1 \input texinfo @c -*-texinfo-*-
5 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
7 @c GNAT DOCUMENTATION o
13 @c Copyright (C) 1995-2002 Free Software Foundation o
16 @c GNAT is maintained by Ada Core Technologies Inc (http://www.gnat.com). o
18 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
20 @setfilename gnat_rm.info
21 @settitle GNAT Reference Manual
22 @setchapternewpage odd
25 @dircategory GNU Ada tools
27 * GNAT Reference Manual: (gnat_rm). Reference Manual for GNU Ada tools.
32 @title GNAT Reference Manual
33 @subtitle GNAT, The GNU Ada 95 Compiler
38 @subtitle Version 3.16w
40 @author Ada Core Technologies, Inc.
43 @vskip 0pt plus 1filll
46 Copyright @copyright{} 1995-2001, Free Software Foundation
48 Permission is granted to copy, distribute and/or modify this document
49 under the terms of the GNU Free Documentation License, Version 1.1
50 or any later version published by the Free Software Foundation;
51 with the Invariant Sections being ``GNU Free Documentation License'', with the
52 Front-Cover Texts being ``GNAT Reference Manual'', and with no Back-Cover Texts.
53 A copy of the license is included in the section entitled ``GNU
54 Free Documentation License''.
59 @node Top, About This Guide, (dir), (dir)
60 @top GNAT Reference Manual
65 GNAT, The GNU Ada 95 Compiler
74 Ada Core Technologies, Inc.
77 Copyright @copyright{} 1995-2001, Free Software Foundation
79 Permission is granted to copy, distribute and/or modify this document
80 under the terms of the GNU Free Documentation License, Version 1.1
81 or any later version published by the Free Software Foundation;
82 with the Invariant Sections being ``GNU Free Documentation License'', with the
83 Front-Cover Texts being ``GNAT Reference Manual'', and with no Back-Cover Texts.
84 A copy of the license is included in the section entitled ``GNU
85 Free Documentation License''.
90 * Implementation Defined Pragmas::
91 * Implementation Defined Attributes::
92 * Implementation Advice::
93 * Implementation Defined Characteristics::
94 * Intrinsic Subprograms::
95 * Representation Clauses and Pragmas::
96 * Standard Library Routines::
97 * The Implementation of Standard I/O::
99 * Interfacing to Other Languages::
100 * Machine Code Insertions::
101 * GNAT Implementation of Tasking::
102 * Code generation for array aggregates::
103 * Specialized Needs Annexes::
104 * Compatibility Guide::
105 * GNU Free Documentation License::
108 --- The Detailed Node Listing ---
112 * What This Reference Manual Contains::
113 * Related Information::
115 The Implementation of Standard I/O
117 * Standard I/O Packages::
126 * Operations on C Streams::
127 * Interfacing to C Streams::
131 * Ada.Characters.Latin_9 (a-chlat9.ads)::
132 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
133 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
134 * Ada.Command_Line.Remove (a-colire.ads)::
135 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
136 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
137 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
138 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
139 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
140 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
141 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
142 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
143 * GNAT.AWK (g-awk.ads)::
144 * GNAT.Bubble_Sort_A (g-busora.ads)::
145 * GNAT.Bubble_Sort_G (g-busorg.ads)::
146 * GNAT.Calendar (g-calend.ads)::
147 * GNAT.Calendar.Time_IO (g-catiio.ads)::
148 * GNAT.Case_Util (g-casuti.ads)::
149 * GNAT.CGI (g-cgi.ads)::
150 * GNAT.CGI.Cookie (g-cgicoo.ads)::
151 * GNAT.CGI.Debug (g-cgideb.ads)::
152 * GNAT.Command_Line (g-comlin.ads)::
153 * GNAT.CRC32 (g-crc32.ads)::
154 * GNAT.Current_Exception (g-curexc.ads)::
155 * GNAT.Debug_Pools (g-debpoo.ads)::
156 * GNAT.Debug_Utilities (g-debuti.ads)::
157 * GNAT.Directory_Operations (g-dirope.ads)::
158 * GNAT.Dynamic_Tables (g-dyntab.ads)::
159 * GNAT.Exception_Traces (g-exctra.ads)::
160 * GNAT.Expect (g-expect.ads)::
161 * GNAT.Float_Control (g-flocon.ads)::
162 * GNAT.Heap_Sort_A (g-hesora.ads)::
163 * GNAT.Heap_Sort_G (g-hesorg.ads)::
164 * GNAT.HTable (g-htable.ads)::
165 * GNAT.IO (g-io.ads)::
166 * GNAT.IO_Aux (g-io_aux.ads)::
167 * GNAT.Lock_Files (g-locfil.ads)::
168 * GNAT.MD5 (g-md5.ads)::
169 * GNAT.Most_Recent_Exception (g-moreex.ads)::
170 * GNAT.OS_Lib (g-os_lib.ads)::
171 * GNAT.Regexp (g-regexp.ads)::
172 * GNAT.Registry (g-regist.ads)::
173 * GNAT.Regpat (g-regpat.ads)::
174 * GNAT.Sockets (g-socket.ads)::
175 * GNAT.Source_Info (g-souinf.ads)::
176 * GNAT.Spell_Checker (g-speche.ads)::
177 * GNAT.Spitbol.Patterns (g-spipat.ads)::
178 * GNAT.Spitbol (g-spitbo.ads)::
179 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
180 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
181 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
182 * GNAT.Table (g-table.ads)::
183 * GNAT.Task_Lock (g-tasloc.ads)::
184 * GNAT.Threads (g-thread.ads)::
185 * GNAT.Traceback (g-traceb.ads)::
186 * GNAT.Traceback.Symbolic (g-trasym.ads)::
187 * Interfaces.C.Extensions (i-cexten.ads)::
188 * Interfaces.C.Streams (i-cstrea.ads)::
189 * Interfaces.CPP (i-cpp.ads)::
190 * Interfaces.Os2lib (i-os2lib.ads)::
191 * Interfaces.Os2lib.Errors (i-os2err.ads)::
192 * Interfaces.Os2lib.Synchronization (i-os2syn.ads)::
193 * Interfaces.Os2lib.Threads (i-os2thr.ads)::
194 * Interfaces.Packed_Decimal (i-pacdec.ads)::
195 * Interfaces.VxWorks (i-vxwork.ads)::
196 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
197 * System.Address_Image (s-addima.ads)::
198 * System.Assertions (s-assert.ads)::
199 * System.Partition_Interface (s-parint.ads)::
200 * System.Task_Info (s-tasinf.ads)::
201 * System.Wch_Cnv (s-wchcnv.ads)::
202 * System.Wch_Con (s-wchcon.ads)::
206 * Text_IO Stream Pointer Positioning::
207 * Text_IO Reading and Writing Non-Regular Files::
209 * Treating Text_IO Files as Streams::
210 * Text_IO Extensions::
211 * Text_IO Facilities for Unbounded Strings::
215 * Wide_Text_IO Stream Pointer Positioning::
216 * Wide_Text_IO Reading and Writing Non-Regular Files::
218 Interfacing to Other Languages
221 * Interfacing to C++::
222 * Interfacing to COBOL::
223 * Interfacing to Fortran::
224 * Interfacing to non-GNAT Ada code::
226 GNAT Implementation of Tasking
228 * Mapping Ada Tasks onto the Underlying Kernel Threads::
229 * Ensuring Compliance with the Real-Time Annex::
234 @node About This Guide
235 @unnumbered About This Guide
238 This manual contains useful information in writing programs using the
239 GNAT compiler. It includes information on implementation dependent
240 characteristics of GNAT, including all the information required by Annex
243 Ada 95 is designed to be highly portable,and guarantees that, for most
244 programs, Ada 95 compilers behave in exactly the same manner on
245 different machines. However, since Ada 95 is designed to be used in a
246 wide variety of applications, it also contains a number of system
247 dependent features to Functbe used in interfacing to the external world.
249 @c Maybe put the following in platform-specific section
252 This reference manual discusses how these features are implemented for
253 use in ProDev Ada running on the IRIX 5.3 or greater operating systems.
256 @cindex Implementation-dependent features
258 Note: Any program that makes use of implementation-dependent features
259 may be non-portable. You should follow good programming practice and
260 isolate and clearly document any sections of your program that make use
261 of these features in a non-portable manner.
264 * What This Reference Manual Contains::
266 * Related Information::
269 @node What This Reference Manual Contains
270 @unnumberedsec What This Reference Manual Contains
272 This reference manual contains the following chapters:
276 @ref{Implementation Defined Pragmas} lists GNAT implementation-dependent
277 pragmas, which can be used to extend and enhance the functionality of the
281 @ref{Implementation Defined Attributes} lists GNAT
282 implementation-dependent attributes which can be used to extend and
283 enhance the functionality of the compiler.
286 @ref{Implementation Advice} provides information on generally
287 desirable behavior which are not requirements that all compilers must
288 follow since it cannot be provided on all systems, or which may be
289 undesirable on some systems.
292 @ref{Implementation Defined Characteristics} provides a guide to
293 minimizing implementation dependent features.
296 @ref{Intrinsic Subprograms} describes the intrinsic subprograms
297 implemented by GNAT, and how they can be imported into user
298 application programs.
301 @ref{Representation Clauses and Pragmas} describes in detail the
302 way that GNAT represents data, and in particular the exact set
303 of representation clauses and pragmas that is accepted.
306 @ref{Standard Library Routines} provides a listing of packages and a
307 brief description of the functionality that is provided by Ada's
308 extensive set of standard library routines as implemented by GNAT@.
311 @ref{The Implementation of Standard I/O} details how the GNAT
312 implementation of the input-output facilities.
315 @ref{Interfacing to Other Languages} describes how programs
316 written in Ada using GNAT can be interfaced to other programming
320 @ref{Specialized Needs Annexes} describes the GNAT implementation of all
321 of the special needs annexes.
324 @ref{Compatibility Guide} includes sections on compatibility of GNAT with
325 other Ada 83 and Ada 95 compilation systems, to assist in porting code
326 from other environments.
329 @cindex Ada 95 ISO/ANSI Standard
330 This reference manual assumes that you are familiar with Ada 95
331 language, as described in the International Standard
332 ANSI/ISO/IEC-8652:1995, Jan 1995.
335 @unnumberedsec Conventions
336 @cindex Conventions, typographical
337 @cindex Typographical conventions
340 Following are examples of the typographical and graphic conventions used
345 @code{Functions}, @code{utility program names}, @code{standard names},
352 @file{File Names}, @samp{button names}, and @samp{field names}.
361 [optional information or parameters]
364 Examples are described by text
366 and then shown this way.
371 Commands that are entered by the user are preceded in this manual by the
372 characters @samp{$ } (dollar sign followed by space). If your system uses this
373 sequence as a prompt, then the commands will appear exactly as you see them
374 in the manual. If your system uses some other prompt, then the command will
375 appear with the @samp{$} replaced by whatever prompt character you are using.
377 @node Related Information
378 @unnumberedsec Related Information
379 See the following documents for further information on GNAT:
383 @cite{GNAT User's Guide}, which provides information on how to use
384 the GNAT compiler system.
387 @cite{Ada 95 Reference Manual}, which contains all reference
388 material for the Ada 95 programming language.
391 @cite{Ada 95 Annotated Reference Manual}, which is an annotated version
392 of the standard reference manual cited above. The annotations describe
393 detailed aspects of the design decision, and in particular contain useful
394 sections on Ada 83 compatibility.
397 @cite{DEC Ada, Technical Overview and Comparison on DIGITAL Platforms},
398 which contains specific information on compatibility between GNAT and
402 @cite{DEC Ada, Language Reference Manual, part number AA-PYZAB-TK} which
403 describes in detail the pragmas and attributes provided by the DEC Ada 83
408 @node Implementation Defined Pragmas
409 @chapter Implementation Defined Pragmas
412 Ada 95 defines a set of pragmas that can be used to supply additional
413 information to the compiler. These language defined pragmas are
414 implemented in GNAT and work as described in the Ada 95 Reference
417 In addition, Ada 95 allows implementations to define additional pragmas
418 whose meaning is defined by the implementation. GNAT provides a number
419 of these implementation-dependent pragmas which can be used to extend
420 and enhance the functionality of the compiler. This section of the GNAT
421 Reference Manual describes these additional pragmas.
423 Note that any program using these pragmas may not be portable to other
424 compilers (although GNAT implements this set of pragmas on all
425 platforms). Therefore if portability to other compilers is an important
426 consideration, the use of these pragmas should be minimized.
431 @cindex Deferring aborts
432 @item pragma Abort_Defer
441 This pragma must appear at the start of the statement sequence of a
442 handled sequence of statements (right after the @code{begin}). It has
443 the effect of deferring aborts for the sequence of statements (but not
444 for the declarations or handlers, if any, associated with this statement
457 A configuration pragma that establishes Ada 83 mode for the unit to
458 which it applies, regardless of the mode set by the command line
459 switches. In Ada 83 mode, GNAT attempts to be as compatible with
460 the syntax and semantics of Ada 83, as defined in the original Ada
461 83 Reference Manual as possible. In particular, the new Ada 95
462 keywords are not recognized, optional package bodies are allowed,
463 and generics may name types with unknown discriminants without using
464 the @code{(<>)} notation. In addition, some but not all of the additional
465 restrictions of Ada 83 are enforced.
467 Ada 83 mode is intended for two purposes. Firstly, it allows existing
468 legacy Ada 83 code to be compiled and adapted to GNAT with less effort.
469 Secondly, it aids in keeping code backwards compatible with Ada 83.
470 However, there is no guarantee that code that is processed correctly
471 by GNAT in Ada 83 mode will in fact compile and execute with an Ada
472 83 compiler, since GNAT does not enforce all the additional checks
485 A configuration pragma that establishes Ada 95 mode for the unit to which
486 it applies, regardless of the mode set by the command line switches.
487 This mode is set automatically for the @code{Ada} and @code{System}
488 packages and their children, so you need not specify it in these
489 contexts. This pragma is useful when writing a reusable component that
490 itself uses Ada 95 features, but which is intended to be usable from
491 either Ada 83 or Ada 95 programs.
494 @item pragma Annotate
499 pragma Annotate (IDENTIFIER @{, ARG@});
501 ARG ::= NAME | EXPRESSION
505 This pragma is used to annotate programs. @var{identifier} identifies
506 the type of annotation. GNAT verifies this is an identifier, but does
507 not otherwise analyze it. The @var{arg} argument
508 can be either a string literal or an
509 expression. String literals are assumed to be of type
510 @code{Standard.String}. Names of entities are simply analyzed as entity
511 names. All other expressions are analyzed as expressions, and must be
514 The analyzed pragma is retained in the tree, but not otherwise processed
515 by any part of the GNAT compiler. This pragma is intended for use by
516 external tools, including ASIS@.
526 [, static_string_EXPRESSION])
530 The effect of this pragma depends on whether the corresponding command
531 line switch is set to activate assertions. The pragma expands into code
532 equivalent to the following:
535 if assertions-enabled then
536 if not boolean_EXPRESSION then
537 System.Assertions.Raise_Assert_Failure
544 The string argument, if given, is the message that will be associated
545 with the exception occurrence if the exception is raised. If no second
546 argument is given, the default message is @samp{@var{file}:@var{nnn}},
547 where @var{file} is the name of the source file containing the assert,
548 and @var{nnn} is the line number of the assert. A pragma is not a
549 statement, so if a statement sequence contains nothing but a pragma
550 assert, then a null statement is required in addition, as in:
555 pragma Assert (K > 3, "Bad value for K");
561 Note that, as with the @code{if} statement to which it is equivalent, the
562 type of the expression is either @code{Standard.Boolean}, or any type derived
563 from this standard type.
565 If assertions are disabled (switch @code{-gnata} not used), then there
566 is no effect (and in particular, any side effects from the expression
567 are suppressed). More precisely it is not quite true that the pragma
568 has no effect, since the expression is analyzed, and may cause types
569 to be frozen if they are mentioned here for the first time.
571 If assertions are enabled, then the given expression is tested, and if
572 it is @code{False} then @code{System.Assertions.Raise_Assert_Failure} is called
573 which results in the raising of @code{Assert_Failure} with the given message.
575 If the boolean expression has side effects, these side effects will turn
576 on and off with the setting of the assertions mode, resulting in
577 assertions that have an effect on the program. You should generally
578 avoid side effects in the expression arguments of this pragma. However,
579 the expressions are analyzed for semantic correctness whether or not
580 assertions are enabled, so turning assertions on and off cannot affect
581 the legality of a program.
585 @item pragma Ast_Entry
590 pragma AST_Entry (entry_IDENTIFIER);
594 This pragma is implemented only in the OpenVMS implementation of GNAT@. The
595 argument is the simple name of a single entry; at most one @code{AST_Entry}
596 pragma is allowed for any given entry. This pragma must be used in
597 conjunction with the @code{AST_Entry} attribute, and is only allowed after
598 the entry declaration and in the same task type specification or single task
599 as the entry to which it applies. This pragma specifies that the given entry
600 may be used to handle an OpenVMS asynchronous system trap (@code{AST})
601 resulting from an OpenVMS system service call. The pragma does not affect
602 normal use of the entry. For further details on this pragma, see the
603 DEC Ada Language Reference Manual, section 9.12a.
605 @cindex Passing by copy
606 @findex C_Pass_By_Copy
607 @item pragma C_Pass_By_Copy
612 pragma C_Pass_By_Copy
613 ([Max_Size =>] static_integer_EXPRESSION);
617 Normally the default mechanism for passing C convention records to C
618 convention subprograms is to pass them by reference, as suggested by RM
619 B.3(69). Use the configuration pragma @code{C_Pass_By_Copy} to change
620 this default, by requiring that record formal parameters be passed by
621 copy if all of the following conditions are met:
625 The size of the record type does not exceed@*@var{static_integer_expression}.
627 The record type has @code{Convention C}.
629 The formal parameter has this record type, and the subprogram has a
630 foreign (non-Ada) convention.
634 If these conditions are met the argument is passed by copy, i.e.@: in a
635 manner consistent with what C expects if the corresponding formal in the
636 C prototype is a struct (rather than a pointer to a struct).
638 You can also pass records by copy by specifying the convention
639 @code{C_Pass_By_Copy} for the record type, or by using the extended
640 @code{Import} and @code{Export} pragmas, which allow specification of
641 passing mechanisms on a parameter by parameter basis.
649 pragma Comment (static_string_EXPRESSION);
653 This is almost identical in effect to pragma @code{Ident}. It allows the
654 placement of a comment into the object file and hence into the
655 executable file if the operating system permits such usage. The
656 difference is that @code{Comment}, unlike @code{Ident}, has no limit on the
657 length of the string argument, and no limitations on placement
658 of the pragma (it can be placed anywhere in the main source unit).
660 @findex Common_Object
661 @item pragma Common_Object
666 pragma Common_Object (
667 [Internal =>] LOCAL_NAME,
668 [, [External =>] EXTERNAL_SYMBOL]
669 [, [Size =>] EXTERNAL_SYMBOL] )
673 | static_string_EXPRESSION
677 This pragma enables the shared use of variables stored in overlaid
678 linker areas corresponding to the use of @code{COMMON}
679 in Fortran. The single
680 object @var{local_name} is assigned to the area designated by
681 the @var{External} argument.
682 You may define a record to correspond to a series
683 of fields. The @var{size} argument
684 is syntax checked in GNAT, but otherwise ignored.
686 @code{Common_Object} is not supported on all platforms. If no
687 support is available, then the code generator will issue a message
688 indicating that the necessary attribute for implementation of this
689 pragma is not available.
691 @findex Complex_Representation
692 @item pragma Complex_Representation
697 pragma Complex_Representation
698 ([Entity =>] LOCAL_NAME);
702 The @var{Entity} argument must be the name of a record type which has
703 two fields of the same floating-point type. The effect of this pragma is
704 to force gcc to use the special internal complex representation form for
705 this record, which may be more efficient. Note that this may result in
706 the code for this type not conforming to standard ABI (application
707 binary interface) requirements for the handling of record types. For
708 example, in some environments, there is a requirement for passing
709 records by pointer, and the use of this pragma may result in passing
710 this type in floating-point registers.
712 @cindex Alignments of components
713 @findex Component_Alignment
714 @item pragma Component_Alignment
719 pragma Component_Alignment (
720 [Form =>] ALIGNMENT_CHOICE
721 [, [Name =>] type_LOCAL_NAME]);
731 Specifies the alignment of components in array or record types.
732 The meaning of the @var{Form} argument is as follows:
735 @findex Component_Size
737 Aligns scalar components and subcomponents of the array or record type
738 on boundaries appropriate to their inherent size (naturally
739 aligned). For example, 1-byte components are aligned on byte boundaries,
740 2-byte integer components are aligned on 2-byte boundaries, 4-byte
741 integer components are aligned on 4-byte boundaries and so on. These
742 alignment rules correspond to the normal rules for C compilers on all
743 machines except the VAX@.
745 @findex Component_Size_4
746 @item Component_Size_4
747 Naturally aligns components with a size of four or fewer
748 bytes. Components that are larger than 4 bytes are placed on the next
753 Specifies that array or record components are byte aligned, i.e.@:
754 aligned on boundaries determined by the value of the constant
755 @code{System.Storage_Unit}.
759 Specifies that array or record components are aligned on default
760 boundaries, appropriate to the underlying hardware or operating system or
761 both. For OpenVMS VAX systems, the @code{Default} choice is the same as
762 the @code{Storage_Unit} choice (byte alignment). For all other systems,
763 the @code{Default} choice is the same as @code{Component_Size} (natural
767 If the @code{Name} parameter is present, @var{type_local_name} must
768 refer to a local record or array type, and the specified alignment
769 choice applies to the specified type. The use of
770 @code{Component_Alignment} together with a pragma @code{Pack} causes the
771 @code{Component_Alignment} pragma to be ignored. The use of
772 @code{Component_Alignment} together with a record representation clause
773 is only effective for fields not specified by the representation clause.
775 If the @code{Name} parameter is absent, the pragma can be used as either
776 a configuration pragma, in which case it applies to one or more units in
777 accordance with the normal rules for configuration pragmas, or it can be
778 used within a declarative part, in which case it applies to types that
779 are declared within this declarative part, or within any nested scope
780 within this declarative part. In either case it specifies the alignment
781 to be applied to any record or array type which has otherwise standard
784 If the alignment for a record or array type is not specified (using
785 pragma @code{Pack}, pragma @code{Component_Alignment}, or a record rep
786 clause), the GNAT uses the default alignment as described previously.
788 @findex Convention_Identifier
789 @cindex Conventions, synonyms
790 @item pragma Convention_Identifier
795 pragma Convention_Identifier (
796 [Name =>] IDENTIFIER,
797 [Convention =>] convention_IDENTIFIER);
801 This pragma provides a mechanism for supplying synonyms for existing
802 convention identifiers. The @code{Name} identifier can subsequently
803 be used as a synonym for the given convention in other pragmas (including
804 for example pragma @code{Import} or another @code{Convention_Identifier}
805 pragma). As an example of the use of this, suppose you had legacy code
806 which used Fortran77 as the identifier for Fortran. Then the pragma:
809 pragma Convention_Indentifier (Fortran77, Fortran);
813 would allow the use of the convention identifier @code{Fortran77} in
814 subsequent code, avoiding the need to modify the sources. As another
815 example, you could use this to parametrize convention requirements
816 according to systems. Suppose you needed to use @code{Stdcall} on
817 windows systems, and @code{C} on some other system, then you could
818 define a convention identifier @code{Library} and use a single
819 @code{Convention_Identifier} pragma to specify which convention
820 would be used system-wide.
823 @cindex Interfacing with C++
824 @item pragma CPP_Class
829 pragma CPP_Class ([Entity =>] LOCAL_NAME);
833 The argument denotes an entity in the current declarative region
834 that is declared as a tagged or untagged record type. It indicates that
835 the type corresponds to an externally declared C++ class type, and is to
836 be laid out the same way that C++ would lay out the type.
838 If (and only if) the type is tagged, at least one component in the
839 record must be of type @code{Interfaces.CPP.Vtable_Ptr}, corresponding
840 to the C++ Vtable (or Vtables in the case of multiple inheritance) used
843 Types for which @code{CPP_Class} is specified do not have assignment or
844 equality operators defined (such operations can be imported or declared
845 as subprograms as required). Initialization is allowed only by
846 constructor functions (see pragma @code{CPP_Constructor}).
848 Pragma @code{CPP_Class} is intended primarily for automatic generation
849 using an automatic binding generator tool.
850 See @ref{Interfacing to C++} for related information.
852 @cindex Interfacing with C++
853 @findex CPP_Constructor
854 @item pragma CPP_Constructor
859 pragma CPP_Constructor ([Entity =>] LOCAL_NAME);
863 This pragma identifies an imported function (imported in the usual way
864 with pragma @code{Import}) as corresponding to a C++
865 constructor. The argument is a name that must have been
866 previously mentioned in a pragma @code{Import}
867 with @code{Convention} = @code{CPP}, and must be of one of the following
872 @code{function @var{Fname} return @var{T}'Class}
875 @code{function @var{Fname} (@dots{}) return @var{T}'Class}
879 where @var{T} is a tagged type to which the pragma @code{CPP_Class} applies.
881 The first form is the default constructor, used when an object of type
882 @var{T} is created on the Ada side with no explicit constructor. Other
883 constructors (including the copy constructor, which is simply a special
884 case of the second form in which the one and only argument is of type
885 @var{T}), can only appear in two contexts:
889 On the right side of an initialization of an object of type @var{T}.
891 In an extension aggregate for an object of a type derived from @var{T}.
894 Although the constructor is described as a function that returns a value
895 on the Ada side, it is typically a procedure with an extra implicit
896 argument (the object being initialized) at the implementation
897 level. GNAT issues the appropriate call, whatever it is, to get the
898 object properly initialized.
900 In the case of derived objects, you may use one of two possible forms
901 for declaring and creating an object:
904 @item @code{New_Object : Derived_T}
905 @item @code{New_Object : Derived_T := (@var{constructor-function-call with} @dots{})}
908 In the first case the default constructor is called and extension fields
909 if any are initialized according to the default initialization
910 expressions in the Ada declaration. In the second case, the given
911 constructor is called and the extension aggregate indicates the explicit
912 values of the extension fields.
914 If no constructors are imported, it is impossible to create any objects
915 on the Ada side. If no default constructor is imported, only the
916 initialization forms using an explicit call to a constructor are
919 Pragma @code{CPP_Constructor} is intended primarily for automatic generation
920 using an automatic binding generator tool.
921 See @ref{Interfacing to C++} for more related information.
923 @cindex Interfacing to C++
925 @item pragma CPP_Virtual
932 [, [Vtable_Ptr =>] vtable_ENTITY,]
933 [, [Position =>] static_integer_EXPRESSION])
936 This pragma serves the same function as pragma @code{Import} in that
937 case of a virtual function imported from C++. The @var{Entity} argument
939 primitive subprogram of a tagged type to which pragma @code{CPP_Class}
940 applies. The @var{Vtable_Ptr} argument specifies
941 the Vtable_Ptr component which contains the
942 entry for this virtual function. The @var{Position} argument
943 is the sequential number
944 counting virtual functions for this Vtable starting at 1.
946 The @code{Vtable_Ptr} and @code{Position} arguments may be omitted if
947 there is one Vtable_Ptr present (single inheritance case) and all
948 virtual functions are imported. In that case the compiler can deduce both
951 No @code{External_Name} or @code{Link_Name} arguments are required for a
952 virtual function, since it is always accessed indirectly via the
953 appropriate Vtable entry.
955 Pragma @code{CPP_Virtual} is intended primarily for automatic generation
956 using an automatic binding generator tool.
957 See @ref{Interfacing to C++} for related information.
959 @cindex Interfacing with C++
961 @item pragma CPP_Vtable
968 [Vtable_Ptr =>] vtable_ENTITY,
969 [Entry_Count =>] static_integer_EXPRESSION);
973 Given a record to which the pragma @code{CPP_Class} applies,
974 this pragma can be specified for each component of type
975 @code{CPP.Interfaces.Vtable_Ptr}.
976 @var{Entity} is the tagged type, @var{Vtable_Ptr}
977 is the record field of type @code{Vtable_Ptr}, and @var{Entry_Count} is
978 the number of virtual functions on the C++ side. Not all of these
979 functions need to be imported on the Ada side.
981 You may omit the @code{CPP_Vtable} pragma if there is only one
982 @code{Vtable_Ptr} component in the record and all virtual functions are
983 imported on the Ada side (the default value for the entry count in this
984 case is simply the total number of virtual functions).
986 Pragma @code{CPP_Vtable} is intended primarily for automatic generation
987 using an automatic binding generator tool.
988 See @ref{Interfacing to C++} for related information.
996 pragma Debug (PROCEDURE_CALL_WITHOUT_SEMICOLON);
998 PROCEDURE_CALL_WITHOUT_SEMICOLON ::=
1000 | PROCEDURE_PREFIX ACTUAL_PARAMETER_PART
1004 The argument has the syntactic form of an expression, meeting the
1005 syntactic requirements for pragmas.
1007 If assertions are not enabled on the command line, this pragma has no
1008 effect. If asserts are enabled, the semantics of the pragma is exactly
1009 equivalent to the procedure call statement corresponding to the argument
1010 with a terminating semicolon. Pragmas are permitted in sequences of
1011 declarations, so you can use pragma @code{Debug} to intersperse calls to
1012 debug procedures in the middle of declarations.
1014 @cindex Elaboration control
1015 @findex Elaboration_Checks
1016 @item pragma Elaboration_Checks
1021 pragma Elaboration_Checks (RM | Static);
1025 This is a configuration pragma that provides control over the
1026 elaboration model used by the compilation affected by the
1027 pragma. If the parameter is RM, then the dynamic elaboration
1028 model described in the Ada Reference Manual is used, as though
1029 the @code{-gnatE} switch had been specified on the command
1030 line. If the parameter is Static, then the default GNAT static
1031 model is used. This configuration pragma overrides the setting
1032 of the command line. For full details on the elaboration models
1033 used by the GNAT compiler, see section ``Elaboration Order
1034 Handling in GNAT'' in the @cite{GNAT User's Guide}.
1036 @cindex Elimination of unused subprograms
1038 @item pragma Eliminate
1044 [Unit_Name =>] IDENTIFIER |
1045 SELECTED_COMPONENT);
1048 [Unit_Name =>] IDENTIFIER |
1050 [Entity =>] IDENTIFIER |
1051 SELECTED_COMPONENT |
1053 [,[Parameter_Types =>] PARAMETER_TYPES]
1054 [,[Result_Type =>] result_SUBTYPE_NAME]
1055 [,[Homonym_Number =>] INTEGER_LITERAL]);
1057 PARAMETER_TYPES ::= (SUBTYPE_NAME @{, SUBTYPE_NAME@})
1058 SUBTYPE_NAME ::= STRING_LITERAL
1062 This pragma indicates that the given entity is not used outside the
1063 compilation unit it is defined in. The entity may be either a subprogram
1066 If the entity to be eliminated is a library level subprogram, then
1067 the first form of pragma @code{Eliminate} is used with only a single argument.
1068 In this form, the @code{Unit_Name} argument specifies the name of the
1069 library level unit to be eliminated.
1071 In all other cases, both @code{Unit_Name} and @code{Entity} arguments
1072 are required. item is an entity of a library package, then the first
1073 argument specifies the unit name, and the second argument specifies
1074 the particular entity. If the second argument is in string form, it must
1075 correspond to the internal manner in which GNAT stores entity names (see
1076 compilation unit Namet in the compiler sources for details).
1078 The remaining parameters are optionally used to distinguish
1079 between overloaded subprograms. There are two ways of doing this.
1081 Use @code{Parameter_Types} and @code{Result_Type} to specify the
1082 profile of the subprogram to be eliminated in a manner similar to that
1084 the extended @code{Import} and @code{Export} pragmas, except that the
1085 subtype names are always given as string literals, again corresponding
1086 to the internal manner in which GNAT stores entity names.
1088 Alternatively, the @code{Homonym_Number} parameter is used to specify
1089 which overloaded alternative is to be eliminated. A value of 1 indicates
1090 the first subprogram (in lexical order), 2 indicates the second etc.
1092 The effect of the pragma is to allow the compiler to eliminate
1093 the code or data associated with the named entity. Any reference to
1094 an eliminated entity outside the compilation unit it is defined in,
1095 causes a compile time or link time error.
1097 The parameters of this pragma may be given in any order, as long as
1098 the usual rules for use of named parameters and position parameters
1101 The intention of pragma @code{Eliminate} is to allow a program to be compiled
1102 in a system independent manner, with unused entities eliminated, without
1103 the requirement of modifying the source text. Normally the required set
1104 of @code{Eliminate} pragmas is constructed automatically using the gnatelim tool.
1105 Elimination of unused entities local to a compilation unit is automatic,
1106 without requiring the use of pragma @code{Eliminate}.
1108 Note that the reason this pragma takes string literals where names might
1109 be expected is that a pragma @code{Eliminate} can appear in a context where the
1110 relevant names are not visible.
1113 @findex Export_Exception
1114 @item pragma Export_Exception
1119 pragma Export_Exception (
1120 [Internal =>] LOCAL_NAME,
1121 [, [External =>] EXTERNAL_SYMBOL,]
1122 [, [Form =>] Ada | VMS]
1123 [, [Code =>] static_integer_EXPRESSION]);
1127 | static_string_EXPRESSION
1131 This pragma is implemented only in the OpenVMS implementation of GNAT@. It
1132 causes the specified exception to be propagated outside of the Ada program,
1133 so that it can be handled by programs written in other OpenVMS languages.
1134 This pragma establishes an external name for an Ada exception and makes the
1135 name available to the OpenVMS Linker as a global symbol. For further details
1136 on this pragma, see the
1137 DEC Ada Language Reference Manual, section 13.9a3.2.
1139 @cindex Argument passing mechanisms
1140 @findex Export_Function
1141 @item pragma Export_Function @dots{}
1147 pragma Export_Function (
1148 [Internal =>] LOCAL_NAME,
1149 [, [External =>] EXTERNAL_SYMBOL]
1150 [, [Parameter_Types =>] PARAMETER_TYPES]
1151 [, [Result_Type =>] result_SUBTYPE_MARK]
1152 [, [Mechanism =>] MECHANISM]
1153 [, [Result_Mechanism =>] MECHANISM_NAME]);
1157 | static_string_EXPRESSION
1161 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1165 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1167 MECHANISM_ASSOCIATION ::=
1168 [formal_parameter_NAME =>] MECHANISM_NAME
1173 | Descriptor [([Class =>] CLASS_NAME)]
1175 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1178 Use this pragma to make a function externally callable and optionally
1179 provide information on mechanisms to be used for passing parameter and
1180 result values. We recommend, for the purposes of improving portability,
1181 this pragma always be used in conjunction with a separate pragma
1182 @code{Export}, which must precede the pragma @code{Export_Function}.
1183 GNAT does not require a separate pragma @code{Export}, but if none is
1184 present, @code{Convention Ada} is assumed, which is usually
1185 not what is wanted, so it is usually appropriate to use this
1186 pragma in conjunction with a @code{Export} or @code{Convention}
1187 pragma that specifies the desired foreign convention.
1188 Pragma @code{Export_Function}
1189 (and @code{Export}, if present) must appear in the same declarative
1190 region as the function to which they apply.
1192 @var{internal_name} must uniquely designate the function to which the
1193 pragma applies. If more than one function name exists of this name in
1194 the declarative part you must use the @code{Parameter_Types} and
1195 @code{Result_Type} parameters is mandatory to achieve the required
1196 unique designation. @var{subtype_ mark}s in these parameters must
1197 exactly match the subtypes in the corresponding function specification,
1198 using positional notation to match parameters with subtype marks.
1200 @cindex Passing by descriptor
1201 Passing by descriptor is supported only on the OpenVMS ports of GNAT@.
1203 @findex Export_Object
1204 @item pragma Export_Object @dots{}
1209 pragma Export_Object
1210 [Internal =>] LOCAL_NAME,
1211 [, [External =>] EXTERNAL_SYMBOL]
1212 [, [Size =>] EXTERNAL_SYMBOL]
1216 | static_string_EXPRESSION
1219 This pragma designates an object as exported, and apart from the
1220 extended rules for external symbols, is identical in effect to the use of
1221 the normal @code{Export} pragma applied to an object. You may use a
1222 separate Export pragma (and you probably should from the point of view
1223 of portability), but it is not required. @var{Size} is syntax checked,
1224 but otherwise ignored by GNAT@.
1226 @findex Export_Procedure
1227 @item pragma Export_Procedure @dots{}
1232 pragma Export_Procedure (
1233 [Internal =>] LOCAL_NAME
1234 [, [External =>] EXTERNAL_SYMBOL]
1235 [, [Parameter_Types =>] PARAMETER_TYPES]
1236 [, [Mechanism =>] MECHANISM]);
1240 | static_string_EXPRESSION
1244 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1248 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1250 MECHANISM_ASSOCIATION ::=
1251 [formal_parameter_NAME =>] MECHANISM_NAME
1256 | Descriptor [([Class =>] CLASS_NAME)]
1258 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1262 This pragma is identical to @code{Export_Function} except that it
1263 applies to a procedure rather than a function and the parameters
1264 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
1265 GNAT does not require a separate pragma @code{Export}, but if none is
1266 present, @code{Convention Ada} is assumed, which is usually
1267 not what is wanted, so it is usually appropriate to use this
1268 pragma in conjunction with a @code{Export} or @code{Convention}
1269 pragma that specifies the desired foreign convention.
1271 @findex Export_Valued_Procedure
1272 @item pragma Export_Valued_Procedure
1277 pragma Export_Valued_Procedure (
1278 [Internal =>] LOCAL_NAME
1279 [, [External =>] EXTERNAL_SYMBOL]
1280 [, [Parameter_Types =>] PARAMETER_TYPES]
1281 [, [Mechanism =>] MECHANISM]);
1285 | static_string_EXPRESSION
1289 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1293 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1295 MECHANISM_ASSOCIATION ::=
1296 [formal_parameter_NAME =>] MECHANISM_NAME
1301 | Descriptor [([Class =>] CLASS_NAME)]
1303 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1306 This pragma is identical to @code{Export_Procedure} except that the
1307 first parameter of @var{local_name}, which must be present, must be of
1308 mode @code{OUT}, and externally the subprogram is treated as a function
1309 with this parameter as the result of the function. GNAT provides for
1310 this capability to allow the use of @code{OUT} and @code{IN OUT}
1311 parameters in interfacing to external functions (which are not permitted
1313 GNAT does not require a separate pragma @code{Export}, but if none is
1314 present, @code{Convention Ada} is assumed, which is almost certainly
1315 not what is wanted since the whole point of this pragma is to interface
1316 with foreign language functions, so it is usually appropriate to use this
1317 pragma in conjunction with a @code{Export} or @code{Convention}
1318 pragma that specifies the desired foreign convention.
1320 @cindex @code{system}, extending
1322 @findex Extend_System
1323 @item pragma Extend_System
1328 pragma Extend_System ([Name =>] IDENTIFIER);
1332 This pragma is used to provide backwards compatibility with other
1333 implementations that extend the facilities of package @code{System}. In
1334 GNAT, @code{System} contains only the definitions that are present in
1335 the Ada 95 RM@. However, other implementations, notably the DEC Ada 83
1336 implementation, provide many extensions to package @code{System}.
1338 For each such implementation accommodated by this pragma, GNAT provides a
1339 package @code{Aux_@var{xxx}}, e.g.@: @code{Aux_DEC} for the DEC Ada 83
1340 implementation, which provides the required additional definitions. You
1341 can use this package in two ways. You can @code{with} it in the normal
1342 way and access entities either by selection or using a @code{use}
1343 clause. In this case no special processing is required.
1345 However, if existing code contains references such as
1346 @code{System.@var{xxx}} where @var{xxx} is an entity in the extended
1347 definitions provided in package @code{System}, you may use this pragma
1348 to extend visibility in @code{System} in a non-standard way that
1349 provides greater compatibility with the existing code. Pragma
1350 @code{Extend_System} is a configuration pragma whose single argument is
1351 the name of the package containing the extended definition
1352 (e.g.@: @code{Aux_DEC} for the DEC Ada case). A unit compiled under
1353 control of this pragma will be processed using special visibility
1354 processing that looks in package @code{System.Aux_@var{xxx}} where
1355 @code{Aux_@var{xxx}} is the pragma argument for any entity referenced in
1356 package @code{System}, but not found in package @code{System}.
1358 You can use this pragma either to access a predefined @code{System}
1359 extension supplied with the compiler, for example @code{Aux_DEC} or
1360 you can construct your own extension unit following the above
1361 definition. Note that such a package is a child of @code{System}
1362 and thus is considered part of the implementation. To compile
1363 it you will have to use the appropriate switch for compiling
1364 system units. See the GNAT User's Guide for details.
1367 @item pragma External
1373 [ Convention =>] convention_IDENTIFIER,
1374 [ Entity =>] local_NAME
1375 [, [External_Name =>] static_string_EXPRESSION ]
1376 [, [Link_Name =>] static_string_EXPRESSION ]);
1380 This pragma is identical in syntax and semantics to pragma
1381 @code{Export} as defined in the Ada Reference Manual. It is
1382 provided for compatibility with some Ada 83 compilers that
1383 used this pragma for exactly the same purposes as pragma
1384 @code{Export} before the latter was standardized.
1386 @cindex Dec Ada 83 casing compatibility
1387 @cindex External Names, casing
1388 @cindex Casing of External names
1389 @findex External_Name_Casing
1390 @item pragma External_Name_Casing
1395 pragma External_Name_Casing (
1396 Uppercase | Lowercase
1397 [, Uppercase | Lowercase | As_Is]);
1401 This pragma provides control over the casing of external names associated
1402 with Import and Export pragmas. There are two cases to consider:
1405 @item Implicit external names
1406 Implicit external names are derived from identifiers. The most common case
1407 arises when a standard Ada 95 Import or Export pragma is used with only two
1411 pragma Import (C, C_Routine);
1415 Since Ada is a case insensitive language, the spelling of the identifier in
1416 the Ada source program does not provide any information on the desired
1417 casing of the external name, and so a convention is needed. In GNAT the
1418 default treatment is that such names are converted to all lower case
1419 letters. This corresponds to the normal C style in many environments.
1420 The first argument of pragma @code{External_Name_Casing} can be used to
1421 control this treatment. If @code{Uppercase} is specified, then the name
1422 will be forced to all uppercase letters. If @code{Lowercase} is specified,
1423 then the normal default of all lower case letters will be used.
1425 This same implicit treatment is also used in the case of extended DEC Ada 83
1426 compatible Import and Export pragmas where an external name is explicitly
1427 specified using an identifier rather than a string.
1429 @item Explicit external names
1430 Explicit external names are given as string literals. The most common case
1431 arises when a standard Ada 95 Import or Export pragma is used with three
1435 pragma Import (C, C_Routine, "C_routine");
1439 In this case, the string literal normally provides the exact casing required
1440 for the external name. The second argument of pragma
1441 @code{External_Name_Casing} may be used to modify this behavior.
1442 If @code{Uppercase} is specified, then the name
1443 will be forced to all uppercase letters. If @code{Lowercase} is specified,
1444 then the name will be forced to all lowercase letters. A specification of
1445 @code{As_Is} provides the normal default behavior in which the casing is
1446 taken from the string provided.
1450 This pragma may appear anywhere that a pragma is valid. In particular, it
1451 can be used as a configuration pragma in the @file{gnat.adc} file, in which
1452 case it applies to all subsequent compilations, or it can be used as a program
1453 unit pragma, in which case it only applies to the current unit, or it can
1454 be used more locally to control individual Import/Export pragmas.
1456 It is primarily intended for use with OpenVMS systems, where many
1457 compilers convert all symbols to upper case by default. For interfacing to
1458 such compilers (e.g.@: the DEC C compiler), it may be convenient to use
1462 pragma External_Name_Casing (Uppercase, Uppercase);
1466 to enforce the upper casing of all external symbols.
1468 @findex Finalize_Storage_Only
1469 @item pragma Finalize_Storage_Only
1474 pragma Finalize_Storage_Only (first_subtype_LOCAL_NAME);
1478 This pragma allows the compiler not to emit a Finalize call for objects
1479 defined at the library level. This is mostly useful for types where
1480 finalization is only used to deal with storage reclamation since in most
1481 environments it is not necessary to reclaim memory just before terminating
1482 execution, hence the name.
1485 @findex Float_Representation
1486 @item pragma Float_Representation
1491 pragma Float_Representation (FLOAT_REP);
1493 FLOAT_REP ::= VAX_Float | IEEE_Float
1497 This pragma is implemented only in the OpenVMS implementation of GNAT@.
1498 It allows control over the internal representation chosen for the predefined
1499 floating point types declared in the packages @code{Standard} and
1500 @code{System}. For further details on this pragma, see the
1501 DEC Ada Language Reference Manual, section 3.5.7a. Note that to use this
1502 pragma, the standard runtime libraries must be recompiled. See the
1503 description of the @code{GNAT LIBRARY} command in the OpenVMS version
1504 of the GNAT Users Guide for details on the use of this command.
1512 pragma Ident (static_string_EXPRESSION);
1516 This pragma provides a string identification in the generated object file,
1517 if the system supports the concept of this kind of identification string.
1518 The maximum permitted length of the string literal is 31 characters.
1519 This pragma is allowed only in the outermost declarative part or
1520 declarative items of a compilation unit.
1522 On OpenVMS systems, the effect of the pragma is identical to the effect of
1523 the DEC Ada 83 pragma of the same name.
