1 \input texinfo @c -*-texinfo-*-
5 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
7 @c GNAT DOCUMENTATION o
11 @c GNAT is maintained by Ada Core Technologies Inc (http://www.gnat.com). o
13 @c oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
15 @setfilename gnat_rm.info
18 Copyright @copyright{} 1995-2012, Free Software Foundation, Inc.
20 Permission is granted to copy, distribute and/or modify this document
21 under the terms of the GNU Free Documentation License, Version 1.3 or
22 any later version published by the Free Software Foundation; with no
23 Invariant Sections, with the Front-Cover Texts being ``GNAT Reference
24 Manual'', and with no Back-Cover Texts. A copy of the license is
25 included in the section entitled ``GNU Free Documentation License''.
30 @settitle GNAT Reference Manual
32 @setchapternewpage odd
35 @include gcc-common.texi
37 @dircategory GNU Ada tools
39 * GNAT Reference Manual: (gnat_rm). Reference Manual for GNU Ada tools.
43 @title GNAT Reference Manual
44 @subtitle GNAT, The GNU Ada Development Environment
48 @vskip 0pt plus 1filll
55 @node Top, About This Guide, (dir), (dir)
56 @top GNAT Reference Manual
62 GNAT, The GNU Ada Development Environment@*
63 GCC version @value{version-GCC}@*
70 * Implementation Defined Pragmas::
71 * Implementation Defined Aspects::
72 * Implementation Defined Attributes::
73 * Standard and Implementation Defined Restrictions::
74 * Implementation Advice::
75 * Implementation Defined Characteristics::
76 * Intrinsic Subprograms::
77 * Representation Clauses and Pragmas::
78 * Standard Library Routines::
79 * The Implementation of Standard I/O::
81 * Interfacing to Other Languages::
82 * Specialized Needs Annexes::
83 * Implementation of Specific Ada Features::
84 * Implementation of Ada 2012 Features::
85 * Obsolescent Features::
86 * GNU Free Documentation License::
87 * Index:Concept Index.
90 --- The Detailed Node Listing ---
94 * What This Reference Manual Contains::
95 * Related Information::
97 Implementation Defined Pragmas
99 * Pragma Abort_Defer::
100 * Pragma Abstract_State::
107 * Pragma Allow_Integer_Address::
110 * Pragma Assert_And_Cut::
111 * Pragma Assertion_Policy::
113 * Pragma Assume_No_Invalid_Values::
115 * Pragma Async_Readers::
116 * Pragma Async_Writers::
117 * Pragma Attribute_Definition::
118 * Pragma C_Pass_By_Copy::
120 * Pragma Check_Float_Overflow::
121 * Pragma Check_Name::
122 * Pragma Check_Policy::
123 * Pragma CIL_Constructor::
125 * Pragma Common_Object::
126 * Pragma Compile_Time_Error::
127 * Pragma Compile_Time_Warning::
128 * Pragma Compiler_Unit::
129 * Pragma Compiler_Unit_Warning::
130 * Pragma Complete_Representation::
131 * Pragma Complex_Representation::
132 * Pragma Component_Alignment::
133 * Pragma Contract_Cases::
134 * Pragma Convention_Identifier::
136 * Pragma CPP_Constructor::
137 * Pragma CPP_Virtual::
138 * Pragma CPP_Vtable::
141 * Pragma Debug_Policy::
142 * Pragma Default_Scalar_Storage_Order::
143 * Pragma Default_Storage_Pool::
145 * Pragma Detect_Blocking::
146 * Pragma Disable_Atomic_Synchronization::
147 * Pragma Dispatching_Domain::
148 * Pragma Effective_Reads::
149 * Pragma Effective_Writes::
150 * Pragma Elaboration_Checks::
152 * Pragma Enable_Atomic_Synchronization::
153 * Pragma Export_Exception::
154 * Pragma Export_Function::
155 * Pragma Export_Object::
156 * Pragma Export_Procedure::
157 * Pragma Export_Value::
158 * Pragma Export_Valued_Procedure::
159 * Pragma Extend_System::
160 * Pragma Extensions_Allowed::
162 * Pragma External_Name_Casing::
164 * Pragma Favor_Top_Level::
165 * Pragma Finalize_Storage_Only::
166 * Pragma Float_Representation::
169 * Pragma Implementation_Defined::
170 * Pragma Implemented::
171 * Pragma Implicit_Packing::
172 * Pragma Import_Exception::
173 * Pragma Import_Function::
174 * Pragma Import_Object::
175 * Pragma Import_Procedure::
176 * Pragma Import_Valued_Procedure::
177 * Pragma Independent::
178 * Pragma Independent_Components::
179 * Pragma Initial_Condition::
180 * Pragma Initialize_Scalars::
181 * Pragma Initializes::
182 * Pragma Inline_Always::
183 * Pragma Inline_Generic::
185 * Pragma Interface_Name::
186 * Pragma Interrupt_Handler::
187 * Pragma Interrupt_State::
189 * Pragma Java_Constructor::
190 * Pragma Java_Interface::
191 * Pragma Keep_Names::
194 * Pragma Linker_Alias::
195 * Pragma Linker_Constructor::
196 * Pragma Linker_Destructor::
197 * Pragma Linker_Section::
199 * Pragma Long_Float::
200 * Pragma Loop_Invariant::
201 * Pragma Loop_Optimize::
202 * Pragma Loop_Variant::
203 * Pragma Machine_Attribute::
205 * Pragma Main_Storage::
207 * Pragma No_Elaboration_Code_All::
210 * Pragma No_Run_Time::
211 * Pragma No_Strict_Aliasing ::
212 * Pragma Normalize_Scalars::
213 * Pragma Obsolescent::
214 * Pragma Optimize_Alignment::
216 * Pragma Overflow_Mode::
217 * Pragma Overriding_Renamings::
218 * Pragma Partition_Elaboration_Policy::
221 * Pragma Persistent_BSS::
224 * Pragma Postcondition::
225 * Pragma Post_Class::
227 * Pragma Precondition::
229 * Pragma Preelaborable_Initialization::
230 * Pragma Prefix_Exception_Messages::
232 * Pragma Priority_Specific_Dispatching::
234 * Pragma Profile_Warnings::
235 * Pragma Propagate_Exceptions::
236 * Pragma Provide_Shift_Operators::
237 * Pragma Psect_Object::
238 * Pragma Pure_Function::
241 * Pragma Refined_Depends::
242 * Pragma Refined_Global::
243 * Pragma Refined_Post::
244 * Pragma Refined_State::
245 * Pragma Relative_Deadline::
246 * Pragma Remote_Access_Type::
247 * Pragma Restricted_Run_Time::
248 * Pragma Restriction_Warnings::
249 * Pragma Reviewable::
250 * Pragma Share_Generic::
252 * Pragma Short_Circuit_And_Or::
253 * Pragma Short_Descriptors::
254 * Pragma Simple_Storage_Pool_Type::
255 * Pragma Source_File_Name::
256 * Pragma Source_File_Name_Project::
257 * Pragma Source_Reference::
258 * Pragma SPARK_Mode::
259 * Pragma Static_Elaboration_Desired::
260 * Pragma Stream_Convert::
261 * Pragma Style_Checks::
264 * Pragma Suppress_All::
265 * Pragma Suppress_Debug_Info::
266 * Pragma Suppress_Exception_Locations::
267 * Pragma Suppress_Initialization::
269 * Pragma Task_Storage::
271 * Pragma Thread_Local_Storage::
272 * Pragma Time_Slice::
274 * Pragma Type_Invariant::
275 * Pragma Type_Invariant_Class::
276 * Pragma Unchecked_Union::
277 * Pragma Unevaluated_Use_Of_Old::
278 * Pragma Unimplemented_Unit::
279 * Pragma Universal_Aliasing ::
280 * Pragma Universal_Data::
281 * Pragma Unmodified::
282 * Pragma Unreferenced::
283 * Pragma Unreferenced_Objects::
284 * Pragma Unreserve_All_Interrupts::
285 * Pragma Unsuppress::
286 * Pragma Use_VADS_Size::
287 * Pragma Validity_Checks::
289 * Pragma Warning_As_Error::
291 * Pragma Weak_External::
292 * Pragma Wide_Character_Encoding::
294 Implementation Defined Aspects
296 * Aspect Abstract_State::
298 * Aspect Async_Readers::
299 * Aspect Async_Writers::
300 * Aspect Contract_Cases::
303 * Aspect Dimension_System::
304 * Aspect Effective_Reads::
305 * Aspect Effective_Writes::
306 * Aspect Favor_Top_Level::
308 * Aspect Initial_Condition::
309 * Aspect Initializes::
310 * Aspect Inline_Always::
312 * Aspect Invariant'Class::
314 * Aspect Linker_Section::
315 * Aspect No_Elaboration_Code_All::
316 * Aspect Object_Size::
317 * Aspect Obsolescent::
319 * Aspect Persistent_BSS::
321 * Aspect Pure_Function::
322 * Aspect Refined_Depends::
323 * Aspect Refined_Global::
324 * Aspect Refined_Post::
325 * Aspect Refined_State::
326 * Aspect Remote_Access_Type::
327 * Aspect Scalar_Storage_Order::
329 * Aspect Simple_Storage_Pool::
330 * Aspect Simple_Storage_Pool_Type::
331 * Aspect SPARK_Mode::
332 * Aspect Suppress_Debug_Info::
334 * Aspect Thread_Local_Storage::
335 * Aspect Universal_Aliasing::
336 * Aspect Universal_Data::
337 * Aspect Unmodified::
338 * Aspect Unreferenced::
339 * Aspect Unreferenced_Objects::
340 * Aspect Value_Size::
343 Implementation Defined Attributes
345 * Attribute Abort_Signal::
346 * Attribute Address_Size::
347 * Attribute Asm_Input::
348 * Attribute Asm_Output::
349 * Attribute AST_Entry::
350 * Attribute Atomic_Always_Lock_Free::
352 * Attribute Bit_Position::
353 * Attribute Code_Address::
354 * Attribute Compiler_Version::
355 * Attribute Constrained::
356 * Attribute Default_Bit_Order::
357 * Attribute Default_Scalar_Storage_Order::
358 * Attribute Descriptor_Size::
359 * Attribute Elaborated::
360 * Attribute Elab_Body::
361 * Attribute Elab_Spec::
362 * Attribute Elab_Subp_Body::
364 * Attribute Enabled::
365 * Attribute Enum_Rep::
366 * Attribute Enum_Val::
367 * Attribute Epsilon::
368 * Attribute Fast_Math::
369 * Attribute Fixed_Value::
370 * Attribute From_Any::
371 * Attribute Has_Access_Values::
372 * Attribute Has_Discriminants::
374 * Attribute Integer_Value::
375 * Attribute Invalid_Value::
376 * Attribute Iterable::
378 * Attribute Library_Level::
379 * Attribute Lock_Free::
380 * Attribute Loop_Entry::
381 * Attribute Machine_Size::
382 * Attribute Mantissa::
383 * Attribute Maximum_Alignment::
384 * Attribute Mechanism_Code::
385 * Attribute Null_Parameter::
386 * Attribute Object_Size::
388 * Attribute Passed_By_Reference::
389 * Attribute Pool_Address::
390 * Attribute Range_Length::
392 * Attribute Restriction_Set::
394 * Attribute Safe_Emax::
395 * Attribute Safe_Large::
396 * Attribute Safe_Small::
397 * Attribute Scalar_Storage_Order::
398 * Attribute Simple_Storage_Pool::
400 * Attribute Storage_Unit::
401 * Attribute Stub_Type::
402 * Attribute System_Allocator_Alignment::
403 * Attribute Target_Name::
404 * Attribute To_Address::
406 * Attribute Type_Class::
407 * Attribute Type_Key::
408 * Attribute TypeCode::
409 * Attribute UET_Address::
410 * Attribute Unconstrained_Array::
411 * Attribute Universal_Literal_String::
412 * Attribute Unrestricted_Access::
414 * Attribute VADS_Size::
415 * Attribute Valid_Scalars::
416 * Attribute Value_Size::
417 * Attribute Wchar_T_Size::
418 * Attribute Word_Size::
420 Standard and Implementation Defined Restrictions
422 * Partition-Wide Restrictions::
423 * Program Unit Level Restrictions::
425 Partition-Wide Restrictions
427 * Immediate_Reclamation::
428 * Max_Asynchronous_Select_Nesting::
429 * Max_Entry_Queue_Length::
430 * Max_Protected_Entries::
431 * Max_Select_Alternatives::
432 * Max_Storage_At_Blocking::
435 * No_Abort_Statements::
436 * No_Access_Parameter_Allocators::
437 * No_Access_Subprograms::
439 * No_Anonymous_Allocators::
442 * No_Default_Initialization::
445 * No_Direct_Boolean_Operators::
447 * No_Dispatching_Calls::
448 * No_Dynamic_Attachment::
449 * No_Dynamic_Priorities::
450 * No_Entry_Calls_In_Elaboration_Code::
451 * No_Enumeration_Maps::
452 * No_Exception_Handlers::
453 * No_Exception_Propagation::
454 * No_Exception_Registration::
458 * No_Floating_Point::
459 * No_Implicit_Conditionals::
460 * No_Implicit_Dynamic_Code::
461 * No_Implicit_Heap_Allocations::
462 * No_Implicit_Loops::
463 * No_Initialize_Scalars::
465 * No_Local_Allocators::
466 * No_Local_Protected_Objects::
467 * No_Local_Timing_Events::
468 * No_Long_Long_Integers::
469 * No_Multiple_Elaboration::
470 * No_Nested_Finalization::
471 * No_Protected_Type_Allocators::
472 * No_Protected_Types::
475 * No_Relative_Delay::
476 * No_Requeue_Statements::
477 * No_Secondary_Stack::
478 * No_Select_Statements::
479 * No_Specific_Termination_Handlers::
480 * No_Specification_of_Aspect::
481 * No_Standard_Allocators_After_Elaboration::
482 * No_Standard_Storage_Pools::
483 * No_Stream_Optimizations::
485 * No_Task_Allocators::
486 * No_Task_Attributes_Package::
487 * No_Task_Hierarchy::
488 * No_Task_Termination::
490 * No_Terminate_Alternatives::
491 * No_Unchecked_Access::
493 * Static_Priorities::
494 * Static_Storage_Size::
496 Program Unit Level Restrictions
498 * No_Elaboration_Code::
500 * No_Implementation_Aspect_Specifications::
501 * No_Implementation_Attributes::
502 * No_Implementation_Identifiers::
503 * No_Implementation_Pragmas::
504 * No_Implementation_Restrictions::
505 * No_Implementation_Units::
506 * No_Implicit_Aliasing::
507 * No_Obsolescent_Features::
508 * No_Wide_Characters::
511 The Implementation of Standard I/O
513 * Standard I/O Packages::
519 * Wide_Wide_Text_IO::
523 * Filenames encoding::
524 * File content encoding::
526 * Operations on C Streams::
527 * Interfacing to C Streams::
531 * Ada.Characters.Latin_9 (a-chlat9.ads)::
532 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
533 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
534 * Ada.Characters.Wide_Wide_Latin_1 (a-chzla1.ads)::
535 * Ada.Characters.Wide_Wide_Latin_9 (a-chzla9.ads)::
536 * Ada.Containers.Formal_Doubly_Linked_Lists (a-cfdlli.ads)::
537 * Ada.Containers.Formal_Hashed_Maps (a-cfhama.ads)::
538 * Ada.Containers.Formal_Hashed_Sets (a-cfhase.ads)::
539 * Ada.Containers.Formal_Ordered_Maps (a-cforma.ads)::
540 * Ada.Containers.Formal_Ordered_Sets (a-cforse.ads)::
541 * Ada.Containers.Formal_Vectors (a-cofove.ads)::
542 * Ada.Command_Line.Environment (a-colien.ads)::
543 * Ada.Command_Line.Remove (a-colire.ads)::
544 * Ada.Command_Line.Response_File (a-clrefi.ads)::
545 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
546 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
547 * Ada.Exceptions.Last_Chance_Handler (a-elchha.ads)::
548 * Ada.Exceptions.Traceback (a-exctra.ads)::
549 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
550 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
551 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
552 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
553 * Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO (a-szuzti.ads)::
554 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
555 * Ada.Text_IO.Reset_Standard_Files (a-tirsfi.ads)::
556 * Ada.Wide_Characters.Unicode (a-wichun.ads)::
557 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
558 * Ada.Wide_Text_IO.Reset_Standard_Files (a-wrstfi.ads)::
559 * Ada.Wide_Wide_Characters.Unicode (a-zchuni.ads)::
560 * Ada.Wide_Wide_Text_IO.C_Streams (a-ztcstr.ads)::
561 * Ada.Wide_Wide_Text_IO.Reset_Standard_Files (a-zrstfi.ads)::
562 * GNAT.Altivec (g-altive.ads)::
563 * GNAT.Altivec.Conversions (g-altcon.ads)::
564 * GNAT.Altivec.Vector_Operations (g-alveop.ads)::
565 * GNAT.Altivec.Vector_Types (g-alvety.ads)::
566 * GNAT.Altivec.Vector_Views (g-alvevi.ads)::
567 * GNAT.Array_Split (g-arrspl.ads)::
568 * GNAT.AWK (g-awk.ads)::
569 * GNAT.Bounded_Buffers (g-boubuf.ads)::
570 * GNAT.Bounded_Mailboxes (g-boumai.ads)::
571 * GNAT.Bubble_Sort (g-bubsor.ads)::
572 * GNAT.Bubble_Sort_A (g-busora.ads)::
573 * GNAT.Bubble_Sort_G (g-busorg.ads)::
574 * GNAT.Byte_Order_Mark (g-byorma.ads)::
575 * GNAT.Byte_Swapping (g-bytswa.ads)::
576 * GNAT.Calendar (g-calend.ads)::
577 * GNAT.Calendar.Time_IO (g-catiio.ads)::
578 * GNAT.Case_Util (g-casuti.ads)::
579 * GNAT.CGI (g-cgi.ads)::
580 * GNAT.CGI.Cookie (g-cgicoo.ads)::
581 * GNAT.CGI.Debug (g-cgideb.ads)::
582 * GNAT.Command_Line (g-comlin.ads)::
583 * GNAT.Compiler_Version (g-comver.ads)::
584 * GNAT.Ctrl_C (g-ctrl_c.ads)::
585 * GNAT.CRC32 (g-crc32.ads)::
586 * GNAT.Current_Exception (g-curexc.ads)::
587 * GNAT.Debug_Pools (g-debpoo.ads)::
588 * GNAT.Debug_Utilities (g-debuti.ads)::
589 * GNAT.Decode_String (g-decstr.ads)::
590 * GNAT.Decode_UTF8_String (g-deutst.ads)::
591 * GNAT.Directory_Operations (g-dirope.ads)::
592 * GNAT.Directory_Operations.Iteration (g-diopit.ads)::
593 * GNAT.Dynamic_HTables (g-dynhta.ads)::
594 * GNAT.Dynamic_Tables (g-dyntab.ads)::
595 * GNAT.Encode_String (g-encstr.ads)::
596 * GNAT.Encode_UTF8_String (g-enutst.ads)::
597 * GNAT.Exception_Actions (g-excact.ads)::
598 * GNAT.Exception_Traces (g-exctra.ads)::
599 * GNAT.Exceptions (g-except.ads)::
600 * GNAT.Expect (g-expect.ads)::
601 * GNAT.Expect.TTY (g-exptty.ads)::
602 * GNAT.Float_Control (g-flocon.ads)::
603 * GNAT.Formatted_String (g-forstr.ads)::
604 * GNAT.Heap_Sort (g-heasor.ads)::
605 * GNAT.Heap_Sort_A (g-hesora.ads)::
606 * GNAT.Heap_Sort_G (g-hesorg.ads)::
607 * GNAT.HTable (g-htable.ads)::
608 * GNAT.IO (g-io.ads)::
609 * GNAT.IO_Aux (g-io_aux.ads)::
610 * GNAT.Lock_Files (g-locfil.ads)::
611 * GNAT.MBBS_Discrete_Random (g-mbdira.ads)::
612 * GNAT.MBBS_Float_Random (g-mbflra.ads)::
613 * GNAT.MD5 (g-md5.ads)::
614 * GNAT.Memory_Dump (g-memdum.ads)::
615 * GNAT.Most_Recent_Exception (g-moreex.ads)::
616 * GNAT.OS_Lib (g-os_lib.ads)::
617 * GNAT.Perfect_Hash_Generators (g-pehage.ads)::
618 * GNAT.Random_Numbers (g-rannum.ads)::
619 * GNAT.Regexp (g-regexp.ads)::
620 * GNAT.Registry (g-regist.ads)::
621 * GNAT.Regpat (g-regpat.ads)::
622 * GNAT.Rewrite_Data (g-rewdat.ads)::
623 * GNAT.Secondary_Stack_Info (g-sestin.ads)::
624 * GNAT.Semaphores (g-semaph.ads)::
625 * GNAT.Serial_Communications (g-sercom.ads)::
626 * GNAT.SHA1 (g-sha1.ads)::
627 * GNAT.SHA224 (g-sha224.ads)::
628 * GNAT.SHA256 (g-sha256.ads)::
629 * GNAT.SHA384 (g-sha384.ads)::
630 * GNAT.SHA512 (g-sha512.ads)::
631 * GNAT.Signals (g-signal.ads)::
632 * GNAT.Sockets (g-socket.ads)::
633 * GNAT.Source_Info (g-souinf.ads)::
634 * GNAT.Spelling_Checker (g-speche.ads)::
635 * GNAT.Spelling_Checker_Generic (g-spchge.ads)::
636 * GNAT.Spitbol.Patterns (g-spipat.ads)::
637 * GNAT.Spitbol (g-spitbo.ads)::
638 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
639 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
640 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
641 * GNAT.SSE (g-sse.ads)::
642 * GNAT.SSE.Vector_Types (g-ssvety.ads)::
643 * GNAT.Strings (g-string.ads)::
644 * GNAT.String_Split (g-strspl.ads)::
645 * GNAT.Table (g-table.ads)::
646 * GNAT.Task_Lock (g-tasloc.ads)::
647 * GNAT.Threads (g-thread.ads)::
648 * GNAT.Time_Stamp (g-timsta.ads)::
649 * GNAT.Traceback (g-traceb.ads)::
650 * GNAT.Traceback.Symbolic (g-trasym.ads)::
651 * GNAT.UTF_32 (g-utf_32.ads)::
652 * GNAT.UTF_32_Spelling_Checker (g-u3spch.ads)::
653 * GNAT.Wide_Spelling_Checker (g-wispch.ads)::
654 * GNAT.Wide_String_Split (g-wistsp.ads)::
655 * GNAT.Wide_Wide_Spelling_Checker (g-zspche.ads)::
656 * GNAT.Wide_Wide_String_Split (g-zistsp.ads)::
657 * Interfaces.C.Extensions (i-cexten.ads)::
658 * Interfaces.C.Streams (i-cstrea.ads)::
659 * Interfaces.Packed_Decimal (i-pacdec.ads)::
660 * Interfaces.VxWorks (i-vxwork.ads)::
661 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
662 * System.Address_Image (s-addima.ads)::
663 * System.Assertions (s-assert.ads)::
664 * System.Atomic_Counters (s-atocou.ads)::
665 * System.Memory (s-memory.ads)::
666 * System.Multiprocessors (s-multip.ads)::
667 * System.Multiprocessors.Dispatching_Domains (s-mudido.ads)::
668 * System.Partition_Interface (s-parint.ads)::
669 * System.Pool_Global (s-pooglo.ads)::
670 * System.Pool_Local (s-pooloc.ads)::
671 * System.Restrictions (s-restri.ads)::
672 * System.Rident (s-rident.ads)::
673 * System.Strings.Stream_Ops (s-ststop.ads)::
674 * System.Unsigned_Types (s-unstyp.ads)::
675 * System.Wch_Cnv (s-wchcnv.ads)::
676 * System.Wch_Con (s-wchcon.ads)::
680 * Text_IO Stream Pointer Positioning::
681 * Text_IO Reading and Writing Non-Regular Files::
683 * Treating Text_IO Files as Streams::
684 * Text_IO Extensions::
685 * Text_IO Facilities for Unbounded Strings::
689 * Wide_Text_IO Stream Pointer Positioning::
690 * Wide_Text_IO Reading and Writing Non-Regular Files::
694 * Wide_Wide_Text_IO Stream Pointer Positioning::
695 * Wide_Wide_Text_IO Reading and Writing Non-Regular Files::
697 Interfacing to Other Languages
700 * Interfacing to C++::
701 * Interfacing to COBOL::
702 * Interfacing to Fortran::
703 * Interfacing to non-GNAT Ada code::
705 Specialized Needs Annexes
707 Implementation of Specific Ada Features
708 * Machine Code Insertions::
709 * GNAT Implementation of Tasking::
710 * GNAT Implementation of Shared Passive Packages::
711 * Code Generation for Array Aggregates::
712 * The Size of Discriminated Records with Default Discriminants::
713 * Strict Conformance to the Ada Reference Manual::
715 Implementation of Ada 2012 Features
719 GNU Free Documentation License
727 @node About This Guide
728 @unnumbered About This Guide
731 This manual contains useful information in writing programs using the
732 @value{EDITION} compiler. It includes information on implementation dependent
733 characteristics of @value{EDITION}, including all the information required by
734 Annex M of the Ada language standard.
736 @value{EDITION} implements Ada 95, Ada 2005 and Ada 2012, and it may also be
737 invoked in Ada 83 compatibility mode.
738 By default, @value{EDITION} assumes Ada 2012,
739 but you can override with a compiler switch
740 to explicitly specify the language version.
741 (Please refer to @ref{Compiling Different Versions of Ada,,, gnat_ugn,
742 @value{EDITION} User's Guide}, for details on these switches.)
743 Throughout this manual, references to ``Ada'' without a year suffix
744 apply to all the Ada versions of the language.
746 Ada is designed to be highly portable.
747 In general, a program will have the same effect even when compiled by
748 different compilers on different platforms.
749 However, since Ada is designed to be used in a
750 wide variety of applications, it also contains a number of system
751 dependent features to be used in interfacing to the external world.
752 @cindex Implementation-dependent features
755 Note: Any program that makes use of implementation-dependent features
756 may be non-portable. You should follow good programming practice and
757 isolate and clearly document any sections of your program that make use
758 of these features in a non-portable manner.
761 For ease of exposition, ``@value{EDITION}'' will be referred to simply as
762 ``GNAT'' in the remainder of this document.
766 * What This Reference Manual Contains::
768 * Related Information::
771 @node What This Reference Manual Contains
772 @unnumberedsec What This Reference Manual Contains
775 This reference manual contains the following chapters:
779 @ref{Implementation Defined Pragmas}, lists GNAT implementation-dependent
780 pragmas, which can be used to extend and enhance the functionality of the
784 @ref{Implementation Defined Attributes}, lists GNAT
785 implementation-dependent attributes, which can be used to extend and
786 enhance the functionality of the compiler.
789 @ref{Standard and Implementation Defined Restrictions}, lists GNAT
790 implementation-dependent restrictions, which can be used to extend and
791 enhance the functionality of the compiler.
794 @ref{Implementation Advice}, provides information on generally
795 desirable behavior which are not requirements that all compilers must
796 follow since it cannot be provided on all systems, or which may be
797 undesirable on some systems.
800 @ref{Implementation Defined Characteristics}, provides a guide to
801 minimizing implementation dependent features.
804 @ref{Intrinsic Subprograms}, describes the intrinsic subprograms
805 implemented by GNAT, and how they can be imported into user
806 application programs.
809 @ref{Representation Clauses and Pragmas}, describes in detail the
810 way that GNAT represents data, and in particular the exact set
811 of representation clauses and pragmas that is accepted.
814 @ref{Standard Library Routines}, provides a listing of packages and a
815 brief description of the functionality that is provided by Ada's
816 extensive set of standard library routines as implemented by GNAT@.
819 @ref{The Implementation of Standard I/O}, details how the GNAT
820 implementation of the input-output facilities.
823 @ref{The GNAT Library}, is a catalog of packages that complement
824 the Ada predefined library.
827 @ref{Interfacing to Other Languages}, describes how programs
828 written in Ada using GNAT can be interfaced to other programming
831 @ref{Specialized Needs Annexes}, describes the GNAT implementation of all
832 of the specialized needs annexes.
835 @ref{Implementation of Specific Ada Features}, discusses issues related
836 to GNAT's implementation of machine code insertions, tasking, and several
840 @ref{Implementation of Ada 2012 Features}, describes the status of the
841 GNAT implementation of the Ada 2012 language standard.
844 @ref{Obsolescent Features} documents implementation dependent features,
845 including pragmas and attributes, which are considered obsolescent, since
846 there are other preferred ways of achieving the same results. These
847 obsolescent forms are retained for backwards compatibility.
851 @cindex Ada 95 Language Reference Manual
852 @cindex Ada 2005 Language Reference Manual
854 This reference manual assumes a basic familiarity with the Ada 95 language, as
855 described in the International Standard ANSI/ISO/IEC-8652:1995,
857 It does not require knowledge of the new features introduced by Ada 2005,
858 (officially known as ISO/IEC 8652:1995 with Technical Corrigendum 1
860 Both reference manuals are included in the GNAT documentation
864 @unnumberedsec Conventions
865 @cindex Conventions, typographical
866 @cindex Typographical conventions
869 Following are examples of the typographical and graphic conventions used
874 @code{Functions}, @code{utility program names}, @code{standard names},
881 @file{File names}, @samp{button names}, and @samp{field names}.
884 @code{Variables}, @env{environment variables}, and @var{metasyntactic
891 [optional information or parameters]
894 Examples are described by text
896 and then shown this way.
901 Commands that are entered by the user are preceded in this manual by the
902 characters @samp{$ } (dollar sign followed by space). If your system uses this
903 sequence as a prompt, then the commands will appear exactly as you see them
904 in the manual. If your system uses some other prompt, then the command will
905 appear with the @samp{$} replaced by whatever prompt character you are using.
907 @node Related Information
908 @unnumberedsec Related Information
910 See the following documents for further information on GNAT:
914 @xref{Top, @value{EDITION} User's Guide, About This Guide, gnat_ugn,
915 @value{EDITION} User's Guide}, which provides information on how to use the
916 GNAT compiler system.
919 @cite{Ada 95 Reference Manual}, which contains all reference
920 material for the Ada 95 programming language.
923 @cite{Ada 95 Annotated Reference Manual}, which is an annotated version
924 of the Ada 95 standard. The annotations describe
925 detailed aspects of the design decision, and in particular contain useful
926 sections on Ada 83 compatibility.
929 @cite{Ada 2005 Reference Manual}, which contains all reference
930 material for the Ada 2005 programming language.
933 @cite{Ada 2005 Annotated Reference Manual}, which is an annotated version
934 of the Ada 2005 standard. The annotations describe
935 detailed aspects of the design decision, and in particular contain useful
936 sections on Ada 83 and Ada 95 compatibility.
939 @cite{DEC Ada, Technical Overview and Comparison on DIGITAL Platforms},
940 which contains specific information on compatibility between GNAT and
944 @cite{DEC Ada, Language Reference Manual, part number AA-PYZAB-TK} which
945 describes in detail the pragmas and attributes provided by the DEC Ada 83
950 @node Implementation Defined Pragmas
951 @chapter Implementation Defined Pragmas
954 Ada defines a set of pragmas that can be used to supply additional
955 information to the compiler. These language defined pragmas are
956 implemented in GNAT and work as described in the Ada Reference Manual.
958 In addition, Ada allows implementations to define additional pragmas
959 whose meaning is defined by the implementation. GNAT provides a number
960 of these implementation-defined pragmas, which can be used to extend
961 and enhance the functionality of the compiler. This section of the GNAT
962 Reference Manual describes these additional pragmas.
964 Note that any program using these pragmas might not be portable to other
965 compilers (although GNAT implements this set of pragmas on all
966 platforms). Therefore if portability to other compilers is an important
967 consideration, the use of these pragmas should be minimized.
970 * Pragma Abort_Defer::
971 * Pragma Abstract_State::
978 * Pragma Allow_Integer_Address::
981 * Pragma Assert_And_Cut::
982 * Pragma Assertion_Policy::
984 * Pragma Assume_No_Invalid_Values::
986 * Pragma Async_Readers::
987 * Pragma Async_Writers::
988 * Pragma Attribute_Definition::
989 * Pragma C_Pass_By_Copy::
991 * Pragma Check_Float_Overflow::
992 * Pragma Check_Name::
993 * Pragma Check_Policy::
994 * Pragma CIL_Constructor::
996 * Pragma Common_Object::
997 * Pragma Compile_Time_Error::
998 * Pragma Compile_Time_Warning::
999 * Pragma Compiler_Unit::
1000 * Pragma Compiler_Unit_Warning::
1001 * Pragma Complete_Representation::
1002 * Pragma Complex_Representation::
1003 * Pragma Component_Alignment::
1004 * Pragma Contract_Cases::
1005 * Pragma Convention_Identifier::
1006 * Pragma CPP_Class::
1007 * Pragma CPP_Constructor::
1008 * Pragma CPP_Virtual::
1009 * Pragma CPP_Vtable::
1012 * Pragma Debug_Policy::
1013 * Pragma Default_Scalar_Storage_Order::
1014 * Pragma Default_Storage_Pool::
1016 * Pragma Detect_Blocking::
1017 * Pragma Disable_Atomic_Synchronization::
1018 * Pragma Dispatching_Domain::
1019 * Pragma Effective_Reads::
1020 * Pragma Effective_Writes::
1021 * Pragma Elaboration_Checks::
1022 * Pragma Eliminate::
1023 * Pragma Enable_Atomic_Synchronization::
1024 * Pragma Export_Exception::
1025 * Pragma Export_Function::
1026 * Pragma Export_Object::
1027 * Pragma Export_Procedure::
1028 * Pragma Export_Value::
1029 * Pragma Export_Valued_Procedure::
1030 * Pragma Extend_System::
1031 * Pragma Extensions_Allowed::
1033 * Pragma External_Name_Casing::
1034 * Pragma Fast_Math::
1035 * Pragma Favor_Top_Level::
1036 * Pragma Finalize_Storage_Only::
1037 * Pragma Float_Representation::
1040 * Pragma Implementation_Defined::
1041 * Pragma Implemented::
1042 * Pragma Implicit_Packing::
1043 * Pragma Import_Exception::
1044 * Pragma Import_Function::
1045 * Pragma Import_Object::
1046 * Pragma Import_Procedure::
1047 * Pragma Import_Valued_Procedure::
1048 * Pragma Independent::
1049 * Pragma Independent_Components::
1050 * Pragma Initial_Condition::
1051 * Pragma Initialize_Scalars::
1052 * Pragma Initializes::
1053 * Pragma Inline_Always::
1054 * Pragma Inline_Generic::
1055 * Pragma Interface::
1056 * Pragma Interface_Name::
1057 * Pragma Interrupt_Handler::
1058 * Pragma Interrupt_State::
1059 * Pragma Invariant::
1060 * Pragma Java_Constructor::
1061 * Pragma Java_Interface::
1062 * Pragma Keep_Names::
1064 * Pragma Link_With::
1065 * Pragma Linker_Alias::
1066 * Pragma Linker_Constructor::
1067 * Pragma Linker_Destructor::
1068 * Pragma Linker_Section::
1069 * Pragma Lock_Free::
1070 * Pragma Long_Float::
1071 * Pragma Loop_Invariant::
1072 * Pragma Loop_Optimize::
1073 * Pragma Loop_Variant::
1074 * Pragma Machine_Attribute::
1076 * Pragma Main_Storage::
1078 * Pragma No_Elaboration_Code_All::
1079 * Pragma No_Inline::
1080 * Pragma No_Return::
1081 * Pragma No_Run_Time::
1082 * Pragma No_Strict_Aliasing::
1083 * Pragma Normalize_Scalars::
1084 * Pragma Obsolescent::
1085 * Pragma Optimize_Alignment::
1087 * Pragma Overflow_Mode::
1088 * Pragma Overriding_Renamings::
1089 * Pragma Partition_Elaboration_Policy::
1092 * Pragma Persistent_BSS::
1095 * Pragma Postcondition::
1096 * Pragma Post_Class::
1098 * Pragma Precondition::
1099 * Pragma Predicate::
1100 * Pragma Preelaborable_Initialization::
1101 * Pragma Prefix_Exception_Messages::
1102 * Pragma Pre_Class::
1103 * Pragma Priority_Specific_Dispatching::
1105 * Pragma Profile_Warnings::
1106 * Pragma Propagate_Exceptions::
1107 * Pragma Provide_Shift_Operators::
1108 * Pragma Psect_Object::
1109 * Pragma Pure_Function::
1111 * Pragma Ravenscar::
1112 * Pragma Refined_Depends::
1113 * Pragma Refined_Global::
1114 * Pragma Refined_Post::
1115 * Pragma Refined_State::
1116 * Pragma Relative_Deadline::
1117 * Pragma Remote_Access_Type::
1118 * Pragma Restricted_Run_Time::
1119 * Pragma Restriction_Warnings::
1120 * Pragma Reviewable::
1121 * Pragma Share_Generic::
1123 * Pragma Short_Circuit_And_Or::
1124 * Pragma Short_Descriptors::
1125 * Pragma Simple_Storage_Pool_Type::
1126 * Pragma Source_File_Name::
1127 * Pragma Source_File_Name_Project::
1128 * Pragma Source_Reference::
1129 * Pragma SPARK_Mode::
1130 * Pragma Static_Elaboration_Desired::
1131 * Pragma Stream_Convert::
1132 * Pragma Style_Checks::
1135 * Pragma Suppress_All::
1136 * Pragma Suppress_Debug_Info::
1137 * Pragma Suppress_Exception_Locations::
1138 * Pragma Suppress_Initialization::
1139 * Pragma Task_Name::
1140 * Pragma Task_Storage::
1141 * Pragma Test_Case::
1142 * Pragma Thread_Local_Storage::
1143 * Pragma Time_Slice::
1145 * Pragma Type_Invariant::
1146 * Pragma Type_Invariant_Class::
1147 * Pragma Unchecked_Union::
1148 * Pragma Unevaluated_Use_Of_Old::
1149 * Pragma Unimplemented_Unit::
1150 * Pragma Universal_Aliasing ::
1151 * Pragma Universal_Data::
1152 * Pragma Unmodified::
1153 * Pragma Unreferenced::
1154 * Pragma Unreferenced_Objects::
1155 * Pragma Unreserve_All_Interrupts::
1156 * Pragma Unsuppress::
1157 * Pragma Use_VADS_Size::
1158 * Pragma Validity_Checks::
1160 * Pragma Warning_As_Error::
1162 * Pragma Weak_External::
1163 * Pragma Wide_Character_Encoding::
1166 @node Pragma Abort_Defer
1167 @unnumberedsec Pragma Abort_Defer
1169 @cindex Deferring aborts
1177 This pragma must appear at the start of the statement sequence of a
1178 handled sequence of statements (right after the @code{begin}). It has
1179 the effect of deferring aborts for the sequence of statements (but not
1180 for the declarations or handlers, if any, associated with this statement
1183 @node Pragma Abstract_State
1184 @unnumberedsec Pragma Abstract_State
1185 @findex Abstract_State
1187 For the description of this pragma, see SPARK 2014 Reference Manual,
1191 @unnumberedsec Pragma Ada_83
1195 @smallexample @c ada
1200 A configuration pragma that establishes Ada 83 mode for the unit to
1201 which it applies, regardless of the mode set by the command line
1202 switches. In Ada 83 mode, GNAT attempts to be as compatible with
1203 the syntax and semantics of Ada 83, as defined in the original Ada
1204 83 Reference Manual as possible. In particular, the keywords added by Ada 95
1205 and Ada 2005 are not recognized, optional package bodies are allowed,
1206 and generics may name types with unknown discriminants without using
1207 the @code{(<>)} notation. In addition, some but not all of the additional
1208 restrictions of Ada 83 are enforced.
1210 Ada 83 mode is intended for two purposes. Firstly, it allows existing
1211 Ada 83 code to be compiled and adapted to GNAT with less effort.
1212 Secondly, it aids in keeping code backwards compatible with Ada 83.
1213 However, there is no guarantee that code that is processed correctly
1214 by GNAT in Ada 83 mode will in fact compile and execute with an Ada
1215 83 compiler, since GNAT does not enforce all the additional checks
1219 @unnumberedsec Pragma Ada_95
1223 @smallexample @c ada
1228 A configuration pragma that establishes Ada 95 mode for the unit to which
1229 it applies, regardless of the mode set by the command line switches.
1230 This mode is set automatically for the @code{Ada} and @code{System}
1231 packages and their children, so you need not specify it in these
1232 contexts. This pragma is useful when writing a reusable component that
1233 itself uses Ada 95 features, but which is intended to be usable from
1234 either Ada 83 or Ada 95 programs.
1237 @unnumberedsec Pragma Ada_05
1241 @smallexample @c ada
1243 pragma Ada_05 (local_NAME);
1247 A configuration pragma that establishes Ada 2005 mode for the unit to which
1248 it applies, regardless of the mode set by the command line switches.
1249 This pragma is useful when writing a reusable component that
1250 itself uses Ada 2005 features, but which is intended to be usable from
1251 either Ada 83 or Ada 95 programs.
1253 The one argument form (which is not a configuration pragma)
1254 is used for managing the transition from
1255 Ada 95 to Ada 2005 in the run-time library. If an entity is marked
1256 as Ada_2005 only, then referencing the entity in Ada_83 or Ada_95
1257 mode will generate a warning. In addition, in Ada_83 or Ada_95
1258 mode, a preference rule is established which does not choose
1259 such an entity unless it is unambiguously specified. This avoids
1260 extra subprograms marked this way from generating ambiguities in
1261 otherwise legal pre-Ada_2005 programs. The one argument form is
1262 intended for exclusive use in the GNAT run-time library.
1264 @node Pragma Ada_2005
1265 @unnumberedsec Pragma Ada_2005
1269 @smallexample @c ada
1274 This configuration pragma is a synonym for pragma Ada_05 and has the
1275 same syntax and effect.
1278 @unnumberedsec Pragma Ada_12
1282 @smallexample @c ada
1284 pragma Ada_12 (local_NAME);
1288 A configuration pragma that establishes Ada 2012 mode for the unit to which
1289 it applies, regardless of the mode set by the command line switches.
1290 This mode is set automatically for the @code{Ada} and @code{System}
1291 packages and their children, so you need not specify it in these
1292 contexts. This pragma is useful when writing a reusable component that
1293 itself uses Ada 2012 features, but which is intended to be usable from
1294 Ada 83, Ada 95, or Ada 2005 programs.
1296 The one argument form, which is not a configuration pragma,
1297 is used for managing the transition from Ada
1298 2005 to Ada 2012 in the run-time library. If an entity is marked
1299 as Ada_201 only, then referencing the entity in any pre-Ada_2012
1300 mode will generate a warning. In addition, in any pre-Ada_2012
1301 mode, a preference rule is established which does not choose
1302 such an entity unless it is unambiguously specified. This avoids
1303 extra subprograms marked this way from generating ambiguities in
1304 otherwise legal pre-Ada_2012 programs. The one argument form is
1305 intended for exclusive use in the GNAT run-time library.
1307 @node Pragma Ada_2012
1308 @unnumberedsec Pragma Ada_2012
1312 @smallexample @c ada
1317 This configuration pragma is a synonym for pragma Ada_12 and has the
1318 same syntax and effect.
1320 @node Pragma Allow_Integer_Address
1321 @unnumberedsec Pragma Allow_Integer_Address
1322 @findex Allow_Integer_Address
1325 @smallexample @c ada
1326 pragma Allow_Integer_Address;
1330 In almost all versions of GNAT, @code{System.Address} is a private
1331 type in accordance with the implementation advice in the RM. This
1332 means that integer values,
1333 in particular integer literals, are not allowed as address values.
1334 If the configuration pragma
1335 @code{Allow_Integer_Address} is given, then integer expressions may
1336 be used anywhere a value of type @code{System.Address} is required.
1337 The effect is to introduce an implicit unchecked conversion from the
1338 integer value to type @code{System.Address}. The reverse case of using
1339 an address where an integer type is required is handled analogously.
1340 The following example compiles without errors:
1342 @smallexample @c ada
1343 pragma Allow_Integer_Address;
1344 with System; use System;
1345 package AddrAsInt is
1348 for X'Address use 16#1240#;
1349 for Y use at 16#3230#;
1350 m : Address := 16#4000#;
1351 n : constant Address := 4000;
1352 p : constant Address := Address (X + Y);
1353 v : Integer := y'Address;
1354 w : constant Integer := Integer (Y'Address);
1355 type R is new integer;
1358 for Z'Address use RR;
1363 Note that pragma @code{Allow_Integer_Address} is ignored if
1364 @code{System.Address}
1365 is not a private type. In implementations of @code{GNAT} where
1366 System.Address is a visible integer type (notably the implementations
1367 for @code{OpenVMS}), this pragma serves no purpose but is ignored
1368 rather than rejected to allow common sets of sources to be used
1369 in the two situations.
1371 @node Pragma Annotate
1372 @unnumberedsec Pragma Annotate
1376 @smallexample @c ada
1377 pragma Annotate (IDENTIFIER [,IDENTIFIER @{, ARG@}] [entity => local_NAME]);
1379 ARG ::= NAME | EXPRESSION
1383 This pragma is used to annotate programs. @var{identifier} identifies
1384 the type of annotation. GNAT verifies that it is an identifier, but does
1385 not otherwise analyze it. The second optional identifier is also left
1386 unanalyzed, and by convention is used to control the action of the tool to
1387 which the annotation is addressed. The remaining @var{arg} arguments
1388 can be either string literals or more generally expressions.
1389 String literals are assumed to be either of type
1390 @code{Standard.String} or else @code{Wide_String} or @code{Wide_Wide_String}
1391 depending on the character literals they contain.
1392 All other kinds of arguments are analyzed as expressions, and must be
1393 unambiguous. The last argument if present must have the identifier
1394 @code{Entity} and GNAT verifies that a local name is given.
1396 The analyzed pragma is retained in the tree, but not otherwise processed
1397 by any part of the GNAT compiler, except to generate corresponding note
1398 lines in the generated ALI file. For the format of these note lines, see
1399 the compiler source file lib-writ.ads. This pragma is intended for use by
1400 external tools, including ASIS@. The use of pragma Annotate does not
1401 affect the compilation process in any way. This pragma may be used as
1402 a configuration pragma.
1405 @unnumberedsec Pragma Assert
1409 @smallexample @c ada
1412 [, string_EXPRESSION]);
1416 The effect of this pragma depends on whether the corresponding command
1417 line switch is set to activate assertions. The pragma expands into code
1418 equivalent to the following:
1420 @smallexample @c ada
1421 if assertions-enabled then
1422 if not boolean_EXPRESSION then
1423 System.Assertions.Raise_Assert_Failure
1424 (string_EXPRESSION);
1430 The string argument, if given, is the message that will be associated
1431 with the exception occurrence if the exception is raised. If no second
1432 argument is given, the default message is @samp{@var{file}:@var{nnn}},
1433 where @var{file} is the name of the source file containing the assert,
1434 and @var{nnn} is the line number of the assert. A pragma is not a
1435 statement, so if a statement sequence contains nothing but a pragma
1436 assert, then a null statement is required in addition, as in:
1438 @smallexample @c ada
1441 pragma Assert (K > 3, "Bad value for K");
1447 Note that, as with the @code{if} statement to which it is equivalent, the
1448 type of the expression is either @code{Standard.Boolean}, or any type derived
1449 from this standard type.
1451 Assert checks can be either checked or ignored. By default they are ignored.
1452 They will be checked if either the command line switch @option{-gnata} is
1453 used, or if an @code{Assertion_Policy} or @code{Check_Policy} pragma is used
1454 to enable @code{Assert_Checks}.
1456 If assertions are ignored, then there
1457 is no run-time effect (and in particular, any side effects from the
1458 expression will not occur at run time). (The expression is still
1459 analyzed at compile time, and may cause types to be frozen if they are
1460 mentioned here for the first time).
1462 If assertions are checked, then the given expression is tested, and if
1463 it is @code{False} then @code{System.Assertions.Raise_Assert_Failure} is called
1464 which results in the raising of @code{Assert_Failure} with the given message.
1466 You should generally avoid side effects in the expression arguments of
1467 this pragma, because these side effects will turn on and off with the
1468 setting of the assertions mode, resulting in assertions that have an
1469 effect on the program. However, the expressions are analyzed for
1470 semantic correctness whether or not assertions are enabled, so turning
1471 assertions on and off cannot affect the legality of a program.
1473 Note that the implementation defined policy @code{DISABLE}, given in a
1474 pragma @code{Assertion_Policy}, can be used to suppress this semantic analysis.
1476 Note: this is a standard language-defined pragma in versions
1477 of Ada from 2005 on. In GNAT, it is implemented in all versions
1478 of Ada, and the DISABLE policy is an implementation-defined
1481 @node Pragma Assert_And_Cut
1482 @unnumberedsec Pragma Assert_And_Cut
1483 @findex Assert_And_Cut
1486 @smallexample @c ada
1487 pragma Assert_And_Cut (
1489 [, string_EXPRESSION]);
1493 The effect of this pragma is identical to that of pragma @code{Assert},
1494 except that in an @code{Assertion_Policy} pragma, the identifier
1495 @code{Assert_And_Cut} is used to control whether it is ignored or checked
1498 The intention is that this be used within a subprogram when the
1499 given test expresion sums up all the work done so far in the
1500 subprogram, so that the rest of the subprogram can be verified
1501 (informally or formally) using only the entry preconditions,
1502 and the expression in this pragma. This allows dividing up
1503 a subprogram into sections for the purposes of testing or
1504 formal verification. The pragma also serves as useful
1507 @node Pragma Assertion_Policy
1508 @unnumberedsec Pragma Assertion_Policy
1509 @findex Assertion_Policy
1512 @smallexample @c ada
1513 pragma Assertion_Policy (CHECK | DISABLE | IGNORE);
1515 pragma Assertion_Policy (
1516 ASSERTION_KIND => POLICY_IDENTIFIER
1517 @{, ASSERTION_KIND => POLICY_IDENTIFIER@});
1519 ASSERTION_KIND ::= RM_ASSERTION_KIND | ID_ASSERTION_KIND
1521 RM_ASSERTION_KIND ::= Assert |
1529 Type_Invariant'Class
1531 ID_ASSERTION_KIND ::= Assertions |
1544 Statement_Assertions
1546 POLICY_IDENTIFIER ::= Check | Disable | Ignore
1550 This is a standard Ada 2012 pragma that is available as an
1551 implementation-defined pragma in earlier versions of Ada.
1552 The assertion kinds @code{RM_ASSERTION_KIND} are those defined in
1553 the Ada standard. The assertion kinds @code{ID_ASSERTION_KIND}
1554 are implementation defined additions recognized by the GNAT compiler.
1556 The pragma applies in both cases to pragmas and aspects with matching
1557 names, e.g. @code{Pre} applies to the Pre aspect, and @code{Precondition}
1558 applies to both the @code{Precondition} pragma
1559 and the aspect @code{Precondition}. Note that the identifiers for
1560 pragmas Pre_Class and Post_Class are Pre'Class and Post'Class (not
1561 Pre_Class and Post_Class), since these pragmas are intended to be
1562 identical to the corresponding aspects).
1564 If the policy is @code{CHECK}, then assertions are enabled, i.e.
1565 the corresponding pragma or aspect is activated.
1566 If the policy is @code{IGNORE}, then assertions are ignored, i.e.
1567 the corresponding pragma or aspect is deactivated.
1568 This pragma overrides the effect of the @option{-gnata} switch on the
1571 The implementation defined policy @code{DISABLE} is like
1572 @code{IGNORE} except that it completely disables semantic
1573 checking of the corresponding pragma or aspect. This is
1574 useful when the pragma or aspect argument references subprograms
1575 in a with'ed package which is replaced by a dummy package
1576 for the final build.
1578 The implementation defined assertion kind @code{Assertions} applies to all
1579 assertion kinds. The form with no assertion kind given implies this
1580 choice, so it applies to all assertion kinds (RM defined, and
1581 implementation defined).
1583 The implementation defined assertion kind @code{Statement_Assertions}
1584 applies to @code{Assert}, @code{Assert_And_Cut},
1585 @code{Assume}, @code{Loop_Invariant}, and @code{Loop_Variant}.
1588 @unnumberedsec Pragma Assume
1592 @smallexample @c ada
1595 [, string_EXPRESSION]);
1599 The effect of this pragma is identical to that of pragma @code{Assert},
1600 except that in an @code{Assertion_Policy} pragma, the identifier
1601 @code{Assume} is used to control whether it is ignored or checked
1604 The intention is that this be used for assumptions about the
1605 external environment. So you cannot expect to verify formally
1606 or informally that the condition is met, this must be
1607 established by examining things outside the program itself.
1608 For example, we may have code that depends on the size of
1609 @code{Long_Long_Integer} being at least 64. So we could write:
1611 @smallexample @c ada
1612 pragma Assume (Long_Long_Integer'Size >= 64);
1616 This assumption cannot be proved from the program itself,
1617 but it acts as a useful run-time check that the assumption
1618 is met, and documents the need to ensure that it is met by
1619 reference to information outside the program.
1621 @node Pragma Assume_No_Invalid_Values
1622 @unnumberedsec Pragma Assume_No_Invalid_Values
1623 @findex Assume_No_Invalid_Values
1624 @cindex Invalid representations
1625 @cindex Invalid values
1628 @smallexample @c ada
1629 pragma Assume_No_Invalid_Values (On | Off);
1633 This is a configuration pragma that controls the assumptions made by the
1634 compiler about the occurrence of invalid representations (invalid values)
1637 The default behavior (corresponding to an Off argument for this pragma), is
1638 to assume that values may in general be invalid unless the compiler can
1639 prove they are valid. Consider the following example:
1641 @smallexample @c ada
1642 V1 : Integer range 1 .. 10;
1643 V2 : Integer range 11 .. 20;
1645 for J in V2 .. V1 loop
1651 if V1 and V2 have valid values, then the loop is known at compile
1652 time not to execute since the lower bound must be greater than the
1653 upper bound. However in default mode, no such assumption is made,
1654 and the loop may execute. If @code{Assume_No_Invalid_Values (On)}
1655 is given, the compiler will assume that any occurrence of a variable
1656 other than in an explicit @code{'Valid} test always has a valid
1657 value, and the loop above will be optimized away.
1659 The use of @code{Assume_No_Invalid_Values (On)} is appropriate if
1660 you know your code is free of uninitialized variables and other
1661 possible sources of invalid representations, and may result in
1662 more efficient code. A program that accesses an invalid representation
1663 with this pragma in effect is erroneous, so no guarantees can be made
1666 It is peculiar though permissible to use this pragma in conjunction
1667 with validity checking (-gnatVa). In such cases, accessing invalid
1668 values will generally give an exception, though formally the program
1669 is erroneous so there are no guarantees that this will always be the
1670 case, and it is recommended that these two options not be used together.
1672 @node Pragma Async_Readers
1673 @unnumberedsec Pragma Async_Readers
1674 @findex Async_Readers
1676 For the description of this pragma, see SPARK 2014 Reference Manual,
1679 @node Pragma Async_Writers
1680 @unnumberedsec Pragma Async_Writers
1681 @findex Async_Writers
1683 For the description of this pragma, see SPARK 2014 Reference Manual,
1686 @node Pragma AST_Entry
1687 @unnumberedsec Pragma AST_Entry
1692 @smallexample @c ada
1693 pragma AST_Entry (entry_IDENTIFIER);
1697 This pragma is implemented only in the OpenVMS implementation of GNAT@. The
1698 argument is the simple name of a single entry; at most one @code{AST_Entry}
1699 pragma is allowed for any given entry. This pragma must be used in
1700 conjunction with the @code{AST_Entry} attribute, and is only allowed after
1701 the entry declaration and in the same task type specification or single task
1702 as the entry to which it applies. This pragma specifies that the given entry
1703 may be used to handle an OpenVMS asynchronous system trap (@code{AST})
1704 resulting from an OpenVMS system service call. The pragma does not affect
1705 normal use of the entry. For further details on this pragma, see the
1706 DEC Ada Language Reference Manual, section 9.12a.
1708 @node Pragma Attribute_Definition
1709 @unnumberedsec Pragma Attribute_Definition
1710 @findex Attribute_Definition
1713 @smallexample @c ada
1714 pragma Attribute_Definition
1715 ([Attribute =>] ATTRIBUTE_DESIGNATOR,
1716 [Entity =>] LOCAL_NAME,
1717 [Expression =>] EXPRESSION | NAME);
1721 If @code{Attribute} is a known attribute name, this pragma is equivalent to
1722 the attribute definition clause:
1724 @smallexample @c ada
1725 for Entity'Attribute use Expression;
1728 If @code{Attribute} is not a recognized attribute name, the pragma is
1729 ignored, and a warning is emitted. This allows source
1730 code to be written that takes advantage of some new attribute, while remaining
1731 compilable with earlier compilers.
1733 @node Pragma C_Pass_By_Copy
1734 @unnumberedsec Pragma C_Pass_By_Copy
1735 @cindex Passing by copy
1736 @findex C_Pass_By_Copy
1739 @smallexample @c ada
1740 pragma C_Pass_By_Copy
1741 ([Max_Size =>] static_integer_EXPRESSION);
1745 Normally the default mechanism for passing C convention records to C
1746 convention subprograms is to pass them by reference, as suggested by RM
1747 B.3(69). Use the configuration pragma @code{C_Pass_By_Copy} to change
1748 this default, by requiring that record formal parameters be passed by
1749 copy if all of the following conditions are met:
1753 The size of the record type does not exceed the value specified for
1756 The record type has @code{Convention C}.
1758 The formal parameter has this record type, and the subprogram has a
1759 foreign (non-Ada) convention.
1763 If these conditions are met the argument is passed by copy, i.e.@: in a
1764 manner consistent with what C expects if the corresponding formal in the
1765 C prototype is a struct (rather than a pointer to a struct).
1767 You can also pass records by copy by specifying the convention
1768 @code{C_Pass_By_Copy} for the record type, or by using the extended
1769 @code{Import} and @code{Export} pragmas, which allow specification of
1770 passing mechanisms on a parameter by parameter basis.
1773 @unnumberedsec Pragma Check
1775 @cindex Named assertions
1779 @smallexample @c ada
1781 [Name =>] CHECK_KIND,
1782 [Check =>] Boolean_EXPRESSION
1783 [, [Message =>] string_EXPRESSION] );
1785 CHECK_KIND ::= IDENTIFIER |
1788 Type_Invariant'Class |
1793 This pragma is similar to the predefined pragma @code{Assert} except that an
1794 extra identifier argument is present. In conjunction with pragma
1795 @code{Check_Policy}, this can be used to define groups of assertions that can
1796 be independently controlled. The identifier @code{Assertion} is special, it
1797 refers to the normal set of pragma @code{Assert} statements.
1799 Checks introduced by this pragma are normally deactivated by default. They can
1800 be activated either by the command line option @option{-gnata}, which turns on
1801 all checks, or individually controlled using pragma @code{Check_Policy}.
1803 The identifiers @code{Assertions} and @code{Statement_Assertions} are not
1804 permitted as check kinds, since this would cause confusion with the use
1805 of these identifiers in @code{Assertion_Policy} and @code{Check_Policy}
1806 pragmas, where they are used to refer to sets of assertions.
1808 @node Pragma Check_Float_Overflow
1809 @unnumberedsec Pragma Check_Float_Overflow
1810 @cindex Floating-point overflow
1811 @findex Check_Float_Overflow
1814 @smallexample @c ada
1815 pragma Check_Float_Overflow;
1819 In Ada, the predefined floating-point types (@code{Short_Float},
1820 @code{Float}, @code{Long_Float}, @code{Long_Long_Float}) are
1821 defined to be @emph{unconstrained}. This means that even though each
1822 has a well-defined base range, an operation that delivers a result
1823 outside this base range is not required to raise an exception.
1824 This implementation permission accommodates the notion
1825 of infinities in IEEE floating-point, and corresponds to the
1826 efficient execution mode on most machines. GNAT will not raise
1827 overflow exceptions on these machines; instead it will generate
1828 infinities and NaN's as defined in the IEEE standard.
1830 Generating infinities, although efficient, is not always desirable.
1831 Often the preferable approach is to check for overflow, even at the
1832 (perhaps considerable) expense of run-time performance.
1833 This can be accomplished by defining your own constrained floating-point subtypes -- i.e., by supplying explicit
1834 range constraints -- and indeed such a subtype
1835 can have the same base range as its base type. For example:
1837 @smallexample @c ada
1838 subtype My_Float is Float range Float'Range;
1842 Here @code{My_Float} has the same range as
1843 @code{Float} but is constrained, so operations on
1844 @code{My_Float} values will be checked for overflow
1847 This style will achieve the desired goal, but
1848 it is often more convenient to be able to simply use
1849 the standard predefined floating-point types as long
1850 as overflow checking could be guaranteed.
1851 The @code{Check_Float_Overflow}
1852 configuration pragma achieves this effect. If a unit is compiled
1853 subject to this configuration pragma, then all operations
1854 on predefined floating-point types including operations on
1855 base types of these floating-point types will be treated as
1856 though those types were constrained, and overflow checks
1857 will be generated. The @code{Constraint_Error}
1858 exception is raised if the result is out of range.
1860 This mode can also be set by use of the compiler
1861 switch @option{-gnateF}.
1863 @node Pragma Check_Name
1864 @unnumberedsec Pragma Check_Name
1865 @cindex Defining check names
1866 @cindex Check names, defining
1870 @smallexample @c ada
1871 pragma Check_Name (check_name_IDENTIFIER);
1875 This is a configuration pragma that defines a new implementation
1876 defined check name (unless IDENTIFIER matches one of the predefined
1877 check names, in which case the pragma has no effect). Check names
1878 are global to a partition, so if two or more configuration pragmas
1879 are present in a partition mentioning the same name, only one new
1880 check name is introduced.
1882 An implementation defined check name introduced with this pragma may
1883 be used in only three contexts: @code{pragma Suppress},
1884 @code{pragma Unsuppress},
1885 and as the prefix of a @code{Check_Name'Enabled} attribute reference. For
1886 any of these three cases, the check name must be visible. A check
1887 name is visible if it is in the configuration pragmas applying to
1888 the current unit, or if it appears at the start of any unit that
1889 is part of the dependency set of the current unit (e.g., units that
1890 are mentioned in @code{with} clauses).
1892 Check names introduced by this pragma are subject to control by compiler
1893 switches (in particular -gnatp) in the usual manner.
1895 @node Pragma Check_Policy
1896 @unnumberedsec Pragma Check_Policy
1897 @cindex Controlling assertions
1898 @cindex Assertions, control
1899 @cindex Check pragma control
1900 @cindex Named assertions
1904 @smallexample @c ada
1906 ([Name =>] CHECK_KIND,
1907 [Policy =>] POLICY_IDENTIFIER);
1909 pragma Check_Policy (
1910 CHECK_KIND => POLICY_IDENTIFIER
1911 @{, CHECK_KIND => POLICY_IDENTIFIER@});
1913 ASSERTION_KIND ::= RM_ASSERTION_KIND | ID_ASSERTION_KIND
1915 CHECK_KIND ::= IDENTIFIER |
1918 Type_Invariant'Class |
1921 The identifiers Name and Policy are not allowed as CHECK_KIND values. This
1922 avoids confusion between the two possible syntax forms for this pragma.
1924 POLICY_IDENTIFIER ::= ON | OFF | CHECK | DISABLE | IGNORE
1928 This pragma is used to set the checking policy for assertions (specified
1929 by aspects or pragmas), the @code{Debug} pragma, or additional checks
1930 to be checked using the @code{Check} pragma. It may appear either as
1931 a configuration pragma, or within a declarative part of package. In the
1932 latter case, it applies from the point where it appears to the end of
1933 the declarative region (like pragma @code{Suppress}).
1935 The @code{Check_Policy} pragma is similar to the
1936 predefined @code{Assertion_Policy} pragma,
1937 and if the check kind corresponds to one of the assertion kinds that
1938 are allowed by @code{Assertion_Policy}, then the effect is identical.
1940 If the first argument is Debug, then the policy applies to Debug pragmas,
1941 disabling their effect if the policy is @code{OFF}, @code{DISABLE}, or
1942 @code{IGNORE}, and allowing them to execute with normal semantics if
1943 the policy is @code{ON} or @code{CHECK}. In addition if the policy is
1944 @code{DISABLE}, then the procedure call in @code{Debug} pragmas will
1945 be totally ignored and not analyzed semantically.
1947 Finally the first argument may be some other identifier than the above
1948 possibilities, in which case it controls a set of named assertions
1949 that can be checked using pragma @code{Check}. For example, if the pragma:
1951 @smallexample @c ada
1952 pragma Check_Policy (Critical_Error, OFF);
1956 is given, then subsequent @code{Check} pragmas whose first argument is also
1957 @code{Critical_Error} will be disabled.
1959 The check policy is @code{OFF} to turn off corresponding checks, and @code{ON}
1960 to turn on corresponding checks. The default for a set of checks for which no
1961 @code{Check_Policy} is given is @code{OFF} unless the compiler switch
1962 @option{-gnata} is given, which turns on all checks by default.
1964 The check policy settings @code{CHECK} and @code{IGNORE} are recognized
1965 as synonyms for @code{ON} and @code{OFF}. These synonyms are provided for
1966 compatibility with the standard @code{Assertion_Policy} pragma. The check
1967 policy setting @code{DISABLE} causes the second argument of a corresponding
1968 @code{Check} pragma to be completely ignored and not analyzed.
1970 @node Pragma CIL_Constructor
1971 @unnumberedsec Pragma CIL_Constructor
1972 @findex CIL_Constructor
1976 @smallexample @c ada
1977 pragma CIL_Constructor ([Entity =>] function_LOCAL_NAME);
1981 This pragma is used to assert that the specified Ada function should be
1982 mapped to the .NET constructor for some Ada tagged record type.
1984 See section 4.1 of the
1985 @code{GNAT User's Guide: Supplement for the .NET Platform.}
1986 for related information.
1988 @node Pragma Comment
1989 @unnumberedsec Pragma Comment
1994 @smallexample @c ada
1995 pragma Comment (static_string_EXPRESSION);
1999 This is almost identical in effect to pragma @code{Ident}. It allows the
2000 placement of a comment into the object file and hence into the
2001 executable file if the operating system permits such usage. The
2002 difference is that @code{Comment}, unlike @code{Ident}, has
2003 no limitations on placement of the pragma (it can be placed
2004 anywhere in the main source unit), and if more than one pragma
2005 is used, all comments are retained.
2007 @node Pragma Common_Object
2008 @unnumberedsec Pragma Common_Object
2009 @findex Common_Object
2013 @smallexample @c ada
2014 pragma Common_Object (
2015 [Internal =>] LOCAL_NAME
2016 [, [External =>] EXTERNAL_SYMBOL]
2017 [, [Size =>] EXTERNAL_SYMBOL] );
2021 | static_string_EXPRESSION
2025 This pragma enables the shared use of variables stored in overlaid
2026 linker areas corresponding to the use of @code{COMMON}
2027 in Fortran. The single
2028 object @var{LOCAL_NAME} is assigned to the area designated by
2029 the @var{External} argument.
2030 You may define a record to correspond to a series
2031 of fields. The @var{Size} argument
2032 is syntax checked in GNAT, but otherwise ignored.
2034 @code{Common_Object} is not supported on all platforms. If no
2035 support is available, then the code generator will issue a message
2036 indicating that the necessary attribute for implementation of this
2037 pragma is not available.
2039 @node Pragma Compile_Time_Error
2040 @unnumberedsec Pragma Compile_Time_Error
2041 @findex Compile_Time_Error
2045 @smallexample @c ada
2046 pragma Compile_Time_Error
2047 (boolean_EXPRESSION, static_string_EXPRESSION);
2051 This pragma can be used to generate additional compile time
2053 is particularly useful in generics, where errors can be issued for
2054 specific problematic instantiations. The first parameter is a boolean
2055 expression. The pragma is effective only if the value of this expression
2056 is known at compile time, and has the value True. The set of expressions
2057 whose values are known at compile time includes all static boolean
2058 expressions, and also other values which the compiler can determine
2059 at compile time (e.g., the size of a record type set by an explicit
2060 size representation clause, or the value of a variable which was
2061 initialized to a constant and is known not to have been modified).
2062 If these conditions are met, an error message is generated using
2063 the value given as the second argument. This string value may contain
2064 embedded ASCII.LF characters to break the message into multiple lines.
2066 @node Pragma Compile_Time_Warning
2067 @unnumberedsec Pragma Compile_Time_Warning
2068 @findex Compile_Time_Warning
2072 @smallexample @c ada
2073 pragma Compile_Time_Warning
2074 (boolean_EXPRESSION, static_string_EXPRESSION);
2078 Same as pragma Compile_Time_Error, except a warning is issued instead
2079 of an error message. Note that if this pragma is used in a package that
2080 is with'ed by a client, the client will get the warning even though it
2081 is issued by a with'ed package (normally warnings in with'ed units are
2082 suppressed, but this is a special exception to that rule).
2084 One typical use is within a generic where compile time known characteristics
2085 of formal parameters are tested, and warnings given appropriately. Another use
2086 with a first parameter of True is to warn a client about use of a package,
2087 for example that it is not fully implemented.
2089 @node Pragma Compiler_Unit
2090 @unnumberedsec Pragma Compiler_Unit
2091 @findex Compiler_Unit
2095 @smallexample @c ada
2096 pragma Compiler_Unit;
2100 This pragma is obsolete. It is equivalent to Compiler_Unit_Warning. It is
2101 retained so that old versions of the GNAT run-time that use this pragma can
2102 be compiled with newer versions of the compiler.
2104 @node Pragma Compiler_Unit_Warning
2105 @unnumberedsec Pragma Compiler_Unit_Warning
2106 @findex Compiler_Unit_Warning
2110 @smallexample @c ada
2111 pragma Compiler_Unit_Warning;
2115 This pragma is intended only for internal use in the GNAT run-time library.
2116 It indicates that the unit is used as part of the compiler build. The effect
2117 is to generate warnings for the use of constructs (for example, conditional
2118 expressions) that would cause trouble when bootstrapping using an older
2119 version of GNAT. For the exact list of restrictions, see the compiler sources
2120 and references to Check_Compiler_Unit.
2122 @node Pragma Complete_Representation
2123 @unnumberedsec Pragma Complete_Representation
2124 @findex Complete_Representation
2128 @smallexample @c ada
2129 pragma Complete_Representation;
2133 This pragma must appear immediately within a record representation
2134 clause. Typical placements are before the first component clause
2135 or after the last component clause. The effect is to give an error
2136 message if any component is missing a component clause. This pragma
2137 may be used to ensure that a record representation clause is
2138 complete, and that this invariant is maintained if fields are
2139 added to the record in the future.
2141 @node Pragma Complex_Representation
2142 @unnumberedsec Pragma Complex_Representation
2143 @findex Complex_Representation
2147 @smallexample @c ada
2148 pragma Complex_Representation
2149 ([Entity =>] LOCAL_NAME);
2153 The @var{Entity} argument must be the name of a record type which has
2154 two fields of the same floating-point type. The effect of this pragma is
2155 to force gcc to use the special internal complex representation form for
2156 this record, which may be more efficient. Note that this may result in
2157 the code for this type not conforming to standard ABI (application
2158 binary interface) requirements for the handling of record types. For
2159 example, in some environments, there is a requirement for passing
2160 records by pointer, and the use of this pragma may result in passing
2161 this type in floating-point registers.
2163 @node Pragma Component_Alignment
2164 @unnumberedsec Pragma Component_Alignment
2165 @cindex Alignments of components
2166 @findex Component_Alignment
2170 @smallexample @c ada
2171 pragma Component_Alignment (
2172 [Form =>] ALIGNMENT_CHOICE
2173 [, [Name =>] type_LOCAL_NAME]);
2175 ALIGNMENT_CHOICE ::=
2183 Specifies the alignment of components in array or record types.
2184 The meaning of the @var{Form} argument is as follows:
2187 @findex Component_Size
2188 @item Component_Size
2189 Aligns scalar components and subcomponents of the array or record type
2190 on boundaries appropriate to their inherent size (naturally
2191 aligned). For example, 1-byte components are aligned on byte boundaries,
2192 2-byte integer components are aligned on 2-byte boundaries, 4-byte
2193 integer components are aligned on 4-byte boundaries and so on. These
2194 alignment rules correspond to the normal rules for C compilers on all
2195 machines except the VAX@.
2197 @findex Component_Size_4
2198 @item Component_Size_4
2199 Naturally aligns components with a size of four or fewer
2200 bytes. Components that are larger than 4 bytes are placed on the next
2203 @findex Storage_Unit
2205 Specifies that array or record components are byte aligned, i.e.@:
2206 aligned on boundaries determined by the value of the constant
2207 @code{System.Storage_Unit}.
2211 Specifies that array or record components are aligned on default
2212 boundaries, appropriate to the underlying hardware or operating system or
2213 both. For OpenVMS VAX systems, the @code{Default} choice is the same as
2214 the @code{Storage_Unit} choice (byte alignment). For all other systems,
2215 the @code{Default} choice is the same as @code{Component_Size} (natural
2220 If the @code{Name} parameter is present, @var{type_LOCAL_NAME} must
2221 refer to a local record or array type, and the specified alignment
2222 choice applies to the specified type. The use of
2223 @code{Component_Alignment} together with a pragma @code{Pack} causes the
2224 @code{Component_Alignment} pragma to be ignored. The use of
2225 @code{Component_Alignment} together with a record representation clause
2226 is only effective for fields not specified by the representation clause.
2228 If the @code{Name} parameter is absent, the pragma can be used as either
2229 a configuration pragma, in which case it applies to one or more units in
2230 accordance with the normal rules for configuration pragmas, or it can be
2231 used within a declarative part, in which case it applies to types that
2232 are declared within this declarative part, or within any nested scope
2233 within this declarative part. In either case it specifies the alignment
2234 to be applied to any record or array type which has otherwise standard
2237 If the alignment for a record or array type is not specified (using
2238 pragma @code{Pack}, pragma @code{Component_Alignment}, or a record rep
2239 clause), the GNAT uses the default alignment as described previously.
2241 @node Pragma Contract_Cases
2242 @unnumberedsec Pragma Contract_Cases
2243 @cindex Contract cases
2244 @findex Contract_Cases
2248 @smallexample @c ada
2249 pragma Contract_Cases (
2250 Condition => Consequence
2251 @{,Condition => Consequence@});
2255 The @code{Contract_Cases} pragma allows defining fine-grain specifications
2256 that can complement or replace the contract given by a precondition and a
2257 postcondition. Additionally, the @code{Contract_Cases} pragma can be used
2258 by testing and formal verification tools. The compiler checks its validity and,
2259 depending on the assertion policy at the point of declaration of the pragma,
2260 it may insert a check in the executable. For code generation, the contract
2263 @smallexample @c ada
2264 pragma Contract_Cases (
2272 @smallexample @c ada
2273 C1 : constant Boolean := Cond1; -- evaluated at subprogram entry
2274 C2 : constant Boolean := Cond2; -- evaluated at subprogram entry
2275 pragma Precondition ((C1 and not C2) or (C2 and not C1));
2276 pragma Postcondition (if C1 then Pred1);
2277 pragma Postcondition (if C2 then Pred2);
2281 The precondition ensures that one and only one of the conditions is
2282 satisfied on entry to the subprogram.
2283 The postcondition ensures that for the condition that was True on entry,
2284 the corrresponding consequence is True on exit. Other consequence expressions
2287 A precondition @code{P} and postcondition @code{Q} can also be
2288 expressed as contract cases:
2290 @smallexample @c ada
2291 pragma Contract_Cases (P => Q);
2294 The placement and visibility rules for @code{Contract_Cases} pragmas are
2295 identical to those described for preconditions and postconditions.
2297 The compiler checks that boolean expressions given in conditions and
2298 consequences are valid, where the rules for conditions are the same as
2299 the rule for an expression in @code{Precondition} and the rules for
2300 consequences are the same as the rule for an expression in
2301 @code{Postcondition}. In particular, attributes @code{'Old} and
2302 @code{'Result} can only be used within consequence expressions.
2303 The condition for the last contract case may be @code{others}, to denote
2304 any case not captured by the previous cases. The
2305 following is an example of use within a package spec:
2307 @smallexample @c ada
2308 package Math_Functions is
2310 function Sqrt (Arg : Float) return Float;
2311 pragma Contract_Cases ((Arg in 0 .. 99) => Sqrt'Result < 10,
2312 Arg >= 100 => Sqrt'Result >= 10,
2313 others => Sqrt'Result = 0);
2319 The meaning of contract cases is that only one case should apply at each
2320 call, as determined by the corresponding condition evaluating to True,
2321 and that the consequence for this case should hold when the subprogram
2324 @node Pragma Convention_Identifier
2325 @unnumberedsec Pragma Convention_Identifier
2326 @findex Convention_Identifier
2327 @cindex Conventions, synonyms
2331 @smallexample @c ada
2332 pragma Convention_Identifier (
2333 [Name =>] IDENTIFIER,
2334 [Convention =>] convention_IDENTIFIER);
2338 This pragma provides a mechanism for supplying synonyms for existing
2339 convention identifiers. The @code{Name} identifier can subsequently
2340 be used as a synonym for the given convention in other pragmas (including
2341 for example pragma @code{Import} or another @code{Convention_Identifier}
2342 pragma). As an example of the use of this, suppose you had legacy code
2343 which used Fortran77 as the identifier for Fortran. Then the pragma:
2345 @smallexample @c ada
2346 pragma Convention_Identifier (Fortran77, Fortran);
2350 would allow the use of the convention identifier @code{Fortran77} in
2351 subsequent code, avoiding the need to modify the sources. As another
2352 example, you could use this to parameterize convention requirements
2353 according to systems. Suppose you needed to use @code{Stdcall} on
2354 windows systems, and @code{C} on some other system, then you could
2355 define a convention identifier @code{Library} and use a single
2356 @code{Convention_Identifier} pragma to specify which convention
2357 would be used system-wide.
2359 @node Pragma CPP_Class
2360 @unnumberedsec Pragma CPP_Class
2362 @cindex Interfacing with C++
2366 @smallexample @c ada
2367 pragma CPP_Class ([Entity =>] LOCAL_NAME);
2371 The argument denotes an entity in the current declarative region that is
2372 declared as a record type. It indicates that the type corresponds to an
2373 externally declared C++ class type, and is to be laid out the same way
2374 that C++ would lay out the type. If the C++ class has virtual primitives
2375 then the record must be declared as a tagged record type.
2377 Types for which @code{CPP_Class} is specified do not have assignment or
2378 equality operators defined (such operations can be imported or declared
2379 as subprograms as required). Initialization is allowed only by constructor
2380 functions (see pragma @code{CPP_Constructor}). Such types are implicitly
2381 limited if not explicitly declared as limited or derived from a limited
2382 type, and an error is issued in that case.
2384 See @ref{Interfacing to C++} for related information.
2386 Note: Pragma @code{CPP_Class} is currently obsolete. It is supported
2387 for backward compatibility but its functionality is available
2388 using pragma @code{Import} with @code{Convention} = @code{CPP}.
2390 @node Pragma CPP_Constructor
2391 @unnumberedsec Pragma CPP_Constructor
2392 @cindex Interfacing with C++
2393 @findex CPP_Constructor
2397 @smallexample @c ada
2398 pragma CPP_Constructor ([Entity =>] LOCAL_NAME
2399 [, [External_Name =>] static_string_EXPRESSION ]
2400 [, [Link_Name =>] static_string_EXPRESSION ]);
2404 This pragma identifies an imported function (imported in the usual way
2405 with pragma @code{Import}) as corresponding to a C++ constructor. If
2406 @code{External_Name} and @code{Link_Name} are not specified then the
2407 @code{Entity} argument is a name that must have been previously mentioned
2408 in a pragma @code{Import} with @code{Convention} = @code{CPP}. Such name
2409 must be of one of the following forms:
2413 @code{function @var{Fname} return @var{T}}
2417 @code{function @var{Fname} return @var{T}'Class}
2420 @code{function @var{Fname} (@dots{}) return @var{T}}
2424 @code{function @var{Fname} (@dots{}) return @var{T}'Class}
2428 where @var{T} is a limited record type imported from C++ with pragma
2429 @code{Import} and @code{Convention} = @code{CPP}.
2431 The first two forms import the default constructor, used when an object
2432 of type @var{T} is created on the Ada side with no explicit constructor.
2433 The latter two forms cover all the non-default constructors of the type.
2434 See the @value{EDITION} User's Guide for details.
2436 If no constructors are imported, it is impossible to create any objects
2437 on the Ada side and the type is implicitly declared abstract.
2439 Pragma @code{CPP_Constructor} is intended primarily for automatic generation
2440 using an automatic binding generator tool (such as the @code{-fdump-ada-spec}
2442 See @ref{Interfacing to C++} for more related information.
2444 Note: The use of functions returning class-wide types for constructors is
2445 currently obsolete. They are supported for backward compatibility. The
2446 use of functions returning the type T leave the Ada sources more clear
2447 because the imported C++ constructors always return an object of type T;
2448 that is, they never return an object whose type is a descendant of type T.
2450 @node Pragma CPP_Virtual
2451 @unnumberedsec Pragma CPP_Virtual
2452 @cindex Interfacing to C++
2455 This pragma is now obsolete and, other than generating a warning if warnings
2456 on obsolescent features are enabled, is completely ignored.
2457 It is retained for compatibility
2458 purposes. It used to be required to ensure compoatibility with C++, but
2459 is no longer required for that purpose because GNAT generates
2460 the same object layout as the G++ compiler by default.
2462 See @ref{Interfacing to C++} for related information.
2464 @node Pragma CPP_Vtable
2465 @unnumberedsec Pragma CPP_Vtable
2466 @cindex Interfacing with C++
2469 This pragma is now obsolete and, other than generating a warning if warnings
2470 on obsolescent features are enabled, is completely ignored.
2471 It used to be required to ensure compatibility with C++, but
2472 is no longer required for that purpose because GNAT generates
2473 the same object layout as the G++ compiler by default.
2475 See @ref{Interfacing to C++} for related information.
2478 @unnumberedsec Pragma CPU
2483 @smallexample @c ada
2484 pragma CPU (EXPRESSION);
2488 This pragma is standard in Ada 2012, but is available in all earlier
2489 versions of Ada as an implementation-defined pragma.
2490 See Ada 2012 Reference Manual for details.
2493 @unnumberedsec Pragma Debug
2498 @smallexample @c ada
2499 pragma Debug ([CONDITION, ]PROCEDURE_CALL_WITHOUT_SEMICOLON);
2501 PROCEDURE_CALL_WITHOUT_SEMICOLON ::=
2503 | PROCEDURE_PREFIX ACTUAL_PARAMETER_PART
2507 The procedure call argument has the syntactic form of an expression, meeting
2508 the syntactic requirements for pragmas.
2510 If debug pragmas are not enabled or if the condition is present and evaluates
2511 to False, this pragma has no effect. If debug pragmas are enabled, the
2512 semantics of the pragma is exactly equivalent to the procedure call statement
2513 corresponding to the argument with a terminating semicolon. Pragmas are
2514 permitted in sequences of declarations, so you can use pragma @code{Debug} to
2515 intersperse calls to debug procedures in the middle of declarations. Debug
2516 pragmas can be enabled either by use of the command line switch @option{-gnata}
2517 or by use of the pragma @code{Check_Policy} with a first argument of
2520 @node Pragma Debug_Policy
2521 @unnumberedsec Pragma Debug_Policy
2522 @findex Debug_Policy
2526 @smallexample @c ada
2527 pragma Debug_Policy (CHECK | DISABLE | IGNORE | ON | OFF);
2531 This pragma is equivalent to a corresponding @code{Check_Policy} pragma
2532 with a first argument of @code{Debug}. It is retained for historical
2533 compatibility reasons.
2535 @node Pragma Default_Scalar_Storage_Order
2536 @unnumberedsec Pragma Default_Scalar_Storage_Order
2537 @cindex Default_Scalar_Storage_Order
2538 @cindex Scalar_Storage_Order
2539 @findex Default_Scalar_Storage_Order
2543 @smallexample @c ada
2544 pragma Default_Scalar_Storage_Order (High_Order_First | Low_Order_First);
2548 Normally if no explicit @code{Scalar_Storage_Order} is given for a record
2549 type or array type, then the scalar storage order defaults to the ordinary
2550 default for the target. But this default may be overridden using this pragma.
2551 The pragma may appear as a configuration pragma, or locally within a package
2552 spec or declarative part. In the latter case, it applies to all subsequent
2553 types declared within that package spec or declarative part.
2555 If this pragma is used as a configuration pragma which appears within a
2556 configuration pragma file (as opposed to appearing explicitly at the start
2557 of a single unit), then the binder will require that all units in a partition
2558 be compiled in a similar manner, including all units in the run-time that
2559 are included in the partition.
2561 The following example shows the use of this pragma:
2563 @smallexample @c ada
2564 pragma Default_Scalar_Storage_Order (High_Order_First);
2565 with System; use System;
2574 for L2'Scalar_Storage_Order use Low_Order_First;
2583 pragma Default_Scalar_Storage_Order (Low_Order_First);
2590 type H4a is new Inner.L4;
2599 In this example record types L.. have @code{Low_Order_First} scalar
2600 storage order, and record types H.. have @code{High_Order_First}.
2601 Note that in the case of @code{H4a}, the order is not inherited
2602 from the parent type. Only an explicitly set @code{Scalar_Storage_Order}
2603 gets inherited on type derivation.
2605 @node Pragma Default_Storage_Pool
2606 @unnumberedsec Pragma Default_Storage_Pool
2607 @cindex Default_Storage_Pool
2608 @findex Default_Storage_Pool
2612 @smallexample @c ada
2613 pragma Default_Storage_Pool (storage_pool_NAME | null);
2617 This pragma is standard in Ada 2012, but is available in all earlier
2618 versions of Ada as an implementation-defined pragma.
2619 See Ada 2012 Reference Manual for details.
2621 @node Pragma Depends
2622 @unnumberedsec Pragma Depends
2625 For the description of this pragma, see SPARK 2014 Reference Manual,
2628 @node Pragma Detect_Blocking
2629 @unnumberedsec Pragma Detect_Blocking
2630 @findex Detect_Blocking
2634 @smallexample @c ada
2635 pragma Detect_Blocking;
2639 This is a standard pragma in Ada 2005, that is available in all earlier
2640 versions of Ada as an implementation-defined pragma.
2642 This is a configuration pragma that forces the detection of potentially
2643 blocking operations within a protected operation, and to raise Program_Error
2646 @node Pragma Disable_Atomic_Synchronization
2647 @unnumberedsec Pragma Disable_Atomic_Synchronization
2648 @cindex Atomic Synchronization
2649 @findex Disable_Atomic_Synchronization
2653 @smallexample @c ada
2654 pragma Disable_Atomic_Synchronization [(Entity)];
2658 Ada requires that accesses (reads or writes) of an atomic variable be
2659 regarded as synchronization points in the case of multiple tasks.
2660 Particularly in the case of multi-processors this may require special
2661 handling, e.g. the generation of memory barriers. This capability may
2662 be turned off using this pragma in cases where it is known not to be
2665 The placement and scope rules for this pragma are the same as those
2666 for @code{pragma Suppress}. In particular it can be used as a
2667 configuration pragma, or in a declaration sequence where it applies
2668 till the end of the scope. If an @code{Entity} argument is present,
2669 the action applies only to that entity.
2671 @node Pragma Dispatching_Domain
2672 @unnumberedsec Pragma Dispatching_Domain
2673 @findex Dispatching_Domain
2677 @smallexample @c ada
2678 pragma Dispatching_Domain (EXPRESSION);
2682 This pragma is standard in Ada 2012, but is available in all earlier
2683 versions of Ada as an implementation-defined pragma.
2684 See Ada 2012 Reference Manual for details.
2686 @node Pragma Effective_Reads
2687 @unnumberedsec Pragma Effective_Reads
2688 @findex Effective_Reads
2690 For the description of this pragma, see SPARK 2014 Reference Manual,
2693 @node Pragma Effective_Writes
2694 @unnumberedsec Pragma Effective_Writes
2695 @findex Effective_Writes
2697 For the description of this pragma, see SPARK 2014 Reference Manual,
2700 @node Pragma Elaboration_Checks
2701 @unnumberedsec Pragma Elaboration_Checks
2702 @cindex Elaboration control
2703 @findex Elaboration_Checks
2707 @smallexample @c ada
2708 pragma Elaboration_Checks (Dynamic | Static);
2712 This is a configuration pragma that provides control over the
2713 elaboration model used by the compilation affected by the
2714 pragma. If the parameter is @code{Dynamic},
2715 then the dynamic elaboration
2716 model described in the Ada Reference Manual is used, as though
2717 the @option{-gnatE} switch had been specified on the command
2718 line. If the parameter is @code{Static}, then the default GNAT static
2719 model is used. This configuration pragma overrides the setting
2720 of the command line. For full details on the elaboration models
2721 used by the GNAT compiler, see @ref{Elaboration Order Handling in GNAT,,,
2722 gnat_ugn, @value{EDITION} User's Guide}.
2724 @node Pragma Eliminate
2725 @unnumberedsec Pragma Eliminate
2726 @cindex Elimination of unused subprograms
2731 @smallexample @c ada
2732 pragma Eliminate ([Entity =>] DEFINING_DESIGNATOR,
2733 [Source_Location =>] STRING_LITERAL);
2737 The string literal given for the source location is a string which
2738 specifies the line number of the occurrence of the entity, using
2739 the syntax for SOURCE_TRACE given below:
2741 @smallexample @c ada
2742 SOURCE_TRACE ::= SOURCE_REFERENCE [LBRACKET SOURCE_TRACE RBRACKET]
2747 SOURCE_REFERENCE ::= FILE_NAME : LINE_NUMBER
2749 LINE_NUMBER ::= DIGIT @{DIGIT@}
2753 Spaces around the colon in a @code{Source_Reference} are optional.
2755 The @code{DEFINING_DESIGNATOR} matches the defining designator used in an
2756 explicit subprogram declaration, where the @code{entity} name in this
2757 designator appears on the source line specified by the source location.
2759 The source trace that is given as the @code{Source_Location} shall obey the
2760 following rules. The @code{FILE_NAME} is the short name (with no directory
2761 information) of an Ada source file, given using exactly the required syntax
2762 for the underlying file system (e.g. case is important if the underlying
2763 operating system is case sensitive). @code{LINE_NUMBER} gives the line
2764 number of the occurrence of the @code{entity}
2765 as a decimal literal without an exponent or point. If an @code{entity} is not
2766 declared in a generic instantiation (this includes generic subprogram
2767 instances), the source trace includes only one source reference. If an entity
2768 is declared inside a generic instantiation, its source trace (when parsing
2769 from left to right) starts with the source location of the declaration of the
2770 entity in the generic unit and ends with the source location of the
2771 instantiation (it is given in square brackets). This approach is recursively
2772 used in case of nested instantiations: the rightmost (nested most deeply in
2773 square brackets) element of the source trace is the location of the outermost
2774 instantiation, the next to left element is the location of the next (first
2775 nested) instantiation in the code of the corresponding generic unit, and so
2776 on, and the leftmost element (that is out of any square brackets) is the
2777 location of the declaration of the entity to eliminate in a generic unit.
2779 Note that the @code{Source_Location} argument specifies which of a set of
2780 similarly named entities is being eliminated, dealing both with overloading,
2781 and also appearance of the same entity name in different scopes.
2783 This pragma indicates that the given entity is not used in the program to be
2784 compiled and built. The effect of the pragma is to allow the compiler to
2785 eliminate the code or data associated with the named entity. Any reference to
2786 an eliminated entity causes a compile-time or link-time error.
2788 The intention of pragma @code{Eliminate} is to allow a program to be compiled
2789 in a system-independent manner, with unused entities eliminated, without
2790 needing to modify the source text. Normally the required set of
2791 @code{Eliminate} pragmas is constructed automatically using the gnatelim tool.
2793 Any source file change that removes, splits, or
2794 adds lines may make the set of Eliminate pragmas invalid because their
2795 @code{Source_Location} argument values may get out of date.
2797 Pragma @code{Eliminate} may be used where the referenced entity is a dispatching
2798 operation. In this case all the subprograms to which the given operation can
2799 dispatch are considered to be unused (are never called as a result of a direct
2800 or a dispatching call).
2802 @node Pragma Enable_Atomic_Synchronization
2803 @unnumberedsec Pragma Enable_Atomic_Synchronization
2804 @cindex Atomic Synchronization
2805 @findex Enable_Atomic_Synchronization
2809 @smallexample @c ada
2810 pragma Enable_Atomic_Synchronization [(Entity)];
2814 Ada requires that accesses (reads or writes) of an atomic variable be
2815 regarded as synchronization points in the case of multiple tasks.
2816 Particularly in the case of multi-processors this may require special
2817 handling, e.g. the generation of memory barriers. This synchronization
2818 is performed by default, but can be turned off using
2819 @code{pragma Disable_Atomic_Synchronization}. The
2820 @code{Enable_Atomic_Synchronization} pragma can be used to turn
2823 The placement and scope rules for this pragma are the same as those
2824 for @code{pragma Unsuppress}. In particular it can be used as a
2825 configuration pragma, or in a declaration sequence where it applies
2826 till the end of the scope. If an @code{Entity} argument is present,
2827 the action applies only to that entity.
2829 @node Pragma Export_Exception
2830 @unnumberedsec Pragma Export_Exception
2832 @findex Export_Exception
2836 @smallexample @c ada
2837 pragma Export_Exception (
2838 [Internal =>] LOCAL_NAME
2839 [, [External =>] EXTERNAL_SYMBOL]
2840 [, [Form =>] Ada | VMS]
2841 [, [Code =>] static_integer_EXPRESSION]);
2845 | static_string_EXPRESSION
2849 This pragma is implemented only in the OpenVMS implementation of GNAT@. It
2850 causes the specified exception to be propagated outside of the Ada program,
2851 so that it can be handled by programs written in other OpenVMS languages.
