180 If you want to be able
13181 to communicate between programs or partitions running on different
13182 architectures, then you should use the XDR versions of the stream attribute
13183 routines, since these are architecture independent.
13185 If active synchronization is required for access to the variables in the
13186 shared passive package, then as described in the Ada Reference Manual, the
13187 package may contain protected objects used for this purpose. In this case
13188 a lock file (whose name is @file{___lock} (three underscores)
13189 is created in the shared memory directory.
13190 @cindex @file{___lock} file (for shared passive packages)
13191 This is used to provide the required locking
13192 semantics for proper protected object synchronization.
13194 As of January 2003, GNAT supports shared passive packages on all platforms
13195 except for OpenVMS.
13197 @node Code Generation for Array Aggregates
13198 @section Code Generation for Array Aggregates
13201 * Static constant aggregates with static bounds::
13202 * Constant aggregates with an unconstrained nominal types::
13203 * Aggregates with static bounds::
13204 * Aggregates with non-static bounds::
13205 * Aggregates in assignment statements::
13209 Aggregate have a rich syntax and allow the user to specify the values of
13210 complex data structures by means of a single construct. As a result, the
13211 code generated for aggregates can be quite complex and involve loops, case
13212 statements and multiple assignments. In the simplest cases, however, the
13213 compiler will recognize aggregates whose components and constraints are
13214 fully static, and in those cases the compiler will generate little or no
13215 executable code. The following is an outline of the code that GNAT generates
13216 for various aggregate constructs. For further details, the user will find it
13217 useful to examine the output produced by the -gnatG flag to see the expanded
13218 source that is input to the code generator. The user will also want to examine
13219 the assembly code generated at various levels of optimization.
13221 The code generated for aggregates depends on the context, the component values,
13222 and the type. In the context of an object declaration the code generated is
13223 generally simpler than in the case of an assignment. As a general rule, static
13224 component values and static subtypes also lead to simpler code.
13226 @node Static constant aggregates with static bounds
13227 @subsection Static constant aggregates with static bounds
13230 For the declarations:
13231 @smallexample @c ada
13232 type One_Dim is array (1..10) of integer;
13233 ar0 : constant One_Dim := ( 1, 2, 3, 4, 5, 6, 7, 8, 9, 0);
13237 GNAT generates no executable code: the constant ar0 is placed in static memory.
13238 The same is true for constant aggregates with named associations:
13240 @smallexample @c ada
13241 Cr1 : constant One_Dim := (4 => 16, 2 => 4, 3 => 9, 1=> 1);
13242 Cr3 : constant One_Dim := (others => 7777);
13246 The same is true for multidimensional constant arrays such as:
13248 @smallexample @c ada
13249 type two_dim is array (1..3, 1..3) of integer;
13250 Unit : constant two_dim := ( (1,0,0), (0,1,0), (0,0,1));
13254 The same is true for arrays of one-dimensional arrays: the following are
13257 @smallexample @c ada
13258 type ar1b is array (1..3) of boolean;
13259 type ar_ar is array (1..3) of ar1b;
13260 None : constant ar1b := (others => false); -- fully static
13261 None2 : constant ar_ar := (1..3 => None); -- fully static
13265 However, for multidimensional aggregates with named associations, GNAT will
13266 generate assignments and loops, even if all associations are static. The
13267 following two declarations generate a loop for the first dimension, and
13268 individual component assignments for the second dimension:
13270 @smallexample @c ada
13271 Zero1: constant two_dim := (1..3 => (1..3 => 0));
13272 Zero2: constant two_dim := (others => (others => 0));
13275 @node Constant aggregates with an unconstrained nominal types
13276 @subsection Constant aggregates with an unconstrained nominal types
13279 In such cases the aggregate itself establishes the subtype, so that
13280 associations with @code{others} cannot be used. GNAT determines the
13281 bounds for the actual subtype of the aggregate, and allocates the
13282 aggregate statically as well. No code is generated for the following:
13284 @smallexample @c ada
13285 type One_Unc is array (natural range <>) of integer;
13286 Cr_Unc : constant One_Unc := (12,24,36);
13289 @node Aggregates with static bounds
13290 @subsection Aggregates with static bounds
13293 In all previous examples the aggregate was the initial (and immutable) value
13294 of a constant. If the aggregate initializes a variable, then code is generated
13295 for it as a combination of individual assignments and loops over the target
13296 object. The declarations
13298 @smallexample @c ada
13299 Cr_Var1 : One_Dim := (2, 5, 7, 11);
13300 Cr_Var2 : One_Dim := (others > -1);
13304 generate the equivalent of
13306 @smallexample @c ada
13312 for I in Cr_Var2'range loop
13313 Cr_Var2 (I) := =-1;
13317 @node Aggregates with non-static bounds
13318 @subsection Aggregates with non-static bounds
13321 If the bounds of the aggregate are not statically compatible with the bounds
13322 of the nominal subtype of the target, then constraint checks have to be
13323 generated on the bounds. For a multidimensional array, constraint checks may
13324 have to be applied to sub-arrays individually, if they do not have statically
13325 compatible subtypes.