1526 @findex Import_Exception
1527 @item pragma Import_Exception
1532 pragma Import_Exception (
1533 [Internal =>] LOCAL_NAME,
1534 [, [External =>] EXTERNAL_SYMBOL,]
1535 [, [Form =>] Ada | VMS]
1536 [, [Code =>] static_integer_EXPRESSION]);
1540 | static_string_EXPRESSION
1544 This pragma is implemented only in the OpenVMS implementation of GNAT@.
1545 It allows OpenVMS conditions (for example, from OpenVMS system services or
1546 other OpenVMS languages) to be propagated to Ada programs as Ada exceptions.
1547 The pragma specifies that the exception associated with an exception
1548 declaration in an Ada program be defined externally (in non-Ada code).
1549 For further details on this pragma, see the
1550 DEC Ada Language Reference Manual, section 13.9a.3.1.
1552 @findex Import_Function
1553 @item pragma Import_Function @dots{}
1558 pragma Import_Function (
1559 [Internal =>] LOCAL_NAME,
1560 [, [External =>] EXTERNAL_SYMBOL]
1561 [, [Parameter_Types =>] PARAMETER_TYPES]
1562 [, [Result_Type =>] SUBTYPE_MARK]
1563 [, [Mechanism =>] MECHANISM]
1564 [, [Result_Mechanism =>] MECHANISM_NAME]
1565 [, [First_Optional_Parameter =>] IDENTIFIER]);
1569 | static_string_EXPRESSION
1573 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1577 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1579 MECHANISM_ASSOCIATION ::=
1580 [formal_parameter_NAME =>] MECHANISM_NAME
1585 | Descriptor [([Class =>] CLASS_NAME)]
1587 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1590 This pragma is used in conjunction with a pragma @code{Import} to
1591 specify additional information for an imported function. The pragma
1592 @code{Import} (or equivalent pragma @code{Interface}) must precede the
1593 @code{Import_Function} pragma and both must appear in the same
1594 declarative part as the function specification.
1596 The @var{Internal_Name} argument must uniquely designate
1597 the function to which the
1598 pragma applies. If more than one function name exists of this name in
1599 the declarative part you must use the @code{Parameter_Types} and
1600 @var{Result_Type} parameters to achieve the required unique
1601 designation. Subtype marks in these parameters must exactly match the
1602 subtypes in the corresponding function specification, using positional
1603 notation to match parameters with subtype marks.
1605 You may optionally use the @var{Mechanism} and @var{Result_Mechanism}
1606 parameters to specify passing mechanisms for the
1607 parameters and result. If you specify a single mechanism name, it
1608 applies to all parameters. Otherwise you may specify a mechanism on a
1609 parameter by parameter basis using either positional or named
1610 notation. If the mechanism is not specified, the default mechanism
1614 @cindex Passing by descriptor
1615 Passing by descriptor is supported only on the to OpenVMS ports of GNAT@.
1617 @code{First_Optional_Parameter} applies only to OpenVMS ports of GNAT@.
1618 It specifies that the designated parameter and all following parameters
1619 are optional, meaning that they are not passed at the generated code
1620 level (this is distinct from the notion of optional parameters in Ada
1621 where the parameters are passed anyway with the designated optional
1622 parameters). All optional parameters must be of mode @code{IN} and have
1623 default parameter values that are either known at compile time
1624 expressions, or uses of the @code{'Null_Parameter} attribute.
1626 @findex Import_Object
1627 @item pragma Import_Object
1632 pragma Import_Object
1633 [Internal =>] LOCAL_NAME,
1634 [, [External =>] EXTERNAL_SYMBOL],
1635 [, [Size =>] EXTERNAL_SYMBOL])
1639 | static_string_EXPRESSION
1643 This pragma designates an object as imported, and apart from the
1644 extended rules for external symbols, is identical in effect to the use of
1645 the normal @code{Import} pragma applied to an object. Unlike the
1646 subprogram case, you need not use a separate @code{Import} pragma,
1647 although you may do so (and probably should do so from a portability
1648 point of view). @var{size} is syntax checked, but otherwise ignored by
1651 @findex Import_Procedure
1652 @item pragma Import_Procedure
1657 pragma Import_Procedure (
1658 [Internal =>] LOCAL_NAME,
1659 [, [External =>] EXTERNAL_SYMBOL]
1660 [, [Parameter_Types =>] PARAMETER_TYPES]
1661 [, [Mechanism =>] MECHANISM]
1662 [, [First_Optional_Parameter =>] IDENTIFIER]);
1666 | static_string_EXPRESSION
1670 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1674 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1676 MECHANISM_ASSOCIATION ::=
1677 [formal_parameter_NAME =>] MECHANISM_NAME
1682 | Descriptor [([Class =>] CLASS_NAME)]
1684 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1688 This pragma is identical to @code{Import_Function} except that it
1689 applies to a procedure rather than a function and the parameters
1690 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
1692 @findex Import_Valued_Procedure
1693 @item pragma Import_Valued_Procedure @dots{}
1698 pragma Import_Valued_Procedure (
1699 [Internal =>] LOCAL_NAME,
1700 [, [External =>] EXTERNAL_SYMBOL]
1701 [, [Parameter_Types =>] PARAMETER_TYPES]
1702 [, [Mechanism =>] MECHANISM]
1703 [, [First_Optional_Parameter =>] IDENTIFIER]);
1707 | static_string_EXPRESSION
1711 | SUBTYPE_MARK @{, SUBTYPE_MARK@}
1715 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
1717 MECHANISM_ASSOCIATION ::=
1718 [formal_parameter_NAME =>] MECHANISM_NAME
1723 | Descriptor [([Class =>] CLASS_NAME)]
1725 CLASS_NAME ::= ubs | ubsb | uba | s | sb | a | nca
1729 This pragma is identical to @code{Import_Procedure} except that the
1730 first parameter of @var{local_name}, which must be present, must be of
1731 mode @code{OUT}, and externally the subprogram is treated as a function
1732 with this parameter as the result of the function. The purpose of this
1733 capability is to allow the use of @code{OUT} and @code{IN OUT}
1734 parameters in interfacing to external functions (which are not permitted
1735 in Ada functions). You may optionally use the @code{Mechanism}
1736 parameters to specify passing mechanisms for the parameters.
1737 If you specify a single mechanism name, it applies to all parameters.
1738 Otherwise you may specify a mechanism on a parameter by parameter
1739 basis using either positional or named notation. If the mechanism is not
1740 specified, the default mechanism is used.
1742 Note that it is important to use this pragma in conjunction with a separate
1743 pragma Import that specifies the desired convention, since otherwise the
1744 default convention is Ada, which is almost certainly not what is required.
1746 @findex Initialize_Scalars
1747 @cindex debugging with Initialize_Scalars
1748 @item pragma Initialize_Scalars
1753 pragma Initialize_Scalars;
1757 This pragma is similar to @code{Normalize_Scalars} conceptually but has
1758 two important differences. First, there is no requirement for the pragma
1759 to be used uniformly in all units of a partition, in particular, it is fine
1760 to use this just for some or all of the application units of a partition,
1761 without needing to recompile the run-time library.
1763 In the case where some units are compiled with the pragma, and some without,
1764 then a declaration of a variable where the type is defined in package
1765 Standard or is locally declared will always be subject to initialization,
1766 as will any declaration of a scalar variable. For composite variables,
1767 whether the variable is initialized may also depend on whether the package
1768 in which the type of the variable is declared is compiled with the pragma.
1770 The other important difference is that there is control over the value used
1771 for initializing scalar objects. At bind time, you can select whether to
1772 initialize with invalid values (like Normalize_Scalars), or with high or
1773 low values, or with a specified bit pattern. See the users guide for binder
1774 options for specifying these cases.
1776 This means that you can compile a program, and then without having to
1777 recompile the program, you can run it with different values being used
1778 for initializing otherwise uninitialized values, to test if your program
1779 behavior depends on the choice. Of course the behavior should not change,
1780 and if it does, then most likely you have an erroneous reference to an
1781 uninitialized value.
1783 Note that pragma @code{Initialize_Scalars} is particularly useful in
1784 conjunction with the enhanced validity checking that is now provided
1785 in GNAT, which checks for invalid values under more conditions.
1786 Using this feature (see description of the @code{-gnatv} flag in the
1787 users guide) in conjunction with pragma @code{Initialize_Scalars}
1788 provides a powerful new tool to assist in the detection of problems
1789 caused by uninitialized variables.
1791 @findex Inline_Always
1792 @item pragma Inline_Always
1797 pragma Inline_Always (NAME [, NAME]);
1801 Similar to pragma @code{Inline} except that inlining is not subject to
1802 the use of option @code{-gnatn} for inter-unit inlining.
1804 @findex Inline_Generic
1805 @item pragma Inline_Generic
1810 pragma Inline_Generic (generic_package_NAME)
1814 This is implemented for compatibility with DEC Ada 83 and is recognized,
1815 but otherwise ignored, by GNAT@. All generic instantiations are inlined
1816 by default when using GNAT@.
1819 @item pragma Interface
1825 [Convention =>] convention_identifier,
1826 [Entity =>] local_name
1827 [, [External_Name =>] static_string_expression],
1828 [, [Link_Name =>] static_string_expression]);
1832 This pragma is identical in syntax and semantics to
1833 the standard Ada 95 pragma @code{Import}. It is provided for compatibility
1834 with Ada 83. The definition is upwards compatible both with pragma
1835 @code{Interface} as defined in the Ada 83 Reference Manual, and also
1836 with some extended implementations of this pragma in certain Ada 83
1839 @findex Interface_Name
1840 @item pragma Interface_Name
1845 pragma Interface_Name (
1846 [Entity =>] LOCAL_NAME
1847 [, [External_Name =>] static_string_EXPRESSION]
1848 [, [Link_Name =>] static_string_EXPRESSION]);
1852 This pragma provides an alternative way of specifying the interface name
1853 for an interfaced subprogram, and is provided for compatibility with Ada
1854 83 compilers that use the pragma for this purpose. You must provide at
1855 least one of @var{External_Name} or @var{Link_Name}.
1858 @item pragma License
1859 @cindex License checking
1864 pragma License (Unrestricted | GPL | Modified_GPL | Restricted);
1868 This pragma is provided to allow automated checking for appropriate license
1869 conditions with respect to the standard and modified GPL@. A pragma @code{License},
1870 which is a configuration pragma that typically appears at the start of a
1871 source file or in a separate @file{gnat.adc} file, specifies the licensing
1872 conditions of a unit as follows:
1876 This is used for a unit that can be freely used with no license restrictions.
1877 Examples of such units are public domain units, and units from the Ada
1881 This is used for a unit that is licensed under the unmodified GPL, and which
1882 therefore cannot be @code{with}'ed by a restricted unit.
1885 This is used for a unit licensed under the GNAT modified GPL that includes
1886 a special exception paragraph that specifically permits the inclusion of
1887 the unit in programs without requiring the entire program to be released
1888 under the GPL@. This is the license used for the GNAT run-time which ensures
1889 that the run-time can be used freely in any program without GPL concerns.
1892 This is used for a unit that is restricted in that it is not permitted to
1893 depend on units that are licensed under the GPL@. Typical examples are
1894 proprietary code that is to be released under more restrictive license
1895 conditions. Note that restricted units are permitted to @code{with} units
1896 which are licensed under the modified GPL (this is the whole point of the
1902 Normally a unit with no @code{License} pragma is considered to have an
1903 unknown license, and no checking is done. However, standard GNAT headers
1904 are recognized, and license information is derived from them as follows.
1908 A GNAT license header starts with a line containing 78 hyphens. The following
1909 comment text is searched for the appearence of any of the following strings.
1911 If the string ``GNU General Public License'' is found, then the unit is assumed
1912 to have GPL license, unless the string ``As a special exception'' follows, in
1913 which case the license is assumed to be modified GPL@.
1915 If one of the strings
1916 ``This specification is adapated from the Ada Semantic Interface'' or
1917 ``This specification is derived from the Ada Reference Manual'' is found
1918 then the unit is assumed to be unrestricted.
1922 These default actions means that a program with a restricted license pragma
1923 will automatically get warnings if a GPL unit is inappropriately
1924 @code{with}'ed. For example, the program:
1929 procedure Secret_Stuff is
1935 if compiled with pragma @code{License} (@code{Restricted}) in a
1936 @file{gnat.adc} file will generate the warning:
1941 >>> license of withed unit "Sem_Ch3" is incompatible
1943 2. with GNAT.Sockets;
1944 3. procedure Secret_Stuff is
1947 Here we get a warning on @code{Sem_Ch3} since it is part of the GNAT
1948 compiler and is licensed under the
1949 GPL, but no warning for @code{GNAT.Sockets} which is part of the GNAT
1950 run time, and is therefore licensed under the modified GPL@.
1953 @item pragma Link_With
1958 pragma Link_With (static_string_EXPRESSION @{,static_string_EXPRESSION@});
1962 This pragma is provided for compatibility with certain Ada 83 compilers.
1963 It has exactly the same effect as pragma @code{Linker_Options} except
1964 that spaces occurring within one of the string expressions are treated
1965 as separators. For example, in the following case:
1968 pragma Link_With ("-labc -ldef");
1972 results in passing the strings @code{-labc} and @code{-ldef} as two
1973 separate arguments to the linker. In addition pragma Link_With allows
1974 multiple arguments, with the same effect as successive pragmas.
1976 @findex Linker_Alias
1977 @item pragma Linker_Alias
1982 pragma Linker_Alias (
1983 [Entity =>] LOCAL_NAME
1984 [Alias =>] static_string_EXPRESSION);
1988 This pragma establishes a linker alias for the given named entity. For
1989 further details on the exact effect, consult the GCC manual.
1991 @findex Linker_Section
1992 @item pragma Linker_Section
1997 pragma Linker_Section (
1998 [Entity =>] LOCAL_NAME
1999 [Section =>] static_string_EXPRESSION);
2003 This pragma specifies the name of the linker section for the given entity.
2004 For further details on the exact effect, consult the GCC manual.
2007 @item pragma No_Run_Time
2016 This is a configuration pragma that makes sure the user code does not
2017 use nor need anything from the GNAT run time. This is mostly useful in
2018 context where code certification is required. Please consult the
2019 @cite{GNAT Pro High-Integrity Edition User's Guide} for additional information.
2021 @findex Normalize_Scalars
2022 @item pragma Normalize_Scalars
2027 pragma Normalize_Scalars;
2031 This is a language defined pragma which is fully implemented in GNAT@. The
2032 effect is to cause all scalar objects that are not otherwise initialized
2033 to be initialized. The initial values are implementation dependent and
2037 @item Standard.Character
2039 Objects whose root type is Standard.Character are initialized to
2040 Character'Last. This will be out of range of the subtype only if
2041 the subtype range excludes this value.
2043 @item Standard.Wide_Character
2045 Objects whose root type is Standard.Wide_Character are initialized to
2046 Wide_Character'Last. This will be out of range of the subtype only if
2047 the subtype range excludes this value.
2051 Objects of an integer type are initialized to base_type'First, where
2052 base_type is the base type of the object type. This will be out of range
2053 of the subtype only if the subtype range excludes this value. For example,
2054 if you declare the subtype:
2057 subtype Ityp is integer range 1 .. 10;
2061 then objects of type x will be initialized to Integer'First, a negative
2062 number that is certainly outside the range of subtype @code{Ityp}.
2065 Objects of all real types (fixed and floating) are initialized to
2066 base_type'First, where base_Type is the base type of the object type.
2067 This will be out of range of the subtype only if the subtype range
2068 excludes this value.
2071 Objects of a modular type are initialized to typ'Last. This will be out
2072 of range of the subtype only if the subtype excludes this value.
2074 @item Enumeration types
2075 Objects of an enumeration type are initialized to all one-bits, i.e.@: to
2076 the value @code{2 ** typ'Size - 1}. This will be out of range of the enumeration
2077 subtype in all cases except where the subtype contains exactly
2078 2**8, 2**16, or 2**32 elements.
2084 @item pragma Long_Float
2089 pragma Long_Float (FLOAT_FORMAT);
2091 FLOAT_FORMAT ::= D_Float | G_Float
2095 This pragma is implemented only in the OpenVMS implementation of GNAT@.
2096 It allows control over the internal representation chosen for the predefined
2097 type @code{Long_Float} and for floating point type representations with
2098 @code{digits} specified in the range 7 through 15.
2099 For further details on this pragma, see the
2100 @cite{DEC Ada Language Reference Manual}, section 3.5.7b. Note that to use this
2101 pragma, the standard runtime libraries must be recompiled. See the
2102 description of the @code{GNAT LIBRARY} command in the OpenVMS version
2103 of the GNAT User's Guide for details on the use of this command.
2105 @findex Machine_Attribute
2106 @item pragma Machine_Attribute @dots{}
2111 pragma Machine_Attribute (
2112 [Attribute_Name =>] string_EXPRESSION,
2113 [Entity =>] LOCAL_NAME);
2116 Machine dependent attributes can be specified for types and/or
2117 declarations. Currently only subprogram entities are supported. This
2118 pragma is semantically equivalent to
2119 @code{__attribute__((@var{string_expression}))} in GNU C,
2120 where @code{@var{string_expression}} is
2121 recognized by the GNU C macros @code{VALID_MACHINE_TYPE_ATTRIBUTE} and
2122 @code{VALID_MACHINE_DECL_ATTRIBUTE} which are defined in the
2123 configuration header file @file{tm.h} for each machine. See the GCC
2124 manual for further information.
2127 @findex Main_Storage
2128 @item pragma Main_Storage
2134 (MAIN_STORAGE_OPTION [, MAIN_STORAGE_OPTION]);
2136 MAIN_STORAGE_OPTION ::=
2137 [WORKING_STORAGE =>] static_SIMPLE_EXPRESSION
2138 | [TOP_GUARD =>] static_SIMPLE_EXPRESSION
2143 This pragma is provided for compatibility with OpenVMS Vax Systems. It has
2144 no effect in GNAT, other than being syntax checked. Note that the pragma
2145 also has no effect in DEC Ada 83 for OpenVMS Alpha Systems.
2148 @item pragma No_Return
2153 pragma No_Return (procedure_LOCAL_NAME);
2157 @var{procedure_local_NAME} must refer to one or more procedure
2158 declarations in the current declarative part. A procedure to which this
2159 pragma is applied may not contain any explicit @code{return} statements,
2160 and also may not contain any implicit return statements from falling off
2161 the end of a statement sequence. One use of this pragma is to identify
2162 procedures whose only purpose is to raise an exception.
2164 Another use of this pragma is to suppress incorrect warnings about
2165 missing returns in functions, where the last statement of a function
2166 statement sequence is a call to such a procedure.
2169 @item pragma Passive
2174 pragma Passive ([Semaphore | No]);
2178 Syntax checked, but otherwise ignored by GNAT@. This is recognized for
2179 compatibility with DEC Ada 83 implementations, where it is used within a
2180 task definition to request that a task be made passive. If the argument
2181 @code{Semaphore} is present, or no argument is omitted, then DEC Ada 83
2182 treats the pragma as an assertion that the containing task is passive
2183 and that optimization of context switch with this task is permitted and
2184 desired. If the argument @code{No} is present, the task must not be
2185 optimized. GNAT does not attempt to optimize any tasks in this manner
2186 (since protected objects are available in place of passive tasks).
2189 @item pragma Polling
2194 pragma Polling (ON | OFF);
2198 This pragma controls the generation of polling code. This is normally off.
2199 If @code{pragma Polling (ON)} is used then periodic calls are generated to
2200 the routine @code{Ada.Exceptions.Poll}. This routine is a separate unit in the
2201 runtime library, and can be found in file @file{a-excpol.adb}.
2203 Pragma @code{Polling} can appear as a configuration pragma (for example it can be
2204 placed in the @file{gnat.adc} file) to enable polling globally, or it can be used
2205 in the statement or declaration sequence to control polling more locally.
2207 A call to the polling routine is generated at the start of every loop and
2208 at the start of every subprogram call. This guarantees that the @code{Poll}
2209 routine is called frequently, and places an upper bound (determined by
2210 the complexity of the code) on the period between two @code{Poll} calls.
2212 The primary purpose of the polling interface is to enable asynchronous
2213 aborts on targets that cannot otherwise support it (for example Windows
2214 NT), but it may be used for any other purpose requiring periodic polling.
2215 The standard version is null, and can be replaced by a user program. This
2216 will require re-compilation of the @code{Ada.Exceptions} package that can be found
2217 in files @file{a-except.ads} and @file{a-except.adb}.
2219 A standard alternative unit (in file @file{4wexcpol.adb} in the standard GNAT
2220 distribution) is used to enable the asynchronous abort capability on
2221 targets that do not normally support the capability. The version of @code{Poll}
2222 in this file makes a call to the appropriate runtime routine to test for
2225 Note that polling can also be enabled by use of the @code{-gnatP} switch. See
2226 the @cite{GNAT User's Guide} for details.
2228 @findex Propagate_Exceptions
2229 @cindex Zero Cost Exceptions
2230 @item pragma Propagate_Exceptions
2235 pragma Propagate_Exceptions (subprogram_LOCAL_NAME);
2239 This pragma indicates that the given entity, which is the name of an
2240 imported foreign-language subprogram may receive an Ada exception,
2241 and that the exception should be propagated. It is relevant only if
2242 zero cost exception handling is in use, and is thus never needed if
2243 the alternative @code{longjmp} / @code{setjmp} implementation of exceptions is used
2244 (although it is harmless to use it in such cases).
2246 The implementation of fast exceptions always properly propagates
2247 exceptions through Ada code, as described in the Ada Reference Manual.
2248 However, this manual is silent about the propagation of exceptions
2249 through foreign code. For example, consider the
2250 situation where @code{P1} calls
2251 @code{P2}, and @code{P2} calls @code{P3}, where
2252 @code{P1} and @code{P3} are in Ada, but @code{P2} is in C@.
2253 @code{P3} raises an Ada exception. The question is whether or not
2254 it will be propagated through @code{P2} and can be handled in
2257 For the @code{longjmp} / @code{setjmp} implementation of exceptions, the answer is
2258 always yes. For some targets on which zero cost exception handling
2259 is implemented, the answer is also always yes. However, there are
2260 some targets, notably in the current version all x86 architecture
2261 targets, in which the answer is that such propagation does not
2262 happen automatically. If such propagation is required on these
2263 targets, it is mandatory to use @code{Propagate_Exceptions} to
2264 name all foreign language routines through which Ada exceptions
2267 @findex Psect_Object
2268 @item pragma Psect_Object
2274 [Internal =>] LOCAL_NAME,
2275 [, [External =>] EXTERNAL_SYMBOL]
2276 [, [Size =>] EXTERNAL_SYMBOL]
2280 | static_string_EXPRESSION
2284 This pragma is identical in effect to pragma @code{Common_Object}.
2286 @findex Pure_Function
2287 @item pragma Pure_Function
2292 pragma Pure_Function ([Entity =>] function_LOCAL_NAME);
2295 This pragma appears in the same declarative part as a function
2296 declaration (or a set of function declarations if more than one
2297 overloaded declaration exists, in which case the pragma applies
2298 to all entities). If specifies that the function @code{Entity} is
2299 to be considered pure for the purposes of code generation. This means
2300 that the compiler can assume that there are no side effects, and
2301 in particular that two calls with identical arguments produce the
2302 same result. It also means that the function can be used in an
2305 Note that, quite deliberately, there are no static checks to try
2306 to ensure that this promise is met, so @code{Pure_Function} can be used
2307 with functions that are conceptually pure, even if they do modify
2308 global variables. For example, a square root function that is
2309 instrumented to count the number of times it is called is still
2310 conceptually pure, and can still be optimized, even though it
2311 modifies a global variable (the count). Memo functions are another
2312 example (where a table of previous calls is kept and consulted to
2313 avoid re-computation).
2316 Note: Most functions in a @code{Pure} package are automatically pure, and
2317 there is no need to use pragma @code{Pure_Function} for such functions. An
2318 exception is any function that has at least one formal of type
2319 @code{System.Address} or a type derived from it. Such functions are not
2320 considered pure by default, since the compiler assumes that the
2321 @code{Address} parameter may be functioning as a pointer and that the
2322 referenced data may change even if the address value does not. The use
2323 of pragma @code{Pure_Function} for such a function will override this default
2324 assumption, and cause the compiler to treat such a function as pure.
2326 Note: If pragma @code{Pure_Function} is applied to a renamed function, it
2327 applies to the underlying renamed function. This can be used to
2328 disambiguate cases of overloading where some but not all functions
2329 in a set of overloaded functions are to be designated as pure.
2332 @item pragma Ravenscar
2341 A configuration pragma that establishes the following set of restrictions:
2344 @item No_Abort_Statements
2345 [RM D.7] There are no abort_statements, and there are
2346 no calls to Task_Identification.Abort_Task.
2348 @item No_Select_Statements
2349 There are no select_statements.
2351 @item No_Task_Hierarchy
2352 [RM D.7] All (non-environment) tasks depend
2353 directly on the environment task of the partition.
2355 @item No_Task_Allocators
2356 [RM D.7] There are no allocators for task types
2357 or types containing task subcomponents.
2359 @item No_Dynamic_Priorities
2360 [RM D.7] There are no semantic dependencies on the package Dynamic_Priorities.
2362 @item No_Terminate_Alternatives
2363 [RM D.7] There are no selective_accepts with terminate_alternatives
2365 @item No_Dynamic_Interrupts
2366 There are no semantic dependencies on Ada.Interrupts.
2368 @item No_Protected_Type_Allocators
2369 There are no allocators for protected types or
2370 types containing protected subcomponents.
2372 @item No_Local_Protected_Objects
2373 Protected objects and access types that designate
2374 such objects shall be declared only at library level.
2377 Requeue statements are not allowed.
2380 There are no semantic dependencies on the package Ada.Calendar.
2382 @item No_Relative_Delay
2383 There are no delay_relative_statements.
2385 @item No_Task_Attributes
2386 There are no semantic dependencies on the Ada.Task_Attributes package and
2387 there are no references to the attributes Callable and Terminated [RM 9.9].
2389 @item Static_Storage_Size
2390 The expression for pragma Storage_Size is static.
2392 @item Boolean_Entry_Barriers
2393 Entry barrier condition expressions shall be boolean
2394 objects which are declared in the protected type
2395 which contains the entry.
2397 @item Max_Asynchronous_Select_Nesting = 0
2398 [RM D.7] Specifies the maximum dynamic nesting level of asynchronous_selects.
2399 A value of zero prevents the use of any asynchronous_select.
2401 @item Max_Task_Entries = 0
2402 [RM D.7] Specifies the maximum number of entries
2403 per task. The bounds of every entry family
2404 of a task unit shall be static, or shall be
2405 defined by a discriminant of a subtype whose
2406 corresponding bound is static. A value of zero
2407 indicates that no rendezvous are possible. For
2408 the Ravenscar pragma, the value of Max_Task_Entries is always
2411 @item Max_Protected_Entries = 1
2412 [RM D.7] Specifies the maximum number of entries per
2413 protected type. The bounds of every entry family of
2414 a protected unit shall be static, or shall be defined
2415 by a discriminant of a subtype whose corresponding
2416 bound is static. For the Ravenscar pragma the value of
2417 Max_Protected_Entries is always 1.
2419 @item Max_Select_Alternatives = 0
2420 [RM D.7] Specifies the maximum number of alternatives in a selective_accept.
2421 For the Ravenscar pragma the value if always 0.
2423 @item No_Task_Termination
2424 Tasks which terminate are erroneous.
2426 @item No_Entry_Queue
2427 No task can be queued on a protected entry. Note that this restrictions is
2428 checked at run time. The violation of this restriction generates a
2429 Program_Error exception.
2433 This set of restrictions corresponds to the definition of the ``Ravenscar
2434 Profile'' for limited tasking, devised and published by the @cite{International
2435 Real-Time Ada Workshop}, 1997.
2437 The above set is a superset of the restrictions provided by pragma
2438 @code{Restricted_Run_Time}, it includes six additional restrictions
2439 (@code{Boolean_Entry_Barriers}, @code{No_Select_Statements},
2440 @code{No_Calendar}, @code{Static_Storage_Size},
2441 @code{No_Relative_Delay} and @code{No_Task_Termination}). This means
2442 that pragma @code{Ravenscar}, like the pragma @code{Restricted_Run_Time}, automatically
2443 causes the use of a simplified, more efficient version of the tasking
2446 @findex Restricted_Run_Time
2447 @item pragma Restricted_Run_Time
2452 pragma Restricted_Run_Time
2456 A configuration pragma that establishes the following set of restrictions:
2459 @item No_Abort_Statements
2460 @item No_Asynchronous_Control
2461 @item No_Entry_Queue
2462 @item No_Task_Hierarchy
2463 @item No_Task_Allocators
2464 @item No_Dynamic_Priorities
2465 @item No_Terminate_Alternatives
2466 @item No_Dynamic_Interrupts
2467 @item No_Protected_Type_Allocators
2468 @item No_Local_Protected_Objects
2470 @item No_Task_Attributes
2471 @item Max_Asynchronous_Select_Nesting = 0
2472 @item Max_Task_Entries = 0
2473 @item Max_Protected_Entries = 1
2474 @item Max_Select_Alternatives = 0
2478 This set of restrictions causes the automatic selection of a simplified
2479 version of the run time that provides improved performance for the
2480 limited set of tasking functionality permitted by this set of restrictions.
2482 @findex Share_Generic
2483 @item pragma Share_Generic
2488 pragma Share_Generic (NAME @{, NAME@});
2492 This pragma is recognized for compatibility with other Ada compilers
2493 but is ignored by GNAT@. GNAT does not provide the capability for
2494 sharing of generic code. All generic instantiations result in making
2495 an inlined copy of the template with appropriate substitutions.
2497 @findex Source_File_Name
2498 @item pragma Source_File_Name
2503 pragma Source_File_Name (
2504 [Unit_Name =>] unit_NAME,
2505 Spec_File_Name => STRING_LITERAL);
2507 pragma Source_File_Name (
2508 [Unit_Name =>] unit_NAME,
2509 Body_File_Name => STRING_LITERAL);
2513 Use this to override the normal naming convention. It is a configuration
2514 pragma, and so has the usual applicability of configuration pragmas
2515 (i.e.@: it applies to either an entire partition, or to all units in a
2516 compilation, or to a single unit, depending on how it is used.
2517 @var{unit_name} is mapped to @var{file_name_literal}. The identifier for
2518 the second argument is required, and indicates whether this is the file
2519 name for the spec or for the body.
2521 Another form of the @code{Source_File_Name} pragma allows
2522 the specification of patterns defining alternative file naming schemes
2523 to apply to all files.
2526 pragma Source_File_Name
2527 (Spec_File_Name => STRING_LITERAL
2528 [,Casing => CASING_SPEC]
2529 [,Dot_Replacement => STRING_LITERAL]);
2531 pragma Source_File_Name
2532 (Body_File_Name => STRING_LITERAL
2533 [,Casing => CASING_SPEC]
2534 [,Dot_Replacement => STRING_LITERAL]);
2536 pragma Source_File_Name
2537 (Subunit_File_Name => STRING_LITERAL
2538 [,Casing => CASING_SPEC]
2539 [,Dot_Replacement => STRING_LITERAL]);
2541 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
2545 The first argument is a pattern that contains a single asterisk indicating
2546 the point at which the unit name is to be inserted in the pattern string
2547 to form the file name. The second argument is optional. If present it
2548 specifies the casing of the unit name in the resulting file name string.
2549 The default is lower case. Finally the third argument allows for systematic
2550 replacement of any dots in the unit name by the specified string literal.
2552 For more details on the use of the @code{Source_File_Name} pragma,
2553 see the sections ``Using Other File Names'' and
2554 ``Alternative File Naming Schemes'' in the @cite{GNAT User's Guide}.
2556 @findex Source_Reference
2557 @item pragma Source_Reference
2562 pragma Source_Reference (INTEGER_LITERAL,
2567 This pragma must appear as the first line of a source file.
2568 @var{integer_literal} is the logical line number of the line following
2569 the pragma line (for use in error messages and debugging
2570 information). @var{string_literal} is a static string constant that
2571 specifies the file name to be used in error messages and debugging
2572 information. This is most notably used for the output of @code{gnatchop}
2573 with the @code{-r} switch, to make sure that the original unchopped
2574 source file is the one referred to.
2576 The second argument must be a string literal, it cannot be a static
2577 string expression other than a string literal. This is because its value
2578 is needed for error messages issued by all phases of the compiler.
2580 @findex Stream_Convert
2581 @item pragma Stream_Convert
2586 pragma Stream_Convert (
2587 [Entity =>] type_LOCAL_NAME,
2588 [Read =>] function_NAME,
2589 [Write =>] function NAME);
2593 This pragma provides an efficient way of providing stream functions for
2594 types defined in packages. Not only is it simpler to use than declaring
2595 the necessary functions with attribute representation clauses, but more
2596 significantly, it allows the declaration to made in such a way that the
2597 stream packages are not loaded unless they are needed. The use of
2598 the Stream_Convert pragma adds no overhead at all, unless the stream
2599 attributes are actually used on the designated type.
2601 The first argument specifies the type for which stream functions are
2602 provided. The second parameter provides a function used to read values
2603 of this type. It must name a function whose argument type may be any
2604 subtype, and whose returned type must be the type given as the first
2605 argument to the pragma.
2607 The meaning of the @var{Read}
2608 parameter is that if a stream attribute directly
2609 or indirectly specifies reading of the type given as the first parameter,
2610 then a value of the type given as the argument to the Read function is
2611 read from the stream, and then the Read function is used to convert this
2612 to the required target type.
2614 Similarly the @var{Write} parameter specifies how to treat write attributes
2615 that directly or indirectly apply to the type given as the first parameter.
2616 It must have an input parameter of the type specified by the first parameter,
2617 and the return type must be the same as the input type of the Read function.
2618 The effect is to first call the Write function to convert to the given stream
2619 type, and then write the result type to the stream.
2621 The Read and Write functions must not be overloaded subprograms. If necessary
2622 renamings can be supplied to meet this requirement.
2623 The usage of this attribute is best illustrated by a simple example, taken
2624 from the GNAT implementation of package Ada.Strings.Unbounded:
2627 function To_Unbounded (S : String)
2628 return Unbounded_String
2629 renames To_Unbounded_String;
2631 pragma Stream_Convert
2632 (Unbounded_String, To_Unbounded, To_String);
2636 The specifications of the referenced functions, as given in the Ada 95
2637 Reference Manual are:
2640 function To_Unbounded_String (Source : String)
2641 return Unbounded_String;
2643 function To_String (Source : Unbounded_String)
2648 The effect is that if the value of an unbounded string is written to a
2649 stream, then the representation of the item in the stream is in the same
2650 format used for @code{Standard.String}, and this same representation is
2651 expected when a value of this type is read from the stream.
2653 @findex Style_Checks
2654 @item pragma Style_Checks
2659 pragma Style_Checks (string_LITERAL | ALL_CHECKS |
2660 On | Off [, LOCAL_NAME]);
2664 This pragma is used in conjunction with compiler switches to control the
2665 built in style checking provided by GNAT@. The compiler switches, if set
2666 provide an initial setting for the switches, and this pragma may be used
2667 to modify these settings, or the settings may be provided entirely by
2668 the use of the pragma. This pragma can be used anywhere that a pragma
2669 is legal, including use as a configuration pragma (including use in
2670 the @file{gnat.adc} file).
2672 The form with a string literal specifies which style options are to be
2673 activated. These are additive, so they apply in addition to any previously
2674 set style check options. The codes for the options are the same as those
2675 used in the @code{-gnaty} switch to @code{gcc} or @code{gnatmake}.
2676 For example the following two methods can be used to enable
2680 pragma Style_Checks ("l");
2681 gcc -c -gnatyl @dots{}
2685 The form ALL_CHECKS activates all standard checks (its use is equivalent
2686 to the use of the @code{gnaty} switch with no options. See GNAT User's
2689 The forms with @code{Off} and @code{On}
2690 can be used to temporarily disable style checks
2691 as shown in the following example:
2697 pragma Style_Checks ("k"); -- requires keywords in lower case
2698 pragma Style_Checks (Off); -- turn off style checks
2699 NULL; -- this will not generate an error message
2700 pragma Style_Checks (On); -- turn style checks back on
2701 NULL; -- this will generate an error message
2705 Finally the two argument form is allowed only if the first argument is
2706 @code{On} or @code{Off}. The effect is to turn of semantic style checks
2707 for the specified entity, as shown in the following example:
2713 pragma Style_Checks ("r"); -- require consistency of identifier casing
2715 Rf1 : Integer := ARG; -- incorrect, wrong case
2716 pragma Style_Checks (Off, Arg);
2717 Rf2 : Integer := ARG; -- OK, no error
2721 @item pragma Subtitle
2726 pragma Subtitle ([Subtitle =>] STRING_LITERAL);
2730 This pragma is recognized for compatibility with other Ada compilers
2731 but is ignored by GNAT@.
2733 @findex Suppress_All
2734 @item pragma Suppress_All
2739 pragma Suppress_All;
2743 This pragma can only appear immediately following a compilation
2744 unit. The effect is to apply @code{Suppress (All_Checks)} to the unit
2745 which it follows. This pragma is implemented for compatibility with DEC
2746 Ada 83 usage. The use of pragma @code{Suppress (All_Checks)} as a normal
2747 configuration pragma is the preferred usage in GNAT@.
2749 @findex Suppress_Initialization
2750 @cindex Suppressing initialization
2751 @cindex Initialization, suppression of
2752 @item pragma Suppress_Initialization
2757 pragma Suppress_Initialization ([Entity =>] type_Name);
2761 This pragma suppresses any implicit or explicit initialization
2762 associated with the given type name for all variables of this type.
2765 @item pragma Task_Info
2770 pragma Task_Info (EXPRESSION);
2774 This pragma appears within a task definition (like pragma
2775 @code{Priority}) and applies to the task in which it appears. The
2776 argument must be of type @code{System.Task_Info.Task_Info_Type}.
2777 The @code{Task_Info} pragma provides system dependent control over
2778 aspect of tasking implementation, for example, the ability to map
2779 tasks to specific processors. For details on the facilities available
2780 for the version of GNAT that you are using, see the documentation
2781 in the specification of package System.Task_Info in the runtime
2785 @item pragma Task_Name
2790 pragma Task_Name (string_EXPRESSION);
2794 This pragma appears within a task definition (like pragma
2795 @code{Priority}) and applies to the task in which it appears. The
2796 argument must be of type String, and provides a name to be used for
2797 the task instance when the task is created. Note that this expression
2798 is not required to be static, and in particular, it can contain
2799 references to task discriminants. This facility can be used to
2800 provide different names for different tasks as they are created,
2801 as illustrated in the example below.
2803 The task name is recorded internally in the run-time structures
2804 and is accessible to tools like the debugger. In addition the
2805 routine @code{Ada.Task_Identification.Image} will return this
2806 string, with a unique task address appended.
2809 -- Example of the use of pragma Task_Name
2811 with Ada.Task_Identification;
2812 use Ada.Task_Identification;
2813 with Text_IO; use Text_IO;
2816 type Astring is access String;
2818 task type Task_Typ (Name : access String) is
2819 pragma Task_Name (Name.all);
2822 task body Task_Typ is
2823 Nam : constant String := Image (Current_Task);
2825 Put_Line ("-->" & Nam (1 .. 14) & "<--");
2828 type Ptr_Task is access Task_Typ;
2829 Task_Var : Ptr_Task;
2833 new Task_Typ (new String'("This is task 1"));
2835 new Task_Typ (new String'("This is task 2"));
2839 @findex Task_Storage
2840 @item pragma Task_Storage
2845 [Task_Type =>] LOCAL_NAME,
2846 [Top_Guard =>] static_integer_EXPRESSION);
2849 This pragma specifies the length of the guard area for tasks. The guard
2850 area is an additional storage area allocated to a task. A value of zero
2851 means that either no guard area is created or a minimal guard area is
2852 created, depending on the target. This pragma can appear anywhere a
2853 @code{Storage_Size} attribute definition clause is allowed for a task
2857 @item pragma Time_Slice
2862 pragma Time_Slice (static_duration_EXPRESSION);
2866 For implementations of GNAT on operating systems where it is possible
2867 to supply a time slice value, this pragma may be used for this purpose.
2868 It is ignored if it is used in a system that does not allow this control,
2869 or if it appears in other than the main program unit.
2871 Note that the effect of this pragma is identical to the effect of the
2872 DEC Ada 83 pragma of the same name when operating under OpenVMS systems.
2880 pragma Title (TITLING_OPTION [, TITLING OPTION]);
2883 [Title =>] STRING_LITERAL,
2884 | [Subtitle =>] STRING_LITERAL
2888 Syntax checked but otherwise ignored by GNAT@. This is a listing control
2889 pragma used in DEC Ada 83 implementations to provide a title and/or
2890 subtitle for the program listing. The program listing generated by GNAT
2891 does not have titles or subtitles.
2893 Unlike other pragmas, the full flexibility of named notation is allowed
2894 for this pragma, i.e.@: the parameters may be given in any order if named
2895 notation is used, and named and positional notation can be mixed
2896 following the normal rules for procedure calls in Ada.
2899 @findex Unchecked_Union
2900 @item pragma Unchecked_Union
2905 pragma Unchecked_Union (first_subtype_LOCAL_NAME)
2909 This pragma is used to declare that the specified type should be represented
2911 equivalent to a C union type, and is intended only for use in
2912 interfacing with C code that uses union types. In Ada terms, the named
2913 type must obey the following rules:
2917 It is a non-tagged non-limited record type.
2919 It has a single discrete discriminant with a default value.
2921 The component list consists of a single variant part.
2923 Each variant has a component list with a single component.
2925 No nested variants are allowed.
2927 No component has an explicit default value.
2929 No component has a non-static constraint.
2932 In addition, given a type that meets the above requirements, the
2933 following restrictions apply to its use throughout the program:
2937 The discriminant name can be mentioned only in an aggregate.
2939 No subtypes may be created of this type.
2941 The type may not be constrained by giving a discriminant value.
2943 The type cannot be passed as the actual for a generic formal with a
2947 Equality and inequality operations on @code{unchecked_unions} are not
2948 available, since there is no discriminant to compare and the compiler
2949 does not even know how many bits to compare. It is implementation
2950 dependent whether this is detected at compile time as an illegality or
2951 whether it is undetected and considered to be an erroneous construct. In
2952 GNAT, a direct comparison is illegal, but GNAT does not attempt to catch
2953 the composite case (where two composites are compared that contain an
2954 unchecked union component), so such comparisons are simply considered
2957 The layout of the resulting type corresponds exactly to a C union, where
2958 each branch of the union corresponds to a single variant in the Ada
2959 record. The semantics of the Ada program is not changed in any way by
2960 the pragma, i.e.@: provided the above restrictions are followed, and no
2961 erroneous incorrect references to fields or erroneous comparisons occur,
2962 the semantics is exactly as described by the Ada reference manual.
2963 Pragma @code{Suppress (Discriminant_Check)} applies implicitly to the
2964 type and the default convention is C
2966 @findex Unimplemented_Unit
2967 @item pragma Unimplemented_Unit
2972 pragma Unimplemented_Unit;
2976 If this pragma occurs in a unit that is processed by the compiler, GNAT
2977 aborts with the message @samp{@var{xxx} not implemented}, where
2978 @var{xxx} is the name of the current compilation unit. This pragma is
2979 intended to allow the compiler to handle unimplemented library units in
2982 The abort only happens if code is being generated. Thus you can use
2983 specs of unimplemented packages in syntax or semantic checking mode.
2985 @findex Unreferenced
2986 @item pragma Unreferenced
2987 @cindex Warnings, unreferenced
2992 pragma Unreferenced (local_Name @{, local_Name@});
2996 This pragma signals that the entities whose names are listed are
2997 deliberately not referenced. This suppresses warnings about the
2998 entities being unreferenced, and in addition a warning will be
2999 generated if one of these entities is in fact referenced.
3001 This is particularly useful for clearly signalling that a particular
3002 parameter is not referenced in some particular subprogram implementation
3003 and that this is deliberate. It can also be useful in the case of
3004 objects declared only for their initialization or finalization side
3007 If @code{local_Name} identifies more than one matching homonym in the
3008 current scope, then the entity most recently declared is the one to which
3011 @findex Unreserve_All_Interrupts
3012 @item pragma Unreserve_All_Interrupts
3017 pragma Unreserve_All_Interrupts;
3021 Normally certain interrupts are reserved to the implementation. Any attempt
3022 to attach an interrupt causes Program_Error to be raised, as described in
3023 RM C.3.2(22). A typical example is the @code{SIGINT} interrupt used in
3024 many systems for an @kbd{Ctrl-C} interrupt. Normally this interrupt is
3025 reserved to the implementation, so that @kbd{Ctrl-C} can be used to
3026 interrupt execution.
3028 If the pragma @code{Unreserve_All_Interrupts} appears anywhere in any unit in
3029 a program, then all such interrupts are unreserved. This allows the
3030 program to handle these interrupts, but disables their standard
3031 functions. For example, if this pragma is used, then pressing
3032 @kbd{Ctrl-C} will not automatically interrupt execution. However,
3033 a program can then handle the @code{SIGINT} interrupt as it chooses.
3035 For a full list of the interrupts handled in a specific implementation,
3036 see the source code for the specification of @code{Ada.Interrupts.Names} in
3037 file @file{a-intnam.ads}. This is a target dependent file that contains the
3038 list of interrupts recognized for a given target. The documentation in
3039 this file also specifies what interrupts are affected by the use of
3040 the @code{Unreserve_All_Interrupts} pragma.
3043 @item pragma Unsuppress
3048 pragma Unsuppress (IDENTIFIER [, [On =>] NAME]);
3052 This pragma undoes the effect of a previous pragma @code{Suppress}. If
3053 there is no corresponding pragma @code{Suppress} in effect, it has no
3054 effect. The range of the effect is the same as for pragma
3055 @code{Suppress}. The meaning of the arguments is identical to that used
3056 in pragma @code{Suppress}.