2852 This pragma establishes an external name for an Ada exception and makes the
2853 name available to the OpenVMS Linker as a global symbol. For further details
2854 on this pragma, see the
2855 DEC Ada Language Reference Manual, section 13.9a3.2.
2857 @node Pragma Export_Function
2858 @unnumberedsec Pragma Export_Function
2859 @cindex Argument passing mechanisms
2860 @findex Export_Function
2865 @smallexample @c ada
2866 pragma Export_Function (
2867 [Internal =>] LOCAL_NAME
2868 [, [External =>] EXTERNAL_SYMBOL]
2869 [, [Parameter_Types =>] PARAMETER_TYPES]
2870 [, [Result_Type =>] result_SUBTYPE_MARK]
2871 [, [Mechanism =>] MECHANISM]
2872 [, [Result_Mechanism =>] MECHANISM_NAME]);
2876 | static_string_EXPRESSION
2881 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
2885 | subtype_Name ' Access
2889 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
2891 MECHANISM_ASSOCIATION ::=
2892 [formal_parameter_NAME =>] MECHANISM_NAME
2894 MECHANISM_NAME ::= Value | Reference
2898 Use this pragma to make a function externally callable and optionally
2899 provide information on mechanisms to be used for passing parameter and
2900 result values. We recommend, for the purposes of improving portability,
2901 this pragma always be used in conjunction with a separate pragma
2902 @code{Export}, which must precede the pragma @code{Export_Function}.
2903 GNAT does not require a separate pragma @code{Export}, but if none is
2904 present, @code{Convention Ada} is assumed, which is usually
2905 not what is wanted, so it is usually appropriate to use this
2906 pragma in conjunction with a @code{Export} or @code{Convention}
2907 pragma that specifies the desired foreign convention.
2908 Pragma @code{Export_Function}
2909 (and @code{Export}, if present) must appear in the same declarative
2910 region as the function to which they apply.
2912 @var{internal_name} must uniquely designate the function to which the
2913 pragma applies. If more than one function name exists of this name in
2914 the declarative part you must use the @code{Parameter_Types} and
2915 @code{Result_Type} parameters is mandatory to achieve the required
2916 unique designation. @var{subtype_mark}s in these parameters must
2917 exactly match the subtypes in the corresponding function specification,
2918 using positional notation to match parameters with subtype marks.
2919 The form with an @code{'Access} attribute can be used to match an
2920 anonymous access parameter.
2922 @cindex Suppressing external name
2923 Special treatment is given if the EXTERNAL is an explicit null
2924 string or a static string expressions that evaluates to the null
2925 string. In this case, no external name is generated. This form
2926 still allows the specification of parameter mechanisms.
2928 @node Pragma Export_Object
2929 @unnumberedsec Pragma Export_Object
2930 @findex Export_Object
2934 @smallexample @c ada
2935 pragma Export_Object
2936 [Internal =>] LOCAL_NAME
2937 [, [External =>] EXTERNAL_SYMBOL]
2938 [, [Size =>] EXTERNAL_SYMBOL]
2942 | static_string_EXPRESSION
2946 This pragma designates an object as exported, and apart from the
2947 extended rules for external symbols, is identical in effect to the use of
2948 the normal @code{Export} pragma applied to an object. You may use a
2949 separate Export pragma (and you probably should from the point of view
2950 of portability), but it is not required. @var{Size} is syntax checked,
2951 but otherwise ignored by GNAT@.
2953 @node Pragma Export_Procedure
2954 @unnumberedsec Pragma Export_Procedure
2955 @findex Export_Procedure
2959 @smallexample @c ada
2960 pragma Export_Procedure (
2961 [Internal =>] LOCAL_NAME
2962 [, [External =>] EXTERNAL_SYMBOL]
2963 [, [Parameter_Types =>] PARAMETER_TYPES]
2964 [, [Mechanism =>] MECHANISM]);
2968 | static_string_EXPRESSION
2973 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
2977 | subtype_Name ' Access
2981 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
2983 MECHANISM_ASSOCIATION ::=
2984 [formal_parameter_NAME =>] MECHANISM_NAME
2986 MECHANISM_NAME ::= Value | Reference
2990 This pragma is identical to @code{Export_Function} except that it
2991 applies to a procedure rather than a function and the parameters
2992 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
2993 GNAT does not require a separate pragma @code{Export}, but if none is
2994 present, @code{Convention Ada} is assumed, which is usually
2995 not what is wanted, so it is usually appropriate to use this
2996 pragma in conjunction with a @code{Export} or @code{Convention}
2997 pragma that specifies the desired foreign convention.
2999 @cindex Suppressing external name
3000 Special treatment is given if the EXTERNAL is an explicit null
3001 string or a static string expressions that evaluates to the null
3002 string. In this case, no external name is generated. This form
3003 still allows the specification of parameter mechanisms.
3005 @node Pragma Export_Value
3006 @unnumberedsec Pragma Export_Value
3007 @findex Export_Value
3011 @smallexample @c ada
3012 pragma Export_Value (
3013 [Value =>] static_integer_EXPRESSION,
3014 [Link_Name =>] static_string_EXPRESSION);
3018 This pragma serves to export a static integer value for external use.
3019 The first argument specifies the value to be exported. The Link_Name
3020 argument specifies the symbolic name to be associated with the integer
3021 value. This pragma is useful for defining a named static value in Ada
3022 that can be referenced in assembly language units to be linked with
3023 the application. This pragma is currently supported only for the
3024 AAMP target and is ignored for other targets.
3026 @node Pragma Export_Valued_Procedure
3027 @unnumberedsec Pragma Export_Valued_Procedure
3028 @findex Export_Valued_Procedure
3032 @smallexample @c ada
3033 pragma Export_Valued_Procedure (
3034 [Internal =>] LOCAL_NAME
3035 [, [External =>] EXTERNAL_SYMBOL]
3036 [, [Parameter_Types =>] PARAMETER_TYPES]
3037 [, [Mechanism =>] MECHANISM]);
3041 | static_string_EXPRESSION
3046 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
3050 | subtype_Name ' Access
3054 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
3056 MECHANISM_ASSOCIATION ::=
3057 [formal_parameter_NAME =>] MECHANISM_NAME
3059 MECHANISM_NAME ::= Value | Reference
3063 This pragma is identical to @code{Export_Procedure} except that the
3064 first parameter of @var{LOCAL_NAME}, which must be present, must be of
3065 mode @code{OUT}, and externally the subprogram is treated as a function
3066 with this parameter as the result of the function. GNAT provides for
3067 this capability to allow the use of @code{OUT} and @code{IN OUT}
3068 parameters in interfacing to external functions (which are not permitted
3070 GNAT does not require a separate pragma @code{Export}, but if none is
3071 present, @code{Convention Ada} is assumed, which is almost certainly
3072 not what is wanted since the whole point of this pragma is to interface
3073 with foreign language functions, so it is usually appropriate to use this
3074 pragma in conjunction with a @code{Export} or @code{Convention}
3075 pragma that specifies the desired foreign convention.
3077 @cindex Suppressing external name
3078 Special treatment is given if the EXTERNAL is an explicit null
3079 string or a static string expressions that evaluates to the null
3080 string. In this case, no external name is generated. This form
3081 still allows the specification of parameter mechanisms.
3083 @node Pragma Extend_System
3084 @unnumberedsec Pragma Extend_System
3085 @cindex @code{system}, extending
3087 @findex Extend_System
3091 @smallexample @c ada
3092 pragma Extend_System ([Name =>] IDENTIFIER);
3096 This pragma is used to provide backwards compatibility with other
3097 implementations that extend the facilities of package @code{System}. In
3098 GNAT, @code{System} contains only the definitions that are present in
3099 the Ada RM@. However, other implementations, notably the DEC Ada 83
3100 implementation, provide many extensions to package @code{System}.
3102 For each such implementation accommodated by this pragma, GNAT provides a
3103 package @code{Aux_@var{xxx}}, e.g.@: @code{Aux_DEC} for the DEC Ada 83
3104 implementation, which provides the required additional definitions. You
3105 can use this package in two ways. You can @code{with} it in the normal
3106 way and access entities either by selection or using a @code{use}
3107 clause. In this case no special processing is required.
3109 However, if existing code contains references such as
3110 @code{System.@var{xxx}} where @var{xxx} is an entity in the extended
3111 definitions provided in package @code{System}, you may use this pragma
3112 to extend visibility in @code{System} in a non-standard way that
3113 provides greater compatibility with the existing code. Pragma
3114 @code{Extend_System} is a configuration pragma whose single argument is
3115 the name of the package containing the extended definition
3116 (e.g.@: @code{Aux_DEC} for the DEC Ada case). A unit compiled under
3117 control of this pragma will be processed using special visibility
3118 processing that looks in package @code{System.Aux_@var{xxx}} where
3119 @code{Aux_@var{xxx}} is the pragma argument for any entity referenced in
3120 package @code{System}, but not found in package @code{System}.
3122 You can use this pragma either to access a predefined @code{System}
3123 extension supplied with the compiler, for example @code{Aux_DEC} or
3124 you can construct your own extension unit following the above
3125 definition. Note that such a package is a child of @code{System}
3126 and thus is considered part of the implementation.
3127 To compile it you will have to use the @option{-gnatg} switch,
3128 or the @option{/GNAT_INTERNAL} qualifier on OpenVMS,
3129 for compiling System units, as explained in the
3130 @value{EDITION} User's Guide.
3132 @node Pragma Extensions_Allowed
3133 @unnumberedsec Pragma Extensions_Allowed
3134 @cindex Ada Extensions
3135 @cindex GNAT Extensions
3136 @findex Extensions_Allowed
3140 @smallexample @c ada
3141 pragma Extensions_Allowed (On | Off);
3145 This configuration pragma enables or disables the implementation
3146 extension mode (the use of Off as a parameter cancels the effect
3147 of the @option{-gnatX} command switch).
3149 In extension mode, the latest version of the Ada language is
3150 implemented (currently Ada 2012), and in addition a small number
3151 of GNAT specific extensions are recognized as follows:
3154 @item Constrained attribute for generic objects
3155 The @code{Constrained} attribute is permitted for objects of
3156 generic types. The result indicates if the corresponding actual
3161 @node Pragma External
3162 @unnumberedsec Pragma External
3167 @smallexample @c ada
3169 [ Convention =>] convention_IDENTIFIER,
3170 [ Entity =>] LOCAL_NAME
3171 [, [External_Name =>] static_string_EXPRESSION ]
3172 [, [Link_Name =>] static_string_EXPRESSION ]);
3176 This pragma is identical in syntax and semantics to pragma
3177 @code{Export} as defined in the Ada Reference Manual. It is
3178 provided for compatibility with some Ada 83 compilers that
3179 used this pragma for exactly the same purposes as pragma
3180 @code{Export} before the latter was standardized.
3182 @node Pragma External_Name_Casing
3183 @unnumberedsec Pragma External_Name_Casing
3184 @cindex Dec Ada 83 casing compatibility
3185 @cindex External Names, casing
3186 @cindex Casing of External names
3187 @findex External_Name_Casing
3191 @smallexample @c ada
3192 pragma External_Name_Casing (
3193 Uppercase | Lowercase
3194 [, Uppercase | Lowercase | As_Is]);
3198 This pragma provides control over the casing of external names associated
3199 with Import and Export pragmas. There are two cases to consider:
3202 @item Implicit external names
3203 Implicit external names are derived from identifiers. The most common case
3204 arises when a standard Ada Import or Export pragma is used with only two
3207 @smallexample @c ada
3208 pragma Import (C, C_Routine);
3212 Since Ada is a case-insensitive language, the spelling of the identifier in
3213 the Ada source program does not provide any information on the desired
3214 casing of the external name, and so a convention is needed. In GNAT the
3215 default treatment is that such names are converted to all lower case
3216 letters. This corresponds to the normal C style in many environments.
3217 The first argument of pragma @code{External_Name_Casing} can be used to
3218 control this treatment. If @code{Uppercase} is specified, then the name
3219 will be forced to all uppercase letters. If @code{Lowercase} is specified,
3220 then the normal default of all lower case letters will be used.
3222 This same implicit treatment is also used in the case of extended DEC Ada 83
3223 compatible Import and Export pragmas where an external name is explicitly
3224 specified using an identifier rather than a string.
3226 @item Explicit external names
3227 Explicit external names are given as string literals. The most common case
3228 arises when a standard Ada Import or Export pragma is used with three
3231 @smallexample @c ada
3232 pragma Import (C, C_Routine, "C_routine");
3236 In this case, the string literal normally provides the exact casing required
3237 for the external name. The second argument of pragma
3238 @code{External_Name_Casing} may be used to modify this behavior.
3239 If @code{Uppercase} is specified, then the name
3240 will be forced to all uppercase letters. If @code{Lowercase} is specified,
3241 then the name will be forced to all lowercase letters. A specification of
3242 @code{As_Is} provides the normal default behavior in which the casing is
3243 taken from the string provided.
3247 This pragma may appear anywhere that a pragma is valid. In particular, it
3248 can be used as a configuration pragma in the @file{gnat.adc} file, in which
3249 case it applies to all subsequent compilations, or it can be used as a program
3250 unit pragma, in which case it only applies to the current unit, or it can
3251 be used more locally to control individual Import/Export pragmas.
3253 It is primarily intended for use with OpenVMS systems, where many
3254 compilers convert all symbols to upper case by default. For interfacing to
3255 such compilers (e.g.@: the DEC C compiler), it may be convenient to use
3258 @smallexample @c ada
3259 pragma External_Name_Casing (Uppercase, Uppercase);
3263 to enforce the upper casing of all external symbols.
3265 @node Pragma Fast_Math
3266 @unnumberedsec Pragma Fast_Math
3271 @smallexample @c ada
3276 This is a configuration pragma which activates a mode in which speed is
3277 considered more important for floating-point operations than absolutely
3278 accurate adherence to the requirements of the standard. Currently the
3279 following operations are affected:
3282 @item Complex Multiplication
3283 The normal simple formula for complex multiplication can result in intermediate
3284 overflows for numbers near the end of the range. The Ada standard requires that
3285 this situation be detected and corrected by scaling, but in Fast_Math mode such
3286 cases will simply result in overflow. Note that to take advantage of this you
3287 must instantiate your own version of @code{Ada.Numerics.Generic_Complex_Types}
3288 under control of the pragma, rather than use the preinstantiated versions.
3291 @node Pragma Favor_Top_Level
3292 @unnumberedsec Pragma Favor_Top_Level
3293 @findex Favor_Top_Level
3297 @smallexample @c ada
3298 pragma Favor_Top_Level (type_NAME);
3302 The named type must be an access-to-subprogram type. This pragma is an
3303 efficiency hint to the compiler, regarding the use of 'Access or
3304 'Unrestricted_Access on nested (non-library-level) subprograms. The
3305 pragma means that nested subprograms are not used with this type, or
3306 are rare, so that the generated code should be efficient in the
3307 top-level case. When this pragma is used, dynamically generated
3308 trampolines may be used on some targets for nested subprograms.
3309 See also the No_Implicit_Dynamic_Code restriction.
3311 @node Pragma Finalize_Storage_Only
3312 @unnumberedsec Pragma Finalize_Storage_Only
3313 @findex Finalize_Storage_Only
3317 @smallexample @c ada
3318 pragma Finalize_Storage_Only (first_subtype_LOCAL_NAME);
3322 This pragma allows the compiler not to emit a Finalize call for objects
3323 defined at the library level. This is mostly useful for types where
3324 finalization is only used to deal with storage reclamation since in most
3325 environments it is not necessary to reclaim memory just before terminating
3326 execution, hence the name.
3328 @node Pragma Float_Representation
3329 @unnumberedsec Pragma Float_Representation
3331 @findex Float_Representation
3335 @smallexample @c ada
3336 pragma Float_Representation (FLOAT_REP[, float_type_LOCAL_NAME]);
3338 FLOAT_REP ::= VAX_Float | IEEE_Float
3342 In the one argument form, this pragma is a configuration pragma which
3343 allows control over the internal representation chosen for the predefined
3344 floating point types declared in the packages @code{Standard} and
3345 @code{System}. On all systems other than OpenVMS, the argument must
3346 be @code{IEEE_Float} and the pragma has no effect. On OpenVMS, the
3347 argument may be @code{VAX_Float} to specify the use of the VAX float
3348 format for the floating-point types in Standard. This requires that
3349 the standard runtime libraries be recompiled.
3351 The two argument form specifies the representation to be used for
3352 the specified floating-point type. On all systems other than OpenVMS,
3354 be @code{IEEE_Float} to specify the use of IEEE format, as follows:
3358 For a digits value of 6, 32-bit IEEE short format will be used.
3360 For a digits value of 15, 64-bit IEEE long format will be used.
3362 No other value of digits is permitted.
3366 argument may be @code{VAX_Float} to specify the use of the VAX float
3371 For digits values up to 6, F float format will be used.
3373 For digits values from 7 to 9, D float format will be used.
3375 For digits values from 10 to 15, G float format will be used.
3377 Digits values above 15 are not allowed.
3381 @unnumberedsec Pragma Global
3384 For the description of this pragma, see SPARK 2014 Reference Manual,
3388 @unnumberedsec Pragma Ident
3393 @smallexample @c ada
3394 pragma Ident (static_string_EXPRESSION);
3398 This pragma is identical in effect to pragma @code{Comment}. It is provided
3399 for compatibility with other Ada compilers providing this pragma.
3401 @node Pragma Implementation_Defined
3402 @unnumberedsec Pragma Implementation_Defined
3403 @findex Implementation_Defined
3407 @smallexample @c ada
3408 pragma Implementation_Defined (local_NAME);
3412 This pragma marks a previously declared entioty as implementation-defined.
3413 For an overloaded entity, applies to the most recent homonym.
3415 @smallexample @c ada
3416 pragma Implementation_Defined;
3420 The form with no arguments appears anywhere within a scope, most
3421 typically a package spec, and indicates that all entities that are
3422 defined within the package spec are Implementation_Defined.
3424 This pragma is used within the GNAT runtime library to identify
3425 implementation-defined entities introduced in language-defined units,
3426 for the purpose of implementing the No_Implementation_Identifiers
3429 @node Pragma Implemented
3430 @unnumberedsec Pragma Implemented
3435 @smallexample @c ada
3436 pragma Implemented (procedure_LOCAL_NAME, implementation_kind);
3438 implementation_kind ::= By_Entry | By_Protected_Procedure | By_Any
3442 This is an Ada 2012 representation pragma which applies to protected, task
3443 and synchronized interface primitives. The use of pragma Implemented provides
3444 a way to impose a static requirement on the overriding operation by adhering
3445 to one of the three implementation kinds: entry, protected procedure or any of
3446 the above. This pragma is available in all earlier versions of Ada as an
3447 implementation-defined pragma.
3449 @smallexample @c ada
3450 type Synch_Iface is synchronized interface;
3451 procedure Prim_Op (Obj : in out Iface) is abstract;
3452 pragma Implemented (Prim_Op, By_Protected_Procedure);
3454 protected type Prot_1 is new Synch_Iface with
3455 procedure Prim_Op; -- Legal
3458 protected type Prot_2 is new Synch_Iface with
3459 entry Prim_Op; -- Illegal
3462 task type Task_Typ is new Synch_Iface with
3463 entry Prim_Op; -- Illegal
3468 When applied to the procedure_or_entry_NAME of a requeue statement, pragma
3469 Implemented determines the runtime behavior of the requeue. Implementation kind
3470 By_Entry guarantees that the action of requeueing will proceed from an entry to
3471 another entry. Implementation kind By_Protected_Procedure transforms the
3472 requeue into a dispatching call, thus eliminating the chance of blocking. Kind
3473 By_Any shares the behavior of By_Entry and By_Protected_Procedure depending on
3474 the target's overriding subprogram kind.
3476 @node Pragma Implicit_Packing
3477 @unnumberedsec Pragma Implicit_Packing
3478 @findex Implicit_Packing
3479 @cindex Rational Profile
3483 @smallexample @c ada
3484 pragma Implicit_Packing;
3488 This is a configuration pragma that requests implicit packing for packed
3489 arrays for which a size clause is given but no explicit pragma Pack or
3490 specification of Component_Size is present. It also applies to records
3491 where no record representation clause is present. Consider this example:
3493 @smallexample @c ada
3494 type R is array (0 .. 7) of Boolean;
3499 In accordance with the recommendation in the RM (RM 13.3(53)), a Size clause
3500 does not change the layout of a composite object. So the Size clause in the
3501 above example is normally rejected, since the default layout of the array uses
3502 8-bit components, and thus the array requires a minimum of 64 bits.
3504 If this declaration is compiled in a region of code covered by an occurrence
3505 of the configuration pragma Implicit_Packing, then the Size clause in this
3506 and similar examples will cause implicit packing and thus be accepted. For
3507 this implicit packing to occur, the type in question must be an array of small
3508 components whose size is known at compile time, and the Size clause must
3509 specify the exact size that corresponds to the number of elements in the array
3510 multiplied by the size in bits of the component type (both single and
3511 multi-dimensioned arrays can be controlled with this pragma).
3513 @cindex Array packing
3515 Similarly, the following example shows the use in the record case
3517 @smallexample @c ada
3519 a, b, c, d, e, f, g, h : boolean;
3526 Without a pragma Pack, each Boolean field requires 8 bits, so the
3527 minimum size is 72 bits, but with a pragma Pack, 16 bits would be
3528 sufficient. The use of pragma Implicit_Packing allows this record
3529 declaration to compile without an explicit pragma Pack.
3530 @node Pragma Import_Exception
3531 @unnumberedsec Pragma Import_Exception
3533 @findex Import_Exception
3537 @smallexample @c ada
3538 pragma Import_Exception (
3539 [Internal =>] LOCAL_NAME
3540 [, [External =>] EXTERNAL_SYMBOL]
3541 [, [Form =>] Ada | VMS]
3542 [, [Code =>] static_integer_EXPRESSION]);
3546 | static_string_EXPRESSION
3550 This pragma is implemented only in the OpenVMS implementation of GNAT@.
3551 It allows OpenVMS conditions (for example, from OpenVMS system services or
3552 other OpenVMS languages) to be propagated to Ada programs as Ada exceptions.
3553 The pragma specifies that the exception associated with an exception
3554 declaration in an Ada program be defined externally (in non-Ada code).
3555 For further details on this pragma, see the
3556 DEC Ada Language Reference Manual, section 13.9a.3.1.
3558 @node Pragma Import_Function
3559 @unnumberedsec Pragma Import_Function
3560 @findex Import_Function
3564 @smallexample @c ada
3565 pragma Import_Function (
3566 [Internal =>] LOCAL_NAME,
3567 [, [External =>] EXTERNAL_SYMBOL]
3568 [, [Parameter_Types =>] PARAMETER_TYPES]
3569 [, [Result_Type =>] SUBTYPE_MARK]
3570 [, [Mechanism =>] MECHANISM]
3571 [, [Result_Mechanism =>] MECHANISM_NAME]);
3575 | static_string_EXPRESSION
3579 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
3583 | subtype_Name ' Access
3587 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
3589 MECHANISM_ASSOCIATION ::=
3590 [formal_parameter_NAME =>] MECHANISM_NAME
3598 This pragma is used in conjunction with a pragma @code{Import} to
3599 specify additional information for an imported function. The pragma
3600 @code{Import} (or equivalent pragma @code{Interface}) must precede the
3601 @code{Import_Function} pragma and both must appear in the same
3602 declarative part as the function specification.
3604 The @var{Internal} argument must uniquely designate
3605 the function to which the
3606 pragma applies. If more than one function name exists of this name in
3607 the declarative part you must use the @code{Parameter_Types} and
3608 @var{Result_Type} parameters to achieve the required unique
3609 designation. Subtype marks in these parameters must exactly match the
3610 subtypes in the corresponding function specification, using positional
3611 notation to match parameters with subtype marks.
3612 The form with an @code{'Access} attribute can be used to match an
3613 anonymous access parameter.
3615 You may optionally use the @var{Mechanism} and @var{Result_Mechanism}
3616 parameters to specify passing mechanisms for the
3617 parameters and result. If you specify a single mechanism name, it
3618 applies to all parameters. Otherwise you may specify a mechanism on a
3619 parameter by parameter basis using either positional or named
3620 notation. If the mechanism is not specified, the default mechanism
3623 @node Pragma Import_Object
3624 @unnumberedsec Pragma Import_Object
3625 @findex Import_Object
3629 @smallexample @c ada
3630 pragma Import_Object
3631 [Internal =>] LOCAL_NAME
3632 [, [External =>] EXTERNAL_SYMBOL]
3633 [, [Size =>] EXTERNAL_SYMBOL]);
3637 | static_string_EXPRESSION
3641 This pragma designates an object as imported, and apart from the
3642 extended rules for external symbols, is identical in effect to the use of
3643 the normal @code{Import} pragma applied to an object. Unlike the
3644 subprogram case, you need not use a separate @code{Import} pragma,
3645 although you may do so (and probably should do so from a portability
3646 point of view). @var{size} is syntax checked, but otherwise ignored by
3649 @node Pragma Import_Procedure
3650 @unnumberedsec Pragma Import_Procedure
3651 @findex Import_Procedure
3655 @smallexample @c ada
3656 pragma Import_Procedure (
3657 [Internal =>] LOCAL_NAME
3658 [, [External =>] EXTERNAL_SYMBOL]
3659 [, [Parameter_Types =>] PARAMETER_TYPES]
3660 [, [Mechanism =>] MECHANISM]);
3664 | static_string_EXPRESSION
3668 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
3672 | subtype_Name ' Access
3676 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
3678 MECHANISM_ASSOCIATION ::=
3679 [formal_parameter_NAME =>] MECHANISM_NAME
3681 MECHANISM_NAME ::= Value | Reference
3685 This pragma is identical to @code{Import_Function} except that it
3686 applies to a procedure rather than a function and the parameters
3687 @code{Result_Type} and @code{Result_Mechanism} are not permitted.
3689 @node Pragma Import_Valued_Procedure
3690 @unnumberedsec Pragma Import_Valued_Procedure
3691 @findex Import_Valued_Procedure
3695 @smallexample @c ada
3696 pragma Import_Valued_Procedure (
3697 [Internal =>] LOCAL_NAME
3698 [, [External =>] EXTERNAL_SYMBOL]
3699 [, [Parameter_Types =>] PARAMETER_TYPES]
3700 [, [Mechanism =>] MECHANISM]);
3704 | static_string_EXPRESSION
3708 | TYPE_DESIGNATOR @{, TYPE_DESIGNATOR@}
3712 | subtype_Name ' Access
3716 | (MECHANISM_ASSOCIATION @{, MECHANISM_ASSOCIATION@})
3718 MECHANISM_ASSOCIATION ::=
3719 [formal_parameter_NAME =>] MECHANISM_NAME
3721 MECHANISM_NAME ::= Value | Reference
3725 This pragma is identical to @code{Import_Procedure} except that the
3726 first parameter of @var{LOCAL_NAME}, which must be present, must be of
3727 mode @code{OUT}, and externally the subprogram is treated as a function
3728 with this parameter as the result of the function. The purpose of this
3729 capability is to allow the use of @code{OUT} and @code{IN OUT}
3730 parameters in interfacing to external functions (which are not permitted
3731 in Ada functions). You may optionally use the @code{Mechanism}
3732 parameters to specify passing mechanisms for the parameters.
3733 If you specify a single mechanism name, it applies to all parameters.
3734 Otherwise you may specify a mechanism on a parameter by parameter
3735 basis using either positional or named notation. If the mechanism is not
3736 specified, the default mechanism is used.
3738 Note that it is important to use this pragma in conjunction with a separate
3739 pragma Import that specifies the desired convention, since otherwise the
3740 default convention is Ada, which is almost certainly not what is required.
3742 @node Pragma Independent
3743 @unnumberedsec Pragma Independent
3748 @smallexample @c ada
3749 pragma Independent (Local_NAME);
3753 This pragma is standard in Ada 2012 mode (which also provides an aspect
3754 of the same name). It is also available as an implementation-defined
3755 pragma in all earlier versions. It specifies that the
3756 designated object or all objects of the designated type must be
3757 independently addressable. This means that separate tasks can safely
3758 manipulate such objects. For example, if two components of a record are
3759 independent, then two separate tasks may access these two components.
3761 constraints on the representation of the object (for instance prohibiting
3764 @node Pragma Independent_Components
3765 @unnumberedsec Pragma Independent_Components
3766 @findex Independent_Components
3770 @smallexample @c ada
3771 pragma Independent_Components (Local_NAME);
3775 This pragma is standard in Ada 2012 mode (which also provides an aspect
3776 of the same name). It is also available as an implementation-defined
3777 pragma in all earlier versions. It specifies that the components of the
3778 designated object, or the components of each object of the designated
3780 independently addressable. This means that separate tasks can safely
3781 manipulate separate components in the composite object. This may place
3782 constraints on the representation of the object (for instance prohibiting
3785 @node Pragma Initial_Condition
3786 @unnumberedsec Pragma Initial_Condition
3787 @findex Initial_Condition
3789 For the description of this pragma, see SPARK 2014 Reference Manual,
3792 @node Pragma Initialize_Scalars
3793 @unnumberedsec Pragma Initialize_Scalars
3794 @findex Initialize_Scalars
3795 @cindex debugging with Initialize_Scalars
3799 @smallexample @c ada
3800 pragma Initialize_Scalars;
3804 This pragma is similar to @code{Normalize_Scalars} conceptually but has
3805 two important differences. First, there is no requirement for the pragma
3806 to be used uniformly in all units of a partition, in particular, it is fine
3807 to use this just for some or all of the application units of a partition,
3808 without needing to recompile the run-time library.
3810 In the case where some units are compiled with the pragma, and some without,
3811 then a declaration of a variable where the type is defined in package
3812 Standard or is locally declared will always be subject to initialization,
3813 as will any declaration of a scalar variable. For composite variables,
3814 whether the variable is initialized may also depend on whether the package
3815 in which the type of the variable is declared is compiled with the pragma.
3817 The other important difference is that you can control the value used
3818 for initializing scalar objects. At bind time, you can select several
3819 options for initialization. You can
3820 initialize with invalid values (similar to Normalize_Scalars, though for
3821 Initialize_Scalars it is not always possible to determine the invalid
3822 values in complex cases like signed component fields with non-standard
3823 sizes). You can also initialize with high or
3824 low values, or with a specified bit pattern. See the @value{EDITION}
3825 User's Guide for binder options for specifying these cases.
3827 This means that you can compile a program, and then without having to
3828 recompile the program, you can run it with different values being used
3829 for initializing otherwise uninitialized values, to test if your program
3830 behavior depends on the choice. Of course the behavior should not change,
3831 and if it does, then most likely you have an incorrect reference to an
3832 uninitialized value.
3834 It is even possible to change the value at execution time eliminating even
3835 the need to rebind with a different switch using an environment variable.
3836 See the @value{EDITION} User's Guide for details.
3838 Note that pragma @code{Initialize_Scalars} is particularly useful in
3839 conjunction with the enhanced validity checking that is now provided
3840 in GNAT, which checks for invalid values under more conditions.
3841 Using this feature (see description of the @option{-gnatV} flag in the
3842 @value{EDITION} User's Guide) in conjunction with
3843 pragma @code{Initialize_Scalars}
3844 provides a powerful new tool to assist in the detection of problems
3845 caused by uninitialized variables.
3847 Note: the use of @code{Initialize_Scalars} has a fairly extensive
3848 effect on the generated code. This may cause your code to be
3849 substantially larger. It may also cause an increase in the amount
3850 of stack required, so it is probably a good idea to turn on stack
3851 checking (see description of stack checking in the @value{EDITION}
3852 User's Guide) when using this pragma.
3854 @node Pragma Initializes
3855 @unnumberedsec Pragma Initializes
3858 For the description of this pragma, see SPARK 2014 Reference Manual,
3861 @node Pragma Inline_Always
3862 @unnumberedsec Pragma Inline_Always
3863 @findex Inline_Always
3867 @smallexample @c ada
3868 pragma Inline_Always (NAME [, NAME]);
3872 Similar to pragma @code{Inline} except that inlining is not subject to
3873 the use of option @option{-gnatn} or @option{-gnatN} and the inlining
3874 happens regardless of whether these options are used.
3876 @node Pragma Inline_Generic
3877 @unnumberedsec Pragma Inline_Generic
3878 @findex Inline_Generic
3882 @smallexample @c ada
3883 pragma Inline_Generic (GNAME @{, GNAME@});
3885 GNAME ::= generic_unit_NAME | generic_instance_NAME
3889 This pragma is provided for compatibility with Dec Ada 83. It has
3890 no effect in @code{GNAT} (which always inlines generics), other
3891 than to check that the given names are all names of generic units or
3894 @node Pragma Interface
3895 @unnumberedsec Pragma Interface
3900 @smallexample @c ada
3902 [Convention =>] convention_identifier,
3903 [Entity =>] local_NAME
3904 [, [External_Name =>] static_string_expression]
3905 [, [Link_Name =>] static_string_expression]);
3909 This pragma is identical in syntax and semantics to
3910 the standard Ada pragma @code{Import}. It is provided for compatibility
3911 with Ada 83. The definition is upwards compatible both with pragma
3912 @code{Interface} as defined in the Ada 83 Reference Manual, and also
3913 with some extended implementations of this pragma in certain Ada 83
3914 implementations. The only difference between pragma @code{Interface}
3915 and pragma @code{Import} is that there is special circuitry to allow
3916 both pragmas to appear for the same subprogram entity (normally it
3917 is illegal to have multiple @code{Import} pragmas. This is useful in
3918 maintaining Ada 83/Ada 95 compatibility and is compatible with other
3921 @node Pragma Interface_Name
3922 @unnumberedsec Pragma Interface_Name
3923 @findex Interface_Name
3927 @smallexample @c ada
3928 pragma Interface_Name (
3929 [Entity =>] LOCAL_NAME
3930 [, [External_Name =>] static_string_EXPRESSION]
3931 [, [Link_Name =>] static_string_EXPRESSION]);
3935 This pragma provides an alternative way of specifying the interface name
3936 for an interfaced subprogram, and is provided for compatibility with Ada
3937 83 compilers that use the pragma for this purpose. You must provide at
3938 least one of @var{External_Name} or @var{Link_Name}.
3940 @node Pragma Interrupt_Handler
3941 @unnumberedsec Pragma Interrupt_Handler
3942 @findex Interrupt_Handler
3946 @smallexample @c ada
3947 pragma Interrupt_Handler (procedure_LOCAL_NAME);
3951 This program unit pragma is supported for parameterless protected procedures
3952 as described in Annex C of the Ada Reference Manual. On the AAMP target
3953 the pragma can also be specified for nonprotected parameterless procedures
3954 that are declared at the library level (which includes procedures
3955 declared at the top level of a library package). In the case of AAMP,
3956 when this pragma is applied to a nonprotected procedure, the instruction
3957 @code{IERET} is generated for returns from the procedure, enabling
3958 maskable interrupts, in place of the normal return instruction.
3960 @node Pragma Interrupt_State
3961 @unnumberedsec Pragma Interrupt_State
3962 @findex Interrupt_State
3966 @smallexample @c ada
3967 pragma Interrupt_State
3969 [State =>] SYSTEM | RUNTIME | USER);
3973 Normally certain interrupts are reserved to the implementation. Any attempt
3974 to attach an interrupt causes Program_Error to be raised, as described in
3975 RM C.3.2(22). A typical example is the @code{SIGINT} interrupt used in
3976 many systems for an @kbd{Ctrl-C} interrupt. Normally this interrupt is
3977 reserved to the implementation, so that @kbd{Ctrl-C} can be used to
3978 interrupt execution. Additionally, signals such as @code{SIGSEGV},
3979 @code{SIGABRT}, @code{SIGFPE} and @code{SIGILL} are often mapped to specific
3980 Ada exceptions, or used to implement run-time functions such as the
3981 @code{abort} statement and stack overflow checking.
3983 Pragma @code{Interrupt_State} provides a general mechanism for overriding
3984 such uses of interrupts. It subsumes the functionality of pragma
3985 @code{Unreserve_All_Interrupts}. Pragma @code{Interrupt_State} is not
3986 available on Windows or VMS. On all other platforms than VxWorks,
3987 it applies to signals; on VxWorks, it applies to vectored hardware interrupts
3988 and may be used to mark interrupts required by the board support package
3991 Interrupts can be in one of three states:
3995 The interrupt is reserved (no Ada handler can be installed), and the
3996 Ada run-time may not install a handler. As a result you are guaranteed
3997 standard system default action if this interrupt is raised.
4001 The interrupt is reserved (no Ada handler can be installed). The run time
4002 is allowed to install a handler for internal control purposes, but is
4003 not required to do so.
4007 The interrupt is unreserved. The user may install a handler to provide
4012 These states are the allowed values of the @code{State} parameter of the
4013 pragma. The @code{Name} parameter is a value of the type
4014 @code{Ada.Interrupts.Interrupt_ID}. Typically, it is a name declared in
4015 @code{Ada.Interrupts.Names}.
4017 This is a configuration pragma, and the binder will check that there
4018 are no inconsistencies between different units in a partition in how a
4019 given interrupt is specified. It may appear anywhere a pragma is legal.
4021 The effect is to move the interrupt to the specified state.
4023 By declaring interrupts to be SYSTEM, you guarantee the standard system
4024 action, such as a core dump.
4026 By declaring interrupts to be USER, you guarantee that you can install
4029 Note that certain signals on many operating systems cannot be caught and
4030 handled by applications. In such cases, the pragma is ignored. See the
4031 operating system documentation, or the value of the array @code{Reserved}
4032 declared in the spec of package @code{System.OS_Interface}.
4034 Overriding the default state of signals used by the Ada runtime may interfere
4035 with an application's runtime behavior in the cases of the synchronous signals,
4036 and in the case of the signal used to implement the @code{abort} statement.
4038 @node Pragma Invariant
4039 @unnumberedsec Pragma Invariant
4044 @smallexample @c ada
4046 ([Entity =>] private_type_LOCAL_NAME,
4047 [Check =>] EXPRESSION
4048 [,[Message =>] String_Expression]);
4052 This pragma provides exactly the same capabilities as the Type_Invariant aspect
4053 defined in AI05-0146-1, and in the Ada 2012 Reference Manual. The
4054 Type_Invariant aspect is fully implemented in Ada 2012 mode, but since it
4055 requires the use of the aspect syntax, which is not available except in 2012
4056 mode, it is not possible to use the Type_Invariant aspect in earlier versions
4057 of Ada. However the Invariant pragma may be used in any version of Ada. Also
4058 note that the aspect Invariant is a synonym in GNAT for the aspect
4059 Type_Invariant, but there is no pragma Type_Invariant.
4061 The pragma must appear within the visible part of the package specification,
4062 after the type to which its Entity argument appears. As with the Invariant
4063 aspect, the Check expression is not analyzed until the end of the visible
4064 part of the package, so it may contain forward references. The Message
4065 argument, if present, provides the exception message used if the invariant
4066 is violated. If no Message parameter is provided, a default message that
4067 identifies the line on which the pragma appears is used.
4069 It is permissible to have multiple Invariants for the same type entity, in
4070 which case they are and'ed together. It is permissible to use this pragma
4071 in Ada 2012 mode, but you cannot have both an invariant aspect and an
4072 invariant pragma for the same entity.
4074 For further details on the use of this pragma, see the Ada 2012 documentation
4075 of the Type_Invariant aspect.
4077 @node Pragma Java_Constructor
4078 @unnumberedsec Pragma Java_Constructor
4079 @findex Java_Constructor
4083 @smallexample @c ada
4084 pragma Java_Constructor ([Entity =>] function_LOCAL_NAME);
4088 This pragma is used to assert that the specified Ada function should be
4089 mapped to the Java constructor for some Ada tagged record type.
4091 See section 7.3.2 of the
4092 @code{GNAT User's Guide: Supplement for the JVM Platform.}
4093 for related information.
4095 @node Pragma Java_Interface
4096 @unnumberedsec Pragma Java_Interface
4097 @findex Java_Interface
4101 @smallexample @c ada
4102 pragma Java_Interface ([Entity =>] abstract_tagged_type_LOCAL_NAME);
4106 This pragma is used to assert that the specified Ada abstract tagged type
4107 is to be mapped to a Java interface name.
4109 See sections 7.1 and 7.2 of the
4110 @code{GNAT User's Guide: Supplement for the JVM Platform.}
4111 for related information.
4113 @node Pragma Keep_Names
4114 @unnumberedsec Pragma Keep_Names
4119 @smallexample @c ada
4120 pragma Keep_Names ([On =>] enumeration_first_subtype_LOCAL_NAME);
4124 The @var{LOCAL_NAME} argument
4125 must refer to an enumeration first subtype
4126 in the current declarative part. The effect is to retain the enumeration
4127 literal names for use by @code{Image} and @code{Value} even if a global
4128 @code{Discard_Names} pragma applies. This is useful when you want to
4129 generally suppress enumeration literal names and for example you therefore
4130 use a @code{Discard_Names} pragma in the @file{gnat.adc} file, but you
4131 want to retain the names for specific enumeration types.
4133 @node Pragma License
4134 @unnumberedsec Pragma License
4136 @cindex License checking
4140 @smallexample @c ada
4141 pragma License (Unrestricted | GPL | Modified_GPL | Restricted);
4145 This pragma is provided to allow automated checking for appropriate license
4146 conditions with respect to the standard and modified GPL@. A pragma
4147 @code{License}, which is a configuration pragma that typically appears at
4148 the start of a source file or in a separate @file{gnat.adc} file, specifies
4149 the licensing conditions of a unit as follows:
4153 This is used for a unit that can be freely used with no license restrictions.
4154 Examples of such units are public domain units, and units from the Ada
4158 This is used for a unit that is licensed under the unmodified GPL, and which
4159 therefore cannot be @code{with}'ed by a restricted unit.
4162 This is used for a unit licensed under the GNAT modified GPL that includes
4163 a special exception paragraph that specifically permits the inclusion of
4164 the unit in programs without requiring the entire program to be released
4168 This is used for a unit that is restricted in that it is not permitted to
4169 depend on units that are licensed under the GPL@. Typical examples are
4170 proprietary code that is to be released under more restrictive license
4171 conditions. Note that restricted units are permitted to @code{with} units
4172 which are licensed under the modified GPL (this is the whole point of the
4178 Normally a unit with no @code{License} pragma is considered to have an
4179 unknown license, and no checking is done. However, standard GNAT headers
4180 are recognized, and license information is derived from them as follows.
4182 A GNAT license header starts with a line containing 78 hyphens. The following
4183 comment text is searched for the appearance of any of the following strings.
4185 If the string ``GNU General Public License'' is found, then the unit is assumed
4186 to have GPL license, unless the string ``As a special exception'' follows, in
4187 which case the license is assumed to be modified GPL@.
4189 If one of the strings
4190 ``This specification is adapted from the Ada Semantic Interface'' or
4191 ``This specification is derived from the Ada Reference Manual'' is found
4192 then the unit is assumed to be unrestricted.
4195 These default actions means that a program with a restricted license pragma
4196 will automatically get warnings if a GPL unit is inappropriately
4197 @code{with}'ed. For example, the program:
4199 @smallexample @c ada
4202 procedure Secret_Stuff is
4208 if compiled with pragma @code{License} (@code{Restricted}) in a
4209 @file{gnat.adc} file will generate the warning:
4214 >>> license of withed unit "Sem_Ch3" is incompatible
4216 2. with GNAT.Sockets;
4217 3. procedure Secret_Stuff is
4221 Here we get a warning on @code{Sem_Ch3} since it is part of the GNAT
4222 compiler and is licensed under the
4223 GPL, but no warning for @code{GNAT.Sockets} which is part of the GNAT
4224 run time, and is therefore licensed under the modified GPL@.
4226 @node Pragma Link_With
4227 @unnumberedsec Pragma Link_With
4232 @smallexample @c ada
4233 pragma Link_With (static_string_EXPRESSION @{,static_string_EXPRESSION@});
4237 This pragma is provided for compatibility with certain Ada 83 compilers.
4238 It has exactly the same effect as pragma @code{Linker_Options} except
4239 that spaces occurring within one of the string expressions are treated
4240 as separators. For example, in the following case:
4242 @smallexample @c ada
4243 pragma Link_With ("-labc -ldef");
4247 results in passing the strings @code{-labc} and @code{-ldef} as two
4248 separate arguments to the linker. In addition pragma Link_With allows
4249 multiple arguments, with the same effect as successive pragmas.
4251 @node Pragma Linker_Alias
4252 @unnumberedsec Pragma Linker_Alias
4253 @findex Linker_Alias
4257 @smallexample @c ada
4258 pragma Linker_Alias (
4259 [Entity =>] LOCAL_NAME,
4260 [Target =>] static_string_EXPRESSION);
4264 @var{LOCAL_NAME} must refer to an object that is declared at the library
4265 level. This pragma establishes the given entity as a linker alias for the
4266 given target. It is equivalent to @code{__attribute__((alias))} in GNU C
4267 and causes @var{LOCAL_NAME} to be emitted as an alias for the symbol
4268 @var{static_string_EXPRESSION} in the object file, that is to say no space
4269 is reserved for @var{LOCAL_NAME} by the assembler and it will be resolved
4270 to the same address as @var{static_string_EXPRESSION} by the linker.
4272 The actual linker name for the target must be used (e.g.@: the fully
4273 encoded name with qualification in Ada, or the mangled name in C++),
4274 or it must be declared using the C convention with @code{pragma Import}
4275 or @code{pragma Export}.
4277 Not all target machines support this pragma. On some of them it is accepted
4278 only if @code{pragma Weak_External} has been applied to @var{LOCAL_NAME}.
4280 @smallexample @c ada
4281 -- Example of the use of pragma Linker_Alias
4285 pragma Export (C, i);
4287 new_name_for_i : Integer;
4288 pragma Linker_Alias (new_name_for_i, "i");
4292 @node Pragma Linker_Constructor
4293 @unnumberedsec Pragma Linker_Constructor
4294 @findex Linker_Constructor
4298 @smallexample @c ada
4299 pragma Linker_Constructor (procedure_LOCAL_NAME);
4303 @var{procedure_LOCAL_NAME} must refer to a parameterless procedure that
4304 is declared at the library level. A procedure to which this pragma is
4305 applied will be treated as an initialization routine by the linker.
4306 It is equivalent to @code{__attribute__((constructor))} in GNU C and
4307 causes @var{procedure_LOCAL_NAME} to be invoked before the entry point
4308 of the executable is called (or immediately after the shared library is
4309 loaded if the procedure is linked in a shared library), in particular
4310 before the Ada run-time environment is set up.
4312 Because of these specific contexts, the set of operations such a procedure
4313 can perform is very limited and the type of objects it can manipulate is
4314 essentially restricted to the elementary types. In particular, it must only
4315 contain code to which pragma Restrictions (No_Elaboration_Code) applies.
4317 This pragma is used by GNAT to implement auto-initialization of shared Stand
4318 Alone Libraries, which provides a related capability without the restrictions
4319 listed above. Where possible, the use of Stand Alone Libraries is preferable
4320 to the use of this pragma.
4322 @node Pragma Linker_Destructor
4323 @unnumberedsec Pragma Linker_Destructor
4324 @findex Linker_Destructor
4328 @smallexample @c ada
4329 pragma Linker_Destructor (procedure_LOCAL_NAME);
4333 @var{procedure_LOCAL_NAME} must refer to a parameterless procedure that
4334 is declared at the library level. A procedure to which this pragma is
4335 applied will be treated as a finalization routine by the linker.
4336 It is equivalent to @code{__attribute__((destructor))} in GNU C and
4337 causes @var{procedure_LOCAL_NAME} to be invoked after the entry point
4338 of the executable has exited (or immediately before the shared library
4339 is unloaded if the procedure is linked in a shared library), in particular
4340 after the Ada run-time environment is shut down.
4342 See @code{pragma Linker_Constructor} for the set of restrictions that apply
4343 because of these specific contexts.
4345 @node Pragma Linker_Section
4346 @unnumberedsec Pragma Linker_Section
4347 @findex Linker_Section
4351 @smallexample @c ada
4352 pragma Linker_Section (
4353 [Entity =>] LOCAL_NAME,
4354 [Section =>] static_string_EXPRESSION);
4358 @var{LOCAL_NAME} must refer to an object, type, or subprogram that is
4359 declared at the library level. This pragma specifies the name of the
4360 linker section for the given entity. It is equivalent to
4361 @code{__attribute__((section))} in GNU C and causes @var{LOCAL_NAME} to
4362 be placed in the @var{static_string_EXPRESSION} section of the
4363 executable (assuming the linker doesn't rename the section).
4364 GNAT also provides an implementation defined aspect of the same name.
4366 In the case of specifying this aspect for a type, the effect is to
4367 specify the corresponding for all library level objects of the type which
4368 do not have an explicit linker section set. Note that this only applies to
4369 whole objects, not to components of composite objects.
4371 In the case of a subprogram, the linker section applies to all previously
4372 declared matching overloaded subprograms in the current declarative part
4373 which do not already have a linker section assigned. The linker section
4374 aspect is useful in this case for specifying different linker sections
4375 for different elements of such an overloaded set.
4377 Note that an empty string specifies that no linker section is specified.
4378 This is not quite the same as omitting the pragma or aspect, since it
4379 can be used to specify that one element of an overloaded set of subprograms
4380 has the default linker section, or that one object of a type for which a
4381 linker section is specified should has the default linker section.
4383 The compiler normally places library-level entities in standard sections
4384 depending on the class: procedures and functions generally go in the
4385 @code{.text} section, initialized variables in the @code{.data} section
4386 and uninitialized variables in the @code{.bss} section.
4388 Other, special sections may exist on given target machines to map special
4389 hardware, for example I/O ports or flash memory. This pragma is a means to
4390 defer the final layout of the executable to the linker, thus fully working
4391 at the symbolic level with the compiler.
4393 Some file formats do not support arbitrary sections so not all target
4394 machines support this pragma. The use of this pragma may cause a program
4395 execution to be erroneous if it is used to place an entity into an
4396 inappropriate section (e.g.@: a modified variable into the @code{.text}
4397 section). See also @code{pragma Persistent_BSS}.
4399 @smallexample @c ada
4400 -- Example of the use of pragma Linker_Section
4404 pragma Volatile (Port_A);
4405 pragma Linker_Section (Port_A, ".bss.port_a");
4408 pragma Volatile (Port_B);
4409 pragma Linker_Section (Port_B, ".bss.port_b");
4411 type Port_Type is new Integer with Linker_Section => ".bss";
4412 PA : Port_Type with Linker_Section => ".bss.PA";
4413 PB : Port_Type; -- ends up in linker section ".bss"
4415 procedure Q with Linker_Section => "Qsection";
4419 @node Pragma Lock_Free
4420 @unnumberedsec Pragma Lock_Free
4424 This pragma may be specified for protected types or objects. It specifies that
4425 the implementation of protected operations must be implemented without locks.
4426 Compilation fails if the compiler cannot generate lock-free code for the
4429 @node Pragma Long_Float
4430 @unnumberedsec Pragma Long_Float
4436 @smallexample @c ada
4437 pragma Long_Float (FLOAT_FORMAT);
4439 FLOAT_FORMAT ::= D_Float | G_Float
4443 This pragma is implemented only in the OpenVMS implementation of GNAT@.
4444 It allows control over the internal representation chosen for the predefined
4445 type @code{Long_Float} and for floating point type representations with
4446 @code{digits} specified in the range 7 through 15.
4447 For further details on this pragma, see the
4448 @cite{DEC Ada Language Reference Manual}, section 3.5.7b. Note that to use
4449 this pragma, the standard runtime libraries must be recompiled.
4451 @node Pragma Loop_Invariant
4452 @unnumberedsec Pragma Loop_Invariant
4453 @findex Loop_Invariant
4457 @smallexample @c ada
4458 pragma Loop_Invariant ( boolean_EXPRESSION );
4462 The effect of this pragma is similar to that of pragma @code{Assert},
4463 except that in an @code{Assertion_Policy} pragma, the identifier
4464 @code{Loop_Invariant} is used to control whether it is ignored or checked
4467 @code{Loop_Invariant} can only appear as one of the items in the sequence
4468 of statements of a loop body, or nested inside block statements that
4469 appear in the sequence of statements of a loop body.
4470 The intention is that it be used to
4471 represent a "loop invariant" assertion, i.e. something that is true each
4472 time through the loop, and which can be used to show that the loop is
4473 achieving its purpose.
4475 Multiple @code{Loop_Invariant} and @code{Loop_Variant} pragmas that
4476 apply to the same loop should be grouped in the same sequence of
4479 To aid in writing such invariants, the special attribute @code{Loop_Entry}
4480 may be used to refer to the value of an expression on entry to the loop. This
4481 attribute can only be used within the expression of a @code{Loop_Invariant}
4482 pragma. For full details, see documentation of attribute @code{Loop_Entry}.
4484 @node Pragma Loop_Optimize
4485 @unnumberedsec Pragma Loop_Optimize
4486 @findex Loop_Optimize
4490 @smallexample @c ada
4491 pragma Loop_Optimize (OPTIMIZATION_HINT @{, OPTIMIZATION_HINT@});
4493 OPTIMIZATION_HINT ::= Ivdep | No_Unroll | Unroll | No_Vector | Vector
4497 This pragma must appear immediately within a loop statement. It allows the
4498 programmer to specify optimization hints for the enclosing loop. The hints
4499 are not mutually exclusive and can be freely mixed, but not all combinations
4500 will yield a sensible outcome.
4502 There are five supported optimization hints for a loop:
4507 The programmer asserts that there are no loop-carried dependencies
4508 which would prevent consecutive iterations of the loop from being
4509 executed simultaneously.
4513 The loop must not be unrolled. This is a strong hint: the compiler will not
4514 unroll a loop marked with this hint.
4518 The loop should be unrolled. This is a weak hint: the compiler will try to
4519 apply unrolling to this loop preferably to other optimizations, notably
4520 vectorization, but there is no guarantee that the loop will be unrolled.
4524 The loop must not be vectorized. This is a strong hint: the compiler will not
4525 vectorize a loop marked with this hint.
4529 The loop should be vectorized. This is a weak hint: the compiler will try to
4530 apply vectorization to this loop preferably to other optimizations, notably
4531 unrolling, but there is no guarantee that the loop will be vectorized.
4535 These hints do not remove the need to pass the appropriate switches to the
4536 compiler in order to enable the relevant optimizations, that is to say
4537 @option{-funroll-loops} for unrolling and @option{-ftree-vectorize} for
4540 @node Pragma Loop_Variant
4541 @unnumberedsec Pragma Loop_Variant
4542 @findex Loop_Variant
4546 @smallexample @c ada
4547 pragma Loop_Variant ( LOOP_VARIANT_ITEM @{, LOOP_VARIANT_ITEM @} );
4548 LOOP_VARIANT_ITEM ::= CHANGE_DIRECTION => discrete_EXPRESSION
4549 CHANGE_DIRECTION ::= Increases | Decreases
4553 @code{Loop_Variant} can only appear as one of the items in the sequence
4554 of statements of a loop body, or nested inside block statements that
4555 appear in the sequence of statements of a loop body.
4556 It allows the specification of quantities which must always
4557 decrease or increase in successive iterations of the loop. In its simplest
4558 form, just one expression is specified, whose value must increase or decrease
4559 on each iteration of the loop.
4561 In a more complex form, multiple arguments can be given which are intepreted
4562 in a nesting lexicographic manner. For example:
4564 @smallexample @c ada
4565 pragma Loop_Variant (Increases => X, Decreases => Y);
4569 specifies that each time through the loop either X increases, or X stays
4570 the same and Y decreases. A @code{Loop_Variant} pragma ensures that the
4571 loop is making progress. It can be useful in helping to show informally
4572 or prove formally that the loop always terminates.
4574 @code{Loop_Variant} is an assertion whose effect can be controlled using
4575 an @code{Assertion_Policy} with a check name of @code{Loop_Variant}. The
4576 policy can be @code{Check} to enable the loop variant check, @code{Ignore}
4577 to ignore the check (in which case the pragma has no effect on the program),
4578 or @code{Disable} in which case the pragma is not even checked for correct
4581 Multiple @code{Loop_Invariant} and @code{Loop_Variant} pragmas that
4582 apply to the same loop should be grouped in the same sequence of
4585 The @code{Loop_Entry} attribute may be used within the expressions of the
4586 @code{Loop_Variant} pragma to refer to values on entry to the loop.
4588 @node Pragma Machine_Attribute
4589 @unnumberedsec Pragma Machine_Attribute
4590 @findex Machine_Attribute
4594 @smallexample @c ada
4595 pragma Machine_Attribute (
4596 [Entity =>] LOCAL_NAME,
4597 [Attribute_Name =>] static_string_EXPRESSION
4598 [, [Info =>] static_EXPRESSION] );
4602 Machine-dependent attributes can be specified for types and/or
4603 declarations. This pragma is semantically equivalent to
4604 @code{__attribute__((@var{attribute_name}))} (if @var{info} is not
4605 specified) or @code{__attribute__((@var{attribute_name}(@var{info})))}
4606 in GNU C, where @code{@var{attribute_name}} is recognized by the
4607 compiler middle-end or the @code{TARGET_ATTRIBUTE_TABLE} machine
4608 specific macro. A string literal for the optional parameter @var{info}
4609 is transformed into an identifier, which may make this pragma unusable
4610 for some attributes. @xref{Target Attributes,, Defining target-specific
4611 uses of @code{__attribute__}, gccint, GNU Compiler Collection (GCC)
4612 Internals}, further information.
4615 @unnumberedsec Pragma Main
4621 @smallexample @c ada
4623 (MAIN_OPTION [, MAIN_OPTION]);
4626 [Stack_Size =>] static_integer_EXPRESSION
4627 | [Task_Stack_Size_Default =>] static_integer_EXPRESSION
4628 | [Time_Slicing_Enabled =>] static_boolean_EXPRESSION
4632 This pragma is provided for compatibility with OpenVMS VAX Systems. It has
4633 no effect in GNAT, other than being syntax checked.
4635 @node Pragma Main_Storage
4636 @unnumberedsec Pragma Main_Storage
4638 @findex Main_Storage
4642 @smallexample @c ada
4644 (MAIN_STORAGE_OPTION [, MAIN_STORAGE_OPTION]);
4646 MAIN_STORAGE_OPTION ::=
4647 [WORKING_STORAGE =>] static_SIMPLE_EXPRESSION
4648 | [TOP_GUARD =>] static_SIMPLE_EXPRESSION
4652 This pragma is provided for compatibility with OpenVMS VAX Systems. It has
4653 no effect in GNAT, other than being syntax checked. Note that the pragma
4654 also has no effect in DEC Ada 83 for OpenVMS Alpha Systems.
4656 @node Pragma No_Body
4657 @unnumberedsec Pragma No_Body
4662 @smallexample @c ada
4667 There are a number of cases in which a package spec does not require a body,
4668 and in fact a body is not permitted. GNAT will not permit the spec to be
4669 compiled if there is a body around. The pragma No_Body allows you to provide
4670 a body file, even in a case where no body is allowed. The body file must
4671 contain only comments and a single No_Body pragma. This is recognized by
4672 the compiler as indicating that no body is logically present.
4674 This is particularly useful during maintenance when a package is modified in
4675 such a way that a body needed before is no longer needed. The provision of a
4676 dummy body with a No_Body pragma ensures that there is no interference from
4677 earlier versions of the package body.
4679 @node Pragma No_Elaboration_Code_All
4680 @unnumberedsec Pragma No_Elaboration_Code_All
4681 @findex No_Elaboration_Code_All
4685 @smallexample @c ada
4686 pragma No_Elaboration_Code_All [(program_unit_NAME)];
4690 This is a program unit pragma (there is also an equivalent aspect of the
4691 same name) that establishes the restriction @code{No_Elaboration_Code} for
4692 the current unit and any extended main source units (body and subunits.
4693 It also has has the effect of enforcing a transitive application of this
4694 aspect, so that if any unit is implicitly or explicitly WITH'ed by the
4695 current unit, it must also have the No_Elaboration_Code_All aspect set.
4697 @node Pragma No_Inline
4698 @unnumberedsec Pragma No_Inline
4703 @smallexample @c ada
4704 pragma No_Inline (NAME @{, NAME@});
4708 This pragma suppresses inlining for the callable entity or the instances of
4709 the generic subprogram designated by @var{NAME}, including inlining that
4710 results from the use of pragma @code{Inline}. This pragma is always active,
4711 in particular it is not subject to the use of option @option{-gnatn} or
4712 @option{-gnatN}. It is illegal to specify both pragma @code{No_Inline} and
4713 pragma @code{Inline_Always} for the same @var{NAME}.
4715 @node Pragma No_Return
4716 @unnumberedsec Pragma No_Return
4721 @smallexample @c ada
4722 pragma No_Return (procedure_LOCAL_NAME @{, procedure_LOCAL_NAME@});
4726 Each @var{procedure_LOCAL_NAME} argument must refer to one or more procedure
4727 declarations in the current declarative part. A procedure to which this
4728 pragma is applied may not contain any explicit @code{return} statements.
4729 In addition, if the procedure contains any implicit returns from falling
4730 off the end of a statement sequence, then execution of that implicit
4731 return will cause Program_Error to be raised.
4733 One use of this pragma is to identify procedures whose only purpose is to raise
4734 an exception. Another use of this pragma is to suppress incorrect warnings
4735 about missing returns in functions, where the last statement of a function
4736 statement sequence is a call to such a procedure.
4738 Note that in Ada 2005 mode, this pragma is part of the language. It is
4739 available in all earlier versions of Ada as an implementation-defined
4742 @node Pragma No_Run_Time
4743 @unnumberedsec Pragma No_Run_Time
4748 @smallexample @c ada
4753 This is an obsolete configuration pragma that historically was used to
4754 set up a runtime library with no object code. It is now used only for
4755 internal testing. The pragma has been superseded by the reconfigurable
4756 runtime capability of @code{GNAT}.
4758 @node Pragma No_Strict_Aliasing
4759 @unnumberedsec Pragma No_Strict_Aliasing
4760 @findex No_Strict_Aliasing
4764 @smallexample @c ada
4765 pragma No_Strict_Aliasing [([Entity =>] type_LOCAL_NAME)];
4769 @var{type_LOCAL_NAME} must refer to an access type
4770 declaration in the current declarative part. The effect is to inhibit
4771 strict aliasing optimization for the given type. The form with no
4772 arguments is a configuration pragma which applies to all access types
4773 declared in units to which the pragma applies. For a detailed
4774 description of the strict aliasing optimization, and the situations
4775 in which it must be suppressed, see @ref{Optimization and Strict
4776 Aliasing,,, gnat_ugn, @value{EDITION} User's Guide}.
4778 This pragma currently has no effects on access to unconstrained array types.
4780 @node Pragma Normalize_Scalars
4781 @unnumberedsec Pragma Normalize_Scalars
4782 @findex Normalize_Scalars
4786 @smallexample @c ada
4787 pragma Normalize_Scalars;
4791 This is a language defined pragma which is fully implemented in GNAT@. The
4792 effect is to cause all scalar objects that are not otherwise initialized
4793 to be initialized. The initial values are implementation dependent and
4797 @item Standard.Character
4799 Objects whose root type is Standard.Character are initialized to
4800 Character'Last unless the subtype range excludes NUL (in which case
4801 NUL is used). This choice will always generate an invalid value if
4804 @item Standard.Wide_Character
4806 Objects whose root type is Standard.Wide_Character are initialized to
4807 Wide_Character'Last unless the subtype range excludes NUL (in which case
4808 NUL is used). This choice will always generate an invalid value if
4811 @item Standard.Wide_Wide_Character
4813 Objects whose root type is Standard.Wide_Wide_Character are initialized to
4814 the invalid value 16#FFFF_FFFF# unless the subtype range excludes NUL (in
4815 which case NUL is used). This choice will always generate an invalid value if
4820 Objects of an integer type are treated differently depending on whether
4821 negative values are present in the subtype. If no negative values are
4822 present, then all one bits is used as the initial value except in the
4823 special case where zero is excluded from the subtype, in which case
4824 all zero bits are used. This choice will always generate an invalid
4825 value if one exists.
4827 For subtypes with negative values present, the largest negative number
4828 is used, except in the unusual case where this largest negative number
4829 is in the subtype, and the largest positive number is not, in which case
4830 the largest positive value is used. This choice will always generate
4831 an invalid value if one exists.
4833 @item Floating-Point Types
4834 Objects of all floating-point types are initialized to all 1-bits. For
4835 standard IEEE format, this corresponds to a NaN (not a number) which is
4836 indeed an invalid value.
4838 @item Fixed-Point Types
4839 Objects of all fixed-point types are treated as described above for integers,
4840 with the rules applying to the underlying integer value used to represent
4841 the fixed-point value.
4844 Objects of a modular type are initialized to all one bits, except in
4845 the special case where zero is excluded from the subtype, in which
4846 case all zero bits are used. This choice will always generate an
4847 invalid value if one exists.
4849 @item Enumeration types
4850 Objects of an enumeration type are initialized to all one-bits, i.e.@: to
4851 the value @code{2 ** typ'Size - 1} unless the subtype excludes the literal
4852 whose Pos value is zero, in which case a code of zero is used. This choice
4853 will always generate an invalid value if one exists.
4857 @node Pragma Obsolescent
4858 @unnumberedsec Pragma Obsolescent
4863 @smallexample @c ada
4866 pragma Obsolescent (
4867 [Message =>] static_string_EXPRESSION
4868 [,[Version =>] Ada_05]]);
4870 pragma Obsolescent (
4872 [,[Message =>] static_string_EXPRESSION
4873 [,[Version =>] Ada_05]] );
4877 This pragma can occur immediately following a declaration of an entity,
4878 including the case of a record component. If no Entity argument is present,
4879 then this declaration is the one to which the pragma applies. If an Entity
4880 parameter is present, it must either match the name of the entity in this
4881 declaration, or alternatively, the pragma can immediately follow an enumeration
4882 type declaration, where the Entity argument names one of the enumeration
4885 This pragma is used to indicate that the named entity
4886 is considered obsolescent and should not be used. Typically this is
4887 used when an API must be modified by eventually removing or modifying
4888 existing subprograms or other entities. The pragma can be used at an
4889 intermediate stage when the entity is still present, but will be
4892 The effect of this pragma is to output a warning message on a reference to
4893 an entity thus marked that the subprogram is obsolescent if the appropriate
4894 warning option in the compiler is activated. If the Message parameter is
4895 present, then a second warning message is given containing this text. In
4896 addition, a reference to the entity is considered to be a violation of pragma
4897 Restrictions (No_Obsolescent_Features).
4899 This pragma can also be used as a program unit pragma for a package,
4900 in which case the entity name is the name of the package, and the
4901 pragma indicates that the entire package is considered
4902 obsolescent. In this case a client @code{with}'ing such a package
4903 violates the restriction, and the @code{with} statement is
4904 flagged with warnings if the warning option is set.
4906 If the Version parameter is present (which must be exactly
4907 the identifier Ada_05, no other argument is allowed), then the
4908 indication of obsolescence applies only when compiling in Ada 2005
4909 mode. This is primarily intended for dealing with the situations
4910 in the predefined library where subprograms or packages
4911 have become defined as obsolescent in Ada 2005
4912 (e.g.@: in Ada.Characters.Handling), but may be used anywhere.
4914 The following examples show typical uses of this pragma:
4916 @smallexample @c ada
4918 pragma Obsolescent (p, Message => "use pp instead of p");
4923 pragma Obsolescent ("use q2new instead");
4925 type R is new integer;
4928 Message => "use RR in Ada 2005",
4938 type E is (a, bc, 'd', quack);
4939 pragma Obsolescent (Entity => bc)
4940 pragma Obsolescent (Entity => 'd')
4943 (a, b : character) return character;
4944 pragma Obsolescent (Entity => "+");
4949 Note that, as for all pragmas, if you use a pragma argument identifier,
4950 then all subsequent parameters must also use a pragma argument identifier.
4951 So if you specify "Entity =>" for the Entity argument, and a Message
4952 argument is present, it must be preceded by "Message =>".
4954 @node Pragma Optimize_Alignment
4955 @unnumberedsec Pragma Optimize_Alignment
4956 @findex Optimize_Alignment
4957 @cindex Alignment, default settings
4961 @smallexample @c ada
4962 pragma Optimize_Alignment (TIME | SPACE | OFF);
4966 This is a configuration pragma which affects the choice of default alignments
4967 for types and objects where no alignment is explicitly specified. There is a
4968 time/space trade-off in the selection of these values. Large alignments result
4969 in more efficient code, at the expense of larger data space, since sizes have
4970 to be increased to match these alignments. Smaller alignments save space, but
4971 the access code is slower. The normal choice of default alignments for types
4972 and individual alignment promotions for objects (which is what you get if you
4973 do not use this pragma, or if you use an argument of OFF), tries to balance
4974 these two requirements.
4976 Specifying SPACE causes smaller default alignments to be chosen in two cases.
4977 First any packed record is given an alignment of 1. Second, if a size is given
4978 for the type, then the alignment is chosen to avoid increasing this size. For
4981 @smallexample @c ada
4991 In the default mode, this type gets an alignment of 4, so that access to the
4992 Integer field X are efficient. But this means that objects of the type end up
4993 with a size of 8 bytes. This is a valid choice, since sizes of objects are
4994 allowed to be bigger than the size of the type, but it can waste space if for
4995 example fields of type R appear in an enclosing record. If the above type is
4996 compiled in @code{Optimize_Alignment (Space)} mode, the alignment is set to 1.
4998 However, there is one case in which SPACE is ignored. If a variable length
4999 record (that is a discriminated record with a component which is an array
5000 whose length depends on a discriminant), has a pragma Pack, then it is not
5001 in general possible to set the alignment of such a record to one, so the
5002 pragma is ignored in this case (with a warning).
5004 Specifying SPACE also disables alignment promotions for standalone objects,
5005 which occur when the compiler increases the alignment of a specific object
5006 without changing the alignment of its type.
5008 Specifying TIME causes larger default alignments to be chosen in the case of
5009 small types with sizes that are not a power of 2. For example, consider:
5011 @smallexample @c ada
5023 The default alignment for this record is normally 1, but if this type is
5024 compiled in @code{Optimize_Alignment (Time)} mode, then the alignment is set
5025 to 4, which wastes space for objects of the type, since they are now 4 bytes
5026 long, but results in more efficient access when the whole record is referenced.
5028 As noted above, this is a configuration pragma, and there is a requirement
5029 that all units in a partition be compiled with a consistent setting of the
5030 optimization setting. This would normally be achieved by use of a configuration
5031 pragma file containing the appropriate setting. The exception to this rule is
5032 that units with an explicit configuration pragma in the same file as the source
5033 unit are excluded from the consistency check, as are all predefined units. The
5034 latter are compiled by default in pragma Optimize_Alignment (Off) mode if no
5035 pragma appears at the start of the file.
5037 @node Pragma Ordered
5038 @unnumberedsec Pragma Ordered
5040 @findex pragma @code{Ordered}
5044 @smallexample @c ada
5045 pragma Ordered (enumeration_first_subtype_LOCAL_NAME);
5049 Most enumeration types are from a conceptual point of view unordered.
5050 For example, consider:
5052 @smallexample @c ada
5053 type Color is (Red, Blue, Green, Yellow);
5057 By Ada semantics @code{Blue > Red} and @code{Green > Blue},
5058 but really these relations make no sense; the enumeration type merely
5059 specifies a set of possible colors, and the order is unimportant.
5061 For unordered enumeration types, it is generally a good idea if
5062 clients avoid comparisons (other than equality or inequality) and
5063 explicit ranges. (A @emph{client} is a unit where the type is referenced,
5064 other than the unit where the type is declared, its body, and its subunits.)
5065 For example, if code buried in some client says:
5067 @smallexample @c ada
5068 if Current_Color < Yellow then ...
5069 if Current_Color in Blue .. Green then ...
5073 then the client code is relying on the order, which is undesirable.
5074 It makes the code hard to read and creates maintenance difficulties if
5075 entries have to be added to the enumeration type. Instead,
5076 the code in the client should list the possibilities, or an
5077 appropriate subtype should be declared in the unit that declares
5078 the original enumeration type. E.g., the following subtype could
5079 be declared along with the type @code{Color}:
5081 @smallexample @c ada
5082 subtype RBG is Color range Red .. Green;
5086 and then the client could write:
5088 @smallexample @c ada
5089 if Current_Color in RBG then ...
5090 if Current_Color = Blue or Current_Color = Green then ...
5094 However, some enumeration types are legitimately ordered from a conceptual
5095 point of view. For example, if you declare:
5097 @smallexample @c ada
5098 type Day is (Mon, Tue, Wed, Thu, Fri, Sat, Sun);
5102 then the ordering imposed by the language is reasonable, and
5103 clients can depend on it, writing for example:
5105 @smallexample @c ada
5106 if D in Mon .. Fri then ...
5111 The pragma @option{Ordered} is provided to mark enumeration types that
5112 are conceptually ordered, alerting the reader that clients may depend
5113 on the ordering. GNAT provides a pragma to mark enumerations as ordered
5114 rather than one to mark them as unordered, since in our experience,
5115 the great majority of enumeration types are conceptually unordered.
5117 The types @code{Boolean}, @code{Character}, @code{Wide_Character},
5118 and @code{Wide_Wide_Character}
5119 are considered to be ordered types, so each is declared with a
5120 pragma @code{Ordered} in package @code{Standard}.
5122 Normally pragma @code{Ordered} serves only as documentation and a guide for
5123 coding standards, but GNAT provides a warning switch @option{-gnatw.u} that
5124 requests warnings for inappropriate uses (comparisons and explicit
5125 subranges) for unordered types. If this switch is used, then any
5126 enumeration type not marked with pragma @code{Ordered} will be considered
5127 as unordered, and will generate warnings for inappropriate uses.
5129 For additional information please refer to the description of the
5130 @option{-gnatw.u} switch in the @value{EDITION} User's Guide.
5132 @node Pragma Overflow_Mode
5133 @unnumberedsec Pragma Overflow_Mode
5134 @findex Overflow checks
5135 @findex Overflow mode
5136 @findex pragma @code{Overflow_Mode}
5140 @smallexample @c ada
5141 pragma Overflow_Mode
5143 [,[Assertions =>] MODE]);
5145 MODE ::= STRICT | MINIMIZED | ELIMINATED
5149 This pragma sets the current overflow mode to the given setting. For details
5150 of the meaning of these modes, please refer to the
5151 ``Overflow Check Handling in GNAT'' appendix in the
5152 @value{EDITION} User's Guide. If only the @code{General} parameter is present,
5153 the given mode applies to all expressions. If both parameters are present,
5154 the @code{General} mode applies to expressions outside assertions, and
5155 the @code{Eliminated} mode applies to expressions within assertions.
5157 The case of the @code{MODE} parameter is ignored,
5158 so @code{MINIMIZED}, @code{Minimized} and
5159 @code{minimized} all have the same effect.
5161 The @code{Overflow_Mode} pragma has the same scoping and placement
5162 rules as pragma @code{Suppress}, so it can occur either as a
5163 configuration pragma, specifying a default for the whole
5164 program, or in a declarative scope, where it applies to the
5165 remaining declarations and statements in that scope.
5167 The pragma @code{Suppress (Overflow_Check)} suppresses
5168 overflow checking, but does not affect the overflow mode.
5170 The pragma @code{Unsuppress (Overflow_Check)} unsuppresses (enables)
5171 overflow checking, but does not affect the overflow mode.
5173 @node Pragma Overriding_Renamings
5174 @unnumberedsec Pragma Overriding_Renamings
5175 @findex Overriding_Renamings
5176 @cindex Rational profile
5177 @cindex Rational compatibility
5181 @smallexample @c ada
5182 pragma Overriding_Renamings;
5186 This is a GNAT configuration pragma to simplify porting
5187 legacy code accepted by the Rational
5188 Ada compiler. In the presence of this pragma, a renaming declaration that
5189 renames an inherited operation declared in the same scope is legal if selected
5190 notation is used as in:
5192 @smallexample @c ada
5193 pragma Overriding_Renamings;
5198 function F (..) renames R.F;
5203 RM 8.3 (15) stipulates that an overridden operation is not visible within the
5204 declaration of the overriding operation.
5206 @node Pragma Partition_Elaboration_Policy
5207 @unnumberedsec Pragma Partition_Elaboration_Policy
5208 @findex Partition_Elaboration_Policy
5212 @smallexample @c ada
5213 pragma Partition_Elaboration_Policy (POLICY_IDENTIFIER);
5215 POLICY_IDENTIFIER ::= Concurrent | Sequential
5219 This pragma is standard in Ada 2005, but is available in all earlier
5220 versions of Ada as an implementation-defined pragma.
5221 See Ada 2012 Reference Manual for details.
5223 @node Pragma Part_Of
5224 @unnumberedsec Pragma Part_Of
5227 For the description of this pragma, see SPARK 2014 Reference Manual,
5230 @node Pragma Passive
5231 @unnumberedsec Pragma Passive
5236 @smallexample @c ada
5237 pragma Passive [(Semaphore | No)];
5241 Syntax checked, but otherwise ignored by GNAT@. This is recognized for
5242 compatibility with DEC Ada 83 implementations, where it is used within a
5243 task definition to request that a task be made passive. If the argument
5244 @code{Semaphore} is present, or the argument is omitted, then DEC Ada 83
5245 treats the pragma as an assertion that the containing task is passive
5246 and that optimization of context switch with this task is permitted and
5247 desired. If the argument @code{No} is present, the task must not be
5248 optimized. GNAT does not attempt to optimize any tasks in this manner
5249 (since protected objects are available in place of passive tasks).
5251 For more information on the subject of passive tasks, see the section
5252 ``Passive Task Optimization'' in the GNAT Users Guide.
5254 @node Pragma Persistent_BSS
5255 @unnumberedsec Pragma Persistent_BSS
5256 @findex Persistent_BSS
5260 @smallexample @c ada
5261 pragma Persistent_BSS [(LOCAL_NAME)]
5265 This pragma allows selected objects to be placed in the @code{.persistent_bss}
5266 section. On some targets the linker and loader provide for special
5267 treatment of this section, allowing a program to be reloaded without
5268 affecting the contents of this data (hence the name persistent).
5270 There are two forms of usage. If an argument is given, it must be the
5271 local name of a library level object, with no explicit initialization
5272 and whose type is potentially persistent. If no argument is given, then
5273 the pragma is a configuration pragma, and applies to all library level
5274 objects with no explicit initialization of potentially persistent types.
5276 A potentially persistent type is a scalar type, or an untagged,
5277 non-discriminated record, all of whose components have no explicit
5278 initialization and are themselves of a potentially persistent type,
5279 or an array, all of whose constraints are static, and whose component
5280 type is potentially persistent.
5282 If this pragma is used on a target where this feature is not supported,
5283 then the pragma will be ignored. See also @code{pragma Linker_Section}.
5285 @node Pragma Polling
5286 @unnumberedsec Pragma Polling
5291 @smallexample @c ada
5292 pragma Polling (ON | OFF);
5296 This pragma controls the generation of polling code. This is normally off.
5297 If @code{pragma Polling (ON)} is used then periodic calls are generated to
5298 the routine @code{Ada.Exceptions.Poll}. This routine is a separate unit in the
5299 runtime library, and can be found in file @file{a-excpol.adb}.
5301 Pragma @code{Polling} can appear as a configuration pragma (for example it
5302 can be placed in the @file{gnat.adc} file) to enable polling globally, or it
5303 can be used in the statement or declaration sequence to control polling
5306 A call to the polling routine is generated at the start of every loop and
5307 at the start of every subprogram call. This guarantees that the @code{Poll}
5308 routine is called frequently, and places an upper bound (determined by
5309 the complexity of the code) on the period between two @code{Poll} calls.
5311 The primary purpose of the polling interface is to enable asynchronous
5312 aborts on targets that cannot otherwise support it (for example Windows
5313 NT), but it may be used for any other purpose requiring periodic polling.
5314 The standard version is null, and can be replaced by a user program. This
5315 will require re-compilation of the @code{Ada.Exceptions} package that can
5316 be found in files @file{a-except.ads} and @file{a-except.adb}.
5318 A standard alternative unit (in file @file{4wexcpol.adb} in the standard GNAT
5319 distribution) is used to enable the asynchronous abort capability on
5320 targets that do not normally support the capability. The version of
5321 @code{Poll} in this file makes a call to the appropriate runtime routine
5322 to test for an abort condition.
5324 Note that polling can also be enabled by use of the @option{-gnatP} switch.
5325 @xref{Switches for gcc,,, gnat_ugn, @value{EDITION} User's Guide}, for
5329 @unnumberedsec Pragma Post
5331 @cindex Checks, postconditions
5332 @findex Postconditions
5336 @smallexample @c ada
5337 pragma Post (Boolean_Expression);
5341 The @code{Post} pragma is intended to be an exact replacement for
5342 the language-defined
5343 @code{Post} aspect, and shares its restrictions and semantics.
5344 It must appear either immediately following the corresponding
5345 subprogram declaration (only other pragmas may intervene), or
5346 if there is no separate subprogram declaration, then it can
5347 appear at the start of the declarations in a subprogram body
5348 (preceded only by other pragmas).
5350 @node Pragma Postcondition
5351 @unnumberedsec Pragma Postcondition
5352 @cindex Postcondition
5353 @cindex Checks, postconditions
5354 @findex Postconditions
5358 @smallexample @c ada
5359 pragma Postcondition (
5360 [Check =>] Boolean_Expression
5361 [,[Message =>] String_Expression]);
5365 The @code{Postcondition} pragma allows specification of automatic
5366 postcondition checks for subprograms. These checks are similar to
5367 assertions, but are automatically inserted just prior to the return
5368 statements of the subprogram with which they are associated (including
5369 implicit returns at the end of procedure bodies and associated
5370 exception handlers).
5372 In addition, the boolean expression which is the condition which
5373 must be true may contain references to function'Result in the case
5374 of a function to refer to the returned value.
5376 @code{Postcondition} pragmas may appear either immediately following the
5377 (separate) declaration of a subprogram, or at the start of the
5378 declarations of a subprogram body. Only other pragmas may intervene
5379 (that is appear between the subprogram declaration and its
5380 postconditions, or appear before the postcondition in the
5381 declaration sequence in a subprogram body). In the case of a
5382 postcondition appearing after a subprogram declaration, the
5383 formal arguments of the subprogram are visible, and can be
5384 referenced in the postcondition expressions.
5386 The postconditions are collected and automatically tested just
5387 before any return (implicit or explicit) in the subprogram body.
5388 A postcondition is only recognized if postconditions are active
5389 at the time the pragma is encountered. The compiler switch @option{gnata}
5390 turns on all postconditions by default, and pragma @code{Check_Policy}
5391 with an identifier of @code{Postcondition} can also be used to
5392 control whether postconditions are active.
5394 The general approach is that postconditions are placed in the spec
5395 if they represent functional aspects which make sense to the client.
5396 For example we might have:
5398 @smallexample @c ada
5399 function Direction return Integer;
5400 pragma Postcondition
5401 (Direction'Result = +1
5403 Direction'Result = -1);
5407 which serves to document that the result must be +1 or -1, and
5408 will test that this is the case at run time if postcondition
5411 Postconditions within the subprogram body can be used to
5412 check that some internal aspect of the implementation,
5413 not visible to the client, is operating as expected.
5414 For instance if a square root routine keeps an internal
5415 counter of the number of times it is called, then we
5416 might have the following postcondition:
5418 @smallexample @c ada
5419 Sqrt_Calls : Natural := 0;
5421 function Sqrt (Arg : Float) return Float is
5422 pragma Postcondition
5423 (Sqrt_Calls = Sqrt_Calls'Old + 1);
5429 As this example, shows, the use of the @code{Old} attribute
5430 is often useful in postconditions to refer to the state on
5431 entry to the subprogram.
5433 Note that postconditions are only checked on normal returns
5434 from the subprogram. If an abnormal return results from
5435 raising an exception, then the postconditions are not checked.
5437 If a postcondition fails, then the exception
5438 @code{System.Assertions.Assert_Failure} is raised. If
5439 a message argument was supplied, then the given string
5440 will be used as the exception message. If no message
5441 argument was supplied, then the default message has
5442 the form "Postcondition failed at file:line". The
5443 exception is raised in the context of the subprogram
5444 body, so it is possible to catch postcondition failures
5445 within the subprogram body itself.
5447 Within a package spec, normal visibility rules
5448 in Ada would prevent forward references within a
5449 postcondition pragma to functions defined later in
5450 the same package. This would introduce undesirable
5451 ordering constraints. To avoid this problem, all
5452 postcondition pragmas are analyzed at the end of
5453 the package spec, allowing forward references.
5455 The following example shows that this even allows
5456 mutually recursive postconditions as in:
5458 @smallexample @c ada
5459 package Parity_Functions is
5460 function Odd (X : Natural) return Boolean;
5461 pragma Postcondition
5465 (x /= 0 and then Even (X - 1))));
5467 function Even (X : Natural) return Boolean;
5468 pragma Postcondition
5472 (x /= 1 and then Odd (X - 1))));
5474 end Parity_Functions;
5478 There are no restrictions on the complexity or form of
5479 conditions used within @code{Postcondition} pragmas.
5480 The following example shows that it is even possible
5481 to verify performance behavior.
5483 @smallexample @c ada
5486 Performance : constant Float;
5487 -- Performance constant set by implementation
5488 -- to match target architecture behavior.
5490 procedure Treesort (Arg : String);
5491 -- Sorts characters of argument using N*logN sort
5492 pragma Postcondition
5493 (Float (Clock - Clock'Old) <=
5494 Float (Arg'Length) *
5495 log (Float (Arg'Length)) *
5501 Note: postcondition pragmas associated with subprograms that are
5502 marked as Inline_Always, or those marked as Inline with front-end
5503 inlining (-gnatN option set) are accepted and legality-checked
5504 by the compiler, but are ignored at run-time even if postcondition
5505 checking is enabled.
5507 Note that pragma @code{Postcondition} differs from the language-defined
5508 @code{Post} aspect (and corresponding @code{Post} pragma) in allowing
5509 multiple occurrences, allowing occurences in the body even if there
5510 is a separate spec, and allowing a second string parameter, and the
5511 use of the pragma identifier @code{Check}. Historically, pragma
5512 @code{Postcondition} was implemented prior to the development of
5513 Ada 2012, and has been retained in its original form for
5514 compatibility purposes.
5516 @node Pragma Post_Class
5517 @unnumberedsec Pragma Post_Class
5519 @cindex Checks, postconditions
5520 @findex Postconditions
5524 @smallexample @c ada
5525 pragma Post_Class (Boolean_Expression);
5529 The @code{Post_Class} pragma is intended to be an exact replacement for
5530 the language-defined
5531 @code{Post'Class} aspect, and shares its restrictions and semantics.
5532 It must appear either immediately following the corresponding
5533 subprogram declaration (only other pragmas may intervene), or
5534 if there is no separate subprogram declaration, then it can
5535 appear at the start of the declarations in a subprogram body
5536 (preceded only by other pragmas).
5538 Note: This pragma is called @code{Post_Class} rather than
5539 @code{Post'Class} because the latter would not be strictly
5540 conforming to the allowed syntax for pragmas. The motivation
5541 for provinding pragmas equivalent to the aspects is to allow a program
5542 to be written using the pragmas, and then compiled if necessary
5543 using an Ada compiler that does not recognize the pragmas or
5544 aspects, but is prepared to ignore the pragmas. The assertion
5545 policy that controls this pragma is @code{Post'Class}, not
5549 @unnumberedsec Pragma Pre
5551 @cindex Checks, preconditions
5552 @findex Preconditions
5556 @smallexample @c ada
5557 pragma Pre (Boolean_Expression);
5561 The @code{Pre} pragma is intended to be an exact replacement for
5562 the language-defined
5563 @code{Pre} aspect, and shares its restrictions and semantics.
5564 It must appear either immediately following the corresponding
5565 subprogram declaration (only other pragmas may intervene), or
5566 if there is no separate subprogram declaration, then it can
5567 appear at the start of the declarations in a subprogram body
5568 (preceded only by other pragmas).
5570 @node Pragma Precondition
5571 @unnumberedsec Pragma Precondition
5572 @cindex Preconditions
5573 @cindex Checks, preconditions
5574 @findex Preconditions
5578 @smallexample @c ada
5579 pragma Precondition (
5580 [Check =>] Boolean_Expression
5581 [,[Message =>] String_Expression]);
5585 The @code{Precondition} pragma is similar to @code{Postcondition}
5586 except that the corresponding checks take place immediately upon
5587 entry to the subprogram, and if a precondition fails, the exception
5588 is raised in the context of the caller, and the attribute 'Result
5589 cannot be used within the precondition expression.
5591 Otherwise, the placement and visibility rules are identical to those
5592 described for postconditions. The following is an example of use
5593 within a package spec:
5595 @smallexample @c ada
5596 package Math_Functions is
5598 function Sqrt (Arg : Float) return Float;
5599 pragma Precondition (Arg >= 0.0)
5605 @code{Precondition} pragmas may appear either immediately following the
5606 (separate) declaration of a subprogram, or at the start of the
5607 declarations of a subprogram body. Only other pragmas may intervene
5608 (that is appear between the subprogram declaration and its
5609 postconditions, or appear before the postcondition in the
5610 declaration sequence in a subprogram body).
5612 Note: precondition pragmas associated with subprograms that are
5613 marked as Inline_Always, or those marked as Inline with front-end
5614 inlining (-gnatN option set) are accepted and legality-checked
5615 by the compiler, but are ignored at run-time even if precondition
5616 checking is enabled.
5618 Note that pragma @code{Precondition} differs from the language-defined
5619 @code{Pre} aspect (and corresponding @code{Pre} pragma) in allowing
5620 multiple occurrences, allowing occurences in the body even if there
5621 is a separate spec, and allowing a second string parameter, and the
5622 use of the pragma identifier @code{Check}. Historically, pragma
5623 @code{Precondition} was implemented prior to the development of
5624 Ada 2012, and has been retained in its original form for
5625 compatibility purposes.
5627 @node Pragma Predicate
5628 @unnumberedsec Pragma Predicate
5630 @findex Predicate pragma
5634 @smallexample @c ada
5636 ([Entity =>] type_LOCAL_NAME,
5637 [Check =>] EXPRESSION);
5641 This pragma (available in all versions of Ada in GNAT) encompasses both
5642 the @code{Static_Predicate} and @code{Dynamic_Predicate} aspects in
5643 Ada 2012. A predicate is regarded as static if it has an allowed form
5644 for @code{Static_Predicate} and is otherwise treated as a
5645 @code{Dynamic_Predicate}. Otherwise, predicates specified by this
5646 pragma behave exactly as described in the Ada 2012 reference manual.
5647 For example, if we have
5649 @smallexample @c ada
5650 type R is range 1 .. 10;
5652 pragma Predicate (Entity => S, Check => S not in 4 .. 6);
5654 pragma Predicate (Entity => Q, Check => F(Q) or G(Q));
5658 the effect is identical to the following Ada 2012 code:
5660 @smallexample @c ada
5661 type R is range 1 .. 10;
5663 Static_Predicate => S not in 4 .. 6;
5665 Dynamic_Predicate => F(Q) or G(Q);
5668 Note that there is are no pragmas @code{Dynamic_Predicate}
5669 or @code{Static_Predicate}. That is
5670 because these pragmas would affect legality and semantics of
5671 the program and thus do not have a neutral effect if ignored.
5672 The motivation behind providing pragmas equivalent to
5673 corresponding aspects is to allow a program to be written
5674 using the pragmas, and then compiled with a compiler that
5675 will ignore the pragmas. That doesn't work in the case of
5676 static and dynamic predicates, since if the corresponding
5677 pragmas are ignored, then the behavior of the program is
5678 fundamentally changed (for example a membership test
5679 @code{A in B} would not take into account a predicate
5680 defined for subtype B). When following this approach, the
5681 use of predicates should be avoided.
5683 @node Pragma Preelaborable_Initialization
5684 @unnumberedsec Pragma Preelaborable_Initialization
5685 @findex Preelaborable_Initialization
5689 @smallexample @c ada
5690 pragma Preelaborable_Initialization (DIRECT_NAME);
5694 This pragma is standard in Ada 2005, but is available in all earlier
5695 versions of Ada as an implementation-defined pragma.
5696 See Ada 2012 Reference Manual for details.
5698 @node Pragma Prefix_Exception_Messages
5699 @unnumberedsec Pragma Prefix_Exception_Messages
5700 @cindex Prefix_Exception_Messages
5702 @cindex Exception_Message
5707 @smallexample @c ada
5708 pragma Prefix_Exception_Messages;
5712 This is an implementation-defined configuration pragma that affects the
5713 behavior of raise statements with a message given as a static string
5714 constant (typically a string literal). In such cases, the string will
5715 be automatically prefixed by the name of the enclosing entity (giving
5716 the package and subprogram containing the raise statement). This helps
5717 to identify where messages are coming from, and this mode is automatic
5718 for the run-time library.
5720 The pragma has no effect if the message is computed with an expression other
5721 than a static string constant, since the assumption in this case is that
5722 the program computes exactly the string it wants. If you still want the
5723 prefixing in this case, you can always call
5724 @code{GNAT.Source_Info.Enclosing_Entity} and prepend the string manually.
5726 @node Pragma Pre_Class
5727 @unnumberedsec Pragma Pre_Class
5729 @cindex Checks, preconditions
5730 @findex Preconditions
5734 @smallexample @c ada
5735 pragma Pre_Class (Boolean_Expression);
5739 The @code{Pre_Class} pragma is intended to be an exact replacement for
5740 the language-defined
5741 @code{Pre'Class} aspect, and shares its restrictions and semantics.