13327 @node Aggregates in assignment statements
13328 @subsection Aggregates in assignment statements
13331 In general, aggregate assignment requires the construction of a temporary,
13332 and a copy from the temporary to the target of the assignment. This is because
13333 it is not always possible to convert the assignment into a series of individual
13334 component assignments. For example, consider the simple case:
13336 @smallexample @c ada
13341 This cannot be converted into:
13343 @smallexample @c ada
13349 So the aggregate has to be built first in a separate location, and then
13350 copied into the target. GNAT recognizes simple cases where this intermediate
13351 step is not required, and the assignments can be performed in place, directly
13352 into the target. The following sufficient criteria are applied:
13356 The bounds of the aggregate are static, and the associations are static.
13358 The components of the aggregate are static constants, names of
13359 simple variables that are not renamings, or expressions not involving
13360 indexed components whose operands obey these rules.
13364 If any of these conditions are violated, the aggregate will be built in
13365 a temporary (created either by the front-end or the code generator) and then
13366 that temporary will be copied onto the target.
13369 @node The Size of Discriminated Records with Default Discriminants
13370 @section The Size of Discriminated Records with Default Discriminants
13373 If a discriminated type @code{T} has discriminants with default values, it is
13374 possible to declare an object of this type without providing an explicit
13377 @smallexample @c ada
13379 type Size is range 1..100;
13381 type Rec (D : Size := 15) is record
13382 Name : String (1..D);
13390 Such an object is said to be @emph{unconstrained}.
13391 The discriminant of the object
13392 can be modified by a full assignment to the object, as long as it preserves the
13393 relation between the value of the discriminant, and the value of the components
13396 @smallexample @c ada
13398 Word := (3, "yes");
13400 Word := (5, "maybe");
13402 Word := (5, "no"); -- raises Constraint_Error
13407 In order to support this behavior efficiently, an unconstrained object is
13408 given the maximum size that any value of the type requires. In the case
13409 above, @code{Word} has storage for the discriminant and for
13410 a @code{String} of length 100.
13411 It is important to note that unconstrained objects do not require dynamic
13412 allocation. It would be an improper implementation to place on the heap those
13413 components whose size depends on discriminants. (This improper implementation
13414 was used by some Ada83 compilers, where the @code{Name} component above
13416 been stored as a pointer to a dynamic string). Following the principle that
13417 dynamic storage management should never be introduced implicitly,
13418 an Ada95 compiler should reserve the full size for an unconstrained declared
13419 object, and place it on the stack.
13421 This maximum size approach
13422 has been a source of surprise to some users, who expect the default
13423 values of the discriminants to determine the size reserved for an
13424 unconstrained object: ``If the default is 15, why should the object occupy
13426 The answer, of course, is that the discriminant may be later modified,
13427 and its full range of values must be taken into account. This is why the
13432 type Rec (D : Positive := 15) is record
13433 Name : String (1..D);
13441 is flagged by the compiler with a warning:
13442 an attempt to create @code{Too_Large} will raise @code{Storage_Error},
13443 because the required size includes @code{Positive'Last}
13444 bytes. As the first example indicates, the proper approach is to declare an
13445 index type of ``reasonable'' range so that unconstrained objects are not too
13448 One final wrinkle: if the object is declared to be @code{aliased}, or if it is
13449 created in the heap by means of an allocator, then it is @emph{not}
13451 it is constrained by the default values of the discriminants, and those values
13452 cannot be modified by full assignment. This is because in the presence of
13453 aliasing all views of the object (which may be manipulated by different tasks,
13454 say) must be consistent, so it is imperative that the object, once created,
13460 @node Project File Reference
13461 @chapter Project File Reference
13464 This chapter describes the syntax and semantics of project files.
13465 Project files specify the options to be used when building a system.
13466 Project files can specify global settings for all tools,
13467 as well as tool-specific settings.