3058 One important application is to ensure that checks are on in cases where
3059 code depends on the checks for its correct functioning, so that the code
3060 will compile correctly even if the compiler switches are set to suppress
3063 @cindex @code{Size}, VADS compatibility
3064 @findex Use_VADS_Size
3065 @item pragma Use_VADS_Size
3070 pragma Use_VADS_Size;
3074 This is a configuration pragma. In a unit to which it applies, any use
3075 of the 'Size attribute is automatically interpreted as a use of the
3076 'VADS_Size attribute. Note that this may result in incorrect semantic
3077 processing of valid Ada 95 programs. This is intended to aid in the
3078 handling of legacy code which depends on the interpretation of Size
3079 as implemented in the VADS compiler. See description of the VADS_Size
3080 attribute for further details.
3082 @findex Validity_Checks
3083 @item pragma Validity_Checks
3088 pragma Validity_Checks (string_LITERAL | ALL_CHECKS | On | Off);
3092 This pragma is used in conjunction with compiler switches to control the
3093 built in validity checking provided by GNAT@. The compiler switches, if set
3094 provide an initial setting for the switches, and this pragma may be used
3095 to modify these settings, or the settings may be provided entirely by
3096 the use of the pragma. This pragma can be used anywhere that a pragma
3097 is legal, including use as a configuration pragma (including use in
3098 the @file{gnat.adc} file).
3100 The form with a string literal specifies which validity options are to be
3101 activated. The validity checks are first set to include only the default
3102 reference manual settings, and then a string of letters in the string
3103 specifies the exact set of options required. The form of this string
3104 is exactly as described for the @code{-gnatVx} compiler switch (see the
3105 GNAT users guide for details). For example the following two methods
3106 can be used to enable validity checking for mode @code{in} and
3107 @code{in out} subprogram parameters:
3110 pragma Validity_Checks ("im");
3111 gcc -c -gnatVim @dots{}
3115 The form ALL_CHECKS activates all standard checks (its use is equivalent
3116 to the use of the @code{gnatva} switch.
3118 The forms with @code{Off} and @code{On}
3119 can be used to temporarily disable validity checks
3120 as shown in the following example:
3126 pragma Validity_Checks ("c"); -- validity checks for copies
3127 pragma Validity_Checks (Off); -- turn off validity checks
3128 A := B; -- B will not be validity checked
3129 pragma Validity_Checks (On); -- turn validity checks back on
3130 A := C; -- C will be validity checked
3134 @item pragma Volatile
3139 pragma Volatile (local_NAME)
3143 This pragma is defined by the Ada 95 Reference Manual, and the GNAT
3144 implementation is fully conformant with this definition. The reason it
3145 is mentioned in this section is that a pragma of the same name was supplied
3146 in some Ada 83 compilers, including DEC Ada 83. The Ada 95 implementation
3147 of pragma Volatile is upwards compatible with the implementation in
3151 @item pragma Warnings
3156 pragma Warnings (On | Off [, LOCAL_NAME]);
3160 Normally warnings are enabled, with the output being controlled by
3161 the command line switch. Warnings (@code{Off}) turns off generation of
3162 warnings until a Warnings (@code{On}) is encountered or the end of the
3163 current unit. If generation of warnings is turned off using this
3164 pragma, then no warning messages are output, regardless of the
3165 setting of the command line switches.
3167 The form with a single argument is a configuration pragma.
3169 If the @var{local_name} parameter is present, warnings are suppressed for
3170 the specified entity. This suppression is effective from the point where
3171 it occurs till the end of the extended scope of the variable (similar to
3172 the scope of @code{Suppress}).
3174 @findex Weak_External
3175 @item pragma Weak_External
3180 pragma Weak_External ([Entity =>] LOCAL_NAME);
3184 This pragma specifies that the given entity should be marked as a weak
3185 external (one that does not have to be resolved) for the linker. For
3186 further details, consult the GCC manual.
3189 @node Implementation Defined Attributes
3190 @chapter Implementation Defined Attributes
3191 Ada 95 defines (throughout the Ada 95 reference manual,
3192 summarized in annex K),
3193 a set of attributes that provide useful additional functionality in all
3194 areas of the language. These language defined attributes are implemented
3195 in GNAT and work as described in the Ada 95 Reference Manual.
3197 In addition, Ada 95 allows implementations to define additional
3198 attributes whose meaning is defined by the implementation. GNAT provides
3199 a number of these implementation-dependent attributes which can be used
3200 to extend and enhance the functionality of the compiler. This section of
3201 the GNAT reference manual describes these additional attributes.
3203 Note that any program using these attributes may not be portable to
3204 other compilers (although GNAT implements this set of attributes on all
3205 platforms). Therefore if portability to other compilers is an important
3206 consideration, you should minimize the use of these attributes.
3209 @findex Abort_Signal
3212 @code{Standard'Abort_Signal} (@code{Standard} is the only allowed
3213 prefix) provides the entity for the special exception used to signal
3214 task abort or asynchronous transfer of control. Normally this attribute
3215 should only be used in the tasking runtime (it is highly peculiar, and
3216 completely outside the normal semantics of Ada, for a user program to
3217 intercept the abort exception).
3219 @cindex Size of @code{Address}
3220 @findex Address_Size
3223 @code{Standard'Address_Size} (@code{Standard} is the only allowed
3224 prefix) is a static constant giving the number of bits in an
3225 @code{Address}. It is used primarily for constructing the definition of
3226 @code{Memory_Size} in package @code{Standard}, but may be freely used in user
3227 programs and has the advantage of being static, while a direct
3228 reference to System.Address'Size is non-static because Address
3234 The @code{Asm_Input} attribute denotes a function that takes two
3235 parameters. The first is a string, the second is an expression of the
3236 type designated by the prefix. The first (string) argument is required
3237 to be a static expression, and is the constraint for the parameter,
3238 (e.g.@: what kind of register is required). The second argument is the
3239 value to be used as the input argument. The possible values for the
3240 constant are the same as those used in the RTL, and are dependent on
3241 the configuration file used to built the GCC back end.
3242 @ref{Machine Code Insertions}
3247 The @code{Asm_Output} attribute denotes a function that takes two
3248 parameters. The first is a string, the second is the name of a variable
3249 of the type designated by the attribute prefix. The first (string)
3250 argument is required to be a static expression and designates the
3251 constraint for the parameter (e.g.@: what kind of register is
3252 required). The second argument is the variable to be updated with the
3253 result. The possible values for constraint are the same as those used in
3254 the RTL, and are dependent on the configuration file used to build the
3255 GCC back end. If there are no output operands, then this argument may
3256 either be omitted, or explicitly given as @code{No_Output_Operands}.
3257 @ref{Machine Code Insertions}
3263 This attribute is implemented only in OpenVMS versions of GNAT@. Applied to
3264 the name of an entry, it yields a value of the predefined type AST_Handler
3265 (declared in the predefined package System, as extended by the use of
3266 pragma @code{Extend_System (Aux_DEC)}). This value enables the given entry to
3267 be called when an AST occurs. For further details, refer to the @cite{DEC Ada
3268 Language Reference Manual}, section 9.12a.
3272 @code{@var{obj}'Bit}, where @var{obj} is any object, yields the bit
3273 offset within the storage unit (byte) that contains the first bit of
3274 storage allocated for the object. The value of this attribute is of the
3275 type @code{Universal_Integer}, and is always a non-negative number not
3276 exceeding the value of @code{System.Storage_Unit}.
3278 For an object that is a variable or a constant allocated in a register,
3279 the value is zero. (The use of this attribute does not force the
3280 allocation of a variable to memory).
3282 For an object that is a formal parameter, this attribute applies
3283 to either the matching actual parameter or to a copy of the
3284 matching actual parameter.
3286 For an access object the value is zero. Note that
3287 @code{@var{obj}.all'Bit} is subject to an @code{Access_Check} for the
3288 designated object. Similarly for a record component
3289 @code{@var{X}.@var{C}'Bit} is subject to a discriminant check and
3290 @code{@var{X}(@var{I}).Bit} and @code{@var{X}(@var{I1}..@var{I2})'Bit}
3291 are subject to index checks.
3293 This attribute is designed to be compatible with the DEC Ada 83 definition
3294 and implementation of the @code{Bit} attribute.
3296 @findex Bit_Position
3299 @code{@var{R.C}'Bit}, where @var{R} is a record object and C is one
3300 of the fields of the record type, yields the bit
3301 offset within the record contains the first bit of
3302 storage allocated for the object. The value of this attribute is of the
3303 type @code{Universal_Integer}. The value depends only on the field
3304 @var{C} and is independent of the alignment of
3305 the containing record @var{R}.
3307 @findex Code_Address
3308 @cindex Subprogram address
3309 @cindex Address of subprogram code
3313 attribute may be applied to subprograms in Ada 95, but the
3314 intended effect from the Ada 95 reference manual seems to be to provide
3315 an address value which can be used to call the subprogram by means of
3316 an address clause as in the following example:
3319 procedure K is @dots{}
3322 for L'Address use K'Address;
3323 pragma Import (Ada, L);
3327 A call to @code{L} is then expected to result in a call to @code{K}@. In Ada 83, where
3328 there were no access-to-subprogram values, this was a common work around
3329 for getting the effect of an indirect call.
3330 GNAT implements the above use of @code{Address} and the technique illustrated
3331 by the example code works correctly.
3333 However, for some purposes, it is useful to have the address of the start
3334 of the generated code for the subprogram. On some architectures, this is
3335 not necessarily the same as the @code{Address} value described above. For example,
3336 the @code{Address} value may reference a subprogram descriptor rather than the
3339 The @code{'Code_Address} attribute, which can only be applied to
3340 subprogram entities, always returns the address of the start of the
3341 generated code of the specified subprogram, which may or may not be
3342 the same value as is returned by the corresponding @code{'Address}
3346 @cindex Little endian
3347 @findex Default_Bit_Order
3348 @item Default_Bit_Order
3350 @code{Standard'Default_Bit_Order} (@code{Standard} is the only
3351 permissible prefix), provides the value @code{System.Default_Bit_Order}
3352 as a @code{Pos} value (0 for @code{High_Order_First}, 1 for
3353 @code{Low_Order_First}). This is used to construct the definition of
3354 @code{Default_Bit_Order} in package @code{System}.
3359 The prefix of the @code{'Elaborated} attribute must be a unit name. The
3360 value is a Boolean which indicates whether or not the given unit has been
3361 elaborated. This attribute is primarily intended for internal use by the
3362 generated code for dynamic elaboration checking, but it can also be used
3363 in user programs. The value will always be True once elaboration of all
3364 units has been completed.
3369 This attribute can only be applied to a program unit name. It returns
3370 the entity for the corresponding elaboration procedure for elaborating
3371 the body of the referenced unit. This is used in the main generated
3372 elaboration procedure by the binder and is not normally used in any
3373 other context. However, there may be specialized situations in which it
3374 is useful to be able to call this elaboration procedure from Ada code,
3375 e.g.@: if it is necessary to do selective re-elaboration to fix some
3381 This attribute can only be applied to a program unit name. It returns
3382 the entity for the corresponding elaboration procedure for elaborating
3383 the specification of the referenced unit. This is used in the main
3384 generated elaboration procedure by the binder and is not normally used
3385 in any other context. However, there may be specialized situations in
3386 which it is useful to be able to call this elaboration procedure from
3387 Ada code, e.g.@: if it is necessary to do selective re-elaboration to fix
3390 @cindex Ada 83 attributes
3394 The @code{Emax} attribute is provided for compatibility with Ada 83. See
3395 the Ada 83 reference manual for an exact description of the semantics of
3398 @cindex Representation of enums
3402 For every enumeration subtype @var{S}, @code{@var{S}'Enum_Rep} denotes a
3403 function with the following specification:
3406 function @var{S}'Enum_Rep (Arg : @var{S}'Base)
3407 return Universal_Integer;
3411 It is also allowable to apply @code{Enum_Rep} directly to an object of an
3412 enumeration type or to a non-overloaded enumeration
3413 literal. In this case @code{@var{S}'Enum_Rep} is equivalent to
3414 @code{@var{typ}'Enum_Rep(@var{S})} where @var{typ} is the type of the
3415 enumeration literal or object.
3417 The function returns the representation value for the given enumeration
3418 value. This will be equal to value of the @code{Pos} attribute in the
3419 absence of an enumeration representation clause. This is a static
3420 attribute (i.e.@: the result is static if the argument is static).
3422 @code{@var{S}'Enum_Rep} can also be used with integer types and objects, in which
3423 case it simply returns the integer value. The reason for this is to allow
3424 it to be used for @code{(<>)} discrete formal arguments in a generic unit that
3425 can be instantiated with either enumeration types or integer types. Note
3426 that if @code{Enum_Rep} is used on a modular type whose upper bound exceeds the
3427 upper bound of the largest signed integer type, and the argument is a
3428 variable, so that the universal integer calculation is done at run-time,
3429 then the call to @code{Enum_Rep} may raise @code{Constraint_Error}.
3431 @cindex Ada 83 attributes
3435 The @code{Epsilon} attribute is provided for compatibility with Ada 83. See
3436 the Ada 83 reference manual for an exact description of the semantics of
3442 For every fixed-point type @var{S}, @code{@var{S}'Fixed_Value} denotes a
3443 function with the following specification:
3446 function @var{S}'Fixed_Value (Arg : Universal_Integer)
3451 The value returned is the fixed-point value @var{V} such that
3454 @var{V} = Arg * @var{S}'Small
3458 The effect is thus equivalent to first converting the argument to the
3459 integer type used to represent @var{S}, and then doing an unchecked
3460 conversion to the fixed-point type. This attribute is primarily intended
3461 for use in implementation of the input-output functions for fixed-point
3464 @cindex Discriminants, testing for
3465 @findex Has_Discriminants
3466 @item Has_Discriminants
3468 The prefix of the @code{Has_Discriminants} attribute is a type. The result
3469 is a Boolean value which is True if the type has discriminants, and False
3470 otherwise. The intended use of this attribute is in conjunction with generic
3471 definitions. If the attribute is applied to a generic private type, it
3472 indicates whether or not the corresponding actual type has discriminants.
3477 The @code{Img} attribute differs from @code{Image} in that it may be
3478 applied to objects as well as types, in which case it gives the
3479 @code{Image} for the subtype of the object. This is convenient for
3483 Put_Line ("X = " & X'Img);
3487 has the same meaning as the more verbose:
3490 Put_Line ("X = " & @var{type}'Image (X));
3493 where @var{type} is the subtype of the object X@.
3495 @findex Integer_Value
3498 For every integer type @var{S}, @code{@var{S}'Integer_Value} denotes a
3499 function with the following specification:
3502 function @var{S}'Integer_Value (Arg : Universal_Fixed)
3507 The value returned is the integer value @var{V}, such that
3510 Arg = @var{V} * @var{type}'Small
3514 The effect is thus equivalent to first doing an unchecked convert from
3515 the fixed-point type to its corresponding implementation type, and then
3516 converting the result to the target integer type. This attribute is
3517 primarily intended for use in implementation of the standard
3518 input-output functions for fixed-point values.
3520 @cindex Ada 83 attributes
3524 The @code{Large} attribute is provided for compatibility with Ada 83. See
3525 the Ada 83 reference manual for an exact description of the semantics of
3528 @findex Machine_Size
3531 This attribute is identical to the @code{Object_Size} attribute. It is
3532 provided for compatibility with the DEC Ada 83 attribute of this name.
3534 @cindex Ada 83 attributes
3538 The @code{Mantissa} attribute is provided for compatibility with Ada 83. See
3539 the Ada 83 reference manual for an exact description of the semantics of
3542 @cindex Interrupt priority, maximum
3543 @findex Max_Interrupt_Priority
3544 @item Max_Interrupt_Priority
3546 @code{Standard'Max_Interrupt_Priority} (@code{Standard} is the only
3547 permissible prefix), provides the value
3548 @code{System.Max_Interrupt_Priority} and is intended primarily for
3549 constructing this definition in package @code{System}.
3551 @cindex Priority, maximum
3552 @findex Max_Priority
3555 @code{Standard'Max_Priority} (@code{Standard} is the only permissible
3556 prefix) provides the value @code{System.Max_Priority} and is intended
3557 primarily for constructing this definition in package @code{System}.
3559 @cindex Alignment, maximum
3560 @findex Maximum_Alignment
3561 @item Maximum_Alignment
3563 @code{Standard'Maximum_Alignment} (@code{Standard} is the only
3564 permissible prefix) provides the maximum useful alignment value for the
3565 target. This is a static value that can be used to specify the alignment
3566 for an object, guaranteeing that it is properly aligned in all
3567 cases. This is useful when an external object is imported and its
3568 alignment requirements are unknown.
3570 @cindex Return values, passing mechanism
3571 @cindex Parameters, passing mechanism
3572 @findex Mechanism_Code
3573 @item Mechanism_Code
3575 @code{@var{function}'Mechanism_Code} yields an integer code for the
3576 mechanism used for the result of function, and
3577 @code{@var{subprogram}'Mechanism_Code (@var{n})} yields the mechanism
3578 used for formal parameter number @var{n} (a static integer value with 1
3579 meaning the first parameter) of @var{subprogram}. The code returned is:
3587 by descriptor (default descriptor class)
3589 by descriptor (UBS: unaligned bit string)
3591 by descriptor (UBSB: aligned bit string with arbitrary bounds)
3593 by descriptor (UBA: unaligned bit array)
3595 by descriptor (S: string, also scalar access type parameter)
3597 by descriptor (SB: string with arbitrary bounds)
3599 by descriptor (A: contiguous array)
3601 by descriptor (NCA: non-contiguous array)
3605 Values from 3 through 10 are only relevant to Digital OpenVMS implementations.
3607 @cindex Zero address, passing
3608 @findex Null_Parameter
3609 @item Null_Parameter
3611 A reference @code{@var{T}'Null_Parameter} denotes an imaginary object of
3612 type or subtype @var{T} allocated at machine address zero. The attribute
3613 is allowed only as the default expression of a formal parameter, or as
3614 an actual expression of a subprogram call. In either case, the
3615 subprogram must be imported.
3617 The identity of the object is represented by the address zero in the
3618 argument list, independent of the passing mechanism (explicit or
3621 This capability is needed to specify that a zero address should be
3622 passed for a record or other composite object passed by reference.
3623 There is no way of indicating this without the @code{Null_Parameter}
3626 @cindex Size, used for objects
3630 The size of an object is not necessarily the same as the size of the type
3631 of an object. This is because by default object sizes are increased to be
3632 a multiple of the alignment of the object. For example,
3633 @code{Natural'Size} is
3634 31, but by default objects of type @code{Natural} will have a size of 32 bits.
3635 Similarly, a record containing an integer and a character:
3645 will have a size of 40 (that is @code{Rec'Size} will be 40. The
3646 alignment will be 4, because of the
3647 integer field, and so the default size of record objects for this type
3648 will be 64 (8 bytes).
3650 The @code{@var{type}'Object_Size} attribute
3651 has been added to GNAT to allow the
3652 default object size of a type to be easily determined. For example,
3653 @code{Natural'Object_Size} is 32, and
3654 @code{Rec'Object_Size} (for the record type in the above example) will be
3655 64. Note also that, unlike the situation with the
3656 @code{Size} attribute as defined in the Ada RM, the
3657 @code{Object_Size} attribute can be specified individually
3658 for different subtypes. For example:
3661 type R is new Integer;
3662 subtype R1 is R range 1 .. 10;
3663 subtype R2 is R range 1 .. 10;
3664 for R2'Object_Size use 8;
3668 In this example, @code{R'Object_Size} and @code{R1'Object_Size} are both
3669 32 since the default object size for a subtype is the same as the object size
3670 for the parent subtype. This means that objects of type @code{R}
3672 by default be 32 bits (four bytes). But objects of type
3673 @code{R2} will be only
3674 8 bits (one byte), since @code{R2'Object_Size} has been set to 8.
3676 @cindex Parameters, when passed by reference
3677 @findex Passed_By_Reference
3678 @item Passed_By_Reference
3680 @code{@var{type}'Passed_By_Reference} for any subtype @var{type} returns
3681 a value of type @code{Boolean} value that is @code{True} if the type is
3682 normally passed by reference and @code{False} if the type is normally
3683 passed by copy in calls. For scalar types, the result is always @code{False}
3684 and is static. For non-scalar types, the result is non-static.
3686 @findex Range_Length
3689 @code{@var{type}'Range_Length} for any discrete type @var{type} yields
3690 the number of values represented by the subtype (zero for a null
3691 range). The result is static for static subtypes. @code{Range_Length}
3692 applied to the index subtype of a one dimensional array always gives the
3693 same result as @code{Range} applied to the array itself.
3695 @cindex Ada 83 attributes
3699 The @code{Safe_Emax} attribute is provided for compatibility with Ada 83. See
3700 the Ada 83 reference manual for an exact description of the semantics of
3703 @cindex Ada 83 attributes
3707 The @code{Safe_Large} attribute is provided for compatibility with Ada 83. See
3708 the Ada 83 reference manual for an exact description of the semantics of
3711 @cindex Ada 83 attributes
3715 The @code{Safe_Large} attribute is provided for compatibility with Ada 83. See
3716 the Ada 83 reference manual for an exact description of the semantics of
3719 @cindex Ada 83 attributes
3723 The @code{Small} attribute is defined in Ada 95 only for fixed-point types.
3724 GNAT also allows this attribute to be applied to floating-point types
3725 for compatibility with Ada 83. See
3726 the Ada 83 reference manual for an exact description of the semantics of
3727 this attribute when applied to floating-point types.
3729 @findex Storage_Unit
3732 @code{Standard'Storage_Unit} (@code{Standard} is the only permissible
3733 prefix) provides the value @code{System.Storage_Unit} and is intended
3734 primarily for constructing this definition in package @code{System}.
3739 @code{Standard'Tick} (@code{Standard} is the only permissible prefix)
3740 provides the value of @code{System.Tick} and is intended primarily for
3741 constructing this definition in package @code{System}.
3746 The @code{System'To_Address}
3747 (@code{System} is the only permissible prefix)
3748 denotes a function identical to
3749 @code{System.Storage_Elements.To_Address} except that
3750 it is a static attribute. This means that if its argument is
3751 a static expression, then the result of the attribute is a
3752 static expression. The result is that such an expression can be
3753 used in contexts (e.g.@: preelaborable packages) which require a
3754 static expression and where the function call could not be used
3755 (since the function call is always non-static, even if its
3756 argument is static).
3761 @code{@var{type}'Type_Class} for any type or subtype @var{type} yields
3762 the value of the type class for the full type of @var{type}. If
3763 @var{type} is a generic formal type, the value is the value for the
3764 corresponding actual subtype. The value of this attribute is of type
3765 @code{System.Aux_DEC.Type_Class}, which has the following definition:
3769 (Type_Class_Enumeration,
3771 Type_Class_Fixed_Point,
3772 Type_Class_Floating_Point,
3777 Type_Class_Address);
3781 Protected types yield the value @code{Type_Class_Task}, which thus
3782 applies to all concurrent types. This attribute is designed to
3783 be compatible with the DEC Ada 83 attribute of the same name.
3788 The @code{UET_Address} attribute can only be used for a prefix which
3789 denotes a library package. It yields the address of the unit exception
3790 table when zero cost exception handling is used. This attribute is
3791 intended only for use within the GNAT implementation. See the unit
3792 @code{Ada.Exceptions} in files @file{a-except.ads} and @file{a-except.adb}
3793 for details on how this attribute is used in the implementation.
3795 @cindex Named numbers, representation of
3796 @findex Universal_Literal_String
3797 @item Universal_Literal_String
3799 The prefix of @code{Universal_Literal_String} must be a named
3800 number. The static result is the string consisting of the characters of
3801 the number as defined in the original source. This allows the user
3802 program to access the actual text of named numbers without intermediate
3803 conversions and without the need to enclose the strings in quotes (which
3804 would preclude their use as numbers). This is used internally for the
3805 construction of values of the floating-point attributes from the file
3806 @file{ttypef.ads}, but may also be used by user programs.
3808 @cindex @code{Access}, unrestricted
3809 @findex Unrestricted_Access
3810 @item Unrestricted_Access
3812 The @code{Unrestricted_Access} attribute is similar to @code{Access}
3813 except that all accessibility and aliased view checks are omitted. This
3814 is a user-beware attribute. It is similar to
3815 @code{Address}, for which it is a desirable replacement where the value
3816 desired is an access type. In other words, its effect is identical to
3817 first applying the @code{Address} attribute and then doing an unchecked
3818 conversion to a desired access type. In GNAT, but not necessarily in
3819 other implementations, the use of static chains for inner level
3820 subprograms means that @code{Unrestricted_Access} applied to a
3821 subprogram yields a value that can be called as long as the subprogram
3822 is in scope (normal Ada 95 accessibility rules restrict this usage).
3824 @cindex @code{Size}, VADS compatibility
3828 The @code{'VADS_Size} attribute is intended to make it easier to port
3829 legacy code which relies on the semantics of @code{'Size} as implemented
3830 by the VADS Ada 83 compiler. GNAT makes a best effort at duplicating the
3831 same semantic interpretation. In particular, @code{'VADS_Size} applied
3832 to a predefined or other primitive type with no Size clause yields the
3833 Object_Size (for example, @code{Natural'Size} is 32 rather than 31 on
3834 typical machines). In addition @code{'VADS_Size} applied to an object
3835 gives the result that would be obtained by applying the attribute to
3836 the corresponding type.
3838 @cindex @code{Size}, setting for not-first subtype
3841 @code{@var{type}'Value_Size} is the number of bits required to represent
3842 a value of the given subtype. It is the same as @code{@var{type}'Size},
3843 but, unlike @code{Size}, may be set for non-first subtypes.
3845 @findex Wchar_T_Size
3847 @code{Standard'Wchar_T_Size} (@code{Standard} is the only permissible
3848 prefix) provides the size in bits of the C @code{wchar_t} type
3849 primarily for constructing the definition of this type in
3850 package @code{Interfaces.C}.
3854 @code{Standard'Word_Size} (@code{Standard} is the only permissible
3855 prefix) provides the value @code{System.Word_Size} and is intended
3856 primarily for constructing this definition in package @code{System}.
3858 @node Implementation Advice
3859 @chapter Implementation Advice
3860 The main text of the Ada 95 Reference Manual describes the required
3861 behavior of all Ada 95 compilers, and the GNAT compiler conforms to
3864 In addition, there are sections throughout the Ada 95
3865 reference manual headed
3866 by the phrase ``implementation advice''. These sections are not normative,
3867 i.e.@: they do not specify requirements that all compilers must
3868 follow. Rather they provide advice on generally desirable behavior. You
3869 may wonder why they are not requirements. The most typical answer is
3870 that they describe behavior that seems generally desirable, but cannot
3871 be provided on all systems, or which may be undesirable on some systems.
3873 As far as practical, GNAT follows the implementation advice sections in
3874 the Ada 95 Reference Manual. This chapter contains a table giving the
3875 reference manual section number, paragraph number and several keywords
3876 for each advice. Each entry consists of the text of the advice followed
3877 by the GNAT interpretation of this advice. Most often, this simply says
3878 ``followed'', which means that GNAT follows the advice. However, in a
3879 number of cases, GNAT deliberately deviates from this advice, in which
3880 case the text describes what GNAT does and why.
3883 @cindex Error detection
3884 @item 1.1.3(20): Error Detection
3887 If an implementation detects the use of an unsupported Specialized Needs
3888 Annex feature at run time, it should raise @code{Program_Error} if
3891 Not relevant. All specialized needs annex features are either supported,
3892 or diagnosed at compile time.
3895 @item 1.1.3(31): Child Units
3898 If an implementation wishes to provide implementation-defined
3899 extensions to the functionality of a language-defined library unit, it
3900 should normally do so by adding children to the library unit.
3904 @cindex Bounded errors
3905 @item 1.1.5(12): Bounded Errors
3908 If an implementation detects a bounded error or erroneous
3909 execution, it should raise @code{Program_Error}.
3911 Followed in all cases in which the implementation detects a bounded
3912 error or erroneous execution. Not all such situations are detected at
3916 @item 2.8(16): Pragmas
3919 Normally, implementation-defined pragmas should have no semantic effect
3920 for error-free programs; that is, if the implementation-defined pragmas
3921 are removed from a working program, the program should still be legal,
3922 and should still have the same semantics.
3924 The following implementation defined pragmas are exceptions to this
3936 @item CPP_Constructor
3944 @item Interface_Name
3946 @item Machine_Attribute
3948 @item Unimplemented_Unit
3950 @item Unchecked_Union
3954 In each of the above cases, it is essential to the purpose of the pragma
3955 that this advice not be followed. For details see the separate section
3956 on implementation defined pragmas.
3958 @item 2.8(17-19): Pragmas
3961 Normally, an implementation should not define pragmas that can
3962 make an illegal program legal, except as follows:
3966 A pragma used to complete a declaration, such as a pragma @code{Import};
3970 A pragma used to configure the environment by adding, removing, or
3971 replacing @code{library_items}.
3973 See response to paragraph 16 of this same section.
3975 @cindex Character Sets
3976 @cindex Alternative Character Sets
3977 @item 3.5.2(5): Alternative Character Sets
3980 If an implementation supports a mode with alternative interpretations
3981 for @code{Character} and @code{Wide_Character}, the set of graphic
3982 characters of @code{Character} should nevertheless remain a proper
3983 subset of the set of graphic characters of @code{Wide_Character}. Any
3984 character set ``localizations'' should be reflected in the results of
3985 the subprograms defined in the language-defined package
3986 @code{Characters.Handling} (see A.3) available in such a mode. In a mode with
3987 an alternative interpretation of @code{Character}, the implementation should
3988 also support a corresponding change in what is a legal
3989 @code{identifier_letter}.
3991 Not all wide character modes follow this advice, in particular the JIS
3992 and IEC modes reflect standard usage in Japan, and in these encoding,
3993 the upper half of the Latin-1 set is not part of the wide-character
3994 subset, since the most significant bit is used for wide character
3995 encoding. However, this only applies to the external forms. Internally
3996 there is no such restriction.
3998 @cindex Integer types
3999 @item 3.5.4(28): Integer Types
4003 An implementation should support @code{Long_Integer} in addition to
4004 @code{Integer} if the target machine supports 32-bit (or longer)
4005 arithmetic. No other named integer subtypes are recommended for package
4006 @code{Standard}. Instead, appropriate named integer subtypes should be
4007 provided in the library package @code{Interfaces} (see B.2).
4009 @code{Long_Integer} is supported. Other standard integer types are supported
4010 so this advice is not fully followed. These types
4011 are supported for convenient interface to C, and so that all hardware
4012 types of the machine are easily available.
4013 @item 3.5.4(29): Integer Types
4017 An implementation for a two's complement machine should support
4018 modular types with a binary modulus up to @code{System.Max_Int*2+2}. An
4019 implementation should support a non-binary modules up to @code{Integer'Last}.
4023 @cindex Enumeration values
4024 @item 3.5.5(8): Enumeration Values
4027 For the evaluation of a call on @code{@var{S}'Pos} for an enumeration
4028 subtype, if the value of the operand does not correspond to the internal
4029 code for any enumeration literal of its type (perhaps due to an
4030 un-initialized variable), then the implementation should raise
4031 @code{Program_Error}. This is particularly important for enumeration
4032 types with noncontiguous internal codes specified by an
4033 enumeration_representation_clause.
4038 @item 3.5.7(17): Float Types
4041 An implementation should support @code{Long_Float} in addition to
4042 @code{Float} if the target machine supports 11 or more digits of
4043 precision. No other named floating point subtypes are recommended for
4044 package @code{Standard}. Instead, appropriate named floating point subtypes
4045 should be provided in the library package @code{Interfaces} (see B.2).
4047 @code{Short_Float} and @code{Long_Long_Float} are also provided. The
4048 former provides improved compatibility with other implementations
4049 supporting this type. The latter corresponds to the highest precision
4050 floating-point type supported by the hardware. On most machines, this
4051 will be the same as @code{Long_Float}, but on some machines, it will
4052 correspond to the IEEE extended form. The notable case is all ia32
4053 (x86) implementations, where @code{Long_Long_Float} corresponds to
4054 the 80-bit extended precision format supported in hardware on this
4055 processor. Note that the 128-bit format on SPARC is not supported,
4056 since this is a software rather than a hardware format.
4058 @cindex Multidimensional arrays
4059 @cindex Arrays, multidimensional
4060 @item 3.6.2(11): Multidimensional Arrays
4063 An implementation should normally represent multidimensional arrays in
4064 row-major order, consistent with the notation used for multidimensional
4065 array aggregates (see 4.3.3). However, if a pragma @code{Convention}
4066 (@code{Fortran}, @dots{}) applies to a multidimensional array type, then
4067 column-major order should be used instead (see B.5, ``Interfacing with
4072 @findex Duration'Small
4073 @item 9.6(30-31): Duration'Small
4076 Whenever possible in an implementation, the value of @code{Duration'Small}
4077 should be no greater than 100 microseconds.
4079 Followed. (@code{Duration'Small} = 10**(@minus{}9)).
4083 The time base for @code{delay_relative_statements} should be monotonic;
4084 it need not be the same time base as used for @code{Calendar.Clock}.
4088 @item 10.2.1(12): Consistent Representation
4091 In an implementation, a type declared in a pre-elaborated package should
4092 have the same representation in every elaboration of a given version of
4093 the package, whether the elaborations occur in distinct executions of
4094 the same program, or in executions of distinct programs or partitions
4095 that include the given version.
4097 Followed, except in the case of tagged types. Tagged types involve
4098 implicit pointers to a local copy of a dispatch table, and these pointers
4099 have representations which thus depend on a particular elaboration of the
4100 package. It is not easy to see how it would be possible to follow this
4101 advice without severely impacting efficiency of execution.
4103 @cindex Exception information
4104 @item 11.4.1(19): Exception Information
4107 @code{Exception_Message} by default and @code{Exception_Information}
4108 should produce information useful for
4109 debugging. @code{Exception_Message} should be short, about one
4110 line. @code{Exception_Information} can be long. @code{Exception_Message}
4111 should not include the
4112 @code{Exception_Name}. @code{Exception_Information} should include both
4113 the @code{Exception_Name} and the @code{Exception_Message}.
4115 Followed. For each exception that doesn't have a specified
4116 @code{Exception_Message}, the compiler generates one containing the location
4117 of the raise statement. This location has the form ``file:line'', where
4118 file is the short file name (without path information) and line is the line
4119 number in the file. Note that in the case of the Zero Cost Exception
4120 mechanism, these messages become redundant with the Exception_Information that
4121 contains a full backtrace of the calling sequence, so they are disabled.
4122 To disable explicitly the generation of the source location message, use the
4123 Pragma @code{Discard_Names}.
4125 @cindex Suppression of checks
4126 @cindex Checks, suppression of
4127 @item 11.5(28): Suppression of Checks
4130 The implementation should minimize the code executed for checks that
4131 have been suppressed.
4135 @cindex Representation clauses
4136 @item 13.1 (21-24): Representation Clauses
4139 The recommended level of support for all representation items is
4140 qualified as follows:
4144 An implementation need not support representation items containing
4145 non-static expressions, except that an implementation should support a
4146 representation item for a given entity if each non-static expression in
4147 the representation item is a name that statically denotes a constant
4148 declared before the entity.
4150 Followed. GNAT does not support non-static expressions in representation
4151 clauses unless they are constants declared before the entity. For
4156 for X'Address use To_address (16#2000#);
4160 will be rejected, since the To_Address expression is non-static. Instead
4164 X_Address : constant Address : =
4165 To_Address ((16#2000#);
4167 for X'Address use X_Address;
4172 An implementation need not support a specification for the @code{Size}
4173 for a given composite subtype, nor the size or storage place for an
4174 object (including a component) of a given composite subtype, unless the
4175 constraints on the subtype and its composite subcomponents (if any) are
4176 all static constraints.
4178 Followed. Size Clauses are not permitted on non-static components, as
4183 An aliased component, or a component whose type is by-reference, should
4184 always be allocated at an addressable location.
4188 @cindex Packed types
4189 @item 13.2(6-8): Packed Types
4192 If a type is packed, then the implementation should try to minimize
4193 storage allocated to objects of the type, possibly at the expense of
4194 speed of accessing components, subject to reasonable complexity in
4195 addressing calculations.
4199 The recommended level of support pragma @code{Pack} is:
4201 For a packed record type, the components should be packed as tightly as
4202 possible subject to the Sizes of the component subtypes, and subject to
4203 any @code{record_representation_clause} that applies to the type; the
4204 implementation may, but need not, reorder components or cross aligned
4205 word boundaries to improve the packing. A component whose @code{Size} is
4206 greater than the word size may be allocated an integral number of words.
4208 Followed. Tight packing of arrays is supported for all component sizes
4213 An implementation should support Address clauses for imported
4217 @cindex @code{Address} clauses
4218 @item 13.3(14-19): Address Clauses
4222 For an array @var{X}, @code{@var{X}'Address} should point at the first
4223 component of the array, and not at the array bounds.
4229 The recommended level of support for the @code{Address} attribute is:
4231 @code{@var{X}'Address} should produce a useful result if @var{X} is an
4232 object that is aliased or of a by-reference type, or is an entity whose
4233 @code{Address} has been specified.
4235 Followed. A valid address will be produced even if none of those
4236 conditions have been met. If necessary, the object is forced into
4237 memory to ensure the address is valid.
4241 An implementation should support @code{Address} clauses for imported
4248 Objects (including subcomponents) that are aliased or of a by-reference
4249 type should be allocated on storage element boundaries.
4255 If the @code{Address} of an object is specified, or it is imported or exported,
4256 then the implementation should not perform optimizations based on
4257 assumptions of no aliases.
4261 @cindex @code{Alignment} clauses
4262 @item 13.3(29-35): Alignment Clauses
4265 The recommended level of support for the @code{Alignment} attribute for
4268 An implementation should support specified Alignments that are factors
4269 and multiples of the number of storage elements per word, subject to the
4276 An implementation need not support specified @code{Alignment}s for
4277 combinations of @code{Size}s and @code{Alignment}s that cannot be easily
4278 loaded and stored by available machine instructions.
4284 An implementation need not support specified @code{Alignment}s that are
4285 greater than the maximum @code{Alignment} the implementation ever returns by
4292 The recommended level of support for the @code{Alignment} attribute for
4295 Same as above, for subtypes, but in addition:
4301 For stand-alone library-level objects of statically constrained
4302 subtypes, the implementation should support all @code{Alignment}s
4303 supported by the target linker. For example, page alignment is likely to
4304 be supported for such objects, but not for subtypes.
4308 @cindex @code{Size} clauses
4309 @item 13.3(42-43): Size Clauses
4312 The recommended level of support for the @code{Size} attribute of
4315 A @code{Size} clause should be supported for an object if the specified
4316 @code{Size} is at least as large as its subtype's @code{Size}, and
4317 corresponds to a size in storage elements that is a multiple of the
4318 object's @code{Alignment} (if the @code{Alignment} is nonzero).
4322 @item 13.3(50-56): Size Clauses
4325 If the @code{Size} of a subtype is specified, and allows for efficient
4326 independent addressability (see 9.10) on the target architecture, then
4327 the @code{Size} of the following objects of the subtype should equal the
4328 @code{Size} of the subtype:
4330 Aliased objects (including components).
4336 @code{Size} clause on a composite subtype should not affect the
4337 internal layout of components.
4343 The recommended level of support for the @code{Size} attribute of subtypes is:
4347 The @code{Size} (if not specified) of a static discrete or fixed point
4348 subtype should be the number of bits needed to represent each value
4349 belonging to the subtype using an unbiased representation, leaving space
4350 for a sign bit only if the subtype contains negative values. If such a
4351 subtype is a first subtype, then an implementation should support a
4352 specified @code{Size} for it that reflects this representation.
4358 For a subtype implemented with levels of indirection, the @code{Size}
4359 should include the size of the pointers, but not the size of what they
4364 @cindex @code{Component_Size} clauses
4365 @item 13.3(71-73): Component Size Clauses
4368 The recommended level of support for the @code{Component_Size}
4373 An implementation need not support specified @code{Component_Sizes} that are
4374 less than the @code{Size} of the component subtype.
4380 An implementation should support specified @code{Component_Size}s that
4381 are factors and multiples of the word size. For such
4382 @code{Component_Size}s, the array should contain no gaps between
4383 components. For other @code{Component_Size}s (if supported), the array
4384 should contain no gaps between components when packing is also
4385 specified; the implementation should forbid this combination in cases
4386 where it cannot support a no-gaps representation.
4390 @cindex Enumeration representation clauses
4391 @cindex Representation clauses, enumeration
4392 @item 13.4(9-10): Enumeration Representation Clauses
4395 The recommended level of support for enumeration representation clauses
4398 An implementation need not support enumeration representation clauses
4399 for boolean types, but should at minimum support the internal codes in
4400 the range @code{System.Min_Int.System.Max_Int}.
4404 @cindex Record representation clauses
4405 @cindex Representation clauses, records
4406 @item 13.5.1(17-22): Record Representation Clauses
4409 The recommended level of support for
4410 @*@code{record_representation_clauses} is:
4412 An implementation should support storage places that can be extracted
4413 with a load, mask, shift sequence of machine code, and set with a load,
4414 shift, mask, store sequence, given the available machine instructions
4421 A storage place should be supported if its size is equal to the
4422 @code{Size} of the component subtype, and it starts and ends on a
4423 boundary that obeys the @code{Alignment} of the component subtype.
4429 If the default bit ordering applies to the declaration of a given type,
4430 then for a component whose subtype's @code{Size} is less than the word
4431 size, any storage place that does not cross an aligned word boundary
4432 should be supported.
4438 An implementation may reserve a storage place for the tag field of a
4439 tagged type, and disallow other components from overlapping that place.
4441 Followed. The storage place for the tag field is the beginning of the tagged
4442 record, and its size is Address'Size. GNAT will reject an explicit component
4443 clause for the tag field.
4447 An implementation need not support a @code{component_clause} for a
4448 component of an extension part if the storage place is not after the
4449 storage places of all components of the parent type, whether or not
4450 those storage places had been specified.
4452 Followed. The above advice on record representation clauses is followed,
4453 and all mentioned features are implemented.
4455 @cindex Storage place attributes
4456 @item 13.5.2(5): Storage Place Attributes
4459 If a component is represented using some form of pointer (such as an
4460 offset) to the actual data of the component, and this data is contiguous
4461 with the rest of the object, then the storage place attributes should
4462 reflect the place of the actual data, not the pointer. If a component is
4463 allocated discontinuously from the rest of the object, then a warning
4464 should be generated upon reference to one of its storage place
4467 Followed. There are no such components in GNAT@.
4469 @cindex Bit ordering
4470 @item 13.5.3(7-8): Bit Ordering
4473 The recommended level of support for the non-default bit ordering is:
4477 If @code{Word_Size} = @code{Storage_Unit}, then the implementation
4478 should support the non-default bit ordering in addition to the default
4481 Followed. Word size does not equal storage size in this implementation.
4482 Thus non-default bit ordering is not supported.
4484 @cindex @code{Address}, as private type
4485 @item 13.7(37): Address as Private
4488 @code{Address} should be of a private type.
4492 @cindex Operations, on @code{Address}
4493 @cindex @code{Address}, operations of
4494 @item 13.7.1(16): Address Operations
4497 Operations in @code{System} and its children should reflect the target
4498 environment semantics as closely as is reasonable. For example, on most
4499 machines, it makes sense for address arithmetic to ``wrap around''.
4500 Operations that do not make sense should raise @code{Program_Error}.
4502 Followed. Address arithmetic is modular arithmetic that wraps around. No
4503 operation raises @code{Program_Error}, since all operations make sense.
4505 @cindex Unchecked conversion
4506 @item 13.9(14-17): Unchecked Conversion
4509 The @code{Size} of an array object should not include its bounds; hence,
4510 the bounds should not be part of the converted data.
4516 The implementation should not generate unnecessary run-time checks to
4517 ensure that the representation of @var{S} is a representation of the
4518 target type. It should take advantage of the permission to return by
4519 reference when possible. Restrictions on unchecked conversions should be
4520 avoided unless required by the target environment.
4522 Followed. There are no restrictions on unchecked conversion. A warning is
4523 generated if the source and target types do not have the same size since
4524 the semantics in this case may be target dependent.
4528 The recommended level of support for unchecked conversions is:
4532 Unchecked conversions should be supported and should be reversible in
4533 the cases where this clause defines the result. To enable meaningful use
4534 of unchecked conversion, a contiguous representation should be used for
4535 elementary subtypes, for statically constrained array subtypes whose
4536 component subtype is one of the subtypes described in this paragraph,
4537 and for record subtypes without discriminants whose component subtypes
4538 are described in this paragraph.
4542 @cindex Heap usage, implicit
4543 @item 13.11(23-25): Implicit Heap Usage
4546 An implementation should document any cases in which it dynamically
4547 allocates heap storage for a purpose other than the evaluation of an
4550 Followed, the only other points at which heap storage is dynamically
4551 allocated are as follows:
4555 At initial elaboration time, to allocate dynamically sized global
4559 To allocate space for a task when a task is created.
4562 To extend the secondary stack dynamically when needed. The secondary
4563 stack is used for returning variable length results.
4568 A default (implementation-provided) storage pool for an
4569 access-to-constant type should not have overhead to support deallocation of
4576 A storage pool for an anonymous access type should be created at the
4577 point of an allocator for the type, and be reclaimed when the designated
4578 object becomes inaccessible.
4582 @cindex Unchecked deallocation
4583 @item 13.11.2(17): Unchecked De-allocation
4586 For a standard storage pool, @code{Free} should actually reclaim the
4591 @cindex Stream oriented attributes
4592 @item 13.13.2(17): Stream Oriented Attributes
4595 If a stream element is the same size as a storage element, then the
4596 normal in-memory representation should be used by @code{Read} and
4597 @code{Write} for scalar objects. Otherwise, @code{Read} and @code{Write}
4598 should use the smallest number of stream elements needed to represent
4599 all values in the base range of the scalar type.