5742 It must appear either immediately following the corresponding
5743 subprogram declaration (only other pragmas may intervene), or
5744 if there is no separate subprogram declaration, then it can
5745 appear at the start of the declarations in a subprogram body
5746 (preceded only by other pragmas).
5748 Note: This pragma is called @code{Pre_Class} rather than
5749 @code{Pre'Class} because the latter would not be strictly
5750 conforming to the allowed syntax for pragmas. The motivation
5751 for providing pragmas equivalent to the aspects is to allow a program
5752 to be written using the pragmas, and then compiled if necessary
5753 using an Ada compiler that does not recognize the pragmas or
5754 aspects, but is prepared to ignore the pragmas. The assertion
5755 policy that controls this pragma is @code{Pre'Class}, not
5758 @node Pragma Priority_Specific_Dispatching
5759 @unnumberedsec Pragma Priority_Specific_Dispatching
5760 @findex Priority_Specific_Dispatching
5764 @smallexample @c ada
5765 pragma Priority_Specific_Dispatching (
5767 first_priority_EXPRESSION,
5768 last_priority_EXPRESSION)
5770 POLICY_IDENTIFIER ::=
5771 EDF_Across_Priorities |
5772 FIFO_Within_Priorities |
5773 Non_Preemptive_Within_Priorities |
5774 Round_Robin_Within_Priorities
5778 This pragma is standard in Ada 2005, but is available in all earlier
5779 versions of Ada as an implementation-defined pragma.
5780 See Ada 2012 Reference Manual for details.
5782 @node Pragma Profile
5783 @unnumberedsec Pragma Profile
5788 @smallexample @c ada
5789 pragma Profile (Ravenscar | Restricted | Rational);
5793 This pragma is standard in Ada 2005, but is available in all earlier
5794 versions of Ada as an implementation-defined pragma. This is a
5795 configuration pragma that establishes a set of configiuration pragmas
5796 that depend on the argument. @code{Ravenscar} is standard in Ada 2005.
5797 The other two possibilities (@code{Restricted} or @code{Rational})
5798 are implementation-defined. The set of configuration pragmas
5799 is defined in the following sections.
5803 @item Pragma Profile (Ravenscar)
5807 The @code{Ravenscar} profile is standard in Ada 2005,
5808 but is available in all earlier
5809 versions of Ada as an implementation-defined pragma. This profile
5810 establishes the following set of configuration pragmas:
5813 @item Task_Dispatching_Policy (FIFO_Within_Priorities)
5814 [RM D.2.2] Tasks are dispatched following a preemptive
5815 priority-ordered scheduling policy.
5817 @item Locking_Policy (Ceiling_Locking)
5818 [RM D.3] While tasks and interrupts execute a protected action, they inherit
5819 the ceiling priority of the corresponding protected object.
5821 @item Detect_Blocking
5822 This pragma forces the detection of potentially blocking operations within a
5823 protected operation, and to raise Program_Error if that happens.
5827 plus the following set of restrictions:
5830 @item Max_Entry_Queue_Length => 1
5831 No task can be queued on a protected entry.
5832 @item Max_Protected_Entries => 1
5833 @item Max_Task_Entries => 0
5834 No rendezvous statements are allowed.
5835 @item No_Abort_Statements
5836 @item No_Dynamic_Attachment
5837 @item No_Dynamic_Priorities
5838 @item No_Implicit_Heap_Allocations
5839 @item No_Local_Protected_Objects
5840 @item No_Local_Timing_Events
5841 @item No_Protected_Type_Allocators
5842 @item No_Relative_Delay
5843 @item No_Requeue_Statements
5844 @item No_Select_Statements
5845 @item No_Specific_Termination_Handlers
5846 @item No_Task_Allocators
5847 @item No_Task_Hierarchy
5848 @item No_Task_Termination
5849 @item Simple_Barriers
5853 The Ravenscar profile also includes the following restrictions that specify
5854 that there are no semantic dependences on the corresponding predefined
5858 @item No_Dependence => Ada.Asynchronous_Task_Control
5859 @item No_Dependence => Ada.Calendar
5860 @item No_Dependence => Ada.Execution_Time.Group_Budget
5861 @item No_Dependence => Ada.Execution_Time.Timers
5862 @item No_Dependence => Ada.Task_Attributes
5863 @item No_Dependence => System.Multiprocessors.Dispatching_Domains
5868 This set of configuration pragmas and restrictions correspond to the
5869 definition of the ``Ravenscar Profile'' for limited tasking, devised and
5870 published by the @cite{International Real-Time Ada Workshop}, 1997,
5871 and whose most recent description is available at
5872 @url{http://www-users.cs.york.ac.uk/~burns/ravenscar.ps}.
5874 The original definition of the profile was revised at subsequent IRTAW
5875 meetings. It has been included in the ISO
5876 @cite{Guide for the Use of the Ada Programming Language in High
5877 Integrity Systems}, and has been approved by ISO/IEC/SC22/WG9 for inclusion in
5878 the next revision of the standard. The formal definition given by
5879 the Ada Rapporteur Group (ARG) can be found in two Ada Issues (AI-249 and
5880 AI-305) available at
5881 @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/ais/ai-00249.txt} and
5882 @url{http://www.ada-auth.org/cgi-bin/cvsweb.cgi/ais/ai-00305.txt}.
5884 The above set is a superset of the restrictions provided by pragma
5885 @code{Profile (Restricted)}, it includes six additional restrictions
5886 (@code{Simple_Barriers}, @code{No_Select_Statements},
5887 @code{No_Calendar}, @code{No_Implicit_Heap_Allocations},
5888 @code{No_Relative_Delay} and @code{No_Task_Termination}). This means
5889 that pragma @code{Profile (Ravenscar)}, like the pragma
5890 @code{Profile (Restricted)},
5891 automatically causes the use of a simplified,
5892 more efficient version of the tasking run-time system.
5894 @item Pragma Profile (Restricted)
5895 @findex Restricted Run Time
5897 This profile corresponds to the GNAT restricted run time. It
5898 establishes the following set of restrictions:
5901 @item No_Abort_Statements
5902 @item No_Entry_Queue
5903 @item No_Task_Hierarchy
5904 @item No_Task_Allocators
5905 @item No_Dynamic_Priorities
5906 @item No_Terminate_Alternatives
5907 @item No_Dynamic_Attachment
5908 @item No_Protected_Type_Allocators
5909 @item No_Local_Protected_Objects
5910 @item No_Requeue_Statements
5911 @item No_Task_Attributes_Package
5912 @item Max_Asynchronous_Select_Nesting = 0
5913 @item Max_Task_Entries = 0
5914 @item Max_Protected_Entries = 1
5915 @item Max_Select_Alternatives = 0
5919 This set of restrictions causes the automatic selection of a simplified
5920 version of the run time that provides improved performance for the
5921 limited set of tasking functionality permitted by this set of restrictions.
5923 @item Pragma Profile (Rational)
5924 @findex Rational compatibility mode
5926 The Rational profile is intended to facilitate porting legacy code that
5927 compiles with the Rational APEX compiler, even when the code includes non-
5928 conforming Ada constructs. The profile enables the following three pragmas:
5931 @item pragma Implicit_Packing
5932 @item pragma Overriding_Renamings
5933 @item pragma Use_VADS_Size
5938 @node Pragma Profile_Warnings
5939 @unnumberedsec Pragma Profile_Warnings
5940 @findex Profile_Warnings
5944 @smallexample @c ada
5945 pragma Profile_Warnings (Ravenscar | Restricted | Rational);
5949 This is an implementation-defined pragma that is similar in
5950 effect to @code{pragma Profile} except that instead of
5951 generating @code{Restrictions} pragmas, it generates
5952 @code{Restriction_Warnings} pragmas. The result is that
5953 violations of the profile generate warning messages instead
5956 @node Pragma Propagate_Exceptions
5957 @unnumberedsec Pragma Propagate_Exceptions
5958 @cindex Interfacing to C++
5959 @findex Propagate_Exceptions
5963 @smallexample @c ada
5964 pragma Propagate_Exceptions;
5968 This pragma is now obsolete and, other than generating a warning if warnings
5969 on obsolescent features are enabled, is ignored.
5970 It is retained for compatibility
5971 purposes. It used to be used in connection with optimization of
5972 a now-obsolete mechanism for implementation of exceptions.
5974 @node Pragma Provide_Shift_Operators
5975 @unnumberedsec Pragma Provide_Shift_Operators
5976 @cindex Shift operators
5977 @findex Provide_Shift_Operators
5981 @smallexample @c ada
5982 pragma Provide_Shift_Operators (integer_first_subtype_LOCAL_NAME);
5986 This pragma can be applied to a first subtype local name that specifies
5987 either an unsigned or signed type. It has the effect of providing the
5988 five shift operators (Shift_Left, Shift_Right, Shift_Right_Arithmetic,
5989 Rotate_Left and Rotate_Right) for the given type. It is similar to
5990 including the function declarations for these five operators, together
5991 with the pragma Import (Intrinsic, ...) statements.
5993 @node Pragma Psect_Object
5994 @unnumberedsec Pragma Psect_Object
5995 @findex Psect_Object
5999 @smallexample @c ada
6000 pragma Psect_Object (
6001 [Internal =>] LOCAL_NAME,
6002 [, [External =>] EXTERNAL_SYMBOL]
6003 [, [Size =>] EXTERNAL_SYMBOL]);
6007 | static_string_EXPRESSION
6011 This pragma is identical in effect to pragma @code{Common_Object}.
6013 @node Pragma Pure_Function
6014 @unnumberedsec Pragma Pure_Function
6015 @findex Pure_Function
6019 @smallexample @c ada
6020 pragma Pure_Function ([Entity =>] function_LOCAL_NAME);
6024 This pragma appears in the same declarative part as a function
6025 declaration (or a set of function declarations if more than one
6026 overloaded declaration exists, in which case the pragma applies
6027 to all entities). It specifies that the function @code{Entity} is
6028 to be considered pure for the purposes of code generation. This means
6029 that the compiler can assume that there are no side effects, and
6030 in particular that two calls with identical arguments produce the
6031 same result. It also means that the function can be used in an
6034 Note that, quite deliberately, there are no static checks to try
6035 to ensure that this promise is met, so @code{Pure_Function} can be used
6036 with functions that are conceptually pure, even if they do modify
6037 global variables. For example, a square root function that is
6038 instrumented to count the number of times it is called is still
6039 conceptually pure, and can still be optimized, even though it
6040 modifies a global variable (the count). Memo functions are another
6041 example (where a table of previous calls is kept and consulted to
6042 avoid re-computation).
6044 Note also that the normal rules excluding optimization of subprograms
6045 in pure units (when parameter types are descended from System.Address,
6046 or when the full view of a parameter type is limited), do not apply
6047 for the Pure_Function case. If you explicitly specify Pure_Function,
6048 the compiler may optimize away calls with identical arguments, and
6049 if that results in unexpected behavior, the proper action is not to
6050 use the pragma for subprograms that are not (conceptually) pure.
6053 Note: Most functions in a @code{Pure} package are automatically pure, and
6054 there is no need to use pragma @code{Pure_Function} for such functions. One
6055 exception is any function that has at least one formal of type
6056 @code{System.Address} or a type derived from it. Such functions are not
6057 considered pure by default, since the compiler assumes that the
6058 @code{Address} parameter may be functioning as a pointer and that the
6059 referenced data may change even if the address value does not.
6060 Similarly, imported functions are not considered to be pure by default,
6061 since there is no way of checking that they are in fact pure. The use
6062 of pragma @code{Pure_Function} for such a function will override these default
6063 assumption, and cause the compiler to treat a designated subprogram as pure
6066 Note: If pragma @code{Pure_Function} is applied to a renamed function, it
6067 applies to the underlying renamed function. This can be used to
6068 disambiguate cases of overloading where some but not all functions
6069 in a set of overloaded functions are to be designated as pure.
6071 If pragma @code{Pure_Function} is applied to a library level function, the
6072 function is also considered pure from an optimization point of view, but the
6073 unit is not a Pure unit in the categorization sense. So for example, a function
6074 thus marked is free to @code{with} non-pure units.
6076 @node Pragma Rational
6077 @unnumberedsec Pragma Rational
6082 @smallexample @c ada
6087 This pragma is considered obsolescent, but is retained for
6088 compatibility purposes. It is equivalent to:
6090 @smallexample @c ada
6091 pragma Profile (Rational);
6094 @node Pragma Ravenscar
6095 @unnumberedsec Pragma Ravenscar
6096 @findex Pragma Ravenscar
6100 @smallexample @c ada
6105 This pragma is considered obsolescent, but is retained for
6106 compatibility purposes. It is equivalent to:
6108 @smallexample @c ada
6109 pragma Profile (Ravenscar);
6113 which is the preferred method of setting the @code{Ravenscar} profile.
6115 @node Pragma Refined_Depends
6116 @unnumberedsec Pragma Refined_Depends
6117 @findex Refined_Depends
6119 For the description of this pragma, see SPARK 2014 Reference Manual,
6122 @node Pragma Refined_Global
6123 @unnumberedsec Pragma Refined_Global
6124 @findex Refined_Global
6126 For the description of this pragma, see SPARK 2014 Reference Manual,
6129 @node Pragma Refined_Post
6130 @unnumberedsec Pragma Refined_Post
6131 @findex Refined_Post
6133 For the description of this pragma, see SPARK 2014 Reference Manual,
6136 @node Pragma Refined_State
6137 @unnumberedsec Pragma Refined_State
6138 @findex Refined_State
6140 For the description of this pragma, see SPARK 2014 Reference Manual,
6143 @node Pragma Relative_Deadline
6144 @unnumberedsec Pragma Relative_Deadline
6145 @findex Relative_Deadline
6149 @smallexample @c ada
6150 pragma Relative_Deadline (time_span_EXPRESSION);
6154 This pragma is standard in Ada 2005, but is available in all earlier
6155 versions of Ada as an implementation-defined pragma.
6156 See Ada 2012 Reference Manual for details.
6158 @node Pragma Remote_Access_Type
6159 @unnumberedsec Pragma Remote_Access_Type
6160 @findex Remote_Access_Type
6164 @smallexample @c ada
6165 pragma Remote_Access_Type ([Entity =>] formal_access_type_LOCAL_NAME);
6169 This pragma appears in the formal part of a generic declaration.
6170 It specifies an exception to the RM rule from E.2.2(17/2), which forbids
6171 the use of a remote access to class-wide type as actual for a formal
6174 When this pragma applies to a formal access type @code{Entity}, that
6175 type is treated as a remote access to class-wide type in the generic.
6176 It must be a formal general access type, and its designated type must
6177 be the class-wide type of a formal tagged limited private type from the
6178 same generic declaration.
6180 In the generic unit, the formal type is subject to all restrictions
6181 pertaining to remote access to class-wide types. At instantiation, the
6182 actual type must be a remote access to class-wide type.
6184 @node Pragma Restricted_Run_Time
6185 @unnumberedsec Pragma Restricted_Run_Time
6186 @findex Pragma Restricted_Run_Time
6190 @smallexample @c ada
6191 pragma Restricted_Run_Time;
6195 This pragma is considered obsolescent, but is retained for
6196 compatibility purposes. It is equivalent to:
6198 @smallexample @c ada
6199 pragma Profile (Restricted);
6203 which is the preferred method of setting the restricted run time
6206 @node Pragma Restriction_Warnings
6207 @unnumberedsec Pragma Restriction_Warnings
6208 @findex Restriction_Warnings
6212 @smallexample @c ada
6213 pragma Restriction_Warnings
6214 (restriction_IDENTIFIER @{, restriction_IDENTIFIER@});
6218 This pragma allows a series of restriction identifiers to be
6219 specified (the list of allowed identifiers is the same as for
6220 pragma @code{Restrictions}). For each of these identifiers
6221 the compiler checks for violations of the restriction, but
6222 generates a warning message rather than an error message
6223 if the restriction is violated.
6225 One use of this is in situations where you want to know
6226 about violations of a restriction, but you want to ignore some of
6227 these violations. Consider this example, where you want to set
6228 Ada_95 mode and enable style checks, but you want to know about
6229 any other use of implementation pragmas:
6231 @smallexample @c ada
6232 pragma Restriction_Warnings (No_Implementation_Pragmas);
6233 7 (Off, "violation of*No_Implementation_Pragmas*");
6235 pragma Style_Checks ("2bfhkM160");
6236 pragma Warnings (On, "violation of*No_Implementation_Pragmas*");
6240 By including the above lines in a configuration pragmas file,
6241 the Ada_95 and Style_Checks pragmas are accepted without
6242 generating a warning, but any other use of implementation
6243 defined pragmas will cause a warning to be generated.
6245 @node Pragma Reviewable
6246 @unnumberedsec Pragma Reviewable
6251 @smallexample @c ada
6256 This pragma is an RM-defined standard pragma, but has no effect on the
6257 program being compiled, or on the code generated for the program.
6259 To obtain the required output specified in RM H.3.1, the compiler must be
6260 run with various special switches as follows:
6264 @item Where compiler-generated run-time checks remain
6266 The switch @option{-gnatGL}
6267 @findex @option{-gnatGL}
6268 may be used to list the expanded code in pseudo-Ada form.
6269 Runtime checks show up in the listing either as explicit
6270 checks or operators marked with @{@} to indicate a check is present.
6272 @item An identification of known exceptions at compile time
6274 If the program is compiled with @option{-gnatwa},
6275 @findex @option{-gnatwa}
6276 the compiler warning messages will indicate all cases where the compiler
6277 detects that an exception is certain to occur at run time.
6279 @item Possible reads of uninitialized variables
6281 The compiler warns of many such cases, but its output is incomplete.
6283 The CodePeer analysis tool
6284 @findex CodePeer static analysis tool
6287 A supplemental static analysis tool
6289 may be used to obtain a comprehensive list of all
6290 possible points at which uninitialized data may be read.
6292 @item Where run-time support routines are implicitly invoked
6294 In the output from @option{-gnatGL},
6295 @findex @option{-gnatGL}
6296 run-time calls are explicitly listed as calls to the relevant
6299 @item Object code listing
6301 This may be obtained either by using the @option{-S} switch,
6303 or the objdump utility.
6306 @item Constructs known to be erroneous at compile time
6308 These are identified by warnings issued by the compiler (use @option{-gnatwa}).
6309 @findex @option{-gnatwa}
6311 @item Stack usage information
6313 Static stack usage data (maximum per-subprogram) can be obtained via the
6314 @option{-fstack-usage} switch to the compiler.
6315 @findex @option{-fstack-usage}
6316 Dynamic stack usage data (per task) can be obtained via the @option{-u} switch
6320 The gnatstack utility
6322 can be used to provide additional information on stack usage.
6325 @item Object code listing of entire partition
6327 This can be obtained by compiling the partition with @option{-S},
6329 or by applying objdump
6331 to all the object files that are part of the partition.
6333 @item A description of the run-time model
6335 The full sources of the run-time are available, and the documentation of
6336 these routines describes how these run-time routines interface to the
6337 underlying operating system facilities.
6339 @item Control and data-flow information
6343 @findex CodePeer static analysis tool
6346 A supplemental static analysis tool
6348 may be used to obtain complete control and data-flow information, as well as
6349 comprehensive messages identifying possible problems based on this
6353 @node Pragma Share_Generic
6354 @unnumberedsec Pragma Share_Generic
6355 @findex Share_Generic
6359 @smallexample @c ada
6360 pragma Share_Generic (GNAME @{, GNAME@});
6362 GNAME ::= generic_unit_NAME | generic_instance_NAME
6366 This pragma is provided for compatibility with Dec Ada 83. It has
6367 no effect in @code{GNAT} (which does not implement shared generics), other
6368 than to check that the given names are all names of generic units or
6372 @unnumberedsec Pragma Shared
6376 This pragma is provided for compatibility with Ada 83. The syntax and
6377 semantics are identical to pragma Atomic.
6379 @node Pragma Short_Circuit_And_Or
6380 @unnumberedsec Pragma Short_Circuit_And_Or
6381 @findex Short_Circuit_And_Or
6385 @smallexample @c ada
6386 pragma Short_Circuit_And_Or;
6390 This configuration pragma causes any occurrence of the AND operator applied to
6391 operands of type Standard.Boolean to be short-circuited (i.e. the AND operator
6392 is treated as if it were AND THEN). Or is similarly treated as OR ELSE. This
6393 may be useful in the context of certification protocols requiring the use of
6394 short-circuited logical operators. If this configuration pragma occurs locally
6395 within the file being compiled, it applies only to the file being compiled.
6396 There is no requirement that all units in a partition use this option.
6398 @node Pragma Short_Descriptors
6399 @unnumberedsec Pragma Short_Descriptors
6400 @findex Short_Descriptors
6404 @smallexample @c ada
6405 pragma Short_Descriptors
6409 This pragma is provided for compatibility with other Ada implementations. It
6410 is recognized but ignored by all current versions of GNAT.
6412 @node Pragma Simple_Storage_Pool_Type
6413 @unnumberedsec Pragma Simple_Storage_Pool_Type
6414 @findex Simple_Storage_Pool_Type
6415 @cindex Storage pool, simple
6416 @cindex Simple storage pool
6420 @smallexample @c ada
6421 pragma Simple_Storage_Pool_Type (type_LOCAL_NAME);
6425 A type can be established as a ``simple storage pool type'' by applying
6426 the representation pragma @code{Simple_Storage_Pool_Type} to the type.
6427 A type named in the pragma must be a library-level immutably limited record
6428 type or limited tagged type declared immediately within a package declaration.
6429 The type can also be a limited private type whose full type is allowed as
6430 a simple storage pool type.
6432 For a simple storage pool type @var{SSP}, nonabstract primitive subprograms
6433 @code{Allocate}, @code{Deallocate}, and @code{Storage_Size} can be declared that
6434 are subtype conformant with the following subprogram declarations:
6436 @smallexample @c ada
6439 Storage_Address : out System.Address;
6440 Size_In_Storage_Elements : System.Storage_Elements.Storage_Count;
6441 Alignment : System.Storage_Elements.Storage_Count);
6443 procedure Deallocate
6445 Storage_Address : System.Address;
6446 Size_In_Storage_Elements : System.Storage_Elements.Storage_Count;
6447 Alignment : System.Storage_Elements.Storage_Count);
6449 function Storage_Size (Pool : SSP)
6450 return System.Storage_Elements.Storage_Count;
6454 Procedure @code{Allocate} must be declared, whereas @code{Deallocate} and
6455 @code{Storage_Size} are optional. If @code{Deallocate} is not declared, then
6456 applying an unchecked deallocation has no effect other than to set its actual
6457 parameter to null. If @code{Storage_Size} is not declared, then the
6458 @code{Storage_Size} attribute applied to an access type associated with
6459 a pool object of type SSP returns zero. Additional operations can be declared
6460 for a simple storage pool type (such as for supporting a mark/release
6461 storage-management discipline).
6463 An object of a simple storage pool type can be associated with an access
6464 type by specifying the attribute @code{Simple_Storage_Pool}. For example:
6466 @smallexample @c ada
6468 My_Pool : My_Simple_Storage_Pool_Type;
6470 type Acc is access My_Data_Type;
6472 for Acc'Simple_Storage_Pool use My_Pool;
6477 See attribute @code{Simple_Storage_Pool} for further details.
6479 @node Pragma Source_File_Name
6480 @unnumberedsec Pragma Source_File_Name
6481 @findex Source_File_Name
6485 @smallexample @c ada
6486 pragma Source_File_Name (
6487 [Unit_Name =>] unit_NAME,
6488 Spec_File_Name => STRING_LITERAL,
6489 [Index => INTEGER_LITERAL]);
6491 pragma Source_File_Name (
6492 [Unit_Name =>] unit_NAME,
6493 Body_File_Name => STRING_LITERAL,
6494 [Index => INTEGER_LITERAL]);
6498 Use this to override the normal naming convention. It is a configuration
6499 pragma, and so has the usual applicability of configuration pragmas
6500 (i.e.@: it applies to either an entire partition, or to all units in a
6501 compilation, or to a single unit, depending on how it is used.
6502 @var{unit_name} is mapped to @var{file_name_literal}. The identifier for
6503 the second argument is required, and indicates whether this is the file
6504 name for the spec or for the body.
6506 The optional Index argument should be used when a file contains multiple
6507 units, and when you do not want to use @code{gnatchop} to separate then
6508 into multiple files (which is the recommended procedure to limit the
6509 number of recompilations that are needed when some sources change).
6510 For instance, if the source file @file{source.ada} contains
6512 @smallexample @c ada
6524 you could use the following configuration pragmas:
6526 @smallexample @c ada
6527 pragma Source_File_Name
6528 (B, Spec_File_Name => "source.ada", Index => 1);
6529 pragma Source_File_Name
6530 (A, Body_File_Name => "source.ada", Index => 2);
6533 Note that the @code{gnatname} utility can also be used to generate those
6534 configuration pragmas.
6536 Another form of the @code{Source_File_Name} pragma allows
6537 the specification of patterns defining alternative file naming schemes
6538 to apply to all files.
6540 @smallexample @c ada
6541 pragma Source_File_Name
6542 ( [Spec_File_Name =>] STRING_LITERAL
6543 [,[Casing =>] CASING_SPEC]
6544 [,[Dot_Replacement =>] STRING_LITERAL]);
6546 pragma Source_File_Name
6547 ( [Body_File_Name =>] STRING_LITERAL
6548 [,[Casing =>] CASING_SPEC]
6549 [,[Dot_Replacement =>] STRING_LITERAL]);
6551 pragma Source_File_Name
6552 ( [Subunit_File_Name =>] STRING_LITERAL
6553 [,[Casing =>] CASING_SPEC]
6554 [,[Dot_Replacement =>] STRING_LITERAL]);
6556 CASING_SPEC ::= Lowercase | Uppercase | Mixedcase
6560 The first argument is a pattern that contains a single asterisk indicating
6561 the point at which the unit name is to be inserted in the pattern string
6562 to form the file name. The second argument is optional. If present it
6563 specifies the casing of the unit name in the resulting file name string.
6564 The default is lower case. Finally the third argument allows for systematic
6565 replacement of any dots in the unit name by the specified string literal.
6567 Note that Source_File_Name pragmas should not be used if you are using
6568 project files. The reason for this rule is that the project manager is not
6569 aware of these pragmas, and so other tools that use the projet file would not
6570 be aware of the intended naming conventions. If you are using project files,
6571 file naming is controlled by Source_File_Name_Project pragmas, which are
6572 usually supplied automatically by the project manager. A pragma
6573 Source_File_Name cannot appear after a @ref{Pragma Source_File_Name_Project}.
6575 For more details on the use of the @code{Source_File_Name} pragma,
6576 @xref{Using Other File Names,,, gnat_ugn, @value{EDITION} User's Guide},
6577 and @ref{Alternative File Naming Schemes,,, gnat_ugn, @value{EDITION}
6580 @node Pragma Source_File_Name_Project
6581 @unnumberedsec Pragma Source_File_Name_Project
6582 @findex Source_File_Name_Project
6585 This pragma has the same syntax and semantics as pragma Source_File_Name.
6586 It is only allowed as a stand alone configuration pragma.
6587 It cannot appear after a @ref{Pragma Source_File_Name}, and
6588 most importantly, once pragma Source_File_Name_Project appears,
6589 no further Source_File_Name pragmas are allowed.
6591 The intention is that Source_File_Name_Project pragmas are always
6592 generated by the Project Manager in a manner consistent with the naming
6593 specified in a project file, and when naming is controlled in this manner,
6594 it is not permissible to attempt to modify this naming scheme using
6595 Source_File_Name or Source_File_Name_Project pragmas (which would not be
6596 known to the project manager).
6598 @node Pragma Source_Reference
6599 @unnumberedsec Pragma Source_Reference
6600 @findex Source_Reference
6604 @smallexample @c ada
6605 pragma Source_Reference (INTEGER_LITERAL, STRING_LITERAL);
6609 This pragma must appear as the first line of a source file.
6610 @var{integer_literal} is the logical line number of the line following
6611 the pragma line (for use in error messages and debugging
6612 information). @var{string_literal} is a static string constant that
6613 specifies the file name to be used in error messages and debugging
6614 information. This is most notably used for the output of @code{gnatchop}
6615 with the @option{-r} switch, to make sure that the original unchopped
6616 source file is the one referred to.
6618 The second argument must be a string literal, it cannot be a static
6619 string expression other than a string literal. This is because its value
6620 is needed for error messages issued by all phases of the compiler.
6622 @node Pragma SPARK_Mode
6623 @unnumberedsec Pragma SPARK_Mode
6628 @smallexample @c ada
6629 pragma SPARK_Mode [(On | Off)] ;
6633 In general a program can have some parts that are in SPARK 2014 (and
6634 follow all the rules in the SPARK Reference Manual), and some parts
6635 that are full Ada 2012.
6637 The SPARK_Mode pragma is used to identify which parts are in SPARK
6638 2014 (by default programs are in full Ada). The SPARK_Mode pragma can
6639 be used in the following places:
6644 As a configuration pragma, in which case it sets the default mode for
6645 all units compiled with this pragma.
6648 Immediately following a library-level subprogram spec
6651 Immediately within a library-level package body
6654 Immediately following the @code{private} keyword of a library-level
6658 Immediately following the @code{begin} keyword of a library-level
6662 Immediately within a library-level subprogram body
6667 Normally a subprogram or package spec/body inherits the current mode
6668 that is active at the point it is declared. But this can be overridden
6669 by pragma within the spec or body as above.
6671 The basic consistency rule is that you can't turn SPARK_Mode back
6672 @code{On}, once you have explicitly (with a pragma) turned if
6673 @code{Off}. So the following rules apply:
6676 If a subprogram spec has SPARK_Mode @code{Off}, then the body must
6677 also have SPARK_Mode @code{Off}.
6680 For a package, we have four parts:
6684 the package public declarations
6686 the package private part
6688 the body of the package
6690 the elaboration code after @code{begin}
6694 For a package, the rule is that if you explicitly turn SPARK_Mode
6695 @code{Off} for any part, then all the following parts must have
6696 SPARK_Mode @code{Off}. Note that this may require repeating a pragma
6697 SPARK_Mode (@code{Off}) in the body. For example, if we have a
6698 configuration pragma SPARK_Mode (@code{On}) that turns the mode on by
6699 default everywhere, and one particular package spec has pragma
6700 SPARK_Mode (@code{Off}), then that pragma will need to be repeated in
6703 @node Pragma Static_Elaboration_Desired
6704 @unnumberedsec Pragma Static_Elaboration_Desired
6705 @findex Static_Elaboration_Desired
6709 @smallexample @c ada
6710 pragma Static_Elaboration_Desired;
6714 This pragma is used to indicate that the compiler should attempt to initialize
6715 statically the objects declared in the library unit to which the pragma applies,
6716 when these objects are initialized (explicitly or implicitly) by an aggregate.
6717 In the absence of this pragma, aggregates in object declarations are expanded
6718 into assignments and loops, even when the aggregate components are static
6719 constants. When the aggregate is present the compiler builds a static expression
6720 that requires no run-time code, so that the initialized object can be placed in
6721 read-only data space. If the components are not static, or the aggregate has
6722 more that 100 components, the compiler emits a warning that the pragma cannot
6723 be obeyed. (See also the restriction No_Implicit_Loops, which supports static
6724 construction of larger aggregates with static components that include an others
6727 @node Pragma Stream_Convert
6728 @unnumberedsec Pragma Stream_Convert
6729 @findex Stream_Convert
6733 @smallexample @c ada
6734 pragma Stream_Convert (
6735 [Entity =>] type_LOCAL_NAME,
6736 [Read =>] function_NAME,
6737 [Write =>] function_NAME);
6741 This pragma provides an efficient way of providing user-defined stream
6742 attributes. Not only is it simpler to use than specifying the attributes
6743 directly, but more importantly, it allows the specification to be made in such
6744 a way that the predefined unit Ada.Streams is not loaded unless it is actually
6745 needed (i.e. unless the stream attributes are actually used); the use of
6746 the Stream_Convert pragma adds no overhead at all, unless the stream
6747 attributes are actually used on the designated type.
6749 The first argument specifies the type for which stream functions are
6750 provided. The second parameter provides a function used to read values
6751 of this type. It must name a function whose argument type may be any
6752 subtype, and whose returned type must be the type given as the first
6753 argument to the pragma.
6755 The meaning of the @var{Read} parameter is that if a stream attribute directly
6756 or indirectly specifies reading of the type given as the first parameter,
6757 then a value of the type given as the argument to the Read function is
6758 read from the stream, and then the Read function is used to convert this
6759 to the required target type.
6761 Similarly the @var{Write} parameter specifies how to treat write attributes
6762 that directly or indirectly apply to the type given as the first parameter.
6763 It must have an input parameter of the type specified by the first parameter,
6764 and the return type must be the same as the input type of the Read function.
6765 The effect is to first call the Write function to convert to the given stream
6766 type, and then write the result type to the stream.
6768 The Read and Write functions must not be overloaded subprograms. If necessary
6769 renamings can be supplied to meet this requirement.
6770 The usage of this attribute is best illustrated by a simple example, taken
6771 from the GNAT implementation of package Ada.Strings.Unbounded:
6773 @smallexample @c ada
6774 function To_Unbounded (S : String)
6775 return Unbounded_String
6776 renames To_Unbounded_String;
6778 pragma Stream_Convert
6779 (Unbounded_String, To_Unbounded, To_String);
6783 The specifications of the referenced functions, as given in the Ada
6784 Reference Manual are:
6786 @smallexample @c ada
6787 function To_Unbounded_String (Source : String)
6788 return Unbounded_String;
6790 function To_String (Source : Unbounded_String)
6795 The effect is that if the value of an unbounded string is written to a stream,
6796 then the representation of the item in the stream is in the same format that
6797 would be used for @code{Standard.String'Output}, and this same representation
6798 is expected when a value of this type is read from the stream. Note that the
6799 value written always includes the bounds, even for Unbounded_String'Write,
6800 since Unbounded_String is not an array type.
6802 Note that the @code{Stream_Convert} pragma is not effective in the case of
6803 a derived type of a non-limited tagged type. If such a type is specified then
6804 the pragma is silently ignored, and the default implementation of the stream
6805 attributes is used instead.
6807 @node Pragma Style_Checks
6808 @unnumberedsec Pragma Style_Checks
6809 @findex Style_Checks
6813 @smallexample @c ada
6814 pragma Style_Checks (string_LITERAL | ALL_CHECKS |
6815 On | Off [, LOCAL_NAME]);
6819 This pragma is used in conjunction with compiler switches to control the
6820 built in style checking provided by GNAT@. The compiler switches, if set,
6821 provide an initial setting for the switches, and this pragma may be used
6822 to modify these settings, or the settings may be provided entirely by
6823 the use of the pragma. This pragma can be used anywhere that a pragma
6824 is legal, including use as a configuration pragma (including use in
6825 the @file{gnat.adc} file).
6827 The form with a string literal specifies which style options are to be
6828 activated. These are additive, so they apply in addition to any previously
6829 set style check options. The codes for the options are the same as those
6830 used in the @option{-gnaty} switch to @command{gcc} or @command{gnatmake}.
6831 For example the following two methods can be used to enable
6836 @smallexample @c ada
6837 pragma Style_Checks ("l");
6842 gcc -c -gnatyl @dots{}
6847 The form ALL_CHECKS activates all standard checks (its use is equivalent
6848 to the use of the @code{gnaty} switch with no options. @xref{Top,
6849 @value{EDITION} User's Guide, About This Guide, gnat_ugn,
6850 @value{EDITION} User's Guide}, for details.)
6852 Note: the behavior is slightly different in GNAT mode (@option{-gnatg} used).
6853 In this case, ALL_CHECKS implies the standard set of GNAT mode style check
6854 options (i.e. equivalent to -gnatyg).
6856 The forms with @code{Off} and @code{On}
6857 can be used to temporarily disable style checks
6858 as shown in the following example:
6860 @smallexample @c ada
6864 pragma Style_Checks ("k"); -- requires keywords in lower case
6865 pragma Style_Checks (Off); -- turn off style checks
6866 NULL; -- this will not generate an error message
6867 pragma Style_Checks (On); -- turn style checks back on
6868 NULL; -- this will generate an error message
6872 Finally the two argument form is allowed only if the first argument is
6873 @code{On} or @code{Off}. The effect is to turn of semantic style checks
6874 for the specified entity, as shown in the following example:
6876 @smallexample @c ada
6880 pragma Style_Checks ("r"); -- require consistency of identifier casing
6882 Rf1 : Integer := ARG; -- incorrect, wrong case
6883 pragma Style_Checks (Off, Arg);
6884 Rf2 : Integer := ARG; -- OK, no error
6887 @node Pragma Subtitle
6888 @unnumberedsec Pragma Subtitle
6893 @smallexample @c ada
6894 pragma Subtitle ([Subtitle =>] STRING_LITERAL);
6898 This pragma is recognized for compatibility with other Ada compilers
6899 but is ignored by GNAT@.
6901 @node Pragma Suppress
6902 @unnumberedsec Pragma Suppress
6907 @smallexample @c ada
6908 pragma Suppress (Identifier [, [On =>] Name]);
6912 This is a standard pragma, and supports all the check names required in
6913 the RM. It is included here because GNAT recognizes some additional check
6914 names that are implementation defined (as permitted by the RM):
6919 @code{Alignment_Check} can be used to suppress alignment checks
6920 on addresses used in address clauses. Such checks can also be suppressed
6921 by suppressing range checks, but the specific use of @code{Alignment_Check}
6922 allows suppression of alignment checks without suppressing other range checks.
6925 @code{Atomic_Synchronization} can be used to suppress the special memory
6926 synchronization instructions that are normally generated for access to
6927 @code{Atomic} variables to ensure correct synchronization between tasks
6928 that use such variables for synchronization purposes.
6931 @code{Duplicated_Tag_Check} Can be used to suppress the check that is generated
6932 for a duplicated tag value when a tagged type is declared.
6935 @code{Predicate_Check} can be used to control whether predicate checks are
6936 active. It is applicable only to predicates for which the policy is
6937 @code{Check}. Unlike @code{Assertion_Policy}, which determines if a given
6938 predicate is ignored or checked for the whole program, the use of
6939 @code{Suppress} and @code{Unsuppress} with this check name allows a given
6940 predicate to be turned on and off at specific points in the program.
6943 @code{Validity_Check} can be used specifically to control validity checks.
6944 If @code{Suppress} is used to suppress validity checks, then no validity
6945 checks are performed, including those specified by the appropriate compiler
6946 switch or the @code{Validity_Checks} pragma.
6949 Additional check names previously introduced by use of the @code{Check_Name}
6950 pragma are also allowed.
6955 Note that pragma Suppress gives the compiler permission to omit
6956 checks, but does not require the compiler to omit checks. The compiler
6957 will generate checks if they are essentially free, even when they are
6958 suppressed. In particular, if the compiler can prove that a certain
6959 check will necessarily fail, it will generate code to do an
6960 unconditional ``raise'', even if checks are suppressed. The compiler
6963 Of course, run-time checks are omitted whenever the compiler can prove
6964 that they will not fail, whether or not checks are suppressed.
6966 @node Pragma Suppress_All
6967 @unnumberedsec Pragma Suppress_All
6968 @findex Suppress_All
6972 @smallexample @c ada
6973 pragma Suppress_All;
6977 This pragma can appear anywhere within a unit.
6978 The effect is to apply @code{Suppress (All_Checks)} to the unit
6979 in which it appears. This pragma is implemented for compatibility with DEC
6980 Ada 83 usage where it appears at the end of a unit, and for compatibility
6981 with Rational Ada, where it appears as a program unit pragma.
6982 The use of the standard Ada pragma @code{Suppress (All_Checks)}
6983 as a normal configuration pragma is the preferred usage in GNAT@.
6985 @node Pragma Suppress_Debug_Info
6986 @unnumberedsec Pragma Suppress_Debug_Info
6987 @findex Suppress_Debug_Info
6991 @smallexample @c ada
6992 Suppress_Debug_Info ([Entity =>] LOCAL_NAME);
6996 This pragma can be used to suppress generation of debug information
6997 for the specified entity. It is intended primarily for use in debugging
6998 the debugger, and navigating around debugger problems.
7000 @node Pragma Suppress_Exception_Locations
7001 @unnumberedsec Pragma Suppress_Exception_Locations
7002 @findex Suppress_Exception_Locations
7006 @smallexample @c ada
7007 pragma Suppress_Exception_Locations;
7011 In normal mode, a raise statement for an exception by default generates
7012 an exception message giving the file name and line number for the location
7013 of the raise. This is useful for debugging and logging purposes, but this
7014 entails extra space for the strings for the messages. The configuration
7015 pragma @code{Suppress_Exception_Locations} can be used to suppress the
7016 generation of these strings, with the result that space is saved, but the
7017 exception message for such raises is null. This configuration pragma may
7018 appear in a global configuration pragma file, or in a specific unit as
7019 usual. It is not required that this pragma be used consistently within
7020 a partition, so it is fine to have some units within a partition compiled
7021 with this pragma and others compiled in normal mode without it.
7023 @node Pragma Suppress_Initialization
7024 @unnumberedsec Pragma Suppress_Initialization
7025 @findex Suppress_Initialization
7026 @cindex Suppressing initialization
7027 @cindex Initialization, suppression of
7031 @smallexample @c ada
7032 pragma Suppress_Initialization ([Entity =>] subtype_Name);
7036 Here subtype_Name is the name introduced by a type declaration
7037 or subtype declaration.
7038 This pragma suppresses any implicit or explicit initialization
7039 for all variables of the given type or subtype,
7040 including initialization resulting from the use of pragmas
7041 Normalize_Scalars or Initialize_Scalars.
7043 This is considered a representation item, so it cannot be given after
7044 the type is frozen. It applies to all subsequent object declarations,
7045 and also any allocator that creates objects of the type.
7047 If the pragma is given for the first subtype, then it is considered
7048 to apply to the base type and all its subtypes. If the pragma is given
7049 for other than a first subtype, then it applies only to the given subtype.
7050 The pragma may not be given after the type is frozen.
7052 Note that this includes eliminating initialization of discriminants
7053 for discriminated types, and tags for tagged types. In these cases,
7054 you will have to use some non-portable mechanism (e.g. address
7055 overlays or unchecked conversion) to achieve required initialization
7056 of these fields before accessing any object of the corresponding type.
7058 @node Pragma Task_Name
7059 @unnumberedsec Pragma Task_Name
7064 @smallexample @c ada
7065 pragma Task_Name (string_EXPRESSION);
7069 This pragma appears within a task definition (like pragma
7070 @code{Priority}) and applies to the task in which it appears. The
7071 argument must be of type String, and provides a name to be used for
7072 the task instance when the task is created. Note that this expression
7073 is not required to be static, and in particular, it can contain
7074 references to task discriminants. This facility can be used to
7075 provide different names for different tasks as they are created,
7076 as illustrated in the example below.
7078 The task name is recorded internally in the run-time structures
7079 and is accessible to tools like the debugger. In addition the
7080 routine @code{Ada.Task_Identification.Image} will return this
7081 string, with a unique task address appended.
7083 @smallexample @c ada
7084 -- Example of the use of pragma Task_Name
7086 with Ada.Task_Identification;
7087 use Ada.Task_Identification;
7088 with Text_IO; use Text_IO;
7091 type Astring is access String;
7093 task type Task_Typ (Name : access String) is
7094 pragma Task_Name (Name.all);
7097 task body Task_Typ is
7098 Nam : constant String := Image (Current_Task);
7100 Put_Line ("-->" & Nam (1 .. 14) & "<--");
7103 type Ptr_Task is access Task_Typ;
7104 Task_Var : Ptr_Task;
7108 new Task_Typ (new String'("This is task 1"));
7110 new Task_Typ (new String'("This is task 2"));
7114 @node Pragma Task_Storage
7115 @unnumberedsec Pragma Task_Storage
7116 @findex Task_Storage
7119 @smallexample @c ada
7120 pragma Task_Storage (
7121 [Task_Type =>] LOCAL_NAME,
7122 [Top_Guard =>] static_integer_EXPRESSION);
7126 This pragma specifies the length of the guard area for tasks. The guard
7127 area is an additional storage area allocated to a task. A value of zero
7128 means that either no guard area is created or a minimal guard area is
7129 created, depending on the target. This pragma can appear anywhere a
7130 @code{Storage_Size} attribute definition clause is allowed for a task
7133 @node Pragma Test_Case
7134 @unnumberedsec Pragma Test_Case
7140 @smallexample @c ada
7142 [Name =>] static_string_Expression
7143 ,[Mode =>] (Nominal | Robustness)
7144 [, Requires => Boolean_Expression]
7145 [, Ensures => Boolean_Expression]);
7149 The @code{Test_Case} pragma allows defining fine-grain specifications
7150 for use by testing tools.
7151 The compiler checks the validity of the @code{Test_Case} pragma, but its
7152 presence does not lead to any modification of the code generated by the
7155 @code{Test_Case} pragmas may only appear immediately following the
7156 (separate) declaration of a subprogram in a package declaration, inside
7157 a package spec unit. Only other pragmas may intervene (that is appear
7158 between the subprogram declaration and a test case).
7160 The compiler checks that boolean expressions given in @code{Requires} and
7161 @code{Ensures} are valid, where the rules for @code{Requires} are the
7162 same as the rule for an expression in @code{Precondition} and the rules
7163 for @code{Ensures} are the same as the rule for an expression in
7164 @code{Postcondition}. In particular, attributes @code{'Old} and
7165 @code{'Result} can only be used within the @code{Ensures}
7166 expression. The following is an example of use within a package spec:
7168 @smallexample @c ada
7169 package Math_Functions is
7171 function Sqrt (Arg : Float) return Float;
7172 pragma Test_Case (Name => "Test 1",
7174 Requires => Arg < 10000,
7175 Ensures => Sqrt'Result < 10);
7181 The meaning of a test case is that there is at least one context where
7182 @code{Requires} holds such that, if the associated subprogram is executed in
7183 that context, then @code{Ensures} holds when the subprogram returns.
7184 Mode @code{Nominal} indicates that the input context should also satisfy the
7185 precondition of the subprogram, and the output context should also satisfy its
7186 postcondition. Mode @code{Robustness} indicates that the precondition and
7187 postcondition of the subprogram should be ignored for this test case.
7189 @node Pragma Thread_Local_Storage
7190 @unnumberedsec Pragma Thread_Local_Storage
7191 @findex Thread_Local_Storage
7192 @cindex Task specific storage
7193 @cindex TLS (Thread Local Storage)
7194 @cindex Task_Attributes
7197 @smallexample @c ada
7198 pragma Thread_Local_Storage ([Entity =>] LOCAL_NAME);
7202 This pragma specifies that the specified entity, which must be
7203 a variable declared in a library level package, is to be marked as
7204 "Thread Local Storage" (@code{TLS}). On systems supporting this (which
7205 include Solaris, GNU/Linux and VxWorks 6), this causes each thread
7206 (and hence each Ada task) to see a distinct copy of the variable.
7208 The variable may not have default initialization, and if there is
7209 an explicit initialization, it must be either @code{null} for an
7210 access variable, or a static expression for a scalar variable.
7211 This provides a low level mechanism similar to that provided by
7212 the @code{Ada.Task_Attributes} package, but much more efficient
7213 and is also useful in writing interface code that will interact
7214 with foreign threads.
7216 If this pragma is used on a system where @code{TLS} is not supported,
7217 then an error message will be generated and the program will be rejected.
7219 @node Pragma Time_Slice
7220 @unnumberedsec Pragma Time_Slice
7225 @smallexample @c ada
7226 pragma Time_Slice (static_duration_EXPRESSION);
7230 For implementations of GNAT on operating systems where it is possible
7231 to supply a time slice value, this pragma may be used for this purpose.
7232 It is ignored if it is used in a system that does not allow this control,
7233 or if it appears in other than the main program unit.
7235 Note that the effect of this pragma is identical to the effect of the
7236 DEC Ada 83 pragma of the same name when operating under OpenVMS systems.
7239 @unnumberedsec Pragma Title
7244 @smallexample @c ada
7245 pragma Title (TITLING_OPTION [, TITLING OPTION]);
7248 [Title =>] STRING_LITERAL,
7249 | [Subtitle =>] STRING_LITERAL
7253 Syntax checked but otherwise ignored by GNAT@. This is a listing control
7254 pragma used in DEC Ada 83 implementations to provide a title and/or
7255 subtitle for the program listing. The program listing generated by GNAT
7256 does not have titles or subtitles.
7258 Unlike other pragmas, the full flexibility of named notation is allowed
7259 for this pragma, i.e.@: the parameters may be given in any order if named
7260 notation is used, and named and positional notation can be mixed
7261 following the normal rules for procedure calls in Ada.
7263 @node Pragma Type_Invariant
7264 @unnumberedsec Pragma Type_Invariant
7266 @findex Type_Invariant pragma
7270 @smallexample @c ada
7271 pragma Type_Invariant
7272 ([Entity =>] type_LOCAL_NAME,
7273 [Check =>] EXPRESSION);
7277 The @code{Type_Invariant} pragma is intended to be an exact
7278 replacement for the language-defined @code{Type_Invariant}
7279 aspect, and shares its restrictions and semantics. It differs
7280 from the language defined @code{Invariant} pragma in that it
7281 does not permit a string parameter, and it is
7282 controlled by the assertion identifier @code{Type_Invariant}
7283 rather than @code{Invariant}.
7285 @node Pragma Type_Invariant_Class
7286 @unnumberedsec Pragma Type_Invariant_Class
7288 @findex Type_Invariant_Class pragma
7292 @smallexample @c ada
7293 pragma Type_Invariant_Class
7294 ([Entity =>] type_LOCAL_NAME,
7295 [Check =>] EXPRESSION);
7299 The @code{Type_Invariant_Class} pragma is intended to be an exact
7300 replacement for the language-defined @code{Type_Invariant'Class}
7301 aspect, and shares its restrictions and semantics.
7303 Note: This pragma is called @code{Type_Invariant_Class} rather than
7304 @code{Type_Invariant'Class} because the latter would not be strictly
7305 conforming to the allowed syntax for pragmas. The motivation
7306 for providing pragmas equivalent to the aspects is to allow a program
7307 to be written using the pragmas, and then compiled if necessary
7308 using an Ada compiler that does not recognize the pragmas or
7309 aspects, but is prepared to ignore the pragmas. The assertion
7310 policy that controls this pragma is @code{Type_Invariant'Class},
7311 not @code{Type_Invariant_Class}.
7313 @node Pragma Unchecked_Union
7314 @unnumberedsec Pragma Unchecked_Union
7316 @findex Unchecked_Union
7320 @smallexample @c ada
7321 pragma Unchecked_Union (first_subtype_LOCAL_NAME);
7325 This pragma is used to specify a representation of a record type that is
7326 equivalent to a C union. It was introduced as a GNAT implementation defined
7327 pragma in the GNAT Ada 95 mode. Ada 2005 includes an extended version of this
7328 pragma, making it language defined, and GNAT fully implements this extended
7329 version in all language modes (Ada 83, Ada 95, and Ada 2005). For full
7330 details, consult the Ada 2012 Reference Manual, section B.3.3.
7332 @node Pragma Unevaluated_Use_Of_Old
7333 @unnumberedsec Pragma Unevaluated_Use_Of_Old
7334 @cindex Attribute Old
7335 @cindex Attribute Loop_Entry
7336 @cindex Unevaluated_Use_Of_Old
7337 @findex Unevaluated_Use_Of_Old
7341 @smallexample @c ada
7342 pragma Unevaluated_Use_Of_Old (Error | Warn | Allow);
7346 This pragma controls the processing of attributes Old and Loop_Entry.
7347 If either of these attributes is used in a potentially unevaluated
7348 expression (e.g. the then or else parts of an if expression), then
7349 normally this usage is considered illegal if the prefix of the attribute
7350 is other than an entity name. The language requires this
7351 behavior for Old, and GNAT copies the same rule for Loop_Entry.
7353 The reason for this rule is that otherwise, we can have a situation
7354 where we save the Old value, and this results in an exception, even
7355 though we might not evaluate the attribute. Consider this example:
7357 @smallexample @c ada
7358 package UnevalOld is
7360 procedure U (A : String; C : Boolean) -- ERROR
7361 with Post => (if C then A(1)'Old = K else True);
7366 If procedure U is called with a string with a lower bound of 2, and
7367 C false, then an exception would be raised trying to evaluate A(1)
7368 on entry even though the value would not be actually used.
7370 Although the rule guarantees against this possibility, it is sometimes
7371 too restrictive. For example if we know that the string has a lower
7372 bound of 1, then we will never raise an exception.
7373 The pragma @code{Unevaluated_Use_Of_Old} can be
7374 used to modify this behavior. If the argument is @code{Error} then an
7375 error is given (this is the default RM behavior). If the argument is
7376 @code{Warn} then the usage is allowed as legal but with a warning
7377 that an exception might be raised. If the argument is @code{Allow}
7378 then the usage is allowed as legal without generating a warning.
7380 This pragma may appear as a configuration pragma, or in a declarative
7381 part or package specification. In the latter case it applies to
7382 uses up to the end of the corresponding statement sequence or
7383 sequence of package declarations.
7385 @node Pragma Unimplemented_Unit
7386 @unnumberedsec Pragma Unimplemented_Unit
7387 @findex Unimplemented_Unit
7391 @smallexample @c ada
7392 pragma Unimplemented_Unit;
7396 If this pragma occurs in a unit that is processed by the compiler, GNAT
7397 aborts with the message @samp{@var{xxx} not implemented}, where
7398 @var{xxx} is the name of the current compilation unit. This pragma is
7399 intended to allow the compiler to handle unimplemented library units in
7402 The abort only happens if code is being generated. Thus you can use
7403 specs of unimplemented packages in syntax or semantic checking mode.
7405 @node Pragma Universal_Aliasing
7406 @unnumberedsec Pragma Universal_Aliasing
7407 @findex Universal_Aliasing
7411 @smallexample @c ada
7412 pragma Universal_Aliasing [([Entity =>] type_LOCAL_NAME)];
7416 @var{type_LOCAL_NAME} must refer to a type declaration in the current
7417 declarative part. The effect is to inhibit strict type-based aliasing
7418 optimization for the given type. In other words, the effect is as though
7419 access types designating this type were subject to pragma No_Strict_Aliasing.
7420 For a detailed description of the strict aliasing optimization, and the
7421 situations in which it must be suppressed, @xref{Optimization and Strict
7422 Aliasing,,, gnat_ugn, @value{EDITION} User's Guide}.
7424 @node Pragma Universal_Data
7425 @unnumberedsec Pragma Universal_Data
7426 @findex Universal_Data
7430 @smallexample @c ada
7431 pragma Universal_Data [(library_unit_Name)];
7435 This pragma is supported only for the AAMP target and is ignored for
7436 other targets. The pragma specifies that all library-level objects
7437 (Counter 0 data) associated with the library unit are to be accessed
7438 and updated using universal addressing (24-bit addresses for AAMP5)
7439 rather than the default of 16-bit Data Environment (DENV) addressing.
7440 Use of this pragma will generally result in less efficient code for
7441 references to global data associated with the library unit, but
7442 allows such data to be located anywhere in memory. This pragma is
7443 a library unit pragma, but can also be used as a configuration pragma
7444 (including use in the @file{gnat.adc} file). The functionality
7445 of this pragma is also available by applying the -univ switch on the
7446 compilations of units where universal addressing of the data is desired.
7448 @node Pragma Unmodified
7449 @unnumberedsec Pragma Unmodified
7451 @cindex Warnings, unmodified
7455 @smallexample @c ada
7456 pragma Unmodified (LOCAL_NAME @{, LOCAL_NAME@});
7460 This pragma signals that the assignable entities (variables,
7461 @code{out} parameters, @code{in out} parameters) whose names are listed are
7462 deliberately not assigned in the current source unit. This
7463 suppresses warnings about the
7464 entities being referenced but not assigned, and in addition a warning will be
7465 generated if one of these entities is in fact assigned in the
7466 same unit as the pragma (or in the corresponding body, or one
7469 This is particularly useful for clearly signaling that a particular
7470 parameter is not modified, even though the spec suggests that it might
7473 For the variable case, warnings are never given for unreferenced variables
7474 whose name contains one of the substrings
7475 @code{DISCARD, DUMMY, IGNORE, JUNK, UNUSED} in any casing. Such names
7476 are typically to be used in cases where such warnings are expected.
7477 Thus it is never necessary to use @code{pragma Unmodified} for such
7478 variables, though it is harmless to do so.
7480 @node Pragma Unreferenced
7481 @unnumberedsec Pragma Unreferenced
7482 @findex Unreferenced
7483 @cindex Warnings, unreferenced
7487 @smallexample @c ada
7488 pragma Unreferenced (LOCAL_NAME @{, LOCAL_NAME@});
7489 pragma Unreferenced (library_unit_NAME @{, library_unit_NAME@});
7493 This pragma signals that the entities whose names are listed are
7494 deliberately not referenced in the current source unit after the
7495 occurrence of the pragma. This
7496 suppresses warnings about the
7497 entities being unreferenced, and in addition a warning will be
7498 generated if one of these entities is in fact subsequently referenced in the
7499 same unit as the pragma (or in the corresponding body, or one
7502 This is particularly useful for clearly signaling that a particular
7503 parameter is not referenced in some particular subprogram implementation
7504 and that this is deliberate. It can also be useful in the case of
7505 objects declared only for their initialization or finalization side
7508 If @code{LOCAL_NAME} identifies more than one matching homonym in the
7509 current scope, then the entity most recently declared is the one to which
7510 the pragma applies. Note that in the case of accept formals, the pragma
7511 Unreferenced may appear immediately after the keyword @code{do} which
7512 allows the indication of whether or not accept formals are referenced
7513 or not to be given individually for each accept statement.
7515 The left hand side of an assignment does not count as a reference for the
7516 purpose of this pragma. Thus it is fine to assign to an entity for which
7517 pragma Unreferenced is given.
7519 Note that if a warning is desired for all calls to a given subprogram,
7520 regardless of whether they occur in the same unit as the subprogram
7521 declaration, then this pragma should not be used (calls from another
7522 unit would not be flagged); pragma Obsolescent can be used instead
7523 for this purpose, see @xref{Pragma Obsolescent}.
7525 The second form of pragma @code{Unreferenced} is used within a context
7526 clause. In this case the arguments must be unit names of units previously
7527 mentioned in @code{with} clauses (similar to the usage of pragma
7528 @code{Elaborate_All}. The effect is to suppress warnings about unreferenced
7529 units and unreferenced entities within these units.
7531 For the variable case, warnings are never given for unreferenced variables
7532 whose name contains one of the substrings
7533 @code{DISCARD, DUMMY, IGNORE, JUNK, UNUSED} in any casing. Such names
7534 are typically to be used in cases where such warnings are expected.
7535 Thus it is never necessary to use @code{pragma Unreferenced} for such
7536 variables, though it is harmless to do so.
7538 @node Pragma Unreferenced_Objects
7539 @unnumberedsec Pragma Unreferenced_Objects
7540 @findex Unreferenced_Objects
7541 @cindex Warnings, unreferenced
7545 @smallexample @c ada
7546 pragma Unreferenced_Objects (local_subtype_NAME @{, local_subtype_NAME@});
7550 This pragma signals that for the types or subtypes whose names are
7551 listed, objects which are declared with one of these types or subtypes may
7552 not be referenced, and if no references appear, no warnings are given.
7554 This is particularly useful for objects which are declared solely for their
7555 initialization and finalization effect. Such variables are sometimes referred
7556 to as RAII variables (Resource Acquisition Is Initialization). Using this
7557 pragma on the relevant type (most typically a limited controlled type), the
7558 compiler will automatically suppress unwanted warnings about these variables
7559 not being referenced.
7561 @node Pragma Unreserve_All_Interrupts
7562 @unnumberedsec Pragma Unreserve_All_Interrupts
7563 @findex Unreserve_All_Interrupts
7567 @smallexample @c ada
7568 pragma Unreserve_All_Interrupts;
7572 Normally certain interrupts are reserved to the implementation. Any attempt
7573 to attach an interrupt causes Program_Error to be raised, as described in
7574 RM C.3.2(22). A typical example is the @code{SIGINT} interrupt used in
7575 many systems for a @kbd{Ctrl-C} interrupt. Normally this interrupt is
7576 reserved to the implementation, so that @kbd{Ctrl-C} can be used to
7577 interrupt execution.
7579 If the pragma @code{Unreserve_All_Interrupts} appears anywhere in any unit in
7580 a program, then all such interrupts are unreserved. This allows the
7581 program to handle these interrupts, but disables their standard
7582 functions. For example, if this pragma is used, then pressing
7583 @kbd{Ctrl-C} will not automatically interrupt execution. However,
7584 a program can then handle the @code{SIGINT} interrupt as it chooses.
7586 For a full list of the interrupts handled in a specific implementation,
7587 see the source code for the spec of @code{Ada.Interrupts.Names} in
7588 file @file{a-intnam.ads}. This is a target dependent file that contains the
7589 list of interrupts recognized for a given target. The documentation in
7590 this file also specifies what interrupts are affected by the use of
7591 the @code{Unreserve_All_Interrupts} pragma.
7593 For a more general facility for controlling what interrupts can be
7594 handled, see pragma @code{Interrupt_State}, which subsumes the functionality
7595 of the @code{Unreserve_All_Interrupts} pragma.
7597 @node Pragma Unsuppress
7598 @unnumberedsec Pragma Unsuppress
7603 @smallexample @c ada
7604 pragma Unsuppress (IDENTIFIER [, [On =>] NAME]);
7608 This pragma undoes the effect of a previous pragma @code{Suppress}. If
7609 there is no corresponding pragma @code{Suppress} in effect, it has no
7610 effect. The range of the effect is the same as for pragma
7611 @code{Suppress}. The meaning of the arguments is identical to that used
7612 in pragma @code{Suppress}.
7614 One important application is to ensure that checks are on in cases where
7615 code depends on the checks for its correct functioning, so that the code
7616 will compile correctly even if the compiler switches are set to suppress
7617 checks. For example, in a program that depends on external names of tagged
7618 types and wants to ensure that the duplicated tag check occurs even if all
7619 run-time checks are suppressed by a compiler switch, the following
7620 configuration pragma will ensure this test is not suppressed:
7622 @smallexample @c ada
7623 pragma Unsuppress (Duplicated_Tag_Check);
7627 This pragma is standard in Ada 2005. It is available in all earlier versions
7628 of Ada as an implementation-defined pragma.
7630 Note that in addition to the checks defined in the Ada RM, GNAT recogizes
7631 a number of implementation-defined check names. See description of pragma
7632 @code{Suppress} for full details.
7634 @node Pragma Use_VADS_Size
7635 @unnumberedsec Pragma Use_VADS_Size
7636 @cindex @code{Size}, VADS compatibility
7637 @cindex Rational profile
7638 @findex Use_VADS_Size
7642 @smallexample @c ada
7643 pragma Use_VADS_Size;
7647 This is a configuration pragma. In a unit to which it applies, any use
7648 of the 'Size attribute is automatically interpreted as a use of the
7649 'VADS_Size attribute. Note that this may result in incorrect semantic
7650 processing of valid Ada 95 or Ada 2005 programs. This is intended to aid in
7651 the handling of existing code which depends on the interpretation of Size
7652 as implemented in the VADS compiler. See description of the VADS_Size
7653 attribute for further details.
7655 @node Pragma Validity_Checks
7656 @unnumberedsec Pragma Validity_Checks
7657 @findex Validity_Checks
7661 @smallexample @c ada
7662 pragma Validity_Checks (string_LITERAL | ALL_CHECKS | On | Off);
7666 This pragma is used in conjunction with compiler switches to control the
7667 built-in validity checking provided by GNAT@. The compiler switches, if set
7668 provide an initial setting for the switches, and this pragma may be used
7669 to modify these settings, or the settings may be provided entirely by
7670 the use of the pragma. This pragma can be used anywhere that a pragma
7671 is legal, including use as a configuration pragma (including use in
7672 the @file{gnat.adc} file).
7674 The form with a string literal specifies which validity options are to be
7675 activated. The validity checks are first set to include only the default
7676 reference manual settings, and then a string of letters in the string
7677 specifies the exact set of options required. The form of this string
7678 is exactly as described for the @option{-gnatVx} compiler switch (see the
7679 @value{EDITION} User's Guide for details). For example the following two
7680 methods can be used to enable validity checking for mode @code{in} and
7681 @code{in out} subprogram parameters:
7685 @smallexample @c ada
7686 pragma Validity_Checks ("im");
7691 gcc -c -gnatVim @dots{}
7696 The form ALL_CHECKS activates all standard checks (its use is equivalent
7697 to the use of the @code{gnatva} switch.
7699 The forms with @code{Off} and @code{On}
7700 can be used to temporarily disable validity checks
7701 as shown in the following example:
7703 @smallexample @c ada
7707 pragma Validity_Checks ("c"); -- validity checks for copies
7708 pragma Validity_Checks (Off); -- turn off validity checks
7709 A := B; -- B will not be validity checked
7710 pragma Validity_Checks (On); -- turn validity checks back on
7711 A := C; -- C will be validity checked
7714 @node Pragma Volatile
7715 @unnumberedsec Pragma Volatile
7720 @smallexample @c ada
7721 pragma Volatile (LOCAL_NAME);
7725 This pragma is defined by the Ada Reference Manual, and the GNAT
7726 implementation is fully conformant with this definition. The reason it
7727 is mentioned in this section is that a pragma of the same name was supplied
7728 in some Ada 83 compilers, including DEC Ada 83. The Ada 95 / Ada 2005
7729 implementation of pragma Volatile is upwards compatible with the
7730 implementation in DEC Ada 83.
7732 @node Pragma Warning_As_Error
7733 @unnumberedsec Pragma Warning_As_Error
7734 @findex Warning_As_Error
7738 @smallexample @c ada
7739 pragma Warning_As_Error (static_string_EXPRESSION);
7743 This configuration pragma allows the programmer to specify a set
7744 of warnings that will be treated as errors. Any warning which
7745 matches the pattern given by the pragma argument will be treated
7746 as an error. This gives much more precise control that -gnatwe
7747 which treats all warnings as errors.
7749 The pattern may contain asterisks, which match zero or more characters in
7750 the message. For example, you can use
7751 @code{pragma Warning_As_Error ("bits of*unused")} to treat the warning
7752 message @code{warning: 960 bits of "a" unused} as an error. No other regular
7753 expression notations are permitted. All characters other than asterisk in
7754 these three specific cases are treated as literal characters in the match.
7755 The match is case insensitive, for example XYZ matches xyz.
7757 Note that the pattern matches if it occurs anywhere within the warning
7758 message string (it is not necessary to put an asterisk at the start and
7759 the end of the message, since this is implied).
7761 Another possibility for the static_string_EXPRESSION which works whether
7762 or not error tags are enabled (@option{-gnatw.d}) is to use the
7763 @option{-gnatw} tag string, enclosed in brackets,
7764 as shown in the example below, to treat a class of warnings as errors.
7766 The above use of patterns to match the message applies only to warning
7767 messages generated by the front end. This pragma can also be applied to
7768 warnings provided by the back end and mentioned in @ref{Pragma Warnings}.
7769 By using a single full @option{-Wxxx} switch in the pragma, such warnings
7770 can also be treated as errors.
7772 The pragma can appear either in a global configuration pragma file
7773 (e.g. @file{gnat.adc}), or at the start of a file. Given a global
7774 configuration pragma file containing:
7776 @smallexample @c ada
7777 pragma Warning_As_Error ("[-gnatwj]");
7781 which will treat all obsolescent feature warnings as errors, the
7782 following program compiles as shown (compile options here are
7783 @option{-gnatwa.d -gnatl -gnatj55}).
7785 @smallexample @c ada
7786 1. pragma Warning_As_Error ("*never assigned*");
7787 2. function Warnerr return String is
7790 >>> error: variable "X" is never read and
7791 never assigned [-gnatwv] [warning-as-error]
7795 >>> warning: variable "Y" is assigned but
7796 never read [-gnatwu]
7802 >>> error: use of "%" is an obsolescent
7803 feature (RM J.2(4)), use """ instead
7804 [-gnatwj] [warning-as-error]
7808 8 lines: No errors, 3 warnings (2 treated as errors)
7812 Note that this pragma does not affect the set of warnings issued in
7813 any way, it merely changes the effect of a matching warning if one
7814 is produced as a result of other warnings options. As shown in this
7815 example, if the pragma results in a warning being treated as an error,
7816 the tag is changed from "warning:" to "error:" and the string
7817 "[warning-as-error]" is appended to the end of the message.
7819 @node Pragma Warnings
7820 @unnumberedsec Pragma Warnings
7825 @smallexample @c ada
7826 pragma Warnings (On | Off [,REASON]);
7827 pragma Warnings (On | Off, LOCAL_NAME [,REASON]);
7828 pragma Warnings (static_string_EXPRESSION [,REASON]);
7829 pragma Warnings (On | Off, static_string_EXPRESSION [,REASON]);
7831 REASON ::= Reason => STRING_LITERAL @{& STRING_LITERAL@}
7833 Note: in Ada 83 mode, a string literal may be used in place of
7834 a static string expression (which does not exist in Ada 83).
7838 Normally warnings are enabled, with the output being controlled by
7839 the command line switch. Warnings (@code{Off}) turns off generation of
7840 warnings until a Warnings (@code{On}) is encountered or the end of the
7841 current unit. If generation of warnings is turned off using this
7842 pragma, then some or all of the warning messages are suppressed,
7843 regardless of the setting of the command line switches.
7845 The @code{Reason} parameter may optionally appear as the last argument
7846 in any of the forms of this pragma. It is intended purely for the
7847 purposes of documenting the reason for the @code{Warnings} pragma.
7848 The compiler will check that the argument is a static string but
7849 otherwise ignore this argument. Other tools may provide specialized
7850 processing for this string.
7852 The form with a single argument (or two arguments if Reason present),
7853 where the first argument is @code{ON} or @code{OFF}
7854 may be used as a configuration pragma.
7856 If the @var{LOCAL_NAME} parameter is present, warnings are suppressed for
7857 the specified entity. This suppression is effective from the point where
7858 it occurs till the end of the extended scope of the variable (similar to
7859 the scope of @code{Suppress}). This form cannot be used as a configuration
7862 In the case where the first argument is other than @code{ON} or
7864 the third form with a single static_string_EXPRESSION argument (and possible
7865 reason) provides more precise
7866 control over which warnings are active. The string is a list of letters
7867 specifying which warnings are to be activated and which deactivated. The
7868 code for these letters is the same as the string used in the command
7869 line switch controlling warnings. For a brief summary, use the gnatmake
7870 command with no arguments, which will generate usage information containing
7871 the list of warnings switches supported. For
7872 full details see @ref{Warning Message Control,,, gnat_ugn, @value{EDITION}
7873 User's Guide}. This form can also be used as a configuration pragma.
7876 The warnings controlled by the @option{-gnatw} switch are generated by the
7877 front end of the compiler. The GCC back end can provide additional warnings
7878 and they are controlled by the @option{-W} switch. Such warnings can be
7879 identified by the appearance of a string of the form @code{[-Wxxx]} in the
7880 message which designates the @option{-Wxxx} switch that controls the message.
7881 The form with a single static_string_EXPRESSION argument also works for these
7882 warnings, but the string must be a single full @option{-Wxxx} switch in this
7883 case. The above reference lists a few examples of these additional warnings.
7886 The specified warnings will be in effect until the end of the program
7887 or another pragma Warnings is encountered. The effect of the pragma is
7888 cumulative. Initially the set of warnings is the standard default set
7889 as possibly modified by compiler switches. Then each pragma Warning
7890 modifies this set of warnings as specified. This form of the pragma may
7891 also be used as a configuration pragma.
7893 The fourth form, with an @code{On|Off} parameter and a string, is used to
7894 control individual messages, based on their text. The string argument
7895 is a pattern that is used to match against the text of individual
7896 warning messages (not including the initial "warning: " tag).
7898 The pattern may contain asterisks, which match zero or more characters in
7899 the message. For example, you can use
7900 @code{pragma Warnings (Off, "bits of*unused")} to suppress the warning
7901 message @code{warning: 960 bits of "a" unused}. No other regular
7902 expression notations are permitted. All characters other than asterisk in
7903 these three specific cases are treated as literal characters in the match.
7904 The match is case insensitive, for example XYZ matches xyz.
7906 Note that the pattern matches if it occurs anywhere within the warning
7907 message string (it is not necessary to put an asterisk at the start and
7908 the end of the message, since this is implied).
7910 The above use of patterns to match the message applies only to warning
7911 messages generated by the front end. This form of the pragma with a string
7912 argument can also be used to control warnings provided by the back end and
7913 mentioned above. By using a single full @option{-Wxxx} switch in the pragma,
7914 such warnings can be turned on and off.
7916 There are two ways to use the pragma in this form. The OFF form can be used
7917 as a configuration pragma. The effect is to suppress all warnings (if any)
7918 that match the pattern string throughout the compilation (or match the
7919 -W switch in the back end case).
7921 The second usage is to suppress a warning locally, and in this case, two
7922 pragmas must appear in sequence:
7924 @smallexample @c ada
7925 pragma Warnings (Off, Pattern);
7926 @dots{} code where given warning is to be suppressed
7927 pragma Warnings (On, Pattern);
7931 In this usage, the pattern string must match in the Off and On pragmas,
7932 and at least one matching warning must be suppressed.
7934 Note: to write a string that will match any warning, use the string
7935 @code{"***"}. It will not work to use a single asterisk or two asterisks
7936 since this looks like an operator name. This form with three asterisks
7937 is similar in effect to specifying @code{pragma Warnings (Off)} except that a
7938 matching @code{pragma Warnings (On, "***")} will be required. This can be
7939 helpful in avoiding forgetting to turn warnings back on.
7941 Note: the debug flag -gnatd.i (@code{/NOWARNINGS_PRAGMAS} in VMS) can be
7942 used to cause the compiler to entirely ignore all WARNINGS pragmas. This can
7943 be useful in checking whether obsolete pragmas in existing programs are hiding
7946 Note: pragma Warnings does not affect the processing of style messages. See
7947 separate entry for pragma Style_Checks for control of style messages.
7949 @node Pragma Weak_External
7950 @unnumberedsec Pragma Weak_External
7951 @findex Weak_External
7955 @smallexample @c ada
7956 pragma Weak_External ([Entity =>] LOCAL_NAME);
7960 @var{LOCAL_NAME} must refer to an object that is declared at the library
7961 level. This pragma specifies that the given entity should be marked as a
7962 weak symbol for the linker. It is equivalent to @code{__attribute__((weak))}
7963 in GNU C and causes @var{LOCAL_NAME} to be emitted as a weak symbol instead
7964 of a regular symbol, that is to say a symbol that does not have to be
7965 resolved by the linker if used in conjunction with a pragma Import.
7967 When a weak symbol is not resolved by the linker, its address is set to
7968 zero. This is useful in writing interfaces to external modules that may
7969 or may not be linked in the final executable, for example depending on
7970 configuration settings.
7972 If a program references at run time an entity to which this pragma has been
7973 applied, and the corresponding symbol was not resolved at link time, then
7974 the execution of the program is erroneous. It is not erroneous to take the
7975 Address of such an entity, for example to guard potential references,
7976 as shown in the example below.
7978 Some file formats do not support weak symbols so not all target machines
7979 support this pragma.
7981 @smallexample @c ada
7982 -- Example of the use of pragma Weak_External
7984 package External_Module is
7986 pragma Import (C, key);
7987 pragma Weak_External (key);
7988 function Present return boolean;
7989 end External_Module;
7991 with System; use System;
7992 package body External_Module is
7993 function Present return boolean is
7995 return key'Address /= System.Null_Address;
7997 end External_Module;
8000 @node Pragma Wide_Character_Encoding
8001 @unnumberedsec Pragma Wide_Character_Encoding
8002 @findex Wide_Character_Encoding
8006 @smallexample @c ada
8007 pragma Wide_Character_Encoding (IDENTIFIER | CHARACTER_LITERAL);
8011 This pragma specifies the wide character encoding to be used in program
8012 source text appearing subsequently. It is a configuration pragma, but may
8013 also be used at any point that a pragma is allowed, and it is permissible
8014 to have more than one such pragma in a file, allowing multiple encodings
8015 to appear within the same file.
8017 The argument can be an identifier or a character literal. In the identifier
8018 case, it is one of @code{HEX}, @code{UPPER}, @code{SHIFT_JIS},
8019 @code{EUC}, @code{UTF8}, or @code{BRACKETS}. In the character literal
8020 case it is correspondingly one of the characters @samp{h}, @samp{u},
8021 @samp{s}, @samp{e}, @samp{8}, or @samp{b}.
8023 Note that when the pragma is used within a file, it affects only the
8024 encoding within that file, and does not affect withed units, specs,
8027 @node Implementation Defined Aspects
8028 @chapter Implementation Defined Aspects
8029 Ada defines (throughout the Ada 2012 reference manual, summarized
8030 in Annex K) a set of aspects that can be specified for certain entities.
8031 These language defined aspects are implemented in GNAT in Ada 2012 mode
8032 and work as described in the Ada 2012 Reference Manual.
8034 In addition, Ada 2012 allows implementations to define additional aspects
8035 whose meaning is defined by the implementation. GNAT provides
8036 a number of these implementation-defined aspects which can be used
8037 to extend and enhance the functionality of the compiler. This section of
8038 the GNAT reference manual describes these additional aspects.
8040 Note that any program using these aspects may not be portable to
8041 other compilers (although GNAT implements this set of aspects on all
8042 platforms). Therefore if portability to other compilers is an important
8043 consideration, you should minimize the use of these aspects.
8045 Note that for many of these aspects, the effect is essentially similar
8046 to the use of a pragma or attribute specification with the same name
8047 applied to the entity. For example, if we write:
8049 @smallexample @c ada
8050 type R is range 1 .. 100
8051 with Value_Size => 10;
8055 then the effect is the same as:
8057 @smallexample @c ada
8058 type R is range 1 .. 100;
8059 for R'Value_Size use 10;
8065 @smallexample @c ada
8066 type R is new Integer
8067 with Shared => True;
8071 then the effect is the same as:
8073 @smallexample @c ada
8074 type R is new Integer;
8079 In the documentation below, such cases are simply marked
8080 as being equivalent to the corresponding pragma or attribute definition
8084 * Aspect Abstract_State::
8086 * Aspect Async_Readers::
8087 * Aspect Async_Writers::
8088 * Aspect Contract_Cases::
8090 * Aspect Dimension::
8091 * Aspect Dimension_System::
8092 * Aspect Effective_Reads::
8093 * Aspect Effective_Writes::
8094 * Aspect Favor_Top_Level::
8096 * Aspect Initial_Condition::
8097 * Aspect Initializes::
8098 * Aspect Inline_Always::
8099 * Aspect Invariant::
8100 * Aspect Invariant'Class::
8102 * Aspect Linker_Section::
8103 * Aspect Lock_Free::
8104 * Aspect No_Elaboration_Code_All::
8105 * Aspect Object_Size::
8106 * Aspect Obsolescent::
8108 * Aspect Persistent_BSS::
8109 * Aspect Predicate::
8110 * Aspect Pure_Function::
8111 * Aspect Refined_Depends::
8112 * Aspect Refined_Global::
8113 * Aspect Refined_Post::
8114 * Aspect Refined_State::
8115 * Aspect Remote_Access_Type::
8116 * Aspect Scalar_Storage_Order::
8118 * Aspect Simple_Storage_Pool::
8119 * Aspect Simple_Storage_Pool_Type::
8120 * Aspect SPARK_Mode::
8121 * Aspect Suppress_Debug_Info::
8122 * Aspect Test_Case::
8123 * Aspect Thread_Local_Storage::
8124 * Aspect Universal_Aliasing::
8125 * Aspect Universal_Data::
8126 * Aspect Unmodified::
8127 * Aspect Unreferenced::
8128 * Aspect Unreferenced_Objects::
8129 * Aspect Value_Size::
8133 @node Aspect Abstract_State
8134 @unnumberedsec Aspect Abstract_State
8135 @findex Abstract_State
8137 This aspect is equivalent to pragma @code{Abstract_State}.
8139 @node Aspect Annotate
8140 @unnumberedsec Annotate
8143 There are three forms of this aspect (where ID is an identifier,
8144 and ARG is a general expression).
8147 @item Annotate => ID
8148 Equivalent to @code{pragma Annotate (ID, Entity => Name);}
8150 @item Annotate => (ID)
8151 Equivalent to @code{pragma Annotate (ID, Entity => Name);}
8153 @item Annotate => (ID ,ID @{, ARG@})
8154 Equivalent to @code{pragma Annotate (ID, ID @{, ARG@}, Entity => Name);}
8157 @node Aspect Async_Readers
8158 @unnumberedsec Aspect Async_Readers
8159 @findex Async_Readers
8161 This aspect is equivalent to pragma @code{Async_Readers}.
8163 @node Aspect Async_Writers
8164 @unnumberedsec Aspect Async_Writers
8165 @findex Async_Writers
8167 This aspect is equivalent to pragma @code{Async_Writers}.
8169 @node Aspect Contract_Cases
8170 @unnumberedsec Aspect Contract_Cases
8171 @findex Contract_Cases
8173 This aspect is equivalent to pragma @code{Contract_Cases}, the sequence
8174 of clauses being enclosed in parentheses so that syntactically it is an
8177 @node Aspect Depends
8178 @unnumberedsec Aspect Depends
8181 This aspect is equivalent to pragma @code{Depends}.
8183 @node Aspect Dimension
8184 @unnumberedsec Aspect Dimension
8187 The @code{Dimension} aspect is used to specify the dimensions of a given
8188 subtype of a dimensioned numeric type. The aspect also specifies a symbol
8189 used when doing formatted output of dimensioned quantities. The syntax is:
8191 @smallexample @c ada
8193 ([Symbol =>] SYMBOL, DIMENSION_VALUE @{, DIMENSION_Value@})
8195 SYMBOL ::= STRING_LITERAL | CHARACTER_LITERAL
8199 | others => RATIONAL
8200 | DISCRETE_CHOICE_LIST => RATIONAL
8202 RATIONAL ::= [-] NUMERIC_LITERAL [/ NUMERIC_LITERAL]
8206 This aspect can only be applied to a subtype whose parent type has
8207 a @code{Dimension_Systen} aspect. The aspect must specify values for
8208 all dimensions of the system. The rational values are the powers of the
8209 corresponding dimensions that are used by the compiler to verify that
8210 physical (numeric) computations are dimensionally consistent. For example,
8211 the computation of a force must result in dimensions (L => 1, M => 1, T => -2).
8212 For further examples of the usage
8213 of this aspect, see package @code{System.Dim.Mks}.
8214 Note that when the dimensioned type is an integer type, then any
8215 dimension value must be an integer literal.
8217 @node Aspect Dimension_System
8218 @unnumberedsec Aspect Dimension_System
8219 @findex Dimension_System
8221 The @code{Dimension_System} aspect is used to define a system of
8222 dimensions that will be used in subsequent subtype declarations with
8223 @code{Dimension} aspects that reference this system. The syntax is:
8225 @smallexample @c ada
8226 with Dimension_System => (DIMENSION @{, DIMENSION@});
8228 DIMENSION ::= ([Unit_Name =>] IDENTIFIER,
8229 [Unit_Symbol =>] SYMBOL,
8230 [Dim_Symbol =>] SYMBOL)
8232 SYMBOL ::= CHARACTER_LITERAL | STRING_LITERAL
8236 This aspect is applied to a type, which must be a numeric derived type
8237 (typically a floating-point type), that
8238 will represent values within the dimension system. Each @code{DIMENSION}
8239 corresponds to one particular dimension. A maximum of 7 dimensions may
8240 be specified. @code{Unit_Name} is the name of the dimension (for example
8241 @code{Meter}). @code{Unit_Symbol} is the shorthand used for quantities
8242 of this dimension (for example @code{m} for @code{Meter}).
8243 @code{Dim_Symbol} gives
8244 the identification within the dimension system (typically this is a
8245 single letter, e.g. @code{L} standing for length for unit name @code{Meter}).
8246 The @code{Unit_Symbol} is used in formatted output of dimensioned quantities.
8247 The @code{Dim_Symbol} is used in error messages when numeric operations have
8248 inconsistent dimensions.
8250 GNAT provides the standard definition of the International MKS system in
8251 the run-time package @code{System.Dim.Mks}. You can easily define
8252 similar packages for cgs units or British units, and define conversion factors
8253 between values in different systems. The MKS system is characterized by the
8256 @smallexample @c ada
8257 type Mks_Type is new Long_Long_Float with
8258 Dimension_System => (
8259 (Unit_Name => Meter, Unit_Symbol => 'm', Dim_Symbol => 'L'),
8260 (Unit_Name => Kilogram, Unit_Symbol => "kg", Dim_Symbol => 'M'),
8261 (Unit_Name => Second, Unit_Symbol => 's', Dim_Symbol => 'T'),
8262 (Unit_Name => Ampere, Unit_Symbol => 'A', Dim_Symbol => 'I'),
8263 (Unit_Name => Kelvin, Unit_Symbol => 'K', Dim_Symbol => '@'),
8264 (Unit_Name => Mole, Unit_Symbol => "mol", Dim_Symbol => 'N'),
8265 (Unit_Name => Candela, Unit_Symbol => "cd", Dim_Symbol => 'J'));
8269 Note that in the above type definition, we use the symbol @code{@@} to
8270 represent a theta character (avoiding the use of extended Latin-1
8271 characters in this context).
8273 See section ``Performing Dimensionality Analysis in GNAT'' in the GNAT Users
8274 Guide for detailed examples of use of the dimension system.
8276 @node Aspect Effective_Reads
8277 @unnumberedsec Aspect Effective_Reads
8278 @findex Effective_Reads
8280 This aspect is equivalent to pragma @code{Effective_Reads}.
8282 @node Aspect Effective_Writes
8283 @unnumberedsec Aspect Effective_Writes
8284 @findex Effective_Writes
8286 This aspect is equivalent to pragma @code{Effective_Writes}.
8288 @node Aspect Favor_Top_Level
8289 @unnumberedsec Aspect Favor_Top_Level
8290 @findex Favor_Top_Level
8292 This aspect is equivalent to pragma @code{Favor_Top_Level}.
8295 @unnumberedsec Aspect Global
8298 This aspect is equivalent to pragma @code{Global}.
8300 @node Aspect Initial_Condition
8301 @unnumberedsec Aspect Initial_Condition
8302 @findex Initial_Condition
8304 This aspect is equivalent to pragma @code{Initial_Condition}.
8306 @node Aspect Initializes
8307 @unnumberedsec Aspect Initializes
8310 This aspect is equivalent to pragma @code{Initializes}.
8312 @node Aspect Inline_Always
8313 @unnumberedsec Aspect Inline_Always
8314 @findex Inline_Always
8316 This aspect is equivalent to pragma @code{Inline_Always}.
8318 @node Aspect Invariant
8319 @unnumberedsec Aspect Invariant
8322 This aspect is equivalent to pragma @code{Invariant}. It is a
8323 synonym for the language defined aspect @code{Type_Invariant} except
8324 that it is separately controllable using pragma @code{Assertion_Policy}.
8326 @node Aspect Invariant'Class
8327 @unnumberedsec Aspect Invariant'Class
8328 @findex Invariant'Class
8330 This aspect is equivalent to pragma @code{Type_Invariant_Class}. It is a
8331 synonym for the language defined aspect @code{Type_Invariant'Class} except
8332 that it is separately controllable using pragma @code{Assertion_Policy}.
8334 @node Aspect Iterable
8335 @unnumberedsec Aspect Iterable
8338 This aspect is used in the GNAT-defined formal container packages, to provide
8339 a light-weight mechanism for loops over such containers, without the overhead
8340 imposed by the tampering checks of standard Ada2012 iterators. The value of the
8341 aspect is a aggregate with four named components: First, Next, Has_Element,
8342 and Element. The following is a typical example of use:
8344 @smallexample @c ada
8345 type List is private with
8346 Iterable => (First => First_Element,
8348 Has_Element => Get_Element,
8349 Element => List_Element);
8352 @item The value denoted by @code{First} must denote a primitive operation of
8353 the container type that returns a Cursor, which must a be a type declared in
8354 the container package.
8355 @item The value of @code{Next} is a primitive operation of the container type
8356 that takes a cursor and yields a cursor.
8357 @item @code{Has_Element} is an operation that applies to a cursor a yields an
8358 element of the container.
8359 @item @code{Element} is the type of the elements of the container type, and
8360 thus the result of the function denoted by Has_Element.
8363 @node Aspect Linker_Section
8364 @unnumberedsec Aspect Linker_Section
8365 @findex Linker_Section
8367 This aspect is equivalent to an @code{Linker_Section} pragma.
8369 @node Aspect Lock_Free
8370 @unnumberedsec Aspect Lock_Free
8373 This aspect is equivalent to pragma @code{Lock_Free}.
8375 @node Aspect No_Elaboration_Code_All
8376 @unnumberedsec Aspect No_Elaboration_Code_All
8377 @findex No_Elaboration_Code_All
8379 This aspect is equivalent to a @code{pragma No_Elaboration_Code_All}
8380 statement for a program unit.
8382 @node Aspect Object_Size
8383 @unnumberedsec Aspect Object_Size
8386 This aspect is equivalent to an @code{Object_Size} attribute definition
8389 @node Aspect Obsolescent
8390 @unnumberedsec Aspect Obsolescent
8393 This aspect is equivalent to an @code{Obsolescent} pragma. Note that the
8394 evaluation of this aspect happens at the point of occurrence, it is not
8395 delayed until the freeze point.
8397 @node Aspect Part_Of
8398 @unnumberedsec Aspect Part_Of
8401 This aspect is equivalent to pragma @code{Part_Of}.
8403 @node Aspect Persistent_BSS
8404 @unnumberedsec Aspect Persistent_BSS
8405 @findex Persistent_BSS
8407 This aspect is equivalent to pragma @code{Persistent_BSS}.
8409 @node Aspect Predicate
8410 @unnumberedsec Aspect Predicate
8413 This aspect is equivalent to pragma @code{Predicate}. It is thus
8414 similar to the language defined aspects @code{Dynamic_Predicate}
8415 and @code{Static_Predicate} except that whether the resulting
8416 predicate is static or dynamic is controlled by the form of the
8417 expression. It is also separately controllable using pragma
8418 @code{Assertion_Policy}.
8420 @node Aspect Pure_Function
8421 @unnumberedsec Aspect Pure_Function
8422 @findex Pure_Function
8424 This aspect is equivalent to pragma @code{Pure_Function}.
8426 @node Aspect Refined_Depends
8427 @unnumberedsec Aspect Refined_Depends
8428 @findex Refined_Depends
8430 This aspect is equivalent to pragma @code{Refined_Depends}.
8432 @node Aspect Refined_Global
8433 @unnumberedsec Aspect Refined_Global
8434 @findex Refined_Global
8436 This aspect is equivalent to pragma @code{Refined_Global}.
8438 @node Aspect Refined_Post
8439 @unnumberedsec Aspect Refined_Post
8440 @findex Refined_Post
8442 This aspect is equivalent to pragma @code{Refined_Post}.
8444 @node Aspect Refined_State
8445 @unnumberedsec Aspect Refined_State
8446 @findex Refined_State
8448 This aspect is equivalent to pragma @code{Refined_State}.
8450 @node Aspect Remote_Access_Type
8451 @unnumberedsec Aspect Remote_Access_Type
8452 @findex Remote_Access_Type
8454 This aspect is equivalent to pragma @code{Remote_Access_Type}.
8456 @node Aspect Scalar_Storage_Order
8457 @unnumberedsec Aspect Scalar_Storage_Order
8458 @findex Scalar_Storage_Order
8460 This aspect is equivalent to a @code{Scalar_Storage_Order}
8461 attribute definition clause.
8464 @unnumberedsec Aspect Shared
8467 This aspect is equivalent to pragma @code{Shared}, and is thus a synonym
8468 for aspect @code{Atomic}.
8470 @node Aspect Simple_Storage_Pool
8471 @unnumberedsec Aspect Simple_Storage_Pool
8472 @findex Simple_Storage_Pool
8474 This aspect is equivalent to a @code{Simple_Storage_Pool}
8475 attribute definition clause.
8477 @node Aspect Simple_Storage_Pool_Type
8478 @unnumberedsec Aspect Simple_Storage_Pool_Type
8479 @findex Simple_Storage_Pool_Type
8481 This aspect is equivalent to pragma @code{Simple_Storage_Pool_Type}.
8483 @node Aspect SPARK_Mode
8484 @unnumberedsec Aspect SPARK_Mode
8487 This aspect is equivalent to pragma @code{SPARK_Mode} and
8488 may be specified for either or both of the specification and body
8489 of a subprogram or package.
8491 @node Aspect Suppress_Debug_Info
8492 @unnumberedsec Aspect Suppress_Debug_Info
8493 @findex Suppress_Debug_Info
8495 This aspect is equivalent to pragma @code{Suppress_Debug_Info}.
8497 @node Aspect Test_Case
8498 @unnumberedsec Aspect Test_Case
8501 This aspect is equivalent to pragma @code{Test_Case}.
8503 @node Aspect Thread_Local_Storage
8504 @unnumberedsec Aspect Thread_Local_Storage
8505 @findex Thread_Local_Storage
8507 This aspect is equivalent to pragma @code{Thread_Local_Storage}.
8509 @node Aspect Universal_Aliasing
8510 @unnumberedsec Aspect Universal_Aliasing
8511 @findex Universal_Aliasing
8513 This aspect is equivalent to pragma @code{Universal_Aliasing}.
8515 @node Aspect Universal_Data
8516 @unnumberedsec Aspect Universal_Data
8517 @findex Universal_Data
8519 This aspect is equivalent to pragma @code{Universal_Data}.
8521 @node Aspect Unmodified
8522 @unnumberedsec Aspect Unmodified
8525 This aspect is equivalent to pragma @code{Unmodified}.
8527 @node Aspect Unreferenced
8528 @unnumberedsec Aspect Unreferenced
8529 @findex Unreferenced
8531 This aspect is equivalent to pragma @code{Unreferenced}.
8533 @node Aspect Unreferenced_Objects
8534 @unnumberedsec Aspect Unreferenced_Objects
8535 @findex Unreferenced_Objects
8537 This aspect is equivalent to pragma @code{Unreferenced_Objects}.
8539 @node Aspect Value_Size
8540 @unnumberedsec Aspect Value_Size
8543 This aspect is equivalent to a @code{Value_Size}
8544 attribute definition clause.