13468 See the chapter on project files in the GNAT Users guide for examples of use.
13472 * Lexical Elements::
13474 * Empty declarations::
13475 * Typed string declarations::
13479 * Project Attributes::
13480 * Attribute References::
13481 * External Values::
13482 * Case Construction::
13484 * Package Renamings::
13486 * Project Extensions::
13487 * Project File Elaboration::
13490 @node Reserved Words
13491 @section Reserved Words
13494 All Ada95 reserved words are reserved in project files, and cannot be used
13495 as variable names or project names. In addition, the following are
13496 also reserved in project files:
13499 @item @code{extends}
13501 @item @code{external}
13503 @item @code{project}
13507 @node Lexical Elements
13508 @section Lexical Elements
13511 Rules for identifiers are the same as in Ada95. Identifiers
13512 are case-insensitive. Strings are case sensitive, except where noted.
13513 Comments have the same form as in Ada95.
13523 simple_name @{. simple_name@}
13527 @section Declarations
13530 Declarations introduce new entities that denote types, variables, attributes,
13531 and packages. Some declarations can only appear immediately within a project
13532 declaration. Others can appear within a project or within a package.
13536 declarative_item ::=
13537 simple_declarative_item |
13538 typed_string_declaration |
13539 package_declaration
13541 simple_declarative_item ::=
13542 variable_declaration |
13543 typed_variable_declaration |
13544 attribute_declaration |
13545 case_construction |
13549 @node Empty declarations
13550 @section Empty declarations
13553 empty_declaration ::=
13557 An empty declaration is allowed anywhere a declaration is allowed.
13560 @node Typed string declarations
13561 @section Typed string declarations
13564 Typed strings are sequences of string literals. Typed strings are the only
13565 named types in project files. They are used in case constructions, where they
13566 provide support for conditional attribute definitions.
13570 typed_string_declaration ::=
13571 @b{type} <typed_string_>_simple_name @b{is}
13572 ( string_literal @{, string_literal@} );
13576 A typed string declaration can only appear immediately within a project
13579 All the string literals in a typed string declaration must be distinct.
13585 Variables denote values, and appear as constituents of expressions.
13588 typed_variable_declaration ::=
13589 <typed_variable_>simple_name : <typed_string_>name := string_expression ;
13591 variable_declaration ::=
13592 <variable_>simple_name := expression;
13596 The elaboration of a variable declaration introduces the variable and
13597 assigns to it the value of the expression. The name of the variable is
13598 available after the assignment symbol.
13601 A typed_variable can only be declare once.
13604 a non typed variable can be declared multiple times.
13607 Before the completion of its first declaration, the value of variable
13608 is the null string.
13611 @section Expressions
13614 An expression is a formula that defines a computation or retrieval of a value.
13615 In a project file the value of an expression is either a string or a list
13616 of strings. A string value in an expression is either a literal, the current
13617 value of a variable, an external value, an attribute reference, or a
13618 concatenation operation.
13631 attribute_reference
13637 ( <string_>expression @{ , <string_>expression @} )
13640 @subsection Concatenation
13642 The following concatenation functions are defined:
13644 @smallexample @c ada
13645 function "&" (X : String; Y : String) return String;
13646 function "&" (X : String_List; Y : String) return String_List;
13647 function "&" (X : String_List; Y : String_List) return String_List;
13651 @section Attributes
13654 An attribute declaration defines a property of a project or package. This
13655 property can later be queried by means of an attribute reference.
13656 Attribute values are strings or string lists.
13658 Some attributes are associative arrays. These attributes are mappings whose
13659 domain is a set of strings. These attributes are declared one association
13660 at a time, by specifying a point in the domain and the corresponding image
13661 of the attribute. They may also be declared as a full associative array,
13662 getting the same associations as the corresponding attribute in an imported
13663 or extended project.