4601 Followed. In particular, the interpretation chosen is that of AI-195,
4602 which specifies that the size to be used is that of the first subtype.
4604 @item A.1(52): Implementation Advice
4607 If an implementation provides additional named predefined integer types,
4608 then the names should end with @samp{Integer} as in
4609 @samp{Long_Integer}. If an implementation provides additional named
4610 predefined floating point types, then the names should end with
4611 @samp{Float} as in @samp{Long_Float}.
4615 @findex Ada.Characters.Handling
4616 @item A.3.2(49): @code{Ada.Characters.Handling}
4619 If an implementation provides a localized definition of @code{Character}
4620 or @code{Wide_Character}, then the effects of the subprograms in
4621 @code{Characters.Handling} should reflect the localizations. See also
4624 Followed. GNAT provides no such localized definitions.
4626 @cindex Bounded-length strings
4627 @item A.4.4(106): Bounded-Length String Handling
4630 Bounded string objects should not be implemented by implicit pointers
4631 and dynamic allocation.
4633 Followed. No implicit pointers or dynamic allocation are used.
4635 @cindex Random number generation
4636 @item A.5.2(46-47): Random Number Generation
4639 Any storage associated with an object of type @code{Generator} should be
4640 reclaimed on exit from the scope of the object.
4646 If the generator period is sufficiently long in relation to the number
4647 of distinct initiator values, then each possible value of
4648 @code{Initiator} passed to @code{Reset} should initiate a sequence of
4649 random numbers that does not, in a practical sense, overlap the sequence
4650 initiated by any other value. If this is not possible, then the mapping
4651 between initiator values and generator states should be a rapidly
4652 varying function of the initiator value.
4654 Followed. The generator period is sufficiently long for the first
4655 condition here to hold true.
4657 @findex Get_Immediate
4658 @item A.10.7(23): @code{Get_Immediate}
4661 The @code{Get_Immediate} procedures should be implemented with
4662 unbuffered input. For a device such as a keyboard, input should be
4663 @dfn{available} if a key has already been typed, whereas for a disk
4664 file, input should always be available except at end of file. For a file
4665 associated with a keyboard-like device, any line-editing features of the
4666 underlying operating system should be disabled during the execution of
4667 @code{Get_Immediate}.
4672 @item B.1(39-41): Pragma @code{Export}
4675 If an implementation supports pragma @code{Export} to a given language,
4676 then it should also allow the main subprogram to be written in that
4677 language. It should support some mechanism for invoking the elaboration
4678 of the Ada library units included in the system, and for invoking the
4679 finalization of the environment task. On typical systems, the
4680 recommended mechanism is to provide two subprograms whose link names are
4681 @code{adainit} and @code{adafinal}. @code{adainit} should contain the
4682 elaboration code for library units. @code{adafinal} should contain the
4683 finalization code. These subprograms should have no effect the second
4684 and subsequent time they are called.
4690 Automatic elaboration of pre-elaborated packages should be
4691 provided when pragma @code{Export} is supported.
4693 Followed when the main program is in Ada. If the main program is in a
4694 foreign language, then
4695 @code{adainit} must be called to elaborate pre-elaborated
4700 For each supported convention @var{L} other than @code{Intrinsic}, an
4701 implementation should support @code{Import} and @code{Export} pragmas
4702 for objects of @var{L}-compatible types and for subprograms, and pragma
4703 @code{Convention} for @var{L}-eligible types and for subprograms,
4704 presuming the other language has corresponding features. Pragma
4705 @code{Convention} need not be supported for scalar types.
4709 @cindex Package @code{Interfaces}
4711 @item B.2(12-13): Package @code{Interfaces}
4714 For each implementation-defined convention identifier, there should be a
4715 child package of package Interfaces with the corresponding name. This
4716 package should contain any declarations that would be useful for
4717 interfacing to the language (implementation) represented by the
4718 convention. Any declarations useful for interfacing to any language on
4719 the given hardware architecture should be provided directly in
4722 Followed. An additional package not defined
4723 in the Ada 95 Reference Manual is @code{Interfaces.CPP}, used
4724 for interfacing to C++.
4728 An implementation supporting an interface to C, COBOL, or Fortran should
4729 provide the corresponding package or packages described in the following
4732 Followed. GNAT provides all the packages described in this section.
4734 @cindex C, interfacing with
4735 @item B.3(63-71): Interfacing with C
4738 An implementation should support the following interface correspondences
4745 An Ada procedure corresponds to a void-returning C function.
4751 An Ada function corresponds to a non-void C function.
4757 An Ada @code{in} scalar parameter is passed as a scalar argument to a C
4764 An Ada @code{in} parameter of an access-to-object type with designated
4765 type @var{T} is passed as a @code{@var{t}*} argument to a C function,
4766 where @var{t} is the C type corresponding to the Ada type @var{T}.
4772 An Ada access @var{T} parameter, or an Ada @code{out} or @code{in out}
4773 parameter of an elementary type @var{T}, is passed as a @code{@var{t}*}
4774 argument to a C function, where @var{t} is the C type corresponding to
4775 the Ada type @var{T}. In the case of an elementary @code{out} or
4776 @code{in out} parameter, a pointer to a temporary copy is used to
4777 preserve by-copy semantics.
4783 An Ada parameter of a record type @var{T}, of any mode, is passed as a
4784 @code{@var{t}*} argument to a C function, where @var{t} is the C
4785 structure corresponding to the Ada type @var{T}.
4787 Followed. This convention may be overridden by the use of the C_Pass_By_Copy
4788 pragma, or Convention, or by explicitly specifying the mechanism for a given
4789 call using an extended import or export pragma.
4793 An Ada parameter of an array type with component type @var{T}, of any
4794 mode, is passed as a @code{@var{t}*} argument to a C function, where
4795 @var{t} is the C type corresponding to the Ada type @var{T}.
4801 An Ada parameter of an access-to-subprogram type is passed as a pointer
4802 to a C function whose prototype corresponds to the designated
4803 subprogram's specification.
4807 @cindex COBOL, interfacing with
4808 @item B.4(95-98): Interfacing with COBOL
4811 An Ada implementation should support the following interface
4812 correspondences between Ada and COBOL@.
4818 An Ada access @var{T} parameter is passed as a @samp{BY REFERENCE} data item of
4819 the COBOL type corresponding to @var{T}.
4825 An Ada in scalar parameter is passed as a @samp{BY CONTENT} data item of
4826 the corresponding COBOL type.
4832 Any other Ada parameter is passed as a @samp{BY REFERENCE} data item of the
4833 COBOL type corresponding to the Ada parameter type; for scalars, a local
4834 copy is used if necessary to ensure by-copy semantics.
4838 @cindex Fortran, interfacing with
4839 @item B.5(22-26): Interfacing with Fortran
4842 An Ada implementation should support the following interface
4843 correspondences between Ada and Fortran:
4849 An Ada procedure corresponds to a Fortran subroutine.
4855 An Ada function corresponds to a Fortran function.
4861 An Ada parameter of an elementary, array, or record type @var{T} is
4862 passed as a @var{T} argument to a Fortran procedure, where @var{T} is
4863 the Fortran type corresponding to the Ada type @var{T}, and where the
4864 INTENT attribute of the corresponding dummy argument matches the Ada
4865 formal parameter mode; the Fortran implementation's parameter passing
4866 conventions are used. For elementary types, a local copy is used if
4867 necessary to ensure by-copy semantics.
4873 An Ada parameter of an access-to-subprogram type is passed as a
4874 reference to a Fortran procedure whose interface corresponds to the
4875 designated subprogram's specification.
4879 @cindex Machine operations
4880 @item C.1(3-5): Access to Machine Operations
4883 The machine code or intrinsic support should allow access to all
4884 operations normally available to assembly language programmers for the
4885 target environment, including privileged instructions, if any.
4891 The interfacing pragmas (see Annex B) should support interface to
4892 assembler; the default assembler should be associated with the
4893 convention identifier @code{Assembler}.
4899 If an entity is exported to assembly language, then the implementation
4900 should allocate it at an addressable location, and should ensure that it
4901 is retained by the linking process, even if not otherwise referenced
4902 from the Ada code. The implementation should assume that any call to a
4903 machine code or assembler subprogram is allowed to read or update every
4904 object that is specified as exported.
4908 @item C.1(10-16): Access to Machine Operations
4911 The implementation should ensure that little or no overhead is
4912 associated with calling intrinsic and machine-code subprograms.
4914 Followed for both intrinsics and machine-code subprograms.
4918 It is recommended that intrinsic subprograms be provided for convenient
4919 access to any machine operations that provide special capabilities or
4920 efficiency and that are not otherwise available through the language
4923 Followed. A full set of machine operation intrinsic subprograms is provided.
4927 Atomic read-modify-write operations---e.g.@:, test and set, compare and
4928 swap, decrement and test, enqueue/dequeue.
4930 Followed on any target supporting such operations.
4934 Standard numeric functions---e.g.@:, sin, log.
4936 Followed on any target supporting such operations.
4940 String manipulation operations---e.g.@:, translate and test.
4942 Followed on any target supporting such operations.
4946 Vector operations---e.g.@:, compare vector against thresholds.
4948 Followed on any target supporting such operations.
4952 Direct operations on I/O ports.
4954 Followed on any target supporting such operations.
4956 @cindex Interrupt support
4957 @item C.3(28): Interrupt Support
4960 If the @code{Ceiling_Locking} policy is not in effect, the
4961 implementation should provide means for the application to specify which
4962 interrupts are to be blocked during protected actions, if the underlying
4963 system allows for a finer-grain control of interrupt blocking.
4965 Followed. The underlying system does not allow for finer-grain control
4966 of interrupt blocking.
4968 @cindex Protected procedure handlers
4969 @item C.3.1(20-21): Protected Procedure Handlers
4972 Whenever possible, the implementation should allow interrupt handlers to
4973 be called directly by the hardware.
4977 This is never possible under IRIX, so this is followed by default.
4979 Followed on any target where the underlying operating system permits
4984 Whenever practical, violations of any
4985 implementation-defined restrictions should be detected before run time.
4987 Followed. Compile time warnings are given when possible.
4989 @cindex Package @code{Interrupts}
4991 @item C.3.2(25): Package @code{Interrupts}
4995 If implementation-defined forms of interrupt handler procedures are
4996 supported, such as protected procedures with parameters, then for each
4997 such form of a handler, a type analogous to @code{Parameterless_Handler}
4998 should be specified in a child package of @code{Interrupts}, with the
4999 same operations as in the predefined package Interrupts.
5003 @cindex Pre-elaboration requirements
5004 @item C.4(14): Pre-elaboration Requirements
5007 It is recommended that pre-elaborated packages be implemented in such a
5008 way that there should be little or no code executed at run time for the
5009 elaboration of entities not already covered by the Implementation
5012 Followed. Executable code is generated in some cases, e.g.@: loops
5013 to initialize large arrays.
5015 @item C.5(8): Pragma @code{Discard_Names}
5019 If the pragma applies to an entity, then the implementation should
5020 reduce the amount of storage used for storing names associated with that
5025 @cindex Package @code{Task_Attributes}
5026 @findex Task_Attributes
5027 @item C.7.2(30): The Package Task_Attributes
5030 Some implementations are targeted to domains in which memory use at run
5031 time must be completely deterministic. For such implementations, it is
5032 recommended that the storage for task attributes will be pre-allocated
5033 statically and not from the heap. This can be accomplished by either
5034 placing restrictions on the number and the size of the task's
5035 attributes, or by using the pre-allocated storage for the first @var{N}
5036 attribute objects, and the heap for the others. In the latter case,
5037 @var{N} should be documented.
5039 Not followed. This implementation is not targeted to such a domain.
5041 @cindex Locking Policies
5042 @item D.3(17): Locking Policies
5046 The implementation should use names that end with @samp{_Locking} for
5047 locking policies defined by the implementation.
5049 Followed. A single implementation-defined locking policy is defined,
5050 whose name (@code{Inheritance_Locking}) follows this suggestion.
5052 @cindex Entry queuing policies
5053 @item D.4(16): Entry Queuing Policies
5056 Names that end with @samp{_Queuing} should be used
5057 for all implementation-defined queuing policies.
5059 Followed. No such implementation-defined queueing policies exist.
5061 @cindex Preemptive abort
5062 @item D.6(9-10): Preemptive Abort
5065 Even though the @code{abort_statement} is included in the list of
5066 potentially blocking operations (see 9.5.1), it is recommended that this
5067 statement be implemented in a way that never requires the task executing
5068 the @code{abort_statement} to block.
5074 On a multi-processor, the delay associated with aborting a task on
5075 another processor should be bounded; the implementation should use
5076 periodic polling, if necessary, to achieve this.
5080 @cindex Tasking restrictions
5081 @item D.7(21): Tasking Restrictions
5084 When feasible, the implementation should take advantage of the specified
5085 restrictions to produce a more efficient implementation.
5087 GNAT currently takes advantage of these restrictions by providing an optimized
5088 run time when the Ravenscar profile and the GNAT restricted run time set
5089 of restrictions are specified. See pragma @code{Ravenscar} and pragma
5090 @code{Restricted_Run_Time} for more details.
5092 @cindex Time, monotonic
5093 @item D.8(47-49): Monotonic Time
5096 When appropriate, implementations should provide configuration
5097 mechanisms to change the value of @code{Tick}.
5099 Such configuration mechanisms are not appropriate to this implementation
5100 and are thus not supported.
5104 It is recommended that @code{Calendar.Clock} and @code{Real_Time.Clock}
5105 be implemented as transformations of the same time base.
5111 It is recommended that the @dfn{best} time base which exists in
5112 the underlying system be available to the application through
5113 @code{Clock}. @dfn{Best} may mean highest accuracy or largest range.
5117 @cindex Partition communication subsystem
5119 @item E.5(28-29): Partition Communication Subsystem
5122 Whenever possible, the PCS on the called partition should allow for
5123 multiple tasks to call the RPC-receiver with different messages and
5124 should allow them to block until the corresponding subprogram body
5127 Followed by GLADE, a separately supplied PCS that can be used with
5132 The @code{Write} operation on a stream of type @code{Params_Stream_Type}
5133 should raise @code{Storage_Error} if it runs out of space trying to
5134 write the @code{Item} into the stream.
5136 Followed by GLADE, a separately supplied PCS that can be used with
5139 @cindex COBOL support
5140 @item F(7): COBOL Support
5143 If COBOL (respectively, C) is widely supported in the target
5144 environment, implementations supporting the Information Systems Annex
5145 should provide the child package @code{Interfaces.COBOL} (respectively,
5146 @code{Interfaces.C}) specified in Annex B and should support a
5147 @code{convention_identifier} of COBOL (respectively, C) in the interfacing
5148 pragmas (see Annex B), thus allowing Ada programs to interface with
5149 programs written in that language.
5153 @cindex Decimal radix support
5154 @item F.1(2): Decimal Radix Support
5157 Packed decimal should be used as the internal representation for objects
5158 of subtype @var{S} when @var{S}'Machine_Radix = 10.
5160 Not followed. GNAT ignores @var{S}'Machine_Radix and always uses binary
5167 If Fortran (respectively, C) is widely supported in the target
5168 environment, implementations supporting the Numerics Annex
5169 should provide the child package @code{Interfaces.Fortran} (respectively,
5170 @code{Interfaces.C}) specified in Annex B and should support a
5171 @code{convention_identifier} of Fortran (respectively, C) in the interfacing
5172 pragmas (see Annex B), thus allowing Ada programs to interface with
5173 programs written in that language.
5177 @cindex Complex types
5178 @item G.1.1(56-58): Complex Types
5181 Because the usual mathematical meaning of multiplication of a complex
5182 operand and a real operand is that of the scaling of both components of
5183 the former by the latter, an implementation should not perform this
5184 operation by first promoting the real operand to complex type and then
5185 performing a full complex multiplication. In systems that, in the
5186 future, support an Ada binding to IEC 559:1989, the latter technique
5187 will not generate the required result when one of the components of the
5188 complex operand is infinite. (Explicit multiplication of the infinite
5189 component by the zero component obtained during promotion yields a NaN
5190 that propagates into the final result.) Analogous advice applies in the
5191 case of multiplication of a complex operand and a pure-imaginary
5192 operand, and in the case of division of a complex operand by a real or
5193 pure-imaginary operand.
5199 Similarly, because the usual mathematical meaning of addition of a
5200 complex operand and a real operand is that the imaginary operand remains
5201 unchanged, an implementation should not perform this operation by first
5202 promoting the real operand to complex type and then performing a full
5203 complex addition. In implementations in which the @code{Signed_Zeros}
5204 attribute of the component type is @code{True} (and which therefore
5205 conform to IEC 559:1989 in regard to the handling of the sign of zero in
5206 predefined arithmetic operations), the latter technique will not
5207 generate the required result when the imaginary component of the complex
5208 operand is a negatively signed zero. (Explicit addition of the negative
5209 zero to the zero obtained during promotion yields a positive zero.)
5210 Analogous advice applies in the case of addition of a complex operand
5211 and a pure-imaginary operand, and in the case of subtraction of a
5212 complex operand and a real or pure-imaginary operand.
5218 Implementations in which @code{Real'Signed_Zeros} is @code{True} should
5219 attempt to provide a rational treatment of the signs of zero results and
5220 result components. As one example, the result of the @code{Argument}
5221 function should have the sign of the imaginary component of the
5222 parameter @code{X} when the point represented by that parameter lies on
5223 the positive real axis; as another, the sign of the imaginary component
5224 of the @code{Compose_From_Polar} function should be the same as
5225 (respectively, the opposite of) that of the @code{Argument} parameter when that
5226 parameter has a value of zero and the @code{Modulus} parameter has a
5227 nonnegative (respectively, negative) value.
5231 @cindex Complex elementary functions
5232 @item G.1.2(49): Complex Elementary Functions
5235 Implementations in which @code{Complex_Types.Real'Signed_Zeros} is
5236 @code{True} should attempt to provide a rational treatment of the signs
5237 of zero results and result components. For example, many of the complex
5238 elementary functions have components that are odd functions of one of
5239 the parameter components; in these cases, the result component should
5240 have the sign of the parameter component at the origin. Other complex
5241 elementary functions have zero components whose sign is opposite that of
5242 a parameter component at the origin, or is always positive or always
5247 @cindex Accuracy requirements
5248 @item G.2.4(19): Accuracy Requirements
5251 The versions of the forward trigonometric functions without a
5252 @code{Cycle} parameter should not be implemented by calling the
5253 corresponding version with a @code{Cycle} parameter of
5254 @code{2.0*Numerics.Pi}, since this will not provide the required
5255 accuracy in some portions of the domain. For the same reason, the
5256 version of @code{Log} without a @code{Base} parameter should not be
5257 implemented by calling the corresponding version with a @code{Base}
5258 parameter of @code{Numerics.e}.
5262 @cindex Complex arithmetic accuracy
5263 @cindex Accuracy, complex arithmetic
5264 @item G.2.6(15): Complex Arithmetic Accuracy
5268 The version of the @code{Compose_From_Polar} function without a
5269 @code{Cycle} parameter should not be implemented by calling the
5270 corresponding version with a @code{Cycle} parameter of
5271 @code{2.0*Numerics.Pi}, since this will not provide the required
5272 accuracy in some portions of the domain.
5277 @node Implementation Defined Characteristics
5278 @chapter Implementation Defined Characteristics
5279 In addition to the implementation dependent pragmas and attributes, and
5280 the implementation advice, there are a number of other features of Ada
5281 95 that are potentially implementation dependent. These are mentioned
5282 throughout the Ada 95 Reference Manual, and are summarized in annex M@.
5284 A requirement for conforming Ada compilers is that they provide
5285 documentation describing how the implementation deals with each of these
5286 issues. In this chapter, you will find each point in annex M listed
5287 followed by a description in italic font of how GNAT
5291 implementation on IRIX 5.3 operating system or greater
5293 handles the implementation dependence.
5295 You can use this chapter as a guide to minimizing implementation
5296 dependent features in your programs if portability to other compilers
5297 and other operating systems is an important consideration. The numbers
5298 in each section below correspond to the paragraph number in the Ada 95
5304 @strong{2}. Whether or not each recommendation given in Implementation
5305 Advice is followed. See 1.1.2(37).
5308 @xref{Implementation Advice}.
5313 @strong{3}. Capacity limitations of the implementation. See 1.1.3(3).
5316 The complexity of programs that can be processed is limited only by the
5317 total amount of available virtual memory, and disk space for the
5318 generated object files.
5323 @strong{4}. Variations from the standard that are impractical to avoid
5324 given the implementation's execution environment. See 1.1.3(6).
5327 There are no variations from the standard.
5332 @strong{5}. Which @code{code_statement}s cause external
5333 interactions. See 1.1.3(10).
5336 Any @code{code_statement} can potentially cause external interactions.
5341 @strong{6}. The coded representation for the text of an Ada
5342 program. See 2.1(4).
5345 See separate section on source representation.
5350 @strong{7}. The control functions allowed in comments. See 2.1(14).
5353 See separate section on source representation.
5358 @strong{8}. The representation for an end of line. See 2.2(2).
5361 See separate section on source representation.
5366 @strong{9}. Maximum supported line length and lexical element
5367 length. See 2.2(15).
5370 The maximum line length is 255 characters an the maximum length of a
5371 lexical element is also 255 characters.
5376 @strong{10}. Implementation defined pragmas. See 2.8(14).
5380 @xref{Implementation Defined Pragmas}.
5385 @strong{11}. Effect of pragma @code{Optimize}. See 2.8(27).
5388 Pragma @code{Optimize}, if given with a @code{Time} or @code{Space}
5389 parameter, checks that the optimization flag is set, and aborts if it is
5395 @strong{12}. The sequence of characters of the value returned by
5396 @code{@var{S}'Image} when some of the graphic characters of
5397 @code{@var{S}'Wide_Image} are not defined in @code{Character}. See
5401 The sequence of characters is as defined by the wide character encoding
5402 method used for the source. See section on source representation for
5408 @strong{13}. The predefined integer types declared in
5409 @code{Standard}. See 3.5.4(25).
5413 @item Short_Short_Integer
5416 (Short) 16 bit signed
5420 64 bit signed (Alpha OpenVMS only)
5421 32 bit signed (all other targets)
5422 @item Long_Long_Integer
5429 @strong{14}. Any nonstandard integer types and the operators defined
5430 for them. See 3.5.4(26).
5433 There are no nonstandard integer types.
5438 @strong{15}. Any nonstandard real types and the operators defined for
5442 There are no nonstandard real types.
5447 @strong{16}. What combinations of requested decimal precision and range
5448 are supported for floating point types. See 3.5.7(7).
5451 The precision and range is as defined by the IEEE standard.
5456 @strong{17}. The predefined floating point types declared in
5457 @code{Standard}. See 3.5.7(16).
5464 (Short) 32 bit IEEE short
5467 @item Long_Long_Float
5468 64 bit IEEE long (80 bit IEEE long on x86 processors)
5474 @strong{18}. The small of an ordinary fixed point type. See 3.5.9(8).
5477 @code{Fine_Delta} is 2**(@minus{}63)
5482 @strong{19}. What combinations of small, range, and digits are
5483 supported for fixed point types. See 3.5.9(10).
5486 Any combinations are permitted that do not result in a small less than
5487 @code{Fine_Delta} and do not result in a mantissa larger than 63 bits.
5488 If the mantissa is larger than 53 bits on machines where Long_Long_Float
5489 is 64 bits (true of all architectures except ia32), then the output from
5490 Text_IO is accurate to only 53 bits, rather than the full mantissa. This
5491 is because floating-point conversions are used to convert fixed point.
5496 @strong{20}. The result of @code{Tags.Expanded_Name} for types declared
5497 within an unnamed @code{block_statement}. See 3.9(10).
5500 Block numbers of the form @code{B@var{nnn}}, where @var{nnn} is a
5501 decimal integer are allocated.
5506 @strong{21}. Implementation-defined attributes. See 4.1.4(12).
5509 @xref{Implementation Defined Attributes}.
5514 @strong{22}. Any implementation-defined time types. See 9.6(6).
5517 There are no implementation-defined time types.
5522 @strong{23}. The time base associated with relative delays.
5525 See 9.6(20). The time base used is that provided by the C library
5526 function @code{gettimeofday}.
5531 @strong{24}. The time base of the type @code{Calendar.Time}. See
5535 The time base used is that provided by the C library function
5536 @code{gettimeofday}.
5541 @strong{25}. The time zone used for package @code{Calendar}
5542 operations. See 9.6(24).
5545 The time zone used by package @code{Calendar} is the current system time zone
5546 setting for local time, as accessed by the C library function
5552 @strong{26}. Any limit on @code{delay_until_statements} of
5553 @code{select_statements}. See 9.6(29).
5556 There are no such limits.
5561 @strong{27}. Whether or not two non overlapping parts of a composite
5562 object are independently addressable, in the case where packing, record
5563 layout, or @code{Component_Size} is specified for the object. See
5567 Separate components are independently addressable if they do not share
5568 overlapping storage units.
5573 @strong{28}. The representation for a compilation. See 10.1(2).
5576 A compilation is represented by a sequence of files presented to the
5577 compiler in a single invocation of the @code{gcc} command.
5582 @strong{29}. Any restrictions on compilations that contain multiple
5583 compilation_units. See 10.1(4).
5586 No single file can contain more than one compilation unit, but any
5587 sequence of files can be presented to the compiler as a single
5593 @strong{30}. The mechanisms for creating an environment and for adding
5594 and replacing compilation units. See 10.1.4(3).
5597 See separate section on compilation model.
5602 @strong{31}. The manner of explicitly assigning library units to a
5603 partition. See 10.2(2).
5606 If a unit contains an Ada main program, then the Ada units for the partition
5607 are determined by recursive application of the rules in the Ada Reference
5608 Manual section 10.2(2-6). In other words, the Ada units will be those that
5609 are needed by the main program, and then this definition of need is applied
5610 recursively to those units, and the partition contains the transitive
5611 closure determined by this relationship. In short, all the necessary units
5612 are included, with no need to explicitly specify the list. If additional
5613 units are required, e.g.@: by foreign language units, then all units must be
5614 mentioned in the context clause of one of the needed Ada units.
5616 If the partition contains no main program, or if the main program is in
5617 a language other than Ada, then GNAT
5618 provides the binder options @code{-z} and @code{-n} respectively, and in this case a
5619 list of units can be explicitly supplied to the binder for inclusion in
5620 the partition (all units needed by these units will also be included
5621 automatically). For full details on the use of these options, refer to
5622 the @cite{GNAT User's Guide} sections on Binding and Linking.
5627 @strong{32}. The implementation-defined means, if any, of specifying
5628 which compilation units are needed by a given compilation unit. See
5632 The units needed by a given compilation unit are as defined in
5633 the Ada Reference Manual section 10.2(2-6). There are no
5634 implementation-defined pragmas or other implementation-defined
5635 means for specifying needed units.
5640 @strong{33}. The manner of designating the main subprogram of a
5641 partition. See 10.2(7).
5644 The main program is designated by providing the name of the
5645 corresponding @file{ALI} file as the input parameter to the binder.
5650 @strong{34}. The order of elaboration of @code{library_items}. See
5654 The first constraint on ordering is that it meets the requirements of
5655 chapter 10 of the Ada 95 Reference Manual. This still leaves some
5656 implementation dependent choices, which are resolved by first
5657 elaborating bodies as early as possible (i.e.@: in preference to specs
5658 where there is a choice), and second by evaluating the immediate with
5659 clauses of a unit to determine the probably best choice, and
5660 third by elaborating in alphabetical order of unit names
5661 where a choice still remains.
5666 @strong{35}. Parameter passing and function return for the main
5667 subprogram. See 10.2(21).
5670 The main program has no parameters. It may be a procedure, or a function
5671 returning an integer type. In the latter case, the returned integer
5672 value is the return code of the program.
5677 @strong{36}. The mechanisms for building and running partitions. See
5681 GNAT itself supports programs with only a single partition. The GNATDIST
5682 tool provided with the GLADE package (which also includes an implementation
5683 of the PCS) provides a completely flexible method for building and running
5684 programs consisting of multiple partitions. See the separate GLADE manual
5690 @strong{37}. The details of program execution, including program
5691 termination. See 10.2(25).
5694 See separate section on compilation model.
5699 @strong{38}. The semantics of any non-active partitions supported by the
5700 implementation. See 10.2(28).
5703 Passive partitions are supported on targets where shared memory is
5704 provided by the operating system. See the GLADE reference manual for
5710 @strong{39}. The information returned by @code{Exception_Message}. See
5714 Exception message returns the null string unless a specific message has
5715 been passed by the program.
5720 @strong{40}. The result of @code{Exceptions.Exception_Name} for types
5721 declared within an unnamed @code{block_statement}. See 11.4.1(12).
5724 Blocks have implementation defined names of the form @code{B@var{nnn}}
5725 where @var{nnn} is an integer.
5730 @strong{41}. The information returned by
5731 @code{Exception_Information}. See 11.4.1(13).
5734 @code{Exception_Information} returns a string in the following format:
5737 @emph{Exception_Name:} nnnnn
5738 @emph{Message:} mmmmm
5740 @emph{Call stack traceback locations:}
5741 0xhhhh 0xhhhh 0xhhhh ... 0xhhh
5749 @code{nnnn} is the fully qualified name of the exception in all upper
5750 case letters. This line is always present.
5753 @code{mmmm} is the message (this line present only if message is non-null)
5756 @code{ppp} is the Process Id value as a decimal integer (this line is
5757 present only if the Process Id is non-zero). Currently we are
5758 not making use of this field.
5761 The Call stack traceback locations line and the following values
5762 are present only if at least one traceback location was recorded.
5763 The values are given in C style format, with lower case letters
5764 for a-f, and only as many digits present as are necessary.
5768 The line terminator sequence at the end of each line, including
5769 the last line is a single @code{LF} character (@code{16#0A#}).
5774 @strong{42}. Implementation-defined check names. See 11.5(27).
5777 No implementation-defined check names are supported.
5782 @strong{43}. The interpretation of each aspect of representation. See
5786 See separate section on data representations.
5791 @strong{44}. Any restrictions placed upon representation items. See
5795 See separate section on data representations.
5800 @strong{45}. The meaning of @code{Size} for indefinite subtypes. See
5804 Size for an indefinite subtype is the maximum possible size, except that
5805 for the case of a subprogram parameter, the size of the parameter object
5811 @strong{46}. The default external representation for a type tag. See
5815 The default external representation for a type tag is the fully expanded
5816 name of the type in upper case letters.
5821 @strong{47}. What determines whether a compilation unit is the same in
5822 two different partitions. See 13.3(76).
5825 A compilation unit is the same in two different partitions if and only
5826 if it derives from the same source file.
5831 @strong{48}. Implementation-defined components. See 13.5.1(15).
5834 The only implementation defined component is the tag for a tagged type,
5835 which contains a pointer to the dispatching table.
5840 @strong{49}. If @code{Word_Size} = @code{Storage_Unit}, the default bit
5841 ordering. See 13.5.3(5).
5844 @code{Word_Size} (32) is not the same as @code{Storage_Unit} (8) for this
5845 implementation, so no non-default bit ordering is supported. The default
5846 bit ordering corresponds to the natural endianness of the target architecture.
5851 @strong{50}. The contents of the visible part of package @code{System}
5852 and its language-defined children. See 13.7(2).
5855 See the definition of these packages in files @file{system.ads} and
5856 @file{s-stoele.ads}.
5861 @strong{51}. The contents of the visible part of package
5862 @code{System.Machine_Code}, and the meaning of
5863 @code{code_statements}. See 13.8(7).
5866 See the definition and documentation in file @file{s-maccod.ads}.
5871 @strong{52}. The effect of unchecked conversion. See 13.9(11).
5874 Unchecked conversion between types of the same size
5875 and results in an uninterpreted transmission of the bits from one type
5876 to the other. If the types are of unequal sizes, then in the case of
5877 discrete types, a shorter source is first zero or sign extended as
5878 necessary, and a shorter target is simply truncated on the left.
5879 For all non-discrete types, the source is first copied if necessary
5880 to ensure that the alignment requirements of the target are met, then
5881 a pointer is constructed to the source value, and the result is obtained
5882 by dereferencing this pointer after converting it to be a pointer to the
5888 @strong{53}. The manner of choosing a storage pool for an access type
5889 when @code{Storage_Pool} is not specified for the type. See 13.11(17).
5892 There are 3 different standard pools used by the compiler when
5893 @code{Storage_Pool} is not specified depending whether the type is local
5894 to a subprogram or defined at the library level and whether
5895 @code{Storage_Size}is specified or not. See documentation in the runtime
5896 library units @code{System.Pool_Global}, @code{System.Pool_Size} and
5897 @code{System.Pool_Local} in files @file{s-poosiz.ads},
5898 @file{s-pooglo.ads} and @file{s-pooloc.ads} for full details on the
5904 @strong{54}. Whether or not the implementation provides user-accessible
5905 names for the standard pool type(s). See 13.11(17).
5909 See documentation in the sources of the run time mentioned in paragraph
5910 @strong{53} . All these pools are accessible by means of @code{with}'ing
5916 @strong{55}. The meaning of @code{Storage_Size}. See 13.11(18).
5919 @code{Storage_Size} is measured in storage units, and refers to the
5920 total space available for an access type collection, or to the primary
5921 stack space for a task.
5926 @strong{56}. Implementation-defined aspects of storage pools. See
5930 See documentation in the sources of the run time mentioned in paragraph
5931 @strong{53} for details on GNAT-defined aspects of storage pools.
5936 @strong{57}. The set of restrictions allowed in a pragma
5937 @code{Restrictions}. See 13.12(7).
5940 All RM defined Restriction identifiers are implemented. The following
5941 additional restriction identifiers are provided. There are two separate
5942 lists of implementation dependent restriction identifiers. The first
5943 set requires consistency throughout a partition (in other words, if the
5944 restriction identifier is used for any compilation unit in the partition,
5945 then all compilation units in the partition must obey the restriction.
5949 @item Boolean_Entry_Barriers
5950 @findex Boolean_Entry_Barriers
5951 This restriction ensures at compile time that barriers in entry declarations
5952 for protected types are restricted to references to simple boolean variables
5953 defined in the private part of the protected type. No other form of entry
5954 barriers is permitted. This is one of the restrictions of the Ravenscar
5955 profile for limited tasking (see also pragma @code{Ravenscar}).
5957 @item Max_Entry_Queue_Depth => Expr
5958 @findex Max_Entry_Queue_Depth
5959 This restriction is a declaration that any protected entry compiled in
5960 the scope of the restriction has at most the specified number of
5961 tasks waiting on the entry
5962 at any one time, and so no queue is required. This restriction is not
5963 checked at compile time. A program execution is erroneous if an attempt
5964 is made to queue more than the specified number of tasks on such an entry.
5968 This restriction ensures at compile time that there is no implicit or
5969 explicit dependence on the package @code{Ada.Calendar}.
5971 @item No_Dynamic_Interrupts
5972 @findex No_Dynamic_Interrupts
5973 This restriction ensures at compile time that there is no attempt to
5974 dynamically associate interrupts. Only static association is allowed.
5976 @item No_Enumeration_Maps
5977 @findex No_Enumeration_Maps
5978 This restriction ensures at compile time that no operations requiring
5979 enumeration maps are used (that is Image and Value attributes applied
5980 to enumeration types).
5982 @item No_Entry_Calls_In_Elaboration_Code
5983 @findex No_Entry_Calls_In_Elaboration_Code
5984 This restriction ensures at compile time that no task or protected entry
5985 calls are made during elaboration code. As a result of the use of this
5986 restriction, the compiler can assume that no code past an accept statement
5987 in a task can be executed at elaboration time.
5989 @item No_Exception_Handlers
5990 @findex No_Exception_Handlers
5991 This restriction ensures at compile time that there are no explicit
5994 @item No_Implicit_Conditionals
5995 @findex No_Implicit_Conditionals
5996 This restriction ensures that the generated code does not contain any
5997 implicit conditionals, either by modifying the generated code where possible,
5998 or by rejecting any construct that would otherwise generate an implicit
5999 conditional. The details and use of this restriction are described in
6000 more detail in the High Integrity product documentation.
6002 @item No_Implicit_Loops
6003 @findex No_Implicit_Loops
6004 This restriction ensures that the generated code does not contain any
6005 implicit @code{for} loops, either by modifying
6006 the generated code where possible,
6007 or by rejecting any construct that would otherwise generate an implicit
6008 @code{for} loop. The details and use of this restriction are described in
6009 more detail in the High Integrity product documentation.
6011 @item No_Local_Protected_Objects
6012 @findex No_Local_Protected_Objects
6013 This restriction ensures at compile time that protected objects are
6014 only declared at the library level.
6016 @item No_Protected_Type_Allocators
6017 @findex No_Protected_Type_Allocators
6018 This restriction ensures at compile time that there are no allocator
6019 expressions that attempt to allocate protected objects.
6021 @item No_Secondary_Stack
6022 @findex No_Secondary_Stack
6023 This restriction ensures at compile time that the generated code does not
6024 contain any reference to the secondary stack. The secondary stack is used
6025 to implement functions returning unconstrained objects (arrays or records)
6027 The details and use of this restriction are described in
6028 more detail in the High Integrity product documentation.
6030 @item No_Select_Statements
6031 @findex No_Select_Statements
6032 This restriction ensures at compile time no select statements of any kind
6033 are permitted, that is the keyword @code{select} may not appear.
6034 This is one of the restrictions of the Ravenscar
6035 profile for limited tasking (see also pragma @code{Ravenscar}).
6037 @item No_Standard_Storage_Pools
6038 @findex No_Standard_Storage_Pools
6039 This restriction ensures at compile time that no access types
6040 use the standard default storage pool. Any access type declared must
6041 have an explicit Storage_Pool attribute defined specifying a
6042 user-defined storage pool.
6046 This restriction ensures at compile time that there are no implicit or
6047 explicit dependencies on the package @code{Ada.Streams}.
6049 @item No_Task_Attributes
6050 @findex No_Task_Attributes
6051 This restriction ensures at compile time that there are no implicit or
6052 explicit dependencies on the package @code{Ada.Task_Attributes}.
6054 @item No_Task_Termination
6055 @findex No_Task_Termination
6056 This restriction ensures at compile time that no terminate alternatives
6057 appear in any task body.
6061 This restriction prevents the declaration of tasks or task types throughout
6062 the partition. It is similar in effect to the use of @code{Max_Tasks => 0}
6063 except that violations are caught at compile time and cause an error message
6064 to be output either by the compiler or binder.
6066 @item No_Wide_Characters
6067 @findex No_Wide_Characters
6068 This restriction ensures at compile time that no uses of the types
6069 @code{Wide_Character} or @code{Wide_String}
6070 appear, and that no wide character literals
6071 appear in the program (that is literals representing characters not in
6072 type @code{Character}.
6074 @item Static_Priorities
6075 @findex Static_Priorities
6076 This restriction ensures at compile time that all priority expressions
6077 are static, and that there are no dependencies on the package
6078 @code{Ada.Dynamic_Priorities}.
6080 @item Static_Storage_Size
6081 @findex Static_Storage_Size
6082 This restriction ensures at compile time that any expression appearing
6083 in a Storage_Size pragma or attribute definition clause is static.
6088 The second set of implementation dependent restriction identifiers
6089 does not require partition-wide consistency.
6090 The restriction may be enforced for a single
6091 compilation unit without any effect on any of the
6092 other compilation units in the partition.
6096 @item No_Elaboration_Code
6097 @findex No_Elaboration_Code
6098 This restriction ensures at compile time that no elaboration code is
6099 generated. Note that this is not the same condition as is enforced
6100 by pragma @code{Preelaborate}. There are cases in which pragma @code{Preelaborate}
6101 still permits code to be generated (e.g.@: code to initialize a large
6102 array to all zeroes), and there are cases of units which do not meet
6103 the requirements for pragma @code{Preelaborate}, but for which no elaboration
6104 code is generated. Generally, it is the case that preelaborable units
6105 will meet the restrictions, with the exception of large aggregates
6106 initialized with an others_clause, and exception declarations (which
6107 generate calls to a run-time registry procedure). Note that this restriction
6108 is enforced on a unit by unit basis, it need not be obeyed consistently
6109 throughout a partition.
6111 @item No_Entry_Queue
6112 @findex No_Entry_Queue
6113 This restriction is a declaration that any protected entry compiled in
6114 the scope of the restriction has at most one task waiting on the entry
6115 at any one time, and so no queue is required. This restriction is not
6116 checked at compile time. A program execution is erroneous if an attempt
6117 is made to queue a second task on such an entry.
6119 @item No_Implementation_Attributes
6120 @findex No_Implementation_Attributes
6121 This restriction checks at compile time that no GNAT-defined attributes
6122 are present. With this restriction, the only attributes that can be used
6123 are those defined in the Ada 95 Reference Manual.
6125 @item No_Implementation_Pragmas
6126 @findex No_Implementation_Pragmas
6127 This restriction checks at compile time that no GNAT-defined pragmas
6128 are present. With this restriction, the only pragmas that can be used
6129 are those defined in the Ada 95 Reference Manual.
6131 @item No_Implementation_Restrictions
6132 @findex No_Implementation_Restrictions
6133 This restriction checks at compile time that no GNAT-defined restriction
6134 identifiers (other than @code{No_Implementation_Restrictions} itself)
6135 are present. With this restriction, the only other restriction identifiers
6136 that can be used are those defined in the Ada 95 Reference Manual.
6143 @strong{58}. The consequences of violating limitations on
6144 @code{Restrictions} pragmas. See 13.12(9).
6147 Restrictions that can be checked at compile time result in illegalities
6148 if violated. Currently there are no other consequences of violating
6154 @strong{59}. The representation used by the @code{Read} and
6155 @code{Write} attributes of elementary types in terms of stream
6156 elements. See 13.13.2(9).
6159 The representation is the in-memory representation of the base type of
6160 the type, using the number of bits corresponding to the
6161 @code{@var{type}'Size} value, and the natural ordering of the machine.
6166 @strong{60}. The names and characteristics of the numeric subtypes
6167 declared in the visible part of package @code{Standard}. See A.1(3).
6170 See items describing the integer and floating-point types supported.
6175 @strong{61}. The accuracy actually achieved by the elementary
6176 functions. See A.5.1(1).
6179 The elementary functions correspond to the functions available in the C
6180 library. Only fast math mode is implemented.
6185 @strong{62}. The sign of a zero result from some of the operators or
6186 functions in @code{Numerics.Generic_Elementary_Functions}, when
6187 @code{Float_Type'Signed_Zeros} is @code{True}. See A.5.1(46).
6190 The sign of zeroes follows the requirements of the IEEE 754 standard on
6196 @strong{63}. The value of
6197 @code{Numerics.Float_Random.Max_Image_Width}. See A.5.2(27).
6200 Maximum image width is 649, see library file @file{a-numran.ads}.
6205 @strong{64}. The value of
6206 @code{Numerics.Discrete_Random.Max_Image_Width}. See A.5.2(27).
6209 Maximum image width is 80, see library file @file{a-nudira.ads}.
6214 @strong{65}. The algorithms for random number generation. See
6218 The algorithm is documented in the source files @file{a-numran.ads} and
6219 @file{a-numran.adb}.
6224 @strong{66}. The string representation of a random number generator's
6225 state. See A.5.2(38).
6228 See the documentation contained in the file @file{a-numran.adb}.
6233 @strong{67}. The minimum time interval between calls to the
6234 time-dependent Reset procedure that are guaranteed to initiate different
6235 random number sequences. See A.5.2(45).
6238 The minimum period between reset calls to guarantee distinct series of
6239 random numbers is one microsecond.
6244 @strong{68}. The values of the @code{Model_Mantissa},
6245 @code{Model_Emin}, @code{Model_Epsilon}, @code{Model},
6246 @code{Safe_First}, and @code{Safe_Last} attributes, if the Numerics
6247 Annex is not supported. See A.5.3(72).
6250 See the source file @file{ttypef.ads} for the values of all numeric
6256 @strong{69}. Any implementation-defined characteristics of the
6257 input-output packages. See A.7(14).
6260 There are no special implementation defined characteristics for these
6266 @strong{70}. The value of @code{Buffer_Size} in @code{Storage_IO}. See
6270 All type representations are contiguous, and the @code{Buffer_Size} is
6271 the value of @code{@var{type}'Size} rounded up to the next storage unit
6277 @strong{71}. External files for standard input, standard output, and
6278 standard error See A.10(5).
6281 These files are mapped onto the files provided by the C streams
6282 libraries. See source file @file{i-cstrea.ads} for further details.
6287 @strong{72}. The accuracy of the value produced by @code{Put}. See
6291 If more digits are requested in the output than are represented by the
6292 precision of the value, zeroes are output in the corresponding least
6293 significant digit positions.
6298 @strong{73}. The meaning of @code{Argument_Count}, @code{Argument}, and
6299 @code{Command_Name}. See A.15(1).
6302 These are mapped onto the @code{argv} and @code{argc} parameters of the
6303 main program in the natural manner.
6308 @strong{74}. Implementation-defined convention names. See B.1(11).
6311 The following convention names are supported
6319 Synonym for Assembler
6321 Synonym for Assembler
6324 @item C_Pass_By_Copy
6325 Allowed only for record types, like C, but also notes that record
6326 is to be passed by copy rather than reference.
6332 Treated the same as C
6334 Treated the same as C
6338 For support of pragma @code{Import} with convention Intrinsic, see
6339 separate section on Intrinsic Subprograms.
6341 Stdcall (used for Windows implementations only). This convention correspond
6342 to the WINAPI (previously called Pascal convention) C/C++ convention under
6343 Windows. A function with this convention cleans the stack before exit.
6349 Stubbed is a special convention used to indicate that the body of the
6350 subprogram will be entirely ignored. Any call to the subprogram
6351 is converted into a raise of the @code{Program_Error} exception. If a
6352 pragma @code{Import} specifies convention @code{stubbed} then no body need
6353 be present at all. This convention is useful during development for the
6354 inclusion of subprograms whose body has not yet been written.