8546 @node Aspect Warnings
8547 @unnumberedsec Aspect Warnings
8550 This aspect is equivalent to the two argument form of pragma @code{Warnings},
8551 where the first argument is @code{ON} or @code{OFF} and the second argument
8555 @node Implementation Defined Attributes
8556 @chapter Implementation Defined Attributes
8557 Ada defines (throughout the Ada reference manual,
8558 summarized in Annex K),
8559 a set of attributes that provide useful additional functionality in all
8560 areas of the language. These language defined attributes are implemented
8561 in GNAT and work as described in the Ada Reference Manual.
8563 In addition, Ada allows implementations to define additional
8564 attributes whose meaning is defined by the implementation. GNAT provides
8565 a number of these implementation-dependent attributes which can be used
8566 to extend and enhance the functionality of the compiler. This section of
8567 the GNAT reference manual describes these additional attributes. It also
8568 describes additional implementation-dependent features of standard
8569 language-defined attributes.
8571 Note that any program using these attributes may not be portable to
8572 other compilers (although GNAT implements this set of attributes on all
8573 platforms). Therefore if portability to other compilers is an important
8574 consideration, you should minimize the use of these attributes.
8577 * Attribute Abort_Signal::
8578 * Attribute Address_Size::
8579 * Attribute Asm_Input::
8580 * Attribute Asm_Output::
8581 * Attribute AST_Entry::
8582 * Attribute Atomic_Always_Lock_Free::
8584 * Attribute Bit_Position::
8585 * Attribute Code_Address::
8586 * Attribute Compiler_Version::
8587 * Attribute Constrained::
8588 * Attribute Default_Bit_Order::
8589 * Attribute Default_Scalar_Storage_Order::
8590 * Attribute Descriptor_Size::
8591 * Attribute Elaborated::
8592 * Attribute Elab_Body::
8593 * Attribute Elab_Spec::
8594 * Attribute Elab_Subp_Body::
8596 * Attribute Enabled::
8597 * Attribute Enum_Rep::
8598 * Attribute Enum_Val::
8599 * Attribute Epsilon::
8600 * Attribute Fast_Math::
8601 * Attribute Fixed_Value::
8602 * Attribute From_Any::
8603 * Attribute Has_Access_Values::
8604 * Attribute Has_Discriminants::
8606 * Attribute Integer_Value::
8607 * Attribute Invalid_Value::
8608 * Attribute Iterable::
8610 * Attribute Library_Level::
8611 * Attribute Lock_Free::
8612 * Attribute Loop_Entry::
8613 * Attribute Machine_Size::
8614 * Attribute Mantissa::
8615 * Attribute Maximum_Alignment::
8616 * Attribute Mechanism_Code::
8617 * Attribute Null_Parameter::
8618 * Attribute Object_Size::
8620 * Attribute Passed_By_Reference::
8621 * Attribute Pool_Address::
8622 * Attribute Range_Length::
8624 * Attribute Restriction_Set::
8625 * Attribute Result::
8626 * Attribute Safe_Emax::
8627 * Attribute Safe_Large::
8628 * Attribute Safe_Small::
8629 * Attribute Scalar_Storage_Order::
8630 * Attribute Simple_Storage_Pool::
8632 * Attribute Storage_Unit::
8633 * Attribute Stub_Type::
8634 * Attribute System_Allocator_Alignment::
8635 * Attribute Target_Name::
8636 * Attribute To_Address::
8637 * Attribute To_Any::
8638 * Attribute Type_Class::
8639 * Attribute Type_Key::
8640 * Attribute TypeCode::
8641 * Attribute UET_Address::
8642 * Attribute Unconstrained_Array::
8643 * Attribute Universal_Literal_String::
8644 * Attribute Unrestricted_Access::
8645 * Attribute Update::
8646 * Attribute VADS_Size::
8647 * Attribute Valid_Scalars::
8648 * Attribute Value_Size::
8649 * Attribute Wchar_T_Size::
8650 * Attribute Word_Size::
8653 @node Attribute Abort_Signal
8654 @unnumberedsec Attribute Abort_Signal
8655 @findex Abort_Signal
8657 @code{Standard'Abort_Signal} (@code{Standard} is the only allowed
8658 prefix) provides the entity for the special exception used to signal
8659 task abort or asynchronous transfer of control. Normally this attribute
8660 should only be used in the tasking runtime (it is highly peculiar, and
8661 completely outside the normal semantics of Ada, for a user program to
8662 intercept the abort exception).
8664 @node Attribute Address_Size
8665 @unnumberedsec Attribute Address_Size
8666 @cindex Size of @code{Address}
8667 @findex Address_Size
8669 @code{Standard'Address_Size} (@code{Standard} is the only allowed
8670 prefix) is a static constant giving the number of bits in an
8671 @code{Address}. It is the same value as System.Address'Size,
8672 but has the advantage of being static, while a direct
8673 reference to System.Address'Size is non-static because Address
8676 @node Attribute Asm_Input
8677 @unnumberedsec Attribute Asm_Input
8680 The @code{Asm_Input} attribute denotes a function that takes two
8681 parameters. The first is a string, the second is an expression of the
8682 type designated by the prefix. The first (string) argument is required
8683 to be a static expression, and is the constraint for the parameter,
8684 (e.g.@: what kind of register is required). The second argument is the
8685 value to be used as the input argument. The possible values for the
8686 constant are the same as those used in the RTL, and are dependent on
8687 the configuration file used to built the GCC back end.
8688 @ref{Machine Code Insertions}
8690 @node Attribute Asm_Output
8691 @unnumberedsec Attribute Asm_Output
8694 The @code{Asm_Output} attribute denotes a function that takes two
8695 parameters. The first is a string, the second is the name of a variable
8696 of the type designated by the attribute prefix. The first (string)
8697 argument is required to be a static expression and designates the
8698 constraint for the parameter (e.g.@: what kind of register is
8699 required). The second argument is the variable to be updated with the
8700 result. The possible values for constraint are the same as those used in
8701 the RTL, and are dependent on the configuration file used to build the
8702 GCC back end. If there are no output operands, then this argument may
8703 either be omitted, or explicitly given as @code{No_Output_Operands}.
8704 @ref{Machine Code Insertions}
8706 @node Attribute AST_Entry
8707 @unnumberedsec Attribute AST_Entry
8711 This attribute is implemented only in OpenVMS versions of GNAT@. Applied to
8712 the name of an entry, it yields a value of the predefined type AST_Handler
8713 (declared in the predefined package System, as extended by the use of
8714 pragma @code{Extend_System (Aux_DEC)}). This value enables the given entry to
8715 be called when an AST occurs. For further details, refer to the @cite{DEC Ada
8716 Language Reference Manual}, section 9.12a.
8718 @node Attribute Atomic_Always_Lock_Free
8719 @unnumberedsec Attribute Atomic_Always_Lock_Free
8720 @findex Atomic_Always_Lock_Free
8723 The prefix of the @code{Atomic_Always_Lock_Free} attribute is a type.
8724 The result is a Boolean value which is True if the type has discriminants,
8725 and False otherwise. The result indicate whether atomic operations are
8726 supported by the target for the given type.
8729 @unnumberedsec Attribute Bit
8731 @code{@var{obj}'Bit}, where @var{obj} is any object, yields the bit
8732 offset within the storage unit (byte) that contains the first bit of
8733 storage allocated for the object. The value of this attribute is of the
8734 type @code{Universal_Integer}, and is always a non-negative number not
8735 exceeding the value of @code{System.Storage_Unit}.
8737 For an object that is a variable or a constant allocated in a register,
8738 the value is zero. (The use of this attribute does not force the
8739 allocation of a variable to memory).
8741 For an object that is a formal parameter, this attribute applies
8742 to either the matching actual parameter or to a copy of the
8743 matching actual parameter.
8745 For an access object the value is zero. Note that
8746 @code{@var{obj}.all'Bit} is subject to an @code{Access_Check} for the
8747 designated object. Similarly for a record component
8748 @code{@var{X}.@var{C}'Bit} is subject to a discriminant check and
8749 @code{@var{X}(@var{I}).Bit} and @code{@var{X}(@var{I1}..@var{I2})'Bit}
8750 are subject to index checks.
8752 This attribute is designed to be compatible with the DEC Ada 83 definition
8753 and implementation of the @code{Bit} attribute.
8755 @node Attribute Bit_Position
8756 @unnumberedsec Attribute Bit_Position
8757 @findex Bit_Position
8759 @code{@var{R.C}'Bit_Position}, where @var{R} is a record object and C is one
8760 of the fields of the record type, yields the bit
8761 offset within the record contains the first bit of
8762 storage allocated for the object. The value of this attribute is of the
8763 type @code{Universal_Integer}. The value depends only on the field
8764 @var{C} and is independent of the alignment of
8765 the containing record @var{R}.
8767 @node Attribute Code_Address
8768 @unnumberedsec Attribute Code_Address
8769 @findex Code_Address
8770 @cindex Subprogram address
8771 @cindex Address of subprogram code
8774 attribute may be applied to subprograms in Ada 95 and Ada 2005, but the
8775 intended effect seems to be to provide
8776 an address value which can be used to call the subprogram by means of
8777 an address clause as in the following example:
8779 @smallexample @c ada
8780 procedure K is @dots{}
8783 for L'Address use K'Address;
8784 pragma Import (Ada, L);
8788 A call to @code{L} is then expected to result in a call to @code{K}@.
8789 In Ada 83, where there were no access-to-subprogram values, this was
8790 a common work-around for getting the effect of an indirect call.
8791 GNAT implements the above use of @code{Address} and the technique
8792 illustrated by the example code works correctly.
8794 However, for some purposes, it is useful to have the address of the start
8795 of the generated code for the subprogram. On some architectures, this is
8796 not necessarily the same as the @code{Address} value described above.
8797 For example, the @code{Address} value may reference a subprogram
8798 descriptor rather than the subprogram itself.
8800 The @code{'Code_Address} attribute, which can only be applied to
8801 subprogram entities, always returns the address of the start of the
8802 generated code of the specified subprogram, which may or may not be
8803 the same value as is returned by the corresponding @code{'Address}
8806 @node Attribute Compiler_Version
8807 @unnumberedsec Attribute Compiler_Version
8808 @findex Compiler_Version
8810 @code{Standard'Compiler_Version} (@code{Standard} is the only allowed
8811 prefix) yields a static string identifying the version of the compiler
8812 being used to compile the unit containing the attribute reference. A
8813 typical result would be something like
8814 "@value{EDITION} @value{gnat_version} (20090221)".
8816 @node Attribute Constrained
8817 @unnumberedsec Attribute Constrained
8820 In addition to the usage of this attribute in the Ada RM, @code{GNAT}
8821 also permits the use of the @code{'Constrained} attribute
8822 in a generic template
8823 for any type, including types without discriminants. The value of this
8824 attribute in the generic instance when applied to a scalar type or a
8825 record type without discriminants is always @code{True}. This usage is
8826 compatible with older Ada compilers, including notably DEC Ada.
8828 @node Attribute Default_Bit_Order
8829 @unnumberedsec Attribute Default_Bit_Order
8831 @cindex Little endian
8832 @findex Default_Bit_Order
8834 @code{Standard'Default_Bit_Order} (@code{Standard} is the only
8835 permissible prefix), provides the value @code{System.Default_Bit_Order}
8836 as a @code{Pos} value (0 for @code{High_Order_First}, 1 for
8837 @code{Low_Order_First}). This is used to construct the definition of
8838 @code{Default_Bit_Order} in package @code{System}.
8840 @node Attribute Default_Scalar_Storage_Order
8841 @unnumberedsec Attribute Default_Scalar_Storage_Order
8843 @cindex Little endian
8844 @findex Default_Scalar_Storage_Order
8846 @code{Standard'Default_Scalar_Storage_Order} (@code{Standard} is the only
8847 permissible prefix), provides the current value of the default scalar storage
8848 order (as specified using pragma @code{Default_Scalar_Storage_Order}, or
8849 equal to @code{Default_Bit_Order} if unspecified) as a
8850 @code{System.Bit_Order} value. This is a static attribute.
8852 @node Attribute Descriptor_Size
8853 @unnumberedsec Attribute Descriptor_Size
8856 @findex Descriptor_Size
8858 Non-static attribute @code{Descriptor_Size} returns the size in bits of the
8859 descriptor allocated for a type. The result is non-zero only for unconstrained
8860 array types and the returned value is of type universal integer. In GNAT, an
8861 array descriptor contains bounds information and is located immediately before
8862 the first element of the array.
8864 @smallexample @c ada
8865 type Unconstr_Array is array (Positive range <>) of Boolean;
8866 Put_Line ("Descriptor size = " & Unconstr_Array'Descriptor_Size'Img);
8870 The attribute takes into account any additional padding due to type alignment.
8871 In the example above, the descriptor contains two values of type
8872 @code{Positive} representing the low and high bound. Since @code{Positive} has
8873 a size of 31 bits and an alignment of 4, the descriptor size is @code{2 *
8874 Positive'Size + 2} or 64 bits.
8876 @node Attribute Elaborated
8877 @unnumberedsec Attribute Elaborated
8880 The prefix of the @code{'Elaborated} attribute must be a unit name. The
8881 value is a Boolean which indicates whether or not the given unit has been
8882 elaborated. This attribute is primarily intended for internal use by the
8883 generated code for dynamic elaboration checking, but it can also be used
8884 in user programs. The value will always be True once elaboration of all
8885 units has been completed. An exception is for units which need no
8886 elaboration, the value is always False for such units.
8888 @node Attribute Elab_Body
8889 @unnumberedsec Attribute Elab_Body
8892 This attribute can only be applied to a program unit name. It returns
8893 the entity for the corresponding elaboration procedure for elaborating
8894 the body of the referenced unit. This is used in the main generated
8895 elaboration procedure by the binder and is not normally used in any
8896 other context. However, there may be specialized situations in which it
8897 is useful to be able to call this elaboration procedure from Ada code,
8898 e.g.@: if it is necessary to do selective re-elaboration to fix some
8901 @node Attribute Elab_Spec
8902 @unnumberedsec Attribute Elab_Spec
8905 This attribute can only be applied to a program unit name. It returns
8906 the entity for the corresponding elaboration procedure for elaborating
8907 the spec of the referenced unit. This is used in the main
8908 generated elaboration procedure by the binder and is not normally used
8909 in any other context. However, there may be specialized situations in
8910 which it is useful to be able to call this elaboration procedure from
8911 Ada code, e.g.@: if it is necessary to do selective re-elaboration to fix
8914 @node Attribute Elab_Subp_Body
8915 @unnumberedsec Attribute Elab_Subp_Body
8916 @findex Elab_Subp_Body
8918 This attribute can only be applied to a library level subprogram
8919 name and is only allowed in CodePeer mode. It returns the entity
8920 for the corresponding elaboration procedure for elaborating the body
8921 of the referenced subprogram unit. This is used in the main generated
8922 elaboration procedure by the binder in CodePeer mode only and is unrecognized
8925 @node Attribute Emax
8926 @unnumberedsec Attribute Emax
8927 @cindex Ada 83 attributes
8930 The @code{Emax} attribute is provided for compatibility with Ada 83. See
8931 the Ada 83 reference manual for an exact description of the semantics of
8934 @node Attribute Enabled
8935 @unnumberedsec Attribute Enabled
8938 The @code{Enabled} attribute allows an application program to check at compile
8939 time to see if the designated check is currently enabled. The prefix is a
8940 simple identifier, referencing any predefined check name (other than
8941 @code{All_Checks}) or a check name introduced by pragma Check_Name. If
8942 no argument is given for the attribute, the check is for the general state
8943 of the check, if an argument is given, then it is an entity name, and the
8944 check indicates whether an @code{Suppress} or @code{Unsuppress} has been
8945 given naming the entity (if not, then the argument is ignored).
8947 Note that instantiations inherit the check status at the point of the
8948 instantiation, so a useful idiom is to have a library package that
8949 introduces a check name with @code{pragma Check_Name}, and then contains
8950 generic packages or subprograms which use the @code{Enabled} attribute
8951 to see if the check is enabled. A user of this package can then issue
8952 a @code{pragma Suppress} or @code{pragma Unsuppress} before instantiating
8953 the package or subprogram, controlling whether the check will be present.
8955 @node Attribute Enum_Rep
8956 @unnumberedsec Attribute Enum_Rep
8957 @cindex Representation of enums
8960 For every enumeration subtype @var{S}, @code{@var{S}'Enum_Rep} denotes a
8961 function with the following spec:
8963 @smallexample @c ada
8964 function @var{S}'Enum_Rep (Arg : @var{S}'Base)
8965 return @i{Universal_Integer};
8969 It is also allowable to apply @code{Enum_Rep} directly to an object of an
8970 enumeration type or to a non-overloaded enumeration
8971 literal. In this case @code{@var{S}'Enum_Rep} is equivalent to
8972 @code{@var{typ}'Enum_Rep(@var{S})} where @var{typ} is the type of the
8973 enumeration literal or object.
8975 The function returns the representation value for the given enumeration
8976 value. This will be equal to value of the @code{Pos} attribute in the
8977 absence of an enumeration representation clause. This is a static
8978 attribute (i.e.@: the result is static if the argument is static).
8980 @code{@var{S}'Enum_Rep} can also be used with integer types and objects,
8981 in which case it simply returns the integer value. The reason for this
8982 is to allow it to be used for @code{(<>)} discrete formal arguments in
8983 a generic unit that can be instantiated with either enumeration types
8984 or integer types. Note that if @code{Enum_Rep} is used on a modular
8985 type whose upper bound exceeds the upper bound of the largest signed
8986 integer type, and the argument is a variable, so that the universal
8987 integer calculation is done at run time, then the call to @code{Enum_Rep}
8988 may raise @code{Constraint_Error}.
8990 @node Attribute Enum_Val
8991 @unnumberedsec Attribute Enum_Val
8992 @cindex Representation of enums
8995 For every enumeration subtype @var{S}, @code{@var{S}'Enum_Val} denotes a
8996 function with the following spec:
8998 @smallexample @c ada
8999 function @var{S}'Enum_Val (Arg : @i{Universal_Integer)
9000 return @var{S}'Base};
9004 The function returns the enumeration value whose representation matches the
9005 argument, or raises Constraint_Error if no enumeration literal of the type
9006 has the matching value.
9007 This will be equal to value of the @code{Val} attribute in the
9008 absence of an enumeration representation clause. This is a static
9009 attribute (i.e.@: the result is static if the argument is static).
9011 @node Attribute Epsilon
9012 @unnumberedsec Attribute Epsilon
9013 @cindex Ada 83 attributes
9016 The @code{Epsilon} attribute is provided for compatibility with Ada 83. See
9017 the Ada 83 reference manual for an exact description of the semantics of
9020 @node Attribute Fast_Math
9021 @unnumberedsec Attribute Fast_Math
9024 @code{Standard'Fast_Math} (@code{Standard} is the only allowed
9025 prefix) yields a static Boolean value that is True if pragma
9026 @code{Fast_Math} is active, and False otherwise.
9028 @node Attribute Fixed_Value
9029 @unnumberedsec Attribute Fixed_Value
9032 For every fixed-point type @var{S}, @code{@var{S}'Fixed_Value} denotes a
9033 function with the following specification:
9035 @smallexample @c ada
9036 function @var{S}'Fixed_Value (Arg : @i{Universal_Integer})
9041 The value returned is the fixed-point value @var{V} such that
9043 @smallexample @c ada
9044 @var{V} = Arg * @var{S}'Small
9048 The effect is thus similar to first converting the argument to the
9049 integer type used to represent @var{S}, and then doing an unchecked
9050 conversion to the fixed-point type. The difference is
9051 that there are full range checks, to ensure that the result is in range.
9052 This attribute is primarily intended for use in implementation of the
9053 input-output functions for fixed-point values.
9055 @node Attribute From_Any
9056 @unnumberedsec Attribute From_Any
9059 This internal attribute is used for the generation of remote subprogram
9060 stubs in the context of the Distributed Systems Annex.
9062 @node Attribute Has_Access_Values
9063 @unnumberedsec Attribute Has_Access_Values
9064 @cindex Access values, testing for
9065 @findex Has_Access_Values
9067 The prefix of the @code{Has_Access_Values} attribute is a type. The result
9068 is a Boolean value which is True if the is an access type, or is a composite
9069 type with a component (at any nesting depth) that is an access type, and is
9071 The intended use of this attribute is in conjunction with generic
9072 definitions. If the attribute is applied to a generic private type, it
9073 indicates whether or not the corresponding actual type has access values.
9075 @node Attribute Has_Discriminants
9076 @unnumberedsec Attribute Has_Discriminants
9077 @cindex Discriminants, testing for
9078 @findex Has_Discriminants
9080 The prefix of the @code{Has_Discriminants} attribute is a type. The result
9081 is a Boolean value which is True if the type has discriminants, and False
9082 otherwise. The intended use of this attribute is in conjunction with generic
9083 definitions. If the attribute is applied to a generic private type, it
9084 indicates whether or not the corresponding actual type has discriminants.
9087 @unnumberedsec Attribute Img
9090 The @code{Img} attribute differs from @code{Image} in that it is applied
9091 directly to an object, and yields the same result as
9092 @code{Image} for the subtype of the object. This is convenient for
9095 @smallexample @c ada
9096 Put_Line ("X = " & X'Img);
9100 has the same meaning as the more verbose:
9102 @smallexample @c ada
9103 Put_Line ("X = " & @var{T}'Image (X));
9107 where @var{T} is the (sub)type of the object @code{X}.
9109 Note that technically, in analogy to @code{Image},
9110 @code{X'Img} returns a parameterless function
9111 that returns the appropriate string when called. This means that
9112 @code{X'Img} can be renamed as a function-returning-string, or used
9113 in an instantiation as a function parameter.
9115 @node Attribute Integer_Value
9116 @unnumberedsec Attribute Integer_Value
9117 @findex Integer_Value
9119 For every integer type @var{S}, @code{@var{S}'Integer_Value} denotes a
9120 function with the following spec:
9122 @smallexample @c ada
9123 function @var{S}'Integer_Value (Arg : @i{Universal_Fixed})
9128 The value returned is the integer value @var{V}, such that
9130 @smallexample @c ada
9131 Arg = @var{V} * @var{T}'Small
9135 where @var{T} is the type of @code{Arg}.
9136 The effect is thus similar to first doing an unchecked conversion from
9137 the fixed-point type to its corresponding implementation type, and then
9138 converting the result to the target integer type. The difference is
9139 that there are full range checks, to ensure that the result is in range.
9140 This attribute is primarily intended for use in implementation of the
9141 standard input-output functions for fixed-point values.
9143 @node Attribute Invalid_Value
9144 @unnumberedsec Attribute Invalid_Value
9145 @findex Invalid_Value
9147 For every scalar type S, S'Invalid_Value returns an undefined value of the
9148 type. If possible this value is an invalid representation for the type. The
9149 value returned is identical to the value used to initialize an otherwise
9150 uninitialized value of the type if pragma Initialize_Scalars is used,
9151 including the ability to modify the value with the binder -Sxx flag and
9152 relevant environment variables at run time.
9154 @node Attribute Iterable
9155 @unnumberedsec Attribute Iterable
9158 Equivalent to Aspect Iterable.
9160 @node Attribute Large
9161 @unnumberedsec Attribute Large
9162 @cindex Ada 83 attributes
9165 The @code{Large} attribute is provided for compatibility with Ada 83. See
9166 the Ada 83 reference manual for an exact description of the semantics of
9169 @node Attribute Library_Level
9170 @unnumberedsec Attribute Library_Level
9171 @findex Library_Level
9174 @code{P'Library_Level}, where P is an entity name,
9175 returns a Boolean value which is True if the entity is declared
9176 at the library level, and False otherwise. Note that within a
9177 generic instantition, the name of the generic unit denotes the
9178 instance, which means that this attribute can be used to test
9179 if a generic is instantiated at the library level, as shown
9182 @smallexample @c ada
9186 pragma Compile_Time_Error
9187 (not Gen'Library_Level,
9188 "Gen can only be instantiated at library level");
9193 @node Attribute Lock_Free
9194 @unnumberedsec Attribute Lock_Free
9197 @code{P'Lock_Free}, where P is a protected object, returns True if a
9198 pragma @code{Lock_Free} applies to P.
9200 @node Attribute Loop_Entry
9201 @unnumberedsec Attribute Loop_Entry
9206 @smallexample @c ada
9207 X'Loop_Entry [(loop_name)]
9211 The @code{Loop_Entry} attribute is used to refer to the value that an
9212 expression had upon entry to a given loop in much the same way that the
9213 @code{Old} attribute in a subprogram postcondition can be used to refer
9214 to the value an expression had upon entry to the subprogram. The
9215 relevant loop is either identified by the given loop name, or it is the
9216 innermost enclosing loop when no loop name is given.
9219 A @code{Loop_Entry} attribute can only occur within a
9220 @code{Loop_Variant} or @code{Loop_Invariant} pragma. A common use of
9221 @code{Loop_Entry} is to compare the current value of objects with their
9222 initial value at loop entry, in a @code{Loop_Invariant} pragma.
9225 The effect of using @code{X'Loop_Entry} is the same as declaring
9226 a constant initialized with the initial value of @code{X} at loop
9227 entry. This copy is not performed if the loop is not entered, or if the
9228 corresponding pragmas are ignored or disabled.
9230 @node Attribute Machine_Size
9231 @unnumberedsec Attribute Machine_Size
9232 @findex Machine_Size
9234 This attribute is identical to the @code{Object_Size} attribute. It is
9235 provided for compatibility with the DEC Ada 83 attribute of this name.
9237 @node Attribute Mantissa
9238 @unnumberedsec Attribute Mantissa
9239 @cindex Ada 83 attributes
9242 The @code{Mantissa} attribute is provided for compatibility with Ada 83. See
9243 the Ada 83 reference manual for an exact description of the semantics of
9246 @node Attribute Maximum_Alignment
9247 @unnumberedsec Attribute Maximum_Alignment
9248 @cindex Alignment, maximum
9249 @findex Maximum_Alignment
9251 @code{Standard'Maximum_Alignment} (@code{Standard} is the only
9252 permissible prefix) provides the maximum useful alignment value for the
9253 target. This is a static value that can be used to specify the alignment
9254 for an object, guaranteeing that it is properly aligned in all
9257 @node Attribute Mechanism_Code
9258 @unnumberedsec Attribute Mechanism_Code
9259 @cindex Return values, passing mechanism
9260 @cindex Parameters, passing mechanism
9261 @findex Mechanism_Code
9263 @code{@var{function}'Mechanism_Code} yields an integer code for the
9264 mechanism used for the result of function, and
9265 @code{@var{subprogram}'Mechanism_Code (@var{n})} yields the mechanism
9266 used for formal parameter number @var{n} (a static integer value with 1
9267 meaning the first parameter) of @var{subprogram}. The code returned is:
9276 @node Attribute Null_Parameter
9277 @unnumberedsec Attribute Null_Parameter
9278 @cindex Zero address, passing
9279 @findex Null_Parameter
9281 A reference @code{@var{T}'Null_Parameter} denotes an imaginary object of
9282 type or subtype @var{T} allocated at machine address zero. The attribute
9283 is allowed only as the default expression of a formal parameter, or as
9284 an actual expression of a subprogram call. In either case, the
9285 subprogram must be imported.
9287 The identity of the object is represented by the address zero in the
9288 argument list, independent of the passing mechanism (explicit or
9291 This capability is needed to specify that a zero address should be
9292 passed for a record or other composite object passed by reference.
9293 There is no way of indicating this without the @code{Null_Parameter}
9296 @node Attribute Object_Size
9297 @unnumberedsec Attribute Object_Size
9298 @cindex Size, used for objects
9301 The size of an object is not necessarily the same as the size of the type
9302 of an object. This is because by default object sizes are increased to be
9303 a multiple of the alignment of the object. For example,
9304 @code{Natural'Size} is
9305 31, but by default objects of type @code{Natural} will have a size of 32 bits.
9306 Similarly, a record containing an integer and a character:
9308 @smallexample @c ada
9316 will have a size of 40 (that is @code{Rec'Size} will be 40). The
9317 alignment will be 4, because of the
9318 integer field, and so the default size of record objects for this type
9319 will be 64 (8 bytes).
9321 If the alignment of the above record is specified to be 1, then the
9322 object size will be 40 (5 bytes). This is true by default, and also
9323 an object size of 40 can be explicitly specified in this case.
9325 A consequence of this capability is that different object sizes can be
9326 given to subtypes that would otherwise be considered in Ada to be
9327 statically matching. But it makes no sense to consider such subtypes
9328 as statically matching. Consequently, in @code{GNAT} we add a rule
9329 to the static matching rules that requires object sizes to match.
9330 Consider this example:
9332 @smallexample @c ada
9333 1. procedure BadAVConvert is
9334 2. type R is new Integer;
9335 3. subtype R1 is R range 1 .. 10;
9336 4. subtype R2 is R range 1 .. 10;
9337 5. for R1'Object_Size use 8;
9338 6. for R2'Object_Size use 16;
9339 7. type R1P is access all R1;
9340 8. type R2P is access all R2;
9341 9. R1PV : R1P := new R1'(4);
9344 12. R2PV := R2P (R1PV);
9346 >>> target designated subtype not compatible with
9347 type "R1" defined at line 3
9353 In the absence of lines 5 and 6,
9354 types @code{R1} and @code{R2} statically match and
9355 hence the conversion on line 12 is legal. But since lines 5 and 6
9356 cause the object sizes to differ, @code{GNAT} considers that types
9357 @code{R1} and @code{R2} are not statically matching, and line 12
9358 generates the diagnostic shown above.
9361 Similar additional checks are performed in other contexts requiring
9362 statically matching subtypes.
9365 @unnumberedsec Attribute Old
9368 In addition to the usage of @code{Old} defined in the Ada 2012 RM (usage
9369 within @code{Post} aspect), GNAT also permits the use of this attribute
9370 in implementation defined pragmas @code{Postcondition},
9371 @code{Contract_Cases} and @code{Test_Case}. Also usages of
9372 @code{Old} which would be illegal according to the Ada 2012 RM
9373 definition are allowed under control of
9374 implementation defined pragma @code{Unevaluated_Use_Of_Old}.
9376 @node Attribute Passed_By_Reference
9377 @unnumberedsec Attribute Passed_By_Reference
9378 @cindex Parameters, when passed by reference
9379 @findex Passed_By_Reference
9381 @code{@var{type}'Passed_By_Reference} for any subtype @var{type} returns
9382 a value of type @code{Boolean} value that is @code{True} if the type is
9383 normally passed by reference and @code{False} if the type is normally
9384 passed by copy in calls. For scalar types, the result is always @code{False}
9385 and is static. For non-scalar types, the result is non-static.
9387 @node Attribute Pool_Address
9388 @unnumberedsec Attribute Pool_Address
9389 @cindex Parameters, when passed by reference
9390 @findex Pool_Address
9392 @code{@var{X}'Pool_Address} for any object @var{X} returns the address
9393 of X within its storage pool. This is the same as
9394 @code{@var{X}'Address}, except that for an unconstrained array whose
9395 bounds are allocated just before the first component,
9396 @code{@var{X}'Pool_Address} returns the address of those bounds,
9397 whereas @code{@var{X}'Address} returns the address of the first
9400 Here, we are interpreting ``storage pool'' broadly to mean ``wherever
9401 the object is allocated'', which could be a user-defined storage pool,
9402 the global heap, on the stack, or in a static memory area. For an
9403 object created by @code{new}, @code{@var{Ptr.all}'Pool_Address} is
9404 what is passed to @code{Allocate} and returned from @code{Deallocate}.
9406 @node Attribute Range_Length
9407 @unnumberedsec Attribute Range_Length
9408 @findex Range_Length
9410 @code{@var{type}'Range_Length} for any discrete type @var{type} yields
9411 the number of values represented by the subtype (zero for a null
9412 range). The result is static for static subtypes. @code{Range_Length}
9413 applied to the index subtype of a one dimensional array always gives the
9414 same result as @code{Length} applied to the array itself.
9417 @unnumberedsec Attribute Ref
9422 @node Attribute Restriction_Set
9423 @unnumberedsec Attribute Restriction_Set
9424 @findex Restriction_Set
9425 @cindex Restrictions
9427 This attribute allows compile time testing of restrictions that
9428 are currently in effect. It is primarily intended for specializing
9429 code in the run-time based on restrictions that are active (e.g.
9430 don't need to save fpt registers if restriction No_Floating_Point
9431 is known to be in effect), but can be used anywhere.
9433 There are two forms:
9435 @smallexample @c ada
9436 System'Restriction_Set (partition_boolean_restriction_NAME)
9437 System'Restriction_Set (No_Dependence => library_unit_NAME);
9441 In the case of the first form, the only restriction names
9442 allowed are parameterless restrictions that are checked
9443 for consistency at bind time. For a complete list see the
9444 subtype @code{System.Rident.Partition_Boolean_Restrictions}.
9446 The result returned is True if the restriction is known to
9447 be in effect, and False if the restriction is known not to
9448 be in effect. An important guarantee is that the value of
9449 a Restriction_Set attribute is known to be consistent throughout
9450 all the code of a partition.
9452 This is trivially achieved if the entire partition is compiled
9453 with a consistent set of restriction pragmas. However, the
9454 compilation model does not require this. It is possible to
9455 compile one set of units with one set of pragmas, and another
9456 set of units with another set of pragmas. It is even possible
9457 to compile a spec with one set of pragmas, and then WITH the
9458 same spec with a different set of pragmas. Inconsistencies
9459 in the actual use of the restriction are checked at bind time.
9461 In order to achieve the guarantee of consistency for the
9462 Restriction_Set pragma, we consider that a use of the pragma
9463 that yields False is equivalent to a violation of the
9466 So for example if you write
9468 @smallexample @c ada
9469 if System'Restriction_Set (No_Floating_Point) then
9477 And the result is False, so that the else branch is executed,
9478 you can assume that this restriction is not set for any unit
9479 in the partition. This is checked by considering this use of
9480 the restriction pragma to be a violation of the restriction
9481 No_Floating_Point. This means that no other unit can attempt
9482 to set this restriction (if some unit does attempt to set it,
9483 the binder will refuse to bind the partition).
9485 Technical note: The restriction name and the unit name are
9486 intepreted entirely syntactically, as in the corresponding
9487 Restrictions pragma, they are not analyzed semantically,
9488 so they do not have a type.
9490 @node Attribute Result
9491 @unnumberedsec Attribute Result
9494 @code{@var{function}'Result} can only be used with in a Postcondition pragma
9495 for a function. The prefix must be the name of the corresponding function. This
9496 is used to refer to the result of the function in the postcondition expression.
9497 For a further discussion of the use of this attribute and examples of its use,
9498 see the description of pragma Postcondition.
9500 @node Attribute Safe_Emax
9501 @unnumberedsec Attribute Safe_Emax
9502 @cindex Ada 83 attributes
9505 The @code{Safe_Emax} attribute is provided for compatibility with Ada 83. See
9506 the Ada 83 reference manual for an exact description of the semantics of
9509 @node Attribute Safe_Large
9510 @unnumberedsec Attribute Safe_Large
9511 @cindex Ada 83 attributes
9514 The @code{Safe_Large} attribute is provided for compatibility with Ada 83. See
9515 the Ada 83 reference manual for an exact description of the semantics of
9518 @node Attribute Safe_Small
9519 @unnumberedsec Attribute Safe_Small
9520 @cindex Ada 83 attributes
9523 The @code{Safe_Small} attribute is provided for compatibility with Ada 83. See
9524 the Ada 83 reference manual for an exact description of the semantics of
9527 @node Attribute Scalar_Storage_Order
9528 @unnumberedsec Attribute Scalar_Storage_Order
9530 @cindex Scalar storage order
9531 @findex Scalar_Storage_Order
9533 For every array or record type @var{S}, the representation attribute
9534 @code{Scalar_Storage_Order} denotes the order in which storage elements
9535 that make up scalar components are ordered within S. The value given must
9536 be a static expression of type System.Bit_Order. The following is an example
9537 of the use of this feature:
9539 @smallexample @c ada
9540 -- Component type definitions
9542 subtype Yr_Type is Natural range 0 .. 127;
9543 subtype Mo_Type is Natural range 1 .. 12;
9544 subtype Da_Type is Natural range 1 .. 31;
9546 -- Record declaration
9549 Years_Since_1980 : Yr_Type;
9551 Day_Of_Month : Da_Type;
9554 -- Record representation clause
9557 Years_Since_1980 at 0 range 0 .. 6;
9558 Month at 0 range 7 .. 10;
9559 Day_Of_Month at 0 range 11 .. 15;
9562 -- Attribute definition clauses
9564 for Date'Bit_Order use System.High_Order_First;
9565 for Date'Scalar_Storage_Order use System.High_Order_First;
9566 -- If Scalar_Storage_Order is specified, it must be consistent with
9567 -- Bit_Order, so it's best to always define the latter explicitly if
9568 -- the former is used.
9572 Other properties are as for standard representation attribute @code{Bit_Order},
9573 as defined by Ada RM 13.5.3(4). The default is @code{System.Default_Bit_Order}.
9575 For a record type @var{T}, if @code{@var{T}'Scalar_Storage_Order} is
9576 specified explicitly, it shall be equal to @code{@var{T}'Bit_Order}. Note:
9577 this means that if a @code{Scalar_Storage_Order} attribute definition
9578 clause is not confirming, then the type's @code{Bit_Order} shall be
9579 specified explicitly and set to the same value.
9581 Derived types inherit an explicitly set scalar storage order from their parent
9582 types. This may be overridden for the derived type by giving an explicit scalar
9583 storage order for the derived type. For a record extension, the derived type
9584 must have the same scalar storage order as the parent type.
9586 If a component of @var{T} is of a record or array type, then that type must
9587 also have a @code{Scalar_Storage_Order} attribute definition clause.
9589 A component of a record or array type that is a packed array, or that
9590 does not start on a byte boundary, must have the same scalar storage order
9591 as the enclosing record or array type.
9593 No component of a type that has an explicit @code{Scalar_Storage_Order}
9594 attribute definition may be aliased.
9596 A confirming @code{Scalar_Storage_Order} attribute definition clause (i.e.
9597 with a value equal to @code{System.Default_Bit_Order}) has no effect.
9599 If the opposite storage order is specified, then whenever the value of
9600 a scalar component of an object of type @var{S} is read, the storage
9601 elements of the enclosing machine scalar are first reversed (before
9602 retrieving the component value, possibly applying some shift and mask
9603 operatings on the enclosing machine scalar), and the opposite operation
9606 In that case, the restrictions set forth in 13.5.1(10.3/2) for scalar components
9607 are relaxed. Instead, the following rules apply:
9610 @item the underlying storage elements are those at positions
9611 @code{(position + first_bit / storage_element_size) ..
9612 (position + (last_bit + storage_element_size - 1) /
9613 storage_element_size)}
9614 @item the sequence of underlying storage elements shall have
9615 a size no greater than the largest machine scalar
9616 @item the enclosing machine scalar is defined as the smallest machine
9617 scalar starting at a position no greater than
9618 @code{position + first_bit / storage_element_size} and covering
9619 storage elements at least up to @code{position + (last_bit +
9620 storage_element_size - 1) / storage_element_size}
9621 @item the position of the component is interpreted relative to that machine
9626 If no scalar storage order is specified for a type (either directly, or by
9627 inheritance in the case of a derived type), then the default is normally
9628 the native ordering of the target, but this default can be overridden using
9629 pragma @code{Default_Scalar_Storage_Order}.
9631 @node Attribute Simple_Storage_Pool
9632 @unnumberedsec Attribute Simple_Storage_Pool
9633 @cindex Storage pool, simple
9634 @cindex Simple storage pool
9635 @findex Simple_Storage_Pool
9637 For every nonformal, nonderived access-to-object type @var{Acc}, the
9638 representation attribute @code{Simple_Storage_Pool} may be specified
9639 via an attribute_definition_clause (or by specifying the equivalent aspect):
9641 @smallexample @c ada
9643 My_Pool : My_Simple_Storage_Pool_Type;
9645 type Acc is access My_Data_Type;
9647 for Acc'Simple_Storage_Pool use My_Pool;
9652 The name given in an attribute_definition_clause for the
9653 @code{Simple_Storage_Pool} attribute shall denote a variable of
9654 a ``simple storage pool type'' (see pragma @code{Simple_Storage_Pool_Type}).
9656 The use of this attribute is only allowed for a prefix denoting a type
9657 for which it has been specified. The type of the attribute is the type
9658 of the variable specified as the simple storage pool of the access type,
9659 and the attribute denotes that variable.
9661 It is illegal to specify both @code{Storage_Pool} and @code{Simple_Storage_Pool}
9662 for the same access type.
9664 If the @code{Simple_Storage_Pool} attribute has been specified for an access
9665 type, then applying the @code{Storage_Pool} attribute to the type is flagged
9666 with a warning and its evaluation raises the exception @code{Program_Error}.
9668 If the Simple_Storage_Pool attribute has been specified for an access
9669 type @var{S}, then the evaluation of the attribute @code{@var{S}'Storage_Size}
9670 returns the result of calling @code{Storage_Size (@var{S}'Simple_Storage_Pool)},
9671 which is intended to indicate the number of storage elements reserved for
9672 the simple storage pool. If the Storage_Size function has not been defined
9673 for the simple storage pool type, then this attribute returns zero.
9675 If an access type @var{S} has a specified simple storage pool of type
9676 @var{SSP}, then the evaluation of an allocator for that access type calls
9677 the primitive @code{Allocate} procedure for type @var{SSP}, passing
9678 @code{@var{S}'Simple_Storage_Pool} as the pool parameter. The detailed
9679 semantics of such allocators is the same as those defined for allocators
9680 in section 13.11 of the Ada Reference Manual, with the term
9681 ``simple storage pool'' substituted for ``storage pool''.
9683 If an access type @var{S} has a specified simple storage pool of type
9684 @var{SSP}, then a call to an instance of the @code{Ada.Unchecked_Deallocation}
9685 for that access type invokes the primitive @code{Deallocate} procedure
9686 for type @var{SSP}, passing @code{@var{S}'Simple_Storage_Pool} as the pool
9687 parameter. The detailed semantics of such unchecked deallocations is the same
9688 as defined in section 13.11.2 of the Ada Reference Manual, except that the
9689 term ``simple storage pool'' is substituted for ``storage pool''.
9691 @node Attribute Small
9692 @unnumberedsec Attribute Small
9693 @cindex Ada 83 attributes
9696 The @code{Small} attribute is defined in Ada 95 (and Ada 2005) only for
9698 GNAT also allows this attribute to be applied to floating-point types
9699 for compatibility with Ada 83. See
9700 the Ada 83 reference manual for an exact description of the semantics of
9701 this attribute when applied to floating-point types.
9703 @node Attribute Storage_Unit
9704 @unnumberedsec Attribute Storage_Unit
9705 @findex Storage_Unit
9707 @code{Standard'Storage_Unit} (@code{Standard} is the only permissible
9708 prefix) provides the same value as @code{System.Storage_Unit}.
9710 @node Attribute Stub_Type
9711 @unnumberedsec Attribute Stub_Type
9714 The GNAT implementation of remote access-to-classwide types is
9715 organized as described in AARM section E.4 (20.t): a value of an RACW type
9716 (designating a remote object) is represented as a normal access
9717 value, pointing to a "stub" object which in turn contains the
9718 necessary information to contact the designated remote object. A
9719 call on any dispatching operation of such a stub object does the
9720 remote call, if necessary, using the information in the stub object
9721 to locate the target partition, etc.
9723 For a prefix @code{T} that denotes a remote access-to-classwide type,
9724 @code{T'Stub_Type} denotes the type of the corresponding stub objects.
9726 By construction, the layout of @code{T'Stub_Type} is identical to that of
9727 type @code{RACW_Stub_Type} declared in the internal implementation-defined
9728 unit @code{System.Partition_Interface}. Use of this attribute will create
9729 an implicit dependency on this unit.
9731 @node Attribute System_Allocator_Alignment
9732 @unnumberedsec Attribute System_Allocator_Alignment
9733 @cindex Alignment, allocator
9734 @findex System_Allocator_Alignment
9736 @code{Standard'System_Allocator_Alignment} (@code{Standard} is the only
9737 permissible prefix) provides the observable guaranted to be honored by
9738 the system allocator (malloc). This is a static value that can be used
9739 in user storage pools based on malloc either to reject allocation
9740 with alignment too large or to enable a realignment circuitry if the
9741 alignment request is larger than this value.
9743 @node Attribute Target_Name
9744 @unnumberedsec Attribute Target_Name
9747 @code{Standard'Target_Name} (@code{Standard} is the only permissible
9748 prefix) provides a static string value that identifies the target
9749 for the current compilation. For GCC implementations, this is the
9750 standard gcc target name without the terminating slash (for
9751 example, GNAT 5.0 on windows yields "i586-pc-mingw32msv").
9753 @node Attribute To_Address
9754 @unnumberedsec Attribute To_Address
9757 The @code{System'To_Address}
9758 (@code{System} is the only permissible prefix)
9759 denotes a function identical to
9760 @code{System.Storage_Elements.To_Address} except that
9761 it is a static attribute. This means that if its argument is
9762 a static expression, then the result of the attribute is a
9763 static expression. This means that such an expression can be
9764 used in contexts (e.g.@: preelaborable packages) which require a
9765 static expression and where the function call could not be used
9766 (since the function call is always non-static, even if its
9767 argument is static). The argument must be in the range
9768 -(2**(m-1) .. 2**m-1, where m is the memory size
9769 (typically 32 or 64). Negative values are intepreted in a
9770 modular manner (e.g. -1 means the same as 16#FFFF_FFFF# on
9773 @node Attribute To_Any
9774 @unnumberedsec Attribute To_Any
9777 This internal attribute is used for the generation of remote subprogram
9778 stubs in the context of the Distributed Systems Annex.
9780 @node Attribute Type_Class
9781 @unnumberedsec Attribute Type_Class
9784 @code{@var{type}'Type_Class} for any type or subtype @var{type} yields
9785 the value of the type class for the full type of @var{type}. If
9786 @var{type} is a generic formal type, the value is the value for the
9787 corresponding actual subtype. The value of this attribute is of type
9788 @code{System.Aux_DEC.Type_Class}, which has the following definition:
9790 @smallexample @c ada
9792 (Type_Class_Enumeration,
9794 Type_Class_Fixed_Point,
9795 Type_Class_Floating_Point,
9800 Type_Class_Address);
9804 Protected types yield the value @code{Type_Class_Task}, which thus
9805 applies to all concurrent types. This attribute is designed to
9806 be compatible with the DEC Ada 83 attribute of the same name.
9808 @node Attribute Type_Key
9809 @unnumberedsec Attribute Type_Key
9812 The @code{Type_Key} attribute is applicable to a type or subtype and
9813 yields a value of type Standard.String containing encoded information
9814 about the type or subtype. This provides improved compatibility with
9815 other implementations that support this attribute.
9817 @node Attribute TypeCode
9818 @unnumberedsec Attribute TypeCode
9821 This internal attribute is used for the generation of remote subprogram
9822 stubs in the context of the Distributed Systems Annex.
9824 @node Attribute UET_Address
9825 @unnumberedsec Attribute UET_Address
9828 The @code{UET_Address} attribute can only be used for a prefix which
9829 denotes a library package. It yields the address of the unit exception
9830 table when zero cost exception handling is used. This attribute is
9831 intended only for use within the GNAT implementation. See the unit
9832 @code{Ada.Exceptions} in files @file{a-except.ads} and @file{a-except.adb}
9833 for details on how this attribute is used in the implementation.
9835 @node Attribute Unconstrained_Array
9836 @unnumberedsec Attribute Unconstrained_Array
9837 @findex Unconstrained_Array
9839 The @code{Unconstrained_Array} attribute can be used with a prefix that
9840 denotes any type or subtype. It is a static attribute that yields
9841 @code{True} if the prefix designates an unconstrained array,
9842 and @code{False} otherwise. In a generic instance, the result is
9843 still static, and yields the result of applying this test to the
9846 @node Attribute Universal_Literal_String
9847 @unnumberedsec Attribute Universal_Literal_String
9848 @cindex Named numbers, representation of
9849 @findex Universal_Literal_String
9851 The prefix of @code{Universal_Literal_String} must be a named
9852 number. The static result is the string consisting of the characters of
9853 the number as defined in the original source. This allows the user
9854 program to access the actual text of named numbers without intermediate
9855 conversions and without the need to enclose the strings in quotes (which
9856 would preclude their use as numbers).
9858 For example, the following program prints the first 50 digits of pi:
9860 @smallexample @c ada
9861 with Text_IO; use Text_IO;
9865 Put (Ada.Numerics.Pi'Universal_Literal_String);
9869 @node Attribute Unrestricted_Access
9870 @unnumberedsec Attribute Unrestricted_Access
9871 @cindex @code{Access}, unrestricted
9872 @findex Unrestricted_Access
9874 The @code{Unrestricted_Access} attribute is similar to @code{Access}
9875 except that all accessibility and aliased view checks are omitted. This
9876 is a user-beware attribute.
9878 For objects, it is similar to @code{Address}, for which it is a
9879 desirable replacement where the value desired is an access type.
9880 In other words, its effect is similar to first applying the
9881 @code{Address} attribute and then doing an unchecked conversion to a
9882 desired access type.
9884 For subprograms, @code{P'Unrestricted_Access} may be used where
9885 @code{P'Access} would be illegal, to construct a value of a
9886 less-nested named access type that designates a more-nested
9887 subprogram. This value may be used in indirect calls, so long as the
9888 more-nested subprogram still exists; once the subprogram containing it
9889 has returned, such calls are erroneous. For example:
9891 @smallexample @c ada
9894 type Less_Nested is not null access procedure;
9895 Global : Less_Nested;
9903 Local_Var : Integer;
9905 procedure More_Nested is
9910 Global := More_Nested'Unrestricted_Access;
9917 When P1 is called from P2, the call via Global is OK, but if P1 were
9918 called after P2 returns, it would be an erroneous use of a dangling
9921 For objects, it is possible to use @code{Unrestricted_Access} for any
9922 type, but care must be exercised if it is used to create pointers to
9923 unconstrained array objects. In this case, the resulting pointer has
9924 the same scope as the context of the attribute, and may not be
9925 returned to some enclosing scope. For instance, a function cannot use
9926 @code{Unrestricted_Access} to create a pointer to unconstrained and
9927 then return that value to the caller. In addition, it is only valid
9928 to create pointers to unconstrained arrays using this attribute if the
9929 pointer has the normal default ``fat'' representation where a pointer
9930 has two components, one points to the array and one points to the
9931 bounds. If a size clause is used to force ``thin'' representation for
9932 a pointer to unconstrained where there is only space for a single
9933 pointer, then the resulting pointer is not usable.
9935 In the simple case where a direct use of Unrestricted_Access attempts
9936 to make a thin pointer for a non-aliased object, the compiler will
9937 reject the use as illegal, as shown in the following example:
9939 @smallexample @c ada
9940 with System; use System;
9941 procedure SliceUA2 is
9942 type A is access all String;
9943 for A'Size use Standard'Address_Size;
9945 procedure P (Arg : A) is
9950 X : String := "hello world!";
9951 X2 : aliased String := "hello world!";
9953 AV : A := X'Unrestricted_Access; -- ERROR
9955 >>> illegal use of Unrestricted_Access attribute
9956 >>> attempt to generate thin pointer to unaliased object
9959 P (X'Unrestricted_Access); -- ERROR
9961 >>> illegal use of Unrestricted_Access attribute
9962 >>> attempt to generate thin pointer to unaliased object
9964 P (X(7 .. 12)'Unrestricted_Access); -- ERROR
9966 >>> illegal use of Unrestricted_Access attribute
9967 >>> attempt to generate thin pointer to unaliased object
9969 P (X2'Unrestricted_Access); -- OK
9974 but other cases cannot be detected by the compiler, and are
9975 considered to be erroneous. Consider the following example:
9977 @smallexample @c ada
9978 with System; use System;
9979 with System; use System;
9980 procedure SliceUA is
9981 type AF is access all String;
9983 type A is access all String;
9984 for A'Size use Standard'Address_Size;
9986 procedure P (Arg : A) is
9988 if Arg'Length /= 6 then
9989 raise Program_Error;
9993 X : String := "hello world!";
9994 Y : AF := X (7 .. 12)'Unrestricted_Access;
10002 A normal unconstrained array value
10003 or a constrained array object marked as aliased has the bounds in memory
10004 just before the array, so a thin pointer can retrieve both the data and
10005 the bounds. But in this case, the non-aliased object @code{X} does not have the
10006 bounds before the string. If the size clause for type @code{A}
10007 were not present, then the pointer
10008 would be a fat pointer, where one component is a pointer to the bounds,
10009 and all would be well. But with the size clause present, the conversion from
10010 fat pointer to thin pointer in the call loses the bounds, and so this
10011 is erroneous, and the program likely raises a @code{Program_Error} exception.
10013 In general, it is advisable to completely
10014 avoid mixing the use of thin pointers and the use of
10015 @code{Unrestricted_Access} where the designated type is an
10016 unconstrained array. The use of thin pointers should be restricted to
10017 cases of porting legacy code that implicitly assumes the size of pointers,
10018 and such code should not in any case be using this attribute.
10020 Another erroneous situation arises if the attribute is
10021 applied to a constant. The resulting pointer can be used to access the
10022 constant, but the effect of trying to modify a constant in this manner
10023 is not well-defined. Consider this example:
10025 @smallexample @c ada
10026 P : constant Integer := 4;
10027 type R is access all Integer;
10028 RV : R := P'Unrestricted_Access;
10034 Here we attempt to modify the constant P from 4 to 3, but the compiler may
10035 or may not notice this attempt, and subsequent references to P may yield
10036 either the value 3 or the value 4 or the assignment may blow up if the
10037 compiler decides to put P in read-only memory. One particular case where
10038 @code{Unrestricted_Access} can be used in this way is to modify the
10039 value of an @code{IN} parameter:
10041 @smallexample @c ada
10042 procedure K (S : in String) is
10043 type R is access all Character;
10044 RV : R := S (3)'Unrestricted_Access;
10051 In general this is a risky approach. It may appear to "work" but such uses of
10052 @code{Unrestricted_Access} are potentially non-portable, even from one version
10053 of @code{GNAT} to another, so are best avoided if possible.
10055 @node Attribute Update
10056 @unnumberedsec Attribute Update
10059 The @code{Update} attribute creates a copy of an array or record value
10060 with one or more modified components. The syntax is:
10062 @smallexample @c ada
10063 PREFIX'Update ( RECORD_COMPONENT_ASSOCIATION_LIST )
10064 PREFIX'Update ( ARRAY_COMPONENT_ASSOCIATION @{, ARRAY_COMPONENT_ASSOCIATION @} )
10065 PREFIX'Update ( MULTIDIMENSIONAL_ARRAY_COMPONENT_ASSOCIATION
10066 @{, MULTIDIMENSIONAL_ARRAY_COMPONENT_ASSOCIATION @} )
10068 MULTIDIMENSIONAL_ARRAY_COMPONENT_ASSOCIATION ::= INDEX_EXPRESSION_LIST_LIST => EXPRESSION
10069 INDEX_EXPRESSION_LIST_LIST ::= INDEX_EXPRESSION_LIST @{| INDEX_EXPRESSION_LIST @}
10070 INDEX_EXPRESSION_LIST ::= ( EXPRESSION @{, EXPRESSION @} )
10074 where @code{PREFIX} is the name of an array or record object, and
10075 the association list in parentheses does not contain an @code{others}
10076 choice. The effect is to yield a copy of the array or record value which
10077 is unchanged apart from the components mentioned in the association list, which
10078 are changed to the indicated value. The original value of the array or
10079 record value is not affected. For example:
10081 @smallexample @c ada
10082 type Arr is Array (1 .. 5) of Integer;
10084 Avar1 : Arr := (1,2,3,4,5);
10085 Avar2 : Arr := Avar1'Update (2 => 10, 3 .. 4 => 20);
10089 yields a value for @code{Avar2} of 1,10,20,20,5 with @code{Avar1}
10090 begin unmodified. Similarly:
10092 @smallexample @c ada
10093 type Rec is A, B, C : Integer;
10095 Rvar1 : Rec := (A => 1, B => 2, C => 3);
10096 Rvar2 : Rec := Rvar1'Update (B => 20);
10100 yields a value for @code{Rvar2} of (A => 1, B => 20, C => 3),
10101 with @code{Rvar1} being unmodifed.
10102 Note that the value of the attribute reference is computed
10103 completely before it is used. This means that if you write:
10105 @smallexample @c ada
10106 Avar1 := Avar1'Update (1 => 10, 2 => Function_Call);
10110 then the value of @code{Avar1} is not modified if @code{Function_Call}
10111 raises an exception, unlike the effect of a series of direct assignments
10112 to elements of @code{Avar1}. In general this requires that
10113 two extra complete copies of the object are required, which should be
10114 kept in mind when considering efficiency.
10116 The @code{Update} attribute cannot be applied to prefixes of a limited
10117 type, and cannot reference discriminants in the case of a record type.
10118 The accessibility level of an Update attribute result object is defined
10119 as for an aggregate.
10121 In the record case, no component can be mentioned more than once. In
10122 the array case, two overlapping ranges can appear in the association list,
10123 in which case the modifications are processed left to right.
10125 Multi-dimensional arrays can be modified, as shown by this example:
10127 @smallexample @c ada
10128 A : array (1 .. 10, 1 .. 10) of Integer;
10130 A := A'Update ((1, 2) => 20, (3, 4) => 30);
10134 which changes element (1,2) to 20 and (3,4) to 30.
10136 @node Attribute Valid_Scalars
10137 @unnumberedsec Attribute Valid_Scalars
10138 @findex Valid_Scalars
10140 The @code{'Valid_Scalars} attribute is intended to make it easier to
10141 check the validity of scalar subcomponents of composite objects. It
10142 is defined for any prefix @code{X} that denotes an object.
10143 The value of this attribute is of the predefined type Boolean.
10144 @code{X'Valid_Scalars} yields True if and only if evaluation of
10145 @code{P'Valid} yields True for every scalar part P of X or if X has
10146 no scalar parts. It is not specified in what order the scalar parts
10147 are checked, nor whether any more are checked after any one of them
10148 is determined to be invalid. If the prefix @code{X} is of a class-wide
10149 type @code{T'Class} (where @code{T} is the associated specific type),
10150 or if the prefix @code{X} is of a specific tagged type @code{T}, then
10151 only the scalar parts of components of @code{T} are traversed; in other
10152 words, components of extensions of @code{T} are not traversed even if
10153 @code{T'Class (X)'Tag /= T'Tag} . The compiler will issue a warning if it can
10154 be determined at compile time that the prefix of the attribute has no
10155 scalar parts (e.g., if the prefix is of an access type, an interface type,
10156 an undiscriminated task type, or an undiscriminated protected type).
10158 For scalar types, @code{Valid_Scalars} is equivalent to @code{Valid}. The use
10159 of this attribute is not permitted for @code{Unchecked_Union} types for which
10160 in general it is not possible to determine the values of the discriminants.
10162 Note: @code{Valid_Scalars} can generate a lot of code, especially in the case
10163 of a large variant record. If the attribute is called in many places in the
10164 same program applied to objects of the same type, it can reduce program size
10165 to write a function with a single use of the attribute, and then call that
10166 function from multiple places.
10168 @node Attribute VADS_Size
10169 @unnumberedsec Attribute VADS_Size
10170 @cindex @code{Size}, VADS compatibility
10173 The @code{'VADS_Size} attribute is intended to make it easier to port
10174 legacy code which relies on the semantics of @code{'Size} as implemented
10175 by the VADS Ada 83 compiler. GNAT makes a best effort at duplicating the
10176 same semantic interpretation. In particular, @code{'VADS_Size} applied
10177 to a predefined or other primitive type with no Size clause yields the
10178 Object_Size (for example, @code{Natural'Size} is 32 rather than 31 on
10179 typical machines). In addition @code{'VADS_Size} applied to an object
10180 gives the result that would be obtained by applying the attribute to
10181 the corresponding type.
10183 @node Attribute Value_Size
10184 @unnumberedsec Attribute Value_Size
10185 @cindex @code{Size}, setting for not-first subtype
10187 @code{@var{type}'Value_Size} is the number of bits required to represent
10188 a value of the given subtype. It is the same as @code{@var{type}'Size},
10189 but, unlike @code{Size}, may be set for non-first subtypes.
10191 @node Attribute Wchar_T_Size
10192 @unnumberedsec Attribute Wchar_T_Size
10193 @findex Wchar_T_Size
10194 @code{Standard'Wchar_T_Size} (@code{Standard} is the only permissible
10195 prefix) provides the size in bits of the C @code{wchar_t} type
10196 primarily for constructing the definition of this type in
10197 package @code{Interfaces.C}. The result is a static constant.
10199 @node Attribute Word_Size
10200 @unnumberedsec Attribute Word_Size
10202 @code{Standard'Word_Size} (@code{Standard} is the only permissible
10203 prefix) provides the value @code{System.Word_Size}. The result is
10206 @node Standard and Implementation Defined Restrictions
10207 @chapter Standard and Implementation Defined Restrictions
10210 All RM defined Restriction identifiers are implemented:
10213 @item language-defined restrictions (see 13.12.1)
10214 @item tasking restrictions (see D.7)
10215 @item high integrity restrictions (see H.4)
10219 GNAT implements additional restriction identifiers. All restrictions, whether
10220 language defined or GNAT-specific, are listed in the following.
10223 * Partition-Wide Restrictions::
10224 * Program Unit Level Restrictions::
10227 @node Partition-Wide Restrictions
10228 @section Partition-Wide Restrictions
10230 There are two separate lists of restriction identifiers. The first
10231 set requires consistency throughout a partition (in other words, if the
10232 restriction identifier is used for any compilation unit in the partition,
10233 then all compilation units in the partition must obey the restriction).
10236 * Immediate_Reclamation::
10237 * Max_Asynchronous_Select_Nesting::
10238 * Max_Entry_Queue_Length::
10239 * Max_Protected_Entries::
10240 * Max_Select_Alternatives::
10241 * Max_Storage_At_Blocking::
10242 * Max_Task_Entries::
10244 * No_Abort_Statements::
10245 * No_Access_Parameter_Allocators::
10246 * No_Access_Subprograms::
10248 * No_Anonymous_Allocators::
10250 * No_Coextensions::
10251 * No_Default_Initialization::
10254 * No_Direct_Boolean_Operators::
10256 * No_Dispatching_Calls::
10257 * No_Dynamic_Attachment::
10258 * No_Dynamic_Priorities::
10259 * No_Entry_Calls_In_Elaboration_Code::
10260 * No_Enumeration_Maps::
10261 * No_Exception_Handlers::
10262 * No_Exception_Propagation::
10263 * No_Exception_Registration::
10265 * No_Finalization::
10267 * No_Floating_Point::
10268 * No_Implicit_Conditionals::
10269 * No_Implicit_Dynamic_Code::
10270 * No_Implicit_Heap_Allocations::
10271 * No_Implicit_Loops::
10272 * No_Initialize_Scalars::
10274 * No_Local_Allocators::
10275 * No_Local_Protected_Objects::
10276 * No_Local_Timing_Events::
10277 * No_Long_Long_Integers::
10278 * No_Multiple_Elaboration::
10279 * No_Nested_Finalization::
10280 * No_Protected_Type_Allocators::
10281 * No_Protected_Types::
10284 * No_Relative_Delay::
10285 * No_Requeue_Statements::
10286 * No_Secondary_Stack::
10287 * No_Select_Statements::
10288 * No_Specific_Termination_Handlers::
10289 * No_Specification_of_Aspect::
10290 * No_Standard_Allocators_After_Elaboration::
10291 * No_Standard_Storage_Pools::
10292 * No_Stream_Optimizations::
10294 * No_Task_Allocators::
10295 * No_Task_Attributes_Package::
10296 * No_Task_Hierarchy::
10297 * No_Task_Termination::
10299 * No_Terminate_Alternatives::
10300 * No_Unchecked_Access::
10301 * Simple_Barriers::
10302 * Static_Priorities::
10303 * Static_Storage_Size::
10306 @node Immediate_Reclamation
10307 @unnumberedsubsec Immediate_Reclamation
10308 @findex Immediate_Reclamation
10309 [RM H.4] This restriction ensures that, except for storage occupied by
10310 objects created by allocators and not deallocated via unchecked
10311 deallocation, any storage reserved at run time for an object is
10312 immediately reclaimed when the object no longer exists.
10314 @node Max_Asynchronous_Select_Nesting
10315 @unnumberedsubsec Max_Asynchronous_Select_Nesting
10316 @findex Max_Asynchronous_Select_Nesting
10317 [RM D.7] Specifies the maximum dynamic nesting level of asynchronous
10318 selects. Violations of this restriction with a value of zero are
10319 detected at compile time. Violations of this restriction with values
10320 other than zero cause Storage_Error to be raised.
10322 @node Max_Entry_Queue_Length
10323 @unnumberedsubsec Max_Entry_Queue_Length
10324 @findex Max_Entry_Queue_Length
10325 [RM D.7] This restriction is a declaration that any protected entry compiled in
10326 the scope of the restriction has at most the specified number of
10327 tasks waiting on the entry at any one time, and so no queue is required.
10328 Note that this restriction is checked at run time. Violation of this
10329 restriction results in the raising of Program_Error exception at the point of
10332 @findex Max_Entry_Queue_Depth
10333 The restriction @code{Max_Entry_Queue_Depth} is recognized as a
10334 synonym for @code{Max_Entry_Queue_Length}. This is retained for historical
10335 compatibility purposes (and a warning will be generated for its use if
10336 warnings on obsolescent features are activated).
10338 @node Max_Protected_Entries
10339 @unnumberedsubsec Max_Protected_Entries
10340 @findex Max_Protected_Entries
10341 [RM D.7] Specifies the maximum number of entries per protected type. The
10342 bounds of every entry family of a protected unit shall be static, or shall be
10343 defined by a discriminant of a subtype whose corresponding bound is static.
10345 @node Max_Select_Alternatives
10346 @unnumberedsubsec Max_Select_Alternatives
10347 @findex Max_Select_Alternatives
10348 [RM D.7] Specifies the maximum number of alternatives in a selective accept.
10350 @node Max_Storage_At_Blocking
10351 @unnumberedsubsec Max_Storage_At_Blocking
10352 @findex Max_Storage_At_Blocking
10353 [RM D.7] Specifies the maximum portion (in storage elements) of a task's
10354 Storage_Size that can be retained by a blocked task. A violation of this
10355 restriction causes Storage_Error to be raised.
10357 @node Max_Task_Entries
10358 @unnumberedsubsec Max_Task_Entries
10359 @findex Max_Task_Entries
10360 [RM D.7] Specifies the maximum number of entries
10361 per task. The bounds of every entry family
10362 of a task unit shall be static, or shall be
10363 defined by a discriminant of a subtype whose
10364 corresponding bound is static.
10367 @unnumberedsubsec Max_Tasks
10369 [RM D.7] Specifies the maximum number of task that may be created, not
10370 counting the creation of the environment task. Violations of this
10371 restriction with a value of zero are detected at compile
10372 time. Violations of this restriction with values other than zero cause
10373 Storage_Error to be raised.
10375 @node No_Abort_Statements
10376 @unnumberedsubsec No_Abort_Statements
10377 @findex No_Abort_Statements
10378 [RM D.7] There are no abort_statements, and there are
10379 no calls to Task_Identification.Abort_Task.
10381 @node No_Access_Parameter_Allocators
10382 @unnumberedsubsec No_Access_Parameter_Allocators
10383 @findex No_Access_Parameter_Allocators
10384 [RM H.4] This restriction ensures at compile time that there are no
10385 occurrences of an allocator as the actual parameter to an access
10388 @node No_Access_Subprograms
10389 @unnumberedsubsec No_Access_Subprograms
10390 @findex No_Access_Subprograms
10391 [RM H.4] This restriction ensures at compile time that there are no
10392 declarations of access-to-subprogram types.
10394 @node No_Allocators
10395 @unnumberedsubsec No_Allocators
10396 @findex No_Allocators
10397 [RM H.4] This restriction ensures at compile time that there are no
10398 occurrences of an allocator.
10400 @node No_Anonymous_Allocators
10401 @unnumberedsubsec No_Anonymous_Allocators
10402 @findex No_Anonymous_Allocators
10403 [RM H.4] This restriction ensures at compile time that there are no
10404 occurrences of an allocator of anonymous access type.
10407 @unnumberedsubsec No_Calendar
10408 @findex No_Calendar
10409 [GNAT] This restriction ensures at compile time that there is no implicit or
10410 explicit dependence on the package @code{Ada.Calendar}.
10412 @node No_Coextensions
10413 @unnumberedsubsec No_Coextensions
10414 @findex No_Coextensions
10415 [RM H.4] This restriction ensures at compile time that there are no
10416 coextensions. See 3.10.2.
10418 @node No_Default_Initialization
10419 @unnumberedsubsec No_Default_Initialization
10420 @findex No_Default_Initialization
10422 [GNAT] This restriction prohibits any instance of default initialization
10423 of variables. The binder implements a consistency rule which prevents
10424 any unit compiled without the restriction from with'ing a unit with the
10425 restriction (this allows the generation of initialization procedures to
10426 be skipped, since you can be sure that no call is ever generated to an
10427 initialization procedure in a unit with the restriction active). If used
10428 in conjunction with Initialize_Scalars or Normalize_Scalars, the effect
10429 is to prohibit all cases of variables declared without a specific
10430 initializer (including the case of OUT scalar parameters).
10433 @unnumberedsubsec No_Delay
10435 [RM H.4] This restriction ensures at compile time that there are no
10436 delay statements and no dependences on package Calendar.
10438 @node No_Dependence
10439 @unnumberedsubsec No_Dependence
10440 @findex No_Dependence
10441 [RM 13.12.1] This restriction checks at compile time that there are no
10442 dependence on a library unit.
10444 @node No_Direct_Boolean_Operators
10445 @unnumberedsubsec No_Direct_Boolean_Operators
10446 @findex No_Direct_Boolean_Operators
10447 [GNAT] This restriction ensures that no logical operators (and/or/xor)
10448 are used on operands of type Boolean (or any type derived from Boolean).
10449 This is intended for use in safety critical programs where the certification
10450 protocol requires the use of short-circuit (and then, or else) forms for all
10451 composite boolean operations.
10454 @unnumberedsubsec No_Dispatch
10455 @findex No_Dispatch
10456 [RM H.4] This restriction ensures at compile time that there are no
10457 occurrences of @code{T'Class}, for any (tagged) subtype @code{T}.
10459 @node No_Dispatching_Calls
10460 @unnumberedsubsec No_Dispatching_Calls
10461 @findex No_Dispatching_Calls
10462 [GNAT] This restriction ensures at compile time that the code generated by the
10463 compiler involves no dispatching calls. The use of this restriction allows the
10464 safe use of record extensions, classwide membership tests and other classwide
10465 features not involving implicit dispatching. This restriction ensures that
10466 the code contains no indirect calls through a dispatching mechanism. Note that
10467 this includes internally-generated calls created by the compiler, for example
10468 in the implementation of class-wide objects assignments. The
10469 membership test is allowed in the presence of this restriction, because its
10470 implementation requires no dispatching.
10471 This restriction is comparable to the official Ada restriction
10472 @code{No_Dispatch} except that it is a bit less restrictive in that it allows
10473 all classwide constructs that do not imply dispatching.
10474 The following example indicates constructs that violate this restriction.
10478 type T is tagged record
10481 procedure P (X : T);
10483 type DT is new T with record
10484 More_Data : Natural;
10486 procedure Q (X : DT);
10490 procedure Example is
10491 procedure Test (O : T'Class) is
10492 N : Natural := O'Size;-- Error: Dispatching call
10493 C : T'Class := O; -- Error: implicit Dispatching Call
10495 if O in DT'Class then -- OK : Membership test
10496 Q (DT (O)); -- OK : Type conversion plus direct call
10498 P (O); -- Error: Dispatching call
10504 P (Obj); -- OK : Direct call
10505 P (T (Obj)); -- OK : Type conversion plus direct call
10506 P (T'Class (Obj)); -- Error: Dispatching call
10508 Test (Obj); -- OK : Type conversion
10510 if Obj in T'Class then -- OK : Membership test
10516 @node No_Dynamic_Attachment
10517 @unnumberedsubsec No_Dynamic_Attachment
10518 @findex No_Dynamic_Attachment
10519 [RM D.7] This restriction ensures that there is no call to any of the
10520 operations defined in package Ada.Interrupts
10521 (Is_Reserved, Is_Attached, Current_Handler, Attach_Handler, Exchange_Handler,
10522 Detach_Handler, and Reference).
10524 @findex No_Dynamic_Interrupts
10525 The restriction @code{No_Dynamic_Interrupts} is recognized as a
10526 synonym for @code{No_Dynamic_Attachment}. This is retained for historical
10527 compatibility purposes (and a warning will be generated for its use if
10528 warnings on obsolescent features are activated).
10530 @node No_Dynamic_Priorities
10531 @unnumberedsubsec No_Dynamic_Priorities
10532 @findex No_Dynamic_Priorities
10533 [RM D.7] There are no semantic dependencies on the package Dynamic_Priorities.
10535 @node No_Entry_Calls_In_Elaboration_Code
10536 @unnumberedsubsec No_Entry_Calls_In_Elaboration_Code
10537 @findex No_Entry_Calls_In_Elaboration_Code
10538 [GNAT] This restriction ensures at compile time that no task or protected entry
10539 calls are made during elaboration code. As a result of the use of this
10540 restriction, the compiler can assume that no code past an accept statement
10541 in a task can be executed at elaboration time.
10543 @node No_Enumeration_Maps
10544 @unnumberedsubsec No_Enumeration_Maps
10545 @findex No_Enumeration_Maps
10546 [GNAT] This restriction ensures at compile time that no operations requiring
10547 enumeration maps are used (that is Image and Value attributes applied
10548 to enumeration types).
10550 @node No_Exception_Handlers
10551 @unnumberedsubsec No_Exception_Handlers
10552 @findex No_Exception_Handlers
10553 [GNAT] This restriction ensures at compile time that there are no explicit
10554 exception handlers. It also indicates that no exception propagation will
10555 be provided. In this mode, exceptions may be raised but will result in
10556 an immediate call to the last chance handler, a routine that the user
10557 must define with the following profile:
10559 @smallexample @c ada
10560 procedure Last_Chance_Handler
10561 (Source_Location : System.Address; Line : Integer);
10562 pragma Export (C, Last_Chance_Handler,
10563 "__gnat_last_chance_handler");
10566 The parameter is a C null-terminated string representing a message to be
10567 associated with the exception (typically the source location of the raise
10568 statement generated by the compiler). The Line parameter when nonzero
10569 represents the line number in the source program where the raise occurs.
10571 @node No_Exception_Propagation
10572 @unnumberedsubsec No_Exception_Propagation
10573 @findex No_Exception_Propagation
10574 [GNAT] This restriction guarantees that exceptions are never propagated
10575 to an outer subprogram scope. The only case in which an exception may
10576 be raised is when the handler is statically in the same subprogram, so
10577 that the effect of a raise is essentially like a goto statement. Any
10578 other raise statement (implicit or explicit) will be considered
10579 unhandled. Exception handlers are allowed, but may not contain an
10580 exception occurrence identifier (exception choice). In addition, use of
10581 the package GNAT.Current_Exception is not permitted, and reraise
10582 statements (raise with no operand) are not permitted.
10584 @node No_Exception_Registration
10585 @unnumberedsubsec No_Exception_Registration
10586 @findex No_Exception_Registration
10587 [GNAT] This restriction ensures at compile time that no stream operations for
10588 types Exception_Id or Exception_Occurrence are used. This also makes it
10589 impossible to pass exceptions to or from a partition with this restriction
10590 in a distributed environment. If this restriction is active, the generated
10591 code is simplified by omitting the otherwise-required global registration
10592 of exceptions when they are declared.
10594 @node No_Exceptions
10595 @unnumberedsubsec No_Exceptions
10596 @findex No_Exceptions
10597 [RM H.4] This restriction ensures at compile time that there are no
10598 raise statements and no exception handlers.
10600 @node No_Finalization
10601 @unnumberedsubsec No_Finalization
10602 @findex No_Finalization
10603 [GNAT] This restriction disables the language features described in
10604 chapter 7.6 of the Ada 2005 RM as well as all form of code generation
10605 performed by the compiler to support these features. The following types
10606 are no longer considered controlled when this restriction is in effect:
10609 @code{Ada.Finalization.Controlled}
10611 @code{Ada.Finalization.Limited_Controlled}
10613 Derivations from @code{Controlled} or @code{Limited_Controlled}
10621 Array and record types with controlled components
10623 The compiler no longer generates code to initialize, finalize or adjust an
10624 object or a nested component, either declared on the stack or on the heap. The
10625 deallocation of a controlled object no longer finalizes its contents.
10627 @node No_Fixed_Point
10628 @unnumberedsubsec No_Fixed_Point
10629 @findex No_Fixed_Point
10630 [RM H.4] This restriction ensures at compile time that there are no
10631 occurrences of fixed point types and operations.
10633 @node No_Floating_Point
10634 @unnumberedsubsec No_Floating_Point
10635 @findex No_Floating_Point
10636 [RM H.4] This restriction ensures at compile time that there are no
10637 occurrences of floating point types and operations.
10639 @node No_Implicit_Conditionals
10640 @unnumberedsubsec No_Implicit_Conditionals
10641 @findex No_Implicit_Conditionals
10642 [GNAT] This restriction ensures that the generated code does not contain any
10643 implicit conditionals, either by modifying the generated code where possible,
10644 or by rejecting any construct that would otherwise generate an implicit
10645 conditional. Note that this check does not include run time constraint
10646 checks, which on some targets may generate implicit conditionals as
10647 well. To control the latter, constraint checks can be suppressed in the
10648 normal manner. Constructs generating implicit conditionals include comparisons
10649 of composite objects and the Max/Min attributes.
10651 @node No_Implicit_Dynamic_Code
10652 @unnumberedsubsec No_Implicit_Dynamic_Code
10653 @findex No_Implicit_Dynamic_Code
10655 [GNAT] This restriction prevents the compiler from building ``trampolines''.
10656 This is a structure that is built on the stack and contains dynamic
10657 code to be executed at run time. On some targets, a trampoline is
10658 built for the following features: @code{Access},
10659 @code{Unrestricted_Access}, or @code{Address} of a nested subprogram;
10660 nested task bodies; primitive operations of nested tagged types.
10661 Trampolines do not work on machines that prevent execution of stack
10662 data. For example, on windows systems, enabling DEP (data execution
10663 protection) will cause trampolines to raise an exception.
10664 Trampolines are also quite slow at run time.
10666 On many targets, trampolines have been largely eliminated. Look at the
10667 version of system.ads for your target --- if it has
10668 Always_Compatible_Rep equal to False, then trampolines are largely
10669 eliminated. In particular, a trampoline is built for the following
10670 features: @code{Address} of a nested subprogram;
10671 @code{Access} or @code{Unrestricted_Access} of a nested subprogram,
10672 but only if pragma Favor_Top_Level applies, or the access type has a
10673 foreign-language convention; primitive operations of nested tagged
10676 @node No_Implicit_Heap_Allocations
10677 @unnumberedsubsec No_Implicit_Heap_Allocations
10678 @findex No_Implicit_Heap_Allocations
10679 [RM D.7] No constructs are allowed to cause implicit heap allocation.
10681 @node No_Implicit_Loops
10682 @unnumberedsubsec No_Implicit_Loops
10683 @findex No_Implicit_Loops
10684 [GNAT] This restriction ensures that the generated code does not contain any
10685 implicit @code{for} loops, either by modifying
10686 the generated code where possible,
10687 or by rejecting any construct that would otherwise generate an implicit
10688 @code{for} loop. If this restriction is active, it is possible to build
10689 large array aggregates with all static components without generating an
10690 intermediate temporary, and without generating a loop to initialize individual
10691 components. Otherwise, a loop is created for arrays larger than about 5000
10694 @node No_Initialize_Scalars
10695 @unnumberedsubsec No_Initialize_Scalars
10696 @findex No_Initialize_Scalars
10697 [GNAT] This restriction ensures that no unit in the partition is compiled with
10698 pragma Initialize_Scalars. This allows the generation of more efficient
10699 code, and in particular eliminates dummy null initialization routines that
10700 are otherwise generated for some record and array types.
10703 @unnumberedsubsec No_IO
10705 [RM H.4] This restriction ensures at compile time that there are no
10706 dependences on any of the library units Sequential_IO, Direct_IO,
10707 Text_IO, Wide_Text_IO, Wide_Wide_Text_IO, or Stream_IO.
10709 @node No_Local_Allocators
10710 @unnumberedsubsec No_Local_Allocators
10711 @findex No_Local_Allocators
10712 [RM H.4] This restriction ensures at compile time that there are no
10713 occurrences of an allocator in subprograms, generic subprograms, tasks,
10716 @node No_Local_Protected_Objects
10717 @unnumberedsubsec No_Local_Protected_Objects
10718 @findex No_Local_Protected_Objects
10719 [RM D.7] This restriction ensures at compile time that protected objects are
10720 only declared at the library level.
10722 @node No_Local_Timing_Events
10723 @unnumberedsubsec No_Local_Timing_Events
10724 @findex No_Local_Timing_Events
10725 [RM D.7] All objects of type Ada.Timing_Events.Timing_Event are
10726 declared at the library level.
10728 @node No_Long_Long_Integers
10729 @unnumberedsubsec No_Long_Long_Integers
10730 @findex No_Long_Long_Integers
10731 [GNAT] This partition-wide restriction forbids any explicit reference to
10732 type Standard.Long_Long_Integer, and also forbids declaring range types whose
10733 implicit base type is Long_Long_Integer, and modular types whose size exceeds
10736 @node No_Multiple_Elaboration
10737 @unnumberedsubsec No_Multiple_Elaboration
10738 @findex No_Multiple_Elaboration
10739 [GNAT] Normally each package contains a 16-bit counter used to check for access
10740 before elaboration, and to control multiple elaboration attempts.
10741 This counter is eliminated for units compiled with the static model
10742 of elaboration if restriction @code{No_Elaboration_Code}
10743 is active but because of
10744 the need to check for multiple elaboration in the general case, these
10745 counters cannot be eliminated if elaboration code may be present. The
10746 restriction @code{No_Multiple_Elaboration}
10747 allows suppression of these counters
10748 in static elaboration units even if they do have elaboration code. If this
10749 restriction is used, then the situations in which multiple elaboration is
10750 possible, including non-Ada main programs, and Stand Alone libraries, are not
10751 permitted, and will be diagnosed by the binder.
10753 @node No_Nested_Finalization
10754 @unnumberedsubsec No_Nested_Finalization
10755 @findex No_Nested_Finalization
10756 [RM D.7] All objects requiring finalization are declared at the library level.
10758 @node No_Protected_Type_Allocators
10759 @unnumberedsubsec No_Protected_Type_Allocators
10760 @findex No_Protected_Type_Allocators
10761 [RM D.7] This restriction ensures at compile time that there are no allocator
10762 expressions that attempt to allocate protected objects.
10764 @node No_Protected_Types
10765 @unnumberedsubsec No_Protected_Types
10766 @findex No_Protected_Types
10767 [RM H.4] This restriction ensures at compile time that there are no
10768 declarations of protected types or protected objects.
10771 @unnumberedsubsec No_Recursion
10772 @findex No_Recursion
10773 [RM H.4] A program execution is erroneous if a subprogram is invoked as
10774 part of its execution.
10776 @node No_Reentrancy
10777 @unnumberedsubsec No_Reentrancy
10778 @findex No_Reentrancy
10779 [RM H.4] A program execution is erroneous if a subprogram is executed by
10780 two tasks at the same time.
10782 @node No_Relative_Delay
10783 @unnumberedsubsec No_Relative_Delay
10784 @findex No_Relative_Delay
10785 [RM D.7] This restriction ensures at compile time that there are no delay
10786 relative statements and prevents expressions such as @code{delay 1.23;} from
10787 appearing in source code.
10789 @node No_Requeue_Statements
10790 @unnumberedsubsec No_Requeue_Statements
10791 @findex No_Requeue_Statements
10792 [RM D.7] This restriction ensures at compile time that no requeue statements
10793 are permitted and prevents keyword @code{requeue} from being used in source
10797 The restriction @code{No_Requeue} is recognized as a
10798 synonym for @code{No_Requeue_Statements}. This is retained for historical
10799 compatibility purposes (and a warning will be generated for its use if
10800 warnings on oNobsolescent features are activated).
10802 @node No_Secondary_Stack
10803 @unnumberedsubsec No_Secondary_Stack
10804 @findex No_Secondary_Stack
10805 [GNAT] This restriction ensures at compile time that the generated code
10806 does not contain any reference to the secondary stack. The secondary
10807 stack is used to implement functions returning unconstrained objects
10808 (arrays or records) on some targets.
10810 @node No_Select_Statements
10811 @unnumberedsubsec No_Select_Statements
10812 @findex No_Select_Statements
10813 [RM D.7] This restriction ensures at compile time no select statements of any
10814 kind are permitted, that is the keyword @code{select} may not appear.
10816 @node No_Specific_Termination_Handlers
10817 @unnumberedsubsec No_Specific_Termination_Handlers
10818 @findex No_Specific_Termination_Handlers
10819 [RM D.7] There are no calls to Ada.Task_Termination.Set_Specific_Handler
10820 or to Ada.Task_Termination.Specific_Handler.
10822 @node No_Specification_of_Aspect
10823 @unnumberedsubsec No_Specification_of_Aspect
10824 @findex No_Specification_of_Aspect
10825 [RM 13.12.1] This restriction checks at compile time that no aspect
10826 specification, attribute definition clause, or pragma is given for a
10829 @node No_Standard_Allocators_After_Elaboration
10830 @unnumberedsubsec No_Standard_Allocators_After_Elaboration
10831 @findex No_Standard_Allocators_After_Elaboration
10832 [RM D.7] Specifies that an allocator using a standard storage pool
10833 should never be evaluated at run time after the elaboration of the
10834 library items of the partition has completed. Otherwise, Storage_Error
10837 @node No_Standard_Storage_Pools
10838 @unnumberedsubsec No_Standard_Storage_Pools
10839 @findex No_Standard_Storage_Pools
10840 [GNAT] This restriction ensures at compile time that no access types
10841 use the standard default storage pool. Any access type declared must
10842 have an explicit Storage_Pool attribute defined specifying a
10843 user-defined storage pool.
10845 @node No_Stream_Optimizations
10846 @unnumberedsubsec No_Stream_Optimizations
10847 @findex No_Stream_Optimizations
10848 [GNAT] This restriction affects the performance of stream operations on types
10849 @code{String}, @code{Wide_String} and @code{Wide_Wide_String}. By default, the
10850 compiler uses block reads and writes when manipulating @code{String} objects
10851 due to their supperior performance. When this restriction is in effect, the
10852 compiler performs all IO operations on a per-character basis.
10855 @unnumberedsubsec No_Streams
10857 [GNAT] This restriction ensures at compile/bind time that there are no
10858 stream objects created and no use of stream attributes.
10859 This restriction does not forbid dependences on the package
10860 @code{Ada.Streams}. So it is permissible to with
10861 @code{Ada.Streams} (or another package that does so itself)
10862 as long as no actual stream objects are created and no
10863 stream attributes are used.
10865 Note that the use of restriction allows optimization of tagged types,
10866 since they do not need to worry about dispatching stream operations.
10867 To take maximum advantage of this space-saving optimization, any
10868 unit declaring a tagged type should be compiled with the restriction,
10869 though this is not required.
10871 @node No_Task_Allocators
10872 @unnumberedsubsec No_Task_Allocators
10873 @findex No_Task_Allocators
10874 [RM D.7] There are no allocators for task types
10875 or types containing task subcomponents.
10877 @node No_Task_Attributes_Package
10878 @unnumberedsubsec No_Task_Attributes_Package
10879 @findex No_Task_Attributes_Package
10880 [GNAT] This restriction ensures at compile time that there are no implicit or
10881 explicit dependencies on the package @code{Ada.Task_Attributes}.
10883 @findex No_Task_Attributes
10884 The restriction @code{No_Task_Attributes} is recognized as a synonym
10885 for @code{No_Task_Attributes_Package}. This is retained for historical
10886 compatibility purposes (and a warning will be generated for its use if
10887 warnings on obsolescent features are activated).
10889 @node No_Task_Hierarchy
10890 @unnumberedsubsec No_Task_Hierarchy
10891 @findex No_Task_Hierarchy
10892 [RM D.7] All (non-environment) tasks depend
10893 directly on the environment task of the partition.
10895 @node No_Task_Termination
10896 @unnumberedsubsec No_Task_Termination
10897 @findex No_Task_Termination
10898 [RM D.7] Tasks which terminate are erroneous.
10901 @unnumberedsubsec No_Tasking
10903 [GNAT] This restriction prevents the declaration of tasks or task types
10904 throughout the partition. It is similar in effect to the use of
10905 @code{Max_Tasks => 0} except that violations are caught at compile time
10906 and cause an error message to be output either by the compiler or
10909 @node No_Terminate_Alternatives
10910 @unnumberedsubsec No_Terminate_Alternatives
10911 @findex No_Terminate_Alternatives
10912 [RM D.7] There are no selective accepts with terminate alternatives.
10914 @node No_Unchecked_Access
10915 @unnumberedsubsec No_Unchecked_Access
10916 @findex No_Unchecked_Access
10917 [RM H.4] This restriction ensures at compile time that there are no
10918 occurrences of the Unchecked_Access attribute.
10920 @node Simple_Barriers
10921 @unnumberedsubsec Simple_Barriers
10922 @findex Simple_Barriers
10923 [RM D.7] This restriction ensures at compile time that barriers in entry
10924 declarations for protected types are restricted to either static boolean
10925 expressions or references to simple boolean variables defined in the private
10926 part of the protected type. No other form of entry barriers is permitted.
10928 @findex Boolean_Entry_Barriers
10929 The restriction @code{Boolean_Entry_Barriers} is recognized as a
10930 synonym for @code{Simple_Barriers}. This is retained for historical
10931 compatibility purposes (and a warning will be generated for its use if
10932 warnings on obsolescent features are activated).
10934 @node Static_Priorities
10935 @unnumberedsubsec Static_Priorities
10936 @findex Static_Priorities
10937 [GNAT] This restriction ensures at compile time that all priority expressions
10938 are static, and that there are no dependences on the package
10939 @code{Ada.Dynamic_Priorities}.
10941 @node Static_Storage_Size
10942 @unnumberedsubsec Static_Storage_Size
10943 @findex Static_Storage_Size
10944 [GNAT] This restriction ensures at compile time that any expression appearing
10945 in a Storage_Size pragma or attribute definition clause is static.
10947 @node Program Unit Level Restrictions
10948 @section Program Unit Level Restrictions
10951 The second set of restriction identifiers
10952 does not require partition-wide consistency.
10953 The restriction may be enforced for a single
10954 compilation unit without any effect on any of the
10955 other compilation units in the partition.
10958 * No_Elaboration_Code::
10960 * No_Implementation_Aspect_Specifications::
10961 * No_Implementation_Attributes::
10962 * No_Implementation_Identifiers::
10963 * No_Implementation_Pragmas::
10964 * No_Implementation_Restrictions::
10965 * No_Implementation_Units::
10966 * No_Implicit_Aliasing::
10967 * No_Obsolescent_Features::
10968 * No_Wide_Characters::
10972 @node No_Elaboration_Code
10973 @unnumberedsubsec No_Elaboration_Code
10974 @findex No_Elaboration_Code
10975 [GNAT] This restriction ensures at compile time that no elaboration code is
10976 generated. Note that this is not the same condition as is enforced
10977 by pragma @code{Preelaborate}. There are cases in which pragma
10978 @code{Preelaborate} still permits code to be generated (e.g.@: code
10979 to initialize a large array to all zeroes), and there are cases of units
10980 which do not meet the requirements for pragma @code{Preelaborate},
10981 but for which no elaboration code is generated. Generally, it is
10982 the case that preelaborable units will meet the restrictions, with
10983 the exception of large aggregates initialized with an others_clause,
10984 and exception declarations (which generate calls to a run-time
10985 registry procedure). This restriction is enforced on
10986 a unit by unit basis, it need not be obeyed consistently
10987 throughout a partition.
10989 In the case of aggregates with others, if the aggregate has a dynamic
10990 size, there is no way to eliminate the elaboration code (such dynamic
10991 bounds would be incompatible with @code{Preelaborate} in any case). If
10992 the bounds are static, then use of this restriction actually modifies
10993 the code choice of the compiler to avoid generating a loop, and instead
10994 generate the aggregate statically if possible, no matter how many times
10995 the data for the others clause must be repeatedly generated.
10997 It is not possible to precisely document
10998 the constructs which are compatible with this restriction, since,
10999 unlike most other restrictions, this is not a restriction on the
11000 source code, but a restriction on the generated object code. For
11001 example, if the source contains a declaration:
11004 Val : constant Integer := X;
11008 where X is not a static constant, it may be possible, depending
11009 on complex optimization circuitry, for the compiler to figure
11010 out the value of X at compile time, in which case this initialization
11011 can be done by the loader, and requires no initialization code. It
11012 is not possible to document the precise conditions under which the
11013 optimizer can figure this out.
11015 Note that this the implementation of this restriction requires full
11016 code generation. If it is used in conjunction with "semantics only"
11017 checking, then some cases of violations may be missed.