13665 Attributes that are not associative arrays are called simple attributes.
13669 attribute_declaration ::=
13670 full_associative_array_declaration |
13671 @b{for} attribute_designator @b{use} expression ;
13673 full_associative_array_declaration ::=
13674 @b{for} <associative_array_attribute_>simple_name @b{use}
13675 <project_>simple_name [ . <package_>simple_Name ] ' <attribute_>simple_name ;
13677 attribute_designator ::=
13678 <simple_attribute_>simple_name |
13679 <associative_array_attribute_>simple_name ( string_literal )
13683 Some attributes are project-specific, and can only appear immediately within
13684 a project declaration. Others are package-specific, and can only appear within
13685 the proper package.
13687 The expression in an attribute definition must be a string or a string_list.
13688 The string literal appearing in the attribute_designator of an associative
13689 array attribute is case-insensitive.
13691 @node Project Attributes
13692 @section Project Attributes
13695 The following attributes apply to a project. All of them are simple
13700 Expression must be a path name. The attribute defines the
13701 directory in which the object files created by the build are to be placed. If
13702 not specified, object files are placed in the project directory.
13705 Expression must be a path name. The attribute defines the
13706 directory in which the executables created by the build are to be placed.
13707 If not specified, executables are placed in the object directory.
13710 Expression must be a list of path names. The attribute
13711 defines the directories in which the source files for the project are to be
13712 found. If not specified, source files are found in the project directory.
13715 Expression must be a list of file names. The attribute
13716 defines the individual files, in the project directory, which are to be used
13717 as sources for the project. File names are path_names that contain no directory
13718 information. If the project has no sources the attribute must be declared
13719 explicitly with an empty list.
13721 @item Source_List_File
13722 Expression must a single path name. The attribute
13723 defines a text file that contains a list of source file names to be used
13724 as sources for the project
13727 Expression must be a path name. The attribute defines the
13728 directory in which a library is to be built. The directory must exist, must
13729 be distinct from the project's object directory, and must be writable.
13732 Expression must be a string that is a legal file name,
13733 without extension. The attribute defines a string that is used to generate
13734 the name of the library to be built by the project.
13737 Argument must be a string value that must be one of the
13738 following @code{"static"}, @code{"dynamic"} or @code{"relocatable"}. This
13739 string is case-insensitive. If this attribute is not specified, the library is
13740 a static library. Otherwise, the library may be dynamic or relocatable. This
13741 distinction is operating-system dependent.
13743 @item Library_Version
13744 Expression must be a string value whose interpretation
13745 is platform dependent. On UNIX, it is used only for dynamic/relocatable
13746 libraries as the internal name of the library (the @code{"soname"}). If the
13747 library file name (built from the @code{Library_Name}) is different from the
13748 @code{Library_Version}, then the library file will be a symbolic link to the
13749 actual file whose name will be @code{Library_Version}.
13751 @item Library_Interface
13752 Expression must be a string list. Each element of the string list
13753 must designate a unit of the project.
13754 If this attribute is present in a Library Project File, then the project
13755 file is a Stand-alone Library_Project_File.
13757 @item Library_Auto_Init
13758 Expression must be a single string "true" or "false", case-insensitive.
13759 If this attribute is present in a Stand-alone Library Project File,
13760 it indicates if initialization is automatic when the dynamic library
13763 @item Library_Options
13764 Expression must be a string list. Indicates additional switches that
13765 are to be used when building a shared library.
13768 Expression must be a single string. Designates an alternative to "gcc"
13769 for building shared libraries.
13771 @item Library_Src_Dir
13772 Expression must be a path name. The attribute defines the
13773 directory in which the sources of the interfaces of a Stand-alone Library will
13774 be copied. The directory must exist, must be distinct from the project's
13775 object directory and source directories, and must be writable.
13778 Expression must be a list of strings that are legal file names.
13779 These file names designate existing compilation units in the source directory
13780 that are legal main subprograms.
13782 When a project file is elaborated, as part of the execution of a gnatmake
13783 command, one or several executables are built and placed in the Exec_Dir.
13784 If the gnatmake command does not include explicit file names, the executables
13785 that are built correspond to the files specified by this attribute.
13787 @item Main_Language
13788 This is a simple attribute. Its value is a string that specifies the
13789 language of the main program.