6358 In addition, all otherwise unrecognized convention names are also
6359 treated as being synonymous with convention C@. In all implementations
6360 except for VMS, use of such other names results in a warning. In VMS
6361 implementations, these names are accepted silently.
6366 @strong{75}. The meaning of link names. See B.1(36).
6369 Link names are the actual names used by the linker.
6374 @strong{76}. The manner of choosing link names when neither the link
6375 name nor the address of an imported or exported entity is specified. See
6379 The default linker name is that which would be assigned by the relevant
6380 external language, interpreting the Ada name as being in all lower case
6386 @strong{77}. The effect of pragma @code{Linker_Options}. See B.1(37).
6389 The string passed to @code{Linker_Options} is presented uninterpreted as
6390 an argument to the link command, unless it contains Ascii.NUL characters.
6391 NUL characters if they appear act as argument separators, so for example
6394 pragma Linker_Options ("-labc" & ASCII.Nul & "-ldef");
6398 causes two separate arguments @code{-labc} and @code{-ldef} to be passed to the
6399 linker. The order of linker options is preserved for a given unit. The final
6400 list of options passed to the linker is in reverse order of the elaboration
6401 order. For example, linker options fo a body always appear before the options
6402 from the corresponding package spec.
6407 @strong{78}. The contents of the visible part of package
6408 @code{Interfaces} and its language-defined descendants. See B.2(1).
6411 See files with prefix @file{i-} in the distributed library.
6416 @strong{79}. Implementation-defined children of package
6417 @code{Interfaces}. The contents of the visible part of package
6418 @code{Interfaces}. See B.2(11).
6421 See files with prefix @file{i-} in the distributed library.
6426 @strong{80}. The types @code{Floating}, @code{Long_Floating},
6427 @code{Binary}, @code{Long_Binary}, @code{Decimal_ Element}, and
6428 @code{COBOL_Character}; and the initialization of the variables
6429 @code{Ada_To_COBOL} and @code{COBOL_To_Ada}, in
6430 @code{Interfaces.COBOL}. See B.4(50).
6437 (Floating) Long_Float
6442 @item Decimal_Element
6444 @item COBOL_Character
6448 For initialization, see the file @file{i-cobol.ads} in the distributed library.
6453 @strong{81}. Support for access to machine instructions. See C.1(1).
6456 See documentation in file @file{s-maccod.ads} in the distributed library.
6461 @strong{82}. Implementation-defined aspects of access to machine
6462 operations. See C.1(9).
6465 See documentation in file @file{s-maccod.ads} in the distributed library.
6470 @strong{83}. Implementation-defined aspects of interrupts. See C.3(2).
6473 Interrupts are mapped to signals or conditions as appropriate. See
6475 @code{Ada.Interrupt_Names} in source file @file{a-intnam.ads} for details
6476 on the interrupts supported on a particular target.
6481 @strong{84}. Implementation-defined aspects of pre-elaboration. See
6485 GNAT does not permit a partition to be restarted without reloading,
6486 except under control of the debugger.
6491 @strong{85}. The semantics of pragma @code{Discard_Names}. See C.5(7).
6494 Pragma @code{Discard_Names} causes names of enumeration literals to
6495 be suppressed. In the presence of this pragma, the Image attribute
6496 provides the image of the Pos of the literal, and Value accepts
6502 @strong{86}. The result of the @code{Task_Identification.Image}
6503 attribute. See C.7.1(7).
6506 The result of this attribute is an 8-digit hexadecimal string
6507 representing the virtual address of the task control block.
6512 @strong{87}. The value of @code{Current_Task} when in a protected entry
6513 or interrupt handler. See C.7.1(17).
6516 Protected entries or interrupt handlers can be executed by any
6517 convenient thread, so the value of @code{Current_Task} is undefined.
6522 @strong{88}. The effect of calling @code{Current_Task} from an entry
6523 body or interrupt handler. See C.7.1(19).
6526 The effect of calling @code{Current_Task} from an entry body or
6527 interrupt handler is to return the identification of the task currently
6533 @strong{89}. Implementation-defined aspects of
6534 @code{Task_Attributes}. See C.7.2(19).
6537 There are no implementation-defined aspects of @code{Task_Attributes}.
6542 @strong{90}. Values of all @code{Metrics}. See D(2).
6545 The metrics information for GNAT depends on the performance of the
6546 underlying operating system. The sources of the run-time for tasking
6547 implementation, together with the output from @code{-gnatG} can be
6548 used to determine the exact sequence of operating systems calls made
6549 to implement various tasking constructs. Together with appropriate
6550 information on the performance of the underlying operating system,
6551 on the exact target in use, this information can be used to determine
6552 the required metrics.
6557 @strong{91}. The declarations of @code{Any_Priority} and
6558 @code{Priority}. See D.1(11).
6561 See declarations in file @file{system.ads}.
6566 @strong{92}. Implementation-defined execution resources. See D.1(15).
6569 There are no implementation-defined execution resources.
6574 @strong{93}. Whether, on a multiprocessor, a task that is waiting for
6575 access to a protected object keeps its processor busy. See D.2.1(3).
6578 On a multi-processor, a task that is waiting for access to a protected
6579 object does not keep its processor busy.
6584 @strong{94}. The affect of implementation defined execution resources
6585 on task dispatching. See D.2.1(9).
6590 Tasks map to IRIX threads, and the dispatching policy is as defined by
6591 the IRIX implementation of threads.
6593 Tasks map to threads in the threads package used by GNAT@. Where possible
6594 and appropriate, these threads correspond to native threads of the
6595 underlying operating system.
6600 @strong{95}. Implementation-defined @code{policy_identifiers} allowed
6601 in a pragma @code{Task_Dispatching_Policy}. See D.2.2(3).
6604 There are no implementation-defined policy-identifiers allowed in this
6610 @strong{96}. Implementation-defined aspects of priority inversion. See
6614 Execution of a task cannot be preempted by the implementation processing
6615 of delay expirations for lower priority tasks.
6620 @strong{97}. Implementation defined task dispatching. See D.2.2(18).
6625 Tasks map to IRIX threads, and the dispatching policy is as defied by
6626 the IRIX implementation of threads.
6628 The policy is the same as that of the underlying threads implementation.
6633 @strong{98}. Implementation-defined @code{policy_identifiers} allowed
6634 in a pragma @code{Locking_Policy}. See D.3(4).
6637 The only implementation defined policy permitted in GNAT is
6638 @code{Inheritance_Locking}. On targets that support this policy, locking
6639 is implemented by inheritance, i.e.@: the task owning the lock operates
6640 at a priority equal to the highest priority of any task currently
6641 requesting the lock.
6646 @strong{99}. Default ceiling priorities. See D.3(10).
6649 The ceiling priority of protected objects of the type
6650 @code{System.Interrupt_Priority'Last} as described in the Ada 95
6651 Reference Manual D.3(10),
6656 @strong{100}. The ceiling of any protected object used internally by
6657 the implementation. See D.3(16).
6660 The ceiling priority of internal protected objects is
6661 @code{System.Priority'Last}.
6666 @strong{101}. Implementation-defined queuing policies. See D.4(1).
6669 There are no implementation-defined queueing policies.
6674 @strong{102}. On a multiprocessor, any conditions that cause the
6675 completion of an aborted construct to be delayed later than what is
6676 specified for a single processor. See D.6(3).
6679 The semantics for abort on a multi-processor is the same as on a single
6680 processor, there are no further delays.
6685 @strong{103}. Any operations that implicitly require heap storage
6686 allocation. See D.7(8).
6689 The only operation that implicitly requires heap storage allocation is
6695 @strong{104}. Implementation-defined aspects of pragma
6696 @code{Restrictions}. See D.7(20).
6699 There are no such implementation-defined aspects.
6704 @strong{105}. Implementation-defined aspects of package
6705 @code{Real_Time}. See D.8(17).
6708 There are no implementation defined aspects of package @code{Real_Time}.
6713 @strong{106}. Implementation-defined aspects of
6714 @code{delay_statements}. See D.9(8).
6717 Any difference greater than one microsecond will cause the task to be
6718 delayed (see D.9(7)).
6723 @strong{107}. The upper bound on the duration of interrupt blocking
6724 caused by the implementation. See D.12(5).
6727 The upper bound is determined by the underlying operating system. In
6728 no cases is it more than 10 milliseconds.
6733 @strong{108}. The means for creating and executing distributed
6737 The GLADE package provides a utility GNATDIST for creating and executing
6738 distributed programs. See the GLADE reference manual for further details.
6743 @strong{109}. Any events that can result in a partition becoming
6744 inaccessible. See E.1(7).
6747 See the GLADE reference manual for full details on such events.
6752 @strong{110}. The scheduling policies, treatment of priorities, and
6753 management of shared resources between partitions in certain cases. See
6757 See the GLADE reference manual for full details on these aspects of
6758 multi-partition execution.
6763 @strong{111}. Events that cause the version of a compilation unit to
6767 Editing the source file of a compilation unit, or the source files of
6768 any units on which it is dependent in a significant way cause the version
6769 to change. No other actions cause the version number to change. All changes
6770 are significant except those which affect only layout, capitalization or
6776 @strong{112}. Whether the execution of the remote subprogram is
6777 immediately aborted as a result of cancellation. See E.4(13).
6780 See the GLADE reference manual for details on the effect of abort in
6781 a distributed application.
6786 @strong{113}. Implementation-defined aspects of the PCS@. See E.5(25).
6789 See the GLADE reference manual for a full description of all implementation
6790 defined aspects of the PCS@.
6795 @strong{114}. Implementation-defined interfaces in the PCS@. See
6799 See the GLADE reference manual for a full description of all
6800 implementation defined interfaces.
6805 @strong{115}. The values of named numbers in the package
6806 @code{Decimal}. See F.2(7).
6818 @item Max_Decimal_Digits
6825 @strong{116}. The value of @code{Max_Picture_Length} in the package
6826 @code{Text_IO.Editing}. See F.3.3(16).
6834 @strong{117}. The value of @code{Max_Picture_Length} in the package
6835 @code{Wide_Text_IO.Editing}. See F.3.4(5).
6843 @strong{118}. The accuracy actually achieved by the complex elementary
6844 functions and by other complex arithmetic operations. See G.1(1).
6847 Standard library functions are used for the complex arithmetic
6848 operations. Only fast math mode is currently supported.
6853 @strong{119}. The sign of a zero result (or a component thereof) from
6854 any operator or function in @code{Numerics.Generic_Complex_Types}, when
6855 @code{Real'Signed_Zeros} is True. See G.1.1(53).
6858 The signs of zero values are as recommended by the relevant
6859 implementation advice.
6864 @strong{120}. The sign of a zero result (or a component thereof) from
6865 any operator or function in
6866 @code{Numerics.Generic_Complex_Elementary_Functions}, when
6867 @code{Real'Signed_Zeros} is @code{True}. See G.1.2(45).
6870 The signs of zero values are as recommended by the relevant
6871 implementation advice.
6876 @strong{121}. Whether the strict mode or the relaxed mode is the
6877 default. See G.2(2).
6880 The strict mode is the default. There is no separate relaxed mode. GNAT
6881 provides a highly efficient implementation of strict mode.
6886 @strong{122}. The result interval in certain cases of fixed-to-float
6887 conversion. See G.2.1(10).
6890 For cases where the result interval is implementation dependent, the
6891 accuracy is that provided by performing all operations in 64-bit IEEE
6892 floating-point format.
6897 @strong{123}. The result of a floating point arithmetic operation in
6898 overflow situations, when the @code{Machine_Overflows} attribute of the
6899 result type is @code{False}. See G.2.1(13).
6902 Infinite and Nan values are produced as dictated by the IEEE
6903 floating-point standard.
6908 @strong{124}. The result interval for division (or exponentiation by a
6909 negative exponent), when the floating point hardware implements division
6910 as multiplication by a reciprocal. See G.2.1(16).
6913 Not relevant, division is IEEE exact.
6918 @strong{125}. The definition of close result set, which determines the
6919 accuracy of certain fixed point multiplications and divisions. See
6923 Operations in the close result set are performed using IEEE long format
6924 floating-point arithmetic. The input operands are converted to
6925 floating-point, the operation is done in floating-point, and the result
6926 is converted to the target type.
6931 @strong{126}. Conditions on a @code{universal_real} operand of a fixed
6932 point multiplication or division for which the result shall be in the
6933 perfect result set. See G.2.3(22).
6936 The result is only defined to be in the perfect result set if the result
6937 can be computed by a single scaling operation involving a scale factor
6938 representable in 64-bits.
6943 @strong{127}. The result of a fixed point arithmetic operation in
6944 overflow situations, when the @code{Machine_Overflows} attribute of the
6945 result type is @code{False}. See G.2.3(27).
6948 Not relevant, @code{Machine_Overflows} is @code{True} for fixed-point
6954 @strong{128}. The result of an elementary function reference in
6955 overflow situations, when the @code{Machine_Overflows} attribute of the
6956 result type is @code{False}. See G.2.4(4).
6959 IEEE infinite and Nan values are produced as appropriate.
6964 @strong{129}. The value of the angle threshold, within which certain
6965 elementary functions, complex arithmetic operations, and complex
6966 elementary functions yield results conforming to a maximum relative
6967 error bound. See G.2.4(10).
6970 Information on this subject is not yet available.
6975 @strong{130}. The accuracy of certain elementary functions for
6976 parameters beyond the angle threshold. See G.2.4(10).
6979 Information on this subject is not yet available.
6984 @strong{131}. The result of a complex arithmetic operation or complex
6985 elementary function reference in overflow situations, when the
6986 @code{Machine_Overflows} attribute of the corresponding real type is
6987 @code{False}. See G.2.6(5).
6990 IEEE infinite and Nan values are produced as appropriate.
6995 @strong{132}. The accuracy of certain complex arithmetic operations and
6996 certain complex elementary functions for parameters (or components
6997 thereof) beyond the angle threshold. See G.2.6(8).
7000 Information on those subjects is not yet available.
7005 @strong{133}. Information regarding bounded errors and erroneous
7006 execution. See H.2(1).
7009 Information on this subject is not yet available.
7014 @strong{134}. Implementation-defined aspects of pragma
7015 @code{Inspection_Point}. See H.3.2(8).
7018 Pragma @code{Inspection_Point} ensures that the variable is live and can
7019 be examined by the debugger at the inspection point.
7024 @strong{135}. Implementation-defined aspects of pragma
7025 @code{Restrictions}. See H.4(25).
7028 There are no implementation-defined aspects of pragma @code{Restrictions}. The
7029 use of pragma @code{Restrictions [No_Exceptions]} has no effect on the
7030 generated code. Checks must suppressed by use of pragma @code{Suppress}.
7035 @strong{136}. Any restrictions on pragma @code{Restrictions}. See
7039 There are no restrictions on pragma @code{Restrictions}.
7041 @node Intrinsic Subprograms
7042 @chapter Intrinsic Subprograms
7043 @cindex Intrinsic Subprograms
7046 * Intrinsic Operators::
7047 * Enclosing_Entity::
7048 * Exception_Information::
7049 * Exception_Message::
7057 * Shift_Right_Arithmetic::
7061 GNAT allows a user application program to write the declaration:
7064 pragma Import (Intrinsic, name);
7068 providing that the name corresponds to one of the implemented intrinsic
7069 subprograms in GNAT, and that the parameter profile of the referenced
7070 subprogram meets the requirements. This chapter describes the set of
7071 implemented intrinsic subprograms, and the requirements on parameter profiles.
7072 Note that no body is supplied; as with other uses of pragma Import, the
7073 body is supplied elsewhere (in this case by the compiler itself). Note
7074 that any use of this feature is potentially non-portable, since the
7075 Ada standard does not require Ada compilers to implement this feature.
7077 @node Intrinsic Operators
7078 @section Intrinsic Operators
7079 @cindex Intrinsic operator
7082 All the predefined numeric operators in package Standard
7083 in @code{pragma Import (Intrinsic,..)}
7084 declarations. In the binary operator case, the operands must have the same
7085 size. The operand or operands must also be appropriate for
7086 the operator. For example, for addition, the operands must
7087 both be floating-point or both be fixed-point, and the
7088 right operand for @code{"**"} must have a root type of
7089 @code{Standard.Integer'Base}.
7090 You can use an intrinsic operator declaration as in the following example:
7093 type Int1 is new Integer;
7094 type Int2 is new Integer;
7096 function "+" (X1 : Int1; X2 : Int2) return Int1;
7097 function "+" (X1 : Int1; X2 : Int2) return Int2;
7098 pragma Import (Intrinsic, "+");
7102 This declaration would permit ``mixed mode'' arithmetic on items
7103 of the differing types @code{Int1} and @code{Int2}.
7104 It is also possible to specify such operators for private types, if the
7105 full views are appropriate arithmetic types.
7107 @node Enclosing_Entity
7108 @section Enclosing_Entity
7109 @cindex Enclosing_Entity
7111 This intrinsic subprogram is used in the implementation of the
7112 library routine @code{GNAT.Source_Info}. The only useful use of the
7113 intrinsic import in this case is the one in this unit, so an
7114 application program should simply call the function
7115 @code{GNAT.Source_Info.Enclosing_Entity} to obtain the name of
7116 the current subprogram, package, task, entry, or protected subprogram.
7118 @node Exception_Information
7119 @section Exception_Information
7120 @cindex Exception_Information'
7122 This intrinsic subprogram is used in the implementation of the
7123 library routine @code{GNAT.Current_Exception}. The only useful
7124 use of the intrinsic import in this case is the one in this unit,
7125 so an application program should simply call the function
7126 @code{GNAT.Current_Exception.Exception_Information} to obtain
7127 the exception information associated with the current exception.
7129 @node Exception_Message
7130 @section Exception_Message
7131 @cindex Exception_Message
7133 This intrinsic subprogram is used in the implementation of the
7134 library routine @code{GNAT.Current_Exception}. The only useful
7135 use of the intrinsic import in this case is the one in this unit,
7136 so an application program should simply call the function
7137 @code{GNAT.Current_Exception.Exception_Message} to obtain
7138 the message associated with the current exception.
7140 @node Exception_Name
7141 @section Exception_Name
7142 @cindex Exception_Name
7144 This intrinsic subprogram is used in the implementation of the
7145 library routine @code{GNAT.Current_Exception}. The only useful
7146 use of the intrinsic import in this case is the one in this unit,
7147 so an application program should simply call the function
7148 @code{GNAT.Current_Exception.Exception_Name} to obtain
7149 the name of the current exception.
7155 This intrinsic subprogram is used in the implementation of the
7156 library routine @code{GNAT.Source_Info}. The only useful use of the
7157 intrinsic import in this case is the one in this unit, so an
7158 application program should simply call the function
7159 @code{GNAT.Source_Info.File} to obtain the name of the current
7166 This intrinsic subprogram is used in the implementation of the
7167 library routine @code{GNAT.Source_Info}. The only useful use of the
7168 intrinsic import in this case is the one in this unit, so an
7169 application program should simply call the function
7170 @code{GNAT.Source_Info.Line} to obtain the number of the current
7174 @section Rotate_Left
7177 In standard Ada 95, the @code{Rotate_Left} function is available only
7178 for the predefined modular types in package @code{Interfaces}. However, in
7179 GNAT it is possible to define a Rotate_Left function for a user
7180 defined modular type or any signed integer type as in this example:
7184 (Value : My_Modular_Type;
7186 return My_Modular_Type;
7190 The requirements are that the profile be exactly as in the example
7191 above. The only modifications allowed are in the formal parameter
7192 names, and in the type of @code{Value} and the return type, which
7193 must be the same, and must be either a signed integer type, or
7194 a modular integer type with a binary modulus, and the size must
7195 be 8. 16, 32 or 64 bits.
7198 @section Rotate_Right
7199 @cindex Rotate_Right
7201 A @code{Rotate_Right} function can be defined for any user defined
7202 binary modular integer type, or signed integer type, as described
7203 above for @code{Rotate_Left}.
7209 A @code{Shift_Left} function can be defined for any user defined
7210 binary modular integer type, or signed integer type, as described
7211 above for @code{Rotate_Left}.
7214 @section Shift_Right
7217 A @code{Shift_Right} function can be defined for any user defined
7218 binary modular integer type, or signed integer type, as described
7219 above for @code{Rotate_Left}.
7221 @node Shift_Right_Arithmetic
7222 @section Shift_Right_Arithmetic
7223 @cindex Shift_Right_Arithmetic
7225 A @code{Shift_Right_Arithmetic} function can be defined for any user
7226 defined binary modular integer type, or signed integer type, as described
7227 above for @code{Rotate_Left}.
7229 @node Source_Location
7230 @section Source_Location
7231 @cindex Source_Location
7233 This intrinsic subprogram is used in the implementation of the
7234 library routine @code{GNAT.Source_Info}. The only useful use of the
7235 intrinsic import in this case is the one in this unit, so an
7236 application program should simply call the function
7237 @code{GNAT.Source_Info.Source_Location} to obtain the current
7238 source file location.
7240 @node Representation Clauses and Pragmas
7241 @chapter Representation Clauses and Pragmas
7242 @cindex Representation Clauses
7245 * Alignment Clauses::
7247 * Storage_Size Clauses::
7248 * Size of Variant Record Objects::
7249 * Biased Representation ::
7250 * Value_Size and Object_Size Clauses::
7251 * Component_Size Clauses::
7252 * Bit_Order Clauses::
7253 * Effect of Bit_Order on Byte Ordering::
7254 * Pragma Pack for Arrays::
7255 * Pragma Pack for Records::
7256 * Record Representation Clauses::
7257 * Enumeration Clauses::
7259 * Effect of Convention on Representation::
7260 * Determining the Representations chosen by GNAT::
7264 @cindex Representation Clause
7265 @cindex Representation Pragma
7266 @cindex Pragma, representation
7267 This section describes the representation clauses accepted by GNAT, and
7268 their effect on the representation of corresponding data objects.
7270 GNAT fully implements Annex C (Systems Programming). This means that all
7271 the implementation advice sections in chapter 13 are fully implemented.
7272 However, these sections only require a minimal level of support for
7273 representation clauses. GNAT provides much more extensive capabilities,
7274 and this section describes the additional capabilities provided.
7276 @node Alignment Clauses
7277 @section Alignment Clauses
7278 @cindex Alignment Clause
7281 GNAT requires that all alignment clauses specify a power of 2, and all
7282 default alignments are always a power of 2. The default alignment
7283 values are as follows:
7286 @item Primitive Types
7287 For primitive types, the alignment is the maximum of the actual size of
7288 objects of the type, and the maximum alignment supported by the target.
7289 For example, for type Long_Float, the object size is 8 bytes, and the
7290 default alignment will be 8 on any target that supports alignments
7291 this large, but on some targets, the maximum alignment may be smaller
7292 than 8, in which case objects of type Long_Float will be maximally
7296 For arrays, the alignment is equal to the alignment of the component type
7297 for the normal case where no packing or component size is given. If the
7298 array is packed, and the packing is effective (see separate section on
7299 packed arrays), then the alignment will be one for long packed arrays,
7300 or arrays whose length is not known at compile time. For short packed
7301 arrays, which are handled internally as modular types, the alignment
7302 will be as described for primitive types, e.g.@: a packed array of length
7303 31 bits will have an object size of four bytes, and an alignment of 4.
7306 For the normal non-packed case, the alignment of a record is equal to
7307 the maximum alignment of any of its components. For tagged records, this
7308 includes the implicit access type used for the tag. If a pragma @code{Pack} is
7309 used and all fields are packable (see separate section on pragma @code{Pack}),
7310 then the resulting alignment is 1.
7312 A special case is when the size of the record is given explicitly, or a
7313 full record representation clause is given, and the size of the record
7314 is 2, 4, or 8 bytes. In this case, an alignment is chosen to match the
7315 size of the record. For example, if we have:
7318 type Small is record
7324 then the default alignment of the record type @code{Small} is 2, not 1. This
7325 leads to more efficient code when the record is treated as a unit, and also
7326 allows the type to specified as @code{Atomic} on architectures requiring
7332 An alignment clause may
7333 always specify a larger alignment than the default value, up to some
7334 maximum value dependent on the target (obtainable by using the
7335 attribute reference System'Maximum_Alignment). The only case in which
7336 it is permissible to specify a smaller alignment than the default value
7337 is in the case of a record for which a record representation clause is
7338 given. In this case, packable fields for which a component clause is
7339 given still result in a default alignment corresponding to the original
7340 type, but this may be overridden, since these components in fact only
7341 require an alignment of one byte. For example, given
7349 a at 0 range 0 .. 31;
7352 for v'alignment use 1;
7356 @cindex Alignment, default
7357 The default alignment for the type @code{v} is 4, as a result of the
7358 integer field in the record, but since this field is placed with a
7359 component clause, it is permissible, as shown, to override the default
7360 alignment of the record to a smaller value.
7363 @section Size Clauses
7367 The default size of types is as specified in the reference manual. For
7368 objects, GNAT will generally increase the type size so that the object
7369 size is a multiple of storage units, and also a multiple of the
7370 alignment. For example
7373 type Smallint is range 1 .. 6;
7382 In this example, @code{Smallint}
7383 has a size of 3, as specified by the RM rules,
7384 but objects of this type will have a size of 8,
7385 since objects by default occupy an integral number
7386 of storage units. On some targets, notably older
7387 versions of the Digital Alpha, the size of stand
7388 alone objects of this type may be 32, reflecting
7389 the inability of the hardware to do byte load/stores.
7391 Similarly, the size of type @code{Rec} is 40 bits, but
7392 the alignment is 4, so objects of this type will have
7393 their size increased to 64 bits so that it is a multiple
7394 of the alignment. The reason for this decision, which is
7395 in accordance with the specific note in RM 13.3(43):
7398 A Size clause should be supported for an object if the specified
7399 Size is at least as large as its subtype's Size, and corresponds
7400 to a size in storage elements that is a multiple of the object's
7401 Alignment (if the Alignment is nonzero).
7405 An explicit size clause may be used to override the default size by
7406 increasing it. For example, if we have:
7409 type My_Boolean is new Boolean;
7410 for My_Boolean'Size use 32;
7414 then objects of this type will always be 32 bits long. In the case of
7415 discrete types, the size can be increased up to 64 bits, with the effect
7416 that the entire specified field is used to hold the value, sign- or
7417 zero-extended as appropriate. If more than 64 bits is specified, then
7418 padding space is allocated after the value, and a warning is issued that
7419 there are unused bits.
7421 Similarly the size of records and arrays may be increased, and the effect
7422 is to add padding bits after the value. This also causes a warning message
7425 The largest Size value permitted in GNAT is 2**32@minus{}1. Since this is a
7426 Size in bits, this corresponds to an object of size 256 megabytes (minus
7427 one). This limitation is true on all targets. The reason for this
7428 limitation is that it improves the quality of the code in many cases
7429 if it is known that a Size value can be accommodated in an object of
7432 @node Storage_Size Clauses
7433 @section Storage_Size Clauses
7434 @cindex Storage_Size Clause
7437 For tasks, the @code{Storage_Size} clause specifies the amount of space
7438 to be allocated for the task stack. This cannot be extended, and if the
7439 stack is exhausted, then @code{Storage_Error} will be raised if stack
7440 checking is enabled. If the default size of 20K bytes is insufficient,
7441 then you need to use a @code{Storage_Size} attribute definition clause,
7442 or a @code{Storage_Size} pragma in the task definition to set the
7443 appropriate required size. A useful technique is to include in every
7444 task definition a pragma of the form:
7447 pragma Storage_Size (Default_Stack_Size);
7451 Then Default_Stack_Size can be defined in a global package, and modified
7452 as required. Any tasks requiring different task stack sizes from the
7453 default can have an appropriate alternative reference in the pragma.
7455 For access types, the @code{Storage_Size} clause specifies the maximum
7456 space available for allocation of objects of the type. If this space is
7457 exceeded then @code{Storage_Error} will be raised by an allocation attempt.
7458 In the case where the access type is declared local to a subprogram, the
7459 use of a @code{Storage_Size} clause triggers automatic use of a special
7460 predefined storage pool (@code{System.Pool_Size}) that ensures that all
7461 space for the pool is automatically reclaimed on exit from the scope in
7462 which the type is declared.
7464 A special case recognized by the compiler is the specification of a
7465 @code{Storage_Size} of zero for an access type. This means that no
7466 items can be allocated from the pool, and this is recognized at compile
7467 time, and all the overhead normally associated with maintaining a fixed
7468 size storage pool is eliminated. Consider the following example:
7472 type R is array (Natural) of Character;
7473 type P is access all R;
7474 for P'Storage_Size use 0;
7475 -- Above access type intended only for interfacing purposes
7479 procedure g (m : P);
7480 pragma Import (C, g);
7491 As indicated in this example, these dummy storage pools are often useful in
7492 connection with interfacing where no object will ever be allocated. If you
7493 compile the above example, you get the warning:
7496 p.adb:16:09: warning: allocation from empty storage pool
7497 p.adb:16:09: warning: Storage_Error will be raised at run time
7501 Of course in practice, there will not be any explicit allocators in the
7502 case of such an access declaration.
7504 @node Size of Variant Record Objects
7505 @section Size of Variant Record Objects
7506 @cindex Size, variant record objects
7507 @cindex Variant record objects, size
7510 An issue arises in the case of variant record objects of whether Size gives
7511 information about a particular variant, or the maximum size required
7512 for any variant. Consider the following program
7515 with Text_IO; use Text_IO;
7517 type R1 (A : Boolean := False) is record
7519 when True => X : Character;
7528 Put_Line (Integer'Image (V1'Size));
7529 Put_Line (Integer'Image (V2'Size));
7534 Here we are dealing with a variant record, where the True variant
7535 requires 16 bits, and the False variant requires 8 bits.
7536 In the above example, both V1 and V2 contain the False variant,
7537 which is only 8 bits long. However, the result of running the
7546 The reason for the difference here is that the discriminant value of
7547 V1 is fixed, and will always be False. It is not possible to assign
7548 a True variant value to V1, therefore 8 bits is sufficient. On the
7549 other hand, in the case of V2, the initial discriminant value is
7550 False (from the default), but it is possible to assign a True
7551 variant value to V2, therefore 16 bits must be allocated for V2
7552 in the general case, even fewer bits may be needed at any particular
7553 point during the program execution.
7555 As can be seen from the output of this program, the @code{'Size}
7556 attribute applied to such an object in GNAT gives the actual allocated
7557 size of the variable, which is the largest size of any of the variants.
7558 The Ada Reference Manual is not completely clear on what choice should
7559 be made here, but the GNAT behavior seems most consistent with the
7560 language in the RM@.
7562 In some cases, it may be desirable to obtain the size of the current
7563 variant, rather than the size of the largest variant. This can be
7564 achieved in GNAT by making use of the fact that in the case of a
7565 subprogram parameter, GNAT does indeed return the size of the current
7566 variant (because a subprogram has no way of knowing how much space
7567 is actually allocated for the actual).
7569 Consider the following modified version of the above program:
7572 with Text_IO; use Text_IO;
7574 type R1 (A : Boolean := False) is record
7576 when True => X : Character;
7583 function Size (V : R1) return Integer is
7589 Put_Line (Integer'Image (V2'Size));
7590 Put_Line (Integer'IMage (Size (V2)));
7592 Put_Line (Integer'Image (V2'Size));
7593 Put_Line (Integer'IMage (Size (V2)));
7598 The output from this program is
7608 Here we see that while the @code{'Size} attribute always returns
7609 the maximum size, regardless of the current variant value, the
7610 @code{Size} function does indeed return the size of the current
7613 @node Biased Representation
7614 @section Biased Representation
7615 @cindex Size for biased representation
7616 @cindex Biased representation
7619 In the case of scalars with a range starting at other than zero, it is
7620 possible in some cases to specify a size smaller than the default minimum
7621 value, and in such cases, GNAT uses an unsigned biased representation,
7622 in which zero is used to represent the lower bound, and successive values
7623 represent successive values of the type.
7625 For example, suppose we have the declaration:
7628 type Small is range -7 .. -4;
7629 for Small'Size use 2;
7633 Although the default size of type @code{Small} is 4, the @code{Size}
7634 clause is accepted by GNAT and results in the following representation
7638 -7 is represented as 2#00#
7639 -6 is represented as 2#01#
7640 -5 is represented as 2#10#
7641 -4 is represented as 2#11#
7645 Biased representation is only used if the specified @code{Size} clause
7646 cannot be accepted in any other manner. These reduced sizes that force
7647 biased representation can be used for all discrete types except for
7648 enumeration types for which a representation clause is given.
7650 @node Value_Size and Object_Size Clauses
7651 @section Value_Size and Object_Size Clauses
7654 @cindex Size, of objects
7657 In Ada 95, the @code{Size} of a discrete type is the minimum number of bits
7658 required to hold values of the type. Although this interpretation was
7659 allowed in Ada 83, it was not required, and this requirement in practice
7660 can cause some significant difficulties. For example, in most Ada 83
7661 compilers, @code{Natural'Size} was 32. However, in Ada-95,
7662 @code{Natural'Size} is
7663 typically 31. This means that code may change in behavior when moving
7664 from Ada 83 to Ada 95. For example, consider:
7673 for A use at 0 range 0 .. Natural'Size - 1;
7674 for B use at 0 range Natural'Size .. 2 * Natural'Size - 1;
7679 In the above code, since the typical size of @code{Natural} objects
7680 is 32 bits and @code{Natural'Size} is 31, the above code can cause
7681 unexpected inefficient packing in Ada 95, and in general there are
7682 surprising cases where the fact that the object size can exceed the
7683 size of the type causes surprises.
7685 To help get around this problem GNAT provides two implementation
7686 dependent attributes @code{Value_Size} and @code{Object_Size}. When
7687 applied to a type, these attributes yield the size of the type
7688 (corresponding to the RM defined size attribute), and the size of
7689 objects of the type respectively.
7691 The @code{Object_Size} is used for determining the default size of
7692 objects and components. This size value can be referred to using the
7693 @code{Object_Size} attribute. The phrase ``is used'' here means that it is
7694 the basis of the determination of the size. The backend is free to
7695 pad this up if necessary for efficiency, e.g.@: an 8-bit stand-alone
7696 character might be stored in 32 bits on a machine with no efficient
7697 byte access instructions such as the Alpha.
7699 The default rules for the value of @code{Object_Size} for fixed-point and
7700 discrete types are as follows:
7704 The @code{Object_Size} for base subtypes reflect the natural hardware
7705 size in bits (run the utility @code{gnatpsta} to find those values for numeric types).
7706 Enumeration types and fixed-point base subtypes have 8, 16, 32 or 64
7707 bits for this size, depending on the range of values to be stored.
7710 The @code{Object_Size} of a subtype is the same as the
7711 @code{Object_Size} of
7712 the type from which it is obtained.
7715 The @code{Object_Size} of a derived base type is copied from the parent
7716 base type, and the @code{Object_Size} of a derived first subtype is copied
7717 from the parent first subtype.
7721 The @code{Value_Size} attribute
7722 is the number of bits required to store a value
7723 of the type. This size can be referred to using the @code{Value_Size}
7724 attribute. This value is used to determine how tightly to pack
7725 records or arrays with components of this type, and also affects
7726 the semantics of unchecked conversion (unchecked conversions where
7727 the @code{Value_Size} values differ generate a warning, and are potentially
7730 The default rules for the value of @code{Value_Size} are as follows:
7734 The @code{Value_Size} for a base subtype is the minimum number of bits
7735 required to store all values of the type (including the sign bit
7736 only if negative values are possible).
7739 If a subtype statically matches the first subtype of a given type, then it has
7740 by default the same @code{Value_Size} as the first subtype. This is a
7741 consequence of RM 13.1(14) (``if two subtypes statically match,
7742 then their subtype-specific aspects are the same''.)
7745 All other subtypes have a @code{Value_Size} corresponding to the minimum
7746 number of bits required to store all values of the subtype. For
7747 dynamic bounds, it is assumed that the value can range down or up
7748 to the corresponding bound of the ancestor
7752 The RM defined attribute @code{Size} corresponds to the
7753 @code{Value_Size} attribute.
7755 The @code{Size} attribute may be defined for a first-named subtype. This sets
7756 the @code{Value_Size} of
7757 the first-named subtype to the given value, and the
7758 @code{Object_Size} of this first-named subtype to the given value padded up
7759 to an appropriate boundary. It is a consequence of the default rules
7760 above that this @code{Object_Size} will apply to all further subtypes. On the
7761 other hand, @code{Value_Size} is affected only for the first subtype, any
7762 dynamic subtypes obtained from it directly, and any statically matching
7763 subtypes. The @code{Value_Size} of any other static subtypes is not affected.
7765 @code{Value_Size} and
7766 @code{Object_Size} may be explicitly set for any subtype using
7767 an attribute definition clause. Note that the use of these attributes
7768 can cause the RM 13.1(14) rule to be violated. If two access types
7769 reference aliased objects whose subtypes have differing @code{Object_Size}
7770 values as a result of explicit attribute definition clauses, then it
7771 is erroneous to convert from one access subtype to the other.
7773 At the implementation level, Esize stores the Object_SIze and the
7774 RM_Size field stores the @code{Value_Size} (and hence the value of the
7775 @code{Size} attribute,
7776 which, as noted above, is equivalent to @code{Value_Size}).
7778 To get a feel for the difference, consider the following examples (note
7779 that in each case the base is short_short_integer with a size of 8):
7782 Object_Size Value_Size
7784 type x1 is range 0 .. 5; 8 3
7786 type x2 is range 0 .. 5;
7787 for x2'size use 12; 12 12
7789 subtype x3 is x2 range 0 .. 3; 12 2
7791 subtype x4 is x2'base range 0 .. 10; 8 4
7793 subtype x5 is x2 range 0 .. dynamic; 12 (7)
7795 subtype x6 is x2'base range 0 .. dynamic; 8 (7)
7800 Note: the entries marked (7) are not actually specified by the Ada 95 RM,
7801 but it seems in the spirit of the RM rules to allocate the minimum number
7802 of bits known to be large enough to hold the given range of values.
7804 So far, so good, but GNAT has to obey the RM rules, so the question is
7805 under what conditions must the RM @code{Size} be used.
7806 The following is a list
7807 of the occasions on which the RM @code{Size} must be used:
7811 Component size for packed arrays or records
7814 Value of the attribute @code{Size} for a type
7817 Warning about sizes not matching for unchecked conversion
7821 For types other than discrete and fixed-point types, the @code{Object_Size}
7822 and Value_Size are the same (and equivalent to the RM attribute @code{Size}).
7823 Only @code{Size} may be specified for such types.
7825 @node Component_Size Clauses
7826 @section Component_Size Clauses
7827 @cindex Component_Size Clause
7830 Normally, the value specified in a component clause must be consistent
7831 with the subtype of the array component with regard to size and alignment.
7832 In other words, the value specified must be at least equal to the size
7833 of this subtype, and must be a multiple of the alignment value.
7835 In addition, component size clauses are allowed which cause the array
7836 to be packed, by specifying a smaller value. The cases in which this
7837 is allowed are for component size values in the range 1 through 63. The value
7838 specified must not be smaller than the Size of the subtype. GNAT will
7839 accurately honor all packing requests in this range. For example, if
7843 type r is array (1 .. 8) of Natural;
7848 then the resulting array has a length of 31 bytes (248 bits = 8 * 31).
7849 Of course access to the components of such an array is considerably
7850 less efficient than if the natural component size of 32 is used.
7852 @node Bit_Order Clauses
7853 @section Bit_Order Clauses
7854 @cindex Bit_Order Clause
7855 @cindex bit ordering
7856 @cindex ordering, of bits
7859 For record subtypes, GNAT permits the specification of the @code{Bit_Order}
7860 attribute. The specification may either correspond to the default bit
7861 order for the target, in which case the specification has no effect and
7862 places no additional restrictions, or it may be for the non-standard
7863 setting (that is the opposite of the default).
7865 In the case where the non-standard value is specified, the effect is
7866 to renumber bits within each byte, but the ordering of bytes is not
7867 affected. There are certain
7868 restrictions placed on component clauses as follows:
7872 @item Components fitting within a single storage unit.
7874 These are unrestricted, and the effect is merely to renumber bits. For
7875 example if we are on a little-endian machine with @code{Low_Order_First}
7876 being the default, then the following two declarations have exactly
7882 B : Integer range 1 .. 120;
7886 A at 0 range 0 .. 0;
7887 B at 0 range 1 .. 7;
7892 B : Integer range 1 .. 120;
7895 for R2'Bit_Order use High_Order_First;
7898 A at 0 range 7 .. 7;
7899 B at 0 range 0 .. 6;
7904 The useful application here is to write the second declaration with the
7905 @code{Bit_Order} attribute definition clause, and know that it will be treated
7906 the same, regardless of whether the target is little-endian or big-endian.
7908 @item Components occupying an integral number of bytes.
7910 These are components that exactly fit in two or more bytes. Such component
7911 declarations are allowed, but have no effect, since it is important to realize
7912 that the @code{Bit_Order} specification does not affect the ordering of bytes.
7913 In particular, the following attempt at getting an endian-independent integer
7921 for R2'Bit_Order use High_Order_First;
7924 A at 0 range 0 .. 31;
7929 This declaration will result in a little-endian integer on a
7930 little-endian machine, and a big-endian integer on a big-endian machine.
7931 If byte flipping is required for interoperability between big- and
7932 little-endian machines, this must be explicitly programmed. This capability
7933 is not provided by @code{Bit_Order}.
7935 @item Components that are positioned across byte boundaries
7937 but do not occupy an integral number of bytes. Given that bytes are not
7938 reordered, such fields would occupy a non-contiguous sequence of bits
7939 in memory, requiring non-trivial code to reassemble. They are for this
7940 reason not permitted, and any component clause specifying such a layout
7941 will be flagged as illegal by GNAT@.
7946 Since the misconception that Bit_Order automatically deals with all
7947 endian-related incompatibilities is a common one, the specification of
7948 a component field that is an integral number of bytes will always
7949 generate a warning. This warning may be suppressed using
7950 @code{pragma Suppress} if desired. The following section contains additional
7951 details regarding the issue of byte ordering.
7953 @node Effect of Bit_Order on Byte Ordering
7954 @section Effect of Bit_Order on Byte Ordering
7955 @cindex byte ordering
7956 @cindex ordering, of bytes
7959 In this section we will review the effect of the @code{Bit_Order} attribute
7960 definition clause on byte ordering. Briefly, it has no effect at all, but
7961 a detailed example will be helpful. Before giving this
7962 example, let us review the precise
7963 definition of the effect of defining @code{Bit_Order}. The effect of a
7964 non-standard bit order is described in section 15.5.3 of the Ada
7968 2 A bit ordering is a method of interpreting the meaning of
7969 the storage place attributes.
7973 To understand the precise definition of storage place attributes in
7974 this context, we visit section 13.5.1 of the manual:
7977 13 A record_representation_clause (without the mod_clause)
7978 specifies the layout. The storage place attributes (see 13.5.2)
7979 are taken from the values of the position, first_bit, and last_bit
7980 expressions after normalizing those values so that first_bit is
7981 less than Storage_Unit.
7985 The critical point here is that storage places are taken from
7986 the values after normalization, not before. So the @code{Bit_Order}
7987 interpretation applies to normalized values. The interpretation
7988 is described in the later part of the 15.5.3 paragraph:
7991 2 A bit ordering is a method of interpreting the meaning of
7992 the storage place attributes. High_Order_First (known in the
7993 vernacular as ``big endian'') means that the first bit of a
7994 storage element (bit 0) is the most significant bit (interpreting
7995 the sequence of bits that represent a component as an unsigned
7996 integer value). Low_Order_First (known in the vernacular as
7997 ``little endian'') means the opposite: the first bit is the
8002 Note that the numbering is with respect to the bits of a storage
8003 unit. In other words, the specification affects only the numbering
8004 of bits within a single storage unit.
8006 We can make the effect clearer by giving an example.
8008 Suppose that we have an external device which presents two bytes, the first
8009 byte presented, which is the first (low addressed byte) of the two byte
8010 record is called Master, and the second byte is called Slave.
8012 The left most (most significant bit is called Control for each byte, and
8013 the remaining 7 bits are called V1, V2, @dots{} V7, where V7 is the rightmost
8014 (least significant) bit.
8016 On a big-endian machine, we can write the following representation clause
8020 Master_Control : Bit;
8028 Slave_Control : Bit;
8039 Master_Control at 0 range 0 .. 0;
8040 Master_V1 at 0 range 1 .. 1;
8041 Master_V2 at 0 range 2 .. 2;
8042 Master_V3 at 0 range 3 .. 3;
8043 Master_V4 at 0 range 4 .. 4;
8044 Master_V5 at 0 range 5 .. 5;
8045 Master_V6 at 0 range 6 .. 6;
8046 Master_V7 at 0 range 7 .. 7;
8047 Slave_Control at 1 range 0 .. 0;
8048 Slave_V1 at 1 range 1 .. 1;
8049 Slave_V2 at 1 range 2 .. 2;
8050 Slave_V3 at 1 range 3 .. 3;
8051 Slave_V4 at 1 range 4 .. 4;
8052 Slave_V5 at 1 range 5 .. 5;
8053 Slave_V6 at 1 range 6 .. 6;
8054 Slave_V7 at 1 range 7 .. 7;
8059 Now if we move this to a little endian machine, then the bit ordering within
8060 the byte is backwards, so we have to rewrite the record rep clause as:
8064 Master_Control at 0 range 7 .. 7;
8065 Master_V1 at 0 range 6 .. 6;
8066 Master_V2 at 0 range 5 .. 5;
8067 Master_V3 at 0 range 4 .. 4;
8068 Master_V4 at 0 range 3 .. 3;
8069 Master_V5 at 0 range 2 .. 2;
8070 Master_V6 at 0 range 1 .. 1;
8071 Master_V7 at 0 range 0 .. 0;
8072 Slave_Control at 1 range 7 .. 7;
8073 Slave_V1 at 1 range 6 .. 6;
8074 Slave_V2 at 1 range 5 .. 5;
8075 Slave_V3 at 1 range 4 .. 4;
8076 Slave_V4 at 1 range 3 .. 3;
8077 Slave_V5 at 1 range 2 .. 2;
8078 Slave_V6 at 1 range 1 .. 1;
8079 Slave_V7 at 1 range 0 .. 0;
8083 It is a nuisance to have to rewrite the clause, especially if
8084 the code has to be maintained on both machines. However,
8085 this is a case that we can handle with the
8086 @code{Bit_Order} attribute if it is implemented.