11019 @node No_Entry_Queue
11020 @unnumberedsubsec No_Entry_Queue
11021 @findex No_Entry_Queue
11022 [GNAT] This restriction is a declaration that any protected entry compiled in
11023 the scope of the restriction has at most one task waiting on the entry
11024 at any one time, and so no queue is required. This restriction is not
11025 checked at compile time. A program execution is erroneous if an attempt
11026 is made to queue a second task on such an entry.
11028 @node No_Implementation_Aspect_Specifications
11029 @unnumberedsubsec No_Implementation_Aspect_Specifications
11030 @findex No_Implementation_Aspect_Specifications
11031 [RM 13.12.1] This restriction checks at compile time that no
11032 GNAT-defined aspects are present. With this restriction, the only
11033 aspects that can be used are those defined in the Ada Reference Manual.
11035 @node No_Implementation_Attributes
11036 @unnumberedsubsec No_Implementation_Attributes
11037 @findex No_Implementation_Attributes
11038 [RM 13.12.1] This restriction checks at compile time that no
11039 GNAT-defined attributes are present. With this restriction, the only
11040 attributes that can be used are those defined in the Ada Reference
11043 @node No_Implementation_Identifiers
11044 @unnumberedsubsec No_Implementation_Identifiers
11045 @findex No_Implementation_Identifiers
11046 [RM 13.12.1] This restriction checks at compile time that no
11047 implementation-defined identifiers (marked with pragma Implementation_Defined)
11048 occur within language-defined packages.
11050 @node No_Implementation_Pragmas
11051 @unnumberedsubsec No_Implementation_Pragmas
11052 @findex No_Implementation_Pragmas
11053 [RM 13.12.1] This restriction checks at compile time that no
11054 GNAT-defined pragmas are present. With this restriction, the only
11055 pragmas that can be used are those defined in the Ada Reference Manual.
11057 @node No_Implementation_Restrictions
11058 @unnumberedsubsec No_Implementation_Restrictions
11059 @findex No_Implementation_Restrictions
11060 [GNAT] This restriction checks at compile time that no GNAT-defined restriction
11061 identifiers (other than @code{No_Implementation_Restrictions} itself)
11062 are present. With this restriction, the only other restriction identifiers
11063 that can be used are those defined in the Ada Reference Manual.
11065 @node No_Implementation_Units
11066 @unnumberedsubsec No_Implementation_Units
11067 @findex No_Implementation_Units
11068 [RM 13.12.1] This restriction checks at compile time that there is no
11069 mention in the context clause of any implementation-defined descendants
11070 of packages Ada, Interfaces, or System.
11072 @node No_Implicit_Aliasing
11073 @unnumberedsubsec No_Implicit_Aliasing
11074 @findex No_Implicit_Aliasing
11075 [GNAT] This restriction, which is not required to be partition-wide consistent,
11076 requires an explicit aliased keyword for an object to which 'Access,
11077 'Unchecked_Access, or 'Address is applied, and forbids entirely the use of
11078 the 'Unrestricted_Access attribute for objects. Note: the reason that
11079 Unrestricted_Access is forbidden is that it would require the prefix
11080 to be aliased, and in such cases, it can always be replaced by
11081 the standard attribute Unchecked_Access which is preferable.
11083 @node No_Obsolescent_Features
11084 @unnumberedsubsec No_Obsolescent_Features
11085 @findex No_Obsolescent_Features
11086 [RM 13.12.1] This restriction checks at compile time that no obsolescent
11087 features are used, as defined in Annex J of the Ada Reference Manual.
11089 @node No_Wide_Characters
11090 @unnumberedsubsec No_Wide_Characters
11091 @findex No_Wide_Characters
11092 [GNAT] This restriction ensures at compile time that no uses of the types
11093 @code{Wide_Character} or @code{Wide_String} or corresponding wide
11095 appear, and that no wide or wide wide string or character literals
11096 appear in the program (that is literals representing characters not in
11097 type @code{Character}).
11100 @unnumberedsubsec SPARK_05
11102 [GNAT] This restriction checks at compile time that some constructs
11103 forbidden in SPARK 2005 are not present. Error messages related to
11104 SPARK restriction have the form:
11107 violation of restriction "SPARK_05" at <source-location>
11112 The restriction @code{SPARK} is recognized as a
11113 synonym for @code{SPARK_05}. This is retained for historical
11114 compatibility purposes (and an unconditional warning will be generated
11115 for its use, advising replacement by @code{SPARK}).
11117 This is not a replacement for the semantic checks performed by the
11118 SPARK Examiner tool, as the compiler currently only deals with code,
11119 not SPARK 2005 annotations, and does not guarantee catching all
11120 cases of constructs forbidden by SPARK 2005.
11122 Thus it may well be the case that code which passes the compiler with
11123 the SPARK restriction is rejected by the SPARK Examiner, e.g. due to
11124 the different visibility rules of the Examiner based on SPARK 2005
11125 @code{inherit} annotations.
11127 This restriction can be useful in providing an initial filter for code
11128 developed using SPARK 2005, or in examining legacy code to see how far
11129 it is from meeting SPARK restrictions.
11131 The list below summarizes the checks that are performed when this
11132 restriction is in force:
11134 @item No block statements
11135 @item No case statements with only an others clause
11136 @item Exit statements in loops must respect the SPARK 2005 language restrictions
11137 @item No goto statements
11138 @item Return can only appear as last statement in function
11139 @item Function must have return statement
11140 @item Loop parameter specification must include subtype mark
11141 @item Prefix of expanded name cannot be a loop statement
11142 @item Abstract subprogram not allowed
11143 @item User-defined operators not allowed
11144 @item Access type parameters not allowed
11145 @item Default expressions for parameters not allowed
11146 @item Default expressions for record fields not allowed
11147 @item No tasking constructs allowed
11148 @item Label needed at end of subprograms and packages
11149 @item No mixing of positional and named parameter association
11150 @item No access types as result type
11151 @item No unconstrained arrays as result types
11152 @item No null procedures
11153 @item Initial and later declarations must be in correct order (declaration can't come after body)
11154 @item No attributes on private types if full declaration not visible
11155 @item No package declaration within package specification
11156 @item No controlled types
11157 @item No discriminant types
11158 @item No overloading
11159 @item Selector name cannot be operator symbol (i.e. operator symbol cannot be prefixed)
11160 @item Access attribute not allowed
11161 @item Allocator not allowed
11162 @item Result of catenation must be String
11163 @item Operands of catenation must be string literal, static char or another catenation
11164 @item No conditional expressions
11165 @item No explicit dereference
11166 @item Quantified expression not allowed
11167 @item Slicing not allowed
11168 @item No exception renaming
11169 @item No generic renaming
11170 @item No object renaming
11171 @item No use clause
11172 @item Aggregates must be qualified
11173 @item Non-static choice in array aggregates not allowed
11174 @item The only view conversions which are allowed as in-out parameters are conversions of a tagged type to an ancestor type
11175 @item No mixing of positional and named association in aggregate, no multi choice
11176 @item AND, OR and XOR for arrays only allowed when operands have same static bounds
11177 @item Fixed point operands to * or / must be qualified or converted
11178 @item Comparison operators not allowed for Booleans or arrays (except strings)
11179 @item Equality not allowed for arrays with non-matching static bounds (except strings)
11180 @item Conversion / qualification not allowed for arrays with non-matching static bounds
11181 @item Subprogram declaration only allowed in package spec (unless followed by import)
11182 @item Access types not allowed
11183 @item Incomplete type declaration not allowed
11184 @item Object and subtype declarations must respect SPARK restrictions
11185 @item Digits or delta constraint not allowed
11186 @item Decimal fixed point type not allowed
11187 @item Aliasing of objects not allowed
11188 @item Modular type modulus must be power of 2
11189 @item Base not allowed on subtype mark
11190 @item Unary operators not allowed on modular types (except not)
11191 @item Untagged record cannot be null
11192 @item No class-wide operations
11193 @item Initialization expressions must respect SPARK restrictions
11194 @item Non-static ranges not allowed except in iteration schemes
11195 @item String subtypes must have lower bound of 1
11196 @item Subtype of Boolean cannot have constraint
11197 @item At most one tagged type or extension per package
11198 @item Interface is not allowed
11199 @item Character literal cannot be prefixed (selector name cannot be character literal)
11200 @item Record aggregate cannot contain 'others'
11201 @item Component association in record aggregate must contain a single choice
11202 @item Ancestor part cannot be a type mark
11203 @item Attributes 'Image, 'Width and 'Value not allowed
11204 @item Functions may not update globals
11205 @item Subprograms may not contain direct calls to themselves (prevents recursion within unit)
11206 @item Call to subprogram not allowed in same unit before body has been seen (prevents recursion within unit)
11209 The following restrictions are enforced, but note that they are actually more
11210 strict that the latest SPARK 2005 language definition:
11213 @item No derived types other than tagged type extensions
11214 @item Subtype of unconstrained array must have constraint
11217 This list summarises the main SPARK 2005 language rules that are not
11218 currently checked by the SPARK_05 restriction:
11221 @item SPARK annotations are treated as comments so are not checked at all
11222 @item Based real literals not allowed
11223 @item Objects cannot be initialized at declaration by calls to user-defined functions
11224 @item Objects cannot be initialized at declaration by assignments from variables
11225 @item Objects cannot be initialized at declaration by assignments from indexed/selected components
11226 @item Ranges shall not be null
11227 @item A fixed point delta expression must be a simple expression
11228 @item Restrictions on where renaming declarations may be placed
11229 @item Externals of mode 'out' cannot be referenced
11230 @item Externals of mode 'in' cannot be updated
11231 @item Loop with no iteration scheme or exits only allowed as last statement in main program or task
11232 @item Subprogram cannot have parent unit name
11233 @item SPARK 2005 inherited subprogram must be prefixed with overriding
11234 @item External variables (or functions that reference them) may not be passed as actual parameters
11235 @item Globals must be explicitly mentioned in contract
11236 @item Deferred constants cannot be completed by pragma Import
11237 @item Package initialization cannot read/write variables from other packages
11238 @item Prefix not allowed for entities that are directly visible
11239 @item Identifier declaration can't override inherited package name
11240 @item Cannot use Standard or other predefined packages as identifiers
11241 @item After renaming, cannot use the original name
11242 @item Subprograms can only be renamed to remove package prefix
11243 @item Pragma import must be immediately after entity it names
11244 @item No mutual recursion between multiple units (this can be checked with gnatcheck)
11247 Note that if a unit is compiled in Ada 95 mode with the SPARK restriction,
11248 violations will be reported for constructs forbidden in SPARK 95,
11249 instead of SPARK 2005.
11251 @c ------------------------
11252 @node Implementation Advice
11253 @chapter Implementation Advice
11255 The main text of the Ada Reference Manual describes the required
11256 behavior of all Ada compilers, and the GNAT compiler conforms to
11257 these requirements.
11259 In addition, there are sections throughout the Ada Reference Manual headed
11260 by the phrase ``Implementation advice''. These sections are not normative,
11261 i.e., they do not specify requirements that all compilers must
11262 follow. Rather they provide advice on generally desirable behavior. You
11263 may wonder why they are not requirements. The most typical answer is
11264 that they describe behavior that seems generally desirable, but cannot
11265 be provided on all systems, or which may be undesirable on some systems.
11267 As far as practical, GNAT follows the implementation advice sections in
11268 the Ada Reference Manual. This chapter contains a table giving the
11269 reference manual section number, paragraph number and several keywords
11270 for each advice. Each entry consists of the text of the advice followed
11271 by the GNAT interpretation of this advice. Most often, this simply says
11272 ``followed'', which means that GNAT follows the advice. However, in a
11273 number of cases, GNAT deliberately deviates from this advice, in which
11274 case the text describes what GNAT does and why.
11276 @cindex Error detection
11277 @unnumberedsec 1.1.3(20): Error Detection
11280 If an implementation detects the use of an unsupported Specialized Needs
11281 Annex feature at run time, it should raise @code{Program_Error} if
11284 Not relevant. All specialized needs annex features are either supported,
11285 or diagnosed at compile time.
11287 @cindex Child Units
11288 @unnumberedsec 1.1.3(31): Child Units
11291 If an implementation wishes to provide implementation-defined
11292 extensions to the functionality of a language-defined library unit, it
11293 should normally do so by adding children to the library unit.
11297 @cindex Bounded errors
11298 @unnumberedsec 1.1.5(12): Bounded Errors
11301 If an implementation detects a bounded error or erroneous
11302 execution, it should raise @code{Program_Error}.
11304 Followed in all cases in which the implementation detects a bounded
11305 error or erroneous execution. Not all such situations are detected at
11309 @unnumberedsec 2.8(16): Pragmas
11312 Normally, implementation-defined pragmas should have no semantic effect
11313 for error-free programs; that is, if the implementation-defined pragmas
11314 are removed from a working program, the program should still be legal,
11315 and should still have the same semantics.
11317 The following implementation defined pragmas are exceptions to this
11329 @item CPP_Constructor
11333 @item Interface_Name
11335 @item Machine_Attribute
11337 @item Unimplemented_Unit
11339 @item Unchecked_Union
11344 In each of the above cases, it is essential to the purpose of the pragma
11345 that this advice not be followed. For details see the separate section
11346 on implementation defined pragmas.
11348 @unnumberedsec 2.8(17-19): Pragmas
11351 Normally, an implementation should not define pragmas that can
11352 make an illegal program legal, except as follows:
11356 A pragma used to complete a declaration, such as a pragma @code{Import};
11360 A pragma used to configure the environment by adding, removing, or
11361 replacing @code{library_items}.
11363 See response to paragraph 16 of this same section.
11365 @cindex Character Sets
11366 @cindex Alternative Character Sets
11367 @unnumberedsec 3.5.2(5): Alternative Character Sets
11370 If an implementation supports a mode with alternative interpretations
11371 for @code{Character} and @code{Wide_Character}, the set of graphic
11372 characters of @code{Character} should nevertheless remain a proper
11373 subset of the set of graphic characters of @code{Wide_Character}. Any
11374 character set ``localizations'' should be reflected in the results of
11375 the subprograms defined in the language-defined package
11376 @code{Characters.Handling} (see A.3) available in such a mode. In a mode with
11377 an alternative interpretation of @code{Character}, the implementation should
11378 also support a corresponding change in what is a legal
11379 @code{identifier_letter}.
11381 Not all wide character modes follow this advice, in particular the JIS
11382 and IEC modes reflect standard usage in Japan, and in these encoding,
11383 the upper half of the Latin-1 set is not part of the wide-character
11384 subset, since the most significant bit is used for wide character
11385 encoding. However, this only applies to the external forms. Internally
11386 there is no such restriction.
11388 @cindex Integer types
11389 @unnumberedsec 3.5.4(28): Integer Types
11393 An implementation should support @code{Long_Integer} in addition to
11394 @code{Integer} if the target machine supports 32-bit (or longer)
11395 arithmetic. No other named integer subtypes are recommended for package
11396 @code{Standard}. Instead, appropriate named integer subtypes should be
11397 provided in the library package @code{Interfaces} (see B.2).
11399 @code{Long_Integer} is supported. Other standard integer types are supported
11400 so this advice is not fully followed. These types
11401 are supported for convenient interface to C, and so that all hardware
11402 types of the machine are easily available.
11403 @unnumberedsec 3.5.4(29): Integer Types
11407 An implementation for a two's complement machine should support
11408 modular types with a binary modulus up to @code{System.Max_Int*2+2}. An
11409 implementation should support a non-binary modules up to @code{Integer'Last}.
11413 @cindex Enumeration values
11414 @unnumberedsec 3.5.5(8): Enumeration Values
11417 For the evaluation of a call on @code{@var{S}'Pos} for an enumeration
11418 subtype, if the value of the operand does not correspond to the internal
11419 code for any enumeration literal of its type (perhaps due to an
11420 un-initialized variable), then the implementation should raise
11421 @code{Program_Error}. This is particularly important for enumeration
11422 types with noncontiguous internal codes specified by an
11423 enumeration_representation_clause.
11427 @cindex Float types
11428 @unnumberedsec 3.5.7(17): Float Types
11431 An implementation should support @code{Long_Float} in addition to
11432 @code{Float} if the target machine supports 11 or more digits of
11433 precision. No other named floating point subtypes are recommended for
11434 package @code{Standard}. Instead, appropriate named floating point subtypes
11435 should be provided in the library package @code{Interfaces} (see B.2).
11437 @code{Short_Float} and @code{Long_Long_Float} are also provided. The
11438 former provides improved compatibility with other implementations
11439 supporting this type. The latter corresponds to the highest precision
11440 floating-point type supported by the hardware. On most machines, this
11441 will be the same as @code{Long_Float}, but on some machines, it will
11442 correspond to the IEEE extended form. The notable case is all ia32
11443 (x86) implementations, where @code{Long_Long_Float} corresponds to
11444 the 80-bit extended precision format supported in hardware on this
11445 processor. Note that the 128-bit format on SPARC is not supported,
11446 since this is a software rather than a hardware format.
11448 @cindex Multidimensional arrays
11449 @cindex Arrays, multidimensional
11450 @unnumberedsec 3.6.2(11): Multidimensional Arrays
11453 An implementation should normally represent multidimensional arrays in
11454 row-major order, consistent with the notation used for multidimensional
11455 array aggregates (see 4.3.3). However, if a pragma @code{Convention}
11456 (@code{Fortran}, @dots{}) applies to a multidimensional array type, then
11457 column-major order should be used instead (see B.5, ``Interfacing with
11462 @findex Duration'Small
11463 @unnumberedsec 9.6(30-31): Duration'Small
11466 Whenever possible in an implementation, the value of @code{Duration'Small}
11467 should be no greater than 100 microseconds.
11469 Followed. (@code{Duration'Small} = 10**(@minus{}9)).
11473 The time base for @code{delay_relative_statements} should be monotonic;
11474 it need not be the same time base as used for @code{Calendar.Clock}.
11478 @unnumberedsec 10.2.1(12): Consistent Representation
11481 In an implementation, a type declared in a pre-elaborated package should
11482 have the same representation in every elaboration of a given version of
11483 the package, whether the elaborations occur in distinct executions of
11484 the same program, or in executions of distinct programs or partitions
11485 that include the given version.
11487 Followed, except in the case of tagged types. Tagged types involve
11488 implicit pointers to a local copy of a dispatch table, and these pointers
11489 have representations which thus depend on a particular elaboration of the
11490 package. It is not easy to see how it would be possible to follow this
11491 advice without severely impacting efficiency of execution.
11493 @cindex Exception information
11494 @unnumberedsec 11.4.1(19): Exception Information
11497 @code{Exception_Message} by default and @code{Exception_Information}
11498 should produce information useful for
11499 debugging. @code{Exception_Message} should be short, about one
11500 line. @code{Exception_Information} can be long. @code{Exception_Message}
11501 should not include the
11502 @code{Exception_Name}. @code{Exception_Information} should include both
11503 the @code{Exception_Name} and the @code{Exception_Message}.
11505 Followed. For each exception that doesn't have a specified
11506 @code{Exception_Message}, the compiler generates one containing the location
11507 of the raise statement. This location has the form ``file:line'', where
11508 file is the short file name (without path information) and line is the line
11509 number in the file. Note that in the case of the Zero Cost Exception
11510 mechanism, these messages become redundant with the Exception_Information that
11511 contains a full backtrace of the calling sequence, so they are disabled.
11512 To disable explicitly the generation of the source location message, use the
11513 Pragma @code{Discard_Names}.
11515 @cindex Suppression of checks
11516 @cindex Checks, suppression of
11517 @unnumberedsec 11.5(28): Suppression of Checks
11520 The implementation should minimize the code executed for checks that
11521 have been suppressed.
11525 @cindex Representation clauses
11526 @unnumberedsec 13.1 (21-24): Representation Clauses
11529 The recommended level of support for all representation items is
11530 qualified as follows:
11534 An implementation need not support representation items containing
11535 non-static expressions, except that an implementation should support a
11536 representation item for a given entity if each non-static expression in
11537 the representation item is a name that statically denotes a constant
11538 declared before the entity.
11540 Followed. In fact, GNAT goes beyond the recommended level of support
11541 by allowing nonstatic expressions in some representation clauses even
11542 without the need to declare constants initialized with the values of
11546 @smallexample @c ada
11549 for Y'Address use X'Address;>>
11554 An implementation need not support a specification for the @code{Size}
11555 for a given composite subtype, nor the size or storage place for an
11556 object (including a component) of a given composite subtype, unless the
11557 constraints on the subtype and its composite subcomponents (if any) are
11558 all static constraints.
11560 Followed. Size Clauses are not permitted on non-static components, as
11565 An aliased component, or a component whose type is by-reference, should
11566 always be allocated at an addressable location.
11570 @cindex Packed types
11571 @unnumberedsec 13.2(6-8): Packed Types
11574 If a type is packed, then the implementation should try to minimize
11575 storage allocated to objects of the type, possibly at the expense of
11576 speed of accessing components, subject to reasonable complexity in
11577 addressing calculations.
11581 The recommended level of support pragma @code{Pack} is:
11583 For a packed record type, the components should be packed as tightly as
11584 possible subject to the Sizes of the component subtypes, and subject to
11585 any @code{record_representation_clause} that applies to the type; the
11586 implementation may, but need not, reorder components or cross aligned
11587 word boundaries to improve the packing. A component whose @code{Size} is
11588 greater than the word size may be allocated an integral number of words.
11590 Followed. Tight packing of arrays is supported for all component sizes
11591 up to 64-bits. If the array component size is 1 (that is to say, if
11592 the component is a boolean type or an enumeration type with two values)
11593 then values of the type are implicitly initialized to zero. This
11594 happens both for objects of the packed type, and for objects that have a
11595 subcomponent of the packed type.
11599 An implementation should support Address clauses for imported
11603 @cindex @code{Address} clauses
11604 @unnumberedsec 13.3(14-19): Address Clauses
11608 For an array @var{X}, @code{@var{X}'Address} should point at the first
11609 component of the array, and not at the array bounds.
11615 The recommended level of support for the @code{Address} attribute is:
11617 @code{@var{X}'Address} should produce a useful result if @var{X} is an
11618 object that is aliased or of a by-reference type, or is an entity whose
11619 @code{Address} has been specified.
11621 Followed. A valid address will be produced even if none of those
11622 conditions have been met. If necessary, the object is forced into
11623 memory to ensure the address is valid.
11627 An implementation should support @code{Address} clauses for imported
11634 Objects (including subcomponents) that are aliased or of a by-reference
11635 type should be allocated on storage element boundaries.
11641 If the @code{Address} of an object is specified, or it is imported or exported,
11642 then the implementation should not perform optimizations based on
11643 assumptions of no aliases.
11647 @cindex @code{Alignment} clauses
11648 @unnumberedsec 13.3(29-35): Alignment Clauses
11651 The recommended level of support for the @code{Alignment} attribute for
11654 An implementation should support specified Alignments that are factors
11655 and multiples of the number of storage elements per word, subject to the
11662 An implementation need not support specified @code{Alignment}s for
11663 combinations of @code{Size}s and @code{Alignment}s that cannot be easily
11664 loaded and stored by available machine instructions.
11670 An implementation need not support specified @code{Alignment}s that are
11671 greater than the maximum @code{Alignment} the implementation ever returns by
11678 The recommended level of support for the @code{Alignment} attribute for
11681 Same as above, for subtypes, but in addition:
11687 For stand-alone library-level objects of statically constrained
11688 subtypes, the implementation should support all @code{Alignment}s
11689 supported by the target linker. For example, page alignment is likely to
11690 be supported for such objects, but not for subtypes.
11694 @cindex @code{Size} clauses
11695 @unnumberedsec 13.3(42-43): Size Clauses
11698 The recommended level of support for the @code{Size} attribute of
11701 A @code{Size} clause should be supported for an object if the specified
11702 @code{Size} is at least as large as its subtype's @code{Size}, and
11703 corresponds to a size in storage elements that is a multiple of the
11704 object's @code{Alignment} (if the @code{Alignment} is nonzero).
11708 @unnumberedsec 13.3(50-56): Size Clauses
11711 If the @code{Size} of a subtype is specified, and allows for efficient
11712 independent addressability (see 9.10) on the target architecture, then
11713 the @code{Size} of the following objects of the subtype should equal the
11714 @code{Size} of the subtype:
11716 Aliased objects (including components).
11722 @code{Size} clause on a composite subtype should not affect the
11723 internal layout of components.
11725 Followed. But note that this can be overridden by use of the implementation
11726 pragma Implicit_Packing in the case of packed arrays.
11730 The recommended level of support for the @code{Size} attribute of subtypes is:
11734 The @code{Size} (if not specified) of a static discrete or fixed point
11735 subtype should be the number of bits needed to represent each value
11736 belonging to the subtype using an unbiased representation, leaving space
11737 for a sign bit only if the subtype contains negative values. If such a
11738 subtype is a first subtype, then an implementation should support a
11739 specified @code{Size} for it that reflects this representation.
11745 For a subtype implemented with levels of indirection, the @code{Size}
11746 should include the size of the pointers, but not the size of what they
11751 @cindex @code{Component_Size} clauses
11752 @unnumberedsec 13.3(71-73): Component Size Clauses
11755 The recommended level of support for the @code{Component_Size}
11760 An implementation need not support specified @code{Component_Sizes} that are
11761 less than the @code{Size} of the component subtype.
11767 An implementation should support specified @code{Component_Size}s that
11768 are factors and multiples of the word size. For such
11769 @code{Component_Size}s, the array should contain no gaps between
11770 components. For other @code{Component_Size}s (if supported), the array
11771 should contain no gaps between components when packing is also
11772 specified; the implementation should forbid this combination in cases
11773 where it cannot support a no-gaps representation.
11777 @cindex Enumeration representation clauses
11778 @cindex Representation clauses, enumeration
11779 @unnumberedsec 13.4(9-10): Enumeration Representation Clauses
11782 The recommended level of support for enumeration representation clauses
11785 An implementation need not support enumeration representation clauses
11786 for boolean types, but should at minimum support the internal codes in
11787 the range @code{System.Min_Int.System.Max_Int}.
11791 @cindex Record representation clauses
11792 @cindex Representation clauses, records
11793 @unnumberedsec 13.5.1(17-22): Record Representation Clauses
11796 The recommended level of support for
11797 @*@code{record_representation_clauses} is:
11799 An implementation should support storage places that can be extracted
11800 with a load, mask, shift sequence of machine code, and set with a load,
11801 shift, mask, store sequence, given the available machine instructions
11802 and run-time model.
11808 A storage place should be supported if its size is equal to the
11809 @code{Size} of the component subtype, and it starts and ends on a
11810 boundary that obeys the @code{Alignment} of the component subtype.
11816 If the default bit ordering applies to the declaration of a given type,
11817 then for a component whose subtype's @code{Size} is less than the word
11818 size, any storage place that does not cross an aligned word boundary
11819 should be supported.
11825 An implementation may reserve a storage place for the tag field of a
11826 tagged type, and disallow other components from overlapping that place.
11828 Followed. The storage place for the tag field is the beginning of the tagged
11829 record, and its size is Address'Size. GNAT will reject an explicit component
11830 clause for the tag field.
11834 An implementation need not support a @code{component_clause} for a
11835 component of an extension part if the storage place is not after the
11836 storage places of all components of the parent type, whether or not
11837 those storage places had been specified.
11839 Followed. The above advice on record representation clauses is followed,
11840 and all mentioned features are implemented.
11842 @cindex Storage place attributes
11843 @unnumberedsec 13.5.2(5): Storage Place Attributes
11846 If a component is represented using some form of pointer (such as an
11847 offset) to the actual data of the component, and this data is contiguous
11848 with the rest of the object, then the storage place attributes should
11849 reflect the place of the actual data, not the pointer. If a component is
11850 allocated discontinuously from the rest of the object, then a warning
11851 should be generated upon reference to one of its storage place
11854 Followed. There are no such components in GNAT@.
11856 @cindex Bit ordering
11857 @unnumberedsec 13.5.3(7-8): Bit Ordering
11860 The recommended level of support for the non-default bit ordering is:
11864 If @code{Word_Size} = @code{Storage_Unit}, then the implementation
11865 should support the non-default bit ordering in addition to the default
11868 Followed. Word size does not equal storage size in this implementation.
11869 Thus non-default bit ordering is not supported.
11871 @cindex @code{Address}, as private type
11872 @unnumberedsec 13.7(37): Address as Private
11875 @code{Address} should be of a private type.
11879 @cindex Operations, on @code{Address}
11880 @cindex @code{Address}, operations of
11881 @unnumberedsec 13.7.1(16): Address Operations
11884 Operations in @code{System} and its children should reflect the target
11885 environment semantics as closely as is reasonable. For example, on most
11886 machines, it makes sense for address arithmetic to ``wrap around''.
11887 Operations that do not make sense should raise @code{Program_Error}.
11889 Followed. Address arithmetic is modular arithmetic that wraps around. No
11890 operation raises @code{Program_Error}, since all operations make sense.
11892 @cindex Unchecked conversion
11893 @unnumberedsec 13.9(14-17): Unchecked Conversion
11896 The @code{Size} of an array object should not include its bounds; hence,
11897 the bounds should not be part of the converted data.
11903 The implementation should not generate unnecessary run-time checks to
11904 ensure that the representation of @var{S} is a representation of the
11905 target type. It should take advantage of the permission to return by
11906 reference when possible. Restrictions on unchecked conversions should be
11907 avoided unless required by the target environment.
11909 Followed. There are no restrictions on unchecked conversion. A warning is
11910 generated if the source and target types do not have the same size since
11911 the semantics in this case may be target dependent.
11915 The recommended level of support for unchecked conversions is:
11919 Unchecked conversions should be supported and should be reversible in
11920 the cases where this clause defines the result. To enable meaningful use
11921 of unchecked conversion, a contiguous representation should be used for
11922 elementary subtypes, for statically constrained array subtypes whose
11923 component subtype is one of the subtypes described in this paragraph,
11924 and for record subtypes without discriminants whose component subtypes
11925 are described in this paragraph.
11929 @cindex Heap usage, implicit
11930 @unnumberedsec 13.11(23-25): Implicit Heap Usage
11933 An implementation should document any cases in which it dynamically
11934 allocates heap storage for a purpose other than the evaluation of an
11937 Followed, the only other points at which heap storage is dynamically
11938 allocated are as follows:
11942 At initial elaboration time, to allocate dynamically sized global
11946 To allocate space for a task when a task is created.
11949 To extend the secondary stack dynamically when needed. The secondary
11950 stack is used for returning variable length results.
11955 A default (implementation-provided) storage pool for an
11956 access-to-constant type should not have overhead to support deallocation of
11957 individual objects.
11963 A storage pool for an anonymous access type should be created at the
11964 point of an allocator for the type, and be reclaimed when the designated
11965 object becomes inaccessible.
11969 @cindex Unchecked deallocation
11970 @unnumberedsec 13.11.2(17): Unchecked De-allocation
11973 For a standard storage pool, @code{Free} should actually reclaim the
11978 @cindex Stream oriented attributes
11979 @unnumberedsec 13.13.2(17): Stream Oriented Attributes
11982 If a stream element is the same size as a storage element, then the
11983 normal in-memory representation should be used by @code{Read} and
11984 @code{Write} for scalar objects. Otherwise, @code{Read} and @code{Write}
11985 should use the smallest number of stream elements needed to represent
11986 all values in the base range of the scalar type.
11989 Followed. By default, GNAT uses the interpretation suggested by AI-195,
11990 which specifies using the size of the first subtype.
11991 However, such an implementation is based on direct binary
11992 representations and is therefore target- and endianness-dependent.
11993 To address this issue, GNAT also supplies an alternate implementation
11994 of the stream attributes @code{Read} and @code{Write},
11995 which uses the target-independent XDR standard representation
11997 @cindex XDR representation
11998 @cindex @code{Read} attribute
11999 @cindex @code{Write} attribute
12000 @cindex Stream oriented attributes
12001 The XDR implementation is provided as an alternative body of the
12002 @code{System.Stream_Attributes} package, in the file
12003 @file{s-stratt-xdr.adb} in the GNAT library.
12004 There is no @file{s-stratt-xdr.ads} file.
12005 In order to install the XDR implementation, do the following:
12007 @item Replace the default implementation of the
12008 @code{System.Stream_Attributes} package with the XDR implementation.
12009 For example on a Unix platform issue the commands:
12011 $ mv s-stratt.adb s-stratt-default.adb
12012 $ mv s-stratt-xdr.adb s-stratt.adb
12016 Rebuild the GNAT run-time library as documented in
12017 @ref{GNAT and Libraries,,, gnat_ugn, @value{EDITION} User's Guide}.
12020 @unnumberedsec A.1(52): Names of Predefined Numeric Types
12023 If an implementation provides additional named predefined integer types,
12024 then the names should end with @samp{Integer} as in
12025 @samp{Long_Integer}. If an implementation provides additional named
12026 predefined floating point types, then the names should end with
12027 @samp{Float} as in @samp{Long_Float}.
12031 @findex Ada.Characters.Handling
12032 @unnumberedsec A.3.2(49): @code{Ada.Characters.Handling}
12035 If an implementation provides a localized definition of @code{Character}
12036 or @code{Wide_Character}, then the effects of the subprograms in
12037 @code{Characters.Handling} should reflect the localizations. See also
12040 Followed. GNAT provides no such localized definitions.
12042 @cindex Bounded-length strings
12043 @unnumberedsec A.4.4(106): Bounded-Length String Handling
12046 Bounded string objects should not be implemented by implicit pointers
12047 and dynamic allocation.
12049 Followed. No implicit pointers or dynamic allocation are used.
12051 @cindex Random number generation
12052 @unnumberedsec A.5.2(46-47): Random Number Generation
12055 Any storage associated with an object of type @code{Generator} should be
12056 reclaimed on exit from the scope of the object.
12062 If the generator period is sufficiently long in relation to the number
12063 of distinct initiator values, then each possible value of
12064 @code{Initiator} passed to @code{Reset} should initiate a sequence of
12065 random numbers that does not, in a practical sense, overlap the sequence
12066 initiated by any other value. If this is not possible, then the mapping
12067 between initiator values and generator states should be a rapidly
12068 varying function of the initiator value.
12070 Followed. The generator period is sufficiently long for the first
12071 condition here to hold true.
12073 @findex Get_Immediate
12074 @unnumberedsec A.10.7(23): @code{Get_Immediate}
12077 The @code{Get_Immediate} procedures should be implemented with
12078 unbuffered input. For a device such as a keyboard, input should be
12079 @dfn{available} if a key has already been typed, whereas for a disk
12080 file, input should always be available except at end of file. For a file
12081 associated with a keyboard-like device, any line-editing features of the
12082 underlying operating system should be disabled during the execution of
12083 @code{Get_Immediate}.
12085 Followed on all targets except VxWorks. For VxWorks, there is no way to
12086 provide this functionality that does not result in the input buffer being
12087 flushed before the @code{Get_Immediate} call. A special unit
12088 @code{Interfaces.Vxworks.IO} is provided that contains routines to enable
12089 this functionality.
12092 @unnumberedsec B.1(39-41): Pragma @code{Export}
12095 If an implementation supports pragma @code{Export} to a given language,
12096 then it should also allow the main subprogram to be written in that
12097 language. It should support some mechanism for invoking the elaboration
12098 of the Ada library units included in the system, and for invoking the
12099 finalization of the environment task. On typical systems, the
12100 recommended mechanism is to provide two subprograms whose link names are
12101 @code{adainit} and @code{adafinal}. @code{adainit} should contain the
12102 elaboration code for library units. @code{adafinal} should contain the
12103 finalization code. These subprograms should have no effect the second
12104 and subsequent time they are called.
12110 Automatic elaboration of pre-elaborated packages should be
12111 provided when pragma @code{Export} is supported.
12113 Followed when the main program is in Ada. If the main program is in a
12114 foreign language, then
12115 @code{adainit} must be called to elaborate pre-elaborated
12120 For each supported convention @var{L} other than @code{Intrinsic}, an
12121 implementation should support @code{Import} and @code{Export} pragmas
12122 for objects of @var{L}-compatible types and for subprograms, and pragma
12123 @code{Convention} for @var{L}-eligible types and for subprograms,
12124 presuming the other language has corresponding features. Pragma
12125 @code{Convention} need not be supported for scalar types.
12129 @cindex Package @code{Interfaces}
12131 @unnumberedsec B.2(12-13): Package @code{Interfaces}
12134 For each implementation-defined convention identifier, there should be a
12135 child package of package Interfaces with the corresponding name. This
12136 package should contain any declarations that would be useful for
12137 interfacing to the language (implementation) represented by the
12138 convention. Any declarations useful for interfacing to any language on
12139 the given hardware architecture should be provided directly in
12146 An implementation supporting an interface to C, COBOL, or Fortran should
12147 provide the corresponding package or packages described in the following
12150 Followed. GNAT provides all the packages described in this section.
12152 @cindex C, interfacing with
12153 @unnumberedsec B.3(63-71): Interfacing with C
12156 An implementation should support the following interface correspondences
12157 between Ada and C@.
12163 An Ada procedure corresponds to a void-returning C function.
12169 An Ada function corresponds to a non-void C function.
12175 An Ada @code{in} scalar parameter is passed as a scalar argument to a C
12182 An Ada @code{in} parameter of an access-to-object type with designated
12183 type @var{T} is passed as a @code{@var{t}*} argument to a C function,
12184 where @var{t} is the C type corresponding to the Ada type @var{T}.
12190 An Ada access @var{T} parameter, or an Ada @code{out} or @code{in out}
12191 parameter of an elementary type @var{T}, is passed as a @code{@var{t}*}
12192 argument to a C function, where @var{t} is the C type corresponding to
12193 the Ada type @var{T}. In the case of an elementary @code{out} or
12194 @code{in out} parameter, a pointer to a temporary copy is used to
12195 preserve by-copy semantics.
12201 An Ada parameter of a record type @var{T}, of any mode, is passed as a
12202 @code{@var{t}*} argument to a C function, where @var{t} is the C
12203 structure corresponding to the Ada type @var{T}.
12205 Followed. This convention may be overridden by the use of the C_Pass_By_Copy
12206 pragma, or Convention, or by explicitly specifying the mechanism for a given
12207 call using an extended import or export pragma.
12211 An Ada parameter of an array type with component type @var{T}, of any
12212 mode, is passed as a @code{@var{t}*} argument to a C function, where
12213 @var{t} is the C type corresponding to the Ada type @var{T}.
12219 An Ada parameter of an access-to-subprogram type is passed as a pointer
12220 to a C function whose prototype corresponds to the designated
12221 subprogram's specification.
12225 @cindex COBOL, interfacing with
12226 @unnumberedsec B.4(95-98): Interfacing with COBOL
12229 An Ada implementation should support the following interface
12230 correspondences between Ada and COBOL@.
12236 An Ada access @var{T} parameter is passed as a @samp{BY REFERENCE} data item of
12237 the COBOL type corresponding to @var{T}.
12243 An Ada in scalar parameter is passed as a @samp{BY CONTENT} data item of
12244 the corresponding COBOL type.
12250 Any other Ada parameter is passed as a @samp{BY REFERENCE} data item of the
12251 COBOL type corresponding to the Ada parameter type; for scalars, a local
12252 copy is used if necessary to ensure by-copy semantics.
12256 @cindex Fortran, interfacing with
12257 @unnumberedsec B.5(22-26): Interfacing with Fortran
12260 An Ada implementation should support the following interface
12261 correspondences between Ada and Fortran:
12267 An Ada procedure corresponds to a Fortran subroutine.
12273 An Ada function corresponds to a Fortran function.
12279 An Ada parameter of an elementary, array, or record type @var{T} is
12280 passed as a @var{T} argument to a Fortran procedure, where @var{T} is
12281 the Fortran type corresponding to the Ada type @var{T}, and where the
12282 INTENT attribute of the corresponding dummy argument matches the Ada
12283 formal parameter mode; the Fortran implementation's parameter passing
12284 conventions are used. For elementary types, a local copy is used if
12285 necessary to ensure by-copy semantics.
12291 An Ada parameter of an access-to-subprogram type is passed as a
12292 reference to a Fortran procedure whose interface corresponds to the
12293 designated subprogram's specification.
12297 @cindex Machine operations
12298 @unnumberedsec C.1(3-5): Access to Machine Operations
12301 The machine code or intrinsic support should allow access to all
12302 operations normally available to assembly language programmers for the
12303 target environment, including privileged instructions, if any.
12309 The interfacing pragmas (see Annex B) should support interface to
12310 assembler; the default assembler should be associated with the
12311 convention identifier @code{Assembler}.
12317 If an entity is exported to assembly language, then the implementation
12318 should allocate it at an addressable location, and should ensure that it
12319 is retained by the linking process, even if not otherwise referenced
12320 from the Ada code. The implementation should assume that any call to a
12321 machine code or assembler subprogram is allowed to read or update every
12322 object that is specified as exported.
12326 @unnumberedsec C.1(10-16): Access to Machine Operations
12329 The implementation should ensure that little or no overhead is
12330 associated with calling intrinsic and machine-code subprograms.
12332 Followed for both intrinsics and machine-code subprograms.
12336 It is recommended that intrinsic subprograms be provided for convenient
12337 access to any machine operations that provide special capabilities or
12338 efficiency and that are not otherwise available through the language
12341 Followed. A full set of machine operation intrinsic subprograms is provided.
12345 Atomic read-modify-write operations---e.g.@:, test and set, compare and
12346 swap, decrement and test, enqueue/dequeue.
12348 Followed on any target supporting such operations.
12352 Standard numeric functions---e.g.@:, sin, log.
12354 Followed on any target supporting such operations.
12358 String manipulation operations---e.g.@:, translate and test.
12360 Followed on any target supporting such operations.
12364 Vector operations---e.g.@:, compare vector against thresholds.
12366 Followed on any target supporting such operations.
12370 Direct operations on I/O ports.
12372 Followed on any target supporting such operations.
12374 @cindex Interrupt support
12375 @unnumberedsec C.3(28): Interrupt Support
12378 If the @code{Ceiling_Locking} policy is not in effect, the
12379 implementation should provide means for the application to specify which
12380 interrupts are to be blocked during protected actions, if the underlying
12381 system allows for a finer-grain control of interrupt blocking.
12383 Followed. The underlying system does not allow for finer-grain control
12384 of interrupt blocking.
12386 @cindex Protected procedure handlers
12387 @unnumberedsec C.3.1(20-21): Protected Procedure Handlers
12390 Whenever possible, the implementation should allow interrupt handlers to
12391 be called directly by the hardware.
12393 Followed on any target where the underlying operating system permits
12398 Whenever practical, violations of any
12399 implementation-defined restrictions should be detected before run time.
12401 Followed. Compile time warnings are given when possible.
12403 @cindex Package @code{Interrupts}
12405 @unnumberedsec C.3.2(25): Package @code{Interrupts}
12409 If implementation-defined forms of interrupt handler procedures are
12410 supported, such as protected procedures with parameters, then for each
12411 such form of a handler, a type analogous to @code{Parameterless_Handler}
12412 should be specified in a child package of @code{Interrupts}, with the
12413 same operations as in the predefined package Interrupts.
12417 @cindex Pre-elaboration requirements
12418 @unnumberedsec C.4(14): Pre-elaboration Requirements
12421 It is recommended that pre-elaborated packages be implemented in such a
12422 way that there should be little or no code executed at run time for the
12423 elaboration of entities not already covered by the Implementation
12426 Followed. Executable code is generated in some cases, e.g.@: loops
12427 to initialize large arrays.
12429 @unnumberedsec C.5(8): Pragma @code{Discard_Names}
12432 If the pragma applies to an entity, then the implementation should
12433 reduce the amount of storage used for storing names associated with that
12438 @cindex Package @code{Task_Attributes}
12439 @findex Task_Attributes
12440 @unnumberedsec C.7.2(30): The Package Task_Attributes
12443 Some implementations are targeted to domains in which memory use at run
12444 time must be completely deterministic. For such implementations, it is
12445 recommended that the storage for task attributes will be pre-allocated
12446 statically and not from the heap. This can be accomplished by either
12447 placing restrictions on the number and the size of the task's
12448 attributes, or by using the pre-allocated storage for the first @var{N}
12449 attribute objects, and the heap for the others. In the latter case,
12450 @var{N} should be documented.
12452 Not followed. This implementation is not targeted to such a domain.
12454 @cindex Locking Policies
12455 @unnumberedsec D.3(17): Locking Policies
12459 The implementation should use names that end with @samp{_Locking} for
12460 locking policies defined by the implementation.
12462 Followed. Two implementation-defined locking policies are defined,
12463 whose names (@code{Inheritance_Locking} and
12464 @code{Concurrent_Readers_Locking}) follow this suggestion.
12466 @cindex Entry queuing policies
12467 @unnumberedsec D.4(16): Entry Queuing Policies
12470 Names that end with @samp{_Queuing} should be used
12471 for all implementation-defined queuing policies.
12473 Followed. No such implementation-defined queuing policies exist.
12475 @cindex Preemptive abort
12476 @unnumberedsec D.6(9-10): Preemptive Abort
12479 Even though the @code{abort_statement} is included in the list of
12480 potentially blocking operations (see 9.5.1), it is recommended that this
12481 statement be implemented in a way that never requires the task executing
12482 the @code{abort_statement} to block.
12488 On a multi-processor, the delay associated with aborting a task on
12489 another processor should be bounded; the implementation should use
12490 periodic polling, if necessary, to achieve this.
12494 @cindex Tasking restrictions
12495 @unnumberedsec D.7(21): Tasking Restrictions
12498 When feasible, the implementation should take advantage of the specified
12499 restrictions to produce a more efficient implementation.
12501 GNAT currently takes advantage of these restrictions by providing an optimized
12502 run time when the Ravenscar profile and the GNAT restricted run time set
12503 of restrictions are specified. See pragma @code{Profile (Ravenscar)} and
12504 pragma @code{Profile (Restricted)} for more details.
12506 @cindex Time, monotonic
12507 @unnumberedsec D.8(47-49): Monotonic Time
12510 When appropriate, implementations should provide configuration
12511 mechanisms to change the value of @code{Tick}.
12513 Such configuration mechanisms are not appropriate to this implementation
12514 and are thus not supported.
12518 It is recommended that @code{Calendar.Clock} and @code{Real_Time.Clock}
12519 be implemented as transformations of the same time base.
12525 It is recommended that the @dfn{best} time base which exists in
12526 the underlying system be available to the application through
12527 @code{Clock}. @dfn{Best} may mean highest accuracy or largest range.
12531 @cindex Partition communication subsystem
12533 @unnumberedsec E.5(28-29): Partition Communication Subsystem
12536 Whenever possible, the PCS on the called partition should allow for
12537 multiple tasks to call the RPC-receiver with different messages and
12538 should allow them to block until the corresponding subprogram body
12541 Followed by GLADE, a separately supplied PCS that can be used with
12546 The @code{Write} operation on a stream of type @code{Params_Stream_Type}
12547 should raise @code{Storage_Error} if it runs out of space trying to
12548 write the @code{Item} into the stream.
12550 Followed by GLADE, a separately supplied PCS that can be used with
12553 @cindex COBOL support
12554 @unnumberedsec F(7): COBOL Support
12557 If COBOL (respectively, C) is widely supported in the target
12558 environment, implementations supporting the Information Systems Annex
12559 should provide the child package @code{Interfaces.COBOL} (respectively,
12560 @code{Interfaces.C}) specified in Annex B and should support a
12561 @code{convention_identifier} of COBOL (respectively, C) in the interfacing
12562 pragmas (see Annex B), thus allowing Ada programs to interface with
12563 programs written in that language.
12567 @cindex Decimal radix support
12568 @unnumberedsec F.1(2): Decimal Radix Support
12571 Packed decimal should be used as the internal representation for objects
12572 of subtype @var{S} when @var{S}'Machine_Radix = 10.
12574 Not followed. GNAT ignores @var{S}'Machine_Radix and always uses binary
12578 @unnumberedsec G: Numerics
12581 If Fortran (respectively, C) is widely supported in the target
12582 environment, implementations supporting the Numerics Annex
12583 should provide the child package @code{Interfaces.Fortran} (respectively,
12584 @code{Interfaces.C}) specified in Annex B and should support a
12585 @code{convention_identifier} of Fortran (respectively, C) in the interfacing
12586 pragmas (see Annex B), thus allowing Ada programs to interface with
12587 programs written in that language.
12591 @cindex Complex types
12592 @unnumberedsec G.1.1(56-58): Complex Types
12595 Because the usual mathematical meaning of multiplication of a complex
12596 operand and a real operand is that of the scaling of both components of
12597 the former by the latter, an implementation should not perform this
12598 operation by first promoting the real operand to complex type and then
12599 performing a full complex multiplication. In systems that, in the
12600 future, support an Ada binding to IEC 559:1989, the latter technique
12601 will not generate the required result when one of the components of the
12602 complex operand is infinite. (Explicit multiplication of the infinite
12603 component by the zero component obtained during promotion yields a NaN
12604 that propagates into the final result.) Analogous advice applies in the
12605 case of multiplication of a complex operand and a pure-imaginary
12606 operand, and in the case of division of a complex operand by a real or
12607 pure-imaginary operand.
12613 Similarly, because the usual mathematical meaning of addition of a
12614 complex operand and a real operand is that the imaginary operand remains
12615 unchanged, an implementation should not perform this operation by first
12616 promoting the real operand to complex type and then performing a full
12617 complex addition. In implementations in which the @code{Signed_Zeros}
12618 attribute of the component type is @code{True} (and which therefore
12619 conform to IEC 559:1989 in regard to the handling of the sign of zero in
12620 predefined arithmetic operations), the latter technique will not
12621 generate the required result when the imaginary component of the complex
12622 operand is a negatively signed zero. (Explicit addition of the negative
12623 zero to the zero obtained during promotion yields a positive zero.)
12624 Analogous advice applies in the case of addition of a complex operand
12625 and a pure-imaginary operand, and in the case of subtraction of a
12626 complex operand and a real or pure-imaginary operand.
12632 Implementations in which @code{Real'Signed_Zeros} is @code{True} should
12633 attempt to provide a rational treatment of the signs of zero results and
12634 result components. As one example, the result of the @code{Argument}
12635 function should have the sign of the imaginary component of the
12636 parameter @code{X} when the point represented by that parameter lies on
12637 the positive real axis; as another, the sign of the imaginary component
12638 of the @code{Compose_From_Polar} function should be the same as
12639 (respectively, the opposite of) that of the @code{Argument} parameter when that
12640 parameter has a value of zero and the @code{Modulus} parameter has a
12641 nonnegative (respectively, negative) value.
12645 @cindex Complex elementary functions
12646 @unnumberedsec G.1.2(49): Complex Elementary Functions
12649 Implementations in which @code{Complex_Types.Real'Signed_Zeros} is
12650 @code{True} should attempt to provide a rational treatment of the signs
12651 of zero results and result components. For example, many of the complex
12652 elementary functions have components that are odd functions of one of
12653 the parameter components; in these cases, the result component should
12654 have the sign of the parameter component at the origin. Other complex
12655 elementary functions have zero components whose sign is opposite that of
12656 a parameter component at the origin, or is always positive or always
12661 @cindex Accuracy requirements
12662 @unnumberedsec G.2.4(19): Accuracy Requirements
12665 The versions of the forward trigonometric functions without a
12666 @code{Cycle} parameter should not be implemented by calling the
12667 corresponding version with a @code{Cycle} parameter of
12668 @code{2.0*Numerics.Pi}, since this will not provide the required
12669 accuracy in some portions of the domain. For the same reason, the
12670 version of @code{Log} without a @code{Base} parameter should not be
12671 implemented by calling the corresponding version with a @code{Base}
12672 parameter of @code{Numerics.e}.
12676 @cindex Complex arithmetic accuracy
12677 @cindex Accuracy, complex arithmetic
12678 @unnumberedsec G.2.6(15): Complex Arithmetic Accuracy
12682 The version of the @code{Compose_From_Polar} function without a
12683 @code{Cycle} parameter should not be implemented by calling the
12684 corresponding version with a @code{Cycle} parameter of
12685 @code{2.0*Numerics.Pi}, since this will not provide the required
12686 accuracy in some portions of the domain.
12690 @cindex Sequential elaboration policy
12691 @unnumberedsec H.6(15/2): Pragma Partition_Elaboration_Policy
12695 If the partition elaboration policy is @code{Sequential} and the
12696 Environment task becomes permanently blocked during elaboration then the
12697 partition is deadlocked and it is recommended that the partition be
12698 immediately terminated.
12702 @c -----------------------------------------
12703 @node Implementation Defined Characteristics
12704 @chapter Implementation Defined Characteristics
12707 In addition to the implementation dependent pragmas and attributes, and the
12708 implementation advice, there are a number of other Ada features that are
12709 potentially implementation dependent and are designated as
12710 implementation-defined. These are mentioned throughout the Ada Reference
12711 Manual, and are summarized in Annex M@.
12713 A requirement for conforming Ada compilers is that they provide
12714 documentation describing how the implementation deals with each of these
12715 issues. In this chapter you will find each point in Annex M listed,
12716 followed by a description of how GNAT
12717 handles the implementation dependence.
12719 You can use this chapter as a guide to minimizing implementation
12720 dependent features in your programs if portability to other compilers
12721 and other operating systems is an important consideration. The numbers
12722 in each entry below correspond to the paragraph numbers in the Ada
12731 Whether or not each recommendation given in Implementation
12732 Advice is followed. See 1.1.2(37).
12735 @xref{Implementation Advice}.
12742 Capacity limitations of the implementation. See 1.1.3(3).
12745 The complexity of programs that can be processed is limited only by the
12746 total amount of available virtual memory, and disk space for the
12747 generated object files.
12754 Variations from the standard that are impractical to avoid
12755 given the implementation's execution environment. See 1.1.3(6).
12758 There are no variations from the standard.
12765 Which @code{code_statement}s cause external
12766 interactions. See 1.1.3(10).
12769 Any @code{code_statement} can potentially cause external interactions.
12775 The coded representation for the text of an Ada
12776 program. See 2.1(4).
12779 See separate section on source representation.
12786 The control functions allowed in comments. See 2.1(14).
12789 See separate section on source representation.
12795 The representation for an end of line. See 2.2(2).
12798 See separate section on source representation.
12804 Maximum supported line length and lexical element
12805 length. See 2.2(15).
12808 The maximum line length is 255 characters and the maximum length of
12809 a lexical element is also 255 characters. This is the default setting
12810 if not overridden by the use of compiler switch @option{-gnaty} (which
12811 sets the maximum to 79) or @option{-gnatyMnn} which allows the maximum
12812 line length to be specified to be any value up to 32767. The maximum
12813 length of a lexical element is the same as the maximum line length.
12819 Implementation defined pragmas. See 2.8(14).
12823 @xref{Implementation Defined Pragmas}.
12829 Effect of pragma @code{Optimize}. See 2.8(27).
12832 Pragma @code{Optimize}, if given with a @code{Time} or @code{Space}
12833 parameter, checks that the optimization flag is set, and aborts if it is
12840 The sequence of characters of the value returned by
12841 @code{@var{S}'Image} when some of the graphic characters of
12842 @code{@var{S}'Wide_Image} are not defined in @code{Character}. See
12846 The sequence of characters is as defined by the wide character encoding
12847 method used for the source. See section on source representation for
12854 The predefined integer types declared in
12855 @code{Standard}. See 3.5.4(25).
12859 @item Short_Short_Integer
12861 @item Short_Integer
12862 (Short) 16 bit signed
12866 64 bit signed (on most 64 bit targets, depending on the C definition of long).
12867 32 bit signed (all other targets)
12868 @item Long_Long_Integer
12876 Any nonstandard integer types and the operators defined
12877 for them. See 3.5.4(26).
12880 There are no nonstandard integer types.
12886 Any nonstandard real types and the operators defined for
12887 them. See 3.5.6(8).
12890 There are no nonstandard real types.
12896 What combinations of requested decimal precision and range
12897 are supported for floating point types. See 3.5.7(7).
12900 The precision and range is as defined by the IEEE standard.
12906 The predefined floating point types declared in
12907 @code{Standard}. See 3.5.7(16).
12914 (Short) 32 bit IEEE short
12917 @item Long_Long_Float
12918 64 bit IEEE long (80 bit IEEE long on x86 processors)
12925 The small of an ordinary fixed point type. See 3.5.9(8).
12928 @code{Fine_Delta} is 2**(@minus{}63)
12934 What combinations of small, range, and digits are
12935 supported for fixed point types. See 3.5.9(10).
12938 Any combinations are permitted that do not result in a small less than
12939 @code{Fine_Delta} and do not result in a mantissa larger than 63 bits.
12940 If the mantissa is larger than 53 bits on machines where Long_Long_Float
12941 is 64 bits (true of all architectures except ia32), then the output from
12942 Text_IO is accurate to only 53 bits, rather than the full mantissa. This
12943 is because floating-point conversions are used to convert fixed point.
12949 The result of @code{Tags.Expanded_Name} for types declared
12950 within an unnamed @code{block_statement}. See 3.9(10).
12953 Block numbers of the form @code{B@var{nnn}}, where @var{nnn} is a
12954 decimal integer are allocated.
12960 Implementation-defined attributes. See 4.1.4(12).
12963 @xref{Implementation Defined Attributes}.
12969 Any implementation-defined time types. See 9.6(6).
12972 There are no implementation-defined time types.
12978 The time base associated with relative delays.
12981 See 9.6(20). The time base used is that provided by the C library
12982 function @code{gettimeofday}.
12988 The time base of the type @code{Calendar.Time}. See
12992 The time base used is that provided by the C library function
12993 @code{gettimeofday}.
12999 The time zone used for package @code{Calendar}
13000 operations. See 9.6(24).
13003 The time zone used by package @code{Calendar} is the current system time zone
13004 setting for local time, as accessed by the C library function
13011 Any limit on @code{delay_until_statements} of
13012 @code{select_statements}. See 9.6(29).
13015 There are no such limits.
13021 Whether or not two non-overlapping parts of a composite
13022 object are independently addressable, in the case where packing, record
13023 layout, or @code{Component_Size} is specified for the object. See
13027 Separate components are independently addressable if they do not share
13028 overlapping storage units.
13034 The representation for a compilation. See 10.1(2).
13037 A compilation is represented by a sequence of files presented to the
13038 compiler in a single invocation of the @command{gcc} command.
13044 Any restrictions on compilations that contain multiple
13045 compilation_units. See 10.1(4).
13048 No single file can contain more than one compilation unit, but any
13049 sequence of files can be presented to the compiler as a single
13056 The mechanisms for creating an environment and for adding
13057 and replacing compilation units. See 10.1.4(3).
13060 See separate section on compilation model.
13066 The manner of explicitly assigning library units to a
13067 partition. See 10.2(2).
13070 If a unit contains an Ada main program, then the Ada units for the partition
13071 are determined by recursive application of the rules in the Ada Reference
13072 Manual section 10.2(2-6). In other words, the Ada units will be those that
13073 are needed by the main program, and then this definition of need is applied
13074 recursively to those units, and the partition contains the transitive
13075 closure determined by this relationship. In short, all the necessary units
13076 are included, with no need to explicitly specify the list. If additional
13077 units are required, e.g.@: by foreign language units, then all units must be
13078 mentioned in the context clause of one of the needed Ada units.
13080 If the partition contains no main program, or if the main program is in
13081 a language other than Ada, then GNAT
13082 provides the binder options @option{-z} and @option{-n} respectively, and in
13083 this case a list of units can be explicitly supplied to the binder for
13084 inclusion in the partition (all units needed by these units will also
13085 be included automatically). For full details on the use of these
13086 options, refer to @ref{The GNAT Make Program gnatmake,,, gnat_ugn,
13087 @value{EDITION} User's Guide}.
13093 The implementation-defined means, if any, of specifying
13094 which compilation units are needed by a given compilation unit. See
13098 The units needed by a given compilation unit are as defined in
13099 the Ada Reference Manual section 10.2(2-6). There are no
13100 implementation-defined pragmas or other implementation-defined
13101 means for specifying needed units.
13107 The manner of designating the main subprogram of a
13108 partition. See 10.2(7).
13111 The main program is designated by providing the name of the
13112 corresponding @file{ALI} file as the input parameter to the binder.
13118 The order of elaboration of @code{library_items}. See
13122 The first constraint on ordering is that it meets the requirements of
13123 Chapter 10 of the Ada Reference Manual. This still leaves some
13124 implementation dependent choices, which are resolved by first
13125 elaborating bodies as early as possible (i.e., in preference to specs
13126 where there is a choice), and second by evaluating the immediate with
13127 clauses of a unit to determine the probably best choice, and
13128 third by elaborating in alphabetical order of unit names
13129 where a choice still remains.
13135 Parameter passing and function return for the main
13136 subprogram. See 10.2(21).
13139 The main program has no parameters. It may be a procedure, or a function
13140 returning an integer type. In the latter case, the returned integer
13141 value is the return code of the program (overriding any value that
13142 may have been set by a call to @code{Ada.Command_Line.Set_Exit_Status}).
13148 The mechanisms for building and running partitions. See
13152 GNAT itself supports programs with only a single partition. The GNATDIST
13153 tool provided with the GLADE package (which also includes an implementation
13154 of the PCS) provides a completely flexible method for building and running
13155 programs consisting of multiple partitions. See the separate GLADE manual
13162 The details of program execution, including program
13163 termination. See 10.2(25).
13166 See separate section on compilation model.
13172 The semantics of any non-active partitions supported by the
13173 implementation. See 10.2(28).
13176 Passive partitions are supported on targets where shared memory is
13177 provided by the operating system. See the GLADE reference manual for
13184 The information returned by @code{Exception_Message}. See
13188 Exception message returns the null string unless a specific message has
13189 been passed by the program.
13195 The result of @code{Exceptions.Exception_Name} for types
13196 declared within an unnamed @code{block_statement}. See 11.4.1(12).
13199 Blocks have implementation defined names of the form @code{B@var{nnn}}
13200 where @var{nnn} is an integer.
13206 The information returned by
13207 @code{Exception_Information}. See 11.4.1(13).
13210 @code{Exception_Information} returns a string in the following format:
13213 @emph{Exception_Name:} nnnnn
13214 @emph{Message:} mmmmm
13216 @emph{Load address:} 0xhhhh
13217 @emph{Call stack traceback locations:}
13218 0xhhhh 0xhhhh 0xhhhh ... 0xhhh
13226 @code{nnnn} is the fully qualified name of the exception in all upper
13227 case letters. This line is always present.
13230 @code{mmmm} is the message (this line present only if message is non-null)
13233 @code{ppp} is the Process Id value as a decimal integer (this line is
13234 present only if the Process Id is nonzero). Currently we are
13235 not making use of this field.
13238 The Load address line, the Call stack traceback locations line and the
13239 following values are present only if at least one traceback location was
13240 recorded. The Load address indicates the address at which the main executable
13241 was loaded; this line may not be present if operating system hasn't relocated
13242 the main executable. The values are given in C style format, with lower case
13243 letters for a-f, and only as many digits present as are necessary.
13247 The line terminator sequence at the end of each line, including
13248 the last line is a single @code{LF} character (@code{16#0A#}).
13254 Implementation-defined check names. See 11.5(27).
13257 The implementation defined check name Alignment_Check controls checking of
13258 address clause values for proper alignment (that is, the address supplied
13259 must be consistent with the alignment of the type).
13261 The implementation defined check name Predicate_Check controls whether
13262 predicate checks are generated.
13264 The implementation defined check name Validity_Check controls whether
13265 validity checks are generated.
13267 In addition, a user program can add implementation-defined check names
13268 by means of the pragma Check_Name.
13274 The interpretation of each aspect of representation. See
13278 See separate section on data representations.
13284 Any restrictions placed upon representation items. See
13288 See separate section on data representations.
13294 The meaning of @code{Size} for indefinite subtypes. See
13298 Size for an indefinite subtype is the maximum possible size, except that
13299 for the case of a subprogram parameter, the size of the parameter object
13300 is the actual size.
13306 The default external representation for a type tag. See
13310 The default external representation for a type tag is the fully expanded
13311 name of the type in upper case letters.
13317 What determines whether a compilation unit is the same in
13318 two different partitions. See 13.3(76).
13321 A compilation unit is the same in two different partitions if and only
13322 if it derives from the same source file.
13328 Implementation-defined components. See 13.5.1(15).
13331 The only implementation defined component is the tag for a tagged type,
13332 which contains a pointer to the dispatching table.
13338 If @code{Word_Size} = @code{Storage_Unit}, the default bit
13339 ordering. See 13.5.3(5).
13342 @code{Word_Size} (32) is not the same as @code{Storage_Unit} (8) for this
13343 implementation, so no non-default bit ordering is supported. The default
13344 bit ordering corresponds to the natural endianness of the target architecture.
13350 The contents of the visible part of package @code{System}
13351 and its language-defined children. See 13.7(2).
13354 See the definition of these packages in files @file{system.ads} and
13355 @file{s-stoele.ads}. Note that two declarations are added to package
13358 @smallexample @c ada
13359 Max_Priority : constant Positive :=
13361 Max_Interrupt_Priority : constant Positive :=
13362 Interrupt_Priority'Last;
13369 The contents of the visible part of package
13370 @code{System.Machine_Code}, and the meaning of
13371 @code{code_statements}. See 13.8(7).
13374 See the definition and documentation in file @file{s-maccod.ads}.
13380 The effect of unchecked conversion. See 13.9(11).
13383 Unchecked conversion between types of the same size
13384 results in an uninterpreted transmission of the bits from one type
13385 to the other. If the types are of unequal sizes, then in the case of
13386 discrete types, a shorter source is first zero or sign extended as
13387 necessary, and a shorter target is simply truncated on the left.
13388 For all non-discrete types, the source is first copied if necessary
13389 to ensure that the alignment requirements of the target are met, then
13390 a pointer is constructed to the source value, and the result is obtained
13391 by dereferencing this pointer after converting it to be a pointer to the
13392 target type. Unchecked conversions where the target subtype is an
13393 unconstrained array are not permitted. If the target alignment is
13394 greater than the source alignment, then a copy of the result is
13395 made with appropriate alignment
13401 The semantics of operations on invalid representations.
13405 For assignments and other operations where the use of invalid values cannot
13406 result in erroneous behavior, the compiler ignores the possibility of invalid
13407 values. An exception is raised at the point where an invalid value would
13408 result in erroneous behavior. For example executing:
13410 @smallexample @c ada
13411 procedure invalidvals is
13413 Y : Natural range 1 .. 10;
13414 for Y'Address use X'Address;
13415 Z : Natural range 1 .. 10;
13416 A : array (Natural range 1 .. 10) of Integer;
13418 Z := Y; -- no exception
13419 A (Z) := 3; -- exception raised;
13424 As indicated, an exception is raised on the array assignment, but not
13425 on the simple assignment of the invalid negative value from Y to Z.
13431 The manner of choosing a storage pool for an access type
13432 when @code{Storage_Pool} is not specified for the type. See 13.11(17).
13435 There are 3 different standard pools used by the compiler when
13436 @code{Storage_Pool} is not specified depending whether the type is local
13437 to a subprogram or defined at the library level and whether
13438 @code{Storage_Size}is specified or not. See documentation in the runtime
13439 library units @code{System.Pool_Global}, @code{System.Pool_Size} and
13440 @code{System.Pool_Local} in files @file{s-poosiz.ads},
13441 @file{s-pooglo.ads} and @file{s-pooloc.ads} for full details on the
13442 default pools used.
13448 Whether or not the implementation provides user-accessible
13449 names for the standard pool type(s). See 13.11(17).
13453 See documentation in the sources of the run time mentioned in the previous
13454 paragraph. All these pools are accessible by means of @code{with}'ing
13461 The meaning of @code{Storage_Size}. See 13.11(18).
13464 @code{Storage_Size} is measured in storage units, and refers to the
13465 total space available for an access type collection, or to the primary
13466 stack space for a task.
13472 Implementation-defined aspects of storage pools. See
13476 See documentation in the sources of the run time mentioned in the
13477 paragraph about standard storage pools above
13478 for details on GNAT-defined aspects of storage pools.
13484 The set of restrictions allowed in a pragma
13485 @code{Restrictions}. See 13.12(7).
13488 @xref{Standard and Implementation Defined Restrictions}.
13494 The consequences of violating limitations on
13495 @code{Restrictions} pragmas. See 13.12(9).
13498 Restrictions that can be checked at compile time result in illegalities
13499 if violated. Currently there are no other consequences of violating
13506 The representation used by the @code{Read} and
13507 @code{Write} attributes of elementary types in terms of stream
13508 elements. See 13.13.2(9).
13511 The representation is the in-memory representation of the base type of
13512 the type, using the number of bits corresponding to the
13513 @code{@var{type}'Size} value, and the natural ordering of the machine.
13519 The names and characteristics of the numeric subtypes
13520 declared in the visible part of package @code{Standard}. See A.1(3).
13523 See items describing the integer and floating-point types supported.
13529 The string returned by @code{Character_Set_Version}.
13533 @code{Ada.Wide_Characters.Handling.Character_Set_Version} returns
13534 the string "Unicode 4.0", referring to version 4.0 of the
13535 Unicode specification.
13541 The accuracy actually achieved by the elementary
13542 functions. See A.5.1(1).
13545 The elementary functions correspond to the functions available in the C
13546 library. Only fast math mode is implemented.
13552 The sign of a zero result from some of the operators or
13553 functions in @code{Numerics.Generic_Elementary_Functions}, when
13554 @code{Float_Type'Signed_Zeros} is @code{True}. See A.5.1(46).
13557 The sign of zeroes follows the requirements of the IEEE 754 standard on
13565 @code{Numerics.Float_Random.Max_Image_Width}. See A.5.2(27).
13568 Maximum image width is 6864, see library file @file{s-rannum.ads}.
13575 @code{Numerics.Discrete_Random.Max_Image_Width}. See A.5.2(27).
13578 Maximum image width is 6864, see library file @file{s-rannum.ads}.
13584 The algorithms for random number generation. See
13588 The algorithm is the Mersenne Twister, as documented in the source file
13589 @file{s-rannum.adb}. This version of the algorithm has a period of
13596 The string representation of a random number generator's
13597 state. See A.5.2(38).
13600 The value returned by the Image function is the concatenation of
13601 the fixed-width decimal representations of the 624 32-bit integers
13602 of the state vector.
13608 The minimum time interval between calls to the
13609 time-dependent Reset procedure that are guaranteed to initiate different
13610 random number sequences. See A.5.2(45).
13613 The minimum period between reset calls to guarantee distinct series of
13614 random numbers is one microsecond.
13620 The values of the @code{Model_Mantissa},
13621 @code{Model_Emin}, @code{Model_Epsilon}, @code{Model},
13622 @code{Safe_First}, and @code{Safe_Last} attributes, if the Numerics
13623 Annex is not supported. See A.5.3(72).
13626 Run the compiler with @option{-gnatS} to produce a listing of package
13627 @code{Standard}, has the values of all numeric attributes.
13633 Any implementation-defined characteristics of the
13634 input-output packages. See A.7(14).
13637 There are no special implementation defined characteristics for these
13644 The value of @code{Buffer_Size} in @code{Storage_IO}. See
13648 All type representations are contiguous, and the @code{Buffer_Size} is
13649 the value of @code{@var{type}'Size} rounded up to the next storage unit
13656 External files for standard input, standard output, and
13657 standard error See A.10(5).
13660 These files are mapped onto the files provided by the C streams
13661 libraries. See source file @file{i-cstrea.ads} for further details.
13667 The accuracy of the value produced by @code{Put}. See
13671 If more digits are requested in the output than are represented by the
13672 precision of the value, zeroes are output in the corresponding least
13673 significant digit positions.
13679 The meaning of @code{Argument_Count}, @code{Argument}, and
13680 @code{Command_Name}. See A.15(1).
13683 These are mapped onto the @code{argv} and @code{argc} parameters of the
13684 main program in the natural manner.
13690 The interpretation of the @code{Form} parameter in procedure
13691 @code{Create_Directory}. See A.16(56).
13694 The @code{Form} parameter is not used.
13700 The interpretation of the @code{Form} parameter in procedure
13701 @code{Create_Path}. See A.16(60).
13704 The @code{Form} parameter is not used.
13710 The interpretation of the @code{Form} parameter in procedure
13711 @code{Copy_File}. See A.16(68).
13714 The @code{Form} parameter is case-insensitive.
13716 Two fields are recognized in the @code{Form} parameter:
13720 @item preserve=<value>
13727 <value> starts immediately after the character '=' and ends with the
13728 character immediately preceding the next comma (',') or with the last
13729 character of the parameter.
13731 The only possible values for preserve= are:
13735 @item no_attributes
13736 Do not try to preserve any file attributes. This is the default if no
13737 preserve= is found in Form.
13739 @item all_attributes
13740 Try to preserve all file attributes (timestamps, access rights).
13743 Preserve the timestamp of the copied file, but not the other file attributes.
13748 The only possible values for mode= are:
13753 Only do the copy if the destination file does not already exist. If it already
13754 exists, Copy_File fails.
13757 Copy the file in all cases. Overwrite an already existing destination file.
13760 Append the original file to the destination file. If the destination file does
13761 not exist, the destination file is a copy of the source file. When mode=append,
13762 the field preserve=, if it exists, is not taken into account.
13767 If the Form parameter includes one or both of the fields and the value or
13768 values are incorrect, Copy_file fails with Use_Error.
13770 Examples of correct Forms:
13773 Form => "preserve=no_attributes,mode=overwrite" (the default)
13774 Form => "mode=append"
13775 Form => "mode=copy, preserve=all_attributes"
13779 Examples of incorrect Forms
13782 Form => "preserve=junk"
13783 Form => "mode=internal, preserve=timestamps"
13790 The interpretation of the @code{Pattern} parameter, when not the null string,
13791 in the @code{Start_Search} and @code{Search} procedures.
13792 See A.16(104) and A.16(112).
13795 When the @code{Pattern} parameter is not the null string, it is interpreted
13796 according to the syntax of regular expressions as defined in the
13797 @code{GNAT.Regexp} package.
13798 @xref{GNAT.Regexp (g-regexp.ads)}.
13804 Implementation-defined convention names. See B.1(11).
13807 The following convention names are supported
13812 @item Ada_Pass_By_Copy
13813 Allowed for any types except by-reference types such as limited
13814 records. Compatible with convention Ada, but causes any parameters
13815 with this convention to be passed by copy.
13816 @item Ada_Pass_By_Reference
13817 Allowed for any types except by-copy types such as scalars.
13818 Compatible with convention Ada, but causes any parameters
13819 with this convention to be passed by reference.
13823 Synonym for Assembler
13825 Synonym for Assembler
13828 @item C_Pass_By_Copy
13829 Allowed only for record types, like C, but also notes that record
13830 is to be passed by copy rather than reference.
13833 @item C_Plus_Plus (or CPP)
13836 Treated the same as C
13838 Treated the same as C
13842 For support of pragma @code{Import} with convention Intrinsic, see
13843 separate section on Intrinsic Subprograms.
13845 Stdcall (used for Windows implementations only). This convention correspond
13846 to the WINAPI (previously called Pascal convention) C/C++ convention under
13847 Windows. A routine with this convention cleans the stack before
13848 exit. This pragma cannot be applied to a dispatching call.
13850 Synonym for Stdcall
13852 Synonym for Stdcall
13854 Stubbed is a special convention used to indicate that the body of the
13855 subprogram will be entirely ignored. Any call to the subprogram
13856 is converted into a raise of the @code{Program_Error} exception. If a
13857 pragma @code{Import} specifies convention @code{stubbed} then no body need
13858 be present at all. This convention is useful during development for the
13859 inclusion of subprograms whose body has not yet been written.
13863 In addition, all otherwise unrecognized convention names are also
13864 treated as being synonymous with convention C@. In all implementations
13865 except for VMS, use of such other names results in a warning. In VMS
13866 implementations, these names are accepted silently.
13872 The meaning of link names. See B.1(36).
13875 Link names are the actual names used by the linker.
13881 The manner of choosing link names when neither the link
13882 name nor the address of an imported or exported entity is specified. See
13886 The default linker name is that which would be assigned by the relevant
13887 external language, interpreting the Ada name as being in all lower case
13894 The effect of pragma @code{Linker_Options}. See B.1(37).
13897 The string passed to @code{Linker_Options} is presented uninterpreted as
13898 an argument to the link command, unless it contains ASCII.NUL characters.
13899 NUL characters if they appear act as argument separators, so for example
13901 @smallexample @c ada
13902 pragma Linker_Options ("-labc" & ASCII.NUL & "-ldef");
13906 causes two separate arguments @code{-labc} and @code{-ldef} to be passed to the
13907 linker. The order of linker options is preserved for a given unit. The final
13908 list of options passed to the linker is in reverse order of the elaboration
13909 order. For example, linker options for a body always appear before the options
13910 from the corresponding package spec.
13916 The contents of the visible part of package
13917 @code{Interfaces} and its language-defined descendants. See B.2(1).
13920 See files with prefix @file{i-} in the distributed library.
13926 Implementation-defined children of package
13927 @code{Interfaces}. The contents of the visible part of package
13928 @code{Interfaces}. See B.2(11).
13931 See files with prefix @file{i-} in the distributed library.
13937 The types @code{Floating}, @code{Long_Floating},
13938 @code{Binary}, @code{Long_Binary}, @code{Decimal_ Element}, and
13939 @code{COBOL_Character}; and the initialization of the variables
13940 @code{Ada_To_COBOL} and @code{COBOL_To_Ada}, in
13941 @code{Interfaces.COBOL}. See B.4(50).
13947 @item Long_Floating
13948 (Floating) Long_Float
13953 @item Decimal_Element
13955 @item COBOL_Character
13960 For initialization, see the file @file{i-cobol.ads} in the distributed library.
13966 Support for access to machine instructions. See C.1(1).
13969 See documentation in file @file{s-maccod.ads} in the distributed library.
13975 Implementation-defined aspects of access to machine
13976 operations. See C.1(9).
13979 See documentation in file @file{s-maccod.ads} in the distributed library.
13985 Implementation-defined aspects of interrupts. See C.3(2).
13988 Interrupts are mapped to signals or conditions as appropriate. See
13990 @code{Ada.Interrupt_Names} in source file @file{a-intnam.ads} for details
13991 on the interrupts supported on a particular target.
13997 Implementation-defined aspects of pre-elaboration. See
14001 GNAT does not permit a partition to be restarted without reloading,
14002 except under control of the debugger.
14008 The semantics of pragma @code{Discard_Names}. See C.5(7).
14011 Pragma @code{Discard_Names} causes names of enumeration literals to
14012 be suppressed. In the presence of this pragma, the Image attribute
14013 provides the image of the Pos of the literal, and Value accepts
14020 The result of the @code{Task_Identification.Image}
14021 attribute. See C.7.1(7).
14024 The result of this attribute is a string that identifies
14025 the object or component that denotes a given task. If a variable @code{Var}
14026 has a task type, the image for this task will have the form @code{Var_@var{XXXXXXXX}},
14028 is the hexadecimal representation of the virtual address of the corresponding
14029 task control block. If the variable is an array of tasks, the image of each
14030 task will have the form of an indexed component indicating the position of a
14031 given task in the array, e.g.@: @code{Group(5)_@var{XXXXXXX}}. If the task is a
14032 component of a record, the image of the task will have the form of a selected
14033 component. These rules are fully recursive, so that the image of a task that
14034 is a subcomponent of a composite object corresponds to the expression that
14035 designates this task.
14037 If a task is created by an allocator, its image depends on the context. If the
14038 allocator is part of an object declaration, the rules described above are used
14039 to construct its image, and this image is not affected by subsequent
14040 assignments. If the allocator appears within an expression, the image
14041 includes only the name of the task type.
14043 If the configuration pragma Discard_Names is present, or if the restriction
14044 No_Implicit_Heap_Allocation is in effect, the image reduces to
14045 the numeric suffix, that is to say the hexadecimal representation of the
14046 virtual address of the control block of the task.
14051 The value of @code{Current_Task} when in a protected entry
14052 or interrupt handler. See C.7.1(17).
14055 Protected entries or interrupt handlers can be executed by any
14056 convenient thread, so the value of @code{Current_Task} is undefined.
14062 The effect of calling @code{Current_Task} from an entry
14063 body or interrupt handler. See C.7.1(19).
14066 The effect of calling @code{Current_Task} from an entry body or
14067 interrupt handler is to return the identification of the task currently
14068 executing the code.
14074 Implementation-defined aspects of
14075 @code{Task_Attributes}. See C.7.2(19).
14078 There are no implementation-defined aspects of @code{Task_Attributes}.
14084 Values of all @code{Metrics}. See D(2).
14087 The metrics information for GNAT depends on the performance of the
14088 underlying operating system. The sources of the run-time for tasking
14089 implementation, together with the output from @option{-gnatG} can be
14090 used to determine the exact sequence of operating systems calls made
14091 to implement various tasking constructs. Together with appropriate
14092 information on the performance of the underlying operating system,
14093 on the exact target in use, this information can be used to determine
14094 the required metrics.
14100 The declarations of @code{Any_Priority} and
14101 @code{Priority}. See D.1(11).
14104 See declarations in file @file{system.ads}.
14110 Implementation-defined execution resources. See D.1(15).
14113 There are no implementation-defined execution resources.
14119 Whether, on a multiprocessor, a task that is waiting for
14120 access to a protected object keeps its processor busy. See D.2.1(3).
14123 On a multi-processor, a task that is waiting for access to a protected
14124 object does not keep its processor busy.
14130 The affect of implementation defined execution resources
14131 on task dispatching. See D.2.1(9).
14134 Tasks map to threads in the threads package used by GNAT@. Where possible
14135 and appropriate, these threads correspond to native threads of the
14136 underlying operating system.
14142 Implementation-defined @code{policy_identifiers} allowed
14143 in a pragma @code{Task_Dispatching_Policy}. See D.2.2(3).
14146 There are no implementation-defined policy-identifiers allowed in this
14153 Implementation-defined aspects of priority inversion. See
14157 Execution of a task cannot be preempted by the implementation processing
14158 of delay expirations for lower priority tasks.
14164 Implementation-defined task dispatching. See D.2.2(18).
14167 The policy is the same as that of the underlying threads implementation.
14173 Implementation-defined @code{policy_identifiers} allowed
14174 in a pragma @code{Locking_Policy}. See D.3(4).
14177 The two implementation defined policies permitted in GNAT are
14178 @code{Inheritance_Locking} and @code{Conccurent_Readers_Locking}. On
14179 targets that support the @code{Inheritance_Locking} policy, locking is
14180 implemented by inheritance, i.e.@: the task owning the lock operates
14181 at a priority equal to the highest priority of any task currently
14182 requesting the lock. On targets that support the
14183 @code{Conccurent_Readers_Locking} policy, locking is implemented with a
14184 read/write lock allowing multiple propected object functions to enter
14191 Default ceiling priorities. See D.3(10).
14194 The ceiling priority of protected objects of the type
14195 @code{System.Interrupt_Priority'Last} as described in the Ada
14196 Reference Manual D.3(10),
14202 The ceiling of any protected object used internally by
14203 the implementation. See D.3(16).
14206 The ceiling priority of internal protected objects is
14207 @code{System.Priority'Last}.
14213 Implementation-defined queuing policies. See D.4(1).
14216 There are no implementation-defined queuing policies.
14222 On a multiprocessor, any conditions that cause the
14223 completion of an aborted construct to be delayed later than what is
14224 specified for a single processor. See D.6(3).
14227 The semantics for abort on a multi-processor is the same as on a single
14228 processor, there are no further delays.
14234 Any operations that implicitly require heap storage
14235 allocation. See D.7(8).
14238 The only operation that implicitly requires heap storage allocation is
14245 Implementation-defined aspects of pragma
14246 @code{Restrictions}. See D.7(20).
14249 There are no such implementation-defined aspects.
14255 Implementation-defined aspects of package
14256 @code{Real_Time}. See D.8(17).
14259 There are no implementation defined aspects of package @code{Real_Time}.
14265 Implementation-defined aspects of
14266 @code{delay_statements}. See D.9(8).
14269 Any difference greater than one microsecond will cause the task to be
14270 delayed (see D.9(7)).
14276 The upper bound on the duration of interrupt blocking
14277 caused by the implementation. See D.12(5).
14280 The upper bound is determined by the underlying operating system. In
14281 no cases is it more than 10 milliseconds.
14287 The means for creating and executing distributed
14288 programs. See E(5).
14291 The GLADE package provides a utility GNATDIST for creating and executing
14292 distributed programs. See the GLADE reference manual for further details.
14298 Any events that can result in a partition becoming
14299 inaccessible. See E.1(7).
14302 See the GLADE reference manual for full details on such events.
14308 The scheduling policies, treatment of priorities, and
14309 management of shared resources between partitions in certain cases. See
14313 See the GLADE reference manual for full details on these aspects of
14314 multi-partition execution.
14320 Events that cause the version of a compilation unit to
14321 change. See E.3(5).
14324 Editing the source file of a compilation unit, or the source files of
14325 any units on which it is dependent in a significant way cause the version
14326 to change. No other actions cause the version number to change. All changes
14327 are significant except those which affect only layout, capitalization or
14334 Whether the execution of the remote subprogram is
14335 immediately aborted as a result of cancellation. See E.4(13).
14338 See the GLADE reference manual for details on the effect of abort in
14339 a distributed application.
14345 Implementation-defined aspects of the PCS@. See E.5(25).
14348 See the GLADE reference manual for a full description of all implementation
14349 defined aspects of the PCS@.
14355 Implementation-defined interfaces in the PCS@. See
14359 See the GLADE reference manual for a full description of all
14360 implementation defined interfaces.
14366 The values of named numbers in the package
14367 @code{Decimal}. See F.2(7).
14379 @item Max_Decimal_Digits
14387 The value of @code{Max_Picture_Length} in the package
14388 @code{Text_IO.Editing}. See F.3.3(16).
14397 The value of @code{Max_Picture_Length} in the package
14398 @code{Wide_Text_IO.Editing}. See F.3.4(5).
14407 The accuracy actually achieved by the complex elementary
14408 functions and by other complex arithmetic operations. See G.1(1).
14411 Standard library functions are used for the complex arithmetic
14412 operations. Only fast math mode is currently supported.
14418 The sign of a zero result (or a component thereof) from
14419 any operator or function in @code{Numerics.Generic_Complex_Types}, when
14420 @code{Real'Signed_Zeros} is True. See G.1.1(53).
14423 The signs of zero values are as recommended by the relevant
14424 implementation advice.
14430 The sign of a zero result (or a component thereof) from
14431 any operator or function in
14432 @code{Numerics.Generic_Complex_Elementary_Functions}, when
14433 @code{Real'Signed_Zeros} is @code{True}. See G.1.2(45).
14436 The signs of zero values are as recommended by the relevant
14437 implementation advice.
14443 Whether the strict mode or the relaxed mode is the
14444 default. See G.2(2).
14447 The strict mode is the default. There is no separate relaxed mode. GNAT
14448 provides a highly efficient implementation of strict mode.
14454 The result interval in certain cases of fixed-to-float
14455 conversion. See G.2.1(10).
14458 For cases where the result interval is implementation dependent, the
14459 accuracy is that provided by performing all operations in 64-bit IEEE
14460 floating-point format.
14466 The result of a floating point arithmetic operation in
14467 overflow situations, when the @code{Machine_Overflows} attribute of the
14468 result type is @code{False}. See G.2.1(13).
14471 Infinite and NaN values are produced as dictated by the IEEE
14472 floating-point standard.
14474 Note that on machines that are not fully compliant with the IEEE
14475 floating-point standard, such as Alpha, the @option{-mieee} compiler flag
14476 must be used for achieving IEEE conforming behavior (although at the cost
14477 of a significant performance penalty), so infinite and NaN values are
14478 properly generated.
14484 The result interval for division (or exponentiation by a
14485 negative exponent), when the floating point hardware implements division
14486 as multiplication by a reciprocal. See G.2.1(16).
14489 Not relevant, division is IEEE exact.
14495 The definition of close result set, which determines the
14496 accuracy of certain fixed point multiplications and divisions. See
14500 Operations in the close result set are performed using IEEE long format
14501 floating-point arithmetic. The input operands are converted to
14502 floating-point, the operation is done in floating-point, and the result
14503 is converted to the target type.
14509 Conditions on a @code{universal_real} operand of a fixed
14510 point multiplication or division for which the result shall be in the
14511 perfect result set. See G.2.3(22).
14514 The result is only defined to be in the perfect result set if the result
14515 can be computed by a single scaling operation involving a scale factor
14516 representable in 64-bits.
14522 The result of a fixed point arithmetic operation in
14523 overflow situations, when the @code{Machine_Overflows} attribute of the
14524 result type is @code{False}. See G.2.3(27).
14527 Not relevant, @code{Machine_Overflows} is @code{True} for fixed-point
14534 The result of an elementary function reference in
14535 overflow situations, when the @code{Machine_Overflows} attribute of the
14536 result type is @code{False}. See G.2.4(4).
14539 IEEE infinite and Nan values are produced as appropriate.
14545 The value of the angle threshold, within which certain
14546 elementary functions, complex arithmetic operations, and complex
14547 elementary functions yield results conforming to a maximum relative
14548 error bound. See G.2.4(10).
14551 Information on this subject is not yet available.
14557 The accuracy of certain elementary functions for
14558 parameters beyond the angle threshold. See G.2.4(10).
14561 Information on this subject is not yet available.
14567 The result of a complex arithmetic operation or complex
14568 elementary function reference in overflow situations, when the
14569 @code{Machine_Overflows} attribute of the corresponding real type is
14570 @code{False}. See G.2.6(5).
14573 IEEE infinite and Nan values are produced as appropriate.
14579 The accuracy of certain complex arithmetic operations and
14580 certain complex elementary functions for parameters (or components
14581 thereof) beyond the angle threshold. See G.2.6(8).
14584 Information on those subjects is not yet available.
14590 Information regarding bounded errors and erroneous
14591 execution. See H.2(1).
14594 Information on this subject is not yet available.
14600 Implementation-defined aspects of pragma
14601 @code{Inspection_Point}. See H.3.2(8).
14604 Pragma @code{Inspection_Point} ensures that the variable is live and can
14605 be examined by the debugger at the inspection point.
14611 Implementation-defined aspects of pragma
14612 @code{Restrictions}. See H.4(25).
14615 There are no implementation-defined aspects of pragma @code{Restrictions}. The
14616 use of pragma @code{Restrictions [No_Exceptions]} has no effect on the
14617 generated code. Checks must suppressed by use of pragma @code{Suppress}.
14623 Any restrictions on pragma @code{Restrictions}. See
14627 There are no restrictions on pragma @code{Restrictions}.
14632 @c =======================
14633 @node Intrinsic Subprograms
14634 @chapter Intrinsic Subprograms
14635 @cindex Intrinsic Subprograms
14638 * Intrinsic Operators::
14639 * Compilation_Date::
14640 * Compilation_Time::
14641 * Enclosing_Entity::
14642 * Exception_Information::
14643 * Exception_Message::
14647 * Shifts and Rotates::
14648 * Source_Location::
14652 GNAT allows a user application program to write the declaration:
14654 @smallexample @c ada
14655 pragma Import (Intrinsic, name);
14659 providing that the name corresponds to one of the implemented intrinsic
14660 subprograms in GNAT, and that the parameter profile of the referenced
14661 subprogram meets the requirements. This chapter describes the set of
14662 implemented intrinsic subprograms, and the requirements on parameter profiles.
14663 Note that no body is supplied; as with other uses of pragma Import, the
14664 body is supplied elsewhere (in this case by the compiler itself). Note
14665 that any use of this feature is potentially non-portable, since the
14666 Ada standard does not require Ada compilers to implement this feature.
14668 @node Intrinsic Operators
14669 @section Intrinsic Operators
14670 @cindex Intrinsic operator
14673 All the predefined numeric operators in package Standard
14674 in @code{pragma Import (Intrinsic,..)}
14675 declarations. In the binary operator case, the operands must have the same
14676 size. The operand or operands must also be appropriate for
14677 the operator. For example, for addition, the operands must
14678 both be floating-point or both be fixed-point, and the
14679 right operand for @code{"**"} must have a root type of
14680 @code{Standard.Integer'Base}.
14681 You can use an intrinsic operator declaration as in the following example:
14683 @smallexample @c ada
14684 type Int1 is new Integer;
14685 type Int2 is new Integer;
14687 function "+" (X1 : Int1; X2 : Int2) return Int1;
14688 function "+" (X1 : Int1; X2 : Int2) return Int2;
14689 pragma Import (Intrinsic, "+");
14693 This declaration would permit ``mixed mode'' arithmetic on items
14694 of the differing types @code{Int1} and @code{Int2}.
14695 It is also possible to specify such operators for private types, if the
14696 full views are appropriate arithmetic types.
14698 @node Compilation_Date
14699 @section Compilation_Date
14700 @cindex Compilation_Date
14702 This intrinsic subprogram is used in the implementation of the
14703 library package @code{GNAT.Source_Info}. The only useful use of the
14704 intrinsic import in this case is the one in this unit, so an
14705 application program should simply call the function
14706 @code{GNAT.Source_Info.Compilation_Date} to obtain the date of
14707 the current compilation (in local time format MMM DD YYYY).
14709 @node Compilation_Time
14710 @section Compilation_Time
14711 @cindex Compilation_Time
14713 This intrinsic subprogram is used in the implementation of the
14714 library package @code{GNAT.Source_Info}. The only useful use of the
14715 intrinsic import in this case is the one in this unit, so an
14716 application program should simply call the function
14717 @code{GNAT.Source_Info.Compilation_Time} to obtain the time of
14718 the current compilation (in local time format HH:MM:SS).
14720 @node Enclosing_Entity
14721 @section Enclosing_Entity
14722 @cindex Enclosing_Entity
14724 This intrinsic subprogram is used in the implementation of the
14725 library package @code{GNAT.Source_Info}. The only useful use of the
14726 intrinsic import in this case is the one in this unit, so an
14727 application program should simply call the function
14728 @code{GNAT.Source_Info.Enclosing_Entity} to obtain the name of
14729 the current subprogram, package, task, entry, or protected subprogram.