13792 Expression must be a string list. Each string designates
13793 a programming language that is known to GNAT. The strings are case-insensitive.
13795 @item Locally_Removed_Files
13796 This attribute is legal only in a project file that extends another.
13797 Expression must be a list of strings that are legal file names.
13798 Each file name must designate a source that would normally be inherited
13799 by the current project file. It cannot designate an immediate source that is
13800 not inherited. Each of the source files in the list are not considered to
13801 be sources of the project file: they are not inherited.
13804 @node Attribute References
13805 @section Attribute References
13808 Attribute references are used to retrieve the value of previously defined
13809 attribute for a package or project.
13812 attribute_reference ::=
13813 attribute_prefix ' <simple_attribute_>simple_name [ ( string_literal ) ]
13815 attribute_prefix ::=
13817 <project_simple_name | package_identifier |
13818 <project_>simple_name . package_identifier
13822 If an attribute has not been specified for a given package or project, its
13823 value is the null string or the empty list.
13825 @node External Values
13826 @section External Values
13829 An external value is an expression whose value is obtained from the command
13830 that invoked the processing of the current project file (typically a
13836 @b{external} ( string_literal [, string_literal] )
13840 The first string_literal is the string to be used on the command line or
13841 in the environment to specify the external value. The second string_literal,
13842 if present, is the default to use if there is no specification for this
13843 external value either on the command line or in the environment.
13845 @node Case Construction
13846 @section Case Construction
13849 A case construction supports attribute declarations that depend on the value of
13850 a previously declared variable.
13854 case_construction ::=
13855 @b{case} <typed_variable_>name @b{is}
13860 @b{when} discrete_choice_list =>
13861 @{case_construction | attribute_declaration | empty_declaration@}
13863 discrete_choice_list ::=
13864 string_literal @{| string_literal@} |
13869 All choices in a choice list must be distinct. The choice lists of two
13870 distinct alternatives must be disjoint. Unlike Ada, the choice lists of all
13871 alternatives do not need to include all values of the type. An @code{others}
13872 choice must appear last in the list of alternatives.
13878 A package provides a grouping of variable declarations and attribute
13879 declarations to be used when invoking various GNAT tools. The name of
13880 the package indicates the tool(s) to which it applies.
13884 package_declaration ::=
13885 package_specification | package_renaming
13887 package_specification ::=
13888 @b{package} package_identifier @b{is}
13889 @{simple_declarative_item@}
13890 @b{end} package_identifier ;
13892 package_identifier ::=
13893 @code{Naming} | @code{Builder} | @code{Compiler} | @code{Binder} |
13894 @code{Linker} | @code{Finder} | @code{Cross_Reference} |
13895 @code{gnatls} | @code{IDE} | @code{Pretty_Printer}
13898 @subsection Package Naming
13901 The attributes of a @code{Naming} package specifies the naming conventions
13902 that apply to the source files in a project. When invoking other GNAT tools,
13903 they will use the sources in the source directories that satisfy these
13904 naming conventions.
13906 The following attributes apply to a @code{Naming} package:
13910 This is a simple attribute whose value is a string. Legal values of this
13911 string are @code{"lowercase"}, @code{"uppercase"} or @code{"mixedcase"}.
13912 These strings are themselves case insensitive.
13915 If @code{Casing} is not specified, then the default is @code{"lowercase"}.
13917 @item Dot_Replacement
13918 This is a simple attribute whose string value satisfies the following
13922 @item It must not be empty
13923 @item It cannot start or end with an alphanumeric character
13924 @item It cannot be a single underscore
13925 @item It cannot start with an underscore followed by an alphanumeric
13926 @item It cannot contain a dot @code{'.'} if longer than one character
13930 If @code{Dot_Replacement} is not specified, then the default is @code{"-"}.
13933 This is an associative array attribute, defined on language names,
13934 whose image is a string that must satisfy the following
13938 @item It must not be empty
13939 @item It cannot start with an alphanumeric character
13940 @item It cannot start with an underscore followed by an alphanumeric character
13944 For Ada, the attribute denotes the suffix used in file names that contain
13945 library unit declarations, that is to say units that are package and
13946 subprogram declarations. If @code{Spec_Suffix ("Ada")} is not
13947 specified, then the default is @code{".ads"}.