8087 Note that the implementation is not required on byte addressed
8088 machines, but it is indeed implemented in GNAT.
8089 This means that we can simply use the
8090 first record clause, together with the declaration
8093 for Data'Bit_Order use High_Order_First;
8097 and the effect is what is desired, namely the layout is exactly the same,
8098 independent of whether the code is compiled on a big-endian or little-endian
8101 The important point to understand is that byte ordering is not affected.
8102 A @code{Bit_Order} attribute definition never affects which byte a field
8103 ends up in, only where it ends up in that byte.
8104 To make this clear, let us rewrite the record rep clause of the previous
8108 for Data'Bit_Order use High_Order_First;
8110 Master_Control at 0 range 0 .. 0;
8111 Master_V1 at 0 range 1 .. 1;
8112 Master_V2 at 0 range 2 .. 2;
8113 Master_V3 at 0 range 3 .. 3;
8114 Master_V4 at 0 range 4 .. 4;
8115 Master_V5 at 0 range 5 .. 5;
8116 Master_V6 at 0 range 6 .. 6;
8117 Master_V7 at 0 range 7 .. 7;
8118 Slave_Control at 0 range 8 .. 8;
8119 Slave_V1 at 0 range 9 .. 9;
8120 Slave_V2 at 0 range 10 .. 10;
8121 Slave_V3 at 0 range 11 .. 11;
8122 Slave_V4 at 0 range 12 .. 12;
8123 Slave_V5 at 0 range 13 .. 13;
8124 Slave_V6 at 0 range 14 .. 14;
8125 Slave_V7 at 0 range 15 .. 15;
8130 This is exactly equivalent to saying (a repeat of the first example):
8133 for Data'Bit_Order use High_Order_First;
8135 Master_Control at 0 range 0 .. 0;
8136 Master_V1 at 0 range 1 .. 1;
8137 Master_V2 at 0 range 2 .. 2;
8138 Master_V3 at 0 range 3 .. 3;
8139 Master_V4 at 0 range 4 .. 4;
8140 Master_V5 at 0 range 5 .. 5;
8141 Master_V6 at 0 range 6 .. 6;
8142 Master_V7 at 0 range 7 .. 7;
8143 Slave_Control at 1 range 0 .. 0;
8144 Slave_V1 at 1 range 1 .. 1;
8145 Slave_V2 at 1 range 2 .. 2;
8146 Slave_V3 at 1 range 3 .. 3;
8147 Slave_V4 at 1 range 4 .. 4;
8148 Slave_V5 at 1 range 5 .. 5;
8149 Slave_V6 at 1 range 6 .. 6;
8150 Slave_V7 at 1 range 7 .. 7;
8155 Why are they equivalent? Well take a specific field, the @code{Slave_V2}
8156 field. The storage place attributes are obtained by normalizing the
8157 values given so that the @code{First_Bit} value is less than 8. After
8158 nromalizing the values (0,10,10) we get (1,2,2) which is exactly what
8159 we specified in the other case.
8161 Now one might expect that the @code{Bit_Order} attribute might affect
8162 bit numbering within the entire record component (two bytes in this
8163 case, thus affecting which byte fields end up in), but that is not
8164 the way this feature is defined, it only affects numbering of bits,
8165 not which byte they end up in.
8167 Consequently it never makes sense to specify a starting bit number
8168 greater than 7 (for a byte addressable field) if an attribute
8169 definition for @code{Bit_Order} has been given, and indeed it
8170 may be actively confusing to specify such a value, so the compiler
8171 generates a warning for such usage.
8173 If you do need to control byte ordering then appropriate conditional
8174 values must be used. If in our example, the slave byte came first on
8175 some machines we might write:
8178 Master_Byte_First constant Boolean := @dots{};
8180 Master_Byte : constant Natural :=
8181 1 - Boolean'Pos (Master_Byte_First);
8182 Slave_Byte : constant Natural :=
8183 Boolean'Pos (Master_Byte_First);
8185 for Data'Bit_Order use High_Order_First;
8187 Master_Control at Master_Byte range 0 .. 0;
8188 Master_V1 at Master_Byte range 1 .. 1;
8189 Master_V2 at Master_Byte range 2 .. 2;
8190 Master_V3 at Master_Byte range 3 .. 3;
8191 Master_V4 at Master_Byte range 4 .. 4;
8192 Master_V5 at Master_Byte range 5 .. 5;
8193 Master_V6 at Master_Byte range 6 .. 6;
8194 Master_V7 at Master_Byte range 7 .. 7;
8195 Slave_Control at Slave_Byte range 0 .. 0;
8196 Slave_V1 at Slave_Byte range 1 .. 1;
8197 Slave_V2 at Slave_Byte range 2 .. 2;
8198 Slave_V3 at Slave_Byte range 3 .. 3;
8199 Slave_V4 at Slave_Byte range 4 .. 4;
8200 Slave_V5 at Slave_Byte range 5 .. 5;
8201 Slave_V6 at Slave_Byte range 6 .. 6;
8202 Slave_V7 at Slave_Byte range 7 .. 7;
8207 Now to switch between machines, all that is necessary is
8208 to set the boolean constant @code{Master_Byte_First} in
8209 an appropriate manner.
8211 @node Pragma Pack for Arrays
8212 @section Pragma Pack for Arrays
8213 @cindex Pragma Pack (for arrays)
8216 Pragma @code{Pack} applied to an array has no effect unless the component type
8217 is packable. For a component type to be packable, it must be one of the
8224 Any fixed-point type
8226 Any type whose size is specified with a size clause
8228 Any packed array type with a static size
8232 For all these cases, if the component subtype size is in the range
8233 1 through 63, then the effect of the pragma @code{Pack} is exactly as though a
8234 component size were specified giving the component subtype size.
8235 For example if we have:
8238 type r is range 0 .. 17;
8240 type ar is array (1 .. 8) of r;
8245 Then the component size of @code{ar} will be set to 5 (i.e.@: to @code{r'size},
8246 and the size of the array @code{ar} will be exactly 40 bits.
8248 Note that in some cases this rather fierce approach to packing can produce
8249 unexpected effects. For example, in Ada 95, type Natural typically has a
8250 size of 31, meaning that if you pack an array of Natural, you get 31-bit
8251 close packing, which saves a few bits, but results in far less efficient
8252 access. Since many other Ada compilers will ignore such a packing request,
8253 GNAT will generate a warning on some uses of pragma @code{Pack} that it guesses
8254 might not be what is intended. You can easily remove this warning by
8255 using an explicit @code{Component_Size} setting instead, which never generates
8256 a warning, since the intention of the programmer is clear in this case.
8258 GNAT treats packed arrays in one of two ways. If the size of the array is
8259 known at compile time and is less than 64 bits, then internally the array
8260 is represented as a single modular type, of exactly the appropriate number
8261 of bits. If the length is greater than 63 bits, or is not known at compile
8262 time, then the packed array is represented as an array of bytes, and the
8263 length is always a multiple of 8 bits.
8265 @node Pragma Pack for Records
8266 @section Pragma Pack for Records
8267 @cindex Pragma Pack (for records)
8270 Pragma @code{Pack} applied to a record will pack the components to reduce wasted
8271 space from alignment gaps and by reducing the amount of space taken by
8272 components. We distinguish between package components and non-packable
8273 components. Components of the following types are considered packable:
8277 All scalar types are packable.
8280 All fixed-point types are represented internally as integers, and
8284 Small packed arrays, whose size does not exceed 64 bits, and where the
8285 size is statically known at compile time, are represented internally
8286 as modular integers, and so they are also packable.
8291 All packable components occupy the exact number of bits corresponding to
8292 their @code{Size} value, and are packed with no padding bits, i.e.@: they
8293 can start on an arbitrary bit boundary.
8295 All other types are non-packable, they occupy an integral number of
8297 are placed at a boundary corresponding to their alignment requirements.
8299 For example, consider the record
8302 type Rb1 is array (1 .. 13) of Boolean;
8305 type Rb2 is array (1 .. 65) of Boolean;
8320 The representation for the record x2 is as follows:
8323 for x2'Size use 224;
8325 l1 at 0 range 0 .. 0;
8326 l2 at 0 range 1 .. 64;
8327 l3 at 12 range 0 .. 31;
8328 l4 at 16 range 0 .. 0;
8329 l5 at 16 range 1 .. 13;
8330 l6 at 18 range 0 .. 71;
8335 Studying this example, we see that the packable fields @code{l1}
8337 of length equal to their sizes, and placed at specific bit boundaries (and
8338 not byte boundaries) to
8339 eliminate padding. But @code{l3} is of a non-packable float type, so
8340 it is on the next appropriate alignment boundary.
8342 The next two fields are fully packable, so @code{l4} and @code{l5} are
8343 minimally packed with no gaps. However, type @code{Rb2} is a packed
8344 array that is longer than 64 bits, so it is itself non-packable. Thus
8345 the @code{l6} field is aligned to the next byte boundary, and takes an
8346 integral number of bytes, i.e.@: 72 bits.
8348 @node Record Representation Clauses
8349 @section Record Representation Clauses
8350 @cindex Record Representation Clause
8353 Record representation clauses may be given for all record types, including
8354 types obtained by record extension. Component clauses are allowed for any
8355 static component. The restrictions on component clauses depend on the type
8358 @cindex Component Clause
8359 For all components of an elementary type, the only restriction on component
8360 clauses is that the size must be at least the 'Size value of the type
8361 (actually the Value_Size). There are no restrictions due to alignment,
8362 and such components may freely cross storage boundaries.
8364 Packed arrays with a size up to and including 64 bits are represented
8365 internally using a modular type with the appropriate number of bits, and
8366 thus the same lack of restriction applies. For example, if you declare:
8369 type R is array (1 .. 49) of Boolean;
8375 then a component clause for a component of type R may start on any
8376 specified bit boundary, and may specify a value of 49 bits or greater.
8378 For non-primitive types, including packed arrays with a size greater than
8379 64 bits, component clauses must respect the alignment requirement of the
8380 type, in particular, always starting on a byte boundary, and the length
8381 must be a multiple of the storage unit.
8383 The tag field of a tagged type always occupies an address sized field at
8384 the start of the record. No component clause may attempt to overlay this
8387 In the case of a record extension T1, of a type T, no component clause applied
8388 to the type T1 can specify a storage location that would overlap the first
8389 T'Size bytes of the record.
8391 @node Enumeration Clauses
8392 @section Enumeration Clauses
8394 The only restriction on enumeration clauses is that the range of values
8395 must be representable. For the signed case, if one or more of the
8396 representation values are negative, all values must be in the range:
8399 System.Min_Int .. System.Max_Int
8403 For the unsigned case, where all values are non negative, the values must
8407 0 .. System.Max_Binary_Modulus;
8411 A @emph{confirming} representation clause is one in which the values range
8412 from 0 in sequence, i.e.@: a clause that confirms the default representation
8413 for an enumeration type.
8414 Such a confirming representation
8415 is permitted by these rules, and is specially recognized by the compiler so
8416 that no extra overhead results from the use of such a clause.
8418 If an array has an index type which is an enumeration type to which an
8419 enumeration clause has been applied, then the array is stored in a compact
8420 manner. Consider the declarations:
8423 type r is (A, B, C);
8424 for r use (A => 1, B => 5, C => 10);
8425 type t is array (r) of Character;
8429 The array type t corresponds to a vector with exactly three elements and
8430 has a default size equal to @code{3*Character'Size}. This ensures efficient
8431 use of space, but means that accesses to elements of the array will incur
8432 the overhead of converting representation values to the corresponding
8433 positional values, (i.e.@: the value delivered by the @code{Pos} attribute).
8435 @node Address Clauses
8436 @section Address Clauses
8437 @cindex Address Clause
8439 The reference manual allows a general restriction on representation clauses,
8440 as found in RM 13.1(22):
8443 An implementation need not support representation
8444 items containing nonstatic expressions, except that
8445 an implementation should support a representation item
8446 for a given entity if each nonstatic expression in the
8447 representation item is a name that statically denotes
8448 a constant declared before the entity.
8452 In practice this is applicable only to address clauses, since this is the
8453 only case in which a non-static expression is permitted by the syntax. As
8454 the AARM notes in sections 13.1 (22.a-22.h):
8457 22.a Reason: This is to avoid the following sort
8460 22.b X : Integer := F(@dots{});
8461 Y : Address := G(@dots{});
8462 for X'Address use Y;
8464 22.c In the above, we have to evaluate the
8465 initialization expression for X before we
8466 know where to put the result. This seems
8467 like an unreasonable implementation burden.
8469 22.d The above code should instead be written
8472 22.e Y : constant Address := G(@dots{});
8473 X : Integer := F(@dots{});
8474 for X'Address use Y;
8476 22.f This allows the expression ``Y'' to be safely
8477 evaluated before X is created.
8479 22.g The constant could be a formal parameter of mode in.
8481 22.h An implementation can support other nonstatic
8482 expressions if it wants to. Expressions of type
8483 Address are hardly ever static, but their value
8484 might be known at compile time anyway in many
8489 GNAT does indeed permit many additional cases of non-static expressions. In
8490 particular, if the type involved is elementary there are no restrictions
8491 (since in this case, holding a temporary copy of the initialization value,
8492 if one is present, is inexpensive). In addition, if there is no implicit or
8493 explicit initialization, then there are no restrictions. GNAT will reject
8494 only the case where all three of these conditions hold:
8499 The type of the item is non-elementary (e.g.@: a record or array).
8502 There is explicit or implicit initialization required for the object.
8505 The address value is non-static. Here GNAT is more permissive than the
8506 RM, and allows the address value to be the address of a previously declared
8507 stand-alone variable, as long as it does not itself have an address clause.
8510 Anchor : Some_Initialized_Type;
8511 Overlay : Some_Initialized_Type;
8512 for Overlay'Address use Anchor'Address;
8515 However, the prefix of the address clause cannot be an array component, or
8516 a component of a discriminated record.
8521 As noted above in section 22.h, address values are typically non-static. In
8522 particular the To_Address function, even if applied to a literal value, is
8523 a non-static function call. To avoid this minor annoyance, GNAT provides
8524 the implementation defined attribute 'To_Address. The following two
8525 expressions have identical values:
8527 Another issue with address clauses is the interaction with alignment
8528 requirements. When an address clause is given for an object, the address
8529 value must be consistent with the alignment of the object (which is usually
8530 the same as the alignment of the type of the object). If an address clause
8531 is given that specifies an inappropriately aligned address value, then the
8532 program execution is erroneous.
8534 Since this source of erroneous behavior can have unfortunate effects, GNAT
8535 checks (at compile time if possible, generating a warning, or at execution
8536 time with a run-time check) that the alignment is appropriate. If the
8537 run-time check fails, then @code{Program_Error} is raised. This run-time
8538 check is suppressed if range checks are suppressed, or if
8539 @code{pragma Restrictions (No_Elaboration_Code)} is in effect.
8544 To_Address (16#1234_0000#)
8545 System'To_Address (16#1234_0000#);
8549 except that the second form is considered to be a static expression, and
8550 thus when used as an address clause value is always permitted.
8553 Additionally, GNAT treats as static an address clause that is an
8554 unchecked_conversion of a static integer value. This simplifies the porting
8555 of legacy code, and provides a portable equivalent to the GNAT attribute
8559 An address clause cannot be given for an exported object. More
8560 understandably the real restriction is that objects with an address
8561 clause cannot be exported. This is because such variables are not
8562 defined by the Ada program, so there is no external object so export.
8565 It is permissible to give an address clause and a pragma Import for the
8566 same object. In this case, the variable is not really defined by the
8567 Ada program, so there is no external symbol to be linked. The link name
8568 and the external name are ignored in this case. The reason that we allow this
8569 combination is that it provides a useful idiom to avoid unwanted
8570 initializations on objects with address clauses.
8572 When an address clause is given for an object that has implicit or
8573 explicit initialization, then by default initialization takes place. This
8574 means that the effect of the object declaration is to overwrite the
8575 memory at the specified address. This is almost always not what the
8576 programmer wants, so GNAT will output a warning:
8586 for Ext'Address use System'To_Address (16#1234_1234#);
8588 >>> warning: implicit initialization of "Ext" may
8589 modify overlaid storage
8590 >>> warning: use pragma Import for "Ext" to suppress
8591 initialization (RM B(24))
8597 As indicated by the warning message, the solution is to use a (dummy) pragma
8598 Import to suppress this initialization. The pragma tell the compiler that the
8599 object is declared and initialized elsewhere. The following package compiles
8600 without warnings (and the initialization is suppressed):
8610 for Ext'Address use System'To_Address (16#1234_1234#);
8611 pragma Import (Ada, Ext);
8615 @node Effect of Convention on Representation
8616 @section Effect of Convention on Representation
8617 @cindex Convention, effect on representation
8620 Normally the specification of a foreign language convention for a type or
8621 an object has no effect on the chosen representation. In particular, the
8622 representation chosen for data in GNAT generally meets the standard system
8623 conventions, and for example records are laid out in a manner that is
8624 consistent with C@. This means that specifying convention C (for example)
8627 There are three exceptions to this general rule:
8631 @item Convention Fortran and array subtypes
8632 If pragma Convention Fortran is specified for an array subtype, then in
8633 accordance with the implementation advice in section 3.6.2(11) of the
8634 Ada Reference Manual, the array will be stored in a Fortran-compatible
8635 column-major manner, instead of the normal default row-major order.
8637 @item Convention C and enumeration types
8638 GNAT normally stores enumeration types in 8, 16, or 32 bits as required
8639 to accommodate all values of the type. For example, for the enumeration
8643 type Color is (Red, Green, Blue);
8647 8 bits is sufficient to store all values of the type, so by default, objects
8648 of type @code{Color} will be represented using 8 bits. However, normal C
8649 convention is to use 32 bits for all enum values in C, since enum values
8650 are essentially of type int. If pragma @code{Convention C} is specified for an
8651 Ada enumeration type, then the size is modified as necessary (usually to
8652 32 bits) to be consistent with the C convention for enum values.
8654 @item Convention C/Fortran and Boolean types
8655 In C, the usual convention for boolean values, that is values used for
8656 conditions, is that zero represents false, and nonzero values represent
8657 true. In Ada, the normal convention is that two specific values, typically
8658 0/1, are used to represent false/true respectively.
8660 Fortran has a similar convention for @code{LOGICAL} values (any nonzero
8661 value represents true).
8663 To accommodate the Fortran and C conventions, if a pragma Convention specifies
8664 C or Fortran convention for a derived Boolean, as in the following example:
8667 type C_Switch is new Boolean;
8668 pragma Convention (C, C_Switch);
8672 then the GNAT generated code will treat any nonzero value as true. For truth
8673 values generated by GNAT, the conventional value 1 will be used for True, but
8674 when one of these values is read, any nonzero value is treated as True.
8678 @node Determining the Representations chosen by GNAT
8679 @section Determining the Representations chosen by GNAT
8680 @cindex Representation, determination of
8681 @cindex @code{-gnatR} switch
8684 Although the descriptions in this section are intended to be complete, it is
8685 often easier to simply experiment to see what GNAT accepts and what the
8686 effect is on the layout of types and objects.
8688 As required by the Ada RM, if a representation clause is not accepted, then
8689 it must be rejected as illegal by the compiler. However, when a representation
8690 clause or pragma is accepted, there can still be questions of what the
8691 compiler actually does. For example, if a partial record representation
8692 clause specifies the location of some components and not others, then where
8693 are the non-specified components placed? Or if pragma @code{Pack} is used on a
8694 record, then exactly where are the resulting fields placed? The section
8695 on pragma @code{Pack} in this chapter can be used to answer the second question,
8696 but it is often easier to just see what the compiler does.
8698 For this purpose, GNAT provides the option @code{-gnatR}. If you compile
8699 with this option, then the compiler will output information on the actual
8700 representations chosen, in a format similar to source representation
8701 clauses. For example, if we compile the package:
8705 type r (x : boolean) is tagged record
8707 when True => S : String (1 .. 100);
8712 type r2 is new r (false) with record
8717 y2 at 16 range 0 .. 31;
8724 type x1 is array (1 .. 10) of x;
8725 for x1'component_size use 11;
8727 type ia is access integer;
8729 type Rb1 is array (1 .. 13) of Boolean;
8732 type Rb2 is array (1 .. 65) of Boolean;
8748 using the switch @code{-gnatR} we obtain the following output:
8751 Representation information for unit q
8752 -------------------------------------
8755 for r'Alignment use 4;
8757 x at 4 range 0 .. 7;
8758 _tag at 0 range 0 .. 31;
8759 s at 5 range 0 .. 799;
8762 for r2'Size use 160;
8763 for r2'Alignment use 4;
8765 x at 4 range 0 .. 7;
8766 _tag at 0 range 0 .. 31;
8767 _parent at 0 range 0 .. 63;
8768 y2 at 16 range 0 .. 31;
8772 for x'Alignment use 1;
8774 y at 0 range 0 .. 7;
8777 for x1'Size use 112;
8778 for x1'Alignment use 1;
8779 for x1'Component_Size use 11;
8781 for rb1'Size use 13;
8782 for rb1'Alignment use 2;
8783 for rb1'Component_Size use 1;
8785 for rb2'Size use 72;
8786 for rb2'Alignment use 1;
8787 for rb2'Component_Size use 1;
8789 for x2'Size use 224;
8790 for x2'Alignment use 4;
8792 l1 at 0 range 0 .. 0;
8793 l2 at 0 range 1 .. 64;
8794 l3 at 12 range 0 .. 31;
8795 l4 at 16 range 0 .. 0;
8796 l5 at 16 range 1 .. 13;
8797 l6 at 18 range 0 .. 71;
8802 The Size values are actually the Object_Size, i.e.@: the default size that
8803 will be allocated for objects of the type.
8804 The ?? size for type r indicates that we have a variant record, and the
8805 actual size of objects will depend on the discriminant value.
8807 The Alignment values show the actual alignment chosen by the compiler
8808 for each record or array type.
8810 The record representation clause for type r shows where all fields
8811 are placed, including the compiler generated tag field (whose location
8812 cannot be controlled by the programmer).
8814 The record representation clause for the type extension r2 shows all the
8815 fields present, including the parent field, which is a copy of the fields
8816 of the parent type of r2, i.e.@: r1.
8818 The component size and size clauses for types rb1 and rb2 show
8819 the exact effect of pragma @code{Pack} on these arrays, and the record
8820 representation clause for type x2 shows how pragma @code{Pack} affects
8823 In some cases, it may be useful to cut and paste the representation clauses
8824 generated by the compiler into the original source to fix and guarantee
8825 the actual representation to be used.
8827 @node Standard Library Routines
8828 @chapter Standard Library Routines
8831 The Ada 95 Reference Manual contains in Annex A a full description of an
8832 extensive set of standard library routines that can be used in any Ada
8833 program, and which must be provided by all Ada compilers. They are
8834 analogous to the standard C library used by C programs.
8836 GNAT implements all of the facilities described in annex A, and for most
8837 purposes the description in the Ada 95
8838 reference manual, or appropriate Ada
8839 text book, will be sufficient for making use of these facilities.
8841 In the case of the input-output facilities, @xref{The Implementation of
8842 Standard I/O}, gives details on exactly how GNAT interfaces to the
8843 file system. For the remaining packages, the Ada 95 reference manual
8844 should be sufficient. The following is a list of the packages included,
8845 together with a brief description of the functionality that is provided.
8847 For completeness, references are included to other predefined library
8848 routines defined in other sections of the Ada 95 reference manual (these are
8849 cross-indexed from annex A).
8853 This is a parent package for all the standard library packages. It is
8854 usually included implicitly in your program, and itself contains no
8855 useful data or routines.
8857 @item Ada.Calendar (9.6)
8858 @code{Calendar} provides time of day access, and routines for
8859 manipulating times and durations.
8861 @item Ada.Characters (A.3.1)
8862 This is a dummy parent package that contains no useful entities
8864 @item Ada.Characters.Handling (A.3.2)
8865 This package provides some basic character handling capabilities,
8866 including classification functions for classes of characters (e.g.@: test
8867 for letters, or digits).
8869 @item Ada.Characters.Latin_1 (A.3.3)
8870 This package includes a complete set of definitions of the characters
8871 that appear in type CHARACTER@. It is useful for writing programs that
8872 will run in international environments. For example, if you want an
8873 upper case E with an acute accent in a string, it is often better to use
8874 the definition of @code{UC_E_Acute} in this package. Then your program
8875 will print in an understandable manner even if your environment does not
8876 support these extended characters.
8878 @item Ada.Command_Line (A.15)
8879 This package provides access to the command line parameters and the name
8880 of the current program (analogous to the use of @code{argc} and @code{argv} in C), and
8881 also allows the exit status for the program to be set in a
8882 system-independent manner.
8884 @item Ada.Decimal (F.2)
8885 This package provides constants describing the range of decimal numbers
8886 implemented, and also a decimal divide routine (analogous to the COBOL
8887 verb DIVIDE .. GIVING .. REMAINDER ..)
8889 @item Ada.Direct_IO (A.8.4)
8890 This package provides input-output using a model of a set of records of
8891 fixed-length, containing an arbitrary definite Ada type, indexed by an
8892 integer record number.
8894 @item Ada.Dynamic_Priorities (D.5)
8895 This package allows the priorities of a task to be adjusted dynamically
8896 as the task is running.
8898 @item Ada.Exceptions (11.4.1)
8899 This package provides additional information on exceptions, and also
8900 contains facilities for treating exceptions as data objects, and raising
8901 exceptions with associated messages.
8903 @item Ada.Finalization (7.6)
8904 This package contains the declarations and subprograms to support the
8905 use of controlled types, providing for automatic initialization and
8906 finalization (analogous to the constructors and destructors of C++)
8908 @item Ada.Interrupts (C.3.2)
8909 This package provides facilities for interfacing to interrupts, which
8910 includes the set of signals or conditions that can be raised and
8911 recognized as interrupts.
8913 @item Ada.Interrupts.Names (C.3.2)
8914 This package provides the set of interrupt names (actually signal
8915 or condition names) that can be handled by GNAT@.
8917 @item Ada.IO_Exceptions (A.13)
8918 This package defines the set of exceptions that can be raised by use of
8919 the standard IO packages.
8922 This package contains some standard constants and exceptions used
8923 throughout the numerics packages. Note that the constants pi and e are
8924 defined here, and it is better to use these definitions than rolling
8927 @item Ada.Numerics.Complex_Elementary_Functions
8928 Provides the implementation of standard elementary functions (such as
8929 log and trigonometric functions) operating on complex numbers using the
8930 standard @code{Float} and the @code{Complex} and @code{Imaginary} types
8931 created by the package @code{Numerics.Complex_Types}.
8933 @item Ada.Numerics.Complex_Types
8934 This is a predefined instantiation of
8935 @code{Numerics.Generic_Complex_Types} using @code{Standard.Float} to
8936 build the type @code{Complex} and @code{Imaginary}.
8938 @item Ada.Numerics.Discrete_Random
8939 This package provides a random number generator suitable for generating
8940 random integer values from a specified range.
8942 @item Ada.Numerics.Float_Random
8943 This package provides a random number generator suitable for generating
8944 uniformly distributed floating point values.
8946 @item Ada.Numerics.Generic_Complex_Elementary_Functions
8947 This is a generic version of the package that provides the
8948 implementation of standard elementary functions (such as log and
8949 trigonometric functions) for an arbitrary complex type.
8951 The following predefined instantiations of this package are provided:
8955 @code{Ada.Numerics.Short_Complex_Elementary_Functions}
8957 @code{Ada.Numerics.Complex_Elementary_Functions}
8960 Long_Complex_Elementary_Functions}
8963 @item Ada.Numerics.Generic_Complex_Types
8964 This is a generic package that allows the creation of complex types,
8965 with associated complex arithmetic operations.
8967 The following predefined instantiations of this package exist
8970 @code{Ada.Numerics.Short_Complex_Complex_Types}
8972 @code{Ada.Numerics.Complex_Complex_Types}
8974 @code{Ada.Numerics.Long_Complex_Complex_Types}
8977 @item Ada.Numerics.Generic_Elementary_Functions
8978 This is a generic package that provides the implementation of standard
8979 elementary functions (such as log an trigonometric functions) for an
8980 arbitrary float type.
8982 The following predefined instantiations of this package exist
8986 @code{Ada.Numerics.Short_Elementary_Functions}
8988 @code{Ada.Numerics.Elementary_Functions}
8990 @code{Ada.Numerics.Long_Elementary_Functions}
8993 @item Ada.Real_Time (D.8)
8994 This package provides facilities similar to those of @code{Calendar}, but
8995 operating with a finer clock suitable for real time control. Note that
8996 annex D requires that there be no backward clock jumps, and GNAT generally
8997 guarantees this behavior, but of course if the external clock on which
8998 the GNAT runtime depends is deliberately reset by some external event,
8999 then such a backward jump may occur.
9001 @item Ada.Sequential_IO (A.8.1)
9002 This package provides input-output facilities for sequential files,
9003 which can contain a sequence of values of a single type, which can be
9004 any Ada type, including indefinite (unconstrained) types.
9006 @item Ada.Storage_IO (A.9)
9007 This package provides a facility for mapping arbitrary Ada types to and
9008 from a storage buffer. It is primarily intended for the creation of new
9011 @item Ada.Streams (13.13.1)
9012 This is a generic package that provides the basic support for the
9013 concept of streams as used by the stream attributes (@code{Input},
9014 @code{Output}, @code{Read} and @code{Write}).
9016 @item Ada.Streams.Stream_IO (A.12.1)
9017 This package is a specialization of the type @code{Streams} defined in
9018 package @code{Streams} together with a set of operations providing
9019 Stream_IO capability. The Stream_IO model permits both random and
9020 sequential access to a file which can contain an arbitrary set of values
9021 of one or more Ada types.
9023 @item Ada.Strings (A.4.1)
9024 This package provides some basic constants used by the string handling
9027 @item Ada.Strings.Bounded (A.4.4)
9028 This package provides facilities for handling variable length
9029 strings. The bounded model requires a maximum length. It is thus
9030 somewhat more limited than the unbounded model, but avoids the use of
9031 dynamic allocation or finalization.
9033 @item Ada.Strings.Fixed (A.4.3)
9034 This package provides facilities for handling fixed length strings.
9036 @item Ada.Strings.Maps (A.4.2)
9037 This package provides facilities for handling character mappings and
9038 arbitrarily defined subsets of characters. For instance it is useful in
9039 defining specialized translation tables.
9041 @item Ada.Strings.Maps.Constants (A.4.6)
9042 This package provides a standard set of predefined mappings and
9043 predefined character sets. For example, the standard upper to lower case
9044 conversion table is found in this package. Note that upper to lower case
9045 conversion is non-trivial if you want to take the entire set of
9046 characters, including extended characters like E with an acute accent,
9047 into account. You should use the mappings in this package (rather than
9048 adding 32 yourself) to do case mappings.
9050 @item Ada.Strings.Unbounded (A.4.5)
9051 This package provides facilities for handling variable length
9052 strings. The unbounded model allows arbitrary length strings, but
9053 requires the use of dynamic allocation and finalization.
9055 @item Ada.Strings.Wide_Bounded (A.4.7)
9056 @itemx Ada.Strings.Wide_Fixed (A.4.7)
9057 @itemx Ada.Strings.Wide_Maps (A.4.7)
9058 @itemx Ada.Strings.Wide_Maps.Constants (A.4.7)
9059 @itemx Ada.Strings.Wide_Unbounded (A.4.7)
9060 These package provide analogous capabilities to the corresponding
9061 packages without @samp{Wide_} in the name, but operate with the types
9062 @code{Wide_String} and @code{Wide_Character} instead of @code{String}
9063 and @code{Character}.
9065 @item Ada.Synchronous_Task_Control (D.10)
9066 This package provides some standard facilities for controlling task
9067 communication in a synchronous manner.
9070 This package contains definitions for manipulation of the tags of tagged
9073 @item Ada.Task_Attributes
9074 This package provides the capability of associating arbitrary
9075 task-specific data with separate tasks.
9078 This package provides basic text input-output capabilities for
9079 character, string and numeric data. The subpackages of this
9080 package are listed next.
9082 @item Ada.Text_IO.Decimal_IO
9083 Provides input-output facilities for decimal fixed-point types
9085 @item Ada.Text_IO.Enumeration_IO
9086 Provides input-output facilities for enumeration types.
9088 @item Ada.Text_IO.Fixed_IO
9089 Provides input-output facilities for ordinary fixed-point types.
9091 @item Ada.Text_IO.Float_IO
9092 Provides input-output facilities for float types. The following
9093 predefined instantiations of this generic package are available:
9097 @code{Short_Float_Text_IO}
9099 @code{Float_Text_IO}
9101 @code{Long_Float_Text_IO}
9104 @item Ada.Text_IO.Integer_IO
9105 Provides input-output facilities for integer types. The following
9106 predefined instantiations of this generic package are available:
9109 @item Short_Short_Integer
9110 @code{Ada.Short_Short_Integer_Text_IO}
9112 @code{Ada.Short_Integer_Text_IO}
9114 @code{Ada.Integer_Text_IO}
9116 @code{Ada.Long_Integer_Text_IO}
9117 @item Long_Long_Integer
9118 @code{Ada.Long_Long_Integer_Text_IO}
9121 @item Ada.Text_IO.Modular_IO
9122 Provides input-output facilities for modular (unsigned) types
9124 @item Ada.Text_IO.Complex_IO (G.1.3)
9125 This package provides basic text input-output capabilities for complex
9128 @item Ada.Text_IO.Editing (F.3.3)
9129 This package contains routines for edited output, analogous to the use
9130 of pictures in COBOL@. The picture formats used by this package are a
9131 close copy of the facility in COBOL@.
9133 @item Ada.Text_IO.Text_Streams (A.12.2)
9134 This package provides a facility that allows Text_IO files to be treated
9135 as streams, so that the stream attributes can be used for writing
9136 arbitrary data, including binary data, to Text_IO files.
9138 @item Ada.Unchecked_Conversion (13.9)
9139 This generic package allows arbitrary conversion from one type to
9140 another of the same size, providing for breaking the type safety in
9141 special circumstances.
9143 If the types have the same Size (more accurately the same Value_Size),
9144 then the effect is simply to transfer the bits from the source to the
9145 target type without any modification. This usage is well defined, and
9146 for simple types whose representation is typically the same across
9147 all implementations, gives a portable method of performing such
9150 If the types do not have the same size, then the result is implementation
9151 defined, and thus may be non-portable. The following describes how GNAT
9152 handles such unchecked conversion cases.
9154 If the types are of different sizes, and are both discrete types, then
9155 the effect is of a normal type conversion without any constraint checking.
9156 In particular if the result type has a larger size, the result will be
9157 zero or sign extended. If the result type has a smaller size, the result
9158 will be truncated by ignoring high order bits.
9160 If the types are of different sizes, and are not both discrete types,
9161 then the conversion works as though pointers were created to the source
9162 and target, and the pointer value is converted. The effect is that bits
9163 are copied from successive low order storage units and bits of the source
9164 up to the length of the target type.
9166 A warning is issued if the lengths differ, since the effect in this
9167 case is implementation dependent, and the above behavior may not match
9168 that of some other compiler.
9170 A pointer to one type may be converted to a pointer to another type using
9171 unchecked conversion. The only case in which the effect is undefined is
9172 when one or both pointers are pointers to unconstrained array types. In
9173 this case, the bounds information may get incorrectly transferred, and in
9174 particular, GNAT uses double size pointers for such types, and it is
9175 meaningless to convert between such pointer types. GNAT will issue a
9176 warning if the alignment of the target designated type is more strict
9177 than the alignment of the source designated type (since the result may
9178 be unaligned in this case).
9180 A pointer other than a pointer to an unconstrained array type may be
9181 converted to and from System.Address. Such usage is common in Ada 83
9182 programs, but note that Ada.Address_To_Access_Conversions is the
9183 preferred method of performing such conversions in Ada 95. Neither
9184 unchecked conversion nor Ada.Address_To_Access_Conversions should be
9185 used in conjunction with pointers to unconstrained objects, since
9186 the bounds information cannot be handled correctly in this case.
9188 @item Ada.Unchecked_Deallocation (13.11.2)
9189 This generic package allows explicit freeing of storage previously
9190 allocated by use of an allocator.
9192 @item Ada.Wide_Text_IO (A.11)
9193 This package is similar to @code{Ada.Text_IO}, except that the external
9194 file supports wide character representations, and the internal types are
9195 @code{Wide_Character} and @code{Wide_String} instead of @code{Character}
9196 and @code{String}. It contains generic subpackages listed next.
9198 @item Ada.Wide_Text_IO.Decimal_IO
9199 Provides input-output facilities for decimal fixed-point types
9201 @item Ada.Wide_Text_IO.Enumeration_IO
9202 Provides input-output facilities for enumeration types.
9204 @item Ada.Wide_Text_IO.Fixed_IO
9205 Provides input-output facilities for ordinary fixed-point types.
9207 @item Ada.Wide_Text_IO.Float_IO
9208 Provides input-output facilities for float types. The following
9209 predefined instantiations of this generic package are available:
9213 @code{Short_Float_Wide_Text_IO}
9215 @code{Float_Wide_Text_IO}
9217 @code{Long_Float_Wide_Text_IO}
9220 @item Ada.Wide_Text_IO.Integer_IO
9221 Provides input-output facilities for integer types. The following
9222 predefined instantiations of this generic package are available:
9225 @item Short_Short_Integer
9226 @code{Ada.Short_Short_Integer_Wide_Text_IO}
9228 @code{Ada.Short_Integer_Wide_Text_IO}
9230 @code{Ada.Integer_Wide_Text_IO}
9232 @code{Ada.Long_Integer_Wide_Text_IO}
9233 @item Long_Long_Integer
9234 @code{Ada.Long_Long_Integer_Wide_Text_IO}
9237 @item Ada.Wide_Text_IO.Modular_IO
9238 Provides input-output facilities for modular (unsigned) types
9240 @item Ada.Wide_Text_IO.Complex_IO (G.1.3)
9241 This package is similar to @code{Ada.Text_IO.Complex_IO}, except that the
9242 external file supports wide character representations.
9244 @item Ada.Wide_Text_IO.Editing (F.3.4)
9245 This package is similar to @code{Ada.Text_IO.Editing}, except that the
9246 types are @code{Wide_Character} and @code{Wide_String} instead of
9247 @code{Character} and @code{String}.
9249 @item Ada.Wide_Text_IO.Streams (A.12.3)
9250 This package is similar to @code{Ada.Text_IO.Streams}, except that the
9251 types are @code{Wide_Character} and @code{Wide_String} instead of
9252 @code{Character} and @code{String}.
9254 @node The Implementation of Standard I/O
9255 @chapter The Implementation of Standard I/O
9258 GNAT implements all the required input-output facilities described in
9259 A.6 through A.14. These sections of the Ada 95 reference manual describe the
9260 required behavior of these packages from the Ada point of view, and if
9261 you are writing a portable Ada program that does not need to know the
9262 exact manner in which Ada maps to the outside world when it comes to
9263 reading or writing external files, then you do not need to read this
9264 chapter. As long as your files are all regular files (not pipes or
9265 devices), and as long as you write and read the files only from Ada, the
9266 description in the Ada 95 reference manual is sufficient.
9268 However, if you want to do input-output to pipes or other devices, such
9269 as the keyboard or screen, or if the files you are dealing with are
9270 either generated by some other language, or to be read by some other
9271 language, then you need to know more about the details of how the GNAT
9272 implementation of these input-output facilities behaves.
9274 In this chapter we give a detailed description of exactly how GNAT
9275 interfaces to the file system. As always, the sources of the system are
9276 available to you for answering questions at an even more detailed level,
9277 but for most purposes the information in this chapter will suffice.
9279 Another reason that you may need to know more about how input-output is
9280 implemented arises when you have a program written in mixed languages
9281 where, for example, files are shared between the C and Ada sections of
9282 the same program. GNAT provides some additional facilities, in the form
9283 of additional child library packages, that facilitate this sharing, and
9284 these additional facilities are also described in this chapter.
9287 * Standard I/O Packages::
9296 * Operations on C Streams::
9297 * Interfacing to C Streams::
9300 @node Standard I/O Packages
9301 @section Standard I/O Packages
9304 The Standard I/O packages described in Annex A for
9310 Ada.Text_IO.Complex_IO
9312 Ada.Text_IO.Text_Streams,
9316 Ada.Wide_Text_IO.Complex_IO,
9318 Ada.Wide_Text_IO.Text_Streams
9328 are implemented using the C
9329 library streams facility; where
9333 All files are opened using @code{fopen}.
9335 All input/output operations use @code{fread}/@code{fwrite}.
9338 There is no internal buffering of any kind at the Ada library level. The
9339 only buffering is that provided at the system level in the
9340 implementation of the C library routines that support streams. This
9341 facilitates shared use of these streams by mixed language programs.
9344 @section FORM Strings
9347 The format of a FORM string in GNAT is:
9350 "keyword=value,keyword=value,@dots{},keyword=value"
9354 where letters may be in upper or lower case, and there are no spaces
9355 between values. The order of the entries is not important. Currently
9356 there are two keywords defined.
9363 The use of these parameters is described later in this section.
9369 Direct_IO can only be instantiated for definite types. This is a
9370 restriction of the Ada language, which means that the records are fixed
9371 length (the length being determined by @code{@var{type}'Size}, rounded
9372 up to the next storage unit boundary if necessary).
9374 The records of a Direct_IO file are simply written to the file in index
9375 sequence, with the first record starting at offset zero, and subsequent
9376 records following. There is no control information of any kind. For
9377 example, if 32-bit integers are being written, each record takes
9378 4-bytes, so the record at index @var{K} starts at offset
9379 (@var{K}@minus{}1)*4.
9381 There is no limit on the size of Direct_IO files, they are expanded as
9382 necessary to accommodate whatever records are written to the file.
9385 @section Sequential_IO
9388 Sequential_IO may be instantiated with either a definite (constrained)
9389 or indefinite (unconstrained) type.
9391 For the definite type case, the elements written to the file are simply
9392 the memory images of the data values with no control information of any
9393 kind. The resulting file should be read using the same type, no validity
9394 checking is performed on input.
9396 For the indefinite type case, the elements written consist of two
9397 parts. First is the size of the data item, written as the memory image
9398 of a @code{Interfaces.C.size_t} value, followed by the memory image of
9399 the data value. The resulting file can only be read using the same
9400 (unconstrained) type. Normal assignment checks are performed on these
9401 read operations, and if these checks fail, @code{Data_Error} is
9402 raised. In particular, in the array case, the lengths must match, and in
9403 the variant record case, if the variable for a particular read operation
9404 is constrained, the discriminants must match.
9406 Note that it is not possible to use Sequential_IO to write variable
9407 length array items, and then read the data back into different length
9408 arrays. For example, the following will raise @code{Data_Error}:
9411 package IO is new Sequential_IO (String);
9416 IO.Write (F, "hello!")
9417 IO.Reset (F, Mode=>In_File);
9423 On some Ada implementations, this will print @samp{hell}, but the program is
9424 clearly incorrect, since there is only one element in the file, and that
9425 element is the string @samp{hello!}.
9427 In Ada 95, this kind of behavior can be legitimately achieved using
9428 Stream_IO, and this is the preferred mechanism. In particular, the above
9429 program fragment rewritten to use Stream_IO will work correctly.
9435 Text_IO files consist of a stream of characters containing the following
9436 special control characters:
9439 LF (line feed, 16#0A#) Line Mark
9440 FF (form feed, 16#0C#) Page Mark
9443 A canonical Text_IO file is defined as one in which the following
9448 The character @code{LF} is used only as a line mark, i.e.@: to mark the end
9452 The character @code{FF} is used only as a page mark, i.e.@: to mark the
9453 end of a page and consequently can appear only immediately following a
9454 @code{LF} (line mark) character.
9457 The file ends with either @code{LF} (line mark) or @code{LF}-@code{FF}
9458 (line mark, page mark). In the former case, the page mark is implicitly
9459 assumed to be present.
9462 A file written using Text_IO will be in canonical form provided that no
9463 explicit @code{LF} or @code{FF} characters are written using @code{Put}
9464 or @code{Put_Line}. There will be no @code{FF} character at the end of
9465 the file unless an explicit @code{New_Page} operation was performed
9466 before closing the file.
9468 A canonical Text_IO file that is a regular file, i.e.@: not a device or a
9469 pipe, can be read using any of the routines in Text_IO@. The
9470 semantics in this case will be exactly as defined in the Ada 95 reference
9471 manual and all the routines in Text_IO are fully implemented.
9473 A text file that does not meet the requirements for a canonical Text_IO
9474 file has one of the following:
9478 The file contains @code{FF} characters not immediately following a
9479 @code{LF} character.
9482 The file contains @code{LF} or @code{FF} characters written by
9483 @code{Put} or @code{Put_Line}, which are not logically considered to be
9484 line marks or page marks.
9487 The file ends in a character other than @code{LF} or @code{FF},
9488 i.e.@: there is no explicit line mark or page mark at the end of the file.