14731 @node Exception_Information
14732 @section Exception_Information
14733 @cindex Exception_Information'
14735 This intrinsic subprogram is used in the implementation of the
14736 library package @code{GNAT.Current_Exception}. The only useful
14737 use of the intrinsic import in this case is the one in this unit,
14738 so an application program should simply call the function
14739 @code{GNAT.Current_Exception.Exception_Information} to obtain
14740 the exception information associated with the current exception.
14742 @node Exception_Message
14743 @section Exception_Message
14744 @cindex Exception_Message
14746 This intrinsic subprogram is used in the implementation of the
14747 library package @code{GNAT.Current_Exception}. The only useful
14748 use of the intrinsic import in this case is the one in this unit,
14749 so an application program should simply call the function
14750 @code{GNAT.Current_Exception.Exception_Message} to obtain
14751 the message associated with the current exception.
14753 @node Exception_Name
14754 @section Exception_Name
14755 @cindex Exception_Name
14757 This intrinsic subprogram is used in the implementation of the
14758 library package @code{GNAT.Current_Exception}. The only useful
14759 use of the intrinsic import in this case is the one in this unit,
14760 so an application program should simply call the function
14761 @code{GNAT.Current_Exception.Exception_Name} to obtain
14762 the name of the current exception.
14768 This intrinsic subprogram is used in the implementation of the
14769 library package @code{GNAT.Source_Info}. The only useful use of the
14770 intrinsic import in this case is the one in this unit, so an
14771 application program should simply call the function
14772 @code{GNAT.Source_Info.File} to obtain the name of the current
14779 This intrinsic subprogram is used in the implementation of the
14780 library package @code{GNAT.Source_Info}. The only useful use of the
14781 intrinsic import in this case is the one in this unit, so an
14782 application program should simply call the function
14783 @code{GNAT.Source_Info.Line} to obtain the number of the current
14786 @node Shifts and Rotates
14787 @section Shifts and Rotates
14789 @cindex Shift_Right
14790 @cindex Shift_Right_Arithmetic
14791 @cindex Rotate_Left
14792 @cindex Rotate_Right
14794 In standard Ada, the shift and rotate functions are available only
14795 for the predefined modular types in package @code{Interfaces}. However, in
14796 GNAT it is possible to define these functions for any integer
14797 type (signed or modular), as in this example:
14799 @smallexample @c ada
14800 function Shift_Left
14802 Amount : Natural) return T;
14806 The function name must be one of
14807 Shift_Left, Shift_Right, Shift_Right_Arithmetic, Rotate_Left, or
14808 Rotate_Right. T must be an integer type. T'Size must be
14809 8, 16, 32 or 64 bits; if T is modular, the modulus
14810 must be 2**8, 2**16, 2**32 or 2**64.
14811 The result type must be the same as the type of @code{Value}.
14812 The shift amount must be Natural.
14813 The formal parameter names can be anything.
14815 A more convenient way of providing these shift operators is to use
14816 the Provide_Shift_Operators pragma, which provides the function declarations
14817 and corresponding pragma Import's for all five shift functions.
14819 @node Source_Location
14820 @section Source_Location
14821 @cindex Source_Location
14823 This intrinsic subprogram is used in the implementation of the
14824 library routine @code{GNAT.Source_Info}. The only useful use of the
14825 intrinsic import in this case is the one in this unit, so an
14826 application program should simply call the function
14827 @code{GNAT.Source_Info.Source_Location} to obtain the current
14828 source file location.
14830 @node Representation Clauses and Pragmas
14831 @chapter Representation Clauses and Pragmas
14832 @cindex Representation Clauses
14835 * Alignment Clauses::
14837 * Storage_Size Clauses::
14838 * Size of Variant Record Objects::
14839 * Biased Representation ::
14840 * Value_Size and Object_Size Clauses::
14841 * Component_Size Clauses::
14842 * Bit_Order Clauses::
14843 * Effect of Bit_Order on Byte Ordering::
14844 * Pragma Pack for Arrays::
14845 * Pragma Pack for Records::
14846 * Record Representation Clauses::
14847 * Handling of Records with Holes::
14848 * Enumeration Clauses::
14849 * Address Clauses::
14850 * Use of Address Clauses for Memory-Mapped I/O::
14851 * Effect of Convention on Representation::
14852 * Conventions and Anonymous Access Types::
14853 * Determining the Representations chosen by GNAT::
14857 @cindex Representation Clause
14858 @cindex Representation Pragma
14859 @cindex Pragma, representation
14860 This section describes the representation clauses accepted by GNAT, and
14861 their effect on the representation of corresponding data objects.
14863 GNAT fully implements Annex C (Systems Programming). This means that all
14864 the implementation advice sections in chapter 13 are fully implemented.
14865 However, these sections only require a minimal level of support for
14866 representation clauses. GNAT provides much more extensive capabilities,
14867 and this section describes the additional capabilities provided.
14869 @node Alignment Clauses
14870 @section Alignment Clauses
14871 @cindex Alignment Clause
14874 GNAT requires that all alignment clauses specify a power of 2, and all
14875 default alignments are always a power of 2. The default alignment
14876 values are as follows:
14879 @item @emph{Primitive Types}.
14880 For primitive types, the alignment is the minimum of the actual size of
14881 objects of the type divided by @code{Storage_Unit},
14882 and the maximum alignment supported by the target.
14883 (This maximum alignment is given by the GNAT-specific attribute
14884 @code{Standard'Maximum_Alignment}; see @ref{Attribute Maximum_Alignment}.)
14885 @cindex @code{Maximum_Alignment} attribute
14886 For example, for type @code{Long_Float}, the object size is 8 bytes, and the
14887 default alignment will be 8 on any target that supports alignments
14888 this large, but on some targets, the maximum alignment may be smaller
14889 than 8, in which case objects of type @code{Long_Float} will be maximally
14892 @item @emph{Arrays}.
14893 For arrays, the alignment is equal to the alignment of the component type
14894 for the normal case where no packing or component size is given. If the
14895 array is packed, and the packing is effective (see separate section on
14896 packed arrays), then the alignment will be one for long packed arrays,
14897 or arrays whose length is not known at compile time. For short packed
14898 arrays, which are handled internally as modular types, the alignment
14899 will be as described for primitive types, e.g.@: a packed array of length
14900 31 bits will have an object size of four bytes, and an alignment of 4.
14902 @item @emph{Records}.
14903 For the normal non-packed case, the alignment of a record is equal to
14904 the maximum alignment of any of its components. For tagged records, this
14905 includes the implicit access type used for the tag. If a pragma @code{Pack}
14906 is used and all components are packable (see separate section on pragma
14907 @code{Pack}), then the resulting alignment is 1, unless the layout of the
14908 record makes it profitable to increase it.
14910 A special case is when:
14913 the size of the record is given explicitly, or a
14914 full record representation clause is given, and
14916 the size of the record is 2, 4, or 8 bytes.
14919 In this case, an alignment is chosen to match the
14920 size of the record. For example, if we have:
14922 @smallexample @c ada
14923 type Small is record
14926 for Small'Size use 16;
14930 then the default alignment of the record type @code{Small} is 2, not 1. This
14931 leads to more efficient code when the record is treated as a unit, and also
14932 allows the type to specified as @code{Atomic} on architectures requiring
14938 An alignment clause may specify a larger alignment than the default value
14939 up to some maximum value dependent on the target (obtainable by using the
14940 attribute reference @code{Standard'Maximum_Alignment}). It may also specify
14941 a smaller alignment than the default value for enumeration, integer and
14942 fixed point types, as well as for record types, for example
14944 @smallexample @c ada
14949 for V'alignment use 1;
14953 @cindex Alignment, default
14954 The default alignment for the type @code{V} is 4, as a result of the
14955 Integer field in the record, but it is permissible, as shown, to
14956 override the default alignment of the record with a smaller value.
14958 @cindex Alignment, subtypes
14959 Note that according to the Ada standard, an alignment clause applies only
14960 to the first named subtype. If additional subtypes are declared, then the
14961 compiler is allowed to choose any alignment it likes, and there is no way
14962 to control this choice. Consider:
14964 @smallexample @c ada
14965 type R is range 1 .. 10_000;
14966 for R'Alignment use 1;
14967 subtype RS is R range 1 .. 1000;
14971 The alignment clause specifies an alignment of 1 for the first named subtype
14972 @code{R} but this does not necessarily apply to @code{RS}. When writing
14973 portable Ada code, you should avoid writing code that explicitly or
14974 implicitly relies on the alignment of such subtypes.
14976 For the GNAT compiler, if an explicit alignment clause is given, this
14977 value is also used for any subsequent subtypes. So for GNAT, in the
14978 above example, you can count on the alignment of @code{RS} being 1. But this
14979 assumption is non-portable, and other compilers may choose different
14980 alignments for the subtype @code{RS}.
14983 @section Size Clauses
14984 @cindex Size Clause
14987 The default size for a type @code{T} is obtainable through the
14988 language-defined attribute @code{T'Size} and also through the
14989 equivalent GNAT-defined attribute @code{T'Value_Size}.
14990 For objects of type @code{T}, GNAT will generally increase the type size
14991 so that the object size (obtainable through the GNAT-defined attribute
14992 @code{T'Object_Size})
14993 is a multiple of @code{T'Alignment * Storage_Unit}.
14996 @smallexample @c ada
14997 type Smallint is range 1 .. 6;
15006 In this example, @code{Smallint'Size} = @code{Smallint'Value_Size} = 3,
15007 as specified by the RM rules,
15008 but objects of this type will have a size of 8
15009 (@code{Smallint'Object_Size} = 8),
15010 since objects by default occupy an integral number
15011 of storage units. On some targets, notably older
15012 versions of the Digital Alpha, the size of stand
15013 alone objects of this type may be 32, reflecting
15014 the inability of the hardware to do byte load/stores.
15016 Similarly, the size of type @code{Rec} is 40 bits
15017 (@code{Rec'Size} = @code{Rec'Value_Size} = 40), but
15018 the alignment is 4, so objects of this type will have
15019 their size increased to 64 bits so that it is a multiple
15020 of the alignment (in bits). This decision is
15021 in accordance with the specific Implementation Advice in RM 13.3(43):
15024 A @code{Size} clause should be supported for an object if the specified
15025 @code{Size} is at least as large as its subtype's @code{Size}, and corresponds
15026 to a size in storage elements that is a multiple of the object's
15027 @code{Alignment} (if the @code{Alignment} is nonzero).
15031 An explicit size clause may be used to override the default size by
15032 increasing it. For example, if we have:
15034 @smallexample @c ada
15035 type My_Boolean is new Boolean;
15036 for My_Boolean'Size use 32;
15040 then values of this type will always be 32 bits long. In the case of
15041 discrete types, the size can be increased up to 64 bits, with the effect
15042 that the entire specified field is used to hold the value, sign- or
15043 zero-extended as appropriate. If more than 64 bits is specified, then
15044 padding space is allocated after the value, and a warning is issued that
15045 there are unused bits.
15047 Similarly the size of records and arrays may be increased, and the effect
15048 is to add padding bits after the value. This also causes a warning message
15051 The largest Size value permitted in GNAT is 2**31@minus{}1. Since this is a
15052 Size in bits, this corresponds to an object of size 256 megabytes (minus
15053 one). This limitation is true on all targets. The reason for this
15054 limitation is that it improves the quality of the code in many cases
15055 if it is known that a Size value can be accommodated in an object of
15058 @node Storage_Size Clauses
15059 @section Storage_Size Clauses
15060 @cindex Storage_Size Clause
15063 For tasks, the @code{Storage_Size} clause specifies the amount of space
15064 to be allocated for the task stack. This cannot be extended, and if the
15065 stack is exhausted, then @code{Storage_Error} will be raised (if stack
15066 checking is enabled). Use a @code{Storage_Size} attribute definition clause,
15067 or a @code{Storage_Size} pragma in the task definition to set the
15068 appropriate required size. A useful technique is to include in every
15069 task definition a pragma of the form:
15071 @smallexample @c ada
15072 pragma Storage_Size (Default_Stack_Size);
15076 Then @code{Default_Stack_Size} can be defined in a global package, and
15077 modified as required. Any tasks requiring stack sizes different from the
15078 default can have an appropriate alternative reference in the pragma.
15080 You can also use the @option{-d} binder switch to modify the default stack
15083 For access types, the @code{Storage_Size} clause specifies the maximum
15084 space available for allocation of objects of the type. If this space is
15085 exceeded then @code{Storage_Error} will be raised by an allocation attempt.
15086 In the case where the access type is declared local to a subprogram, the
15087 use of a @code{Storage_Size} clause triggers automatic use of a special
15088 predefined storage pool (@code{System.Pool_Size}) that ensures that all
15089 space for the pool is automatically reclaimed on exit from the scope in
15090 which the type is declared.
15092 A special case recognized by the compiler is the specification of a
15093 @code{Storage_Size} of zero for an access type. This means that no
15094 items can be allocated from the pool, and this is recognized at compile
15095 time, and all the overhead normally associated with maintaining a fixed
15096 size storage pool is eliminated. Consider the following example:
15098 @smallexample @c ada
15100 type R is array (Natural) of Character;
15101 type P is access all R;
15102 for P'Storage_Size use 0;
15103 -- Above access type intended only for interfacing purposes
15107 procedure g (m : P);
15108 pragma Import (C, g);
15119 As indicated in this example, these dummy storage pools are often useful in
15120 connection with interfacing where no object will ever be allocated. If you
15121 compile the above example, you get the warning:
15124 p.adb:16:09: warning: allocation from empty storage pool
15125 p.adb:16:09: warning: Storage_Error will be raised at run time
15129 Of course in practice, there will not be any explicit allocators in the
15130 case of such an access declaration.
15132 @node Size of Variant Record Objects
15133 @section Size of Variant Record Objects
15134 @cindex Size, variant record objects
15135 @cindex Variant record objects, size
15138 In the case of variant record objects, there is a question whether Size gives
15139 information about a particular variant, or the maximum size required
15140 for any variant. Consider the following program
15142 @smallexample @c ada
15143 with Text_IO; use Text_IO;
15145 type R1 (A : Boolean := False) is record
15147 when True => X : Character;
15148 when False => null;
15156 Put_Line (Integer'Image (V1'Size));
15157 Put_Line (Integer'Image (V2'Size));
15162 Here we are dealing with a variant record, where the True variant
15163 requires 16 bits, and the False variant requires 8 bits.
15164 In the above example, both V1 and V2 contain the False variant,
15165 which is only 8 bits long. However, the result of running the
15174 The reason for the difference here is that the discriminant value of
15175 V1 is fixed, and will always be False. It is not possible to assign
15176 a True variant value to V1, therefore 8 bits is sufficient. On the
15177 other hand, in the case of V2, the initial discriminant value is
15178 False (from the default), but it is possible to assign a True
15179 variant value to V2, therefore 16 bits must be allocated for V2
15180 in the general case, even fewer bits may be needed at any particular
15181 point during the program execution.
15183 As can be seen from the output of this program, the @code{'Size}
15184 attribute applied to such an object in GNAT gives the actual allocated
15185 size of the variable, which is the largest size of any of the variants.
15186 The Ada Reference Manual is not completely clear on what choice should
15187 be made here, but the GNAT behavior seems most consistent with the
15188 language in the RM@.
15190 In some cases, it may be desirable to obtain the size of the current
15191 variant, rather than the size of the largest variant. This can be
15192 achieved in GNAT by making use of the fact that in the case of a
15193 subprogram parameter, GNAT does indeed return the size of the current
15194 variant (because a subprogram has no way of knowing how much space
15195 is actually allocated for the actual).
15197 Consider the following modified version of the above program:
15199 @smallexample @c ada
15200 with Text_IO; use Text_IO;
15202 type R1 (A : Boolean := False) is record
15204 when True => X : Character;
15205 when False => null;
15211 function Size (V : R1) return Integer is
15217 Put_Line (Integer'Image (V2'Size));
15218 Put_Line (Integer'IMage (Size (V2)));
15220 Put_Line (Integer'Image (V2'Size));
15221 Put_Line (Integer'IMage (Size (V2)));
15226 The output from this program is
15236 Here we see that while the @code{'Size} attribute always returns
15237 the maximum size, regardless of the current variant value, the
15238 @code{Size} function does indeed return the size of the current
15241 @node Biased Representation
15242 @section Biased Representation
15243 @cindex Size for biased representation
15244 @cindex Biased representation
15247 In the case of scalars with a range starting at other than zero, it is
15248 possible in some cases to specify a size smaller than the default minimum
15249 value, and in such cases, GNAT uses an unsigned biased representation,
15250 in which zero is used to represent the lower bound, and successive values
15251 represent successive values of the type.
15253 For example, suppose we have the declaration:
15255 @smallexample @c ada
15256 type Small is range -7 .. -4;
15257 for Small'Size use 2;
15261 Although the default size of type @code{Small} is 4, the @code{Size}
15262 clause is accepted by GNAT and results in the following representation
15266 -7 is represented as 2#00#
15267 -6 is represented as 2#01#
15268 -5 is represented as 2#10#
15269 -4 is represented as 2#11#
15273 Biased representation is only used if the specified @code{Size} clause
15274 cannot be accepted in any other manner. These reduced sizes that force
15275 biased representation can be used for all discrete types except for
15276 enumeration types for which a representation clause is given.
15278 @node Value_Size and Object_Size Clauses
15279 @section Value_Size and Object_Size Clauses
15281 @findex Object_Size
15282 @cindex Size, of objects
15285 In Ada 95 and Ada 2005, @code{T'Size} for a type @code{T} is the minimum
15286 number of bits required to hold values of type @code{T}.
15287 Although this interpretation was allowed in Ada 83, it was not required,
15288 and this requirement in practice can cause some significant difficulties.
15289 For example, in most Ada 83 compilers, @code{Natural'Size} was 32.
15290 However, in Ada 95 and Ada 2005,
15291 @code{Natural'Size} is
15292 typically 31. This means that code may change in behavior when moving
15293 from Ada 83 to Ada 95 or Ada 2005. For example, consider:
15295 @smallexample @c ada
15296 type Rec is record;
15302 at 0 range 0 .. Natural'Size - 1;
15303 at 0 range Natural'Size .. 2 * Natural'Size - 1;
15308 In the above code, since the typical size of @code{Natural} objects
15309 is 32 bits and @code{Natural'Size} is 31, the above code can cause
15310 unexpected inefficient packing in Ada 95 and Ada 2005, and in general
15311 there are cases where the fact that the object size can exceed the
15312 size of the type causes surprises.
15314 To help get around this problem GNAT provides two implementation
15315 defined attributes, @code{Value_Size} and @code{Object_Size}. When
15316 applied to a type, these attributes yield the size of the type
15317 (corresponding to the RM defined size attribute), and the size of
15318 objects of the type respectively.
15320 The @code{Object_Size} is used for determining the default size of
15321 objects and components. This size value can be referred to using the
15322 @code{Object_Size} attribute. The phrase ``is used'' here means that it is
15323 the basis of the determination of the size. The backend is free to
15324 pad this up if necessary for efficiency, e.g.@: an 8-bit stand-alone
15325 character might be stored in 32 bits on a machine with no efficient
15326 byte access instructions such as the Alpha.
15328 The default rules for the value of @code{Object_Size} for
15329 discrete types are as follows:
15333 The @code{Object_Size} for base subtypes reflect the natural hardware
15334 size in bits (run the compiler with @option{-gnatS} to find those values
15335 for numeric types). Enumeration types and fixed-point base subtypes have
15336 8, 16, 32 or 64 bits for this size, depending on the range of values
15340 The @code{Object_Size} of a subtype is the same as the
15341 @code{Object_Size} of
15342 the type from which it is obtained.
15345 The @code{Object_Size} of a derived base type is copied from the parent
15346 base type, and the @code{Object_Size} of a derived first subtype is copied
15347 from the parent first subtype.
15351 The @code{Value_Size} attribute
15352 is the (minimum) number of bits required to store a value
15354 This value is used to determine how tightly to pack
15355 records or arrays with components of this type, and also affects
15356 the semantics of unchecked conversion (unchecked conversions where
15357 the @code{Value_Size} values differ generate a warning, and are potentially
15360 The default rules for the value of @code{Value_Size} are as follows:
15364 The @code{Value_Size} for a base subtype is the minimum number of bits
15365 required to store all values of the type (including the sign bit
15366 only if negative values are possible).
15369 If a subtype statically matches the first subtype of a given type, then it has
15370 by default the same @code{Value_Size} as the first subtype. This is a
15371 consequence of RM 13.1(14) (``if two subtypes statically match,
15372 then their subtype-specific aspects are the same''.)
15375 All other subtypes have a @code{Value_Size} corresponding to the minimum
15376 number of bits required to store all values of the subtype. For
15377 dynamic bounds, it is assumed that the value can range down or up
15378 to the corresponding bound of the ancestor
15382 The RM defined attribute @code{Size} corresponds to the
15383 @code{Value_Size} attribute.
15385 The @code{Size} attribute may be defined for a first-named subtype. This sets
15386 the @code{Value_Size} of
15387 the first-named subtype to the given value, and the
15388 @code{Object_Size} of this first-named subtype to the given value padded up
15389 to an appropriate boundary. It is a consequence of the default rules
15390 above that this @code{Object_Size} will apply to all further subtypes. On the
15391 other hand, @code{Value_Size} is affected only for the first subtype, any
15392 dynamic subtypes obtained from it directly, and any statically matching
15393 subtypes. The @code{Value_Size} of any other static subtypes is not affected.
15395 @code{Value_Size} and
15396 @code{Object_Size} may be explicitly set for any subtype using
15397 an attribute definition clause. Note that the use of these attributes
15398 can cause the RM 13.1(14) rule to be violated. If two access types
15399 reference aliased objects whose subtypes have differing @code{Object_Size}
15400 values as a result of explicit attribute definition clauses, then it
15401 is illegal to convert from one access subtype to the other. For a more
15402 complete description of this additional legality rule, see the
15403 description of the @code{Object_Size} attribute.
15405 At the implementation level, Esize stores the Object_Size and the
15406 RM_Size field stores the @code{Value_Size} (and hence the value of the
15407 @code{Size} attribute,
15408 which, as noted above, is equivalent to @code{Value_Size}).
15410 To get a feel for the difference, consider the following examples (note
15411 that in each case the base is @code{Short_Short_Integer} with a size of 8):
15414 Object_Size Value_Size
15416 type x1 is range 0 .. 5; 8 3
15418 type x2 is range 0 .. 5;
15419 for x2'size use 12; 16 12
15421 subtype x3 is x2 range 0 .. 3; 16 2
15423 subtype x4 is x2'base range 0 .. 10; 8 4
15425 subtype x5 is x2 range 0 .. dynamic; 16 3*
15427 subtype x6 is x2'base range 0 .. dynamic; 8 3*
15432 Note: the entries marked ``3*'' are not actually specified by the Ada
15433 Reference Manual, but it seems in the spirit of the RM rules to allocate
15434 the minimum number of bits (here 3, given the range for @code{x2})
15435 known to be large enough to hold the given range of values.
15437 So far, so good, but GNAT has to obey the RM rules, so the question is
15438 under what conditions must the RM @code{Size} be used.
15439 The following is a list
15440 of the occasions on which the RM @code{Size} must be used:
15444 Component size for packed arrays or records
15447 Value of the attribute @code{Size} for a type
15450 Warning about sizes not matching for unchecked conversion
15454 For record types, the @code{Object_Size} is always a multiple of the
15455 alignment of the type (this is true for all types). In some cases the
15456 @code{Value_Size} can be smaller. Consider:
15466 On a typical 32-bit architecture, the X component will be four bytes, and
15467 require four-byte alignment, and the Y component will be one byte. In this
15468 case @code{R'Value_Size} will be 40 (bits) since this is the minimum size
15469 required to store a value of this type, and for example, it is permissible
15470 to have a component of type R in an outer array whose component size is
15471 specified to be 48 bits. However, @code{R'Object_Size} will be 64 (bits),
15472 since it must be rounded up so that this value is a multiple of the
15473 alignment (4 bytes = 32 bits).
15476 For all other types, the @code{Object_Size}
15477 and Value_Size are the same (and equivalent to the RM attribute @code{Size}).
15478 Only @code{Size} may be specified for such types.
15480 Note that @code{Value_Size} can be used to force biased representation
15481 for a particular subtype. Consider this example:
15484 type R is (A, B, C, D, E, F);
15485 subtype RAB is R range A .. B;
15486 subtype REF is R range E .. F;
15490 By default, @code{RAB}
15491 has a size of 1 (sufficient to accommodate the representation
15492 of @code{A} and @code{B}, 0 and 1), and @code{REF}
15493 has a size of 3 (sufficient to accommodate the representation
15494 of @code{E} and @code{F}, 4 and 5). But if we add the
15495 following @code{Value_Size} attribute definition clause:
15498 for REF'Value_Size use 1;
15502 then biased representation is forced for @code{REF},
15503 and 0 will represent @code{E} and 1 will represent @code{F}.
15504 A warning is issued when a @code{Value_Size} attribute
15505 definition clause forces biased representation. This
15506 warning can be turned off using @code{-gnatw.B}.
15508 @node Component_Size Clauses
15509 @section Component_Size Clauses
15510 @cindex Component_Size Clause
15513 Normally, the value specified in a component size clause must be consistent
15514 with the subtype of the array component with regard to size and alignment.
15515 In other words, the value specified must be at least equal to the size
15516 of this subtype, and must be a multiple of the alignment value.
15518 In addition, component size clauses are allowed which cause the array
15519 to be packed, by specifying a smaller value. A first case is for
15520 component size values in the range 1 through 63. The value specified
15521 must not be smaller than the Size of the subtype. GNAT will accurately
15522 honor all packing requests in this range. For example, if we have:
15524 @smallexample @c ada
15525 type r is array (1 .. 8) of Natural;
15526 for r'Component_Size use 31;
15530 then the resulting array has a length of 31 bytes (248 bits = 8 * 31).
15531 Of course access to the components of such an array is considerably
15532 less efficient than if the natural component size of 32 is used.
15533 A second case is when the subtype of the component is a record type
15534 padded because of its default alignment. For example, if we have:
15536 @smallexample @c ada
15543 type a is array (1 .. 8) of r;
15544 for a'Component_Size use 72;
15548 then the resulting array has a length of 72 bytes, instead of 96 bytes
15549 if the alignment of the record (4) was obeyed.
15551 Note that there is no point in giving both a component size clause
15552 and a pragma Pack for the same array type. if such duplicate
15553 clauses are given, the pragma Pack will be ignored.
15555 @node Bit_Order Clauses
15556 @section Bit_Order Clauses
15557 @cindex Bit_Order Clause
15558 @cindex bit ordering
15559 @cindex ordering, of bits
15562 For record subtypes, GNAT permits the specification of the @code{Bit_Order}
15563 attribute. The specification may either correspond to the default bit
15564 order for the target, in which case the specification has no effect and
15565 places no additional restrictions, or it may be for the non-standard
15566 setting (that is the opposite of the default).
15568 In the case where the non-standard value is specified, the effect is
15569 to renumber bits within each byte, but the ordering of bytes is not
15570 affected. There are certain
15571 restrictions placed on component clauses as follows:
15575 @item Components fitting within a single storage unit.
15577 These are unrestricted, and the effect is merely to renumber bits. For
15578 example if we are on a little-endian machine with @code{Low_Order_First}
15579 being the default, then the following two declarations have exactly
15582 @smallexample @c ada
15585 B : Integer range 1 .. 120;
15589 A at 0 range 0 .. 0;
15590 B at 0 range 1 .. 7;
15595 B : Integer range 1 .. 120;
15598 for R2'Bit_Order use High_Order_First;
15601 A at 0 range 7 .. 7;
15602 B at 0 range 0 .. 6;
15607 The useful application here is to write the second declaration with the
15608 @code{Bit_Order} attribute definition clause, and know that it will be treated
15609 the same, regardless of whether the target is little-endian or big-endian.
15611 @item Components occupying an integral number of bytes.
15613 These are components that exactly fit in two or more bytes. Such component
15614 declarations are allowed, but have no effect, since it is important to realize
15615 that the @code{Bit_Order} specification does not affect the ordering of bytes.
15616 In particular, the following attempt at getting an endian-independent integer
15619 @smallexample @c ada
15624 for R2'Bit_Order use High_Order_First;
15627 A at 0 range 0 .. 31;
15632 This declaration will result in a little-endian integer on a
15633 little-endian machine, and a big-endian integer on a big-endian machine.
15634 If byte flipping is required for interoperability between big- and
15635 little-endian machines, this must be explicitly programmed. This capability
15636 is not provided by @code{Bit_Order}.
15638 @item Components that are positioned across byte boundaries
15640 but do not occupy an integral number of bytes. Given that bytes are not
15641 reordered, such fields would occupy a non-contiguous sequence of bits
15642 in memory, requiring non-trivial code to reassemble. They are for this
15643 reason not permitted, and any component clause specifying such a layout
15644 will be flagged as illegal by GNAT@.
15649 Since the misconception that Bit_Order automatically deals with all
15650 endian-related incompatibilities is a common one, the specification of
15651 a component field that is an integral number of bytes will always
15652 generate a warning. This warning may be suppressed using @code{pragma
15653 Warnings (Off)} if desired. The following section contains additional
15654 details regarding the issue of byte ordering.
15656 @node Effect of Bit_Order on Byte Ordering
15657 @section Effect of Bit_Order on Byte Ordering
15658 @cindex byte ordering
15659 @cindex ordering, of bytes
15662 In this section we will review the effect of the @code{Bit_Order} attribute
15663 definition clause on byte ordering. Briefly, it has no effect at all, but
15664 a detailed example will be helpful. Before giving this
15665 example, let us review the precise
15666 definition of the effect of defining @code{Bit_Order}. The effect of a
15667 non-standard bit order is described in section 15.5.3 of the Ada
15671 2 A bit ordering is a method of interpreting the meaning of
15672 the storage place attributes.
15676 To understand the precise definition of storage place attributes in
15677 this context, we visit section 13.5.1 of the manual:
15680 13 A record_representation_clause (without the mod_clause)
15681 specifies the layout. The storage place attributes (see 13.5.2)
15682 are taken from the values of the position, first_bit, and last_bit
15683 expressions after normalizing those values so that first_bit is
15684 less than Storage_Unit.
15688 The critical point here is that storage places are taken from
15689 the values after normalization, not before. So the @code{Bit_Order}
15690 interpretation applies to normalized values. The interpretation
15691 is described in the later part of the 15.5.3 paragraph:
15694 2 A bit ordering is a method of interpreting the meaning of
15695 the storage place attributes. High_Order_First (known in the
15696 vernacular as ``big endian'') means that the first bit of a
15697 storage element (bit 0) is the most significant bit (interpreting
15698 the sequence of bits that represent a component as an unsigned
15699 integer value). Low_Order_First (known in the vernacular as
15700 ``little endian'') means the opposite: the first bit is the
15705 Note that the numbering is with respect to the bits of a storage
15706 unit. In other words, the specification affects only the numbering
15707 of bits within a single storage unit.
15709 We can make the effect clearer by giving an example.
15711 Suppose that we have an external device which presents two bytes, the first
15712 byte presented, which is the first (low addressed byte) of the two byte
15713 record is called Master, and the second byte is called Slave.
15715 The left most (most significant bit is called Control for each byte, and
15716 the remaining 7 bits are called V1, V2, @dots{} V7, where V7 is the rightmost
15717 (least significant) bit.
15719 On a big-endian machine, we can write the following representation clause
15721 @smallexample @c ada
15722 type Data is record
15723 Master_Control : Bit;
15731 Slave_Control : Bit;
15741 for Data use record
15742 Master_Control at 0 range 0 .. 0;
15743 Master_V1 at 0 range 1 .. 1;
15744 Master_V2 at 0 range 2 .. 2;
15745 Master_V3 at 0 range 3 .. 3;
15746 Master_V4 at 0 range 4 .. 4;
15747 Master_V5 at 0 range 5 .. 5;
15748 Master_V6 at 0 range 6 .. 6;
15749 Master_V7 at 0 range 7 .. 7;
15750 Slave_Control at 1 range 0 .. 0;
15751 Slave_V1 at 1 range 1 .. 1;
15752 Slave_V2 at 1 range 2 .. 2;
15753 Slave_V3 at 1 range 3 .. 3;
15754 Slave_V4 at 1 range 4 .. 4;
15755 Slave_V5 at 1 range 5 .. 5;
15756 Slave_V6 at 1 range 6 .. 6;
15757 Slave_V7 at 1 range 7 .. 7;
15762 Now if we move this to a little endian machine, then the bit ordering within
15763 the byte is backwards, so we have to rewrite the record rep clause as:
15765 @smallexample @c ada
15766 for Data use record
15767 Master_Control at 0 range 7 .. 7;
15768 Master_V1 at 0 range 6 .. 6;
15769 Master_V2 at 0 range 5 .. 5;
15770 Master_V3 at 0 range 4 .. 4;
15771 Master_V4 at 0 range 3 .. 3;
15772 Master_V5 at 0 range 2 .. 2;
15773 Master_V6 at 0 range 1 .. 1;
15774 Master_V7 at 0 range 0 .. 0;
15775 Slave_Control at 1 range 7 .. 7;
15776 Slave_V1 at 1 range 6 .. 6;
15777 Slave_V2 at 1 range 5 .. 5;
15778 Slave_V3 at 1 range 4 .. 4;
15779 Slave_V4 at 1 range 3 .. 3;
15780 Slave_V5 at 1 range 2 .. 2;
15781 Slave_V6 at 1 range 1 .. 1;
15782 Slave_V7 at 1 range 0 .. 0;
15787 It is a nuisance to have to rewrite the clause, especially if
15788 the code has to be maintained on both machines. However,
15789 this is a case that we can handle with the
15790 @code{Bit_Order} attribute if it is implemented.
15791 Note that the implementation is not required on byte addressed
15792 machines, but it is indeed implemented in GNAT.
15793 This means that we can simply use the
15794 first record clause, together with the declaration
15796 @smallexample @c ada
15797 for Data'Bit_Order use High_Order_First;
15801 and the effect is what is desired, namely the layout is exactly the same,
15802 independent of whether the code is compiled on a big-endian or little-endian
15805 The important point to understand is that byte ordering is not affected.
15806 A @code{Bit_Order} attribute definition never affects which byte a field
15807 ends up in, only where it ends up in that byte.
15808 To make this clear, let us rewrite the record rep clause of the previous
15811 @smallexample @c ada
15812 for Data'Bit_Order use High_Order_First;
15813 for Data use record
15814 Master_Control at 0 range 0 .. 0;
15815 Master_V1 at 0 range 1 .. 1;
15816 Master_V2 at 0 range 2 .. 2;
15817 Master_V3 at 0 range 3 .. 3;
15818 Master_V4 at 0 range 4 .. 4;
15819 Master_V5 at 0 range 5 .. 5;
15820 Master_V6 at 0 range 6 .. 6;
15821 Master_V7 at 0 range 7 .. 7;
15822 Slave_Control at 0 range 8 .. 8;
15823 Slave_V1 at 0 range 9 .. 9;
15824 Slave_V2 at 0 range 10 .. 10;
15825 Slave_V3 at 0 range 11 .. 11;
15826 Slave_V4 at 0 range 12 .. 12;
15827 Slave_V5 at 0 range 13 .. 13;
15828 Slave_V6 at 0 range 14 .. 14;
15829 Slave_V7 at 0 range 15 .. 15;
15834 This is exactly equivalent to saying (a repeat of the first example):
15836 @smallexample @c ada
15837 for Data'Bit_Order use High_Order_First;
15838 for Data use record
15839 Master_Control at 0 range 0 .. 0;
15840 Master_V1 at 0 range 1 .. 1;
15841 Master_V2 at 0 range 2 .. 2;
15842 Master_V3 at 0 range 3 .. 3;
15843 Master_V4 at 0 range 4 .. 4;
15844 Master_V5 at 0 range 5 .. 5;
15845 Master_V6 at 0 range 6 .. 6;
15846 Master_V7 at 0 range 7 .. 7;
15847 Slave_Control at 1 range 0 .. 0;
15848 Slave_V1 at 1 range 1 .. 1;
15849 Slave_V2 at 1 range 2 .. 2;
15850 Slave_V3 at 1 range 3 .. 3;
15851 Slave_V4 at 1 range 4 .. 4;
15852 Slave_V5 at 1 range 5 .. 5;
15853 Slave_V6 at 1 range 6 .. 6;
15854 Slave_V7 at 1 range 7 .. 7;
15859 Why are they equivalent? Well take a specific field, the @code{Slave_V2}
15860 field. The storage place attributes are obtained by normalizing the
15861 values given so that the @code{First_Bit} value is less than 8. After
15862 normalizing the values (0,10,10) we get (1,2,2) which is exactly what
15863 we specified in the other case.
15865 Now one might expect that the @code{Bit_Order} attribute might affect
15866 bit numbering within the entire record component (two bytes in this
15867 case, thus affecting which byte fields end up in), but that is not
15868 the way this feature is defined, it only affects numbering of bits,
15869 not which byte they end up in.
15871 Consequently it never makes sense to specify a starting bit number
15872 greater than 7 (for a byte addressable field) if an attribute
15873 definition for @code{Bit_Order} has been given, and indeed it
15874 may be actively confusing to specify such a value, so the compiler
15875 generates a warning for such usage.
15877 If you do need to control byte ordering then appropriate conditional
15878 values must be used. If in our example, the slave byte came first on
15879 some machines we might write:
15881 @smallexample @c ada
15882 Master_Byte_First constant Boolean := @dots{};
15884 Master_Byte : constant Natural :=
15885 1 - Boolean'Pos (Master_Byte_First);
15886 Slave_Byte : constant Natural :=
15887 Boolean'Pos (Master_Byte_First);
15889 for Data'Bit_Order use High_Order_First;
15890 for Data use record
15891 Master_Control at Master_Byte range 0 .. 0;
15892 Master_V1 at Master_Byte range 1 .. 1;
15893 Master_V2 at Master_Byte range 2 .. 2;
15894 Master_V3 at Master_Byte range 3 .. 3;
15895 Master_V4 at Master_Byte range 4 .. 4;
15896 Master_V5 at Master_Byte range 5 .. 5;
15897 Master_V6 at Master_Byte range 6 .. 6;
15898 Master_V7 at Master_Byte range 7 .. 7;
15899 Slave_Control at Slave_Byte range 0 .. 0;
15900 Slave_V1 at Slave_Byte range 1 .. 1;
15901 Slave_V2 at Slave_Byte range 2 .. 2;
15902 Slave_V3 at Slave_Byte range 3 .. 3;
15903 Slave_V4 at Slave_Byte range 4 .. 4;
15904 Slave_V5 at Slave_Byte range 5 .. 5;
15905 Slave_V6 at Slave_Byte range 6 .. 6;
15906 Slave_V7 at Slave_Byte range 7 .. 7;
15911 Now to switch between machines, all that is necessary is
15912 to set the boolean constant @code{Master_Byte_First} in
15913 an appropriate manner.
15915 @node Pragma Pack for Arrays
15916 @section Pragma Pack for Arrays
15917 @cindex Pragma Pack (for arrays)
15920 Pragma @code{Pack} applied to an array has no effect unless the component type
15921 is packable. For a component type to be packable, it must be one of the
15928 Any type whose size is specified with a size clause
15930 Any packed array type with a static size
15932 Any record type padded because of its default alignment
15936 For all these cases, if the component subtype size is in the range
15937 1 through 63, then the effect of the pragma @code{Pack} is exactly as though a
15938 component size were specified giving the component subtype size.
15939 For example if we have:
15941 @smallexample @c ada
15942 type r is range 0 .. 17;
15944 type ar is array (1 .. 8) of r;
15949 Then the component size of @code{ar} will be set to 5 (i.e.@: to @code{r'size},
15950 and the size of the array @code{ar} will be exactly 40 bits.
15952 Note that in some cases this rather fierce approach to packing can produce
15953 unexpected effects. For example, in Ada 95 and Ada 2005,
15954 subtype @code{Natural} typically has a size of 31, meaning that if you
15955 pack an array of @code{Natural}, you get 31-bit
15956 close packing, which saves a few bits, but results in far less efficient
15957 access. Since many other Ada compilers will ignore such a packing request,
15958 GNAT will generate a warning on some uses of pragma @code{Pack} that it guesses
15959 might not be what is intended. You can easily remove this warning by
15960 using an explicit @code{Component_Size} setting instead, which never generates
15961 a warning, since the intention of the programmer is clear in this case.
15963 GNAT treats packed arrays in one of two ways. If the size of the array is
15964 known at compile time and is less than 64 bits, then internally the array
15965 is represented as a single modular type, of exactly the appropriate number
15966 of bits. If the length is greater than 63 bits, or is not known at compile
15967 time, then the packed array is represented as an array of bytes, and the
15968 length is always a multiple of 8 bits.
15970 Note that to represent a packed array as a modular type, the alignment must
15971 be suitable for the modular type involved. For example, on typical machines
15972 a 32-bit packed array will be represented by a 32-bit modular integer with
15973 an alignment of four bytes. If you explicitly override the default alignment
15974 with an alignment clause that is too small, the modular representation
15975 cannot be used. For example, consider the following set of declarations:
15977 @smallexample @c ada
15978 type R is range 1 .. 3;
15979 type S is array (1 .. 31) of R;
15980 for S'Component_Size use 2;
15982 for S'Alignment use 1;
15986 If the alignment clause were not present, then a 62-bit modular
15987 representation would be chosen (typically with an alignment of 4 or 8
15988 bytes depending on the target). But the default alignment is overridden
15989 with the explicit alignment clause. This means that the modular
15990 representation cannot be used, and instead the array of bytes
15991 representation must be used, meaning that the length must be a multiple
15992 of 8. Thus the above set of declarations will result in a diagnostic
15993 rejecting the size clause and noting that the minimum size allowed is 64.
15995 @cindex Pragma Pack (for type Natural)
15996 @cindex Pragma Pack warning
15998 One special case that is worth noting occurs when the base type of the
15999 component size is 8/16/32 and the subtype is one bit less. Notably this
16000 occurs with subtype @code{Natural}. Consider:
16002 @smallexample @c ada
16003 type Arr is array (1 .. 32) of Natural;
16008 In all commonly used Ada 83 compilers, this pragma Pack would be ignored,
16009 since typically @code{Natural'Size} is 32 in Ada 83, and in any case most
16010 Ada 83 compilers did not attempt 31 bit packing.
16012 In Ada 95 and Ada 2005, @code{Natural'Size} is required to be 31. Furthermore,
16013 GNAT really does pack 31-bit subtype to 31 bits. This may result in a
16014 substantial unintended performance penalty when porting legacy Ada 83 code.
16015 To help prevent this, GNAT generates a warning in such cases. If you really
16016 want 31 bit packing in a case like this, you can set the component size
16019 @smallexample @c ada
16020 type Arr is array (1 .. 32) of Natural;
16021 for Arr'Component_Size use 31;
16025 Here 31-bit packing is achieved as required, and no warning is generated,
16026 since in this case the programmer intention is clear.
16028 @node Pragma Pack for Records
16029 @section Pragma Pack for Records
16030 @cindex Pragma Pack (for records)
16033 Pragma @code{Pack} applied to a record will pack the components to reduce
16034 wasted space from alignment gaps and by reducing the amount of space
16035 taken by components. We distinguish between @emph{packable} components and
16036 @emph{non-packable} components.
16037 Components of the following types are considered packable:
16040 Components of a primitive type are packable unless they are aliased
16041 or of an atomic type.
16044 Small packed arrays, whose size does not exceed 64 bits, and where the
16045 size is statically known at compile time, are represented internally
16046 as modular integers, and so they are also packable.
16051 All packable components occupy the exact number of bits corresponding to
16052 their @code{Size} value, and are packed with no padding bits, i.e.@: they
16053 can start on an arbitrary bit boundary.
16055 All other types are non-packable, they occupy an integral number of
16057 are placed at a boundary corresponding to their alignment requirements.
16059 For example, consider the record
16061 @smallexample @c ada
16062 type Rb1 is array (1 .. 13) of Boolean;
16065 type Rb2 is array (1 .. 65) of Boolean;
16068 type AF is new Float with Atomic;
16082 The representation for the record X2 is as follows:
16084 @smallexample @c ada
16085 for X2'Size use 224;
16087 L1 at 0 range 0 .. 0;
16088 L2 at 0 range 1 .. 64;
16089 L3 at 12 range 0 .. 31;
16090 L4 at 16 range 0 .. 0;
16091 L5 at 16 range 1 .. 13;
16092 L6 at 18 range 0 .. 71;
16097 Studying this example, we see that the packable fields @code{L1}
16099 of length equal to their sizes, and placed at specific bit boundaries (and
16100 not byte boundaries) to
16101 eliminate padding. But @code{L3} is of a non-packable float type (because
16102 it is aliased), so it is on the next appropriate alignment boundary.
16104 The next two fields are fully packable, so @code{L4} and @code{L5} are
16105 minimally packed with no gaps. However, type @code{Rb2} is a packed
16106 array that is longer than 64 bits, so it is itself non-packable. Thus
16107 the @code{L6} field is aligned to the next byte boundary, and takes an
16108 integral number of bytes, i.e.@: 72 bits.
16110 @node Record Representation Clauses
16111 @section Record Representation Clauses
16112 @cindex Record Representation Clause
16115 Record representation clauses may be given for all record types, including
16116 types obtained by record extension. Component clauses are allowed for any
16117 static component. The restrictions on component clauses depend on the type
16120 @cindex Component Clause
16121 For all components of an elementary type, the only restriction on component
16122 clauses is that the size must be at least the 'Size value of the type
16123 (actually the Value_Size). There are no restrictions due to alignment,
16124 and such components may freely cross storage boundaries.
16126 Packed arrays with a size up to and including 64 bits are represented
16127 internally using a modular type with the appropriate number of bits, and
16128 thus the same lack of restriction applies. For example, if you declare:
16130 @smallexample @c ada
16131 type R is array (1 .. 49) of Boolean;
16137 then a component clause for a component of type R may start on any
16138 specified bit boundary, and may specify a value of 49 bits or greater.
16140 For packed bit arrays that are longer than 64 bits, there are two
16141 cases. If the component size is a power of 2 (1,2,4,8,16,32 bits),
16142 including the important case of single bits or boolean values, then
16143 there are no limitations on placement of such components, and they
16144 may start and end at arbitrary bit boundaries.
16146 If the component size is not a power of 2 (e.g.@: 3 or 5), then
16147 an array of this type longer than 64 bits must always be placed on
16148 on a storage unit (byte) boundary and occupy an integral number
16149 of storage units (bytes). Any component clause that does not
16150 meet this requirement will be rejected.
16152 Any aliased component, or component of an aliased type, must
16153 have its normal alignment and size. A component clause that
16154 does not meet this requirement will be rejected.
16156 The tag field of a tagged type always occupies an address sized field at
16157 the start of the record. No component clause may attempt to overlay this
16158 tag. When a tagged type appears as a component, the tag field must have
16161 In the case of a record extension T1, of a type T, no component clause applied
16162 to the type T1 can specify a storage location that would overlap the first
16163 T'Size bytes of the record.
16165 For all other component types, including non-bit-packed arrays,
16166 the component can be placed at an arbitrary bit boundary,
16167 so for example, the following is permitted:
16169 @smallexample @c ada
16170 type R is array (1 .. 10) of Boolean;
16179 G at 0 range 0 .. 0;
16180 H at 0 range 1 .. 1;
16181 L at 0 range 2 .. 81;
16182 R at 0 range 82 .. 161;
16187 Note: the above rules apply to recent releases of GNAT 5.
16188 In GNAT 3, there are more severe restrictions on larger components.
16189 For non-primitive types, including packed arrays with a size greater than
16190 64 bits, component clauses must respect the alignment requirement of the
16191 type, in particular, always starting on a byte boundary, and the length
16192 must be a multiple of the storage unit.
16194 @node Handling of Records with Holes
16195 @section Handling of Records with Holes
16196 @cindex Handling of Records with Holes
16198 As a result of alignment considerations, records may contain "holes"
16200 which do not correspond to the data bits of any of the components.
16201 Record representation clauses can also result in holes in records.
16203 GNAT does not attempt to clear these holes, so in record objects,
16204 they should be considered to hold undefined rubbish. The generated
16205 equality routine just tests components so does not access these
16206 undefined bits, and assignment and copy operations may or may not
16207 preserve the contents of these holes (for assignments, the holes
16208 in the target will in practice contain either the bits that are
16209 present in the holes in the source, or the bits that were present
16210 in the target before the assignment).
16212 If it is necessary to ensure that holes in records have all zero
16213 bits, then record objects for which this initialization is desired
16214 should be explicitly set to all zero values using Unchecked_Conversion
16215 or address overlays. For example
16217 @smallexample @c ada
16218 type HRec is record
16225 On typical machines, integers need to be aligned on a four-byte
16226 boundary, resulting in three bytes of undefined rubbish following
16227 the 8-bit field for C. To ensure that the hole in a variable of
16228 type HRec is set to all zero bits,
16229 you could for example do:
16231 @smallexample @c ada
16232 type Base is record
16233 Dummy1, Dummy2 : Integer := 0;
16238 for RealVar'Address use BaseVar'Address;
16242 Now the 8-bytes of the value of RealVar start out containing all zero
16243 bits. A safer approach is to just define dummy fields, avoiding the
16246 @smallexample @c ada
16247 type HRec is record
16249 Dummy1 : Short_Short_Integer := 0;
16250 Dummy2 : Short_Short_Integer := 0;
16251 Dummy3 : Short_Short_Integer := 0;
16257 And to make absolutely sure that the intent of this is followed, you
16258 can use representation clauses:
16260 @smallexample @c ada
16261 for Hrec use record
16262 C at 0 range 0 .. 7;
16263 Dummy1 at 1 range 0 .. 7;
16264 Dummy2 at 2 range 0 .. 7;
16265 Dummy3 at 3 range 0 .. 7;
16266 I at 4 range 0 .. 31;
16268 for Hrec'Size use 64;
16271 @node Enumeration Clauses
16272 @section Enumeration Clauses
16274 The only restriction on enumeration clauses is that the range of values
16275 must be representable. For the signed case, if one or more of the
16276 representation values are negative, all values must be in the range:
16278 @smallexample @c ada
16279 System.Min_Int .. System.Max_Int
16283 For the unsigned case, where all values are nonnegative, the values must
16286 @smallexample @c ada
16287 0 .. System.Max_Binary_Modulus;
16291 A @emph{confirming} representation clause is one in which the values range
16292 from 0 in sequence, i.e.@: a clause that confirms the default representation
16293 for an enumeration type.
16294 Such a confirming representation
16295 is permitted by these rules, and is specially recognized by the compiler so
16296 that no extra overhead results from the use of such a clause.
16298 If an array has an index type which is an enumeration type to which an
16299 enumeration clause has been applied, then the array is stored in a compact
16300 manner. Consider the declarations:
16302 @smallexample @c ada
16303 type r is (A, B, C);
16304 for r use (A => 1, B => 5, C => 10);
16305 type t is array (r) of Character;
16309 The array type t corresponds to a vector with exactly three elements and
16310 has a default size equal to @code{3*Character'Size}. This ensures efficient
16311 use of space, but means that accesses to elements of the array will incur
16312 the overhead of converting representation values to the corresponding
16313 positional values, (i.e.@: the value delivered by the @code{Pos} attribute).
16315 @node Address Clauses
16316 @section Address Clauses
16317 @cindex Address Clause
16319 The reference manual allows a general restriction on representation clauses,
16320 as found in RM 13.1(22):
16323 An implementation need not support representation
16324 items containing nonstatic expressions, except that
16325 an implementation should support a representation item
16326 for a given entity if each nonstatic expression in the
16327 representation item is a name that statically denotes
16328 a constant declared before the entity.
16332 In practice this is applicable only to address clauses, since this is the
16333 only case in which a non-static expression is permitted by the syntax. As
16334 the AARM notes in sections 13.1 (22.a-22.h):
16337 22.a Reason: This is to avoid the following sort of thing:
16339 22.b X : Integer := F(@dots{});
16340 Y : Address := G(@dots{});
16341 for X'Address use Y;
16343 22.c In the above, we have to evaluate the
16344 initialization expression for X before we
16345 know where to put the result. This seems
16346 like an unreasonable implementation burden.
16348 22.d The above code should instead be written
16351 22.e Y : constant Address := G(@dots{});
16352 X : Integer := F(@dots{});
16353 for X'Address use Y;
16355 22.f This allows the expression ``Y'' to be safely
16356 evaluated before X is created.
16358 22.g The constant could be a formal parameter of mode in.
16360 22.h An implementation can support other nonstatic
16361 expressions if it wants to. Expressions of type
16362 Address are hardly ever static, but their value
16363 might be known at compile time anyway in many
16368 GNAT does indeed permit many additional cases of non-static expressions. In
16369 particular, if the type involved is elementary there are no restrictions
16370 (since in this case, holding a temporary copy of the initialization value,
16371 if one is present, is inexpensive). In addition, if there is no implicit or
16372 explicit initialization, then there are no restrictions. GNAT will reject
16373 only the case where all three of these conditions hold:
16378 The type of the item is non-elementary (e.g.@: a record or array).
16381 There is explicit or implicit initialization required for the object.
16382 Note that access values are always implicitly initialized.
16385 The address value is non-static. Here GNAT is more permissive than the
16386 RM, and allows the address value to be the address of a previously declared
16387 stand-alone variable, as long as it does not itself have an address clause.
16389 @smallexample @c ada
16390 Anchor : Some_Initialized_Type;
16391 Overlay : Some_Initialized_Type;
16392 for Overlay'Address use Anchor'Address;
16396 However, the prefix of the address clause cannot be an array component, or
16397 a component of a discriminated record.
16402 As noted above in section 22.h, address values are typically non-static. In
16403 particular the To_Address function, even if applied to a literal value, is
16404 a non-static function call. To avoid this minor annoyance, GNAT provides
16405 the implementation defined attribute 'To_Address. The following two
16406 expressions have identical values:
16410 @smallexample @c ada
16411 To_Address (16#1234_0000#)
16412 System'To_Address (16#1234_0000#);
16416 except that the second form is considered to be a static expression, and
16417 thus when used as an address clause value is always permitted.
16420 Additionally, GNAT treats as static an address clause that is an
16421 unchecked_conversion of a static integer value. This simplifies the porting
16422 of legacy code, and provides a portable equivalent to the GNAT attribute
16425 Another issue with address clauses is the interaction with alignment
16426 requirements. When an address clause is given for an object, the address
16427 value must be consistent with the alignment of the object (which is usually
16428 the same as the alignment of the type of the object). If an address clause
16429 is given that specifies an inappropriately aligned address value, then the
16430 program execution is erroneous.
16432 Since this source of erroneous behavior can have unfortunate effects, GNAT
16433 checks (at compile time if possible, generating a warning, or at execution
16434 time with a run-time check) that the alignment is appropriate. If the
16435 run-time check fails, then @code{Program_Error} is raised. This run-time
16436 check is suppressed if range checks are suppressed, or if the special GNAT
16437 check Alignment_Check is suppressed, or if
16438 @code{pragma Restrictions (No_Elaboration_Code)} is in effect.
16440 Finally, GNAT does not permit overlaying of objects of controlled types or
16441 composite types containing a controlled component. In most cases, the compiler
16442 can detect an attempt at such overlays and will generate a warning at compile
16443 time and a Program_Error exception at run time.
16446 An address clause cannot be given for an exported object. More
16447 understandably the real restriction is that objects with an address
16448 clause cannot be exported. This is because such variables are not
16449 defined by the Ada program, so there is no external object to export.
16452 It is permissible to give an address clause and a pragma Import for the
16453 same object. In this case, the variable is not really defined by the
16454 Ada program, so there is no external symbol to be linked. The link name
16455 and the external name are ignored in this case. The reason that we allow this
16456 combination is that it provides a useful idiom to avoid unwanted
16457 initializations on objects with address clauses.
16459 When an address clause is given for an object that has implicit or
16460 explicit initialization, then by default initialization takes place. This
16461 means that the effect of the object declaration is to overwrite the
16462 memory at the specified address. This is almost always not what the
16463 programmer wants, so GNAT will output a warning:
16473 for Ext'Address use System'To_Address (16#1234_1234#);
16475 >>> warning: implicit initialization of "Ext" may
16476 modify overlaid storage
16477 >>> warning: use pragma Import for "Ext" to suppress
16478 initialization (RM B(24))
16484 As indicated by the warning message, the solution is to use a (dummy) pragma
16485 Import to suppress this initialization. The pragma tell the compiler that the
16486 object is declared and initialized elsewhere. The following package compiles
16487 without warnings (and the initialization is suppressed):
16489 @smallexample @c ada
16497 for Ext'Address use System'To_Address (16#1234_1234#);
16498 pragma Import (Ada, Ext);
16503 A final issue with address clauses involves their use for overlaying
16504 variables, as in the following example:
16505 @cindex Overlaying of objects
16507 @smallexample @c ada
16510 for B'Address use A'Address;
16514 or alternatively, using the form recommended by the RM:
16516 @smallexample @c ada
16518 Addr : constant Address := A'Address;
16520 for B'Address use Addr;
16524 In both of these cases, @code{A}
16525 and @code{B} become aliased to one another via the
16526 address clause. This use of address clauses to overlay
16527 variables, achieving an effect similar to unchecked
16528 conversion was erroneous in Ada 83, but in Ada 95 and Ada 2005
16529 the effect is implementation defined. Furthermore, the
16530 Ada RM specifically recommends that in a situation
16531 like this, @code{B} should be subject to the following
16532 implementation advice (RM 13.3(19)):
16535 19 If the Address of an object is specified, or it is imported
16536 or exported, then the implementation should not perform
16537 optimizations based on assumptions of no aliases.
16541 GNAT follows this recommendation, and goes further by also applying
16542 this recommendation to the overlaid variable (@code{A}
16543 in the above example) in this case. This means that the overlay
16544 works "as expected", in that a modification to one of the variables
16545 will affect the value of the other.
16547 Note that when address clause overlays are used in this way, there is an
16548 issue of unintentional initialization, as shown by this example:
16550 @smallexample @c ada
16551 package Overwrite_Record is
16553 A : Character := 'C';
16554 B : Character := 'A';
16556 X : Short_Integer := 3;
16558 for Y'Address use X'Address;
16560 >>> warning: default initialization of "Y" may
16561 modify "X", use pragma Import for "Y" to
16562 suppress initialization (RM B.1(24))
16564 end Overwrite_Record;
16568 Here the default initialization of @code{Y} will clobber the value
16569 of @code{X}, which justifies the warning. The warning notes that
16570 this effect can be eliminated by adding a @code{pragma Import}
16571 which suppresses the initialization:
16573 @smallexample @c ada
16574 package Overwrite_Record is
16576 A : Character := 'C';
16577 B : Character := 'A';
16579 X : Short_Integer := 3;
16581 for Y'Address use X'Address;
16582 pragma Import (Ada, Y);
16583 end Overwrite_Record;
16587 Note that the use of @code{pragma Initialize_Scalars} may cause variables to
16588 be initialized when they would not otherwise have been in the absence
16589 of the use of this pragma. This may cause an overlay to have this
16590 unintended clobbering effect. The compiler avoids this for scalar
16591 types, but not for composite objects (where in general the effect
16592 of @code{Initialize_Scalars} is part of the initialization routine
16593 for the composite object:
16595 @smallexample @c ada
16596 pragma Initialize_Scalars;
16597 with Ada.Text_IO; use Ada.Text_IO;
16598 procedure Overwrite_Array is
16599 type Arr is array (1 .. 5) of Integer;
16600 X : Arr := (others => 1);
16602 for A'Address use X'Address;
16604 >>> warning: default initialization of "A" may
16605 modify "X", use pragma Import for "A" to
16606 suppress initialization (RM B.1(24))
16609 if X /= Arr'(others => 1) then
16610 Put_Line ("X was clobbered");
16612 Put_Line ("X was not clobbered");
16614 end Overwrite_Array;
16618 The above program generates the warning as shown, and at execution
16619 time, prints @code{X was clobbered}. If the @code{pragma Import} is
16620 added as suggested:
16622 @smallexample @c ada
16623 pragma Initialize_Scalars;
16624 with Ada.Text_IO; use Ada.Text_IO;
16625 procedure Overwrite_Array is
16626 type Arr is array (1 .. 5) of Integer;
16627 X : Arr := (others => 1);
16629 for A'Address use X'Address;
16630 pragma Import (Ada, A);
16632 if X /= Arr'(others => 1) then
16633 Put_Line ("X was clobbered");
16635 Put_Line ("X was not clobbered");
16637 end Overwrite_Array;
16641 then the program compiles without the warning and when run will generate
16642 the output @code{X was not clobbered}.
16644 @node Use of Address Clauses for Memory-Mapped I/O
16645 @section Use of Address Clauses for Memory-Mapped I/O
16646 @cindex Memory-mapped I/O
16648 A common pattern is to use an address clause to map an atomic variable to
16649 a location in memory that corresponds to a memory-mapped I/O operation or
16650 operations, for example:
16652 @smallexample @c ada
16653 type Mem_Word is record
16656 pragma Atomic (Mem_Word);
16657 for Mem_Word_Size use 32;
16660 for Mem'Address use some-address;
16668 For a full access (reference or modification) of the variable (Mem) in
16669 this case, as in the above examples, GNAT guarantees that the entire atomic
16670 word will be accessed. It is not clear whether the RM requires this. For
16671 example in the above, can the compiler reference only the Mem.A field as
16672 an optimization? Whatever the answer to this question is, GNAT makes the
16673 guarantee that for such a reference, the entire word is read or written.
16675 A problem arises with a component access such as:
16677 @smallexample @c ada
16682 Note that the component A is not declared as atomic. This means that it is
16683 not clear what this assignment means. It could correspond to full word read
16684 and write as given in the first example, or on architectures that supported
16685 such an operation it might be a single byte store instruction. The RM does
16686 not have anything to say in this situation, and GNAT does not make any
16687 guarantee. The code generated may vary from target to target. GNAT will issue
16688 a warning in such a case:
16690 @smallexample @c ada
16693 >>> warning: access to non-atomic component of atomic array,
16694 may cause unexpected accesses to atomic object
16698 It is best to be explicit in this situation, by either declaring the
16699 components to be atomic if you want the byte store, or explicitly writing
16700 the full word access sequence if that is what the hardware requires.
16702 @node Effect of Convention on Representation
16703 @section Effect of Convention on Representation
16704 @cindex Convention, effect on representation
16707 Normally the specification of a foreign language convention for a type or
16708 an object has no effect on the chosen representation. In particular, the
16709 representation chosen for data in GNAT generally meets the standard system
16710 conventions, and for example records are laid out in a manner that is
16711 consistent with C@. This means that specifying convention C (for example)
16714 There are four exceptions to this general rule:
16718 @item Convention Fortran and array subtypes
16719 If pragma Convention Fortran is specified for an array subtype, then in
16720 accordance with the implementation advice in section 3.6.2(11) of the
16721 Ada Reference Manual, the array will be stored in a Fortran-compatible
16722 column-major manner, instead of the normal default row-major order.
16724 @item Convention C and enumeration types
16725 GNAT normally stores enumeration types in 8, 16, or 32 bits as required
16726 to accommodate all values of the type. For example, for the enumeration
16729 @smallexample @c ada
16730 type Color is (Red, Green, Blue);
16734 8 bits is sufficient to store all values of the type, so by default, objects
16735 of type @code{Color} will be represented using 8 bits. However, normal C
16736 convention is to use 32 bits for all enum values in C, since enum values
16737 are essentially of type int. If pragma @code{Convention C} is specified for an
16738 Ada enumeration type, then the size is modified as necessary (usually to
16739 32 bits) to be consistent with the C convention for enum values.
16741 Note that this treatment applies only to types. If Convention C is given for
16742 an enumeration object, where the enumeration type is not Convention C, then
16743 Object_Size bits are allocated. For example, for a normal enumeration type,
16744 with less than 256 elements, only 8 bits will be allocated for the object.
16745 Since this may be a surprise in terms of what C expects, GNAT will issue a
16746 warning in this situation. The warning can be suppressed by giving an explicit
16747 size clause specifying the desired size.
16749 @item Convention C/Fortran and Boolean types
16750 In C, the usual convention for boolean values, that is values used for
16751 conditions, is that zero represents false, and nonzero values represent
16752 true. In Ada, the normal convention is that two specific values, typically
16753 0/1, are used to represent false/true respectively.
16755 Fortran has a similar convention for @code{LOGICAL} values (any nonzero
16756 value represents true).
16758 To accommodate the Fortran and C conventions, if a pragma Convention specifies
16759 C or Fortran convention for a derived Boolean, as in the following example:
16761 @smallexample @c ada
16762 type C_Switch is new Boolean;
16763 pragma Convention (C, C_Switch);
16767 then the GNAT generated code will treat any nonzero value as true. For truth
16768 values generated by GNAT, the conventional value 1 will be used for True, but
16769 when one of these values is read, any nonzero value is treated as True.
16771 @item Access types on OpenVMS
16772 For 64-bit OpenVMS systems, access types (other than those for unconstrained
16773 arrays) are 64-bits long. An exception to this rule is for the case of
16774 C-convention access types where there is no explicit size clause present (or
16775 inherited for derived types). In this case, GNAT chooses to make these
16776 pointers 32-bits, which provides an easier path for migration of 32-bit legacy
16777 code. size clause specifying 64-bits must be used to obtain a 64-bit pointer.
16781 @node Conventions and Anonymous Access Types
16782 @section Conventions and Anonymous Access Types
16783 @cindex Anonymous access types
16784 @cindex Convention for anonymous access types
16786 The RM is not entirely clear on convention handling in a number of cases,
16787 and in particular, it is not clear on the convention to be given to
16788 anonymous access types in general, and in particular what is to be
16789 done for the case of anonymous access-to-subprogram.
16791 In GNAT, we decide that if an explicit Convention is applied
16792 to an object or component, and its type is such an anonymous type,
16793 then the convention will apply to this anonymous type as well. This
16794 seems to make sense since it is anomolous in any case to have a
16795 different convention for an object and its type, and there is clearly
16796 no way to explicitly specify a convention for an anonymous type, since
16797 it doesn't have a name to specify!
16799 Furthermore, we decide that if a convention is applied to a record type,
16800 then this convention is inherited by any of its components that are of an
16801 anonymous access type which do not have an explicitly specified convention.
16803 The following program shows these conventions in action:
16805 @smallexample @c ada
16806 package ConvComp is
16807 type Foo is range 1 .. 10;
16809 A : access function (X : Foo) return Integer;
16812 pragma Convention (C, T1);
16815 A : access function (X : Foo) return Integer;
16816 pragma Convention (C, A);
16819 pragma Convention (COBOL, T2);
16822 A : access function (X : Foo) return Integer;
16823 pragma Convention (COBOL, A);
16826 pragma Convention (C, T3);
16829 A : access function (X : Foo) return Integer;
16832 pragma Convention (COBOL, T4);
16834 function F (X : Foo) return Integer;
16835 pragma Convention (C, F);
16837 function F (X : Foo) return Integer is (13);
16839 TV1 : T1 := (F'Access, 12); -- OK
16840 TV2 : T2 := (F'Access, 13); -- OK
16842 TV3 : T3 := (F'Access, 13); -- ERROR
16844 >>> subprogram "F" has wrong convention
16845 >>> does not match access to subprogram declared at line 17
16846 38. TV4 : T4 := (F'Access, 13); -- ERROR
16848 >>> subprogram "F" has wrong convention
16849 >>> does not match access to subprogram declared at line 24
16853 @node Determining the Representations chosen by GNAT
16854 @section Determining the Representations chosen by GNAT
16855 @cindex Representation, determination of
16856 @cindex @option{-gnatR} switch
16859 Although the descriptions in this section are intended to be complete, it is
16860 often easier to simply experiment to see what GNAT accepts and what the
16861 effect is on the layout of types and objects.
16863 As required by the Ada RM, if a representation clause is not accepted, then
16864 it must be rejected as illegal by the compiler. However, when a
16865 representation clause or pragma is accepted, there can still be questions
16866 of what the compiler actually does. For example, if a partial record
16867 representation clause specifies the location of some components and not
16868 others, then where are the non-specified components placed? Or if pragma
16869 @code{Pack} is used on a record, then exactly where are the resulting
16870 fields placed? The section on pragma @code{Pack} in this chapter can be
16871 used to answer the second question, but it is often easier to just see
16872 what the compiler does.
16874 For this purpose, GNAT provides the option @option{-gnatR}. If you compile
16875 with this option, then the compiler will output information on the actual
16876 representations chosen, in a format similar to source representation
16877 clauses. For example, if we compile the package:
16879 @smallexample @c ada
16881 type r (x : boolean) is tagged record
16883 when True => S : String (1 .. 100);
16884 when False => null;
16888 type r2 is new r (false) with record
16893 y2 at 16 range 0 .. 31;
16900 type x1 is array (1 .. 10) of x;
16901 for x1'component_size use 11;
16903 type ia is access integer;
16905 type Rb1 is array (1 .. 13) of Boolean;
16908 type Rb2 is array (1 .. 65) of Boolean;
16924 using the switch @option{-gnatR} we obtain the following output:
16927 Representation information for unit q
16928 -------------------------------------
16931 for r'Alignment use 4;
16933 x at 4 range 0 .. 7;
16934 _tag at 0 range 0 .. 31;
16935 s at 5 range 0 .. 799;
16938 for r2'Size use 160;
16939 for r2'Alignment use 4;
16941 x at 4 range 0 .. 7;
16942 _tag at 0 range 0 .. 31;
16943 _parent at 0 range 0 .. 63;
16944 y2 at 16 range 0 .. 31;
16948 for x'Alignment use 1;
16950 y at 0 range 0 .. 7;
16953 for x1'Size use 112;
16954 for x1'Alignment use 1;
16955 for x1'Component_Size use 11;
16957 for rb1'Size use 13;
16958 for rb1'Alignment use 2;
16959 for rb1'Component_Size use 1;
16961 for rb2'Size use 72;
16962 for rb2'Alignment use 1;
16963 for rb2'Component_Size use 1;
16965 for x2'Size use 224;
16966 for x2'Alignment use 4;
16968 l1 at 0 range 0 .. 0;
16969 l2 at 0 range 1 .. 64;
16970 l3 at 12 range 0 .. 31;
16971 l4 at 16 range 0 .. 0;
16972 l5 at 16 range 1 .. 13;
16973 l6 at 18 range 0 .. 71;
16978 The Size values are actually the Object_Size, i.e.@: the default size that
16979 will be allocated for objects of the type.
16980 The ?? size for type r indicates that we have a variant record, and the
16981 actual size of objects will depend on the discriminant value.
16983 The Alignment values show the actual alignment chosen by the compiler
16984 for each record or array type.
16986 The record representation clause for type r shows where all fields
16987 are placed, including the compiler generated tag field (whose location
16988 cannot be controlled by the programmer).
16990 The record representation clause for the type extension r2 shows all the
16991 fields present, including the parent field, which is a copy of the fields
16992 of the parent type of r2, i.e.@: r1.
16994 The component size and size clauses for types rb1 and rb2 show
16995 the exact effect of pragma @code{Pack} on these arrays, and the record
16996 representation clause for type x2 shows how pragma @code{Pack} affects
16999 In some cases, it may be useful to cut and paste the representation clauses
17000 generated by the compiler into the original source to fix and guarantee
17001 the actual representation to be used.
17003 @node Standard Library Routines
17004 @chapter Standard Library Routines
17007 The Ada Reference Manual contains in Annex A a full description of an
17008 extensive set of standard library routines that can be used in any Ada
17009 program, and which must be provided by all Ada compilers. They are
17010 analogous to the standard C library used by C programs.
17012 GNAT implements all of the facilities described in annex A, and for most
17013 purposes the description in the Ada Reference Manual, or appropriate Ada
17014 text book, will be sufficient for making use of these facilities.
17016 In the case of the input-output facilities,
17017 @xref{The Implementation of Standard I/O},
17018 gives details on exactly how GNAT interfaces to the
17019 file system. For the remaining packages, the Ada Reference Manual
17020 should be sufficient. The following is a list of the packages included,
17021 together with a brief description of the functionality that is provided.
17023 For completeness, references are included to other predefined library
17024 routines defined in other sections of the Ada Reference Manual (these are
17025 cross-indexed from Annex A). For further details see the relevant
17026 package declarations in the run-time library. In particular, a few units
17027 are not implemented, as marked by the presence of pragma Unimplemented_Unit,
17028 and in this case the package declaration contains comments explaining why
17029 the unit is not implemented.
17033 This is a parent package for all the standard library packages. It is
17034 usually included implicitly in your program, and itself contains no
17035 useful data or routines.
17037 @item Ada.Assertions (11.4.2)
17038 @code{Assertions} provides the @code{Assert} subprograms, and also
17039 the declaration of the @code{Assertion_Error} exception.
17041 @item Ada.Asynchronous_Task_Control (D.11)
17042 @code{Asynchronous_Task_Control} provides low level facilities for task
17043 synchronization. It is typically not implemented. See package spec for details.
17045 @item Ada.Calendar (9.6)
17046 @code{Calendar} provides time of day access, and routines for
17047 manipulating times and durations.
17049 @item Ada.Calendar.Arithmetic (9.6.1)
17050 This package provides additional arithmetic
17051 operations for @code{Calendar}.
17053 @item Ada.Calendar.Formatting (9.6.1)
17054 This package provides formatting operations for @code{Calendar}.
17056 @item Ada.Calendar.Time_Zones (9.6.1)
17057 This package provides additional @code{Calendar} facilities
17058 for handling time zones.
17060 @item Ada.Characters (A.3.1)
17061 This is a dummy parent package that contains no useful entities
17063 @item Ada.Characters.Conversions (A.3.2)
17064 This package provides character conversion functions.
17066 @item Ada.Characters.Handling (A.3.2)
17067 This package provides some basic character handling capabilities,
17068 including classification functions for classes of characters (e.g.@: test
17069 for letters, or digits).
17071 @item Ada.Characters.Latin_1 (A.3.3)
17072 This package includes a complete set of definitions of the characters
17073 that appear in type CHARACTER@. It is useful for writing programs that
17074 will run in international environments. For example, if you want an
17075 upper case E with an acute accent in a string, it is often better to use
17076 the definition of @code{UC_E_Acute} in this package. Then your program
17077 will print in an understandable manner even if your environment does not
17078 support these extended characters.
17080 @item Ada.Command_Line (A.15)
17081 This package provides access to the command line parameters and the name
17082 of the current program (analogous to the use of @code{argc} and @code{argv}
17083 in C), and also allows the exit status for the program to be set in a
17084 system-independent manner.
17086 @item Ada.Complex_Text_IO (G.1.3)
17087 This package provides text input and output of complex numbers.
17089 @item Ada.Containers (A.18.1)
17090 A top level package providing a few basic definitions used by all the
17091 following specific child packages that provide specific kinds of
17094 @item Ada.Containers.Bounded_Priority_Queues (A.18.31)
17096 @item Ada.Containers.Bounded_Synchronized_Queues (A.18.29)
17098 @item Ada.Containers.Doubly_Linked_Lists (A.18.3)
17100 @item Ada.Containers.Generic_Array_Sort (A.18.26)
17102 @item Ada.Containers.Generic_Constrained_Array_Sort (A.18.26)
17104 @item Ada.Containers.Generic_Sort (A.18.26)
17106 @item Ada.Containers.Hashed_Maps (A.18.5)
17108 @item Ada.Containers.Hashed_Sets (A.18.8)
17110 @item Ada.Containers.Indefinite_Doubly_Linked_Lists (A.18.12)
17112 @item Ada.Containers.Indefinite_Hashed_Maps (A.18.13)
17114 @item Ada.Containers.Indefinite_Hashed_Sets (A.18.15)
17116 @item Ada.Containers.Indefinite_Holders (A.18.18)
17118 @item Ada.Containers.Indefinite_Multiway_Trees (A.18.17)
17120 @item Ada.Containers.Indefinite_Ordered_Maps (A.18.14)
17122 @item Ada.Containers.Indefinite_Ordered_Sets (A.18.16)
17124 @item Ada.Containers.Indefinite_Vectors (A.18.11)
17126 @item Ada.Containers.Multiway_Trees (A.18.10)
17128 @item Ada.Containers.Ordered_Maps (A.18.6)
17130 @item Ada.Containers.Ordered_Sets (A.18.9)
17132 @item Ada.Containers.Synchronized_Queue_Interfaces (A.18.27)
17134 @item Ada.Containers.Unbounded_Priority_Queues (A.18.30)
17136 @item Ada.Containers.Unbounded_Synchronized_Queues (A.18.28)
17138 @item Ada.Containers.Vectors (A.18.2)
17140 @item Ada.Directories (A.16)
17141 This package provides operations on directories.
17143 @item Ada.Directories.Hierarchical_File_Names (A.16.1)
17144 This package provides additional directory operations handling
17145 hiearchical file names.
17147 @item Ada.Directories.Information (A.16)
17148 This is an implementation defined package for additional directory
17149 operations, which is not implemented in GNAT.
17151 @item Ada.Decimal (F.2)
17152 This package provides constants describing the range of decimal numbers
17153 implemented, and also a decimal divide routine (analogous to the COBOL
17154 verb DIVIDE @dots{} GIVING @dots{} REMAINDER @dots{})
17156 @item Ada.Direct_IO (A.8.4)
17157 This package provides input-output using a model of a set of records of
17158 fixed-length, containing an arbitrary definite Ada type, indexed by an
17159 integer record number.
17161 @item Ada.Dispatching (D.2.1)
17162 A parent package containing definitions for task dispatching operations.
17164 @item Ada.Dispatching.EDF (D.2.6)
17165 Not implemented in GNAT.
17167 @item Ada.Dispatching.Non_Preemptive (D.2.4)
17168 Not implemented in GNAT.
17170 @item Ada.Dispatching.Round_Robin (D.2.5)
17171 Not implemented in GNAT.
17173 @item Ada.Dynamic_Priorities (D.5)
17174 This package allows the priorities of a task to be adjusted dynamically
17175 as the task is running.
17177 @item Ada.Environment_Variables (A.17)
17178 This package provides facilities for accessing environment variables.
17180 @item Ada.Exceptions (11.4.1)
17181 This package provides additional information on exceptions, and also
17182 contains facilities for treating exceptions as data objects, and raising
17183 exceptions with associated messages.
17185 @item Ada.Execution_Time (D.14)
17186 Not implemented in GNAT.
17188 @item Ada.Execution_Time.Group_Budgets (D.14.2)
17189 Not implemented in GNAT.
17191 @item Ada.Execution_Time.Timers (D.14.1)'
17192 Not implemented in GNAT.
17194 @item Ada.Finalization (7.6)
17195 This package contains the declarations and subprograms to support the
17196 use of controlled types, providing for automatic initialization and
17197 finalization (analogous to the constructors and destructors of C++).
17199 @item Ada.Float_Text_IO (A.10.9)
17200 A library level instantiation of Text_IO.Float_IO for type Float.
17202 @item Ada.Float_Wide_Text_IO (A.10.9)
17203 A library level instantiation of Wide_Text_IO.Float_IO for type Float.
17205 @item Ada.Float_Wide_Wide_Text_IO (A.10.9)
17206 A library level instantiation of Wide_Wide_Text_IO.Float_IO for type Float.
17208 @item Ada.Integer_Text_IO (A.10.9)
17209 A library level instantiation of Text_IO.Integer_IO for type Integer.
17211 @item Ada.Integer_Wide_Text_IO (A.10.9)
17212 A library level instantiation of Wide_Text_IO.Integer_IO for type Integer.
17214 @item Ada.Integer_Wide_Wide_Text_IO (A.10.9)
17215 A library level instantiation of Wide_Wide_Text_IO.Integer_IO for type Integer.
17217 @item Ada.Interrupts (C.3.2)
17218 This package provides facilities for interfacing to interrupts, which
17219 includes the set of signals or conditions that can be raised and
17220 recognized as interrupts.
17222 @item Ada.Interrupts.Names (C.3.2)
17223 This package provides the set of interrupt names (actually signal
17224 or condition names) that can be handled by GNAT@.
17226 @item Ada.IO_Exceptions (A.13)
17227 This package defines the set of exceptions that can be raised by use of
17228 the standard IO packages.
17230 @item Ada.Iterator_Interfaces (5.5.1)
17231 This package provides a generic interface to generalized iterators.
17233 @item Ada.Locales (A.19)
17234 This package provides declarations providing information (Language
17235 and Country) about the current locale.
17238 This package contains some standard constants and exceptions used
17239 throughout the numerics packages. Note that the constants pi and e are
17240 defined here, and it is better to use these definitions than rolling
17243 @item Ada.Numerics.Complex_Arrays (G.3.2)
17244 Provides operations on arrays of complex numbers.
17246 @item Ada.Numerics.Complex_Elementary_Functions
17247 Provides the implementation of standard elementary functions (such as
17248 log and trigonometric functions) operating on complex numbers using the
17249 standard @code{Float} and the @code{Complex} and @code{Imaginary} types
17250 created by the package @code{Numerics.Complex_Types}.
17252 @item Ada.Numerics.Complex_Types
17253 This is a predefined instantiation of
17254 @code{Numerics.Generic_Complex_Types} using @code{Standard.Float} to
17255 build the type @code{Complex} and @code{Imaginary}.
17257 @item Ada.Numerics.Discrete_Random
17258 This generic package provides a random number generator suitable for generating
17259 uniformly distributed values of a specified discrete subtype.
17261 @item Ada.Numerics.Float_Random
17262 This package provides a random number generator suitable for generating
17263 uniformly distributed floating point values in the unit interval.
17265 @item Ada.Numerics.Generic_Complex_Elementary_Functions
17266 This is a generic version of the package that provides the
17267 implementation of standard elementary functions (such as log and
17268 trigonometric functions) for an arbitrary complex type.
17270 The following predefined instantiations of this package are provided:
17274 @code{Ada.Numerics.Short_Complex_Elementary_Functions}
17276 @code{Ada.Numerics.Complex_Elementary_Functions}
17278 @code{Ada.Numerics.Long_Complex_Elementary_Functions}
17281 @item Ada.Numerics.Generic_Complex_Types
17282 This is a generic package that allows the creation of complex types,
17283 with associated complex arithmetic operations.
17285 The following predefined instantiations of this package exist
17288 @code{Ada.Numerics.Short_Complex_Complex_Types}
17290 @code{Ada.Numerics.Complex_Complex_Types}
17292 @code{Ada.Numerics.Long_Complex_Complex_Types}
17295 @item Ada.Numerics.Generic_Elementary_Functions
17296 This is a generic package that provides the implementation of standard
17297 elementary functions (such as log an trigonometric functions) for an
17298 arbitrary float type.
17300 The following predefined instantiations of this package exist
17304 @code{Ada.Numerics.Short_Elementary_Functions}
17306 @code{Ada.Numerics.Elementary_Functions}
17308 @code{Ada.Numerics.Long_Elementary_Functions}
17311 @item Ada.Numerics.Generic_Real_Arrays (G.3.1)
17312 Generic operations on arrays of reals
17314 @item Ada.Numerics.Real_Arrays (G.3.1)
17315 Preinstantiation of Ada.Numerics.Generic_Real_Arrays (Float).
17317 @item Ada.Real_Time (D.8)
17318 This package provides facilities similar to those of @code{Calendar}, but
17319 operating with a finer clock suitable for real time control. Note that
17320 annex D requires that there be no backward clock jumps, and GNAT generally
17321 guarantees this behavior, but of course if the external clock on which
17322 the GNAT runtime depends is deliberately reset by some external event,
17323 then such a backward jump may occur.
17325 @item Ada.Real_Time.Timing_Events (D.15)
17326 Not implemented in GNAT.
17328 @item Ada.Sequential_IO (A.8.1)
17329 This package provides input-output facilities for sequential files,
17330 which can contain a sequence of values of a single type, which can be
17331 any Ada type, including indefinite (unconstrained) types.
17333 @item Ada.Storage_IO (A.9)
17334 This package provides a facility for mapping arbitrary Ada types to and
17335 from a storage buffer. It is primarily intended for the creation of new
17338 @item Ada.Streams (13.13.1)
17339 This is a generic package that provides the basic support for the
17340 concept of streams as used by the stream attributes (@code{Input},
17341 @code{Output}, @code{Read} and @code{Write}).
17343 @item Ada.Streams.Stream_IO (A.12.1)
17344 This package is a specialization of the type @code{Streams} defined in
17345 package @code{Streams} together with a set of operations providing
17346 Stream_IO capability. The Stream_IO model permits both random and
17347 sequential access to a file which can contain an arbitrary set of values
17348 of one or more Ada types.
17350 @item Ada.Strings (A.4.1)
17351 This package provides some basic constants used by the string handling
17354 @item Ada.Strings.Bounded (A.4.4)
17355 This package provides facilities for handling variable length
17356 strings. The bounded model requires a maximum length. It is thus
17357 somewhat more limited than the unbounded model, but avoids the use of
17358 dynamic allocation or finalization.
17360 @item Ada.Strings.Bounded.Equal_Case_Insensitive (A.4.10)
17361 Provides case-insensitive comparisons of bounded strings
17363 @item Ada.Strings.Bounded.Hash (A.4.9)
17364 This package provides a generic hash function for bounded strings
17366 @item Ada.Strings.Bounded.Hash_Case_Insensitive (A.4.9)
17367 This package provides a generic hash function for bounded strings that
17368 converts the string to be hashed to lower case.
17370 @item Ada.Strings.Bounded.Less_Case_Insensitive (A.4.10)
17371 This package provides a comparison function for bounded strings that works
17372 in a case insensitive manner by converting to lower case before the comparison.
17374 @item Ada.Strings.Fixed (A.4.3)
17375 This package provides facilities for handling fixed length strings.
17377 @item Ada.Strings.Fixed.Equal_Case_Insensitive (A.4.10)
17378 This package provides an equality function for fixed strings that compares
17379 the strings after converting both to lower case.
17381 @item Ada.Strings.Fixed.Hash_Case_Insensitive (A.4.9)
17382 This package provides a case insensitive hash function for fixed strings that
17383 converts the string to lower case before computing the hash.
17385 @item Ada.Strings.Fixed.Less_Case_Insensitive (A.4.10)
17386 This package provides a comparison function for fixed strings that works
17387 in a case insensitive manner by converting to lower case before the comparison.
17389 Ada.Strings.Hash (A.4.9)
17390 This package provides a hash function for strings.
17392 Ada.Strings.Hash_Case_Insensitive (A.4.9)
17393 This package provides a hash function for strings that is case insensitive.
17394 The string is converted to lower case before computing the hash.
17396 @item Ada.Strings.Less_Case_Insensitive (A.4.10)
17397 This package provides a comparison function for\strings that works
17398 in a case insensitive manner by converting to lower case before the comparison.
17400 @item Ada.Strings.Maps (A.4.2)
17401 This package provides facilities for handling character mappings and
17402 arbitrarily defined subsets of characters. For instance it is useful in
17403 defining specialized translation tables.
17405 @item Ada.Strings.Maps.Constants (A.4.6)
17406 This package provides a standard set of predefined mappings and
17407 predefined character sets. For example, the standard upper to lower case
17408 conversion table is found in this package. Note that upper to lower case
17409 conversion is non-trivial if you want to take the entire set of
17410 characters, including extended characters like E with an acute accent,
17411 into account. You should use the mappings in this package (rather than
17412 adding 32 yourself) to do case mappings.
17414 @item Ada.Strings.Unbounded (A.4.5)
17415 This package provides facilities for handling variable length
17416 strings. The unbounded model allows arbitrary length strings, but
17417 requires the use of dynamic allocation and finalization.
17419 @item Ada.Strings.Unbounded.Equal_Case_Insensitive (A.4.10)
17420 Provides case-insensitive comparisons of unbounded strings
17422 @item Ada.Strings.Unbounded.Hash (A.4.9)
17423 This package provides a generic hash function for unbounded strings
17425 @item Ada.Strings.Unbounded.Hash_Case_Insensitive (A.4.9)
17426 This package provides a generic hash function for unbounded strings that
17427 converts the string to be hashed to lower case.
17429 @item Ada.Strings.Unbounded.Less_Case_Insensitive (A.4.10)
17430 This package provides a comparison function for unbounded strings that works
17431 in a case insensitive manner by converting to lower case before the comparison.
17433 @item Ada.Strings.UTF_Encoding (A.4.11)
17434 This package provides basic definitions for dealing with UTF-encoded strings.
17436 @item Ada.Strings.UTF_Encoding.Conversions (A.4.11)
17437 This package provides conversion functions for UTF-encoded strings.
17439 @item Ada.Strings.UTF_Encoding.Strings (A.4.11)
17440 @itemx Ada.Strings.UTF_Encoding.Wide_Strings (A.4.11)
17441 @itemx Ada.Strings.UTF_Encoding.Wide_Wide_Strings (A.4.11)
17442 These packages provide facilities for handling UTF encodings for
17443 Strings, Wide_Strings and Wide_Wide_Strings.
17445 @item Ada.Strings.Wide_Bounded (A.4.7)
17446 @itemx Ada.Strings.Wide_Fixed (A.4.7)
17447 @itemx Ada.Strings.Wide_Maps (A.4.7)
17448 @itemx Ada.Strings.Wide_Unbounded (A.4.7)
17449 These packages provide analogous capabilities to the corresponding
17450 packages without @samp{Wide_} in the name, but operate with the types
17451 @code{Wide_String} and @code{Wide_Character} instead of @code{String}
17452 and @code{Character}. Versions of all the child packages are available.
17454 @item Ada.Strings.Wide_Wide_Bounded (A.4.7)
17455 @itemx Ada.Strings.Wide_Wide_Fixed (A.4.7)
17456 @itemx Ada.Strings.Wide_Wide_Maps (A.4.7)
17457 @itemx Ada.Strings.Wide_Wide_Unbounded (A.4.7)
17458 These packages provide analogous capabilities to the corresponding
17459 packages without @samp{Wide_} in the name, but operate with the types
17460 @code{Wide_Wide_String} and @code{Wide_Wide_Character} instead
17461 of @code{String} and @code{Character}.
17463 @item Ada.Synchronous_Barriers (D.10.1)
17464 This package provides facilities for synchronizing tasks at a low level
17467 @item Ada.Synchronous_Task_Control (D.10)
17468 This package provides some standard facilities for controlling task
17469 communication in a synchronous manner.
17471 @item Ada.Synchronous_Task_Control.EDF (D.10)
17472 Not implemented in GNAT.
17475 This package contains definitions for manipulation of the tags of tagged
17478 @item Ada.Tags.Generic_Dispatching_Constructor (3.9)
17479 This package provides a way of constructing tagged class-wide values given
17480 only the tag value.
17482 @item Ada.Task_Attributes (C.7.2)
17483 This package provides the capability of associating arbitrary
17484 task-specific data with separate tasks.
17486 @item Ada.Task_Identifification (C.7.1)
17487 This package provides capabilities for task identification.
17489 @item Ada.Task_Termination (C.7.3)
17490 This package provides control over task termination.
17493 This package provides basic text input-output capabilities for
17494 character, string and numeric data. The subpackages of this
17495 package are listed next. Note that although these are defined
17496 as subpackages in the RM, they are actually transparently
17497 implemented as child packages in GNAT, meaning that they
17498 are only loaded if needed.
17500 @item Ada.Text_IO.Decimal_IO
17501 Provides input-output facilities for decimal fixed-point types
17503 @item Ada.Text_IO.Enumeration_IO
17504 Provides input-output facilities for enumeration types.
17506 @item Ada.Text_IO.Fixed_IO
17507 Provides input-output facilities for ordinary fixed-point types.
17509 @item Ada.Text_IO.Float_IO
17510 Provides input-output facilities for float types. The following
17511 predefined instantiations of this generic package are available:
17515 @code{Short_Float_Text_IO}
17517 @code{Float_Text_IO}
17519 @code{Long_Float_Text_IO}
17522 @item Ada.Text_IO.Integer_IO
17523 Provides input-output facilities for integer types. The following
17524 predefined instantiations of this generic package are available:
17527 @item Short_Short_Integer
17528 @code{Ada.Short_Short_Integer_Text_IO}
17529 @item Short_Integer
17530 @code{Ada.Short_Integer_Text_IO}
17532 @code{Ada.Integer_Text_IO}
17534 @code{Ada.Long_Integer_Text_IO}
17535 @item Long_Long_Integer
17536 @code{Ada.Long_Long_Integer_Text_IO}
17539 @item Ada.Text_IO.Modular_IO
17540 Provides input-output facilities for modular (unsigned) types.
17542 @item Ada.Text_IO.Bounded_IO (A.10.11)
17543 Provides input-output facilities for bounded strings.
17545 @item Ada.Text_IO.Complex_IO (G.1.3)
17546 This package provides basic text input-output capabilities for complex
17549 @item Ada.Text_IO.Editing (F.3.3)
17550 This package contains routines for edited output, analogous to the use
17551 of pictures in COBOL@. The picture formats used by this package are a
17552 close copy of the facility in COBOL@.
17554 @item Ada.Text_IO.Text_Streams (A.12.2)
17555 This package provides a facility that allows Text_IO files to be treated
17556 as streams, so that the stream attributes can be used for writing
17557 arbitrary data, including binary data, to Text_IO files.
17559 @item Ada.Text_IO.Unbounded_IO (A.10.12)
17560 This package provides input-output facilities for unbounded strings.
17562 @item Ada.Unchecked_Conversion (13.9)
17563 This generic package allows arbitrary conversion from one type to
17564 another of the same size, providing for breaking the type safety in
17565 special circumstances.
17567 If the types have the same Size (more accurately the same Value_Size),
17568 then the effect is simply to transfer the bits from the source to the
17569 target type without any modification. This usage is well defined, and
17570 for simple types whose representation is typically the same across
17571 all implementations, gives a portable method of performing such
17574 If the types do not have the same size, then the result is implementation
17575 defined, and thus may be non-portable. The following describes how GNAT
17576 handles such unchecked conversion cases.