13949 For C and C++, the attribute denotes the suffix used in file names that
13950 contain prototypes.
13953 This is an associative array attribute defined on language names,
13954 whose image is a string that must satisfy the following
13958 @item It must not be empty
13959 @item It cannot start with an alphanumeric character
13960 @item It cannot start with an underscore followed by an alphanumeric character
13961 @item It cannot be a suffix of @code{Spec_Suffix}
13965 For Ada, the attribute denotes the suffix used in file names that contain
13966 library bodies, that is to say units that are package and subprogram bodies.
13967 If @code{Body_Suffix ("Ada")} is not specified, then the default is
13970 For C and C++, the attribute denotes the suffix used in file names that contain
13973 @item Separate_Suffix
13974 This is a simple attribute whose value satisfies the same conditions as
13975 @code{Body_Suffix}.
13977 This attribute is specific to Ada. It denotes the suffix used in file names
13978 that contain separate bodies. If it is not specified, then it defaults to same
13979 value as @code{Body_Suffix ("Ada")}.
13982 This is an associative array attribute, specific to Ada, defined over
13983 compilation unit names. The image is a string that is the name of the file
13984 that contains that library unit. The file name is case sensitive if the
13985 conventions of the host operating system require it.
13988 This is an associative array attribute, specific to Ada, defined over
13989 compilation unit names. The image is a string that is the name of the file
13990 that contains the library unit body for the named unit. The file name is case
13991 sensitive if the conventions of the host operating system require it.
13993 @item Specification_Exceptions
13994 This is an associative array attribute defined on language names,
13995 whose value is a list of strings.
13997 This attribute is not significant for Ada.
13999 For C and C++, each string in the list denotes the name of a file that
14000 contains prototypes, but whose suffix is not necessarily the
14001 @code{Spec_Suffix} for the language.
14003 @item Implementation_Exceptions
14004 This is an associative array attribute defined on language names,
14005 whose value is a list of strings.
14007 This attribute is not significant for Ada.
14009 For C and C++, each string in the list denotes the name of a file that
14010 contains source code, but whose suffix is not necessarily the
14011 @code{Body_Suffix} for the language.
14014 The following attributes of package @code{Naming} are obsolescent. They are
14015 kept as synonyms of other attributes for compatibility with previous versions
14016 of the Project Manager.
14019 @item Specification_Suffix
14020 This is a synonym of @code{Spec_Suffix}.
14022 @item Implementation_Suffix
14023 This is a synonym of @code{Body_Suffix}.
14025 @item Specification
14026 This is a synonym of @code{Spec}.
14028 @item Implementation
14029 This is a synonym of @code{Body}.
14032 @subsection package Compiler
14035 The attributes of the @code{Compiler} package specify the compilation options
14036 to be used by the underlying compiler.
14039 @item Default_Switches
14040 This is an associative array attribute. Its
14041 domain is a set of language names. Its range is a string list that
14042 specifies the compilation options to be used when compiling a component
14043 written in that language, for which no file-specific switches have been
14047 This is an associative array attribute. Its domain is
14048 a set of file names. Its range is a string list that specifies the
14049 compilation options to be used when compiling the named file. If a file
14050 is not specified in the Switches attribute, it is compiled with the
14051 settings specified by Default_Switches.
14053 @item Local_Configuration_Pragmas.
14054 This is a simple attribute, whose
14055 value is a path name that designates a file containing configuration pragmas
14056 to be used for all invocations of the compiler for immediate sources of the
14060 This is an associative array attribute. Its domain is
14061 a set of main source file names. Its range is a simple string that specifies
14062 the executable file name to be used when linking the specified main source.
14063 If a main source is not specified in the Executable attribute, the executable
14064 file name is deducted from the main source file name.
14067 @subsection package Builder
14070 The attributes of package @code{Builder} specify the compilation, binding, and
14071 linking options to be used when building an executable for a project. The
14072 following attributes apply to package @code{Builder}:
14075 @item Default_Switches
14081 @item Global_Configuration_Pragmas
14082 This is a simple attribute, whose
14083 value is a path name that designates a file that contains configuration pragmas
14084 to be used in every build of an executable. If both local and global
14085 configuration pragmas are specified, a compilation makes use of both sets.