9491 Text_IO can be used to read such non-standard text files but subprograms
9492 to do with line or page numbers do not have defined meanings. In
9493 particular, a @code{FF} character that does not follow a @code{LF}
9494 character may or may not be treated as a page mark from the point of
9495 view of page and line numbering. Every @code{LF} character is considered
9496 to end a line, and there is an implied @code{LF} character at the end of
9500 * Text_IO Stream Pointer Positioning::
9501 * Text_IO Reading and Writing Non-Regular Files::
9503 * Treating Text_IO Files as Streams::
9504 * Text_IO Extensions::
9505 * Text_IO Facilities for Unbounded Strings::
9507 @node Text_IO Stream Pointer Positioning
9509 @subsection Stream Pointer Positioning
9512 @code{Ada.Text_IO} has a definition of current position for a file that
9513 is being read. No internal buffering occurs in Text_IO, and usually the
9514 physical position in the stream used to implement the file corresponds
9515 to this logical position defined by Text_IO@. There are two exceptions:
9519 After a call to @code{End_Of_Page} that returns @code{True}, the stream
9520 is positioned past the @code{LF} (line mark) that precedes the page
9521 mark. Text_IO maintains an internal flag so that subsequent read
9522 operations properly handle the logical position which is unchanged by
9523 the @code{End_Of_Page} call.
9526 After a call to @code{End_Of_File} that returns @code{True}, if the
9527 Text_IO file was positioned before the line mark at the end of file
9528 before the call, then the logical position is unchanged, but the stream
9529 is physically positioned right at the end of file (past the line mark,
9530 and past a possible page mark following the line mark. Again Text_IO
9531 maintains internal flags so that subsequent read operations properly
9532 handle the logical position.
9535 These discrepancies have no effect on the observable behavior of
9536 Text_IO, but if a single Ada stream is shared between a C program and
9537 Ada program, or shared (using @samp{shared=yes} in the form string)
9538 between two Ada files, then the difference may be observable in some
9541 @node Text_IO Reading and Writing Non-Regular Files
9542 @subsection Reading and Writing Non-Regular Files
9545 A non-regular file is a device (such as a keyboard), or a pipe. Text_IO
9546 can be used for reading and writing. Writing is not affected and the
9547 sequence of characters output is identical to the normal file case, but
9548 for reading, the behavior of Text_IO is modified to avoid undesirable
9549 look-ahead as follows:
9551 An input file that is not a regular file is considered to have no page
9552 marks. Any @code{Ascii.FF} characters (the character normally used for a
9553 page mark) appearing in the file are considered to be data
9554 characters. In particular:
9558 @code{Get_Line} and @code{Skip_Line} do not test for a page mark
9559 following a line mark. If a page mark appears, it will be treated as a
9563 This avoids the need to wait for an extra character to be typed or
9564 entered from the pipe to complete one of these operations.
9567 @code{End_Of_Page} always returns @code{False}
9570 @code{End_Of_File} will return @code{False} if there is a page mark at
9571 the end of the file.
9574 Output to non-regular files is the same as for regular files. Page marks
9575 may be written to non-regular files using @code{New_Page}, but as noted
9576 above they will not be treated as page marks on input if the output is
9577 piped to another Ada program.
9579 Another important discrepancy when reading non-regular files is that the end
9580 of file indication is not ``sticky''. If an end of file is entered, e.g.@: by
9581 pressing the @key{EOT} key,
9583 is signalled once (i.e.@: the test @code{End_Of_File}
9584 will yield @code{True}, or a read will
9585 raise @code{End_Error}), but then reading can resume
9586 to read data past that end of
9587 file indication, until another end of file indication is entered.
9590 @subsection Get_Immediate
9591 @cindex Get_Immediate
9594 Get_Immediate returns the next character (including control characters)
9595 from the input file. In particular, Get_Immediate will return LF or FF
9596 characters used as line marks or page marks. Such operations leave the
9597 file positioned past the control character, and it is thus not treated
9598 as having its normal function. This means that page, line and column
9599 counts after this kind of Get_Immediate call are set as though the mark
9600 did not occur. In the case where a Get_Immediate leaves the file
9601 positioned between the line mark and page mark (which is not normally
9602 possible), it is undefined whether the FF character will be treated as a
9605 @node Treating Text_IO Files as Streams
9606 @subsection Treating Text_IO Files as Streams
9607 @cindex Stream files
9610 The package @code{Text_IO.Streams} allows a Text_IO file to be treated
9611 as a stream. Data written to a Text_IO file in this stream mode is
9612 binary data. If this binary data contains bytes 16#0A# (@code{LF}) or
9613 16#0C# (@code{FF}), the resulting file may have non-standard
9614 format. Similarly if read operations are used to read from a Text_IO
9615 file treated as a stream, then @code{LF} and @code{FF} characters may be
9616 skipped and the effect is similar to that described above for
9617 @code{Get_Immediate}.
9619 @node Text_IO Extensions
9620 @subsection Text_IO Extensions
9621 @cindex Text_IO extensions
9624 A package GNAT.IO_Aux in the GNAT library provides some useful extensions
9625 to the standard @code{Text_IO} package:
9628 @item function File_Exists (Name : String) return Boolean;
9629 Determines if a file of the given name exists and can be successfully
9630 opened (without actually performing the open operation).
9632 @item function Get_Line return String;
9633 Reads a string from the standard input file. The value returned is exactly
9634 the length of the line that was read.
9636 @item function Get_Line (File : Ada.Text_IO.File_Type) return String;
9637 Similar, except that the parameter File specifies the file from which
9638 the string is to be read.
9642 @node Text_IO Facilities for Unbounded Strings
9643 @subsection Text_IO Facilities for Unbounded Strings
9644 @cindex Text_IO for unbounded strings
9645 @cindex Unbounded_String, Text_IO operations
9648 The package @code{Ada.Strings.Unbounded.Text_IO}
9649 in library files @code{a-suteio.ads/adb} contains some GNAT-specific
9650 subprograms useful for Text_IO operations on unbounded strings:
9654 @item function Get_Line (File : File_Type) return Unbounded_String;
9655 Reads a line from the specified file
9656 and returns the result as an unbounded string.
9658 @item procedure Put (File : File_Type; U : Unbounded_String);
9659 Writes the value of the given unbounded string to the specified file
9660 Similar to the effect of
9661 @code{Put (To_String (U))} except that an extra copy is avoided.
9663 @item procedure Put_Line (File : File_Type; U : Unbounded_String);
9664 Writes the value of the given unbounded string to the specified file,
9665 followed by a @code{New_Line}.
9666 Similar to the effect of @code{Put_Line (To_String (U))} except
9667 that an extra copy is avoided.
9671 In the above procedures, @code{File} is of type @code{Ada.Text_IO.File_Type}
9672 and is optional. If the parameter is omitted, then the standard input or
9673 output file is referenced as appropriate.
9675 The package @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} in library
9676 files @file{a-swuwti.ads} and @file{a-swuwti.adb} provides similar extended @code{Wide_Text_IO}
9677 functionality for unbounded wide strings.
9680 @section Wide_Text_IO
9683 @code{Wide_Text_IO} is similar in most respects to Text_IO, except that
9684 both input and output files may contain special sequences that represent
9685 wide character values. The encoding scheme for a given file may be
9686 specified using a FORM parameter:
9693 as part of the FORM string (WCEM = wide character encoding method),
9694 where @var{x} is one of the following characters
9711 The encoding methods match those that
9712 can be used in a source
9713 program, but there is no requirement that the encoding method used for
9714 the source program be the same as the encoding method used for files,
9715 and different files may use different encoding methods.
9717 The default encoding method for the standard files, and for opened files
9718 for which no WCEM parameter is given in the FORM string matches the
9719 wide character encoding specified for the main program (the default
9720 being brackets encoding if no coding method was specified with -gnatW).
9724 In this encoding, a wide character is represented by a five character
9731 where @var{a}, @var{b}, @var{c}, @var{d} are the four hexadecimal
9732 characters (using upper case letters) of the wide character code. For
9733 example, ESC A345 is used to represent the wide character with code
9734 16#A345#. This scheme is compatible with use of the full
9735 @code{Wide_Character} set.
9737 @item Upper Half Coding
9738 The wide character with encoding 16#abcd#, where the upper bit is on
9739 (i.e.@: a is in the range 8-F) is represented as two bytes 16#ab# and
9740 16#cd#. The second byte may never be a format control character, but is
9741 not required to be in the upper half. This method can be also used for
9742 shift-JIS or EUC where the internal coding matches the external coding.
9744 @item Shift JIS Coding
9745 A wide character is represented by a two character sequence 16#ab# and
9746 16#cd#, with the restrictions described for upper half encoding as
9747 described above. The internal character code is the corresponding JIS
9748 character according to the standard algorithm for Shift-JIS
9749 conversion. Only characters defined in the JIS code set table can be
9750 used with this encoding method.
9753 A wide character is represented by a two character sequence 16#ab# and
9754 16#cd#, with both characters being in the upper half. The internal
9755 character code is the corresponding JIS character according to the EUC
9756 encoding algorithm. Only characters defined in the JIS code set table
9757 can be used with this encoding method.
9760 A wide character is represented using
9761 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
9762 10646-1/Am.2. Depending on the character value, the representation
9763 is a one, two, or three byte sequence:
9766 16#0000#-16#007f#: 2#0xxxxxxx#
9767 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
9768 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
9771 where the xxx bits correspond to the left-padded bits of the
9772 16-bit character value. Note that all lower half ASCII characters
9773 are represented as ASCII bytes and all upper half characters and
9774 other wide characters are represented as sequences of upper-half
9775 (The full UTF-8 scheme allows for encoding 31-bit characters as
9776 6-byte sequences, but in this implementation, all UTF-8 sequences
9777 of four or more bytes length will raise a Constraint_Error, as
9778 will all invalid UTF-8 sequences.)
9780 @item Brackets Coding
9781 In this encoding, a wide character is represented by the following eight
9788 Where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
9789 characters (using uppercase letters) of the wide character code. For
9790 example, @code{["A345"]} is used to represent the wide character with code
9792 This scheme is compatible with use of the full Wide_Character set.
9793 On input, brackets coding can also be used for upper half characters,
9794 e.g.@: @code{["C1"]} for lower case a. However, on output, brackets notation
9795 is only used for wide characters with a code greater than @code{16#FF#}.
9799 For the coding schemes other than Hex and Brackets encoding,
9800 not all wide character
9801 values can be represented. An attempt to output a character that cannot
9802 be represented using the encoding scheme for the file causes
9803 Constraint_Error to be raised. An invalid wide character sequence on
9804 input also causes Constraint_Error to be raised.
9807 * Wide_Text_IO Stream Pointer Positioning::
9808 * Wide_Text_IO Reading and Writing Non-Regular Files::
9811 @node Wide_Text_IO Stream Pointer Positioning
9812 @subsection Stream Pointer Positioning
9815 @code{Ada.Wide_Text_IO} is similar to @code{Ada.Text_IO} in its handling
9816 of stream pointer positioning (@pxref{Text_IO}). There is one additional
9819 If @code{Ada.Wide_Text_IO.Look_Ahead} reads a character outside the
9820 normal lower ASCII set (i.e.@: a character in the range:
9823 Wide_Character'Val (16#0080#) .. Wide_Character'Val (16#FFFF#)
9827 then although the logical position of the file pointer is unchanged by
9828 the @code{Look_Ahead} call, the stream is physically positioned past the
9829 wide character sequence. Again this is to avoid the need for buffering
9830 or backup, and all @code{Wide_Text_IO} routines check the internal
9831 indication that this situation has occurred so that this is not visible
9832 to a normal program using @code{Wide_Text_IO}. However, this discrepancy
9833 can be observed if the wide text file shares a stream with another file.
9835 @node Wide_Text_IO Reading and Writing Non-Regular Files
9836 @subsection Reading and Writing Non-Regular Files
9839 As in the case of Text_IO, when a non-regular file is read, it is
9840 assumed that the file contains no page marks (any form characters are
9841 treated as data characters), and @code{End_Of_Page} always returns
9842 @code{False}. Similarly, the end of file indication is not sticky, so
9843 it is possible to read beyond an end of file.
9849 A stream file is a sequence of bytes, where individual elements are
9850 written to the file as described in the Ada 95 reference manual. The type
9851 @code{Stream_Element} is simply a byte. There are two ways to read or
9852 write a stream file.
9856 The operations @code{Read} and @code{Write} directly read or write a
9857 sequence of stream elements with no control information.
9860 The stream attributes applied to a stream file transfer data in the
9861 manner described for stream attributes.
9865 @section Shared Files
9868 Section A.14 of the Ada 95 Reference Manual allows implementations to
9869 provide a wide variety of behavior if an attempt is made to access the
9870 same external file with two or more internal files.
9872 To provide a full range of functionality, while at the same time
9873 minimizing the problems of portability caused by this implementation
9874 dependence, GNAT handles file sharing as follows:
9878 In the absence of a @samp{shared=@var{xxx}} form parameter, an attempt
9879 to open two or more files with the same full name is considered an error
9880 and is not supported. The exception @code{Use_Error} will be
9881 raised. Note that a file that is not explicitly closed by the program
9882 remains open until the program terminates.
9885 If the form parameter @samp{shared=no} appears in the form string, the
9886 file can be opened or created with its own separate stream identifier,
9887 regardless of whether other files sharing the same external file are
9888 opened. The exact effect depends on how the C stream routines handle
9889 multiple accesses to the same external files using separate streams.
9892 If the form parameter @samp{shared=yes} appears in the form string for
9893 each of two or more files opened using the same full name, the same
9894 stream is shared between these files, and the semantics are as described
9895 in Ada 95 Reference Manual, Section A.14.
9898 When a program that opens multiple files with the same name is ported
9899 from another Ada compiler to GNAT, the effect will be that
9900 @code{Use_Error} is raised.
9902 The documentation of the original compiler and the documentation of the
9903 program should then be examined to determine if file sharing was
9904 expected, and @samp{shared=@var{xxx}} parameters added to @code{Open}
9905 and @code{Create} calls as required.
9907 When a program is ported from GNAT to some other Ada compiler, no
9908 special attention is required unless the @samp{shared=@var{xxx}} form
9909 parameter is used in the program. In this case, you must examine the
9910 documentation of the new compiler to see if it supports the required
9911 file sharing semantics, and form strings modified appropriately. Of
9912 course it may be the case that the program cannot be ported if the
9913 target compiler does not support the required functionality. The best
9914 approach in writing portable code is to avoid file sharing (and hence
9915 the use of the @samp{shared=@var{xxx}} parameter in the form string)
9918 One common use of file sharing in Ada 83 is the use of instantiations of
9919 Sequential_IO on the same file with different types, to achieve
9920 heterogeneous input-output. Although this approach will work in GNAT if
9921 @samp{shared=yes} is specified, it is preferable in Ada 95 to use Stream_IO
9922 for this purpose (using the stream attributes)
9928 @code{Open} and @code{Create} calls result in a call to @code{fopen}
9929 using the mode shown in Table 6.1
9932 @center Table 6-1 @code{Open} and @code{Create} Call Modes
9934 @b{OPEN } @b{CREATE}
9935 Append_File "r+" "w+"
9937 Out_File (Direct_IO) "r+" "w"
9938 Out_File (all other cases) "w" "w"
9939 Inout_File "r+" "w+"
9942 If text file translation is required, then either @samp{b} or @samp{t}
9943 is added to the mode, depending on the setting of Text. Text file
9944 translation refers to the mapping of CR/LF sequences in an external file
9945 to LF characters internally. This mapping only occurs in DOS and
9946 DOS-like systems, and is not relevant to other systems.
9948 A special case occurs with Stream_IO@. As shown in the above table, the
9949 file is initially opened in @samp{r} or @samp{w} mode for the
9950 @code{In_File} and @code{Out_File} cases. If a @code{Set_Mode} operation
9951 subsequently requires switching from reading to writing or vice-versa,
9952 then the file is reopened in @samp{r+} mode to permit the required operation.
9954 @node Operations on C Streams
9955 @section Operations on C Streams
9956 The package @code{Interfaces.C_Streams} provides an Ada program with direct
9957 access to the C library functions for operations on C streams:
9960 package Interfaces.C_Streams is
9961 -- Note: the reason we do not use the types that are in
9962 -- Interfaces.C is that we want to avoid dragging in the
9963 -- code in this unit if possible.
9964 subtype chars is System.Address;
9965 -- Pointer to null-terminated array of characters
9966 subtype FILEs is System.Address;
9967 -- Corresponds to the C type FILE*
9968 subtype voids is System.Address;
9969 -- Corresponds to the C type void*
9970 subtype int is Integer;
9971 subtype long is Long_Integer;
9972 -- Note: the above types are subtypes deliberately, and it
9973 -- is part of this spec that the above correspondences are
9974 -- guaranteed. This means that it is legitimate to, for
9975 -- example, use Integer instead of int. We provide these
9976 -- synonyms for clarity, but in some cases it may be
9977 -- convenient to use the underlying types (for example to
9978 -- avoid an unnecessary dependency of a spec on the spec
9980 type size_t is mod 2 ** Standard'Address_Size;
9981 NULL_Stream : constant FILEs;
9982 -- Value returned (NULL in C) to indicate an
9983 -- fdopen/fopen/tmpfile error
9984 ----------------------------------
9985 -- Constants Defined in stdio.h --
9986 ----------------------------------
9988 -- Used by a number of routines to indicate error or
9990 IOFBF : constant int;
9991 IOLBF : constant int;
9992 IONBF : constant int;
9993 -- Used to indicate buffering mode for setvbuf call
9994 SEEK_CUR : constant int;
9995 SEEK_END : constant int;
9996 SEEK_SET : constant int;
9997 -- Used to indicate origin for fseek call
9998 function stdin return FILEs;
9999 function stdout return FILEs;
10000 function stderr return FILEs;
10001 -- Streams associated with standard files
10002 --------------------------
10003 -- Standard C functions --
10004 --------------------------
10005 -- The functions selected below are ones that are
10006 -- available in DOS, OS/2, UNIX and Xenix (but not
10007 -- necessarily in ANSI C). These are very thin interfaces
10008 -- which copy exactly the C headers. For more
10009 -- documentation on these functions, see the Microsoft C
10010 -- "Run-Time Library Reference" (Microsoft Press, 1990,
10011 -- ISBN 1-55615-225-6), which includes useful information
10012 -- on system compatibility.
10013 procedure clearerr (stream : FILEs);
10014 function fclose (stream : FILEs) return int;
10015 function fdopen (handle : int; mode : chars) return FILEs;
10016 function feof (stream : FILEs) return int;
10017 function ferror (stream : FILEs) return int;
10018 function fflush (stream : FILEs) return int;
10019 function fgetc (stream : FILEs) return int;
10020 function fgets (strng : chars; n : int; stream : FILEs)
10022 function fileno (stream : FILEs) return int;
10023 function fopen (filename : chars; Mode : chars)
10025 -- Note: to maintain target independence, use
10026 -- text_translation_required, a boolean variable defined in
10027 -- a-sysdep.c to deal with the target dependent text
10028 -- translation requirement. If this variable is set,
10029 -- then b/t should be appended to the standard mode
10030 -- argument to set the text translation mode off or on
10032 function fputc (C : int; stream : FILEs) return int;
10033 function fputs (Strng : chars; Stream : FILEs) return int;
10050 function ftell (stream : FILEs) return long;
10057 function isatty (handle : int) return int;
10058 procedure mktemp (template : chars);
10059 -- The return value (which is just a pointer to template)
10061 procedure rewind (stream : FILEs);
10062 function rmtmp return int;
10070 function tmpfile return FILEs;
10071 function ungetc (c : int; stream : FILEs) return int;
10072 function unlink (filename : chars) return int;
10073 ---------------------
10074 -- Extra functions --
10075 ---------------------
10076 -- These functions supply slightly thicker bindings than
10077 -- those above. They are derived from functions in the
10078 -- C Run-Time Library, but may do a bit more work than
10079 -- just directly calling one of the Library functions.
10080 function is_regular_file (handle : int) return int;
10081 -- Tests if given handle is for a regular file (result 1)
10082 -- or for a non-regular file (pipe or device, result 0).
10083 ---------------------------------
10084 -- Control of Text/Binary Mode --
10085 ---------------------------------
10086 -- If text_translation_required is true, then the following
10087 -- functions may be used to dynamically switch a file from
10088 -- binary to text mode or vice versa. These functions have
10089 -- no effect if text_translation_required is false (i.e. in
10090 -- normal UNIX mode). Use fileno to get a stream handle.
10091 procedure set_binary_mode (handle : int);
10092 procedure set_text_mode (handle : int);
10093 ----------------------------
10094 -- Full Path Name support --
10095 ----------------------------
10096 procedure full_name (nam : chars; buffer : chars);
10097 -- Given a NUL terminated string representing a file
10098 -- name, returns in buffer a NUL terminated string
10099 -- representing the full path name for the file name.
10100 -- On systems where it is relevant the drive is also
10101 -- part of the full path name. It is the responsibility
10102 -- of the caller to pass an actual parameter for buffer
10103 -- that is big enough for any full path name. Use
10104 -- max_path_len given below as the size of buffer.
10105 max_path_len : integer;
10106 -- Maximum length of an allowable full path name on the
10107 -- system, including a terminating NUL character.
10108 end Interfaces.C_Streams;
10111 @node Interfacing to C Streams
10112 @section Interfacing to C Streams
10115 The packages in this section permit interfacing Ada files to C Stream
10119 with Interfaces.C_Streams;
10120 package Ada.Sequential_IO.C_Streams is
10121 function C_Stream (F : File_Type)
10122 return Interfaces.C_Streams.FILEs;
10124 (File : in out File_Type;
10125 Mode : in File_Mode;
10126 C_Stream : in Interfaces.C_Streams.FILEs;
10127 Form : in String := "");
10128 end Ada.Sequential_IO.C_Streams;
10130 with Interfaces.C_Streams;
10131 package Ada.Direct_IO.C_Streams is
10132 function C_Stream (F : File_Type)
10133 return Interfaces.C_Streams.FILEs;
10135 (File : in out File_Type;
10136 Mode : in File_Mode;
10137 C_Stream : in Interfaces.C_Streams.FILEs;
10138 Form : in String := "");
10139 end Ada.Direct_IO.C_Streams;
10141 with Interfaces.C_Streams;
10142 package Ada.Text_IO.C_Streams is
10143 function C_Stream (F : File_Type)
10144 return Interfaces.C_Streams.FILEs;
10146 (File : in out File_Type;
10147 Mode : in File_Mode;
10148 C_Stream : in Interfaces.C_Streams.FILEs;
10149 Form : in String := "");
10150 end Ada.Text_IO.C_Streams;
10152 with Interfaces.C_Streams;
10153 package Ada.Wide_Text_IO.C_Streams is
10154 function C_Stream (F : File_Type)
10155 return Interfaces.C_Streams.FILEs;
10157 (File : in out File_Type;
10158 Mode : in File_Mode;
10159 C_Stream : in Interfaces.C_Streams.FILEs;
10160 Form : in String := "");
10161 end Ada.Wide_Text_IO.C_Streams;
10163 with Interfaces.C_Streams;
10164 package Ada.Stream_IO.C_Streams is
10165 function C_Stream (F : File_Type)
10166 return Interfaces.C_Streams.FILEs;
10168 (File : in out File_Type;
10169 Mode : in File_Mode;
10170 C_Stream : in Interfaces.C_Streams.FILEs;
10171 Form : in String := "");
10172 end Ada.Stream_IO.C_Streams;
10175 In each of these five packages, the @code{C_Stream} function obtains the
10176 @code{FILE} pointer from a currently opened Ada file. It is then
10177 possible to use the @code{Interfaces.C_Streams} package to operate on
10178 this stream, or the stream can be passed to a C program which can
10179 operate on it directly. Of course the program is responsible for
10180 ensuring that only appropriate sequences of operations are executed.
10182 One particular use of relevance to an Ada program is that the
10183 @code{setvbuf} function can be used to control the buffering of the
10184 stream used by an Ada file. In the absence of such a call the standard
10185 default buffering is used.
10187 The @code{Open} procedures in these packages open a file giving an
10188 existing C Stream instead of a file name. Typically this stream is
10189 imported from a C program, allowing an Ada file to operate on an
10192 @node The GNAT Library
10193 @chapter The GNAT Library
10196 The GNAT library contains a number of general and special purpose packages.
10197 It represents functionality that the GNAT developers have found useful, and
10198 which is made available to GNAT users. The packages described here are fully
10199 supported, and upwards compatibility will be maintained in future releases,
10200 so you can use these facilities with the confidence that the same functionality
10201 will be available in future releases.
10203 The chapter here simply gives a brief summary of the facilities available.
10204 The full documentation is found in the spec file for the package. The full
10205 sources of these library packages, including both spec and body, are provided
10206 with all GNAT releases. For example, to find out the full specifications of
10207 the SPITBOL pattern matching capability, including a full tutorial and
10208 extensive examples, look in the @file{g-spipat.ads} file in the library.
10210 For each entry here, the package name (as it would appear in a @code{with}
10211 clause) is given, followed by the name of the corresponding spec file in
10212 parentheses. The packages are children in four hierarchies, @code{Ada},
10213 @code{Interfaces}, @code{System}, and @code{GNAT}, the latter being a
10214 GNAT-specific hierarchy.
10216 Note that an application program should only use packages in one of these
10217 four hierarchies if the package is defined in the Ada Reference Manual,
10218 or is listed in this section of the GNAT Programmers Reference Manual.
10219 All other units should be considered internal implementation units and
10220 should not be directly @code{with}'ed by application code. The use of
10221 a @code{with} statement that references one of these internal implementation
10222 units makes an application potentially dependent on changes in versions
10223 of GNAT, and will generate a warning message.
10226 * Ada.Characters.Latin_9 (a-chlat9.ads)::
10227 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
10228 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
10229 * Ada.Command_Line.Remove (a-colire.ads)::
10230 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
10231 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
10232 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
10233 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
10234 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
10235 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
10236 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
10237 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
10238 * GNAT.AWK (g-awk.ads)::
10239 * GNAT.Bubble_Sort_A (g-busora.ads)::
10240 * GNAT.Bubble_Sort_G (g-busorg.ads)::
10241 * GNAT.Calendar (g-calend.ads)::
10242 * GNAT.Calendar.Time_IO (g-catiio.ads)::
10243 * GNAT.CRC32 (g-crc32.ads)::
10244 * GNAT.Case_Util (g-casuti.ads)::
10245 * GNAT.CGI (g-cgi.ads)::
10246 * GNAT.CGI.Cookie (g-cgicoo.ads)::
10247 * GNAT.CGI.Debug (g-cgideb.ads)::
10248 * GNAT.Command_Line (g-comlin.ads)::
10249 * GNAT.Current_Exception (g-curexc.ads)::
10250 * GNAT.Debug_Pools (g-debpoo.ads)::
10251 * GNAT.Debug_Utilities (g-debuti.ads)::
10252 * GNAT.Directory_Operations (g-dirope.ads)::
10253 * GNAT.Dynamic_Tables (g-dyntab.ads)::
10254 * GNAT.Exception_Traces (g-exctra.ads)::
10255 * GNAT.Expect (g-expect.ads)::
10256 * GNAT.Float_Control (g-flocon.ads)::
10257 * GNAT.Heap_Sort_A (g-hesora.ads)::
10258 * GNAT.Heap_Sort_G (g-hesorg.ads)::
10259 * GNAT.HTable (g-htable.ads)::
10260 * GNAT.IO (g-io.ads)::
10261 * GNAT.IO_Aux (g-io_aux.ads)::
10262 * GNAT.Lock_Files (g-locfil.ads)::
10263 * GNAT.MD5 (g-md5.ads)::
10264 * GNAT.Most_Recent_Exception (g-moreex.ads)::
10265 * GNAT.OS_Lib (g-os_lib.ads)::
10266 * GNAT.Regexp (g-regexp.ads)::
10267 * GNAT.Registry (g-regist.ads)::
10268 * GNAT.Regpat (g-regpat.ads)::
10269 * GNAT.Sockets (g-socket.ads)::
10270 * GNAT.Source_Info (g-souinf.ads)::
10271 * GNAT.Spell_Checker (g-speche.ads)::
10272 * GNAT.Spitbol.Patterns (g-spipat.ads)::
10273 * GNAT.Spitbol (g-spitbo.ads)::
10274 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
10275 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
10276 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
10277 * GNAT.Table (g-table.ads)::
10278 * GNAT.Task_Lock (g-tasloc.ads)::
10279 * GNAT.Threads (g-thread.ads)::
10280 * GNAT.Traceback (g-traceb.ads)::
10281 * GNAT.Traceback.Symbolic (g-trasym.ads)::
10282 * Interfaces.C.Extensions (i-cexten.ads)::
10283 * Interfaces.C.Streams (i-cstrea.ads)::
10284 * Interfaces.CPP (i-cpp.ads)::
10285 * Interfaces.Os2lib (i-os2lib.ads)::
10286 * Interfaces.Os2lib.Errors (i-os2err.ads)::
10287 * Interfaces.Os2lib.Synchronization (i-os2syn.ads)::
10288 * Interfaces.Os2lib.Threads (i-os2thr.ads)::
10289 * Interfaces.Packed_Decimal (i-pacdec.ads)::
10290 * Interfaces.VxWorks (i-vxwork.ads)::
10291 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
10292 * System.Address_Image (s-addima.ads)::
10293 * System.Assertions (s-assert.ads)::
10294 * System.Partition_Interface (s-parint.ads)::
10295 * System.Task_Info (s-tasinf.ads)::
10296 * System.Wch_Cnv (s-wchcnv.ads)::
10297 * System.Wch_Con (s-wchcon.ads)::
10300 @node Ada.Characters.Latin_9 (a-chlat9.ads)
10301 @section @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
10302 @cindex @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
10303 @cindex Latin_9 constants for Character
10306 This child of @code{Ada.Characters}
10307 provides a set of definitions corresponding to those in the
10308 RM-defined package @code{Ada.Characters.Latin_1} but with the
10309 few modifications required for @code{Latin-9}
10310 The provision of such a package
10311 is specifically authorized by the Ada Reference Manual
10314 @node Ada.Characters.Wide_Latin_1 (a-cwila1.ads)
10315 @section @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
10316 @cindex @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
10317 @cindex Latin_1 constants for Wide_Character
10320 This child of @code{Ada.Characters}
10321 provides a set of definitions corresponding to those in the
10322 RM-defined package @code{Ada.Characters.Latin_1} but with the
10323 types of the constants being @code{Wide_Character}
10324 instead of @code{Character}. The provision of such a package
10325 is specifically authorized by the Ada Reference Manual
10328 @node Ada.Characters.Wide_Latin_9 (a-cwila9.ads)
10329 @section @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
10330 @cindex @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
10331 @cindex Latin_9 constants for Wide_Character
10334 This child of @code{Ada.Characters}
10335 provides a set of definitions corresponding to those in the
10336 GNAT defined package @code{Ada.Characters.Latin_9} but with the
10337 types of the constants being @code{Wide_Character}
10338 instead of @code{Character}. The provision of such a package
10339 is specifically authorized by the Ada Reference Manual
10342 @node Ada.Command_Line.Remove (a-colire.ads)
10343 @section @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
10344 @cindex @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
10345 @cindex Removing command line arguments
10346 @cindex Command line, argument removal
10349 This child of @code{Ada.Command_Line}
10350 provides a mechanism for logically removing
10351 arguments from the argument list. Once removed, an argument is not visible
10352 to further calls on the subprograms in @code{Ada.Command_Line} will not
10353 see the removed argument.
10355 @node Ada.Direct_IO.C_Streams (a-diocst.ads)
10356 @section @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
10357 @cindex @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
10358 @cindex C Streams, Interfacing with Direct_IO
10361 This package provides subprograms that allow interfacing between
10362 C streams and @code{Direct_IO}. The stream identifier can be
10363 extracted from a file opened on the Ada side, and an Ada file
10364 can be constructed from a stream opened on the C side.
10366 @node Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)
10367 @section @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
10368 @cindex @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
10369 @cindex Null_Occurrence, testing for
10372 This child subprogram provides a way of testing for the null
10373 exception occurrence (@code{Null_Occurrence}) without raising
10376 @node Ada.Sequential_IO.C_Streams (a-siocst.ads)
10377 @section @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
10378 @cindex @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
10379 @cindex C Streams, Interfacing with Sequential_IO
10382 This package provides subprograms that allow interfacing between
10383 C streams and @code{Sequential_IO}. The stream identifier can be
10384 extracted from a file opened on the Ada side, and an Ada file
10385 can be constructed from a stream opened on the C side.
10387 @node Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)
10388 @section @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
10389 @cindex @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
10390 @cindex C Streams, Interfacing with Stream_IO
10393 This package provides subprograms that allow interfacing between
10394 C streams and @code{Stream_IO}. The stream identifier can be
10395 extracted from a file opened on the Ada side, and an Ada file
10396 can be constructed from a stream opened on the C side.
10398 @node Ada.Strings.Unbounded.Text_IO (a-suteio.ads)
10399 @section @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
10400 @cindex @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
10401 @cindex @code{Unbounded_String}, IO support
10402 @cindex @code{Text_IO}, extensions for unbounded strings
10405 This package provides subprograms for Text_IO for unbounded
10406 strings, avoiding the necessity for an intermediate operation
10407 with ordinary strings.
10409 @node Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)
10410 @section @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
10411 @cindex @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
10412 @cindex @code{Unbounded_Wide_String}, IO support
10413 @cindex @code{Text_IO}, extensions for unbounded wide strings
10416 This package provides subprograms for Text_IO for unbounded
10417 wide strings, avoiding the necessity for an intermediate operation
10418 with ordinary wide strings.
10420 @node Ada.Text_IO.C_Streams (a-tiocst.ads)
10421 @section @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
10422 @cindex @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
10423 @cindex C Streams, Interfacing with @code{Text_IO}
10426 This package provides subprograms that allow interfacing between
10427 C streams and @code{Text_IO}. The stream identifier can be
10428 extracted from a file opened on the Ada side, and an Ada file
10429 can be constructed from a stream opened on the C side.
10431 @node Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)
10432 @section @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
10433 @cindex @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
10434 @cindex C Streams, Interfacing with @code{Wide_Text_IO}
10437 This package provides subprograms that allow interfacing between
10438 C streams and @code{Wide_Text_IO}. The stream identifier can be
10439 extracted from a file opened on the Ada side, and an Ada file
10440 can be constructed from a stream opened on the C side.
10442 @node GNAT.AWK (g-awk.ads)
10443 @section @code{GNAT.AWK} (@file{g-awk.ads})
10444 @cindex @code{GNAT.AWK} (@file{g-awk.ads})
10449 Provides AWK-like parsing functions, with an easy interface for parsing one
10450 or more files containing formatted data. The file is viewed as a database
10451 where each record is a line and a field is a data element in this line.
10453 @node GNAT.Bubble_Sort_A (g-busora.ads)
10454 @section @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
10455 @cindex @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
10457 @cindex Bubble sort
10460 Provides a general implementation of bubble sort usable for sorting arbitrary
10461 data items. Move and comparison procedures are provided by passing
10462 access-to-procedure values.
10464 @node GNAT.Bubble_Sort_G (g-busorg.ads)
10465 @section @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
10466 @cindex @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
10468 @cindex Bubble sort
10471 Similar to @code{Bubble_Sort_A} except that the move and sorting procedures
10472 are provided as generic parameters, this improves efficiency, especially
10473 if the procedures can be inlined, at the expense of duplicating code for
10474 multiple instantiations.
10476 @node GNAT.Calendar (g-calend.ads)
10477 @section @code{GNAT.Calendar} (@file{g-calend.ads})
10478 @cindex @code{GNAT.Calendar} (@file{g-calend.ads})
10479 @cindex @code{Calendar}
10482 Extends the facilities provided by @code{Ada.Calendar} to include handling
10483 of days of the week, an extended @code{Split} and @code{Time_Of} capability.
10484 Also provides conversion of @code{Ada.Calendar.Time} values to and from the
10485 C @code{timeval} format.
10487 @node GNAT.Calendar.Time_IO (g-catiio.ads)
10488 @section @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
10489 @cindex @code{Calendar}
10491 @cindex @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
10493 @node GNAT.CRC32 (g-crc32.ads)
10494 @section @code{GNAT.CRC32} (@file{g-crc32.ads})
10495 @cindex @code{GNAT.CRC32} (@file{g-crc32.ads})
10497 @cindex Cyclic Redundancy Check
10500 This package implements the CRC-32 algorithm. For a full description
10501 of this algorithm you should have a look at:
10502 ``Computation of Cyclic Redundancy Checks via Table Look-Up'', @cite{Communications
10503 of the ACM}, Vol.@: 31 No.@: 8, pp.@: 1008-1013, Aug.@: 1988. Sarwate, D.V@.
10506 Provides an extended capability for formatted output of time values with
10507 full user control over the format. Modeled on the GNU Date specification.
10509 @node GNAT.Case_Util (g-casuti.ads)
10510 @section @code{GNAT.Case_Util} (@file{g-casuti.ads})
10511 @cindex @code{GNAT.Case_Util} (@file{g-casuti.ads})
10512 @cindex Casing utilities
10513 @cindex Character handling (@code{GNAT.Case_Util})
10516 A set of simple routines for handling upper and lower casing of strings
10517 without the overhead of the full casing tables
10518 in @code{Ada.Characters.Handling}.
10520 @node GNAT.CGI (g-cgi.ads)
10521 @section @code{GNAT.CGI} (@file{g-cgi.ads})
10522 @cindex @code{GNAT.CGI} (@file{g-cgi.ads})
10523 @cindex CGI (Common Gateway Interface)
10526 This is a package for interfacing a GNAT program with a Web server via the
10527 Common Gateway Interface (CGI)@. Basically this package parses the CGI
10528 parameters, which are a set of key/value pairs sent by the Web server. It
10529 builds a table whose index is the key and provides some services to deal
10532 @node GNAT.CGI.Cookie (g-cgicoo.ads)
10533 @section @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
10534 @cindex @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
10535 @cindex CGI (Common Gateway Interface) cookie support
10536 @cindex Cookie support in CGI
10539 This is a package to interface a GNAT program with a Web server via the
10540 Common Gateway Interface (CGI). It exports services to deal with Web
10541 cookies (piece of information kept in the Web client software).
10543 @node GNAT.CGI.Debug (g-cgideb.ads)
10544 @section @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
10545 @cindex @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
10546 @cindex CGI (Common Gateway Interface) debugging
10549 This is a package to help debugging CGI (Common Gateway Interface)
10550 programs written in Ada.
10552 @node GNAT.Command_Line (g-comlin.ads)
10553 @section @code{GNAT.Command_Line} (@file{g-comlin.ads})
10554 @cindex @code{GNAT.Command_Line} (@file{g-comlin.ads})
10555 @cindex Command line
10558 Provides a high level interface to @code{Ada.Command_Line} facilities,
10559 including the ability to scan for named switches with optional parameters
10560 and expand file names using wild card notations.
10562 @node GNAT.Current_Exception (g-curexc.ads)
10563 @section @code{GNAT.Current_Exception} (@file{g-curexc.ads})
10564 @cindex @code{GNAT.Current_Exception} (@file{g-curexc.ads})
10565 @cindex Current exception
10566 @cindex Exception retrieval
10569 Provides access to information on the current exception that has been raised
10570 without the need for using the Ada-95 exception choice parameter specification
10571 syntax. This is particularly useful in simulating typical facilities for
10572 obtaining information about exceptions provided by Ada 83 compilers.
10574 @node GNAT.Debug_Pools (g-debpoo.ads)
10575 @section @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
10576 @cindex @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
10578 @cindex Debug pools
10579 @cindex Memory corruption debugging
10582 Provide a debugging storage pools that helps tracking memory corruption
10583 problems. See section ``Finding memory problems with GNAT Debug Pool'' in
10584 the @cite{GNAT User's Guide}.
10586 @node GNAT.Debug_Utilities (g-debuti.ads)
10587 @section @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
10588 @cindex @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
10592 Provides a few useful utilities for debugging purposes, including conversion
10593 to and from string images of address values.
10595 @node GNAT.Directory_Operations (g-dirope.ads)
10596 @section @code{GNAT.Directory_Operations} (g-dirope.ads)
10597 @cindex @code{GNAT.Directory_Operations} (g-dirope.ads)
10598 @cindex Directory operations
10601 Provides a set of routines for manipulating directories, including changing
10602 the current directory, making new directories, and scanning the files in a
10605 @node GNAT.Dynamic_Tables (g-dyntab.ads)
10606 @section @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
10607 @cindex @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
10608 @cindex Table implementation
10609 @cindex Arrays, extendable
10612 A generic package providing a single dimension array abstraction where the
10613 length of the array can be dynamically modified.
10616 This package provides a facility similar to that of GNAT.Table, except
10617 that this package declares a type that can be used to define dynamic
10618 instances of the table, while an instantiation of GNAT.Table creates a
10619 single instance of the table type.
10621 @node GNAT.Exception_Traces (g-exctra.ads)
10622 @section @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
10623 @cindex @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
10624 @cindex Exception traces
10628 Provides an interface allowing to control automatic output upon exception
10631 @node GNAT.Expect (g-expect.ads)
10632 @section @code{GNAT.Expect} (@file{g-expect.ads})
10633 @cindex @code{GNAT.Expect} (@file{g-expect.ads})
10636 Provides a set of subprograms similar to what is available
10637 with the standard Tcl Expect tool.
10638 It allows you to easily spawn and communicate with an external process.
10639 You can send commands or inputs to the process, and compare the output
10640 with some expected regular expression.
10641 Currently GNAT.Expect is implemented on all native GNAT ports except for
10642 OpenVMS@. It is not implemented for cross ports, and in particular is not
10643 implemented for VxWorks or LynxOS@.
10645 @node GNAT.Float_Control (g-flocon.ads)
10646 @section @code{GNAT.Float_Control} (@file{g-flocon.ads})
10647 @cindex @code{GNAT.Float_Control} (@file{g-flocon.ads})
10648 @cindex Floating-Point Processor
10651 Provides an interface for resetting the floating-point processor into the
10652 mode required for correct semantic operation in Ada. Some third party
10653 library calls may cause this mode to be modified, and the Reset procedure
10654 in this package can be used to reestablish the required mode.
10656 @node GNAT.Heap_Sort_A (g-hesora.ads)
10657 @section @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
10658 @cindex @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
10662 Provides a general implementation of heap sort usable for sorting arbitrary
10663 data items. Move and comparison procedures are provided by passing
10664 access-to-procedure values. The algorithm used is a modified heap sort
10665 that performs approximately N*log(N) comparisons in the worst case.
10667 @node GNAT.Heap_Sort_G (g-hesorg.ads)
10668 @section @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
10669 @cindex @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
10673 Similar to @code{Heap_Sort_A} except that the move and sorting procedures
10674 are provided as generic parameters, this improves efficiency, especially
10675 if the procedures can be inlined, at the expense of duplicating code for
10676 multiple instantiations.
10678 @node GNAT.HTable (g-htable.ads)
10679 @section @code{GNAT.HTable} (@file{g-htable.ads})
10680 @cindex @code{GNAT.HTable} (@file{g-htable.ads})
10681 @cindex Hash tables
10684 A generic implementation of hash tables that can be used to hash arbitrary
10685 data. Provides two approaches, one a simple static approach, and the other
10686 allowing arbitrary dynamic hash tables.
10688 @node GNAT.IO (g-io.ads)
10689 @section @code{GNAT.IO} (@file{g-io.ads})
10690 @cindex @code{GNAT.IO} (@file{g-io.ads})
10692 @cindex Input/Output facilities
10695 A simple preealborable input-output package that provides a subset of
10696 simple Text_IO functions for reading characters and strings from
10697 Standard_Input, and writing characters, strings and integers to either
10698 Standard_Output or Standard_Error.
10700 @node GNAT.IO_Aux (g-io_aux.ads)
10701 @section @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
10702 @cindex @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
10704 @cindex Input/Output facilities
10706 Provides some auxiliary functions for use with Text_IO, including a test
10707 for whether a file exists, and functions for reading a line of text.
10709 @node GNAT.Lock_Files (g-locfil.ads)
10710 @section @code{GNAT.Lock_Files} (@file{g-locfil.ads})
10711 @cindex @code{GNAT.Lock_Files} (@file{g-locfil.ads})
10712 @cindex File locking
10713 @cindex Locking using files
10716 Provides a general interface for using files as locks. Can be used for
10717 providing program level synchronization.
10719 @node GNAT.MD5 (g-md5.ads)
10720 @section @code{GNAT.MD5} (@file{g-md5.ads})
10721 @cindex @code{GNAT.MD5} (@file{g-md5.ads})
10722 @cindex Message Digest MD5
10725 Implements the MD5 Message-Digest Algorithm as described in RFC 1321.
10727 @node GNAT.Most_Recent_Exception (g-moreex.ads)
10728 @section @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
10729 @cindex @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
10730 @cindex Exception, obtaining most recent
10733 Provides access to the most recently raised exception. Can be used for
10734 various logging purposes, including duplicating functionality of some
10735 Ada 83 implementation dependent extensions.
10737 @node GNAT.OS_Lib (g-os_lib.ads)
10738 @section @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
10739 @cindex @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
10740 @cindex Operating System interface
10741 @cindex Spawn capability
10744 Provides a range of target independent operating system interface functions,
10745 including time/date management, file operations, subprocess management,
10746 including a portable spawn procedure, and access to environment variables
10747 and error return codes.
10749 @node GNAT.Regexp (g-regexp.ads)
10750 @section @code{GNAT.Regexp} (@file{g-regexp.ads})
10751 @cindex @code{GNAT.Regexp} (@file{g-regexp.ads})
10752 @cindex Regular expressions
10753 @cindex Pattern matching
10756 A simple implementation of regular expressions, using a subset of regular
10757 expression syntax copied from familiar Unix style utilities. This is the
10758 simples of the three pattern matching packages provided, and is particularly
10759 suitable for ``file globbing'' applications.