17578 If the types are of different sizes, and are both discrete types, then
17579 the effect is of a normal type conversion without any constraint checking.
17580 In particular if the result type has a larger size, the result will be
17581 zero or sign extended. If the result type has a smaller size, the result
17582 will be truncated by ignoring high order bits.
17584 If the types are of different sizes, and are not both discrete types,
17585 then the conversion works as though pointers were created to the source
17586 and target, and the pointer value is converted. The effect is that bits
17587 are copied from successive low order storage units and bits of the source
17588 up to the length of the target type.
17590 A warning is issued if the lengths differ, since the effect in this
17591 case is implementation dependent, and the above behavior may not match
17592 that of some other compiler.
17594 A pointer to one type may be converted to a pointer to another type using
17595 unchecked conversion. The only case in which the effect is undefined is
17596 when one or both pointers are pointers to unconstrained array types. In
17597 this case, the bounds information may get incorrectly transferred, and in
17598 particular, GNAT uses double size pointers for such types, and it is
17599 meaningless to convert between such pointer types. GNAT will issue a
17600 warning if the alignment of the target designated type is more strict
17601 than the alignment of the source designated type (since the result may
17602 be unaligned in this case).
17604 A pointer other than a pointer to an unconstrained array type may be
17605 converted to and from System.Address. Such usage is common in Ada 83
17606 programs, but note that Ada.Address_To_Access_Conversions is the
17607 preferred method of performing such conversions in Ada 95 and Ada 2005.
17609 unchecked conversion nor Ada.Address_To_Access_Conversions should be
17610 used in conjunction with pointers to unconstrained objects, since
17611 the bounds information cannot be handled correctly in this case.
17613 @item Ada.Unchecked_Deallocation (13.11.2)
17614 This generic package allows explicit freeing of storage previously
17615 allocated by use of an allocator.
17617 @item Ada.Wide_Text_IO (A.11)
17618 This package is similar to @code{Ada.Text_IO}, except that the external
17619 file supports wide character representations, and the internal types are
17620 @code{Wide_Character} and @code{Wide_String} instead of @code{Character}
17621 and @code{String}. The corresponding set of nested packages and child
17622 packages are defined.
17624 @item Ada.Wide_Wide_Text_IO (A.11)
17625 This package is similar to @code{Ada.Text_IO}, except that the external
17626 file supports wide character representations, and the internal types are
17627 @code{Wide_Character} and @code{Wide_String} instead of @code{Character}
17628 and @code{String}. The corresponding set of nested packages and child
17629 packages are defined.
17633 For packages in Interfaces and System, all the RM defined packages are
17634 available in GNAT, see the Ada 2012 RM for full details.
17636 @node The Implementation of Standard I/O
17637 @chapter The Implementation of Standard I/O
17640 GNAT implements all the required input-output facilities described in
17641 A.6 through A.14. These sections of the Ada Reference Manual describe the
17642 required behavior of these packages from the Ada point of view, and if
17643 you are writing a portable Ada program that does not need to know the
17644 exact manner in which Ada maps to the outside world when it comes to
17645 reading or writing external files, then you do not need to read this
17646 chapter. As long as your files are all regular files (not pipes or
17647 devices), and as long as you write and read the files only from Ada, the
17648 description in the Ada Reference Manual is sufficient.
17650 However, if you want to do input-output to pipes or other devices, such
17651 as the keyboard or screen, or if the files you are dealing with are
17652 either generated by some other language, or to be read by some other
17653 language, then you need to know more about the details of how the GNAT
17654 implementation of these input-output facilities behaves.
17656 In this chapter we give a detailed description of exactly how GNAT
17657 interfaces to the file system. As always, the sources of the system are
17658 available to you for answering questions at an even more detailed level,
17659 but for most purposes the information in this chapter will suffice.
17661 Another reason that you may need to know more about how input-output is
17662 implemented arises when you have a program written in mixed languages
17663 where, for example, files are shared between the C and Ada sections of
17664 the same program. GNAT provides some additional facilities, in the form
17665 of additional child library packages, that facilitate this sharing, and
17666 these additional facilities are also described in this chapter.
17669 * Standard I/O Packages::
17675 * Wide_Wide_Text_IO::
17677 * Text Translation::
17679 * Filenames encoding::
17680 * File content encoding::
17682 * Operations on C Streams::
17683 * Interfacing to C Streams::
17686 @node Standard I/O Packages
17687 @section Standard I/O Packages
17690 The Standard I/O packages described in Annex A for
17696 Ada.Text_IO.Complex_IO
17698 Ada.Text_IO.Text_Streams
17702 Ada.Wide_Text_IO.Complex_IO
17704 Ada.Wide_Text_IO.Text_Streams
17706 Ada.Wide_Wide_Text_IO
17708 Ada.Wide_Wide_Text_IO.Complex_IO
17710 Ada.Wide_Wide_Text_IO.Text_Streams
17720 are implemented using the C
17721 library streams facility; where
17725 All files are opened using @code{fopen}.
17727 All input/output operations use @code{fread}/@code{fwrite}.
17731 There is no internal buffering of any kind at the Ada library level. The only
17732 buffering is that provided at the system level in the implementation of the
17733 library routines that support streams. This facilitates shared use of these
17734 streams by mixed language programs. Note though that system level buffering is
17735 explicitly enabled at elaboration of the standard I/O packages and that can
17736 have an impact on mixed language programs, in particular those using I/O before
17737 calling the Ada elaboration routine (e.g.@: adainit). It is recommended to call
17738 the Ada elaboration routine before performing any I/O or when impractical,
17739 flush the common I/O streams and in particular Standard_Output before
17740 elaborating the Ada code.
17743 @section FORM Strings
17746 The format of a FORM string in GNAT is:
17749 "keyword=value,keyword=value,@dots{},keyword=value"
17753 where letters may be in upper or lower case, and there are no spaces
17754 between values. The order of the entries is not important. Currently
17755 the following keywords defined.
17758 TEXT_TRANSLATION=[YES|NO|TEXT|BINARY|U8TEXT|WTEXT|U16TEXT]
17760 WCEM=[n|h|u|s|e|8|b]
17761 ENCODING=[UTF8|8BITS]
17765 The use of these parameters is described later in this section. If an
17766 unrecognized keyword appears in a form string, it is silently ignored
17767 and not considered invalid.
17770 For OpenVMS additional FORM string keywords are available for use with
17771 RMS services. The syntax is:
17774 VMS_RMS_Keys=(keyword=value,@dots{},keyword=value)
17778 The following RMS keywords and values are currently defined:
17781 Context=Force_Stream_Mode|Force_Record_Mode
17785 VMS RMS keys are silently ignored on non-VMS systems. On OpenVMS
17786 unimplented RMS keywords, values, or invalid syntax will raise Use_Error.
17792 Direct_IO can only be instantiated for definite types. This is a
17793 restriction of the Ada language, which means that the records are fixed
17794 length (the length being determined by @code{@var{type}'Size}, rounded
17795 up to the next storage unit boundary if necessary).
17797 The records of a Direct_IO file are simply written to the file in index
17798 sequence, with the first record starting at offset zero, and subsequent
17799 records following. There is no control information of any kind. For
17800 example, if 32-bit integers are being written, each record takes
17801 4-bytes, so the record at index @var{K} starts at offset
17802 (@var{K}@minus{}1)*4.
17804 There is no limit on the size of Direct_IO files, they are expanded as
17805 necessary to accommodate whatever records are written to the file.
17807 @node Sequential_IO
17808 @section Sequential_IO
17811 Sequential_IO may be instantiated with either a definite (constrained)
17812 or indefinite (unconstrained) type.
17814 For the definite type case, the elements written to the file are simply
17815 the memory images of the data values with no control information of any
17816 kind. The resulting file should be read using the same type, no validity
17817 checking is performed on input.
17819 For the indefinite type case, the elements written consist of two
17820 parts. First is the size of the data item, written as the memory image
17821 of a @code{Interfaces.C.size_t} value, followed by the memory image of
17822 the data value. The resulting file can only be read using the same
17823 (unconstrained) type. Normal assignment checks are performed on these
17824 read operations, and if these checks fail, @code{Data_Error} is
17825 raised. In particular, in the array case, the lengths must match, and in
17826 the variant record case, if the variable for a particular read operation
17827 is constrained, the discriminants must match.
17829 Note that it is not possible to use Sequential_IO to write variable
17830 length array items, and then read the data back into different length
17831 arrays. For example, the following will raise @code{Data_Error}:
17833 @smallexample @c ada
17834 package IO is new Sequential_IO (String);
17839 IO.Write (F, "hello!")
17840 IO.Reset (F, Mode=>In_File);
17847 On some Ada implementations, this will print @code{hell}, but the program is
17848 clearly incorrect, since there is only one element in the file, and that
17849 element is the string @code{hello!}.
17851 In Ada 95 and Ada 2005, this kind of behavior can be legitimately achieved
17852 using Stream_IO, and this is the preferred mechanism. In particular, the
17853 above program fragment rewritten to use Stream_IO will work correctly.
17859 Text_IO files consist of a stream of characters containing the following
17860 special control characters:
17863 LF (line feed, 16#0A#) Line Mark
17864 FF (form feed, 16#0C#) Page Mark
17868 A canonical Text_IO file is defined as one in which the following
17869 conditions are met:
17873 The character @code{LF} is used only as a line mark, i.e.@: to mark the end
17877 The character @code{FF} is used only as a page mark, i.e.@: to mark the
17878 end of a page and consequently can appear only immediately following a
17879 @code{LF} (line mark) character.
17882 The file ends with either @code{LF} (line mark) or @code{LF}-@code{FF}
17883 (line mark, page mark). In the former case, the page mark is implicitly
17884 assumed to be present.
17888 A file written using Text_IO will be in canonical form provided that no
17889 explicit @code{LF} or @code{FF} characters are written using @code{Put}
17890 or @code{Put_Line}. There will be no @code{FF} character at the end of
17891 the file unless an explicit @code{New_Page} operation was performed
17892 before closing the file.
17894 A canonical Text_IO file that is a regular file (i.e., not a device or a
17895 pipe) can be read using any of the routines in Text_IO@. The
17896 semantics in this case will be exactly as defined in the Ada Reference
17897 Manual, and all the routines in Text_IO are fully implemented.
17899 A text file that does not meet the requirements for a canonical Text_IO
17900 file has one of the following:
17904 The file contains @code{FF} characters not immediately following a
17905 @code{LF} character.
17908 The file contains @code{LF} or @code{FF} characters written by
17909 @code{Put} or @code{Put_Line}, which are not logically considered to be
17910 line marks or page marks.
17913 The file ends in a character other than @code{LF} or @code{FF},
17914 i.e.@: there is no explicit line mark or page mark at the end of the file.
17918 Text_IO can be used to read such non-standard text files but subprograms
17919 to do with line or page numbers do not have defined meanings. In
17920 particular, a @code{FF} character that does not follow a @code{LF}
17921 character may or may not be treated as a page mark from the point of
17922 view of page and line numbering. Every @code{LF} character is considered
17923 to end a line, and there is an implied @code{LF} character at the end of
17927 * Text_IO Stream Pointer Positioning::
17928 * Text_IO Reading and Writing Non-Regular Files::
17930 * Treating Text_IO Files as Streams::
17931 * Text_IO Extensions::
17932 * Text_IO Facilities for Unbounded Strings::
17935 @node Text_IO Stream Pointer Positioning
17936 @subsection Stream Pointer Positioning
17939 @code{Ada.Text_IO} has a definition of current position for a file that
17940 is being read. No internal buffering occurs in Text_IO, and usually the
17941 physical position in the stream used to implement the file corresponds
17942 to this logical position defined by Text_IO@. There are two exceptions:
17946 After a call to @code{End_Of_Page} that returns @code{True}, the stream
17947 is positioned past the @code{LF} (line mark) that precedes the page
17948 mark. Text_IO maintains an internal flag so that subsequent read
17949 operations properly handle the logical position which is unchanged by
17950 the @code{End_Of_Page} call.
17953 After a call to @code{End_Of_File} that returns @code{True}, if the
17954 Text_IO file was positioned before the line mark at the end of file
17955 before the call, then the logical position is unchanged, but the stream
17956 is physically positioned right at the end of file (past the line mark,
17957 and past a possible page mark following the line mark. Again Text_IO
17958 maintains internal flags so that subsequent read operations properly
17959 handle the logical position.
17963 These discrepancies have no effect on the observable behavior of
17964 Text_IO, but if a single Ada stream is shared between a C program and
17965 Ada program, or shared (using @samp{shared=yes} in the form string)
17966 between two Ada files, then the difference may be observable in some
17969 @node Text_IO Reading and Writing Non-Regular Files
17970 @subsection Reading and Writing Non-Regular Files
17973 A non-regular file is a device (such as a keyboard), or a pipe. Text_IO
17974 can be used for reading and writing. Writing is not affected and the
17975 sequence of characters output is identical to the normal file case, but
17976 for reading, the behavior of Text_IO is modified to avoid undesirable
17977 look-ahead as follows:
17979 An input file that is not a regular file is considered to have no page
17980 marks. Any @code{Ascii.FF} characters (the character normally used for a
17981 page mark) appearing in the file are considered to be data
17982 characters. In particular:
17986 @code{Get_Line} and @code{Skip_Line} do not test for a page mark
17987 following a line mark. If a page mark appears, it will be treated as a
17991 This avoids the need to wait for an extra character to be typed or
17992 entered from the pipe to complete one of these operations.
17995 @code{End_Of_Page} always returns @code{False}
17998 @code{End_Of_File} will return @code{False} if there is a page mark at
17999 the end of the file.
18003 Output to non-regular files is the same as for regular files. Page marks
18004 may be written to non-regular files using @code{New_Page}, but as noted
18005 above they will not be treated as page marks on input if the output is
18006 piped to another Ada program.
18008 Another important discrepancy when reading non-regular files is that the end
18009 of file indication is not ``sticky''. If an end of file is entered, e.g.@: by
18010 pressing the @key{EOT} key,
18012 is signaled once (i.e.@: the test @code{End_Of_File}
18013 will yield @code{True}, or a read will
18014 raise @code{End_Error}), but then reading can resume
18015 to read data past that end of
18016 file indication, until another end of file indication is entered.
18018 @node Get_Immediate
18019 @subsection Get_Immediate
18020 @cindex Get_Immediate
18023 Get_Immediate returns the next character (including control characters)
18024 from the input file. In particular, Get_Immediate will return LF or FF
18025 characters used as line marks or page marks. Such operations leave the
18026 file positioned past the control character, and it is thus not treated
18027 as having its normal function. This means that page, line and column
18028 counts after this kind of Get_Immediate call are set as though the mark
18029 did not occur. In the case where a Get_Immediate leaves the file
18030 positioned between the line mark and page mark (which is not normally
18031 possible), it is undefined whether the FF character will be treated as a
18034 @node Treating Text_IO Files as Streams
18035 @subsection Treating Text_IO Files as Streams
18036 @cindex Stream files
18039 The package @code{Text_IO.Streams} allows a Text_IO file to be treated
18040 as a stream. Data written to a Text_IO file in this stream mode is
18041 binary data. If this binary data contains bytes 16#0A# (@code{LF}) or
18042 16#0C# (@code{FF}), the resulting file may have non-standard
18043 format. Similarly if read operations are used to read from a Text_IO
18044 file treated as a stream, then @code{LF} and @code{FF} characters may be
18045 skipped and the effect is similar to that described above for
18046 @code{Get_Immediate}.
18048 @node Text_IO Extensions
18049 @subsection Text_IO Extensions
18050 @cindex Text_IO extensions
18053 A package GNAT.IO_Aux in the GNAT library provides some useful extensions
18054 to the standard @code{Text_IO} package:
18057 @item function File_Exists (Name : String) return Boolean;
18058 Determines if a file of the given name exists.
18060 @item function Get_Line return String;
18061 Reads a string from the standard input file. The value returned is exactly
18062 the length of the line that was read.
18064 @item function Get_Line (File : Ada.Text_IO.File_Type) return String;
18065 Similar, except that the parameter File specifies the file from which
18066 the string is to be read.
18070 @node Text_IO Facilities for Unbounded Strings
18071 @subsection Text_IO Facilities for Unbounded Strings
18072 @cindex Text_IO for unbounded strings
18073 @cindex Unbounded_String, Text_IO operations
18076 The package @code{Ada.Strings.Unbounded.Text_IO}
18077 in library files @code{a-suteio.ads/adb} contains some GNAT-specific
18078 subprograms useful for Text_IO operations on unbounded strings:
18082 @item function Get_Line (File : File_Type) return Unbounded_String;
18083 Reads a line from the specified file
18084 and returns the result as an unbounded string.
18086 @item procedure Put (File : File_Type; U : Unbounded_String);
18087 Writes the value of the given unbounded string to the specified file
18088 Similar to the effect of
18089 @code{Put (To_String (U))} except that an extra copy is avoided.
18091 @item procedure Put_Line (File : File_Type; U : Unbounded_String);
18092 Writes the value of the given unbounded string to the specified file,
18093 followed by a @code{New_Line}.
18094 Similar to the effect of @code{Put_Line (To_String (U))} except
18095 that an extra copy is avoided.
18099 In the above procedures, @code{File} is of type @code{Ada.Text_IO.File_Type}
18100 and is optional. If the parameter is omitted, then the standard input or
18101 output file is referenced as appropriate.
18103 The package @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} in library
18104 files @file{a-swuwti.ads} and @file{a-swuwti.adb} provides similar extended
18105 @code{Wide_Text_IO} functionality for unbounded wide strings.
18107 The package @code{Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO} in library
18108 files @file{a-szuzti.ads} and @file{a-szuzti.adb} provides similar extended
18109 @code{Wide_Wide_Text_IO} functionality for unbounded wide wide strings.
18112 @section Wide_Text_IO
18115 @code{Wide_Text_IO} is similar in most respects to Text_IO, except that
18116 both input and output files may contain special sequences that represent
18117 wide character values. The encoding scheme for a given file may be
18118 specified using a FORM parameter:
18125 as part of the FORM string (WCEM = wide character encoding method),
18126 where @var{x} is one of the following characters
18132 Upper half encoding
18144 The encoding methods match those that
18145 can be used in a source
18146 program, but there is no requirement that the encoding method used for
18147 the source program be the same as the encoding method used for files,
18148 and different files may use different encoding methods.
18150 The default encoding method for the standard files, and for opened files
18151 for which no WCEM parameter is given in the FORM string matches the
18152 wide character encoding specified for the main program (the default
18153 being brackets encoding if no coding method was specified with -gnatW).
18157 In this encoding, a wide character is represented by a five character
18165 where @var{a}, @var{b}, @var{c}, @var{d} are the four hexadecimal
18166 characters (using upper case letters) of the wide character code. For
18167 example, ESC A345 is used to represent the wide character with code
18168 16#A345#. This scheme is compatible with use of the full
18169 @code{Wide_Character} set.
18171 @item Upper Half Coding
18172 The wide character with encoding 16#abcd#, where the upper bit is on
18173 (i.e.@: a is in the range 8-F) is represented as two bytes 16#ab# and
18174 16#cd#. The second byte may never be a format control character, but is
18175 not required to be in the upper half. This method can be also used for
18176 shift-JIS or EUC where the internal coding matches the external coding.
18178 @item Shift JIS Coding
18179 A wide character is represented by a two character sequence 16#ab# and
18180 16#cd#, with the restrictions described for upper half encoding as
18181 described above. The internal character code is the corresponding JIS
18182 character according to the standard algorithm for Shift-JIS
18183 conversion. Only characters defined in the JIS code set table can be
18184 used with this encoding method.
18187 A wide character is represented by a two character sequence 16#ab# and
18188 16#cd#, with both characters being in the upper half. The internal
18189 character code is the corresponding JIS character according to the EUC
18190 encoding algorithm. Only characters defined in the JIS code set table
18191 can be used with this encoding method.
18194 A wide character is represented using
18195 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
18196 10646-1/Am.2. Depending on the character value, the representation
18197 is a one, two, or three byte sequence:
18200 16#0000#-16#007f#: 2#0xxxxxxx#
18201 16#0080#-16#07ff#: 2#110xxxxx# 2#10xxxxxx#
18202 16#0800#-16#ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
18206 where the @var{xxx} bits correspond to the left-padded bits of the
18207 16-bit character value. Note that all lower half ASCII characters
18208 are represented as ASCII bytes and all upper half characters and
18209 other wide characters are represented as sequences of upper-half
18210 (The full UTF-8 scheme allows for encoding 31-bit characters as
18211 6-byte sequences, but in this implementation, all UTF-8 sequences
18212 of four or more bytes length will raise a Constraint_Error, as
18213 will all invalid UTF-8 sequences.)
18215 @item Brackets Coding
18216 In this encoding, a wide character is represented by the following eight
18217 character sequence:
18224 where @code{a}, @code{b}, @code{c}, @code{d} are the four hexadecimal
18225 characters (using uppercase letters) of the wide character code. For
18226 example, @code{["A345"]} is used to represent the wide character with code
18228 This scheme is compatible with use of the full Wide_Character set.
18229 On input, brackets coding can also be used for upper half characters,
18230 e.g.@: @code{["C1"]} for lower case a. However, on output, brackets notation
18231 is only used for wide characters with a code greater than @code{16#FF#}.
18233 Note that brackets coding is not normally used in the context of
18234 Wide_Text_IO or Wide_Wide_Text_IO, since it is really just designed as
18235 a portable way of encoding source files. In the context of Wide_Text_IO
18236 or Wide_Wide_Text_IO, it can only be used if the file does not contain
18237 any instance of the left bracket character other than to encode wide
18238 character values using the brackets encoding method. In practice it is
18239 expected that some standard wide character encoding method such
18240 as UTF-8 will be used for text input output.
18242 If brackets notation is used, then any occurrence of a left bracket
18243 in the input file which is not the start of a valid wide character
18244 sequence will cause Constraint_Error to be raised. It is possible to
18245 encode a left bracket as ["5B"] and Wide_Text_IO and Wide_Wide_Text_IO
18246 input will interpret this as a left bracket.
18248 However, when a left bracket is output, it will be output as a left bracket
18249 and not as ["5B"]. We make this decision because for normal use of
18250 Wide_Text_IO for outputting messages, it is unpleasant to clobber left
18251 brackets. For example, if we write:
18254 Put_Line ("Start of output [first run]");
18258 we really do not want to have the left bracket in this message clobbered so
18259 that the output reads:
18262 Start of output ["5B"]first run]
18266 In practice brackets encoding is reasonably useful for normal Put_Line use
18267 since we won't get confused between left brackets and wide character
18268 sequences in the output. But for input, or when files are written out
18269 and read back in, it really makes better sense to use one of the standard
18270 encoding methods such as UTF-8.
18275 For the coding schemes other than UTF-8, Hex, or Brackets encoding,
18276 not all wide character
18277 values can be represented. An attempt to output a character that cannot
18278 be represented using the encoding scheme for the file causes
18279 Constraint_Error to be raised. An invalid wide character sequence on
18280 input also causes Constraint_Error to be raised.
18283 * Wide_Text_IO Stream Pointer Positioning::
18284 * Wide_Text_IO Reading and Writing Non-Regular Files::
18287 @node Wide_Text_IO Stream Pointer Positioning
18288 @subsection Stream Pointer Positioning
18291 @code{Ada.Wide_Text_IO} is similar to @code{Ada.Text_IO} in its handling
18292 of stream pointer positioning (@pxref{Text_IO}). There is one additional
18295 If @code{Ada.Wide_Text_IO.Look_Ahead} reads a character outside the
18296 normal lower ASCII set (i.e.@: a character in the range:
18298 @smallexample @c ada
18299 Wide_Character'Val (16#0080#) .. Wide_Character'Val (16#FFFF#)
18303 then although the logical position of the file pointer is unchanged by
18304 the @code{Look_Ahead} call, the stream is physically positioned past the
18305 wide character sequence. Again this is to avoid the need for buffering
18306 or backup, and all @code{Wide_Text_IO} routines check the internal
18307 indication that this situation has occurred so that this is not visible
18308 to a normal program using @code{Wide_Text_IO}. However, this discrepancy
18309 can be observed if the wide text file shares a stream with another file.
18311 @node Wide_Text_IO Reading and Writing Non-Regular Files
18312 @subsection Reading and Writing Non-Regular Files
18315 As in the case of Text_IO, when a non-regular file is read, it is
18316 assumed that the file contains no page marks (any form characters are
18317 treated as data characters), and @code{End_Of_Page} always returns
18318 @code{False}. Similarly, the end of file indication is not sticky, so
18319 it is possible to read beyond an end of file.
18321 @node Wide_Wide_Text_IO
18322 @section Wide_Wide_Text_IO
18325 @code{Wide_Wide_Text_IO} is similar in most respects to Text_IO, except that
18326 both input and output files may contain special sequences that represent
18327 wide wide character values. The encoding scheme for a given file may be
18328 specified using a FORM parameter:
18335 as part of the FORM string (WCEM = wide character encoding method),
18336 where @var{x} is one of the following characters
18342 Upper half encoding
18354 The encoding methods match those that
18355 can be used in a source
18356 program, but there is no requirement that the encoding method used for
18357 the source program be the same as the encoding method used for files,
18358 and different files may use different encoding methods.
18360 The default encoding method for the standard files, and for opened files
18361 for which no WCEM parameter is given in the FORM string matches the
18362 wide character encoding specified for the main program (the default
18363 being brackets encoding if no coding method was specified with -gnatW).
18368 A wide character is represented using
18369 UCS Transformation Format 8 (UTF-8) as defined in Annex R of ISO
18370 10646-1/Am.2. Depending on the character value, the representation
18371 is a one, two, three, or four byte sequence:
18374 16#000000#-16#00007f#: 2#0xxxxxxx#
18375 16#000080#-16#0007ff#: 2#110xxxxx# 2#10xxxxxx#
18376 16#000800#-16#00ffff#: 2#1110xxxx# 2#10xxxxxx# 2#10xxxxxx#
18377 16#010000#-16#10ffff#: 2#11110xxx# 2#10xxxxxx# 2#10xxxxxx# 2#10xxxxxx#
18381 where the @var{xxx} bits correspond to the left-padded bits of the
18382 21-bit character value. Note that all lower half ASCII characters
18383 are represented as ASCII bytes and all upper half characters and
18384 other wide characters are represented as sequences of upper-half
18387 @item Brackets Coding
18388 In this encoding, a wide wide character is represented by the following eight
18389 character sequence if is in wide character range
18395 and by the following ten character sequence if not
18398 [ " a b c d e f " ]
18402 where @code{a}, @code{b}, @code{c}, @code{d}, @code{e}, and @code{f}
18403 are the four or six hexadecimal
18404 characters (using uppercase letters) of the wide wide character code. For
18405 example, @code{["01A345"]} is used to represent the wide wide character
18406 with code @code{16#01A345#}.
18408 This scheme is compatible with use of the full Wide_Wide_Character set.
18409 On input, brackets coding can also be used for upper half characters,
18410 e.g.@: @code{["C1"]} for lower case a. However, on output, brackets notation
18411 is only used for wide characters with a code greater than @code{16#FF#}.
18416 If is also possible to use the other Wide_Character encoding methods,
18417 such as Shift-JIS, but the other schemes cannot support the full range
18418 of wide wide characters.
18419 An attempt to output a character that cannot
18420 be represented using the encoding scheme for the file causes
18421 Constraint_Error to be raised. An invalid wide character sequence on
18422 input also causes Constraint_Error to be raised.
18425 * Wide_Wide_Text_IO Stream Pointer Positioning::
18426 * Wide_Wide_Text_IO Reading and Writing Non-Regular Files::
18429 @node Wide_Wide_Text_IO Stream Pointer Positioning
18430 @subsection Stream Pointer Positioning
18433 @code{Ada.Wide_Wide_Text_IO} is similar to @code{Ada.Text_IO} in its handling
18434 of stream pointer positioning (@pxref{Text_IO}). There is one additional
18437 If @code{Ada.Wide_Wide_Text_IO.Look_Ahead} reads a character outside the
18438 normal lower ASCII set (i.e.@: a character in the range:
18440 @smallexample @c ada
18441 Wide_Wide_Character'Val (16#0080#) .. Wide_Wide_Character'Val (16#10FFFF#)
18445 then although the logical position of the file pointer is unchanged by
18446 the @code{Look_Ahead} call, the stream is physically positioned past the
18447 wide character sequence. Again this is to avoid the need for buffering
18448 or backup, and all @code{Wide_Wide_Text_IO} routines check the internal
18449 indication that this situation has occurred so that this is not visible
18450 to a normal program using @code{Wide_Wide_Text_IO}. However, this discrepancy
18451 can be observed if the wide text file shares a stream with another file.
18453 @node Wide_Wide_Text_IO Reading and Writing Non-Regular Files
18454 @subsection Reading and Writing Non-Regular Files
18457 As in the case of Text_IO, when a non-regular file is read, it is
18458 assumed that the file contains no page marks (any form characters are
18459 treated as data characters), and @code{End_Of_Page} always returns
18460 @code{False}. Similarly, the end of file indication is not sticky, so
18461 it is possible to read beyond an end of file.
18467 A stream file is a sequence of bytes, where individual elements are
18468 written to the file as described in the Ada Reference Manual. The type
18469 @code{Stream_Element} is simply a byte. There are two ways to read or
18470 write a stream file.
18474 The operations @code{Read} and @code{Write} directly read or write a
18475 sequence of stream elements with no control information.
18478 The stream attributes applied to a stream file transfer data in the
18479 manner described for stream attributes.
18482 @node Text Translation
18483 @section Text Translation
18486 @samp{Text_Translation=@var{xxx}} may be used as the Form parameter
18487 passed to Text_IO.Create and Text_IO.Open. @samp{Text_Translation=@var{xxx}}
18488 has no effect on Unix systems. Possible values are:
18493 @samp{@var{Yes} or @var{Text}} is the default, which means to
18494 translate LF to/from CR/LF on Windows systems.
18496 @samp{@var{No}} disables this translation; i.e. it
18497 uses binary mode. For output files, @samp{Text_Translation=@var{No}}
18498 may be used to create Unix-style files on
18502 @samp{@var{wtext}} translation enabled in Unicode mode.
18503 (corresponds to _O_WTEXT).
18506 @samp{@var{u8text}} translation enabled in Unicode UTF-8 mode.
18507 (corresponds to O_U8TEXT).
18510 @samp{@var{u16text}} translation enabled in Unicode UTF-16
18511 mode. (corresponds to_O_U16TEXT).
18518 @section Shared Files
18521 Section A.14 of the Ada Reference Manual allows implementations to
18522 provide a wide variety of behavior if an attempt is made to access the
18523 same external file with two or more internal files.
18525 To provide a full range of functionality, while at the same time
18526 minimizing the problems of portability caused by this implementation
18527 dependence, GNAT handles file sharing as follows:
18531 In the absence of a @samp{shared=@var{xxx}} form parameter, an attempt
18532 to open two or more files with the same full name is considered an error
18533 and is not supported. The exception @code{Use_Error} will be
18534 raised. Note that a file that is not explicitly closed by the program
18535 remains open until the program terminates.
18538 If the form parameter @samp{shared=no} appears in the form string, the
18539 file can be opened or created with its own separate stream identifier,
18540 regardless of whether other files sharing the same external file are
18541 opened. The exact effect depends on how the C stream routines handle
18542 multiple accesses to the same external files using separate streams.
18545 If the form parameter @samp{shared=yes} appears in the form string for
18546 each of two or more files opened using the same full name, the same
18547 stream is shared between these files, and the semantics are as described
18548 in Ada Reference Manual, Section A.14.
18552 When a program that opens multiple files with the same name is ported
18553 from another Ada compiler to GNAT, the effect will be that
18554 @code{Use_Error} is raised.
18556 The documentation of the original compiler and the documentation of the
18557 program should then be examined to determine if file sharing was
18558 expected, and @samp{shared=@var{xxx}} parameters added to @code{Open}
18559 and @code{Create} calls as required.
18561 When a program is ported from GNAT to some other Ada compiler, no
18562 special attention is required unless the @samp{shared=@var{xxx}} form
18563 parameter is used in the program. In this case, you must examine the
18564 documentation of the new compiler to see if it supports the required
18565 file sharing semantics, and form strings modified appropriately. Of
18566 course it may be the case that the program cannot be ported if the
18567 target compiler does not support the required functionality. The best
18568 approach in writing portable code is to avoid file sharing (and hence
18569 the use of the @samp{shared=@var{xxx}} parameter in the form string)
18572 One common use of file sharing in Ada 83 is the use of instantiations of
18573 Sequential_IO on the same file with different types, to achieve
18574 heterogeneous input-output. Although this approach will work in GNAT if
18575 @samp{shared=yes} is specified, it is preferable in Ada to use Stream_IO
18576 for this purpose (using the stream attributes)
18578 @node Filenames encoding
18579 @section Filenames encoding
18582 An encoding form parameter can be used to specify the filename
18583 encoding @samp{encoding=@var{xxx}}.
18587 If the form parameter @samp{encoding=utf8} appears in the form string, the
18588 filename must be encoded in UTF-8.
18591 If the form parameter @samp{encoding=8bits} appears in the form
18592 string, the filename must be a standard 8bits string.
18595 In the absence of a @samp{encoding=@var{xxx}} form parameter, the
18596 encoding is controlled by the @samp{GNAT_CODE_PAGE} environment
18597 variable. And if not set @samp{utf8} is assumed.
18601 The current system Windows ANSI code page.
18606 This encoding form parameter is only supported on the Windows
18607 platform. On the other Operating Systems the run-time is supporting
18610 @node File content encoding
18611 @section File content encoding
18614 For text files it is possible to specify the encoding to use. This is
18615 controlled by the by the @samp{GNAT_CCS_ENCODING} environment
18616 variable. And if not set @samp{TEXT} is assumed.
18618 The possible values are those supported on Windows:
18622 Translated text mode
18624 Translated unicode encoding
18626 Unicode 16-bit encoding
18628 Unicode 8-bit encoding
18631 This encoding is only supported on the Windows platform.
18634 @section Open Modes
18637 @code{Open} and @code{Create} calls result in a call to @code{fopen}
18638 using the mode shown in the following table:
18641 @center @code{Open} and @code{Create} Call Modes
18643 @b{OPEN } @b{CREATE}
18644 Append_File "r+" "w+"
18646 Out_File (Direct_IO) "r+" "w"
18647 Out_File (all other cases) "w" "w"
18648 Inout_File "r+" "w+"
18652 If text file translation is required, then either @samp{b} or @samp{t}
18653 is added to the mode, depending on the setting of Text. Text file
18654 translation refers to the mapping of CR/LF sequences in an external file
18655 to LF characters internally. This mapping only occurs in DOS and
18656 DOS-like systems, and is not relevant to other systems.
18658 A special case occurs with Stream_IO@. As shown in the above table, the
18659 file is initially opened in @samp{r} or @samp{w} mode for the
18660 @code{In_File} and @code{Out_File} cases. If a @code{Set_Mode} operation
18661 subsequently requires switching from reading to writing or vice-versa,
18662 then the file is reopened in @samp{r+} mode to permit the required operation.
18664 @node Operations on C Streams
18665 @section Operations on C Streams
18666 The package @code{Interfaces.C_Streams} provides an Ada program with direct
18667 access to the C library functions for operations on C streams:
18669 @smallexample @c adanocomment
18670 package Interfaces.C_Streams is
18671 -- Note: the reason we do not use the types that are in
18672 -- Interfaces.C is that we want to avoid dragging in the
18673 -- code in this unit if possible.
18674 subtype chars is System.Address;
18675 -- Pointer to null-terminated array of characters
18676 subtype FILEs is System.Address;
18677 -- Corresponds to the C type FILE*
18678 subtype voids is System.Address;
18679 -- Corresponds to the C type void*
18680 subtype int is Integer;
18681 subtype long is Long_Integer;
18682 -- Note: the above types are subtypes deliberately, and it
18683 -- is part of this spec that the above correspondences are
18684 -- guaranteed. This means that it is legitimate to, for
18685 -- example, use Integer instead of int. We provide these
18686 -- synonyms for clarity, but in some cases it may be
18687 -- convenient to use the underlying types (for example to
18688 -- avoid an unnecessary dependency of a spec on the spec
18690 type size_t is mod 2 ** Standard'Address_Size;
18691 NULL_Stream : constant FILEs;
18692 -- Value returned (NULL in C) to indicate an
18693 -- fdopen/fopen/tmpfile error
18694 ----------------------------------
18695 -- Constants Defined in stdio.h --
18696 ----------------------------------
18697 EOF : constant int;
18698 -- Used by a number of routines to indicate error or
18700 IOFBF : constant int;
18701 IOLBF : constant int;
18702 IONBF : constant int;
18703 -- Used to indicate buffering mode for setvbuf call
18704 SEEK_CUR : constant int;
18705 SEEK_END : constant int;
18706 SEEK_SET : constant int;
18707 -- Used to indicate origin for fseek call
18708 function stdin return FILEs;
18709 function stdout return FILEs;
18710 function stderr return FILEs;
18711 -- Streams associated with standard files
18712 --------------------------
18713 -- Standard C functions --
18714 --------------------------
18715 -- The functions selected below are ones that are
18716 -- available in UNIX (but not necessarily in ANSI C).
18717 -- These are very thin interfaces
18718 -- which copy exactly the C headers. For more
18719 -- documentation on these functions, see the Microsoft C
18720 -- "Run-Time Library Reference" (Microsoft Press, 1990,
18721 -- ISBN 1-55615-225-6), which includes useful information
18722 -- on system compatibility.
18723 procedure clearerr (stream : FILEs);
18724 function fclose (stream : FILEs) return int;
18725 function fdopen (handle : int; mode : chars) return FILEs;
18726 function feof (stream : FILEs) return int;
18727 function ferror (stream : FILEs) return int;
18728 function fflush (stream : FILEs) return int;
18729 function fgetc (stream : FILEs) return int;
18730 function fgets (strng : chars; n : int; stream : FILEs)
18732 function fileno (stream : FILEs) return int;
18733 function fopen (filename : chars; Mode : chars)
18735 -- Note: to maintain target independence, use
18736 -- text_translation_required, a boolean variable defined in
18737 -- a-sysdep.c to deal with the target dependent text
18738 -- translation requirement. If this variable is set,
18739 -- then b/t should be appended to the standard mode
18740 -- argument to set the text translation mode off or on
18742 function fputc (C : int; stream : FILEs) return int;
18743 function fputs (Strng : chars; Stream : FILEs) return int;
18760 function ftell (stream : FILEs) return long;
18767 function isatty (handle : int) return int;
18768 procedure mktemp (template : chars);
18769 -- The return value (which is just a pointer to template)
18771 procedure rewind (stream : FILEs);
18772 function rmtmp return int;
18780 function tmpfile return FILEs;
18781 function ungetc (c : int; stream : FILEs) return int;
18782 function unlink (filename : chars) return int;
18783 ---------------------
18784 -- Extra functions --
18785 ---------------------
18786 -- These functions supply slightly thicker bindings than
18787 -- those above. They are derived from functions in the
18788 -- C Run-Time Library, but may do a bit more work than
18789 -- just directly calling one of the Library functions.
18790 function is_regular_file (handle : int) return int;
18791 -- Tests if given handle is for a regular file (result 1)
18792 -- or for a non-regular file (pipe or device, result 0).
18793 ---------------------------------
18794 -- Control of Text/Binary Mode --
18795 ---------------------------------
18796 -- If text_translation_required is true, then the following
18797 -- functions may be used to dynamically switch a file from
18798 -- binary to text mode or vice versa. These functions have
18799 -- no effect if text_translation_required is false (i.e.@: in
18800 -- normal UNIX mode). Use fileno to get a stream handle.
18801 procedure set_binary_mode (handle : int);
18802 procedure set_text_mode (handle : int);
18803 ----------------------------
18804 -- Full Path Name support --
18805 ----------------------------
18806 procedure full_name (nam : chars; buffer : chars);
18807 -- Given a NUL terminated string representing a file
18808 -- name, returns in buffer a NUL terminated string
18809 -- representing the full path name for the file name.
18810 -- On systems where it is relevant the drive is also
18811 -- part of the full path name. It is the responsibility
18812 -- of the caller to pass an actual parameter for buffer
18813 -- that is big enough for any full path name. Use
18814 -- max_path_len given below as the size of buffer.
18815 max_path_len : integer;
18816 -- Maximum length of an allowable full path name on the
18817 -- system, including a terminating NUL character.
18818 end Interfaces.C_Streams;
18821 @node Interfacing to C Streams
18822 @section Interfacing to C Streams
18825 The packages in this section permit interfacing Ada files to C Stream
18828 @smallexample @c ada
18829 with Interfaces.C_Streams;
18830 package Ada.Sequential_IO.C_Streams is
18831 function C_Stream (F : File_Type)
18832 return Interfaces.C_Streams.FILEs;
18834 (File : in out File_Type;
18835 Mode : in File_Mode;
18836 C_Stream : in Interfaces.C_Streams.FILEs;
18837 Form : in String := "");
18838 end Ada.Sequential_IO.C_Streams;
18840 with Interfaces.C_Streams;
18841 package Ada.Direct_IO.C_Streams is
18842 function C_Stream (F : File_Type)
18843 return Interfaces.C_Streams.FILEs;
18845 (File : in out File_Type;
18846 Mode : in File_Mode;
18847 C_Stream : in Interfaces.C_Streams.FILEs;
18848 Form : in String := "");
18849 end Ada.Direct_IO.C_Streams;
18851 with Interfaces.C_Streams;
18852 package Ada.Text_IO.C_Streams is
18853 function C_Stream (F : File_Type)
18854 return Interfaces.C_Streams.FILEs;
18856 (File : in out File_Type;
18857 Mode : in File_Mode;
18858 C_Stream : in Interfaces.C_Streams.FILEs;
18859 Form : in String := "");
18860 end Ada.Text_IO.C_Streams;
18862 with Interfaces.C_Streams;
18863 package Ada.Wide_Text_IO.C_Streams is
18864 function C_Stream (F : File_Type)
18865 return Interfaces.C_Streams.FILEs;
18867 (File : in out File_Type;
18868 Mode : in File_Mode;
18869 C_Stream : in Interfaces.C_Streams.FILEs;
18870 Form : in String := "");
18871 end Ada.Wide_Text_IO.C_Streams;
18873 with Interfaces.C_Streams;
18874 package Ada.Wide_Wide_Text_IO.C_Streams is
18875 function C_Stream (F : File_Type)
18876 return Interfaces.C_Streams.FILEs;
18878 (File : in out File_Type;
18879 Mode : in File_Mode;
18880 C_Stream : in Interfaces.C_Streams.FILEs;
18881 Form : in String := "");
18882 end Ada.Wide_Wide_Text_IO.C_Streams;
18884 with Interfaces.C_Streams;
18885 package Ada.Stream_IO.C_Streams is
18886 function C_Stream (F : File_Type)
18887 return Interfaces.C_Streams.FILEs;
18889 (File : in out File_Type;
18890 Mode : in File_Mode;
18891 C_Stream : in Interfaces.C_Streams.FILEs;
18892 Form : in String := "");
18893 end Ada.Stream_IO.C_Streams;
18897 In each of these six packages, the @code{C_Stream} function obtains the
18898 @code{FILE} pointer from a currently opened Ada file. It is then
18899 possible to use the @code{Interfaces.C_Streams} package to operate on
18900 this stream, or the stream can be passed to a C program which can
18901 operate on it directly. Of course the program is responsible for
18902 ensuring that only appropriate sequences of operations are executed.
18904 One particular use of relevance to an Ada program is that the
18905 @code{setvbuf} function can be used to control the buffering of the
18906 stream used by an Ada file. In the absence of such a call the standard
18907 default buffering is used.
18909 The @code{Open} procedures in these packages open a file giving an
18910 existing C Stream instead of a file name. Typically this stream is
18911 imported from a C program, allowing an Ada file to operate on an
18914 @node The GNAT Library
18915 @chapter The GNAT Library
18918 The GNAT library contains a number of general and special purpose packages.
18919 It represents functionality that the GNAT developers have found useful, and
18920 which is made available to GNAT users. The packages described here are fully
18921 supported, and upwards compatibility will be maintained in future releases,
18922 so you can use these facilities with the confidence that the same functionality
18923 will be available in future releases.
18925 The chapter here simply gives a brief summary of the facilities available.
18926 The full documentation is found in the spec file for the package. The full
18927 sources of these library packages, including both spec and body, are provided
18928 with all GNAT releases. For example, to find out the full specifications of
18929 the SPITBOL pattern matching capability, including a full tutorial and
18930 extensive examples, look in the @file{g-spipat.ads} file in the library.
18932 For each entry here, the package name (as it would appear in a @code{with}
18933 clause) is given, followed by the name of the corresponding spec file in
18934 parentheses. The packages are children in four hierarchies, @code{Ada},
18935 @code{Interfaces}, @code{System}, and @code{GNAT}, the latter being a
18936 GNAT-specific hierarchy.
18938 Note that an application program should only use packages in one of these
18939 four hierarchies if the package is defined in the Ada Reference Manual,
18940 or is listed in this section of the GNAT Programmers Reference Manual.
18941 All other units should be considered internal implementation units and
18942 should not be directly @code{with}'ed by application code. The use of
18943 a @code{with} statement that references one of these internal implementation
18944 units makes an application potentially dependent on changes in versions
18945 of GNAT, and will generate a warning message.
18948 * Ada.Characters.Latin_9 (a-chlat9.ads)::
18949 * Ada.Characters.Wide_Latin_1 (a-cwila1.ads)::
18950 * Ada.Characters.Wide_Latin_9 (a-cwila9.ads)::
18951 * Ada.Characters.Wide_Wide_Latin_1 (a-chzla1.ads)::
18952 * Ada.Characters.Wide_Wide_Latin_9 (a-chzla9.ads)::
18953 * Ada.Containers.Formal_Doubly_Linked_Lists (a-cfdlli.ads)::
18954 * Ada.Containers.Formal_Hashed_Maps (a-cfhama.ads)::
18955 * Ada.Containers.Formal_Hashed_Sets (a-cfhase.ads)::
18956 * Ada.Containers.Formal_Ordered_Maps (a-cforma.ads)::
18957 * Ada.Containers.Formal_Ordered_Sets (a-cforse.ads)::
18958 * Ada.Containers.Formal_Vectors (a-cofove.ads)::
18959 * Ada.Command_Line.Environment (a-colien.ads)::
18960 * Ada.Command_Line.Remove (a-colire.ads)::
18961 * Ada.Command_Line.Response_File (a-clrefi.ads)::
18962 * Ada.Direct_IO.C_Streams (a-diocst.ads)::
18963 * Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)::
18964 * Ada.Exceptions.Last_Chance_Handler (a-elchha.ads)::
18965 * Ada.Exceptions.Traceback (a-exctra.ads)::
18966 * Ada.Sequential_IO.C_Streams (a-siocst.ads)::
18967 * Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)::
18968 * Ada.Strings.Unbounded.Text_IO (a-suteio.ads)::
18969 * Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)::
18970 * Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO (a-szuzti.ads)::
18971 * Ada.Text_IO.C_Streams (a-tiocst.ads)::
18972 * Ada.Text_IO.Reset_Standard_Files (a-tirsfi.ads)::
18973 * Ada.Wide_Characters.Unicode (a-wichun.ads)::
18974 * Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)::
18975 * Ada.Wide_Text_IO.Reset_Standard_Files (a-wrstfi.ads)::
18976 * Ada.Wide_Wide_Characters.Unicode (a-zchuni.ads)::
18977 * Ada.Wide_Wide_Text_IO.C_Streams (a-ztcstr.ads)::
18978 * Ada.Wide_Wide_Text_IO.Reset_Standard_Files (a-zrstfi.ads)::
18979 * GNAT.Altivec (g-altive.ads)::
18980 * GNAT.Altivec.Conversions (g-altcon.ads)::
18981 * GNAT.Altivec.Vector_Operations (g-alveop.ads)::
18982 * GNAT.Altivec.Vector_Types (g-alvety.ads)::
18983 * GNAT.Altivec.Vector_Views (g-alvevi.ads)::
18984 * GNAT.Array_Split (g-arrspl.ads)::
18985 * GNAT.AWK (g-awk.ads)::
18986 * GNAT.Bounded_Buffers (g-boubuf.ads)::
18987 * GNAT.Bounded_Mailboxes (g-boumai.ads)::
18988 * GNAT.Bubble_Sort (g-bubsor.ads)::
18989 * GNAT.Bubble_Sort_A (g-busora.ads)::
18990 * GNAT.Bubble_Sort_G (g-busorg.ads)::
18991 * GNAT.Byte_Order_Mark (g-byorma.ads)::
18992 * GNAT.Byte_Swapping (g-bytswa.ads)::
18993 * GNAT.Calendar (g-calend.ads)::
18994 * GNAT.Calendar.Time_IO (g-catiio.ads)::
18995 * GNAT.Case_Util (g-casuti.ads)::
18996 * GNAT.CGI (g-cgi.ads)::
18997 * GNAT.CGI.Cookie (g-cgicoo.ads)::
18998 * GNAT.CGI.Debug (g-cgideb.ads)::
18999 * GNAT.Command_Line (g-comlin.ads)::
19000 * GNAT.Compiler_Version (g-comver.ads)::
19001 * GNAT.Ctrl_C (g-ctrl_c.ads)::
19002 * GNAT.CRC32 (g-crc32.ads)::
19003 * GNAT.Current_Exception (g-curexc.ads)::
19004 * GNAT.Debug_Pools (g-debpoo.ads)::
19005 * GNAT.Debug_Utilities (g-debuti.ads)::
19006 * GNAT.Decode_String (g-decstr.ads)::
19007 * GNAT.Decode_UTF8_String (g-deutst.ads)::
19008 * GNAT.Directory_Operations (g-dirope.ads)::
19009 * GNAT.Directory_Operations.Iteration (g-diopit.ads)::
19010 * GNAT.Dynamic_HTables (g-dynhta.ads)::
19011 * GNAT.Dynamic_Tables (g-dyntab.ads)::
19012 * GNAT.Encode_String (g-encstr.ads)::
19013 * GNAT.Encode_UTF8_String (g-enutst.ads)::
19014 * GNAT.Exception_Actions (g-excact.ads)::
19015 * GNAT.Exception_Traces (g-exctra.ads)::
19016 * GNAT.Exceptions (g-except.ads)::
19017 * GNAT.Expect (g-expect.ads)::
19018 * GNAT.Expect.TTY (g-exptty.ads)::
19019 * GNAT.Float_Control (g-flocon.ads)::
19020 * GNAT.Formatted_String (g-forstr.ads)::
19021 * GNAT.Heap_Sort (g-heasor.ads)::
19022 * GNAT.Heap_Sort_A (g-hesora.ads)::
19023 * GNAT.Heap_Sort_G (g-hesorg.ads)::
19024 * GNAT.HTable (g-htable.ads)::
19025 * GNAT.IO (g-io.ads)::
19026 * GNAT.IO_Aux (g-io_aux.ads)::
19027 * GNAT.Lock_Files (g-locfil.ads)::
19028 * GNAT.MBBS_Discrete_Random (g-mbdira.ads)::
19029 * GNAT.MBBS_Float_Random (g-mbflra.ads)::
19030 * GNAT.MD5 (g-md5.ads)::
19031 * GNAT.Memory_Dump (g-memdum.ads)::
19032 * GNAT.Most_Recent_Exception (g-moreex.ads)::
19033 * GNAT.OS_Lib (g-os_lib.ads)::
19034 * GNAT.Perfect_Hash_Generators (g-pehage.ads)::
19035 * GNAT.Random_Numbers (g-rannum.ads)::
19036 * GNAT.Regexp (g-regexp.ads)::
19037 * GNAT.Registry (g-regist.ads)::
19038 * GNAT.Regpat (g-regpat.ads)::
19039 * GNAT.Rewrite_Data (g-rewdat.ads)::
19040 * GNAT.Secondary_Stack_Info (g-sestin.ads)::
19041 * GNAT.Semaphores (g-semaph.ads)::
19042 * GNAT.Serial_Communications (g-sercom.ads)::
19043 * GNAT.SHA1 (g-sha1.ads)::
19044 * GNAT.SHA224 (g-sha224.ads)::
19045 * GNAT.SHA256 (g-sha256.ads)::
19046 * GNAT.SHA384 (g-sha384.ads)::
19047 * GNAT.SHA512 (g-sha512.ads)::
19048 * GNAT.Signals (g-signal.ads)::
19049 * GNAT.Sockets (g-socket.ads)::
19050 * GNAT.Source_Info (g-souinf.ads)::
19051 * GNAT.Spelling_Checker (g-speche.ads)::
19052 * GNAT.Spelling_Checker_Generic (g-spchge.ads)::
19053 * GNAT.Spitbol.Patterns (g-spipat.ads)::
19054 * GNAT.Spitbol (g-spitbo.ads)::
19055 * GNAT.Spitbol.Table_Boolean (g-sptabo.ads)::
19056 * GNAT.Spitbol.Table_Integer (g-sptain.ads)::
19057 * GNAT.Spitbol.Table_VString (g-sptavs.ads)::
19058 * GNAT.SSE (g-sse.ads)::
19059 * GNAT.SSE.Vector_Types (g-ssvety.ads)::
19060 * GNAT.Strings (g-string.ads)::
19061 * GNAT.String_Split (g-strspl.ads)::
19062 * GNAT.Table (g-table.ads)::
19063 * GNAT.Task_Lock (g-tasloc.ads)::
19064 * GNAT.Threads (g-thread.ads)::
19065 * GNAT.Time_Stamp (g-timsta.ads)::
19066 * GNAT.Traceback (g-traceb.ads)::
19067 * GNAT.Traceback.Symbolic (g-trasym.ads)::
19068 * GNAT.UTF_32 (g-utf_32.ads)::
19069 * GNAT.UTF_32_Spelling_Checker (g-u3spch.ads)::
19070 * GNAT.Wide_Spelling_Checker (g-wispch.ads)::
19071 * GNAT.Wide_String_Split (g-wistsp.ads)::
19072 * GNAT.Wide_Wide_Spelling_Checker (g-zspche.ads)::
19073 * GNAT.Wide_Wide_String_Split (g-zistsp.ads)::
19074 * Interfaces.C.Extensions (i-cexten.ads)::
19075 * Interfaces.C.Streams (i-cstrea.ads)::
19076 * Interfaces.Packed_Decimal (i-pacdec.ads)::
19077 * Interfaces.VxWorks (i-vxwork.ads)::
19078 * Interfaces.VxWorks.IO (i-vxwoio.ads)::
19079 * System.Address_Image (s-addima.ads)::
19080 * System.Assertions (s-assert.ads)::
19081 * System.Atomic_Counters (s-atocou.ads)::
19082 * System.Memory (s-memory.ads)::
19083 * System.Multiprocessors (s-multip.ads)::
19084 * System.Multiprocessors.Dispatching_Domains (s-mudido.ads)::
19085 * System.Partition_Interface (s-parint.ads)::
19086 * System.Pool_Global (s-pooglo.ads)::
19087 * System.Pool_Local (s-pooloc.ads)::
19088 * System.Restrictions (s-restri.ads)::
19089 * System.Rident (s-rident.ads)::
19090 * System.Strings.Stream_Ops (s-ststop.ads)::
19091 * System.Unsigned_Types (s-unstyp.ads)::
19092 * System.Wch_Cnv (s-wchcnv.ads)::
19093 * System.Wch_Con (s-wchcon.ads)::
19096 @node Ada.Characters.Latin_9 (a-chlat9.ads)
19097 @section @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
19098 @cindex @code{Ada.Characters.Latin_9} (@file{a-chlat9.ads})
19099 @cindex Latin_9 constants for Character
19102 This child of @code{Ada.Characters}
19103 provides a set of definitions corresponding to those in the
19104 RM-defined package @code{Ada.Characters.Latin_1} but with the
19105 few modifications required for @code{Latin-9}
19106 The provision of such a package
19107 is specifically authorized by the Ada Reference Manual
19110 @node Ada.Characters.Wide_Latin_1 (a-cwila1.ads)
19111 @section @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
19112 @cindex @code{Ada.Characters.Wide_Latin_1} (@file{a-cwila1.ads})
19113 @cindex Latin_1 constants for Wide_Character
19116 This child of @code{Ada.Characters}
19117 provides a set of definitions corresponding to those in the
19118 RM-defined package @code{Ada.Characters.Latin_1} but with the
19119 types of the constants being @code{Wide_Character}
19120 instead of @code{Character}. The provision of such a package
19121 is specifically authorized by the Ada Reference Manual
19124 @node Ada.Characters.Wide_Latin_9 (a-cwila9.ads)
19125 @section @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
19126 @cindex @code{Ada.Characters.Wide_Latin_9} (@file{a-cwila1.ads})
19127 @cindex Latin_9 constants for Wide_Character
19130 This child of @code{Ada.Characters}
19131 provides a set of definitions corresponding to those in the
19132 GNAT defined package @code{Ada.Characters.Latin_9} but with the
19133 types of the constants being @code{Wide_Character}
19134 instead of @code{Character}. The provision of such a package
19135 is specifically authorized by the Ada Reference Manual
19138 @node Ada.Characters.Wide_Wide_Latin_1 (a-chzla1.ads)
19139 @section @code{Ada.Characters.Wide_Wide_Latin_1} (@file{a-chzla1.ads})
19140 @cindex @code{Ada.Characters.Wide_Wide_Latin_1} (@file{a-chzla1.ads})
19141 @cindex Latin_1 constants for Wide_Wide_Character
19144 This child of @code{Ada.Characters}
19145 provides a set of definitions corresponding to those in the
19146 RM-defined package @code{Ada.Characters.Latin_1} but with the
19147 types of the constants being @code{Wide_Wide_Character}
19148 instead of @code{Character}. The provision of such a package
19149 is specifically authorized by the Ada Reference Manual
19152 @node Ada.Characters.Wide_Wide_Latin_9 (a-chzla9.ads)
19153 @section @code{Ada.Characters.Wide_Wide_Latin_9} (@file{a-chzla9.ads})
19154 @cindex @code{Ada.Characters.Wide_Wide_Latin_9} (@file{a-chzla9.ads})
19155 @cindex Latin_9 constants for Wide_Wide_Character
19158 This child of @code{Ada.Characters}
19159 provides a set of definitions corresponding to those in the
19160 GNAT defined package @code{Ada.Characters.Latin_9} but with the
19161 types of the constants being @code{Wide_Wide_Character}
19162 instead of @code{Character}. The provision of such a package
19163 is specifically authorized by the Ada Reference Manual
19166 @node Ada.Containers.Formal_Doubly_Linked_Lists (a-cfdlli.ads)
19167 @section @code{Ada.Containers.Formal_Doubly_Linked_Lists} (@file{a-cfdlli.ads})
19168 @cindex @code{Ada.Containers.Formal_Doubly_Linked_Lists} (@file{a-cfdlli.ads})
19169 @cindex Formal container for doubly linked lists
19172 This child of @code{Ada.Containers} defines a modified version of the
19173 Ada 2005 container for doubly linked lists, meant to facilitate formal
19174 verification of code using such containers. The specification of this
19175 unit is compatible with SPARK 2014.
19177 Note that although this container was designed with formal verification
19178 in mind, it may well be generally useful in that it is a simplified more
19179 efficient version than the one defined in the standard. In particular it
19180 does not have the complex overhead required to detect cursor tampering.
19182 @node Ada.Containers.Formal_Hashed_Maps (a-cfhama.ads)
19183 @section @code{Ada.Containers.Formal_Hashed_Maps} (@file{a-cfhama.ads})
19184 @cindex @code{Ada.Containers.Formal_Hashed_Maps} (@file{a-cfhama.ads})
19185 @cindex Formal container for hashed maps
19188 This child of @code{Ada.Containers} defines a modified version of the
19189 Ada 2005 container for hashed maps, meant to facilitate formal
19190 verification of code using such containers. The specification of this
19191 unit is compatible with SPARK 2014.
19193 Note that although this container was designed with formal verification
19194 in mind, it may well be generally useful in that it is a simplified more
19195 efficient version than the one defined in the standard. In particular it
19196 does not have the complex overhead required to detect cursor tampering.
19198 @node Ada.Containers.Formal_Hashed_Sets (a-cfhase.ads)
19199 @section @code{Ada.Containers.Formal_Hashed_Sets} (@file{a-cfhase.ads})
19200 @cindex @code{Ada.Containers.Formal_Hashed_Sets} (@file{a-cfhase.ads})
19201 @cindex Formal container for hashed sets
19204 This child of @code{Ada.Containers} defines a modified version of the
19205 Ada 2005 container for hashed sets, meant to facilitate formal
19206 verification of code using such containers. The specification of this
19207 unit is compatible with SPARK 2014.
19209 Note that although this container was designed with formal verification
19210 in mind, it may well be generally useful in that it is a simplified more
19211 efficient version than the one defined in the standard. In particular it
19212 does not have the complex overhead required to detect cursor tampering.
19214 @node Ada.Containers.Formal_Ordered_Maps (a-cforma.ads)
19215 @section @code{Ada.Containers.Formal_Ordered_Maps} (@file{a-cforma.ads})
19216 @cindex @code{Ada.Containers.Formal_Ordered_Maps} (@file{a-cforma.ads})
19217 @cindex Formal container for ordered maps
19220 This child of @code{Ada.Containers} defines a modified version of the
19221 Ada 2005 container for ordered maps, meant to facilitate formal
19222 verification of code using such containers. The specification of this
19223 unit is compatible with SPARK 2014.
19225 Note that although this container was designed with formal verification
19226 in mind, it may well be generally useful in that it is a simplified more
19227 efficient version than the one defined in the standard. In particular it
19228 does not have the complex overhead required to detect cursor tampering.
19230 @node Ada.Containers.Formal_Ordered_Sets (a-cforse.ads)
19231 @section @code{Ada.Containers.Formal_Ordered_Sets} (@file{a-cforse.ads})
19232 @cindex @code{Ada.Containers.Formal_Ordered_Sets} (@file{a-cforse.ads})
19233 @cindex Formal container for ordered sets
19236 This child of @code{Ada.Containers} defines a modified version of the
19237 Ada 2005 container for ordered sets, meant to facilitate formal
19238 verification of code using such containers. The specification of this
19239 unit is compatible with SPARK 2014.
19241 Note that although this container was designed with formal verification
19242 in mind, it may well be generally useful in that it is a simplified more
19243 efficient version than the one defined in the standard. In particular it
19244 does not have the complex overhead required to detect cursor tampering.
19246 @node Ada.Containers.Formal_Vectors (a-cofove.ads)
19247 @section @code{Ada.Containers.Formal_Vectors} (@file{a-cofove.ads})
19248 @cindex @code{Ada.Containers.Formal_Vectors} (@file{a-cofove.ads})
19249 @cindex Formal container for vectors
19252 This child of @code{Ada.Containers} defines a modified version of the
19253 Ada 2005 container for vectors, meant to facilitate formal
19254 verification of code using such containers. The specification of this
19255 unit is compatible with SPARK 2014.
19257 Note that although this container was designed with formal verification
19258 in mind, it may well be generally useful in that it is a simplified more
19259 efficient version than the one defined in the standard. In particular it
19260 does not have the complex overhead required to detect cursor tampering.
19262 @node Ada.Command_Line.Environment (a-colien.ads)
19263 @section @code{Ada.Command_Line.Environment} (@file{a-colien.ads})
19264 @cindex @code{Ada.Command_Line.Environment} (@file{a-colien.ads})
19265 @cindex Environment entries
19268 This child of @code{Ada.Command_Line}
19269 provides a mechanism for obtaining environment values on systems
19270 where this concept makes sense.
19272 @node Ada.Command_Line.Remove (a-colire.ads)
19273 @section @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
19274 @cindex @code{Ada.Command_Line.Remove} (@file{a-colire.ads})
19275 @cindex Removing command line arguments
19276 @cindex Command line, argument removal
19279 This child of @code{Ada.Command_Line}
19280 provides a mechanism for logically removing
19281 arguments from the argument list. Once removed, an argument is not visible
19282 to further calls on the subprograms in @code{Ada.Command_Line} will not
19283 see the removed argument.
19285 @node Ada.Command_Line.Response_File (a-clrefi.ads)
19286 @section @code{Ada.Command_Line.Response_File} (@file{a-clrefi.ads})
19287 @cindex @code{Ada.Command_Line.Response_File} (@file{a-clrefi.ads})
19288 @cindex Response file for command line
19289 @cindex Command line, response file
19290 @cindex Command line, handling long command lines
19293 This child of @code{Ada.Command_Line} provides a mechanism facilities for
19294 getting command line arguments from a text file, called a "response file".
19295 Using a response file allow passing a set of arguments to an executable longer
19296 than the maximum allowed by the system on the command line.
19298 @node Ada.Direct_IO.C_Streams (a-diocst.ads)
19299 @section @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
19300 @cindex @code{Ada.Direct_IO.C_Streams} (@file{a-diocst.ads})
19301 @cindex C Streams, Interfacing with Direct_IO
19304 This package provides subprograms that allow interfacing between
19305 C streams and @code{Direct_IO}. The stream identifier can be
19306 extracted from a file opened on the Ada side, and an Ada file
19307 can be constructed from a stream opened on the C side.
19309 @node Ada.Exceptions.Is_Null_Occurrence (a-einuoc.ads)
19310 @section @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
19311 @cindex @code{Ada.Exceptions.Is_Null_Occurrence} (@file{a-einuoc.ads})
19312 @cindex Null_Occurrence, testing for
19315 This child subprogram provides a way of testing for the null
19316 exception occurrence (@code{Null_Occurrence}) without raising
19319 @node Ada.Exceptions.Last_Chance_Handler (a-elchha.ads)
19320 @section @code{Ada.Exceptions.Last_Chance_Handler} (@file{a-elchha.ads})
19321 @cindex @code{Ada.Exceptions.Last_Chance_Handler} (@file{a-elchha.ads})
19322 @cindex Null_Occurrence, testing for
19325 This child subprogram is used for handling otherwise unhandled
19326 exceptions (hence the name last chance), and perform clean ups before
19327 terminating the program. Note that this subprogram never returns.
19329 @node Ada.Exceptions.Traceback (a-exctra.ads)
19330 @section @code{Ada.Exceptions.Traceback} (@file{a-exctra.ads})
19331 @cindex @code{Ada.Exceptions.Traceback} (@file{a-exctra.ads})
19332 @cindex Traceback for Exception Occurrence
19335 This child package provides the subprogram (@code{Tracebacks}) to
19336 give a traceback array of addresses based on an exception
19339 @node Ada.Sequential_IO.C_Streams (a-siocst.ads)
19340 @section @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
19341 @cindex @code{Ada.Sequential_IO.C_Streams} (@file{a-siocst.ads})
19342 @cindex C Streams, Interfacing with Sequential_IO
19345 This package provides subprograms that allow interfacing between
19346 C streams and @code{Sequential_IO}. The stream identifier can be
19347 extracted from a file opened on the Ada side, and an Ada file
19348 can be constructed from a stream opened on the C side.
19350 @node Ada.Streams.Stream_IO.C_Streams (a-ssicst.ads)
19351 @section @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
19352 @cindex @code{Ada.Streams.Stream_IO.C_Streams} (@file{a-ssicst.ads})
19353 @cindex C Streams, Interfacing with Stream_IO
19356 This package provides subprograms that allow interfacing between
19357 C streams and @code{Stream_IO}. The stream identifier can be
19358 extracted from a file opened on the Ada side, and an Ada file
19359 can be constructed from a stream opened on the C side.
19361 @node Ada.Strings.Unbounded.Text_IO (a-suteio.ads)
19362 @section @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
19363 @cindex @code{Ada.Strings.Unbounded.Text_IO} (@file{a-suteio.ads})
19364 @cindex @code{Unbounded_String}, IO support
19365 @cindex @code{Text_IO}, extensions for unbounded strings
19368 This package provides subprograms for Text_IO for unbounded
19369 strings, avoiding the necessity for an intermediate operation
19370 with ordinary strings.
19372 @node Ada.Strings.Wide_Unbounded.Wide_Text_IO (a-swuwti.ads)
19373 @section @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
19374 @cindex @code{Ada.Strings.Wide_Unbounded.Wide_Text_IO} (@file{a-swuwti.ads})
19375 @cindex @code{Unbounded_Wide_String}, IO support
19376 @cindex @code{Text_IO}, extensions for unbounded wide strings
19379 This package provides subprograms for Text_IO for unbounded
19380 wide strings, avoiding the necessity for an intermediate operation
19381 with ordinary wide strings.
19383 @node Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO (a-szuzti.ads)
19384 @section @code{Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO} (@file{a-szuzti.ads})
19385 @cindex @code{Ada.Strings.Wide_Wide_Unbounded.Wide_Wide_Text_IO} (@file{a-szuzti.ads})
19386 @cindex @code{Unbounded_Wide_Wide_String}, IO support
19387 @cindex @code{Text_IO}, extensions for unbounded wide wide strings
19390 This package provides subprograms for Text_IO for unbounded
19391 wide wide strings, avoiding the necessity for an intermediate operation
19392 with ordinary wide wide strings.
19394 @node Ada.Text_IO.C_Streams (a-tiocst.ads)
19395 @section @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
19396 @cindex @code{Ada.Text_IO.C_Streams} (@file{a-tiocst.ads})
19397 @cindex C Streams, Interfacing with @code{Text_IO}
19400 This package provides subprograms that allow interfacing between
19401 C streams and @code{Text_IO}. The stream identifier can be
19402 extracted from a file opened on the Ada side, and an Ada file
19403 can be constructed from a stream opened on the C side.
19405 @node Ada.Text_IO.Reset_Standard_Files (a-tirsfi.ads)
19406 @section @code{Ada.Text_IO.Reset_Standard_Files} (@file{a-tirsfi.ads})
19407 @cindex @code{Ada.Text_IO.Reset_Standard_Files} (@file{a-tirsfi.ads})
19408 @cindex @code{Text_IO} resetting standard files
19411 This procedure is used to reset the status of the standard files used
19412 by Ada.Text_IO. This is useful in a situation (such as a restart in an
19413 embedded application) where the status of the files may change during
19414 execution (for example a standard input file may be redefined to be
19417 @node Ada.Wide_Characters.Unicode (a-wichun.ads)
19418 @section @code{Ada.Wide_Characters.Unicode} (@file{a-wichun.ads})
19419 @cindex @code{Ada.Wide_Characters.Unicode} (@file{a-wichun.ads})
19420 @cindex Unicode categorization, Wide_Character
19423 This package provides subprograms that allow categorization of
19424 Wide_Character values according to Unicode categories.
19426 @node Ada.Wide_Text_IO.C_Streams (a-wtcstr.ads)
19427 @section @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
19428 @cindex @code{Ada.Wide_Text_IO.C_Streams} (@file{a-wtcstr.ads})
19429 @cindex C Streams, Interfacing with @code{Wide_Text_IO}
19432 This package provides subprograms that allow interfacing between
19433 C streams and @code{Wide_Text_IO}. The stream identifier can be
19434 extracted from a file opened on the Ada side, and an Ada file
19435 can be constructed from a stream opened on the C side.
19437 @node Ada.Wide_Text_IO.Reset_Standard_Files (a-wrstfi.ads)
19438 @section @code{Ada.Wide_Text_IO.Reset_Standard_Files} (@file{a-wrstfi.ads})
19439 @cindex @code{Ada.Wide_Text_IO.Reset_Standard_Files} (@file{a-wrstfi.ads})
19440 @cindex @code{Wide_Text_IO} resetting standard files
19443 This procedure is used to reset the status of the standard files used
19444 by Ada.Wide_Text_IO. This is useful in a situation (such as a restart in an
19445 embedded application) where the status of the files may change during
19446 execution (for example a standard input file may be redefined to be
19449 @node Ada.Wide_Wide_Characters.Unicode (a-zchuni.ads)
19450 @section @code{Ada.Wide_Wide_Characters.Unicode} (@file{a-zchuni.ads})
19451 @cindex @code{Ada.Wide_Wide_Characters.Unicode} (@file{a-zchuni.ads})
19452 @cindex Unicode categorization, Wide_Wide_Character
19455 This package provides subprograms that allow categorization of
19456 Wide_Wide_Character values according to Unicode categories.
19458 @node Ada.Wide_Wide_Text_IO.C_Streams (a-ztcstr.ads)
19459 @section @code{Ada.Wide_Wide_Text_IO.C_Streams} (@file{a-ztcstr.ads})
19460 @cindex @code{Ada.Wide_Wide_Text_IO.C_Streams} (@file{a-ztcstr.ads})
19461 @cindex C Streams, Interfacing with @code{Wide_Wide_Text_IO}
19464 This package provides subprograms that allow interfacing between
19465 C streams and @code{Wide_Wide_Text_IO}. The stream identifier can be
19466 extracted from a file opened on the Ada side, and an Ada file
19467 can be constructed from a stream opened on the C side.
19469 @node Ada.Wide_Wide_Text_IO.Reset_Standard_Files (a-zrstfi.ads)
19470 @section @code{Ada.Wide_Wide_Text_IO.Reset_Standard_Files} (@file{a-zrstfi.ads})
19471 @cindex @code{Ada.Wide_Wide_Text_IO.Reset_Standard_Files} (@file{a-zrstfi.ads})
19472 @cindex @code{Wide_Wide_Text_IO} resetting standard files
19475 This procedure is used to reset the status of the standard files used
19476 by Ada.Wide_Wide_Text_IO. This is useful in a situation (such as a
19477 restart in an embedded application) where the status of the files may
19478 change during execution (for example a standard input file may be
19479 redefined to be interactive).
19481 @node GNAT.Altivec (g-altive.ads)
19482 @section @code{GNAT.Altivec} (@file{g-altive.ads})
19483 @cindex @code{GNAT.Altivec} (@file{g-altive.ads})
19487 This is the root package of the GNAT AltiVec binding. It provides
19488 definitions of constants and types common to all the versions of the
19491 @node GNAT.Altivec.Conversions (g-altcon.ads)
19492 @section @code{GNAT.Altivec.Conversions} (@file{g-altcon.ads})
19493 @cindex @code{GNAT.Altivec.Conversions} (@file{g-altcon.ads})
19497 This package provides the Vector/View conversion routines.
19499 @node GNAT.Altivec.Vector_Operations (g-alveop.ads)
19500 @section @code{GNAT.Altivec.Vector_Operations} (@file{g-alveop.ads})
19501 @cindex @code{GNAT.Altivec.Vector_Operations} (@file{g-alveop.ads})
19505 This package exposes the Ada interface to the AltiVec operations on
19506 vector objects. A soft emulation is included by default in the GNAT
19507 library. The hard binding is provided as a separate package. This unit
19508 is common to both bindings.
19510 @node GNAT.Altivec.Vector_Types (g-alvety.ads)
19511 @section @code{GNAT.Altivec.Vector_Types} (@file{g-alvety.ads})
19512 @cindex @code{GNAT.Altivec.Vector_Types} (@file{g-alvety.ads})
19516 This package exposes the various vector types part of the Ada binding
19517 to AltiVec facilities.
19519 @node GNAT.Altivec.Vector_Views (g-alvevi.ads)
19520 @section @code{GNAT.Altivec.Vector_Views} (@file{g-alvevi.ads})
19521 @cindex @code{GNAT.Altivec.Vector_Views} (@file{g-alvevi.ads})
19525 This package provides public 'View' data types from/to which private
19526 vector representations can be converted via
19527 GNAT.Altivec.Conversions. This allows convenient access to individual
19528 vector elements and provides a simple way to initialize vector
19531 @node GNAT.Array_Split (g-arrspl.ads)
19532 @section @code{GNAT.Array_Split} (@file{g-arrspl.ads})
19533 @cindex @code{GNAT.Array_Split} (@file{g-arrspl.ads})
19534 @cindex Array splitter
19537 Useful array-manipulation routines: given a set of separators, split
19538 an array wherever the separators appear, and provide direct access
19539 to the resulting slices.
19541 @node GNAT.AWK (g-awk.ads)
19542 @section @code{GNAT.AWK} (@file{g-awk.ads})
19543 @cindex @code{GNAT.AWK} (@file{g-awk.ads})
19548 Provides AWK-like parsing functions, with an easy interface for parsing one
19549 or more files containing formatted data. The file is viewed as a database
19550 where each record is a line and a field is a data element in this line.
19552 @node GNAT.Bounded_Buffers (g-boubuf.ads)
19553 @section @code{GNAT.Bounded_Buffers} (@file{g-boubuf.ads})
19554 @cindex @code{GNAT.Bounded_Buffers} (@file{g-boubuf.ads})
19556 @cindex Bounded Buffers
19559 Provides a concurrent generic bounded buffer abstraction. Instances are
19560 useful directly or as parts of the implementations of other abstractions,
19563 @node GNAT.Bounded_Mailboxes (g-boumai.ads)
19564 @section @code{GNAT.Bounded_Mailboxes} (@file{g-boumai.ads})
19565 @cindex @code{GNAT.Bounded_Mailboxes} (@file{g-boumai.ads})
19570 Provides a thread-safe asynchronous intertask mailbox communication facility.
19572 @node GNAT.Bubble_Sort (g-bubsor.ads)
19573 @section @code{GNAT.Bubble_Sort} (@file{g-bubsor.ads})
19574 @cindex @code{GNAT.Bubble_Sort} (@file{g-bubsor.ads})
19576 @cindex Bubble sort
19579 Provides a general implementation of bubble sort usable for sorting arbitrary
19580 data items. Exchange and comparison procedures are provided by passing
19581 access-to-procedure values.
19583 @node GNAT.Bubble_Sort_A (g-busora.ads)
19584 @section @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
19585 @cindex @code{GNAT.Bubble_Sort_A} (@file{g-busora.ads})
19587 @cindex Bubble sort
19590 Provides a general implementation of bubble sort usable for sorting arbitrary
19591 data items. Move and comparison procedures are provided by passing
19592 access-to-procedure values. This is an older version, retained for
19593 compatibility. Usually @code{GNAT.Bubble_Sort} will be preferable.
19595 @node GNAT.Bubble_Sort_G (g-busorg.ads)
19596 @section @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
19597 @cindex @code{GNAT.Bubble_Sort_G} (@file{g-busorg.ads})
19599 @cindex Bubble sort
19602 Similar to @code{Bubble_Sort_A} except that the move and sorting procedures
19603 are provided as generic parameters, this improves efficiency, especially
19604 if the procedures can be inlined, at the expense of duplicating code for
19605 multiple instantiations.
19607 @node GNAT.Byte_Order_Mark (g-byorma.ads)
19608 @section @code{GNAT.Byte_Order_Mark} (@file{g-byorma.ads})
19609 @cindex @code{GNAT.Byte_Order_Mark} (@file{g-byorma.ads})
19610 @cindex UTF-8 representation
19611 @cindex Wide characte representations
19614 Provides a routine which given a string, reads the start of the string to
19615 see whether it is one of the standard byte order marks (BOM's) which signal
19616 the encoding of the string. The routine includes detection of special XML
19617 sequences for various UCS input formats.
19619 @node GNAT.Byte_Swapping (g-bytswa.ads)
19620 @section @code{GNAT.Byte_Swapping} (@file{g-bytswa.ads})
19621 @cindex @code{GNAT.Byte_Swapping} (@file{g-bytswa.ads})
19622 @cindex Byte swapping
19626 General routines for swapping the bytes in 2-, 4-, and 8-byte quantities.
19627 Machine-specific implementations are available in some cases.
19629 @node GNAT.Calendar (g-calend.ads)
19630 @section @code{GNAT.Calendar} (@file{g-calend.ads})
19631 @cindex @code{GNAT.Calendar} (@file{g-calend.ads})
19632 @cindex @code{Calendar}
19635 Extends the facilities provided by @code{Ada.Calendar} to include handling
19636 of days of the week, an extended @code{Split} and @code{Time_Of} capability.
19637 Also provides conversion of @code{Ada.Calendar.Time} values to and from the
19638 C @code{timeval} format.
19640 @node GNAT.Calendar.Time_IO (g-catiio.ads)
19641 @section @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
19642 @cindex @code{Calendar}
19644 @cindex @code{GNAT.Calendar.Time_IO} (@file{g-catiio.ads})
19646 @node GNAT.Case_Util (g-casuti.ads)
19647 @section @code{GNAT.Case_Util} (@file{g-casuti.ads})
19648 @cindex @code{GNAT.Case_Util} (@file{g-casuti.ads})
19649 @cindex Casing utilities
19650 @cindex Character handling (@code{GNAT.Case_Util})
19653 A set of simple routines for handling upper and lower casing of strings
19654 without the overhead of the full casing tables
19655 in @code{Ada.Characters.Handling}.
19657 @node GNAT.CGI (g-cgi.ads)
19658 @section @code{GNAT.CGI} (@file{g-cgi.ads})
19659 @cindex @code{GNAT.CGI} (@file{g-cgi.ads})
19660 @cindex CGI (Common Gateway Interface)
19663 This is a package for interfacing a GNAT program with a Web server via the
19664 Common Gateway Interface (CGI)@. Basically this package parses the CGI
19665 parameters, which are a set of key/value pairs sent by the Web server. It
19666 builds a table whose index is the key and provides some services to deal
19669 @node GNAT.CGI.Cookie (g-cgicoo.ads)
19670 @section @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
19671 @cindex @code{GNAT.CGI.Cookie} (@file{g-cgicoo.ads})
19672 @cindex CGI (Common Gateway Interface) cookie support
19673 @cindex Cookie support in CGI
19676 This is a package to interface a GNAT program with a Web server via the
19677 Common Gateway Interface (CGI). It exports services to deal with Web
19678 cookies (piece of information kept in the Web client software).
19680 @node GNAT.CGI.Debug (g-cgideb.ads)
19681 @section @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
19682 @cindex @code{GNAT.CGI.Debug} (@file{g-cgideb.ads})
19683 @cindex CGI (Common Gateway Interface) debugging
19686 This is a package to help debugging CGI (Common Gateway Interface)
19687 programs written in Ada.
19689 @node GNAT.Command_Line (g-comlin.ads)
19690 @section @code{GNAT.Command_Line} (@file{g-comlin.ads})
19691 @cindex @code{GNAT.Command_Line} (@file{g-comlin.ads})
19692 @cindex Command line
19695 Provides a high level interface to @code{Ada.Command_Line} facilities,
19696 including the ability to scan for named switches with optional parameters
19697 and expand file names using wild card notations.
19699 @node GNAT.Compiler_Version (g-comver.ads)
19700 @section @code{GNAT.Compiler_Version} (@file{g-comver.ads})
19701 @cindex @code{GNAT.Compiler_Version} (@file{g-comver.ads})
19702 @cindex Compiler Version
19703 @cindex Version, of compiler
19706 Provides a routine for obtaining the version of the compiler used to
19707 compile the program. More accurately this is the version of the binder
19708 used to bind the program (this will normally be the same as the version
19709 of the compiler if a consistent tool set is used to compile all units
19712 @node GNAT.Ctrl_C (g-ctrl_c.ads)
19713 @section @code{GNAT.Ctrl_C} (@file{g-ctrl_c.ads})
19714 @cindex @code{GNAT.Ctrl_C} (@file{g-ctrl_c.ads})
19718 Provides a simple interface to handle Ctrl-C keyboard events.
19720 @node GNAT.CRC32 (g-crc32.ads)
19721 @section @code{GNAT.CRC32} (@file{g-crc32.ads})
19722 @cindex @code{GNAT.CRC32} (@file{g-crc32.ads})
19724 @cindex Cyclic Redundancy Check
19727 This package implements the CRC-32 algorithm. For a full description
19728 of this algorithm see
19729 ``Computation of Cyclic Redundancy Checks via Table Look-Up'',
19730 @cite{Communications of the ACM}, Vol.@: 31 No.@: 8, pp.@: 1008-1013,
19731 Aug.@: 1988. Sarwate, D.V@.
19733 @node GNAT.Current_Exception (g-curexc.ads)
19734 @section @code{GNAT.Current_Exception} (@file{g-curexc.ads})
19735 @cindex @code{GNAT.Current_Exception} (@file{g-curexc.ads})
19736 @cindex Current exception
19737 @cindex Exception retrieval
19740 Provides access to information on the current exception that has been raised
19741 without the need for using the Ada 95 / Ada 2005 exception choice parameter
19742 specification syntax.
19743 This is particularly useful in simulating typical facilities for
19744 obtaining information about exceptions provided by Ada 83 compilers.
19746 @node GNAT.Debug_Pools (g-debpoo.ads)
19747 @section @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
19748 @cindex @code{GNAT.Debug_Pools} (@file{g-debpoo.ads})
19750 @cindex Debug pools
19751 @cindex Memory corruption debugging
19754 Provide a debugging storage pools that helps tracking memory corruption
19755 problems. @xref{The GNAT Debug Pool Facility,,, gnat_ugn,
19756 @value{EDITION} User's Guide}.
19758 @node GNAT.Debug_Utilities (g-debuti.ads)
19759 @section @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
19760 @cindex @code{GNAT.Debug_Utilities} (@file{g-debuti.ads})
19764 Provides a few useful utilities for debugging purposes, including conversion
19765 to and from string images of address values. Supports both C and Ada formats
19766 for hexadecimal literals.
19768 @node GNAT.Decode_String (g-decstr.ads)
19769 @section @code{GNAT.Decode_String} (@file{g-decstr.ads})
19770 @cindex @code{GNAT.Decode_String} (@file{g-decstr.ads})
19771 @cindex Decoding strings
19772 @cindex String decoding
19773 @cindex Wide character encoding
19778 A generic package providing routines for decoding wide character and wide wide
19779 character strings encoded as sequences of 8-bit characters using a specified
19780 encoding method. Includes validation routines, and also routines for stepping
19781 to next or previous encoded character in an encoded string.
19782 Useful in conjunction with Unicode character coding. Note there is a
19783 preinstantiation for UTF-8. See next entry.
19785 @node GNAT.Decode_UTF8_String (g-deutst.ads)
19786 @section @code{GNAT.Decode_UTF8_String} (@file{g-deutst.ads})
19787 @cindex @code{GNAT.Decode_UTF8_String} (@file{g-deutst.ads})
19788 @cindex Decoding strings
19789 @cindex Decoding UTF-8 strings
19790 @cindex UTF-8 string decoding
19791 @cindex Wide character decoding
19796 A preinstantiation of GNAT.Decode_Strings for UTF-8 encoding.
19798 @node GNAT.Directory_Operations (g-dirope.ads)
19799 @section @code{GNAT.Directory_Operations} (@file{g-dirope.ads})
19800 @cindex @code{GNAT.Directory_Operations} (@file{g-dirope.ads})
19801 @cindex Directory operations
19804 Provides a set of routines for manipulating directories, including changing
19805 the current directory, making new directories, and scanning the files in a
19808 @node GNAT.Directory_Operations.Iteration (g-diopit.ads)
19809 @section @code{GNAT.Directory_Operations.Iteration} (@file{g-diopit.ads})
19810 @cindex @code{GNAT.Directory_Operations.Iteration} (@file{g-diopit.ads})
19811 @cindex Directory operations iteration
19814 A child unit of GNAT.Directory_Operations providing additional operations
19815 for iterating through directories.
19817 @node GNAT.Dynamic_HTables (g-dynhta.ads)
19818 @section @code{GNAT.Dynamic_HTables} (@file{g-dynhta.ads})
19819 @cindex @code{GNAT.Dynamic_HTables} (@file{g-dynhta.ads})
19820 @cindex Hash tables
19823 A generic implementation of hash tables that can be used to hash arbitrary
19824 data. Provided in two forms, a simple form with built in hash functions,
19825 and a more complex form in which the hash function is supplied.
19828 This package provides a facility similar to that of @code{GNAT.HTable},
19829 except that this package declares a type that can be used to define
19830 dynamic instances of the hash table, while an instantiation of
19831 @code{GNAT.HTable} creates a single instance of the hash table.
19833 @node GNAT.Dynamic_Tables (g-dyntab.ads)
19834 @section @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
19835 @cindex @code{GNAT.Dynamic_Tables} (@file{g-dyntab.ads})
19836 @cindex Table implementation
19837 @cindex Arrays, extendable
19840 A generic package providing a single dimension array abstraction where the
19841 length of the array can be dynamically modified.
19844 This package provides a facility similar to that of @code{GNAT.Table},
19845 except that this package declares a type that can be used to define
19846 dynamic instances of the table, while an instantiation of
19847 @code{GNAT.Table} creates a single instance of the table type.
19849 @node GNAT.Encode_String (g-encstr.ads)
19850 @section @code{GNAT.Encode_String} (@file{g-encstr.ads})
19851 @cindex @code{GNAT.Encode_String} (@file{g-encstr.ads})
19852 @cindex Encoding strings
19853 @cindex String encoding
19854 @cindex Wide character encoding
19859 A generic package providing routines for encoding wide character and wide
19860 wide character strings as sequences of 8-bit characters using a specified
19861 encoding method. Useful in conjunction with Unicode character coding.
19862 Note there is a preinstantiation for UTF-8. See next entry.
19864 @node GNAT.Encode_UTF8_String (g-enutst.ads)
19865 @section @code{GNAT.Encode_UTF8_String} (@file{g-enutst.ads})
19866 @cindex @code{GNAT.Encode_UTF8_String} (@file{g-enutst.ads})
19867 @cindex Encoding strings
19868 @cindex Encoding UTF-8 strings
19869 @cindex UTF-8 string encoding
19870 @cindex Wide character encoding
19875 A preinstantiation of GNAT.Encode_Strings for UTF-8 encoding.
19877 @node GNAT.Exception_Actions (g-excact.ads)
19878 @section @code{GNAT.Exception_Actions} (@file{g-excact.ads})
19879 @cindex @code{GNAT.Exception_Actions} (@file{g-excact.ads})
19880 @cindex Exception actions
19883 Provides callbacks when an exception is raised. Callbacks can be registered
19884 for specific exceptions, or when any exception is raised. This
19885 can be used for instance to force a core dump to ease debugging.
19887 @node GNAT.Exception_Traces (g-exctra.ads)
19888 @section @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
19889 @cindex @code{GNAT.Exception_Traces} (@file{g-exctra.ads})
19890 @cindex Exception traces
19894 Provides an interface allowing to control automatic output upon exception
19897 @node GNAT.Exceptions (g-except.ads)
19898 @section @code{GNAT.Exceptions} (@file{g-expect.ads})
19899 @cindex @code{GNAT.Exceptions} (@file{g-expect.ads})
19900 @cindex Exceptions, Pure
19901 @cindex Pure packages, exceptions
19904 Normally it is not possible to raise an exception with
19905 a message from a subprogram in a pure package, since the
19906 necessary types and subprograms are in @code{Ada.Exceptions}
19907 which is not a pure unit. @code{GNAT.Exceptions} provides a
19908 facility for getting around this limitation for a few
19909 predefined exceptions, and for example allow raising
19910 @code{Constraint_Error} with a message from a pure subprogram.
19912 @node GNAT.Expect (g-expect.ads)
19913 @section @code{GNAT.Expect} (@file{g-expect.ads})
19914 @cindex @code{GNAT.Expect} (@file{g-expect.ads})
19917 Provides a set of subprograms similar to what is available
19918 with the standard Tcl Expect tool.
19919 It allows you to easily spawn and communicate with an external process.
19920 You can send commands or inputs to the process, and compare the output
19921 with some expected regular expression. Currently @code{GNAT.Expect}
19922 is implemented on all native GNAT ports except for OpenVMS@.
19923 It is not implemented for cross ports, and in particular is not
19924 implemented for VxWorks or LynxOS@.
19926 @node GNAT.Expect.TTY (g-exptty.ads)
19927 @section @code{GNAT.Expect.TTY} (@file{g-exptty.ads})
19928 @cindex @code{GNAT.Expect.TTY} (@file{g-exptty.ads})
19931 As GNAT.Expect but using pseudo-terminal.
19932 Currently @code{GNAT.Expect.TTY} is implemented on all native GNAT
19933 ports except for OpenVMS@. It is not implemented for cross ports, and
19934 in particular is not implemented for VxWorks or LynxOS@.
19936 @node GNAT.Float_Control (g-flocon.ads)
19937 @section @code{GNAT.Float_Control} (@file{g-flocon.ads})
19938 @cindex @code{GNAT.Float_Control} (@file{g-flocon.ads})
19939 @cindex Floating-Point Processor
19942 Provides an interface for resetting the floating-point processor into the
19943 mode required for correct semantic operation in Ada. Some third party
19944 library calls may cause this mode to be modified, and the Reset procedure
19945 in this package can be used to reestablish the required mode.
19947 @node GNAT.Formatted_String (g-forstr.ads)
19948 @section @code{GNAT.Formatted_String} (@file{g-forstr.ads})
19949 @cindex @code{GNAT.Formatted_String} (@file{g-forstr.ads})
19950 @cindex Formatted String
19953 Provides support for C/C++ printf() formatted strings. The format is
19954 copied from the printf() routine and should therefore gives identical
19955 output. Some generic routines are provided to be able to use types
19956 derived from Integer, Float or enumerations as values for the
19959 @node GNAT.Heap_Sort (g-heasor.ads)
19960 @section @code{GNAT.Heap_Sort} (@file{g-heasor.ads})
19961 @cindex @code{GNAT.Heap_Sort} (@file{g-heasor.ads})
19965 Provides a general implementation of heap sort usable for sorting arbitrary
19966 data items. Exchange and comparison procedures are provided by passing
19967 access-to-procedure values. The algorithm used is a modified heap sort
19968 that performs approximately N*log(N) comparisons in the worst case.
19970 @node GNAT.Heap_Sort_A (g-hesora.ads)
19971 @section @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
19972 @cindex @code{GNAT.Heap_Sort_A} (@file{g-hesora.ads})
19976 Provides a general implementation of heap sort usable for sorting arbitrary
19977 data items. Move and comparison procedures are provided by passing
19978 access-to-procedure values. The algorithm used is a modified heap sort
19979 that performs approximately N*log(N) comparisons in the worst case.