14088 This is an associative array attribute, defined over
14089 compilation unit names. The image is a string that is the name of the
14090 executable file corresponding to the main source file index.
14091 This attribute has no effect if its value is the empty string.
14093 @item Executable_Suffix
14094 This is a simple attribute whose value is a suffix to be added to
14095 the executables that don't have an attribute Executable specified.
14098 @subsection package Gnatls
14101 The attributes of package @code{Gnatls} specify the tool options to be used
14102 when invoking the library browser @command{gnatls}.
14103 The following attributes apply to package @code{Gnatls}:
14110 @subsection package Binder
14113 The attributes of package @code{Binder} specify the options to be used
14114 when invoking the binder in the construction of an executable.
14115 The following attributes apply to package @code{Binder}:
14118 @item Default_Switches
14124 @subsection package Linker
14127 The attributes of package @code{Linker} specify the options to be used when
14128 invoking the linker in the construction of an executable.
14129 The following attributes apply to package @code{Linker}:
14132 @item Default_Switches
14138 @subsection package Cross_Reference
14141 The attributes of package @code{Cross_Reference} specify the tool options
14143 when invoking the library tool @command{gnatxref}.
14144 The following attributes apply to package @code{Cross_Reference}:
14147 @item Default_Switches
14153 @subsection package Finder
14156 The attributes of package @code{Finder} specify the tool options to be used
14157 when invoking the search tool @command{gnatfind}.
14158 The following attributes apply to package @code{Finder}:
14161 @item Default_Switches
14167 @subsection package Pretty_Printer
14170 The attributes of package @code{Pretty_Printer}
14171 specify the tool options to be used
14172 when invoking the formatting tool @command{gnatpp}.
14173 The following attributes apply to package @code{Pretty_Printer}:
14176 @item Default_switches
14182 @subsection package IDE
14185 The attributes of package @code{IDE} specify the options to be used when using
14186 an Integrated Development Environment such as @command{GPS}.
14190 This is a simple attribute. Its value is a string that designates the remote
14191 host in a cross-compilation environment, to be used for remote compilation and
14192 debugging. This field should not be specified when running on the local
14196 This is a simple attribute. Its value is a string that specifies the
14197 name of IP address of the embedded target in a cross-compilation environment,
14198 on which the program should execute.
14200 @item Communication_Protocol
14201 This is a simple string attribute. Its value is the name of the protocol
14202 to use to communicate with the target in a cross-compilation environment,
14203 e.g. @code{"wtx"} or @code{"vxworks"}.
14205 @item Compiler_Command
14206 This is an associative array attribute, whose domain is a language name. Its
14207 value is string that denotes the command to be used to invoke the compiler.
14208 The value of @code{Compiler_Command ("Ada")} is expected to be compatible with
14209 gnatmake, in particular in the handling of switches.
14211 @item Debugger_Command
14212 This is simple attribute, Its value is a string that specifies the name of
14213 the debugger to be used, such as gdb, powerpc-wrs-vxworks-gdb or gdb-4.
14215 @item Default_Switches
14216 This is an associative array attribute. Its indexes are the name of the
14217 external tools that the GNAT Programming System (GPS) is supporting. Its
14218 value is a list of switches to use when invoking that tool.
14221 This is a simple attribute. Its value is a string that specifies the name
14222 of the @command{gnatls} utility to be used to retrieve information about the
14223 predefined path; e.g., @code{"gnatls"}, @code{"powerpc-wrs-vxworks-gnatls"}.
14226 This is a simple atribute. Is value is a string used to specify the
14227 Version Control System (VCS) to be used for this project, e.g CVS, RCS
14228 ClearCase or Perforce.
14230 @item VCS_File_Check
14231 This is a simple attribute. Its value is a string that specifies the
14232 command used by the VCS to check the validity of a file, either
14233 when the user explicitly asks for a check, or as a sanity check before
14234 doing the check-in.
14236 @item VCS_Log_Check
14237 This is a simple attribute. Its value is a string that specifies
14238 the command used by the VCS to check the validity of a log file.
14242 @node Package Renamings
14243 @section Package Renamings
14246 A package can be defined by a renaming declaration. The new package renames
14247 a package declared in a different project file, and has the same attributes
14248 as the package it renames.