10761 @node GNAT.Registry (g-regist.ads)
10762 @section @code{GNAT.Registry} (@file{g-regist.ads})
10763 @cindex @code{GNAT.Registry} (@file{g-regist.ads})
10764 @cindex Windows Registry
10767 This is a high level binding to the Windows registry. It is possible to
10768 do simple things like reading a key value, creating a new key. For full
10769 registry API, but at a lower level of abstraction, refer to the Win32.Winreg
10770 package provided with the Win32Ada binding
10772 @node GNAT.Regpat (g-regpat.ads)
10773 @section @code{GNAT.Regpat} (@file{g-regpat.ads})
10774 @cindex @code{GNAT.Regpat} (@file{g-regpat.ads})
10775 @cindex Regular expressions
10776 @cindex Pattern matching
10779 A complete implementation of Unix-style regular expression matching, copied
10780 from the original V7 style regular expression library written in C by
10781 Henry Spencer (and binary compatible with this C library).
10783 @node GNAT.Sockets (g-socket.ads)
10784 @section @code{GNAT.Sockets} (@file{g-socket.ads})
10785 @cindex @code{GNAT.Sockets} (@file{g-socket.ads})
10789 A high level and portable interface to develop sockets based applications.
10790 This package is based on the sockets thin binding found in GNAT.Sockets.Thin.
10791 Currently GNAT.Sockets is implemented on all native GNAT ports except for
10792 OpenVMS@. It is not implemented for the LynxOS@ cross port.
10794 @node GNAT.Source_Info (g-souinf.ads)
10795 @section @code{GNAT.Source_Info} (@file{g-souinf.ads})
10796 @cindex @code{GNAT.Source_Info} (@file{g-souinf.ads})
10797 @cindex Source Information
10800 Provides subprograms that give access to source code information known at
10801 compile time, such as the current file name and line number.
10803 @node GNAT.Spell_Checker (g-speche.ads)
10804 @section @code{GNAT.Spell_Checker} (@file{g-speche.ads})
10805 @cindex @code{GNAT.Spell_Checker} (@file{g-speche.ads})
10806 @cindex Spell checking
10809 Provides a function for determining whether one string is a plausible
10810 near misspelling of another string.
10812 @node GNAT.Spitbol.Patterns (g-spipat.ads)
10813 @section @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
10814 @cindex @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
10815 @cindex SPITBOL pattern matching
10816 @cindex Pattern matching
10819 A complete implementation of SNOBOL4 style pattern matching. This is the
10820 most elaborate of the pattern matching packages provided. It fully duplicates
10821 the SNOBOL4 dynamic pattern construction and matching capabilities, using the
10822 efficient algorithm developed by Robert Dewar for the SPITBOL system.
10824 @node GNAT.Spitbol (g-spitbo.ads)
10825 @section @code{GNAT.Spitbol} (@file{g-spitbo.ads})
10826 @cindex @code{GNAT.Spitbol} (@file{g-spitbo.ads})
10827 @cindex SPITBOL interface
10830 The top level package of the collection of SPITBOL-style functionality, this
10831 package provides basic SNOBOL4 string manipulation functions, such as
10832 Pad, Reverse, Trim, Substr capability, as well as a generic table function
10833 useful for constructing arbitrary mappings from strings in the style of
10834 the SNOBOL4 TABLE function.
10836 @node GNAT.Spitbol.Table_Boolean (g-sptabo.ads)
10837 @section @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
10838 @cindex @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
10839 @cindex Sets of strings
10840 @cindex SPITBOL Tables
10843 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
10844 for type @code{Standard.Boolean}, giving an implementation of sets of
10847 @node GNAT.Spitbol.Table_Integer (g-sptain.ads)
10848 @section @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
10849 @cindex @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
10850 @cindex Integer maps
10852 @cindex SPITBOL Tables
10855 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
10856 for type @code{Standard.Integer}, giving an implementation of maps
10857 from string to integer values.
10859 @node GNAT.Spitbol.Table_VString (g-sptavs.ads)
10860 @section @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
10861 @cindex @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
10862 @cindex String maps
10864 @cindex SPITBOL Tables
10867 A library level of instantiation of GNAT.Spitbol.Patterns.Table for
10868 a variable length string type, giving an implementation of general
10869 maps from strings to strings.
10871 @node GNAT.Table (g-table.ads)
10872 @section @code{GNAT.Table} (@file{g-table.ads})
10873 @cindex @code{GNAT.Table} (@file{g-table.ads})
10874 @cindex Table implementation
10875 @cindex Arrays, extendable
10878 A generic package providing a single dimension array abstraction where the
10879 length of the array can be dynamically modified.
10882 This package provides a facility similar to that of GNAT.Dynamic_Tables,
10883 except that this package declares a single instance of the table type,
10884 while an instantiation of GNAT.Dynamic_Tables creates a type that can be
10885 used to define dynamic instances of the table.
10887 @node GNAT.Task_Lock (g-tasloc.ads)
10888 @section @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
10889 @cindex @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
10890 @cindex Task synchronization
10891 @cindex Task locking
10895 A very simple facility for locking and unlocking sections of code using a
10896 single global task lock. Appropriate for use in situations where contention
10897 between tasks is very rarely expected.
10899 @node GNAT.Threads (g-thread.ads)
10900 @section @code{GNAT.Threads} (@file{g-thread.ads})
10901 @cindex @code{GNAT.Threads} (@file{g-thread.ads})
10902 @cindex Foreign threads
10903 @cindex Threads, foreign
10906 Provides facilities for creating and destroying threads with explicit calls.
10907 These threads are known to the GNAT run-time system. These subprograms are
10908 exported C-convention procedures intended to be called from foreign code.
10909 By using these primitives rather than directly calling operating systems
10910 routines, compatibility with the Ada tasking runt-time is provided.
10912 @node GNAT.Traceback (g-traceb.ads)
10913 @section @code{GNAT.Traceback} (@file{g-traceb.ads})
10914 @cindex @code{GNAT.Traceback} (@file{g-traceb.ads})
10915 @cindex Trace back facilities
10918 Provides a facility for obtaining non-symbolic traceback information, useful
10919 in various debugging situations.
10921 @node GNAT.Traceback.Symbolic (g-trasym.ads)
10922 @section @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
10923 @cindex @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
10924 @cindex Trace back facilities
10927 Provides symbolic traceback information that includes the subprogram
10928 name and line number information.
10930 @node Interfaces.C.Extensions (i-cexten.ads)
10931 @section @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
10932 @cindex @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
10935 This package contains additional C-related definitions, intended
10936 for use with either manually or automatically generated bindings
10939 @node Interfaces.C.Streams (i-cstrea.ads)
10940 @section @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
10941 @cindex @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
10942 @cindex C streams, interfacing
10945 This package is a binding for the most commonly used operations
10948 @node Interfaces.CPP (i-cpp.ads)
10949 @section @code{Interfaces.CPP} (@file{i-cpp.ads})
10950 @cindex @code{Interfaces.CPP} (@file{i-cpp.ads})
10951 @cindex C++ interfacing
10952 @cindex Interfacing, to C++
10955 This package provides facilities for use in interfacing to C++. It
10956 is primarily intended to be used in connection with automated tools
10957 for the generation of C++ interfaces.
10959 @node Interfaces.Os2lib (i-os2lib.ads)
10960 @section @code{Interfaces.Os2lib} (@file{i-os2lib.ads})
10961 @cindex @code{Interfaces.Os2lib} (@file{i-os2lib.ads})
10962 @cindex Interfacing, to OS/2
10963 @cindex OS/2 interfacing
10966 This package provides interface definitions to the OS/2 library.
10967 It is a thin binding which is a direct translation of the
10968 various @file{<bse@.h>} files.
10970 @node Interfaces.Os2lib.Errors (i-os2err.ads)
10971 @section @code{Interfaces.Os2lib.Errors} (@file{i-os2err.ads})
10972 @cindex @code{Interfaces.Os2lib.Errors} (@file{i-os2err.ads})
10973 @cindex OS/2 Error codes
10974 @cindex Interfacing, to OS/2
10975 @cindex OS/2 interfacing
10978 This package provides definitions of the OS/2 error codes.
10980 @node Interfaces.Os2lib.Synchronization (i-os2syn.ads)
10981 @section @code{Interfaces.Os2lib.Synchronization} (@file{i-os2syn.ads})
10982 @cindex @code{Interfaces.Os2lib.Synchronization} (@file{i-os2syn.ads})
10983 @cindex Interfacing, to OS/2
10984 @cindex Synchronization, OS/2
10985 @cindex OS/2 synchronization primitives
10988 This is a child package that provides definitions for interfacing
10989 to the @code{OS/2} synchronization primitives.
10991 @node Interfaces.Os2lib.Threads (i-os2thr.ads)
10992 @section @code{Interfaces.Os2lib.Threads} (@file{i-os2thr.ads})
10993 @cindex @code{Interfaces.Os2lib.Threads} (@file{i-os2thr.ads})
10994 @cindex Interfacing, to OS/2
10995 @cindex Thread control, OS/2
10996 @cindex OS/2 thread interfacing
10999 This is a child package that provides definitions for interfacing
11000 to the @code{OS/2} thread primitives.
11002 @node Interfaces.Packed_Decimal (i-pacdec.ads)
11003 @section @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
11004 @cindex @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
11005 @cindex IBM Packed Format
11006 @cindex Packed Decimal
11009 This package provides a set of routines for conversions to and
11010 from a packed decimal format compatible with that used on IBM
11013 @node Interfaces.VxWorks (i-vxwork.ads)
11014 @section @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
11015 @cindex @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
11016 @cindex Interfacing to VxWorks
11017 @cindex VxWorks, interfacing
11020 This package provides a limited binding to the VxWorks API.
11021 In particular, it interfaces with the
11022 VxWorks hardware interrupt facilities.
11024 @node Interfaces.VxWorks.IO (i-vxwoio.ads)
11025 @section @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
11026 @cindex @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
11027 @cindex Interfacing to VxWorks' I/O
11028 @cindex VxWorks, I/O interfacing
11029 @cindex VxWorks, Get_Immediate
11032 This package provides a limited binding to the VxWorks' I/O API.
11033 In particular, it provides procedures that enable the use of
11034 Get_Immediate under VxWorks.
11036 @node System.Address_Image (s-addima.ads)
11037 @section @code{System.Address_Image} (@file{s-addima.ads})
11038 @cindex @code{System.Address_Image} (@file{s-addima.ads})
11039 @cindex Address image
11040 @cindex Image, of an address
11043 This function provides a useful debugging
11044 function that gives an (implementation dependent)
11045 string which identifies an address.
11047 @node System.Assertions (s-assert.ads)
11048 @section @code{System.Assertions} (@file{s-assert.ads})
11049 @cindex @code{System.Assertions} (@file{s-assert.ads})
11051 @cindex Assert_Failure, exception
11054 This package provides the declaration of the exception raised
11055 by an run-time assertion failure, as well as the routine that
11056 is used internally to raise this assertion.
11058 @node System.Partition_Interface (s-parint.ads)
11059 @section @code{System.Partition_Interface} (@file{s-parint.ads})
11060 @cindex @code{System.Partition_Interface} (@file{s-parint.ads})
11061 @cindex Partition intefacing functions
11064 This package provides facilities for partition interfacing. It
11065 is used primarily in a distribution context when using Annex E
11068 @node System.Task_Info (s-tasinf.ads)
11069 @section @code{System.Task_Info} (@file{s-tasinf.ads})
11070 @cindex @code{System.Task_Info} (@file{s-tasinf.ads})
11071 @cindex Task_Info pragma
11074 This package provides target dependent functionality that is used
11075 to support the @code{Task_Info} pragma
11077 @node System.Wch_Cnv (s-wchcnv.ads)
11078 @section @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
11079 @cindex @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
11080 @cindex Wide Character, Representation
11081 @cindex Wide String, Conversion
11082 @cindex Representation of wide characters
11085 This package provides routines for converting between
11086 wide characters and a representation as a value of type
11087 @code{Standard.String}, using a specified wide character
11088 encoding method. It uses definitions in
11089 package @code{System.Wch_Con}.
11091 @node System.Wch_Con (s-wchcon.ads)
11092 @section @code{System.Wch_Con} (@file{s-wchcon.ads})
11093 @cindex @code{System.Wch_Con} (@file{s-wchcon.ads})
11096 This package provides definitions and descriptions of
11097 the various methods used for encoding wide characters
11098 in ordinary strings. These definitions are used by
11099 the package @code{System.Wch_Cnv}.
11101 @node Interfacing to Other Languages
11102 @chapter Interfacing to Other Languages
11104 The facilities in annex B of the Ada 95 Reference Manual are fully
11105 implemented in GNAT, and in addition, a full interface to C++ is
11109 * Interfacing to C::
11110 * Interfacing to C++::
11111 * Interfacing to COBOL::
11112 * Interfacing to Fortran::
11113 * Interfacing to non-GNAT Ada code::
11116 @node Interfacing to C
11117 @section Interfacing to C
11120 Interfacing to C with GNAT can use one of two approaches:
11124 The types in the package @code{Interfaces.C} may be used.
11126 Standard Ada types may be used directly. This may be less portable to
11127 other compilers, but will work on all GNAT compilers, which guarantee
11128 correspondence between the C and Ada types.
11132 Pragma @code{Convention C} maybe applied to Ada types, but mostly has no
11133 effect, since this is the default. The following table shows the
11134 correspondence between Ada scalar types and the corresponding C types.
11139 @item Short_Integer
11141 @item Short_Short_Integer
11145 @item Long_Long_Integer
11153 @item Long_Long_Float
11154 This is the longest floating-point type supported by the hardware.
11159 Ada enumeration types map to C enumeration types directly if pragma
11160 @code{Convention C} is specified, which causes them to have int
11161 length. Without pragma @code{Convention C}, Ada enumeration types map to
11162 8, 16, or 32 bits (i.e.@: C types @code{signed char}, @code{short}, @code{int}, respectively)
11163 depending on the number of values passed. This is the only case in which
11164 pragma @code{Convention C} affects the representation of an Ada type.
11167 Ada access types map to C pointers, except for the case of pointers to
11168 unconstrained types in Ada, which have no direct C equivalent.
11171 Ada arrays map directly to C arrays.
11174 Ada records map directly to C structures.
11177 Packed Ada records map to C structures where all members are bit fields
11178 of the length corresponding to the @code{@var{type}'Size} value in Ada.
11181 @node Interfacing to C++
11182 @section Interfacing to C++
11185 The interface to C++ makes use of the following pragmas, which are
11186 primarily intended to be constructed automatically using a binding generator
11187 tool, although it is possible to construct them by hand. Ada Core
11188 Technologies does not currently supply a suitable binding generator tool.
11190 Using these pragmas it is possible to achieve complete
11191 inter-operability between Ada tagged types and C class definitions.
11192 See @ref{Implementation Defined Pragmas} for more details.
11195 @item pragma CPP_Class ([Entity =>] @var{local_name})
11196 The argument denotes an entity in the current declarative region that is
11197 declared as a tagged or untagged record type. It indicates that the type
11198 corresponds to an externally declared C++ class type, and is to be laid
11199 out the same way that C++ would lay out the type.
11201 @item pragma CPP_Constructor ([Entity =>] @var{local_name})
11202 This pragma identifies an imported function (imported in the usual way
11203 with pragma @code{Import}) as corresponding to a C++ constructor.
11205 @item pragma CPP_Vtable @dots{}
11206 One @code{CPP_Vtable} pragma can be present for each component of type
11207 @code{CPP.Interfaces.Vtable_Ptr} in a record to which pragma @code{CPP_Class}
11211 @node Interfacing to COBOL
11212 @section Interfacing to COBOL
11215 Interfacing to COBOL is achieved as described in section B.4 of
11216 the Ada 95 reference manual.
11218 @node Interfacing to Fortran
11219 @section Interfacing to Fortran
11222 Interfacing to Fortran is achieved as described in section B.5 of the
11223 reference manual. The pragma @code{Convention Fortran}, applied to a
11224 multi-dimensional array causes the array to be stored in column-major
11225 order as required for convenient interface to Fortran.
11227 @node Interfacing to non-GNAT Ada code
11228 @section Interfacing to non-GNAT Ada code
11230 It is possible to specify the convention @code{Ada} in a pragma @code{Import} or
11231 pragma @code{Export}. However this refers to the calling conventions used
11232 by GNAT, which may or may not be similar enough to those used by
11233 some other Ada 83 or Ada 95 compiler to allow interoperation.
11235 If arguments types are kept simple, and if the foreign compiler generally
11236 follows system calling conventions, then it may be possible to integrate
11237 files compiled by other Ada compilers, provided that the elaboration
11238 issues are adequately addressed (for example by eliminating the
11239 need for any load time elaboration).
11241 In particular, GNAT running on VMS is designed to
11242 be highly compatible with the DEC Ada 83 compiler, so this is one
11243 case in which it is possible to import foreign units of this type,
11244 provided that the data items passed are restricted to simple scalar
11245 values or simple record types without variants, or simple array
11246 types with fixed bounds.
11248 @node Machine Code Insertions
11249 @chapter Machine Code Insertions
11252 Package @code{Machine_Code} provides machine code support as described
11253 in the Ada 95 Reference Manual in two separate forms:
11256 Machine code statements, consisting of qualified expressions that
11257 fit the requirements of RM section 13.8.
11259 An intrinsic callable procedure, providing an alternative mechanism of
11260 including machine instructions in a subprogram.
11263 The two features are similar, and both closely related to the mechanism
11264 provided by the asm instruction in the GNU C compiler. Full understanding
11265 and use of the facilities in this package requires understanding the asm
11266 instruction as described in
11267 @cite{Using and Porting the GNU Compiler Collection (GCC)} by Richard
11268 Stallman. Calls to the function @code{Asm} and the procedure @code{Asm}
11269 have identical semantic restrictions and effects as described below.
11270 Both are provided so that the procedure call can be used as a statement,
11271 and the function call can be used to form a code_statement.
11273 The first example given in the GCC documentation is the C @code{asm}
11276 asm ("fsinx %1 %0" : "=f" (result) : "f" (angle));
11280 The equivalent can be written for GNAT as:
11283 Asm ("fsinx %1 %0",
11284 My_Float'Asm_Output ("=f", result),
11285 My_Float'Asm_Input ("f", angle));
11288 The first argument to @code{Asm} is the assembler template, and is
11289 identical to what is used in GNU C@. This string must be a static
11290 expression. The second argument is the output operand list. It is
11291 either a single @code{Asm_Output} attribute reference, or a list of such
11292 references enclosed in parentheses (technically an array aggregate of
11295 The @code{Asm_Output} attribute denotes a function that takes two
11296 parameters. The first is a string, the second is the name of a variable
11297 of the type designated by the attribute prefix. The first (string)
11298 argument is required to be a static expression and designates the
11299 constraint for the parameter (e.g.@: what kind of register is
11300 required). The second argument is the variable to be updated with the
11301 result. The possible values for constraint are the same as those used in
11302 the RTL, and are dependent on the configuration file used to build the
11303 GCC back end. If there are no output operands, then this argument may
11304 either be omitted, or explicitly given as @code{No_Output_Operands}.
11306 The second argument of @code{@var{my_float}'Asm_Output} functions as
11307 though it were an @code{out} parameter, which is a little curious, but
11308 all names have the form of expressions, so there is no syntactic
11309 irregularity, even though normally functions would not be permitted
11310 @code{out} parameters. The third argument is the list of input
11311 operands. It is either a single @code{Asm_Input} attribute reference, or
11312 a list of such references enclosed in parentheses (technically an array
11313 aggregate of such references).
11315 The @code{Asm_Input} attribute denotes a function that takes two
11316 parameters. The first is a string, the second is an expression of the
11317 type designated by the prefix. The first (string) argument is required
11318 to be a static expression, and is the constraint for the parameter,
11319 (e.g.@: what kind of register is required). The second argument is the
11320 value to be used as the input argument. The possible values for the
11321 constant are the same as those used in the RTL, and are dependent on
11322 the configuration file used to built the GCC back end.
11324 If there are no input operands, this argument may either be omitted, or
11325 explicitly given as @code{No_Input_Operands}. The fourth argument, not
11326 present in the above example, is a list of register names, called the
11327 @dfn{clobber} argument. This argument, if given, must be a static string
11328 expression, and is a space or comma separated list of names of registers
11329 that must be considered destroyed as a result of the @code{Asm} call. If
11330 this argument is the null string (the default value), then the code
11331 generator assumes that no additional registers are destroyed.
11333 The fifth argument, not present in the above example, called the
11334 @dfn{volatile} argument, is by default @code{False}. It can be set to
11335 the literal value @code{True} to indicate to the code generator that all
11336 optimizations with respect to the instruction specified should be
11337 suppressed, and that in particular, for an instruction that has outputs,
11338 the instruction will still be generated, even if none of the outputs are
11339 used. See the full description in the GCC manual for further details.
11341 The @code{Asm} subprograms may be used in two ways. First the procedure
11342 forms can be used anywhere a procedure call would be valid, and
11343 correspond to what the RM calls ``intrinsic'' routines. Such calls can
11344 be used to intersperse machine instructions with other Ada statements.
11345 Second, the function forms, which return a dummy value of the limited
11346 private type @code{Asm_Insn}, can be used in code statements, and indeed
11347 this is the only context where such calls are allowed. Code statements
11348 appear as aggregates of the form:
11351 Asm_Insn'(Asm (@dots{}));
11352 Asm_Insn'(Asm_Volatile (@dots{}));
11355 In accordance with RM rules, such code statements are allowed only
11356 within subprograms whose entire body consists of such statements. It is
11357 not permissible to intermix such statements with other Ada statements.
11359 Typically the form using intrinsic procedure calls is more convenient
11360 and more flexible. The code statement form is provided to meet the RM
11361 suggestion that such a facility should be made available. The following
11362 is the exact syntax of the call to @code{Asm} (of course if named notation is
11363 used, the arguments may be given in arbitrary order, following the
11364 normal rules for use of positional and named arguments)
11368 [Template =>] static_string_EXPRESSION
11369 [,[Outputs =>] OUTPUT_OPERAND_LIST ]
11370 [,[Inputs =>] INPUT_OPERAND_LIST ]
11371 [,[Clobber =>] static_string_EXPRESSION ]
11372 [,[Volatile =>] static_boolean_EXPRESSION] )
11373 OUTPUT_OPERAND_LIST ::=
11375 | OUTPUT_OPERAND_ATTRIBUTE
11376 | (OUTPUT_OPERAND_ATTRIBUTE @{,OUTPUT_OPERAND_ATTRIBUTE@})
11377 OUTPUT_OPERAND_ATTRIBUTE ::=
11378 SUBTYPE_MARK'Asm_Output (static_string_EXPRESSION, NAME)
11379 INPUT_OPERAND_LIST ::=
11381 | INPUT_OPERAND_ATTRIBUTE
11382 | (INPUT_OPERAND_ATTRIBUTE @{,INPUT_OPERAND_ATTRIBUTE@})
11383 INPUT_OPERAND_ATTRIBUTE ::=
11384 SUBTYPE_MARK'Asm_Input (static_string_EXPRESSION, EXPRESSION)
11387 @node GNAT Implementation of Tasking
11388 @chapter GNAT Implementation of Tasking
11390 * Mapping Ada Tasks onto the Underlying Kernel Threads::
11391 * Ensuring Compliance with the Real-Time Annex::
11394 @node Mapping Ada Tasks onto the Underlying Kernel Threads
11395 @section Mapping Ada Tasks onto the Underlying Kernel Threads
11397 GNAT run-time system comprises two layers:
11400 @item GNARL (GNAT Run-time Layer)
11401 @item GNULL (GNAT Low-level Library)
11404 In GNAT, Ada's tasking services rely on a platform and OS independent
11405 layer known as GNARL@. This code is responsible for implementing the
11406 correct semantics of Ada's task creation, rendezvous, protected
11409 GNARL decomposes Ada's tasking semantics into simpler lower level
11410 operations such as create a thread, set the priority of a thread,
11411 yield, create a lock, lock/unlock, etc. The spec for these low-level
11412 operations constitutes GNULLI, the GNULL Interface. This interface is
11413 directly inspired from the POSIX real-time API@.
11415 If the underlying executive or OS implements the POSIX standard
11416 faithfully, the GNULL Interface maps as is to the services offered by
11417 the underlying kernel. Otherwise, some target dependent glue code maps
11418 the services offered by the underlying kernel to the semantics expected
11421 Whatever the underlying OS (VxWorks, UNIX, OS/2, Windows NT, etc.) the
11422 key point is that each Ada task is mapped on a thread in the underlying
11423 kernel. For example, in the case of VxWorks, one Ada task = one VxWorks task.
11425 In addition Ada task priorities map onto the underlying thread priorities.
11426 Mapping Ada tasks onto the underlying kernel threads has several advantages:
11431 The underlying scheduler is used to schedule the Ada tasks. This
11432 makes Ada tasks as efficient as kernel threads from a scheduling
11436 Interaction with code written in C containing threads is eased
11437 since at the lowest level Ada tasks and C threads map onto the same
11438 underlying kernel concept.
11441 When an Ada task is blocked during I/O the remaining Ada tasks are
11445 On multi-processor systems Ada Tasks can execute in parallel.
11448 @node Ensuring Compliance with the Real-Time Annex
11449 @section Ensuring Compliance with the Real-Time Annex
11451 The reader will be quick to notice that while mapping Ada tasks onto
11452 the underlying threads has significant advantages, it does create some
11453 complications when it comes to respecting the scheduling semantics
11454 specified in the real-time annex (Annex D).
11456 For instance Annex D requires that for the FIFO_Within_Priorities
11457 scheduling policy we have:
11460 When the active priority of a ready task that is not running
11461 changes, or the setting of its base priority takes effect, the
11462 task is removed from the ready queue for its old active priority
11463 and is added at the tail of the ready queue for its new active
11464 priority, except in the case where the active priority is lowered
11465 due to the loss of inherited priority, in which case the task is
11466 added at the head of the ready queue for its new active priority.
11469 While most kernels do put tasks at the end of the priority queue when
11470 a task changes its priority, (which respects the main
11471 FIFO_Within_Priorities requirement), almost none keep a thread at the
11472 beginning of its priority queue when its priority drops from the loss
11473 of inherited priority.
11475 As a result most vendors have provided incomplete Annex D implementations.
11477 The GNAT run-time, has a nice cooperative solution to this problem
11478 which ensures that accurate FIFO_Within_Priorities semantics are
11481 The principle is as follows. When an Ada task T is about to start
11482 running, it checks whether some other Ada task R with the same
11483 priority as T has been suspended due to the loss of priority
11484 inheritance. If this is the case, T yields and is placed at the end of
11485 its priority queue. When R arrives at the front of the queue it
11488 Note that this simple scheme preserves the relative order of the tasks
11489 that were ready to execute in the priority queue where R has been
11492 @node Code generation for array aggregates
11493 @chapter Code generation for array aggregates
11496 * Static constant aggregates with static bounds::
11497 * Constant aggregates with an unconstrained nominal types::
11498 * Aggregates with static bounds::
11499 * Aggregates with non-static bounds::
11500 * Aggregates in assignments statements::
11503 Aggregate have a rich syntax and allow the user to specify the values of
11504 complex data structures by means of a single construct. As a result, the
11505 code generated for aggregates can be quite complex and involve loops, case
11506 statements and multiple assignments. In the simplest cases, however, the
11507 compiler will recognize aggregates whose components and constraints are
11508 fully static, and in those cases the compiler will generate little or no
11509 executable code. The following is an outline of the code that GNAT generates
11510 for various aggregate constructs. For further details, the user will find it
11511 useful to examine the output produced by the -gnatG flag to see the expanded
11512 source that is input to the code generator. The user will also want to examine
11513 the assembly code generated at various levels of optimization.
11515 The code generated for aggregates depends on the context, the component values,
11516 and the type. In the context of an object declaration the code generated is
11517 generally simpler than in the case of an assignment. As a general rule, static
11518 component values and static subtypes also lead to simpler code.
11520 @node Static constant aggregates with static bounds
11521 @section Static constant aggregates with static bounds
11523 For the declarations:
11525 type One_Dim is array (1..10) of integer;
11526 ar0 : constant One_Dim := ( 1, 2, 3, 4, 5, 6, 7, 8, 9, 0);
11529 GNAT generates no executable code: the constant ar0 is placed in static memory.
11530 The same is true for constant aggregates with named associations:
11533 Cr1 : constant One_Dim := (4 => 16, 2 => 4, 3 => 9, 1=> 1);
11534 Cr3 : constant One_Dim := (others => 7777);
11537 The same is true for multidimensional constant arrays such as:
11540 type two_dim is array (1..3, 1..3) of integer;
11541 Unit : constant two_dim := ( (1,0,0), (0,1,0), (0,0,1));
11544 The same is true for arrays of one-dimensional arrays: the following are
11548 type ar1b is array (1..3) of boolean;
11549 type ar_ar is array (1..3) of ar1b;
11550 None : constant ar1b := (others => false); -- fully static
11551 None2 : constant ar_ar := (1..3 => None); -- fully static
11554 However, for multidimensional aggregates with named associations, GNAT will
11555 generate assignments and loops, even if all associations are static. The
11556 following two declarations generate a loop for the first dimension, and
11557 individual component assignments for the second dimension:
11560 Zero1: constant two_dim := (1..3 => (1..3 => 0));
11561 Zero2: constant two_dim := (others => (others => 0));
11564 @node Constant aggregates with an unconstrained nominal types
11565 @section Constant aggregates with an unconstrained nominal types
11567 In such cases the aggregate itself establishes the subtype, so that associations
11568 with @code{others} cannot be used. GNAT determines the bounds for the actual
11569 subtype of the aggregate, and allocates the aggregate statically as well. No
11570 code is generated for the following:
11573 type One_Unc is array (natural range <>) of integer;
11574 Cr_Unc : constant One_Unc := (12,24,36);
11577 @node Aggregates with static bounds
11578 @section Aggregates with static bounds
11580 In all previous examples the aggregate was the initial (and immutable) value
11581 of a constant. If the aggregate initializes a variable, then code is generated
11582 for it as a combination of individual assignments and loops over the target
11583 object. The declarations
11586 Cr_Var1 : One_Dim := (2, 5, 7, 11);
11587 Cr_Var2 : One_Dim := (others > -1);
11590 generate the equivalent of
11598 for I in Cr_Var2'range loop
11599 Cr_Var2 (I) := =-1;
11603 @node Aggregates with non-static bounds
11604 @section Aggregates with non-static bounds
11606 If the bounds of the aggregate are not statically compatible with the bounds
11607 of the nominal subtype of the target, then constraint checks have to be
11608 generated on the bounds. For a multidimensional array, constraint checks may
11609 have to be applied to sub-arrays individually, if they do not have statically
11610 compatible subtypes.
11612 @node Aggregates in assignments statements
11613 @section Aggregates in assignments statements
11615 In general, aggregate assignment requires the construction of a temporary,
11616 and a copy from the temporary to the target of the assignment. This is because
11617 it is not always possible to convert the assignment into a series of individual
11618 component assignments. For example, consider the simple case:
11624 This cannot be converted into:
11631 So the aggregate has to be built first in a separate location, and then
11632 copied into the target. GNAT recognizes simple cases where this intermediate
11633 step is not required, and the assignments can be performed in place, directly
11634 into the target. The following sufficient criteria are applied:
11637 @item The bounds of the aggregate are static, and the associations are static.
11638 @item The components of the aggregate are static constants, names of
11639 simple variables that are not renamings, or expressions not involving
11640 indexed components whose operands obey these rules.
11643 If any of these conditions are violated, the aggregate will be built in
11644 a temporary (created either by the front-end or the code generator) and then
11645 that temporary will be copied onto the target.
11648 @node Specialized Needs Annexes
11649 @chapter Specialized Needs Annexes
11652 Ada 95 defines a number of specialized needs annexes, which are not
11653 required in all implementations. However, as described in this chapter,
11654 GNAT implements all of these special needs annexes:
11657 @item Systems Programming (Annex C)
11658 The Systems Programming Annex is fully implemented.
11660 @item Real-Time Systems (Annex D)
11661 The Real-Time Systems Annex is fully implemented.
11663 @item Distributed Systems (Annex E)
11664 Stub generation is fully implemented in the GNAT compiler. In addition,
11665 a complete compatible PCS is available as part of the GLADE system,
11666 a separate product. When the two
11667 products are used in conjunction, this annex is fully implemented.
11669 @item Information Systems (Annex F)
11670 The Information Systems annex is fully implemented.
11672 @item Numerics (Annex G)
11673 The Numerics Annex is fully implemented.
11675 @item Safety and Security (Annex H)
11676 The Safety and Security annex is fully implemented.
11680 @node Compatibility Guide
11681 @chapter Compatibility Guide
11684 This chapter contains sections that describe compatibility issues between
11685 GNAT and other Ada 83 and Ada 95 compilation systems, to aid in porting
11686 applications developed in other Ada environments.
11689 * Compatibility with Ada 83::
11690 * Compatibility with DEC Ada 83::
11691 * Compatibility with Other Ada 95 Systems::
11692 * Representation Clauses::
11695 @node Compatibility with Ada 83
11696 @section Compatibility with Ada 83
11697 @cindex Compatibility (between Ada 83 and Ada 95)
11700 Ada 95 is designed to be highly upwards compatible with Ada 83. In
11701 particular, the design intention is that the difficulties associated
11702 with moving from Ada 83 to Ada 95 should be no greater than those
11703 that occur when moving from one Ada 83 system to another.
11705 However, there are a number of points at which there are minor
11706 incompatibilities. The Ada 95 Annotated Reference Manual contains
11707 full details of these issues,
11708 and should be consulted for a complete treatment.
11710 following are the most likely issues to be encountered.
11713 @item Character range
11714 The range of @code{Standard.Character} is now the full 256 characters of Latin-1,
11715 whereas in most Ada 83 implementations it was restricted to 128 characters.
11716 This may show up as compile time or runtime errors. The desirable fix is to
11717 adapt the program to accommodate the full character set, but in some cases
11718 it may be convenient to define a subtype or derived type of Character that
11719 covers only the restricted range.
11722 @item New reserved words
11723 The identifiers @code{abstract}, @code{aliased}, @code{protected},
11724 @code{requeue}, @code{tagged}, and @code{until} are reserved in Ada 95.
11725 Existing Ada 83 code using any of these identifiers must be edited to
11726 use some alternative name.
11728 @item Freezing rules
11729 The rules in Ada 95 are slightly different with regard to the point at
11730 which entities are frozen, and representation pragmas and clauses are
11731 not permitted past the freeze point. This shows up most typically in
11732 the form of an error message complaining that a representation item
11733 appears too late, and the appropriate corrective action is to move
11734 the item nearer to the declaration of the entity to which it refers.
11736 A particular case is that representation pragmas (including the
11737 extended DEC Ada 83 compatibility pragmas such as @code{Export_Procedure}), cannot
11738 be applied to a subprogram body. If necessary, a separate subprogram
11739 declaration must be introduced to which the pragma can be applied.
11741 @item Optional bodies for library packages
11742 In Ada 83, a package that did not require a package body was nevertheless
11743 allowed to have one. This lead to certain surprises in compiling large
11744 systems (situations in which the body could be unexpectedly ignored). In
11745 Ada 95, if a package does not require a body then it is not permitted to
11746 have a body. To fix this problem, simply remove a redundant body if it
11747 is empty, or, if it is non-empty, introduce a dummy declaration into the
11748 spec that makes the body required. One approach is to add a private part
11749 to the package declaration (if necessary), and define a parameterless
11750 procedure called Requires_Body, which must then be given a dummy
11751 procedure body in the package body, which then becomes required.
11753 @item @code{Numeric_Error} is now the same as @code{Constraint_Error}
11754 In Ada 95, the exception @code{Numeric_Error} is a renaming of @code{Constraint_Error}.
11755 This means that it is illegal to have separate exception handlers for
11756 the two exceptions. The fix is simply to remove the handler for the
11757 @code{Numeric_Error} case (since even in Ada 83, a compiler was free to raise
11758 @code{Constraint_Error} in place of @code{Numeric_Error} in all cases).
11760 @item Indefinite subtypes in generics
11761 In Ada 83, it was permissible to pass an indefinite type (e.g.@: @code{String}) as
11762 the actual for a generic formal private type, but then the instantiation
11763 would be illegal if there were any instances of declarations of variables
11764 of this type in the generic body. In Ada 95, to avoid this clear violation
11765 of the contract model, the generic declaration clearly indicates whether
11766 or not such instantiations are permitted. If a generic formal parameter
11767 has explicit unknown discriminants, indicated by using @code{(<>)} after the
11768 type name, then it can be instantiated with indefinite types, but no
11769 variables can be declared of this type. Any attempt to declare a variable
11770 will result in an illegality at the time the generic is declared. If the
11771 @code{(<>)} notation is not used, then it is illegal to instantiate the generic
11772 with an indefinite type. This will show up as a compile time error, and
11773 the fix is usually simply to add the @code{(<>)} to the generic declaration.
11776 All implementations of GNAT provide a switch that causes GNAT to operate
11777 in Ada 83 mode. In this mode, some but not all compatibility problems
11778 of the type described above are handled automatically. For example, the
11779 new Ada 95 protected keywords are not recognized in this mode. However,
11780 in practice, it is usually advisable to make the necessary modifications
11781 to the program to remove the need for using this switch.
11783 @node Compatibility with Other Ada 95 Systems
11784 @section Compatibility with Other Ada 95 Systems
11787 Providing that programs avoid the use of implementation dependent and
11788 implementation defined features of Ada 95, as documented in the Ada 95
11789 reference manual, there should be a high degree of portability between
11790 GNAT and other Ada 95 systems. The following are specific items which
11791 have proved troublesome in moving GNAT programs to other Ada 95
11792 compilers, but do not affect porting code to GNAT@.
11795 @item Ada 83 Pragmas and Attributes
11796 Ada 95 compilers are allowed, but not required, to implement the missing
11797 Ada 83 pragmas and attributes that are no longer defined in Ada 95.
11798 GNAT implements all such pragmas and attributes, eliminating this as
11799 a compatibility concern, but some other Ada 95 compilers reject these
11800 pragmas and attributes.
11802 @item Special-needs Annexes
11803 GNAT implements the full set of special needs annexes. At the
11804 current time, it is the only Ada 95 compiler to do so. This means that
11805 programs making use of these features may not be portable to other Ada
11806 95 compilation systems.
11808 @item Representation Clauses
11809 Some other Ada 95 compilers implement only the minimal set of
11810 representation clauses required by the Ada 95 reference manual. GNAT goes
11811 far beyond this minimal set, as described in the next section.
11814 @node Representation Clauses
11815 @section Representation Clauses
11818 The Ada 83 reference manual was quite vague in describing both the minimal
11819 required implementation of representation clauses, and also their precise
11820 effects. The Ada 95 reference manual is much more explicit, but the minimal
11821 set of capabilities required in Ada 95 is quite limited.
11823 GNAT implements the full required set of capabilities described in the
11824 Ada 95 reference manual, but also goes much beyond this, and in particular
11825 an effort has been made to be compatible with existing Ada 83 usage to the
11826 greatest extent possible.
11828 A few cases exist in which Ada 83 compiler behavior is incompatible with
11829 requirements in the Ada 95 reference manual. These are instances of
11830 intentional or accidental dependence on specific implementation dependent
11831 characteristics of these Ada 83 compilers. The following is a list of
11832 the cases most likely to arise in existing legacy Ada 83 code.
11835 @item Implicit Packing
11836 Some Ada 83 compilers allowed a Size specification to cause implicit
11837 packing of an array or record. This could cause expensive implicit
11838 conversions for change of representation in the presence of derived
11839 types, and the Ada design intends to avoid this possibility.
11840 Subsequent AI's were issued to make it clear that such implicit
11841 change of representation in response to a Size clause is inadvisable,
11842 and this recommendation is represented explicitly in the Ada 95 RM
11843 as implementation advice that is followed by GNAT@.
11844 The problem will show up as an error
11845 message rejecting the size clause. The fix is simply to provide
11846 the explicit pragma @code{Pack}, or for more fine tuned control, provide
11847 a Component_Size clause.
11849 @item Meaning of Size Attribute
11850 The Size attribute in Ada 95 for discrete types is defined as being the
11851 minimal number of bits required to hold values of the type. For example,
11852 on a 32-bit machine, the size of Natural will typically be 31 and not
11853 32 (since no sign bit is required). Some Ada 83 compilers gave 31, and
11854 some 32 in this situation. This problem will usually show up as a compile
11855 time error, but not always. It is a good idea to check all uses of the
11856 'Size attribute when porting Ada 83 code. The GNAT specific attribute
11857 Object_Size can provide a useful way of duplicating the behavior of
11858 some Ada 83 compiler systems.
11860 @item Size of Access Types
11861 A common assumption in Ada 83 code is that an access type is in fact a pointer,
11862 and that therefore it will be the same size as a System.Address value. This
11863 assumption is true for GNAT in most cases with one exception. For the case of
11864 a pointer to an unconstrained array type (where the bounds may vary from one
11865 value of the access type to another), the default is to use a ``fat pointer'',
11866 which is represented as two separate pointers, one to the bounds, and one to
11867 the array. This representation has a number of advantages, including improved
11868 efficiency. However, it may cause some difficulties in porting existing Ada 83
11869 code which makes the assumption that, for example, pointers fit in 32 bits on
11870 a machine with 32-bit addressing.
11872 To get around this problem, GNAT also permits the use of ``thin pointers'' for
11873 access types in this case (where the designated type is an unconstrained array
11874 type). These thin pointers are indeed the same size as a System.Address value.
11875 To specify a thin pointer, use a size clause for the type, for example:
11878 type X is access all String;
11879 for X'Size use Standard'Address_Size;
11883 which will cause the type X to be represented using a single pointer. When using
11884 this representation, the bounds are right behind the array. This representation
11885 is slightly less efficient, and does not allow quite such flexibility in the
11886 use of foreign pointers or in using the Unrestricted_Access attribute to create
11887 pointers to non-aliased objects. But for any standard portable use of the access
11888 type it will work in a functionally correct manner and allow porting of existing
11889 code. Note that another way of forcing a thin pointer representation is to use
11890 a component size clause for the element size in an array, or a record
11891 representation clause for an access field in a record.
11894 @node Compatibility with DEC Ada 83
11895 @section Compatibility with DEC Ada 83
11898 The VMS version of GNAT fully implements all the pragmas and attributes
11899 provided by DEC Ada 83, as well as providing the standard DEC Ada 83
11900 libraries, including Starlet. In addition, data layouts and parameter
11901 passing conventions are highly compatible. This means that porting
11902 existing DEC Ada 83 code to GNAT in VMS systems should be easier than
11903 most other porting efforts. The following are some of the most
11904 significant differences between GNAT and DEC Ada 83.
11907 @item Default floating-point representation
11908 In GNAT, the default floating-point format is IEEE, whereas in DEC Ada 83,
11909 it is VMS format. GNAT does implement the necessary pragmas
11910 (Long_Float, Float_Representation) for changing this default.
11913 The package System in GNAT exactly corresponds to the definition in the
11914 Ada 95 reference manual, which means that it excludes many of the
11915 DEC Ada 83 extensions. However, a separate package Aux_DEC is provided
11916 that contains the additional definitions, and a special pragma,
11917 Extend_System allows this package to be treated transparently as an
11918 extension of package System.
11921 The definitions provided by Aux_DEC are exactly compatible with those
11922 in the DEC Ada 83 version of System, with one exception. DEC Ada provides
11923 the following declarations:
11926 TO_ADDRESS(INTEGER)
11927 TO_ADDRESS(UNSIGNED_LONGWORD)
11928 TO_ADDRESS(universal_integer)
11932 The version of TO_ADDRESS taking a universal integer argument is in fact
11933 an extension to Ada 83 not strictly compatible with the reference manual.
11934 In GNAT, we are constrained to be exactly compatible with the standard,
11935 and this means we cannot provide this capability. In DEC Ada 83, the
11936 point of this definition is to deal with a call like:
11939 TO_ADDRESS (16#12777#);
11943 Normally, according to the Ada 83 standard, one would expect this to be
11944 ambiguous, since it matches both the INTEGER and UNSIGNED_LONGWORD forms
11945 of TO_ADDRESS@. However, in DEC Ada 83, there is no ambiguity, since the
11946 definition using universal_integer takes precedence.
11948 In GNAT, since the version with universal_integer cannot be supplied, it is
11949 not possible to be 100% compatible. Since there are many programs using
11950 numeric constants for the argument to TO_ADDRESS, the decision in GNAT was
11951 to change the name of the function in the UNSIGNED_LONGWORD case, so the
11952 declarations provided in the GNAT version of AUX_Dec are:
11955 function To_Address (X : Integer) return Address;
11956 pragma Pure_Function (To_Address);
11958 function To_Address_Long (X : Unsigned_Longword)
11960 pragma Pure_Function (To_Address_Long);
11964 This means that programs using TO_ADDRESS for UNSIGNED_LONGWORD must
11965 change the name to TO_ADDRESS_LONG@.
11967 @item Task_Id values
11968 The Task_Id values assigned will be different in the two systems, and GNAT
11969 does not provide a specified value for the Task_Id of the environment task,
11970 which in GNAT is treated like any other declared task.
11973 For full details on these and other less significant compatibility issues,
11974 see appendix E of the Digital publication entitled @cite{DEC Ada, Technical
11975 Overview and Comparison on DIGITAL Platforms}.
11977 For GNAT running on other than VMS systems, all the DEC Ada 83 pragmas and
11978 attributes are recognized, although only a subset of them can sensibly
11979 be implemented. The description of pragmas in this reference manual
11980 indicates whether or not they are applicable to non-VMS systems.
11983 @c GNU Free Documentation License
11985 @node Index,,GNU Free Documentation License, Top
OSDN Git Service