14251 package_renaming ::==
14252 @b{package} package_identifier @b{renames}
14253 <project_>simple_name.package_identifier ;
14257 The package_identifier of the renamed package must be the same as the
14258 package_identifier. The project whose name is the prefix of the renamed
14259 package must contain a package declaration with this name. This project
14260 must appear in the context_clause of the enclosing project declaration,
14261 or be the parent project of the enclosing child project.
14267 A project file specifies a set of rules for constructing a software system.
14268 A project file can be self-contained, or depend on other project files.
14269 Dependencies are expressed through a context clause that names other projects.
14275 context_clause project_declaration
14277 project_declaration ::=
14278 simple_project_declaration | project_extension
14280 simple_project_declaration ::=
14281 @b{project} <project_>simple_name @b{is}
14282 @{declarative_item@}
14283 @b{end} <project_>simple_name;
14289 [@b{limited}] @b{with} path_name @{ , path_name @} ;
14296 A path name denotes a project file. A path name can be absolute or relative.
14297 An absolute path name includes a sequence of directories, in the syntax of
14298 the host operating system, that identifies uniquely the project file in the
14299 file system. A relative path name identifies the project file, relative
14300 to the directory that contains the current project, or relative to a
14301 directory listed in the environment variable ADA_PROJECT_PATH.
14302 Path names are case sensitive if file names in the host operating system
14303 are case sensitive.
14305 The syntax of the environment variable ADA_PROJECT_PATH is a list of
14306 directory names separated by colons (semicolons on Windows).
14308 A given project name can appear only once in a context_clause.
14310 It is illegal for a project imported by a context clause to refer, directly
14311 or indirectly, to the project in which this context clause appears (the
14312 dependency graph cannot contain cycles), except when one of the with_clause
14313 in the cycle is a @code{limited with}.
14315 @node Project Extensions
14316 @section Project Extensions
14319 A project extension introduces a new project, which inherits the declarations
14320 of another project.
14324 project_extension ::=
14325 @b{project} <project_>simple_name @b{extends} path_name @b{is}
14326 @{declarative_item@}
14327 @b{end} <project_>simple_name;
14331 The project extension declares a child project. The child project inherits
14332 all the declarations and all the files of the parent project, These inherited
14333 declaration can be overridden in the child project, by means of suitable
14336 @node Project File Elaboration
14337 @section Project File Elaboration
14340 A project file is processed as part of the invocation of a gnat tool that
14341 uses the project option. Elaboration of the process file consists in the
14342 sequential elaboration of all its declarations. The computed values of
14343 attributes and variables in the project are then used to establish the
14344 environment in which the gnat tool will execute.
14346 @node Obsolescent Features
14347 @chapter Obsolescent Features
14350 This chapter describes features that are provided by GNAT, but are
14351 considered obsolescent since there are preferred ways of achieving
14352 the same effect. These features are provided solely for historical
14353 compatibility purposes.
14356 * pragma No_Run_Time::
14357 * pragma Ravenscar::
14358 * pragma Restricted_Run_Time::
14361 @node pragma No_Run_Time
14362 @section pragma No_Run_Time
14364 The pragma @code{No_Run_Time} is used to achieve an affect similar
14365 to the use of the "Zero Foot Print" configurable run time, but without
14366 requiring a specially configured run time. The result of using this
14367 pragma, which must be used for all units in a partition, is to restrict
14368 the use of any language features requiring run-time support code. The
14369 preferred usage is to use an appropriately configured run-time that
14370 includes just those features that are to be made accessible.
14372 @node pragma Ravenscar
14373 @section pragma Ravenscar
14375 The pragma @code{Ravenscar} has exactly the same effect as pragma
14376 @code{Profile (Ravenscar)}. The latter usage is preferred since it
14377 is part of the new Ada 2005 standard.
14379 @node pragma Restricted_Run_Time
14380 @section pragma Restricted_Run_Time
14382 The pragma @code{Restricted_Run_Time} has exactly the same effect as
14383 pragma @code{Profile (Restricted)}. The latter usage is
14384 preferred since the Ada 2005 pragma @code{Profile} is intended for
14385 this kind of implementation dependent addition.
14388 @c GNU Free Documentation License
14390 @node Index,,GNU Free Documentation License, Top