19980 This differs from @code{GNAT.Heap_Sort} in having a less convenient
19981 interface, but may be slightly more efficient.
19983 @node GNAT.Heap_Sort_G (g-hesorg.ads)
19984 @section @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
19985 @cindex @code{GNAT.Heap_Sort_G} (@file{g-hesorg.ads})
19989 Similar to @code{Heap_Sort_A} except that the move and sorting procedures
19990 are provided as generic parameters, this improves efficiency, especially
19991 if the procedures can be inlined, at the expense of duplicating code for
19992 multiple instantiations.
19994 @node GNAT.HTable (g-htable.ads)
19995 @section @code{GNAT.HTable} (@file{g-htable.ads})
19996 @cindex @code{GNAT.HTable} (@file{g-htable.ads})
19997 @cindex Hash tables
20000 A generic implementation of hash tables that can be used to hash arbitrary
20001 data. Provides two approaches, one a simple static approach, and the other
20002 allowing arbitrary dynamic hash tables.
20004 @node GNAT.IO (g-io.ads)
20005 @section @code{GNAT.IO} (@file{g-io.ads})
20006 @cindex @code{GNAT.IO} (@file{g-io.ads})
20008 @cindex Input/Output facilities
20011 A simple preelaborable input-output package that provides a subset of
20012 simple Text_IO functions for reading characters and strings from
20013 Standard_Input, and writing characters, strings and integers to either
20014 Standard_Output or Standard_Error.
20016 @node GNAT.IO_Aux (g-io_aux.ads)
20017 @section @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
20018 @cindex @code{GNAT.IO_Aux} (@file{g-io_aux.ads})
20020 @cindex Input/Output facilities
20022 Provides some auxiliary functions for use with Text_IO, including a test
20023 for whether a file exists, and functions for reading a line of text.
20025 @node GNAT.Lock_Files (g-locfil.ads)
20026 @section @code{GNAT.Lock_Files} (@file{g-locfil.ads})
20027 @cindex @code{GNAT.Lock_Files} (@file{g-locfil.ads})
20028 @cindex File locking
20029 @cindex Locking using files
20032 Provides a general interface for using files as locks. Can be used for
20033 providing program level synchronization.
20035 @node GNAT.MBBS_Discrete_Random (g-mbdira.ads)
20036 @section @code{GNAT.MBBS_Discrete_Random} (@file{g-mbdira.ads})
20037 @cindex @code{GNAT.MBBS_Discrete_Random} (@file{g-mbdira.ads})
20038 @cindex Random number generation
20041 The original implementation of @code{Ada.Numerics.Discrete_Random}. Uses
20042 a modified version of the Blum-Blum-Shub generator.
20044 @node GNAT.MBBS_Float_Random (g-mbflra.ads)
20045 @section @code{GNAT.MBBS_Float_Random} (@file{g-mbflra.ads})
20046 @cindex @code{GNAT.MBBS_Float_Random} (@file{g-mbflra.ads})
20047 @cindex Random number generation
20050 The original implementation of @code{Ada.Numerics.Float_Random}. Uses
20051 a modified version of the Blum-Blum-Shub generator.
20053 @node GNAT.MD5 (g-md5.ads)
20054 @section @code{GNAT.MD5} (@file{g-md5.ads})
20055 @cindex @code{GNAT.MD5} (@file{g-md5.ads})
20056 @cindex Message Digest MD5
20059 Implements the MD5 Message-Digest Algorithm as described in RFC 1321, and
20060 the HMAC-MD5 message authentication function as described in RFC 2104 and
20063 @node GNAT.Memory_Dump (g-memdum.ads)
20064 @section @code{GNAT.Memory_Dump} (@file{g-memdum.ads})
20065 @cindex @code{GNAT.Memory_Dump} (@file{g-memdum.ads})
20066 @cindex Dump Memory
20069 Provides a convenient routine for dumping raw memory to either the
20070 standard output or standard error files. Uses GNAT.IO for actual
20073 @node GNAT.Most_Recent_Exception (g-moreex.ads)
20074 @section @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
20075 @cindex @code{GNAT.Most_Recent_Exception} (@file{g-moreex.ads})
20076 @cindex Exception, obtaining most recent
20079 Provides access to the most recently raised exception. Can be used for
20080 various logging purposes, including duplicating functionality of some
20081 Ada 83 implementation dependent extensions.
20083 @node GNAT.OS_Lib (g-os_lib.ads)
20084 @section @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
20085 @cindex @code{GNAT.OS_Lib} (@file{g-os_lib.ads})
20086 @cindex Operating System interface
20087 @cindex Spawn capability
20090 Provides a range of target independent operating system interface functions,
20091 including time/date management, file operations, subprocess management,
20092 including a portable spawn procedure, and access to environment variables
20093 and error return codes.
20095 @node GNAT.Perfect_Hash_Generators (g-pehage.ads)
20096 @section @code{GNAT.Perfect_Hash_Generators} (@file{g-pehage.ads})
20097 @cindex @code{GNAT.Perfect_Hash_Generators} (@file{g-pehage.ads})
20098 @cindex Hash functions
20101 Provides a generator of static minimal perfect hash functions. No
20102 collisions occur and each item can be retrieved from the table in one
20103 probe (perfect property). The hash table size corresponds to the exact
20104 size of the key set and no larger (minimal property). The key set has to
20105 be know in advance (static property). The hash functions are also order
20106 preserving. If w2 is inserted after w1 in the generator, their
20107 hashcode are in the same order. These hashing functions are very
20108 convenient for use with realtime applications.
20110 @node GNAT.Random_Numbers (g-rannum.ads)
20111 @section @code{GNAT.Random_Numbers} (@file{g-rannum.ads})
20112 @cindex @code{GNAT.Random_Numbers} (@file{g-rannum.ads})
20113 @cindex Random number generation
20116 Provides random number capabilities which extend those available in the
20117 standard Ada library and are more convenient to use.
20119 @node GNAT.Regexp (g-regexp.ads)
20120 @section @code{GNAT.Regexp} (@file{g-regexp.ads})
20121 @cindex @code{GNAT.Regexp} (@file{g-regexp.ads})
20122 @cindex Regular expressions
20123 @cindex Pattern matching
20126 A simple implementation of regular expressions, using a subset of regular
20127 expression syntax copied from familiar Unix style utilities. This is the
20128 simplest of the three pattern matching packages provided, and is particularly
20129 suitable for ``file globbing'' applications.
20131 @node GNAT.Registry (g-regist.ads)
20132 @section @code{GNAT.Registry} (@file{g-regist.ads})
20133 @cindex @code{GNAT.Registry} (@file{g-regist.ads})
20134 @cindex Windows Registry
20137 This is a high level binding to the Windows registry. It is possible to
20138 do simple things like reading a key value, creating a new key. For full
20139 registry API, but at a lower level of abstraction, refer to the Win32.Winreg
20140 package provided with the Win32Ada binding
20142 @node GNAT.Regpat (g-regpat.ads)
20143 @section @code{GNAT.Regpat} (@file{g-regpat.ads})
20144 @cindex @code{GNAT.Regpat} (@file{g-regpat.ads})
20145 @cindex Regular expressions
20146 @cindex Pattern matching
20149 A complete implementation of Unix-style regular expression matching, copied
20150 from the original V7 style regular expression library written in C by
20151 Henry Spencer (and binary compatible with this C library).
20153 @node GNAT.Rewrite_Data (g-rewdat.ads)
20154 @section @code{GNAT.Rewrite_Data} (@file{g-rewdat.ads})
20155 @cindex @code{GNAT.Rewrite_Data} (@file{g-rewdat.ads})
20156 @cindex Rewrite data
20159 A unit to rewrite on-the-fly string occurrences in a stream of
20160 data. The implementation has a very minimal memory footprint as the
20161 full content to be processed is not loaded into memory all at once. This makes
20162 this interface usable for large files or socket streams.
20164 @node GNAT.Secondary_Stack_Info (g-sestin.ads)
20165 @section @code{GNAT.Secondary_Stack_Info} (@file{g-sestin.ads})
20166 @cindex @code{GNAT.Secondary_Stack_Info} (@file{g-sestin.ads})
20167 @cindex Secondary Stack Info
20170 Provide the capability to query the high water mark of the current task's
20173 @node GNAT.Semaphores (g-semaph.ads)
20174 @section @code{GNAT.Semaphores} (@file{g-semaph.ads})
20175 @cindex @code{GNAT.Semaphores} (@file{g-semaph.ads})
20179 Provides classic counting and binary semaphores using protected types.
20181 @node GNAT.Serial_Communications (g-sercom.ads)
20182 @section @code{GNAT.Serial_Communications} (@file{g-sercom.ads})
20183 @cindex @code{GNAT.Serial_Communications} (@file{g-sercom.ads})
20184 @cindex Serial_Communications
20187 Provides a simple interface to send and receive data over a serial
20188 port. This is only supported on GNU/Linux and Windows.
20190 @node GNAT.SHA1 (g-sha1.ads)
20191 @section @code{GNAT.SHA1} (@file{g-sha1.ads})
20192 @cindex @code{GNAT.SHA1} (@file{g-sha1.ads})
20193 @cindex Secure Hash Algorithm SHA-1
20196 Implements the SHA-1 Secure Hash Algorithm as described in FIPS PUB 180-3
20197 and RFC 3174, and the HMAC-SHA1 message authentication function as described
20198 in RFC 2104 and FIPS PUB 198.
20200 @node GNAT.SHA224 (g-sha224.ads)
20201 @section @code{GNAT.SHA224} (@file{g-sha224.ads})
20202 @cindex @code{GNAT.SHA224} (@file{g-sha224.ads})
20203 @cindex Secure Hash Algorithm SHA-224
20206 Implements the SHA-224 Secure Hash Algorithm as described in FIPS PUB 180-3,
20207 and the HMAC-SHA224 message authentication function as described
20208 in RFC 2104 and FIPS PUB 198.
20210 @node GNAT.SHA256 (g-sha256.ads)
20211 @section @code{GNAT.SHA256} (@file{g-sha256.ads})
20212 @cindex @code{GNAT.SHA256} (@file{g-sha256.ads})
20213 @cindex Secure Hash Algorithm SHA-256
20216 Implements the SHA-256 Secure Hash Algorithm as described in FIPS PUB 180-3,
20217 and the HMAC-SHA256 message authentication function as described
20218 in RFC 2104 and FIPS PUB 198.
20220 @node GNAT.SHA384 (g-sha384.ads)
20221 @section @code{GNAT.SHA384} (@file{g-sha384.ads})
20222 @cindex @code{GNAT.SHA384} (@file{g-sha384.ads})
20223 @cindex Secure Hash Algorithm SHA-384
20226 Implements the SHA-384 Secure Hash Algorithm as described in FIPS PUB 180-3,
20227 and the HMAC-SHA384 message authentication function as described
20228 in RFC 2104 and FIPS PUB 198.
20230 @node GNAT.SHA512 (g-sha512.ads)
20231 @section @code{GNAT.SHA512} (@file{g-sha512.ads})
20232 @cindex @code{GNAT.SHA512} (@file{g-sha512.ads})
20233 @cindex Secure Hash Algorithm SHA-512
20236 Implements the SHA-512 Secure Hash Algorithm as described in FIPS PUB 180-3,
20237 and the HMAC-SHA512 message authentication function as described
20238 in RFC 2104 and FIPS PUB 198.
20240 @node GNAT.Signals (g-signal.ads)
20241 @section @code{GNAT.Signals} (@file{g-signal.ads})
20242 @cindex @code{GNAT.Signals} (@file{g-signal.ads})
20246 Provides the ability to manipulate the blocked status of signals on supported
20249 @node GNAT.Sockets (g-socket.ads)
20250 @section @code{GNAT.Sockets} (@file{g-socket.ads})
20251 @cindex @code{GNAT.Sockets} (@file{g-socket.ads})
20255 A high level and portable interface to develop sockets based applications.
20256 This package is based on the sockets thin binding found in
20257 @code{GNAT.Sockets.Thin}. Currently @code{GNAT.Sockets} is implemented
20258 on all native GNAT ports except for OpenVMS@. It is not implemented
20259 for the LynxOS@ cross port.
20261 @node GNAT.Source_Info (g-souinf.ads)
20262 @section @code{GNAT.Source_Info} (@file{g-souinf.ads})
20263 @cindex @code{GNAT.Source_Info} (@file{g-souinf.ads})
20264 @cindex Source Information
20267 Provides subprograms that give access to source code information known at
20268 compile time, such as the current file name and line number. Also provides
20269 subprograms yielding the date and time of the current compilation (like the
20270 C macros @code{__DATE__} and @code{__TIME__})
20272 @node GNAT.Spelling_Checker (g-speche.ads)
20273 @section @code{GNAT.Spelling_Checker} (@file{g-speche.ads})
20274 @cindex @code{GNAT.Spelling_Checker} (@file{g-speche.ads})
20275 @cindex Spell checking
20278 Provides a function for determining whether one string is a plausible
20279 near misspelling of another string.
20281 @node GNAT.Spelling_Checker_Generic (g-spchge.ads)
20282 @section @code{GNAT.Spelling_Checker_Generic} (@file{g-spchge.ads})
20283 @cindex @code{GNAT.Spelling_Checker_Generic} (@file{g-spchge.ads})
20284 @cindex Spell checking
20287 Provides a generic function that can be instantiated with a string type for
20288 determining whether one string is a plausible near misspelling of another
20291 @node GNAT.Spitbol.Patterns (g-spipat.ads)
20292 @section @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
20293 @cindex @code{GNAT.Spitbol.Patterns} (@file{g-spipat.ads})
20294 @cindex SPITBOL pattern matching
20295 @cindex Pattern matching
20298 A complete implementation of SNOBOL4 style pattern matching. This is the
20299 most elaborate of the pattern matching packages provided. It fully duplicates
20300 the SNOBOL4 dynamic pattern construction and matching capabilities, using the
20301 efficient algorithm developed by Robert Dewar for the SPITBOL system.
20303 @node GNAT.Spitbol (g-spitbo.ads)
20304 @section @code{GNAT.Spitbol} (@file{g-spitbo.ads})
20305 @cindex @code{GNAT.Spitbol} (@file{g-spitbo.ads})
20306 @cindex SPITBOL interface
20309 The top level package of the collection of SPITBOL-style functionality, this
20310 package provides basic SNOBOL4 string manipulation functions, such as
20311 Pad, Reverse, Trim, Substr capability, as well as a generic table function
20312 useful for constructing arbitrary mappings from strings in the style of
20313 the SNOBOL4 TABLE function.
20315 @node GNAT.Spitbol.Table_Boolean (g-sptabo.ads)
20316 @section @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
20317 @cindex @code{GNAT.Spitbol.Table_Boolean} (@file{g-sptabo.ads})
20318 @cindex Sets of strings
20319 @cindex SPITBOL Tables
20322 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
20323 for type @code{Standard.Boolean}, giving an implementation of sets of
20326 @node GNAT.Spitbol.Table_Integer (g-sptain.ads)
20327 @section @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
20328 @cindex @code{GNAT.Spitbol.Table_Integer} (@file{g-sptain.ads})
20329 @cindex Integer maps
20331 @cindex SPITBOL Tables
20334 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table}
20335 for type @code{Standard.Integer}, giving an implementation of maps
20336 from string to integer values.
20338 @node GNAT.Spitbol.Table_VString (g-sptavs.ads)
20339 @section @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
20340 @cindex @code{GNAT.Spitbol.Table_VString} (@file{g-sptavs.ads})
20341 @cindex String maps
20343 @cindex SPITBOL Tables
20346 A library level of instantiation of @code{GNAT.Spitbol.Patterns.Table} for
20347 a variable length string type, giving an implementation of general
20348 maps from strings to strings.
20350 @node GNAT.SSE (g-sse.ads)
20351 @section @code{GNAT.SSE} (@file{g-sse.ads})
20352 @cindex @code{GNAT.SSE} (@file{g-sse.ads})
20355 Root of a set of units aimed at offering Ada bindings to a subset of
20356 the Intel(r) Streaming SIMD Extensions with GNAT on the x86 family of
20357 targets. It exposes vector component types together with a general
20358 introduction to the binding contents and use.
20360 @node GNAT.SSE.Vector_Types (g-ssvety.ads)
20361 @section @code{GNAT.SSE.Vector_Types} (@file{g-ssvety.ads})
20362 @cindex @code{GNAT.SSE.Vector_Types} (@file{g-ssvety.ads})
20365 SSE vector types for use with SSE related intrinsics.
20367 @node GNAT.Strings (g-string.ads)
20368 @section @code{GNAT.Strings} (@file{g-string.ads})
20369 @cindex @code{GNAT.Strings} (@file{g-string.ads})
20372 Common String access types and related subprograms. Basically it
20373 defines a string access and an array of string access types.
20375 @node GNAT.String_Split (g-strspl.ads)
20376 @section @code{GNAT.String_Split} (@file{g-strspl.ads})
20377 @cindex @code{GNAT.String_Split} (@file{g-strspl.ads})
20378 @cindex String splitter
20381 Useful string manipulation routines: given a set of separators, split
20382 a string wherever the separators appear, and provide direct access
20383 to the resulting slices. This package is instantiated from
20384 @code{GNAT.Array_Split}.
20386 @node GNAT.Table (g-table.ads)
20387 @section @code{GNAT.Table} (@file{g-table.ads})
20388 @cindex @code{GNAT.Table} (@file{g-table.ads})
20389 @cindex Table implementation
20390 @cindex Arrays, extendable
20393 A generic package providing a single dimension array abstraction where the
20394 length of the array can be dynamically modified.
20397 This package provides a facility similar to that of @code{GNAT.Dynamic_Tables},
20398 except that this package declares a single instance of the table type,
20399 while an instantiation of @code{GNAT.Dynamic_Tables} creates a type that can be
20400 used to define dynamic instances of the table.
20402 @node GNAT.Task_Lock (g-tasloc.ads)
20403 @section @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
20404 @cindex @code{GNAT.Task_Lock} (@file{g-tasloc.ads})
20405 @cindex Task synchronization
20406 @cindex Task locking
20410 A very simple facility for locking and unlocking sections of code using a
20411 single global task lock. Appropriate for use in situations where contention
20412 between tasks is very rarely expected.
20414 @node GNAT.Threads (g-thread.ads)
20415 @section @code{GNAT.Threads} (@file{g-thread.ads})
20416 @cindex @code{GNAT.Threads} (@file{g-thread.ads})
20417 @cindex Foreign threads
20418 @cindex Threads, foreign
20421 Provides facilities for dealing with foreign threads which need to be known
20422 by the GNAT run-time system. Consult the documentation of this package for
20423 further details if your program has threads that are created by a non-Ada
20424 environment which then accesses Ada code.
20426 @node GNAT.Time_Stamp (g-timsta.ads)
20427 @section @code{GNAT.Time_Stamp} (@file{g-timsta.ads})
20428 @cindex @code{GNAT.Time_Stamp} (@file{g-timsta.ads})
20430 @cindex Current time
20433 Provides a simple function that returns a string YYYY-MM-DD HH:MM:SS.SS that
20434 represents the current date and time in ISO 8601 format. This is a very simple
20435 routine with minimal code and there are no dependencies on any other unit.
20437 @node GNAT.Traceback (g-traceb.ads)
20438 @section @code{GNAT.Traceback} (@file{g-traceb.ads})
20439 @cindex @code{GNAT.Traceback} (@file{g-traceb.ads})
20440 @cindex Trace back facilities
20443 Provides a facility for obtaining non-symbolic traceback information, useful
20444 in various debugging situations.
20446 @node GNAT.Traceback.Symbolic (g-trasym.ads)
20447 @section @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
20448 @cindex @code{GNAT.Traceback.Symbolic} (@file{g-trasym.ads})
20449 @cindex Trace back facilities
20451 @node GNAT.UTF_32 (g-utf_32.ads)
20452 @section @code{GNAT.UTF_32} (@file{g-table.ads})
20453 @cindex @code{GNAT.UTF_32} (@file{g-table.ads})
20454 @cindex Wide character codes
20457 This is a package intended to be used in conjunction with the
20458 @code{Wide_Character} type in Ada 95 and the
20459 @code{Wide_Wide_Character} type in Ada 2005 (available
20460 in @code{GNAT} in Ada 2005 mode). This package contains
20461 Unicode categorization routines, as well as lexical
20462 categorization routines corresponding to the Ada 2005
20463 lexical rules for identifiers and strings, and also a
20464 lower case to upper case fold routine corresponding to
20465 the Ada 2005 rules for identifier equivalence.
20467 @node GNAT.UTF_32_Spelling_Checker (g-u3spch.ads)
20468 @section @code{GNAT.Wide_Spelling_Checker} (@file{g-u3spch.ads})
20469 @cindex @code{GNAT.Wide_Spelling_Checker} (@file{g-u3spch.ads})
20470 @cindex Spell checking
20473 Provides a function for determining whether one wide wide string is a plausible
20474 near misspelling of another wide wide string, where the strings are represented
20475 using the UTF_32_String type defined in System.Wch_Cnv.
20477 @node GNAT.Wide_Spelling_Checker (g-wispch.ads)
20478 @section @code{GNAT.Wide_Spelling_Checker} (@file{g-wispch.ads})
20479 @cindex @code{GNAT.Wide_Spelling_Checker} (@file{g-wispch.ads})
20480 @cindex Spell checking
20483 Provides a function for determining whether one wide string is a plausible
20484 near misspelling of another wide string.
20486 @node GNAT.Wide_String_Split (g-wistsp.ads)
20487 @section @code{GNAT.Wide_String_Split} (@file{g-wistsp.ads})
20488 @cindex @code{GNAT.Wide_String_Split} (@file{g-wistsp.ads})
20489 @cindex Wide_String splitter
20492 Useful wide string manipulation routines: given a set of separators, split
20493 a wide string wherever the separators appear, and provide direct access
20494 to the resulting slices. This package is instantiated from
20495 @code{GNAT.Array_Split}.
20497 @node GNAT.Wide_Wide_Spelling_Checker (g-zspche.ads)
20498 @section @code{GNAT.Wide_Wide_Spelling_Checker} (@file{g-zspche.ads})
20499 @cindex @code{GNAT.Wide_Wide_Spelling_Checker} (@file{g-zspche.ads})
20500 @cindex Spell checking
20503 Provides a function for determining whether one wide wide string is a plausible
20504 near misspelling of another wide wide string.
20506 @node GNAT.Wide_Wide_String_Split (g-zistsp.ads)
20507 @section @code{GNAT.Wide_Wide_String_Split} (@file{g-zistsp.ads})
20508 @cindex @code{GNAT.Wide_Wide_String_Split} (@file{g-zistsp.ads})
20509 @cindex Wide_Wide_String splitter
20512 Useful wide wide string manipulation routines: given a set of separators, split
20513 a wide wide string wherever the separators appear, and provide direct access
20514 to the resulting slices. This package is instantiated from
20515 @code{GNAT.Array_Split}.
20517 @node Interfaces.C.Extensions (i-cexten.ads)
20518 @section @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
20519 @cindex @code{Interfaces.C.Extensions} (@file{i-cexten.ads})
20522 This package contains additional C-related definitions, intended
20523 for use with either manually or automatically generated bindings
20526 @node Interfaces.C.Streams (i-cstrea.ads)
20527 @section @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
20528 @cindex @code{Interfaces.C.Streams} (@file{i-cstrea.ads})
20529 @cindex C streams, interfacing
20532 This package is a binding for the most commonly used operations
20535 @node Interfaces.Packed_Decimal (i-pacdec.ads)
20536 @section @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
20537 @cindex @code{Interfaces.Packed_Decimal} (@file{i-pacdec.ads})
20538 @cindex IBM Packed Format
20539 @cindex Packed Decimal
20542 This package provides a set of routines for conversions to and
20543 from a packed decimal format compatible with that used on IBM
20546 @node Interfaces.VxWorks (i-vxwork.ads)
20547 @section @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
20548 @cindex @code{Interfaces.VxWorks} (@file{i-vxwork.ads})
20549 @cindex Interfacing to VxWorks
20550 @cindex VxWorks, interfacing
20553 This package provides a limited binding to the VxWorks API.
20554 In particular, it interfaces with the
20555 VxWorks hardware interrupt facilities.
20557 @node Interfaces.VxWorks.IO (i-vxwoio.ads)
20558 @section @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
20559 @cindex @code{Interfaces.VxWorks.IO} (@file{i-vxwoio.ads})
20560 @cindex Interfacing to VxWorks' I/O
20561 @cindex VxWorks, I/O interfacing
20562 @cindex VxWorks, Get_Immediate
20563 @cindex Get_Immediate, VxWorks
20566 This package provides a binding to the ioctl (IO/Control)
20567 function of VxWorks, defining a set of option values and
20568 function codes. A particular use of this package is
20569 to enable the use of Get_Immediate under VxWorks.
20571 @node System.Address_Image (s-addima.ads)
20572 @section @code{System.Address_Image} (@file{s-addima.ads})
20573 @cindex @code{System.Address_Image} (@file{s-addima.ads})
20574 @cindex Address image
20575 @cindex Image, of an address
20578 This function provides a useful debugging
20579 function that gives an (implementation dependent)
20580 string which identifies an address.
20582 @node System.Assertions (s-assert.ads)
20583 @section @code{System.Assertions} (@file{s-assert.ads})
20584 @cindex @code{System.Assertions} (@file{s-assert.ads})
20586 @cindex Assert_Failure, exception
20589 This package provides the declaration of the exception raised
20590 by an run-time assertion failure, as well as the routine that
20591 is used internally to raise this assertion.
20593 @node System.Atomic_Counters (s-atocou.ads)
20594 @section @code{System.Atomic_Counters} (@file{s-atocou.ads})
20595 @cindex @code{System.Atomic_Counters} (@file{s-atocou.ads})
20598 This package provides the declaration of an atomic counter type,
20599 together with efficient routines (using hardware
20600 synchronization primitives) for incrementing, decrementing,
20601 and testing of these counters. This package is implemented
20602 on most targets, including all Alpha, ia64, PowerPC, SPARC V9,
20603 x86, and x86_64 platforms.
20605 @node System.Memory (s-memory.ads)
20606 @section @code{System.Memory} (@file{s-memory.ads})
20607 @cindex @code{System.Memory} (@file{s-memory.ads})
20608 @cindex Memory allocation
20611 This package provides the interface to the low level routines used
20612 by the generated code for allocation and freeing storage for the
20613 default storage pool (analogous to the C routines malloc and free.
20614 It also provides a reallocation interface analogous to the C routine
20615 realloc. The body of this unit may be modified to provide alternative
20616 allocation mechanisms for the default pool, and in addition, direct
20617 calls to this unit may be made for low level allocation uses (for
20618 example see the body of @code{GNAT.Tables}).
20620 @node System.Multiprocessors (s-multip.ads)
20621 @section @code{System.Multiprocessors} (@file{s-multip.ads})
20622 @cindex @code{System.Multiprocessors} (@file{s-multip.ads})
20623 @cindex Multiprocessor interface
20624 This is an Ada 2012 unit defined in the Ada 2012 Reference Manual, but
20625 in GNAT we also make it available in Ada 95 and Ada 2005 (where it is
20626 technically an implementation-defined addition).
20628 @node System.Multiprocessors.Dispatching_Domains (s-mudido.ads)
20629 @section @code{System.Multiprocessors.Dispatching_Domains} (@file{s-mudido.ads})
20630 @cindex @code{System.Multiprocessors.Dispatching_Domains} (@file{s-mudido.ads})
20631 @cindex Multiprocessor interface
20632 This is an Ada 2012 unit defined in the Ada 2012 Reference Manual, but
20633 in GNAT we also make it available in Ada 95 and Ada 2005 (where it is
20634 technically an implementation-defined addition).
20636 @node System.Partition_Interface (s-parint.ads)
20637 @section @code{System.Partition_Interface} (@file{s-parint.ads})
20638 @cindex @code{System.Partition_Interface} (@file{s-parint.ads})
20639 @cindex Partition interfacing functions
20642 This package provides facilities for partition interfacing. It
20643 is used primarily in a distribution context when using Annex E
20646 @node System.Pool_Global (s-pooglo.ads)
20647 @section @code{System.Pool_Global} (@file{s-pooglo.ads})
20648 @cindex @code{System.Pool_Global} (@file{s-pooglo.ads})
20649 @cindex Storage pool, global
20650 @cindex Global storage pool
20653 This package provides a storage pool that is equivalent to the default
20654 storage pool used for access types for which no pool is specifically
20655 declared. It uses malloc/free to allocate/free and does not attempt to
20656 do any automatic reclamation.
20658 @node System.Pool_Local (s-pooloc.ads)
20659 @section @code{System.Pool_Local} (@file{s-pooloc.ads})
20660 @cindex @code{System.Pool_Local} (@file{s-pooloc.ads})
20661 @cindex Storage pool, local
20662 @cindex Local storage pool
20665 This package provides a storage pool that is intended for use with locally
20666 defined access types. It uses malloc/free for allocate/free, and maintains
20667 a list of allocated blocks, so that all storage allocated for the pool can
20668 be freed automatically when the pool is finalized.
20670 @node System.Restrictions (s-restri.ads)
20671 @section @code{System.Restrictions} (@file{s-restri.ads})
20672 @cindex @code{System.Restrictions} (@file{s-restri.ads})
20673 @cindex Run-time restrictions access
20676 This package provides facilities for accessing at run time
20677 the status of restrictions specified at compile time for
20678 the partition. Information is available both with regard
20679 to actual restrictions specified, and with regard to
20680 compiler determined information on which restrictions
20681 are violated by one or more packages in the partition.
20683 @node System.Rident (s-rident.ads)
20684 @section @code{System.Rident} (@file{s-rident.ads})
20685 @cindex @code{System.Rident} (@file{s-rident.ads})
20686 @cindex Restrictions definitions
20689 This package provides definitions of the restrictions
20690 identifiers supported by GNAT, and also the format of
20691 the restrictions provided in package System.Restrictions.
20692 It is not normally necessary to @code{with} this generic package
20693 since the necessary instantiation is included in
20694 package System.Restrictions.
20696 @node System.Strings.Stream_Ops (s-ststop.ads)
20697 @section @code{System.Strings.Stream_Ops} (@file{s-ststop.ads})
20698 @cindex @code{System.Strings.Stream_Ops} (@file{s-ststop.ads})
20699 @cindex Stream operations
20700 @cindex String stream operations
20703 This package provides a set of stream subprograms for standard string types.
20704 It is intended primarily to support implicit use of such subprograms when
20705 stream attributes are applied to string types, but the subprograms in this
20706 package can be used directly by application programs.
20708 @node System.Unsigned_Types (s-unstyp.ads)
20709 @section @code{System.Unsigned_Types} (@file{s-unstyp.ads})
20710 @cindex @code{System.Unsigned_Types} (@file{s-unstyp.ads})
20713 This package contains definitions of standard unsigned types that
20714 correspond in size to the standard signed types declared in Standard,
20715 and (unlike the types in Interfaces) have corresponding names. It
20716 also contains some related definitions for other specialized types
20717 used by the compiler in connection with packed array types.
20719 @node System.Wch_Cnv (s-wchcnv.ads)
20720 @section @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
20721 @cindex @code{System.Wch_Cnv} (@file{s-wchcnv.ads})
20722 @cindex Wide Character, Representation
20723 @cindex Wide String, Conversion
20724 @cindex Representation of wide characters
20727 This package provides routines for converting between
20728 wide and wide wide characters and a representation as a value of type
20729 @code{Standard.String}, using a specified wide character
20730 encoding method. It uses definitions in
20731 package @code{System.Wch_Con}.
20733 @node System.Wch_Con (s-wchcon.ads)
20734 @section @code{System.Wch_Con} (@file{s-wchcon.ads})
20735 @cindex @code{System.Wch_Con} (@file{s-wchcon.ads})
20738 This package provides definitions and descriptions of
20739 the various methods used for encoding wide characters
20740 in ordinary strings. These definitions are used by
20741 the package @code{System.Wch_Cnv}.
20743 @node Interfacing to Other Languages
20744 @chapter Interfacing to Other Languages
20746 The facilities in annex B of the Ada Reference Manual are fully
20747 implemented in GNAT, and in addition, a full interface to C++ is
20751 * Interfacing to C::
20752 * Interfacing to C++::
20753 * Interfacing to COBOL::
20754 * Interfacing to Fortran::
20755 * Interfacing to non-GNAT Ada code::
20758 @node Interfacing to C
20759 @section Interfacing to C
20762 Interfacing to C with GNAT can use one of two approaches:
20766 The types in the package @code{Interfaces.C} may be used.
20768 Standard Ada types may be used directly. This may be less portable to
20769 other compilers, but will work on all GNAT compilers, which guarantee
20770 correspondence between the C and Ada types.
20774 Pragma @code{Convention C} may be applied to Ada types, but mostly has no
20775 effect, since this is the default. The following table shows the
20776 correspondence between Ada scalar types and the corresponding C types.
20781 @item Short_Integer
20783 @item Short_Short_Integer
20787 @item Long_Long_Integer
20795 @item Long_Long_Float
20796 This is the longest floating-point type supported by the hardware.
20800 Additionally, there are the following general correspondences between Ada
20804 Ada enumeration types map to C enumeration types directly if pragma
20805 @code{Convention C} is specified, which causes them to have int
20806 length. Without pragma @code{Convention C}, Ada enumeration types map to
20807 8, 16, or 32 bits (i.e.@: C types @code{signed char}, @code{short},
20808 @code{int}, respectively) depending on the number of values passed.
20809 This is the only case in which pragma @code{Convention C} affects the
20810 representation of an Ada type.
20813 Ada access types map to C pointers, except for the case of pointers to
20814 unconstrained types in Ada, which have no direct C equivalent.
20817 Ada arrays map directly to C arrays.
20820 Ada records map directly to C structures.
20823 Packed Ada records map to C structures where all members are bit fields
20824 of the length corresponding to the @code{@var{type}'Size} value in Ada.
20827 @node Interfacing to C++
20828 @section Interfacing to C++
20831 The interface to C++ makes use of the following pragmas, which are
20832 primarily intended to be constructed automatically using a binding generator
20833 tool, although it is possible to construct them by hand.
20835 Using these pragmas it is possible to achieve complete
20836 inter-operability between Ada tagged types and C++ class definitions.
20837 See @ref{Implementation Defined Pragmas}, for more details.
20840 @item pragma CPP_Class ([Entity =>] @var{LOCAL_NAME})
20841 The argument denotes an entity in the current declarative region that is
20842 declared as a tagged or untagged record type. It indicates that the type
20843 corresponds to an externally declared C++ class type, and is to be laid
20844 out the same way that C++ would lay out the type.
20846 Note: Pragma @code{CPP_Class} is currently obsolete. It is supported
20847 for backward compatibility but its functionality is available
20848 using pragma @code{Import} with @code{Convention} = @code{CPP}.
20850 @item pragma CPP_Constructor ([Entity =>] @var{LOCAL_NAME})
20851 This pragma identifies an imported function (imported in the usual way
20852 with pragma @code{Import}) as corresponding to a C++ constructor.
20855 A few restrictions are placed on the use of the @code{Access} attribute
20856 in conjunction with subprograms subject to convention @code{CPP}: the
20857 attribute may be used neither on primitive operations of a tagged
20858 record type with convention @code{CPP}, imported or not, nor on
20859 subprograms imported with pragma @code{CPP_Constructor}.
20861 In addition, C++ exceptions are propagated and can be handled in an
20862 @code{others} choice of an exception handler. The corresponding Ada
20863 occurrence has no message, and the simple name of the exception identity
20864 contains @samp{Foreign_Exception}. Finalization and awaiting dependent
20865 tasks works properly when such foreign exceptions are propagated.
20867 It is also possible to import a C++ exception using the following syntax:
20869 @smallexample @c ada
20870 LOCAL_NAME : exception;
20871 pragma Import (Cpp,
20872 [Entity =>] LOCAL_NAME,
20873 [External_Name =>] static_string_EXPRESSION);
20877 The @code{External_Name} is the name of the C++ RTTI symbol. You can then
20878 cover a specific C++ exception in an exception handler.
20880 @node Interfacing to COBOL
20881 @section Interfacing to COBOL
20884 Interfacing to COBOL is achieved as described in section B.4 of
20885 the Ada Reference Manual.
20887 @node Interfacing to Fortran
20888 @section Interfacing to Fortran
20891 Interfacing to Fortran is achieved as described in section B.5 of the
20892 Ada Reference Manual. The pragma @code{Convention Fortran}, applied to a
20893 multi-dimensional array causes the array to be stored in column-major
20894 order as required for convenient interface to Fortran.
20896 @node Interfacing to non-GNAT Ada code
20897 @section Interfacing to non-GNAT Ada code
20899 It is possible to specify the convention @code{Ada} in a pragma
20900 @code{Import} or pragma @code{Export}. However this refers to
20901 the calling conventions used by GNAT, which may or may not be
20902 similar enough to those used by some other Ada 83 / Ada 95 / Ada 2005
20903 compiler to allow interoperation.
20905 If arguments types are kept simple, and if the foreign compiler generally
20906 follows system calling conventions, then it may be possible to integrate
20907 files compiled by other Ada compilers, provided that the elaboration
20908 issues are adequately addressed (for example by eliminating the
20909 need for any load time elaboration).
20911 In particular, GNAT running on VMS is designed to
20912 be highly compatible with the DEC Ada 83 compiler, so this is one
20913 case in which it is possible to import foreign units of this type,
20914 provided that the data items passed are restricted to simple scalar
20915 values or simple record types without variants, or simple array
20916 types with fixed bounds.
20918 @node Specialized Needs Annexes
20919 @chapter Specialized Needs Annexes
20922 Ada 95 and Ada 2005 define a number of Specialized Needs Annexes, which are not
20923 required in all implementations. However, as described in this chapter,
20924 GNAT implements all of these annexes:
20927 @item Systems Programming (Annex C)
20928 The Systems Programming Annex is fully implemented.
20930 @item Real-Time Systems (Annex D)
20931 The Real-Time Systems Annex is fully implemented.
20933 @item Distributed Systems (Annex E)
20934 Stub generation is fully implemented in the GNAT compiler. In addition,
20935 a complete compatible PCS is available as part of the GLADE system,
20936 a separate product. When the two
20937 products are used in conjunction, this annex is fully implemented.
20939 @item Information Systems (Annex F)
20940 The Information Systems annex is fully implemented.
20942 @item Numerics (Annex G)
20943 The Numerics Annex is fully implemented.
20945 @item Safety and Security / High-Integrity Systems (Annex H)
20946 The Safety and Security Annex (termed the High-Integrity Systems Annex
20947 in Ada 2005) is fully implemented.
20950 @node Implementation of Specific Ada Features
20951 @chapter Implementation of Specific Ada Features
20954 This chapter describes the GNAT implementation of several Ada language
20958 * Machine Code Insertions::
20959 * GNAT Implementation of Tasking::
20960 * GNAT Implementation of Shared Passive Packages::
20961 * Code Generation for Array Aggregates::
20962 * The Size of Discriminated Records with Default Discriminants::
20963 * Strict Conformance to the Ada Reference Manual::
20966 @node Machine Code Insertions
20967 @section Machine Code Insertions
20968 @cindex Machine Code insertions
20971 Package @code{Machine_Code} provides machine code support as described
20972 in the Ada Reference Manual in two separate forms:
20975 Machine code statements, consisting of qualified expressions that
20976 fit the requirements of RM section 13.8.
20978 An intrinsic callable procedure, providing an alternative mechanism of
20979 including machine instructions in a subprogram.
20983 The two features are similar, and both are closely related to the mechanism
20984 provided by the asm instruction in the GNU C compiler. Full understanding
20985 and use of the facilities in this package requires understanding the asm
20986 instruction, see @ref{Extended Asm,,, gcc, Using the GNU Compiler
20989 Calls to the function @code{Asm} and the procedure @code{Asm} have identical
20990 semantic restrictions and effects as described below. Both are provided so
20991 that the procedure call can be used as a statement, and the function call
20992 can be used to form a code_statement.
20994 Consider this C @code{asm} instruction:
20996 asm ("fsinx %1 %0" : "=f" (result) : "f" (angle));
21000 The equivalent can be written for GNAT as:
21002 @smallexample @c ada
21003 Asm ("fsinx %1 %0",
21004 My_Float'Asm_Output ("=f", result),
21005 My_Float'Asm_Input ("f", angle));
21009 The first argument to @code{Asm} is the assembler template, and is
21010 identical to what is used in GNU C@. This string must be a static
21011 expression. The second argument is the output operand list. It is
21012 either a single @code{Asm_Output} attribute reference, or a list of such
21013 references enclosed in parentheses (technically an array aggregate of
21016 The @code{Asm_Output} attribute denotes a function that takes two
21017 parameters. The first is a string, the second is the name of a variable
21018 of the type designated by the attribute prefix. The first (string)
21019 argument is required to be a static expression and designates the
21020 constraint (@pxref{Constraints,,, gcc, Using the GNU Compiler
21022 for the parameter (e.g.@: what kind of register is required). The second
21023 argument is the variable to be written or updated with the
21024 result. The possible values for constraint are the same as those used in
21025 the RTL, and are dependent on the configuration file used to build the
21026 GCC back end. If there are no output operands, then this argument may
21027 either be omitted, or explicitly given as @code{No_Output_Operands}.
21028 No support is provided for GNU C's symbolic names for output parameters.
21030 The second argument of @code{@var{my_float}'Asm_Output} functions as
21031 though it were an @code{out} parameter, which is a little curious, but
21032 all names have the form of expressions, so there is no syntactic
21033 irregularity, even though normally functions would not be permitted
21034 @code{out} parameters. The third argument is the list of input
21035 operands. It is either a single @code{Asm_Input} attribute reference, or
21036 a list of such references enclosed in parentheses (technically an array
21037 aggregate of such references).
21039 The @code{Asm_Input} attribute denotes a function that takes two
21040 parameters. The first is a string, the second is an expression of the
21041 type designated by the prefix. The first (string) argument is required
21042 to be a static expression, and is the constraint for the parameter,
21043 (e.g.@: what kind of register is required). The second argument is the
21044 value to be used as the input argument. The possible values for the
21045 constraint are the same as those used in the RTL, and are dependent on
21046 the configuration file used to built the GCC back end.
21047 No support is provided for GNU C's symbolic names for input parameters.
21049 If there are no input operands, this argument may either be omitted, or
21050 explicitly given as @code{No_Input_Operands}. The fourth argument, not
21051 present in the above example, is a list of register names, called the
21052 @dfn{clobber} argument. This argument, if given, must be a static string
21053 expression, and is a space or comma separated list of names of registers
21054 that must be considered destroyed as a result of the @code{Asm} call. If
21055 this argument is the null string (the default value), then the code
21056 generator assumes that no additional registers are destroyed.
21057 In addition to registers, the special clobbers @code{memory} and
21058 @code{cc} as described in the GNU C docs are both supported.
21060 The fifth argument, not present in the above example, called the
21061 @dfn{volatile} argument, is by default @code{False}. It can be set to
21062 the literal value @code{True} to indicate to the code generator that all
21063 optimizations with respect to the instruction specified should be
21064 suppressed, and in particular an instruction that has outputs
21065 will still be generated, even if none of the outputs are
21066 used. @xref{Volatile,,,
21067 gcc, Using the GNU Compiler Collection (GCC)}, for the full description.
21068 Generally it is strongly advisable to use Volatile for any ASM statement
21069 that is missing either input or output operands or to avoid unwanted
21070 optimizations. A warning is generated if this advice is not followed.
21072 No support is provided for GNU C's @code{asm goto} feature.
21074 The @code{Asm} subprograms may be used in two ways. First the procedure
21075 forms can be used anywhere a procedure call would be valid, and
21076 correspond to what the RM calls ``intrinsic'' routines. Such calls can
21077 be used to intersperse machine instructions with other Ada statements.
21078 Second, the function forms, which return a dummy value of the limited
21079 private type @code{Asm_Insn}, can be used in code statements, and indeed
21080 this is the only context where such calls are allowed. Code statements
21081 appear as aggregates of the form:
21083 @smallexample @c ada
21084 Asm_Insn'(Asm (@dots{}));
21085 Asm_Insn'(Asm_Volatile (@dots{}));
21089 In accordance with RM rules, such code statements are allowed only
21090 within subprograms whose entire body consists of such statements. It is
21091 not permissible to intermix such statements with other Ada statements.
21093 Typically the form using intrinsic procedure calls is more convenient
21094 and more flexible. The code statement form is provided to meet the RM
21095 suggestion that such a facility should be made available. The following
21096 is the exact syntax of the call to @code{Asm}. As usual, if named notation
21097 is used, the arguments may be given in arbitrary order, following the
21098 normal rules for use of positional and named arguments:
21102 [Template =>] static_string_EXPRESSION
21103 [,[Outputs =>] OUTPUT_OPERAND_LIST ]
21104 [,[Inputs =>] INPUT_OPERAND_LIST ]
21105 [,[Clobber =>] static_string_EXPRESSION ]
21106 [,[Volatile =>] static_boolean_EXPRESSION] )
21108 OUTPUT_OPERAND_LIST ::=
21109 [PREFIX.]No_Output_Operands
21110 | OUTPUT_OPERAND_ATTRIBUTE
21111 | (OUTPUT_OPERAND_ATTRIBUTE @{,OUTPUT_OPERAND_ATTRIBUTE@})
21113 OUTPUT_OPERAND_ATTRIBUTE ::=
21114 SUBTYPE_MARK'Asm_Output (static_string_EXPRESSION, NAME)
21116 INPUT_OPERAND_LIST ::=
21117 [PREFIX.]No_Input_Operands
21118 | INPUT_OPERAND_ATTRIBUTE
21119 | (INPUT_OPERAND_ATTRIBUTE @{,INPUT_OPERAND_ATTRIBUTE@})
21121 INPUT_OPERAND_ATTRIBUTE ::=
21122 SUBTYPE_MARK'Asm_Input (static_string_EXPRESSION, EXPRESSION)
21126 The identifiers @code{No_Input_Operands} and @code{No_Output_Operands}
21127 are declared in the package @code{Machine_Code} and must be referenced
21128 according to normal visibility rules. In particular if there is no
21129 @code{use} clause for this package, then appropriate package name
21130 qualification is required.
21132 @node GNAT Implementation of Tasking
21133 @section GNAT Implementation of Tasking
21136 This chapter outlines the basic GNAT approach to tasking (in particular,
21137 a multi-layered library for portability) and discusses issues related
21138 to compliance with the Real-Time Systems Annex.
21141 * Mapping Ada Tasks onto the Underlying Kernel Threads::
21142 * Ensuring Compliance with the Real-Time Annex::
21145 @node Mapping Ada Tasks onto the Underlying Kernel Threads
21146 @subsection Mapping Ada Tasks onto the Underlying Kernel Threads
21149 GNAT's run-time support comprises two layers:
21152 @item GNARL (GNAT Run-time Layer)
21153 @item GNULL (GNAT Low-level Library)
21157 In GNAT, Ada's tasking services rely on a platform and OS independent
21158 layer known as GNARL@. This code is responsible for implementing the
21159 correct semantics of Ada's task creation, rendezvous, protected
21162 GNARL decomposes Ada's tasking semantics into simpler lower level
21163 operations such as create a thread, set the priority of a thread,
21164 yield, create a lock, lock/unlock, etc. The spec for these low-level
21165 operations constitutes GNULLI, the GNULL Interface. This interface is
21166 directly inspired from the POSIX real-time API@.
21168 If the underlying executive or OS implements the POSIX standard
21169 faithfully, the GNULL Interface maps as is to the services offered by
21170 the underlying kernel. Otherwise, some target dependent glue code maps
21171 the services offered by the underlying kernel to the semantics expected
21174 Whatever the underlying OS (VxWorks, UNIX, Windows, etc.) the
21175 key point is that each Ada task is mapped on a thread in the underlying
21176 kernel. For example, in the case of VxWorks, one Ada task = one VxWorks task.
21178 In addition Ada task priorities map onto the underlying thread priorities.
21179 Mapping Ada tasks onto the underlying kernel threads has several advantages:
21183 The underlying scheduler is used to schedule the Ada tasks. This
21184 makes Ada tasks as efficient as kernel threads from a scheduling
21188 Interaction with code written in C containing threads is eased
21189 since at the lowest level Ada tasks and C threads map onto the same
21190 underlying kernel concept.
21193 When an Ada task is blocked during I/O the remaining Ada tasks are
21197 On multiprocessor systems Ada tasks can execute in parallel.
21201 Some threads libraries offer a mechanism to fork a new process, with the
21202 child process duplicating the threads from the parent.
21204 support this functionality when the parent contains more than one task.
21205 @cindex Forking a new process
21207 @node Ensuring Compliance with the Real-Time Annex
21208 @subsection Ensuring Compliance with the Real-Time Annex
21209 @cindex Real-Time Systems Annex compliance
21212 Although mapping Ada tasks onto
21213 the underlying threads has significant advantages, it does create some
21214 complications when it comes to respecting the scheduling semantics
21215 specified in the real-time annex (Annex D).
21217 For instance the Annex D requirement for the @code{FIFO_Within_Priorities}
21218 scheduling policy states:
21221 @emph{When the active priority of a ready task that is not running
21222 changes, or the setting of its base priority takes effect, the
21223 task is removed from the ready queue for its old active priority
21224 and is added at the tail of the ready queue for its new active
21225 priority, except in the case where the active priority is lowered
21226 due to the loss of inherited priority, in which case the task is
21227 added at the head of the ready queue for its new active priority.}
21231 While most kernels do put tasks at the end of the priority queue when
21232 a task changes its priority, (which respects the main
21233 FIFO_Within_Priorities requirement), almost none keep a thread at the
21234 beginning of its priority queue when its priority drops from the loss
21235 of inherited priority.
21237 As a result most vendors have provided incomplete Annex D implementations.
21239 The GNAT run-time, has a nice cooperative solution to this problem
21240 which ensures that accurate FIFO_Within_Priorities semantics are
21243 The principle is as follows. When an Ada task T is about to start
21244 running, it checks whether some other Ada task R with the same
21245 priority as T has been suspended due to the loss of priority
21246 inheritance. If this is the case, T yields and is placed at the end of
21247 its priority queue. When R arrives at the front of the queue it
21250 Note that this simple scheme preserves the relative order of the tasks
21251 that were ready to execute in the priority queue where R has been
21254 @node GNAT Implementation of Shared Passive Packages
21255 @section GNAT Implementation of Shared Passive Packages
21256 @cindex Shared passive packages
21259 GNAT fully implements the pragma @code{Shared_Passive} for
21260 @cindex pragma @code{Shared_Passive}
21261 the purpose of designating shared passive packages.
21262 This allows the use of passive partitions in the
21263 context described in the Ada Reference Manual; i.e., for communication
21264 between separate partitions of a distributed application using the
21265 features in Annex E.
21267 @cindex Distribution Systems Annex
21269 However, the implementation approach used by GNAT provides for more
21270 extensive usage as follows:
21273 @item Communication between separate programs
21275 This allows separate programs to access the data in passive
21276 partitions, using protected objects for synchronization where
21277 needed. The only requirement is that the two programs have a
21278 common shared file system. It is even possible for programs
21279 running on different machines with different architectures
21280 (e.g.@: different endianness) to communicate via the data in
21281 a passive partition.
21283 @item Persistence between program runs
21285 The data in a passive package can persist from one run of a
21286 program to another, so that a later program sees the final
21287 values stored by a previous run of the same program.
21292 The implementation approach used is to store the data in files. A
21293 separate stream file is created for each object in the package, and
21294 an access to an object causes the corresponding file to be read or
21297 The environment variable @code{SHARED_MEMORY_DIRECTORY} should be
21298 @cindex @code{SHARED_MEMORY_DIRECTORY} environment variable
21299 set to the directory to be used for these files.
21300 The files in this directory
21301 have names that correspond to their fully qualified names. For
21302 example, if we have the package
21304 @smallexample @c ada
21306 pragma Shared_Passive (X);
21313 and the environment variable is set to @code{/stemp/}, then the files created
21314 will have the names:
21322 These files are created when a value is initially written to the object, and
21323 the files are retained until manually deleted. This provides the persistence
21324 semantics. If no file exists, it means that no partition has assigned a value
21325 to the variable; in this case the initial value declared in the package
21326 will be used. This model ensures that there are no issues in synchronizing
21327 the elaboration process, since elaboration of passive packages elaborates the
21328 initial values, but does not create the files.
21330 The files are written using normal @code{Stream_IO} access.
21331 If you want to be able
21332 to communicate between programs or partitions running on different
21333 architectures, then you should use the XDR versions of the stream attribute
21334 routines, since these are architecture independent.
21336 If active synchronization is required for access to the variables in the
21337 shared passive package, then as described in the Ada Reference Manual, the
21338 package may contain protected objects used for this purpose. In this case
21339 a lock file (whose name is @file{___lock} (three underscores)
21340 is created in the shared memory directory.
21341 @cindex @file{___lock} file (for shared passive packages)
21342 This is used to provide the required locking
21343 semantics for proper protected object synchronization.
21345 As of January 2003, GNAT supports shared passive packages on all platforms
21346 except for OpenVMS.
21348 @node Code Generation for Array Aggregates
21349 @section Code Generation for Array Aggregates
21352 * Static constant aggregates with static bounds::
21353 * Constant aggregates with unconstrained nominal types::
21354 * Aggregates with static bounds::
21355 * Aggregates with non-static bounds::
21356 * Aggregates in assignment statements::
21360 Aggregates have a rich syntax and allow the user to specify the values of
21361 complex data structures by means of a single construct. As a result, the
21362 code generated for aggregates can be quite complex and involve loops, case
21363 statements and multiple assignments. In the simplest cases, however, the
21364 compiler will recognize aggregates whose components and constraints are
21365 fully static, and in those cases the compiler will generate little or no
21366 executable code. The following is an outline of the code that GNAT generates
21367 for various aggregate constructs. For further details, you will find it
21368 useful to examine the output produced by the -gnatG flag to see the expanded
21369 source that is input to the code generator. You may also want to examine
21370 the assembly code generated at various levels of optimization.
21372 The code generated for aggregates depends on the context, the component values,
21373 and the type. In the context of an object declaration the code generated is
21374 generally simpler than in the case of an assignment. As a general rule, static
21375 component values and static subtypes also lead to simpler code.
21377 @node Static constant aggregates with static bounds
21378 @subsection Static constant aggregates with static bounds
21381 For the declarations:
21382 @smallexample @c ada
21383 type One_Dim is array (1..10) of integer;
21384 ar0 : constant One_Dim := (1, 2, 3, 4, 5, 6, 7, 8, 9, 0);
21388 GNAT generates no executable code: the constant ar0 is placed in static memory.
21389 The same is true for constant aggregates with named associations:
21391 @smallexample @c ada
21392 Cr1 : constant One_Dim := (4 => 16, 2 => 4, 3 => 9, 1 => 1, 5 .. 10 => 0);
21393 Cr3 : constant One_Dim := (others => 7777);
21397 The same is true for multidimensional constant arrays such as:
21399 @smallexample @c ada
21400 type two_dim is array (1..3, 1..3) of integer;
21401 Unit : constant two_dim := ( (1,0,0), (0,1,0), (0,0,1));
21405 The same is true for arrays of one-dimensional arrays: the following are
21408 @smallexample @c ada
21409 type ar1b is array (1..3) of boolean;
21410 type ar_ar is array (1..3) of ar1b;
21411 None : constant ar1b := (others => false); -- fully static
21412 None2 : constant ar_ar := (1..3 => None); -- fully static
21416 However, for multidimensional aggregates with named associations, GNAT will
21417 generate assignments and loops, even if all associations are static. The
21418 following two declarations generate a loop for the first dimension, and
21419 individual component assignments for the second dimension:
21421 @smallexample @c ada
21422 Zero1: constant two_dim := (1..3 => (1..3 => 0));
21423 Zero2: constant two_dim := (others => (others => 0));
21426 @node Constant aggregates with unconstrained nominal types
21427 @subsection Constant aggregates with unconstrained nominal types
21430 In such cases the aggregate itself establishes the subtype, so that
21431 associations with @code{others} cannot be used. GNAT determines the
21432 bounds for the actual subtype of the aggregate, and allocates the
21433 aggregate statically as well. No code is generated for the following:
21435 @smallexample @c ada
21436 type One_Unc is array (natural range <>) of integer;
21437 Cr_Unc : constant One_Unc := (12,24,36);
21440 @node Aggregates with static bounds
21441 @subsection Aggregates with static bounds
21444 In all previous examples the aggregate was the initial (and immutable) value
21445 of a constant. If the aggregate initializes a variable, then code is generated
21446 for it as a combination of individual assignments and loops over the target
21447 object. The declarations
21449 @smallexample @c ada
21450 Cr_Var1 : One_Dim := (2, 5, 7, 11, 0, 0, 0, 0, 0, 0);
21451 Cr_Var2 : One_Dim := (others > -1);
21455 generate the equivalent of
21457 @smallexample @c ada
21463 for I in Cr_Var2'range loop
21468 @node Aggregates with non-static bounds
21469 @subsection Aggregates with non-static bounds
21472 If the bounds of the aggregate are not statically compatible with the bounds
21473 of the nominal subtype of the target, then constraint checks have to be
21474 generated on the bounds. For a multidimensional array, constraint checks may
21475 have to be applied to sub-arrays individually, if they do not have statically
21476 compatible subtypes.
21478 @node Aggregates in assignment statements
21479 @subsection Aggregates in assignment statements
21482 In general, aggregate assignment requires the construction of a temporary,
21483 and a copy from the temporary to the target of the assignment. This is because
21484 it is not always possible to convert the assignment into a series of individual
21485 component assignments. For example, consider the simple case:
21487 @smallexample @c ada
21492 This cannot be converted into:
21494 @smallexample @c ada
21500 So the aggregate has to be built first in a separate location, and then
21501 copied into the target. GNAT recognizes simple cases where this intermediate
21502 step is not required, and the assignments can be performed in place, directly
21503 into the target. The following sufficient criteria are applied:
21507 The bounds of the aggregate are static, and the associations are static.
21509 The components of the aggregate are static constants, names of
21510 simple variables that are not renamings, or expressions not involving
21511 indexed components whose operands obey these rules.
21515 If any of these conditions are violated, the aggregate will be built in
21516 a temporary (created either by the front-end or the code generator) and then
21517 that temporary will be copied onto the target.
21519 @node The Size of Discriminated Records with Default Discriminants
21520 @section The Size of Discriminated Records with Default Discriminants
21523 If a discriminated type @code{T} has discriminants with default values, it is
21524 possible to declare an object of this type without providing an explicit
21527 @smallexample @c ada
21529 type Size is range 1..100;
21531 type Rec (D : Size := 15) is record
21532 Name : String (1..D);
21540 Such an object is said to be @emph{unconstrained}.
21541 The discriminant of the object
21542 can be modified by a full assignment to the object, as long as it preserves the
21543 relation between the value of the discriminant, and the value of the components
21546 @smallexample @c ada
21548 Word := (3, "yes");
21550 Word := (5, "maybe");
21552 Word := (5, "no"); -- raises Constraint_Error
21557 In order to support this behavior efficiently, an unconstrained object is
21558 given the maximum size that any value of the type requires. In the case
21559 above, @code{Word} has storage for the discriminant and for
21560 a @code{String} of length 100.
21561 It is important to note that unconstrained objects do not require dynamic
21562 allocation. It would be an improper implementation to place on the heap those
21563 components whose size depends on discriminants. (This improper implementation
21564 was used by some Ada83 compilers, where the @code{Name} component above
21566 been stored as a pointer to a dynamic string). Following the principle that
21567 dynamic storage management should never be introduced implicitly,
21568 an Ada compiler should reserve the full size for an unconstrained declared
21569 object, and place it on the stack.
21571 This maximum size approach
21572 has been a source of surprise to some users, who expect the default
21573 values of the discriminants to determine the size reserved for an
21574 unconstrained object: ``If the default is 15, why should the object occupy
21576 The answer, of course, is that the discriminant may be later modified,
21577 and its full range of values must be taken into account. This is why the
21582 type Rec (D : Positive := 15) is record
21583 Name : String (1..D);
21591 is flagged by the compiler with a warning:
21592 an attempt to create @code{Too_Large} will raise @code{Storage_Error},
21593 because the required size includes @code{Positive'Last}
21594 bytes. As the first example indicates, the proper approach is to declare an
21595 index type of ``reasonable'' range so that unconstrained objects are not too
21598 One final wrinkle: if the object is declared to be @code{aliased}, or if it is
21599 created in the heap by means of an allocator, then it is @emph{not}
21601 it is constrained by the default values of the discriminants, and those values
21602 cannot be modified by full assignment. This is because in the presence of
21603 aliasing all views of the object (which may be manipulated by different tasks,
21604 say) must be consistent, so it is imperative that the object, once created,
21607 @node Strict Conformance to the Ada Reference Manual
21608 @section Strict Conformance to the Ada Reference Manual
21611 The dynamic semantics defined by the Ada Reference Manual impose a set of
21612 run-time checks to be generated. By default, the GNAT compiler will insert many
21613 run-time checks into the compiled code, including most of those required by the
21614 Ada Reference Manual. However, there are three checks that are not enabled
21615 in the default mode for efficiency reasons: arithmetic overflow checking for
21616 integer operations (including division by zero), checks for access before
21617 elaboration on subprogram calls, and stack overflow checking (most operating
21618 systems do not perform this check by default).
21620 Strict conformance to the Ada Reference Manual can be achieved by adding
21621 three compiler options for overflow checking for integer operations
21622 (@option{-gnato}), dynamic checks for access-before-elaboration on subprogram
21623 calls and generic instantiations (@option{-gnatE}), and stack overflow
21624 checking (@option{-fstack-check}).
21626 Note that the result of a floating point arithmetic operation in overflow and
21627 invalid situations, when the @code{Machine_Overflows} attribute of the result
21628 type is @code{False}, is to generate IEEE NaN and infinite values. This is the
21629 case for machines compliant with the IEEE floating-point standard, but on
21630 machines that are not fully compliant with this standard, such as Alpha, the
21631 @option{-mieee} compiler flag must be used for achieving IEEE confirming
21632 behavior (although at the cost of a significant performance penalty), so
21633 infinite and NaN values are properly generated.
21636 @node Implementation of Ada 2012 Features
21637 @chapter Implementation of Ada 2012 Features
21638 @cindex Ada 2012 implementation status
21640 This chapter contains a complete list of Ada 2012 features that have been
21641 implemented as of GNAT version 6.4. Generally, these features are only
21642 available if the @option{-gnat12} (Ada 2012 features enabled) flag is set
21643 @cindex @option{-gnat12} option
21644 or if the configuration pragma @code{Ada_2012} is used.
21645 @cindex pragma @code{Ada_2012}
21646 @cindex configuration pragma @code{Ada_2012}
21647 @cindex @code{Ada_2012} configuration pragma
21648 However, new pragmas, attributes, and restrictions are
21649 unconditionally available, since the Ada 95 standard allows the addition of
21650 new pragmas, attributes, and restrictions (there are exceptions, which are
21651 documented in the individual descriptions), and also certain packages
21652 were made available in earlier versions of Ada.
21654 An ISO date (YYYY-MM-DD) appears in parentheses on the description line.
21655 This date shows the implementation date of the feature. Any wavefront
21656 subsequent to this date will contain the indicated feature, as will any
21657 subsequent releases. A date of 0000-00-00 means that GNAT has always
21658 implemented the feature, or implemented it as soon as it appeared as a
21659 binding interpretation.
21661 Each feature corresponds to an Ada Issue (``AI'') approved by the Ada
21662 standardization group (ISO/IEC JTC1/SC22/WG9) for inclusion in Ada 2012.
21663 The features are ordered based on the relevant sections of the Ada
21664 Reference Manual (``RM''). When a given AI relates to multiple points
21665 in the RM, the earliest is used.
21667 A complete description of the AIs may be found in
21668 @url{www.ada-auth.org/ai05-summary.html}.
21673 @emph{AI-0176 Quantified expressions (2010-09-29)}
21674 @cindex AI-0176 (Ada 2012 feature)
21677 Both universally and existentially quantified expressions are implemented.
21678 They use the new syntax for iterators proposed in AI05-139-2, as well as
21679 the standard Ada loop syntax.
21682 RM References: 1.01.04 (12) 2.09 (2/2) 4.04 (7) 4.05.09 (0)
21685 @emph{AI-0079 Allow @i{other_format} characters in source (2010-07-10)}
21686 @cindex AI-0079 (Ada 2012 feature)
21689 Wide characters in the unicode category @i{other_format} are now allowed in
21690 source programs between tokens, but not within a token such as an identifier.
21693 RM References: 2.01 (4/2) 2.02 (7)
21696 @emph{AI-0091 Do not allow @i{other_format} in identifiers (0000-00-00)}
21697 @cindex AI-0091 (Ada 2012 feature)
21700 Wide characters in the unicode category @i{other_format} are not permitted
21701 within an identifier, since this can be a security problem. The error
21702 message for this case has been improved to be more specific, but GNAT has
21703 never allowed such characters to appear in identifiers.
21706 RM References: 2.03 (3.1/2) 2.03 (4/2) 2.03 (5/2) 2.03 (5.1/2) 2.03 (5.2/2) 2.03 (5.3/2) 2.09 (2/2)
21709 @emph{AI-0100 Placement of pragmas (2010-07-01)}
21710 @cindex AI-0100 (Ada 2012 feature)
21713 This AI is an earlier version of AI-163. It simplifies the rules
21714 for legal placement of pragmas. In the case of lists that allow pragmas, if
21715 the list may have no elements, then the list may consist solely of pragmas.
21718 RM References: 2.08 (7)
21721 @emph{AI-0163 Pragmas in place of null (2010-07-01)}
21722 @cindex AI-0163 (Ada 2012 feature)
21725 A statement sequence may be composed entirely of pragmas. It is no longer
21726 necessary to add a dummy @code{null} statement to make the sequence legal.
21729 RM References: 2.08 (7) 2.08 (16)
21733 @emph{AI-0080 ``View of'' not needed if clear from context (0000-00-00)}
21734 @cindex AI-0080 (Ada 2012 feature)
21737 This is an editorial change only, described as non-testable in the AI.
21740 RM References: 3.01 (7)
21744 @emph{AI-0183 Aspect specifications (2010-08-16)}
21745 @cindex AI-0183 (Ada 2012 feature)
21748 Aspect specifications have been fully implemented except for pre and post-
21749 conditions, and type invariants, which have their own separate AI's. All
21750 forms of declarations listed in the AI are supported. The following is a
21751 list of the aspects supported (with GNAT implementation aspects marked)
21753 @multitable {@code{Preelaborable_Initialization}} {--GNAT}
21754 @item @code{Ada_2005} @tab -- GNAT
21755 @item @code{Ada_2012} @tab -- GNAT
21756 @item @code{Address} @tab
21757 @item @code{Alignment} @tab
21758 @item @code{Atomic} @tab
21759 @item @code{Atomic_Components} @tab
21760 @item @code{Bit_Order} @tab
21761 @item @code{Component_Size} @tab
21762 @item @code{Contract_Cases} @tab -- GNAT
21763 @item @code{Discard_Names} @tab
21764 @item @code{External_Tag} @tab
21765 @item @code{Favor_Top_Level} @tab -- GNAT
21766 @item @code{Inline} @tab
21767 @item @code{Inline_Always} @tab -- GNAT
21768 @item @code{Invariant} @tab -- GNAT
21769 @item @code{Machine_Radix} @tab
21770 @item @code{No_Return} @tab
21771 @item @code{Object_Size} @tab -- GNAT
21772 @item @code{Pack} @tab
21773 @item @code{Persistent_BSS} @tab -- GNAT
21774 @item @code{Post} @tab
21775 @item @code{Pre} @tab
21776 @item @code{Predicate} @tab
21777 @item @code{Preelaborable_Initialization} @tab
21778 @item @code{Pure_Function} @tab -- GNAT
21779 @item @code{Remote_Access_Type} @tab -- GNAT
21780 @item @code{Shared} @tab -- GNAT
21781 @item @code{Size} @tab
21782 @item @code{Storage_Pool} @tab
21783 @item @code{Storage_Size} @tab
21784 @item @code{Stream_Size} @tab
21785 @item @code{Suppress} @tab
21786 @item @code{Suppress_Debug_Info} @tab -- GNAT
21787 @item @code{Test_Case} @tab -- GNAT
21788 @item @code{Thread_Local_Storage} @tab -- GNAT
21789 @item @code{Type_Invariant} @tab
21790 @item @code{Unchecked_Union} @tab
21791 @item @code{Universal_Aliasing} @tab -- GNAT
21792 @item @code{Unmodified} @tab -- GNAT
21793 @item @code{Unreferenced} @tab -- GNAT
21794 @item @code{Unreferenced_Objects} @tab -- GNAT
21795 @item @code{Unsuppress} @tab
21796 @item @code{Value_Size} @tab -- GNAT
21797 @item @code{Volatile} @tab
21798 @item @code{Volatile_Components}
21799 @item @code{Warnings} @tab -- GNAT
21803 Note that for aspects with an expression, e.g. @code{Size}, the expression is
21804 treated like a default expression (visibility is analyzed at the point of
21805 occurrence of the aspect, but evaluation of the expression occurs at the
21806 freeze point of the entity involved).
21809 RM References: 3.02.01 (3) 3.02.02 (2) 3.03.01 (2/2) 3.08 (6)
21810 3.09.03 (1.1/2) 6.01 (2/2) 6.07 (2/2) 9.05.02 (2/2) 7.01 (3) 7.03
21811 (2) 7.03 (3) 9.01 (2/2) 9.01 (3/2) 9.04 (2/2) 9.04 (3/2)
21812 9.05.02 (2/2) 11.01 (2) 12.01 (3) 12.03 (2/2) 12.04 (2/2) 12.05 (2)
21813 12.06 (2.1/2) 12.06 (2.2/2) 12.07 (2) 13.01 (0.1/2) 13.03 (5/1)
21818 @emph{AI-0128 Inequality is a primitive operation (0000-00-00)}
21819 @cindex AI-0128 (Ada 2012 feature)
21822 If an equality operator ("=") is declared for a type, then the implicitly
21823 declared inequality operator ("/=") is a primitive operation of the type.
21824 This is the only reasonable interpretation, and is the one always implemented
21825 by GNAT, but the RM was not entirely clear in making this point.
21828 RM References: 3.02.03 (6) 6.06 (6)
21831 @emph{AI-0003 Qualified expressions as names (2010-07-11)}
21832 @cindex AI-0003 (Ada 2012 feature)
21835 In Ada 2012, a qualified expression is considered to be syntactically a name,
21836 meaning that constructs such as @code{A'(F(X)).B} are now legal. This is
21837 useful in disambiguating some cases of overloading.
21840 RM References: 3.03 (11) 3.03 (21) 4.01 (2) 4.04 (7) 4.07 (3)
21844 @emph{AI-0120 Constant instance of protected object (0000-00-00)}
21845 @cindex AI-0120 (Ada 2012 feature)
21848 This is an RM editorial change only. The section that lists objects that are
21849 constant failed to include the current instance of a protected object
21850 within a protected function. This has always been treated as a constant
21854 RM References: 3.03 (21)
21857 @emph{AI-0008 General access to constrained objects (0000-00-00)}
21858 @cindex AI-0008 (Ada 2012 feature)
21861 The wording in the RM implied that if you have a general access to a
21862 constrained object, it could be used to modify the discriminants. This was
21863 obviously not intended. @code{Constraint_Error} should be raised, and GNAT
21864 has always done so in this situation.
21867 RM References: 3.03 (23) 3.10.02 (26/2) 4.01 (9) 6.04.01 (17) 8.05.01 (5/2)
21871 @emph{AI-0093 Additional rules use immutably limited (0000-00-00)}
21872 @cindex AI-0093 (Ada 2012 feature)
21875 This is an editorial change only, to make more widespread use of the Ada 2012
21876 ``immutably limited''.
21879 RM References: 3.03 (23.4/3)
21884 @emph{AI-0096 Deriving from formal private types (2010-07-20)}
21885 @cindex AI-0096 (Ada 2012 feature)
21888 In general it is illegal for a type derived from a formal limited type to be
21889 nonlimited. This AI makes an exception to this rule: derivation is legal
21890 if it appears in the private part of the generic, and the formal type is not
21891 tagged. If the type is tagged, the legality check must be applied to the
21892 private part of the package.
21895 RM References: 3.04 (5.1/2) 6.02 (7)
21899 @emph{AI-0181 Soft hyphen is a non-graphic character (2010-07-23)}
21900 @cindex AI-0181 (Ada 2012 feature)
21903 From Ada 2005 on, soft hyphen is considered a non-graphic character, which
21904 means that it has a special name (@code{SOFT_HYPHEN}) in conjunction with the
21905 @code{Image} and @code{Value} attributes for the character types. Strictly
21906 speaking this is an inconsistency with Ada 95, but in practice the use of
21907 these attributes is so obscure that it will not cause problems.
21910 RM References: 3.05.02 (2/2) A.01 (35/2) A.03.03 (21)
21914 @emph{AI-0182 Additional forms for @code{Character'Value} (0000-00-00)}
21915 @cindex AI-0182 (Ada 2012 feature)
21918 This AI allows @code{Character'Value} to accept the string @code{'?'} where
21919 @code{?} is any character including non-graphic control characters. GNAT has
21920 always accepted such strings. It also allows strings such as
21921 @code{HEX_00000041} to be accepted, but GNAT does not take advantage of this
21922 permission and raises @code{Constraint_Error}, as is certainly still
21926 RM References: 3.05 (56/2)
21930 @emph{AI-0214 Defaulted discriminants for limited tagged (2010-10-01)}
21931 @cindex AI-0214 (Ada 2012 feature)
21934 Ada 2012 relaxes the restriction that forbids discriminants of tagged types
21935 to have default expressions by allowing them when the type is limited. It
21936 is often useful to define a default value for a discriminant even though
21937 it can't be changed by assignment.
21940 RM References: 3.07 (9.1/2) 3.07.02 (3)
21944 @emph{AI-0102 Some implicit conversions are illegal (0000-00-00)}
21945 @cindex AI-0102 (Ada 2012 feature)
21948 It is illegal to assign an anonymous access constant to an anonymous access
21949 variable. The RM did not have a clear rule to prevent this, but GNAT has
21950 always generated an error for this usage.
21953 RM References: 3.07 (16) 3.07.01 (9) 6.04.01 (6) 8.06 (27/2)
21957 @emph{AI-0158 Generalizing membership tests (2010-09-16)}
21958 @cindex AI-0158 (Ada 2012 feature)
21961 This AI extends the syntax of membership tests to simplify complex conditions
21962 that can be expressed as membership in a subset of values of any type. It
21963 introduces syntax for a list of expressions that may be used in loop contexts
21967 RM References: 3.08.01 (5) 4.04 (3) 4.05.02 (3) 4.05.02 (5) 4.05.02 (27)
21971 @emph{AI-0173 Testing if tags represent abstract types (2010-07-03)}
21972 @cindex AI-0173 (Ada 2012 feature)
21975 The function @code{Ada.Tags.Type_Is_Abstract} returns @code{True} if invoked
21976 with the tag of an abstract type, and @code{False} otherwise.
21979 RM References: 3.09 (7.4/2) 3.09 (12.4/2)
21984 @emph{AI-0076 function with controlling result (0000-00-00)}
21985 @cindex AI-0076 (Ada 2012 feature)
21988 This is an editorial change only. The RM defines calls with controlling
21989 results, but uses the term ``function with controlling result'' without an
21990 explicit definition.
21993 RM References: 3.09.02 (2/2)
21997 @emph{AI-0126 Dispatching with no declared operation (0000-00-00)}
21998 @cindex AI-0126 (Ada 2012 feature)
22001 This AI clarifies dispatching rules, and simply confirms that dispatching
22002 executes the operation of the parent type when there is no explicitly or
22003 implicitly declared operation for the descendant type. This has always been
22004 the case in all versions of GNAT.
22007 RM References: 3.09.02 (20/2) 3.09.02 (20.1/2) 3.09.02 (20.2/2)
22011 @emph{AI-0097 Treatment of abstract null extension (2010-07-19)}
22012 @cindex AI-0097 (Ada 2012 feature)
22015 The RM as written implied that in some cases it was possible to create an
22016 object of an abstract type, by having an abstract extension inherit a non-
22017 abstract constructor from its parent type. This mistake has been corrected
22018 in GNAT and in the RM, and this construct is now illegal.
22021 RM References: 3.09.03 (4/2)
22025 @emph{AI-0203 Extended return cannot be abstract (0000-00-00)}
22026 @cindex AI-0203 (Ada 2012 feature)
22029 A return_subtype_indication cannot denote an abstract subtype. GNAT has never
22030 permitted such usage.
22033 RM References: 3.09.03 (8/3)
22037 @emph{AI-0198 Inheriting abstract operators (0000-00-00)}
22038 @cindex AI-0198 (Ada 2012 feature)
22041 This AI resolves a conflict between two rules involving inherited abstract
22042 operations and predefined operators. If a derived numeric type inherits
22043 an abstract operator, it overrides the predefined one. This interpretation
22044 was always the one implemented in GNAT.
22047 RM References: 3.09.03 (4/3)
22050 @emph{AI-0073 Functions returning abstract types (2010-07-10)}
22051 @cindex AI-0073 (Ada 2012 feature)
22054 This AI covers a number of issues regarding returning abstract types. In
22055 particular generic functions cannot have abstract result types or access
22056 result types designated an abstract type. There are some other cases which
22057 are detailed in the AI. Note that this binding interpretation has not been
22058 retrofitted to operate before Ada 2012 mode, since it caused a significant
22059 number of regressions.
22062 RM References: 3.09.03 (8) 3.09.03 (10) 6.05 (8/2)
22066 @emph{AI-0070 Elaboration of interface types (0000-00-00)}
22067 @cindex AI-0070 (Ada 2012 feature)
22070 This is an editorial change only, there are no testable consequences short of
22071 checking for the absence of generated code for an interface declaration.
22074 RM References: 3.09.04 (18/2)
22078 @emph{AI-0208 Characteristics of incomplete views (0000-00-00)}
22079 @cindex AI-0208 (Ada 2012 feature)
22082 The wording in the Ada 2005 RM concerning characteristics of incomplete views
22083 was incorrect and implied that some programs intended to be legal were now
22084 illegal. GNAT had never considered such programs illegal, so it has always
22085 implemented the intent of this AI.
22088 RM References: 3.10.01 (2.4/2) 3.10.01 (2.6/2)
22092 @emph{AI-0162 Incomplete type completed by partial view (2010-09-15)}
22093 @cindex AI-0162 (Ada 2012 feature)
22096 Incomplete types are made more useful by allowing them to be completed by
22097 private types and private extensions.
22100 RM References: 3.10.01 (2.5/2) 3.10.01 (2.6/2) 3.10.01 (3) 3.10.01 (4/2)
22105 @emph{AI-0098 Anonymous subprogram access restrictions (0000-00-00)}
22106 @cindex AI-0098 (Ada 2012 feature)
22109 An unintentional omission in the RM implied some inconsistent restrictions on
22110 the use of anonymous access to subprogram values. These restrictions were not
22111 intentional, and have never been enforced by GNAT.
22114 RM References: 3.10.01 (6) 3.10.01 (9.2/2)
22118 @emph{AI-0199 Aggregate with anonymous access components (2010-07-14)}
22119 @cindex AI-0199 (Ada 2012 feature)
22122 A choice list in a record aggregate can include several components of
22123 (distinct) anonymous access types as long as they have matching designated
22127 RM References: 4.03.01 (16)
22131 @emph{AI-0220 Needed components for aggregates (0000-00-00)}
22132 @cindex AI-0220 (Ada 2012 feature)
22135 This AI addresses a wording problem in the RM that appears to permit some
22136 complex cases of aggregates with non-static discriminants. GNAT has always
22137 implemented the intended semantics.
22140 RM References: 4.03.01 (17)
22143 @emph{AI-0147 Conditional expressions (2009-03-29)}
22144 @cindex AI-0147 (Ada 2012 feature)
22147 Conditional expressions are permitted. The form of such an expression is:
22150 (@b{if} @i{expr} @b{then} @i{expr} @{@b{elsif} @i{expr} @b{then} @i{expr}@} [@b{else} @i{expr}])
22153 The parentheses can be omitted in contexts where parentheses are present
22154 anyway, such as subprogram arguments and pragma arguments. If the @b{else}
22155 clause is omitted, @b{else True} is assumed;
22156 thus @code{(@b{if} A @b{then} B)} is a way to conveniently represent
22157 @emph{(A implies B)} in standard logic.
22160 RM References: 4.03.03 (15) 4.04 (1) 4.04 (7) 4.05.07 (0) 4.07 (2)
22161 4.07 (3) 4.09 (12) 4.09 (33) 5.03 (3) 5.03 (4) 7.05 (2.1/2)
22165 @emph{AI-0037 Out-of-range box associations in aggregate (0000-00-00)}
22166 @cindex AI-0037 (Ada 2012 feature)
22169 This AI confirms that an association of the form @code{Indx => <>} in an
22170 array aggregate must raise @code{Constraint_Error} if @code{Indx}
22171 is out of range. The RM specified a range check on other associations, but
22172 not when the value of the association was defaulted. GNAT has always inserted
22173 a constraint check on the index value.
22176 RM References: 4.03.03 (29)
22180 @emph{AI-0123 Composability of equality (2010-04-13)}
22181 @cindex AI-0123 (Ada 2012 feature)
22184 Equality of untagged record composes, so that the predefined equality for a
22185 composite type that includes a component of some untagged record type
22186 @code{R} uses the equality operation of @code{R} (which may be user-defined
22187 or predefined). This makes the behavior of untagged records identical to that
22188 of tagged types in this respect.
22190 This change is an incompatibility with previous versions of Ada, but it
22191 corrects a non-uniformity that was often a source of confusion. Analysis of
22192 a large number of industrial programs indicates that in those rare cases
22193 where a composite type had an untagged record component with a user-defined
22194 equality, either there was no use of the composite equality, or else the code
22195 expected the same composability as for tagged types, and thus had a bug that
22196 would be fixed by this change.
22199 RM References: 4.05.02 (9.7/2) 4.05.02 (14) 4.05.02 (15) 4.05.02 (24)
22204 @emph{AI-0088 The value of exponentiation (0000-00-00)}
22205 @cindex AI-0088 (Ada 2012 feature)
22208 This AI clarifies the equivalence rule given for the dynamic semantics of
22209 exponentiation: the value of the operation can be obtained by repeated
22210 multiplication, but the operation can be implemented otherwise (for example
22211 using the familiar divide-by-two-and-square algorithm, even if this is less
22212 accurate), and does not imply repeated reads of a volatile base.
22215 RM References: 4.05.06 (11)
22218 @emph{AI-0188 Case expressions (2010-01-09)}
22219 @cindex AI-0188 (Ada 2012 feature)
22222 Case expressions are permitted. This allows use of constructs such as:
22224 X := (@b{case} Y @b{is when} 1 => 2, @b{when} 2 => 3, @b{when others} => 31)
22228 RM References: 4.05.07 (0) 4.05.08 (0) 4.09 (12) 4.09 (33)
22231 @emph{AI-0104 Null exclusion and uninitialized allocator (2010-07-15)}
22232 @cindex AI-0104 (Ada 2012 feature)
22235 The assignment @code{Ptr := @b{new not null} Some_Ptr;} will raise
22236 @code{Constraint_Error} because the default value of the allocated object is
22237 @b{null}. This useless construct is illegal in Ada 2012.
22240 RM References: 4.08 (2)
22243 @emph{AI-0157 Allocation/Deallocation from empty pool (2010-07-11)}
22244 @cindex AI-0157 (Ada 2012 feature)
22247 Allocation and Deallocation from an empty storage pool (i.e. allocation or
22248 deallocation of a pointer for which a static storage size clause of zero
22249 has been given) is now illegal and is detected as such. GNAT
22250 previously gave a warning but not an error.
22253 RM References: 4.08 (5.3/2) 13.11.02 (4) 13.11.02 (17)
22256 @emph{AI-0179 Statement not required after label (2010-04-10)}
22257 @cindex AI-0179 (Ada 2012 feature)
22260 It is not necessary to have a statement following a label, so a label
22261 can appear at the end of a statement sequence without the need for putting a
22262 null statement afterwards, but it is not allowable to have only labels and
22263 no real statements in a statement sequence.
22266 RM References: 5.01 (2)
22270 @emph{AI-139-2 Syntactic sugar for iterators (2010-09-29)}
22271 @cindex AI-139-2 (Ada 2012 feature)
22274 The new syntax for iterating over arrays and containers is now implemented.
22275 Iteration over containers is for now limited to read-only iterators. Only
22276 default iterators are supported, with the syntax: @code{@b{for} Elem @b{of} C}.
22279 RM References: 5.05
22282 @emph{AI-0134 Profiles must match for full conformance (0000-00-00)}
22283 @cindex AI-0134 (Ada 2012 feature)
22286 For full conformance, the profiles of anonymous-access-to-subprogram
22287 parameters must match. GNAT has always enforced this rule.
22290 RM References: 6.03.01 (18)
22293 @emph{AI-0207 Mode conformance and access constant (0000-00-00)}
22294 @cindex AI-0207 (Ada 2012 feature)
22297 This AI confirms that access_to_constant indication must match for mode
22298 conformance. This was implemented in GNAT when the qualifier was originally
22299 introduced in Ada 2005.
22302 RM References: 6.03.01 (16/2)
22306 @emph{AI-0046 Null exclusion match for full conformance (2010-07-17)}
22307 @cindex AI-0046 (Ada 2012 feature)
22310 For full conformance, in the case of access parameters, the null exclusion
22311 must match (either both or neither must have @code{@b{not null}}).
22314 RM References: 6.03.02 (18)
22318 @emph{AI-0118 The association of parameter associations (0000-00-00)}
22319 @cindex AI-0118 (Ada 2012 feature)
22322 This AI clarifies the rules for named associations in subprogram calls and
22323 generic instantiations. The rules have been in place since Ada 83.
22326 RM References: 6.04.01 (2) 12.03 (9)
22330 @emph{AI-0196 Null exclusion tests for out parameters (0000-00-00)}
22331 @cindex AI-0196 (Ada 2012 feature)
22334 Null exclusion checks are not made for @code{@b{out}} parameters when
22335 evaluating the actual parameters. GNAT has never generated these checks.
22338 RM References: 6.04.01 (13)
22341 @emph{AI-0015 Constant return objects (0000-00-00)}
22342 @cindex AI-0015 (Ada 2012 feature)
22345 The return object declared in an @i{extended_return_statement} may be
22346 declared constant. This was always intended, and GNAT has always allowed it.
22349 RM References: 6.05 (2.1/2) 3.03 (10/2) 3.03 (21) 6.05 (5/2)
22354 @emph{AI-0032 Extended return for class-wide functions (0000-00-00)}
22355 @cindex AI-0032 (Ada 2012 feature)
22358 If a function returns a class-wide type, the object of an extended return
22359 statement can be declared with a specific type that is covered by the class-
22360 wide type. This has been implemented in GNAT since the introduction of
22361 extended returns. Note AI-0103 complements this AI by imposing matching
22362 rules for constrained return types.
22365 RM References: 6.05 (5.2/2) 6.05 (5.3/2) 6.05 (5.6/2) 6.05 (5.8/2)
22369 @emph{AI-0103 Static matching for extended return (2010-07-23)}
22370 @cindex AI-0103 (Ada 2012 feature)
22373 If the return subtype of a function is an elementary type or a constrained
22374 type, the subtype indication in an extended return statement must match
22375 statically this return subtype.
22378 RM References: 6.05 (5.2/2)
22382 @emph{AI-0058 Abnormal completion of an extended return (0000-00-00)}
22383 @cindex AI-0058 (Ada 2012 feature)
22386 The RM had some incorrect wording implying wrong treatment of abnormal
22387 completion in an extended return. GNAT has always implemented the intended
22388 correct semantics as described by this AI.
22391 RM References: 6.05 (22/2)
22395 @emph{AI-0050 Raising Constraint_Error early for function call (0000-00-00)}
22396 @cindex AI-0050 (Ada 2012 feature)
22399 The implementation permissions for raising @code{Constraint_Error} early on a function call when it was clear an exception would be raised were over-permissive and allowed mishandling of discriminants in some cases. GNAT did
22400 not take advantage of these incorrect permissions in any case.
22403 RM References: 6.05 (24/2)
22407 @emph{AI-0125 Nonoverridable operations of an ancestor (2010-09-28)}
22408 @cindex AI-0125 (Ada 2012 feature)
22411 In Ada 2012, the declaration of a primitive operation of a type extension
22412 or private extension can also override an inherited primitive that is not
22413 visible at the point of this declaration.
22416 RM References: 7.03.01 (6) 8.03 (23) 8.03.01 (5/2) 8.03.01 (6/2)
22419 @emph{AI-0062 Null exclusions and deferred constants (0000-00-00)}
22420 @cindex AI-0062 (Ada 2012 feature)
22423 A full constant may have a null exclusion even if its associated deferred
22424 constant does not. GNAT has always allowed this.
22427 RM References: 7.04 (6/2) 7.04 (7.1/2)
22431 @emph{AI-0178 Incomplete views are limited (0000-00-00)}
22432 @cindex AI-0178 (Ada 2012 feature)
22435 This AI clarifies the role of incomplete views and plugs an omission in the
22436 RM. GNAT always correctly restricted the use of incomplete views and types.
22439 RM References: 7.05 (3/2) 7.05 (6/2)
22442 @emph{AI-0087 Actual for formal nonlimited derived type (2010-07-15)}
22443 @cindex AI-0087 (Ada 2012 feature)
22446 The actual for a formal nonlimited derived type cannot be limited. In
22447 particular, a formal derived type that extends a limited interface but which
22448 is not explicitly limited cannot be instantiated with a limited type.
22451 RM References: 7.05 (5/2) 12.05.01 (5.1/2)
22454 @emph{AI-0099 Tag determines whether finalization needed (0000-00-00)}
22455 @cindex AI-0099 (Ada 2012 feature)
22458 This AI clarifies that ``needs finalization'' is part of dynamic semantics,
22459 and therefore depends on the run-time characteristics of an object (i.e. its
22460 tag) and not on its nominal type. As the AI indicates: ``we do not expect
22461 this to affect any implementation''.
22464 RM References: 7.06.01 (6) 7.06.01 (7) 7.06.01 (8) 7.06.01 (9/2)
22469 @emph{AI-0064 Redundant finalization rule (0000-00-00)}
22470 @cindex AI-0064 (Ada 2012 feature)
22473 This is an editorial change only. The intended behavior is already checked
22474 by an existing ACATS test, which GNAT has always executed correctly.
22477 RM References: 7.06.01 (17.1/1)
22480 @emph{AI-0026 Missing rules for Unchecked_Union (2010-07-07)}
22481 @cindex AI-0026 (Ada 2012 feature)
22484 Record representation clauses concerning Unchecked_Union types cannot mention
22485 the discriminant of the type. The type of a component declared in the variant
22486 part of an Unchecked_Union cannot be controlled, have controlled components,
22487 nor have protected or task parts. If an Unchecked_Union type is declared
22488 within the body of a generic unit or its descendants, then the type of a
22489 component declared in the variant part cannot be a formal private type or a
22490 formal private extension declared within the same generic unit.
22493 RM References: 7.06 (9.4/2) B.03.03 (9/2) B.03.03 (10/2)
22497 @emph{AI-0205 Extended return declares visible name (0000-00-00)}
22498 @cindex AI-0205 (Ada 2012 feature)
22501 This AI corrects a simple omission in the RM. Return objects have always
22502 been visible within an extended return statement.
22505 RM References: 8.03 (17)
22509 @emph{AI-0042 Overriding versus implemented-by (0000-00-00)}
22510 @cindex AI-0042 (Ada 2012 feature)
22513 This AI fixes a wording gap in the RM. An operation of a synchronized
22514 interface can be implemented by a protected or task entry, but the abstract
22515 operation is not being overridden in the usual sense, and it must be stated
22516 separately that this implementation is legal. This has always been the case
22520 RM References: 9.01 (9.2/2) 9.04 (11.1/2)
22523 @emph{AI-0030 Requeue on synchronized interfaces (2010-07-19)}
22524 @cindex AI-0030 (Ada 2012 feature)
22527 Requeue is permitted to a protected, synchronized or task interface primitive
22528 providing it is known that the overriding operation is an entry. Otherwise
22529 the requeue statement has the same effect as a procedure call. Use of pragma
22530 @code{Implemented} provides a way to impose a static requirement on the
22531 overriding operation by adhering to one of the implementation kinds: entry,
22532 protected procedure or any of the above.
22535 RM References: 9.05 (9) 9.05.04 (2) 9.05.04 (3) 9.05.04 (5)
22536 9.05.04 (6) 9.05.04 (7) 9.05.04 (12)
22540 @emph{AI-0201 Independence of atomic object components (2010-07-22)}
22541 @cindex AI-0201 (Ada 2012 feature)
22544 If an Atomic object has a pragma @code{Pack} or a @code{Component_Size}
22545 attribute, then individual components may not be addressable by independent
22546 tasks. However, if the representation clause has no effect (is confirming),
22547 then independence is not compromised. Furthermore, in GNAT, specification of
22548 other appropriately addressable component sizes (e.g. 16 for 8-bit
22549 characters) also preserves independence. GNAT now gives very clear warnings
22550 both for the declaration of such a type, and for any assignment to its components.
22553 RM References: 9.10 (1/3) C.06 (22/2) C.06 (23/2)
22556 @emph{AI-0009 Pragma Independent[_Components] (2010-07-23)}
22557 @cindex AI-0009 (Ada 2012 feature)
22560 This AI introduces the new pragmas @code{Independent} and
22561 @code{Independent_Components},
22562 which control guaranteeing independence of access to objects and components.
22563 The AI also requires independence not unaffected by confirming rep clauses.
22566 RM References: 9.10 (1) 13.01 (15/1) 13.02 (9) 13.03 (13) C.06 (2)
22567 C.06 (4) C.06 (6) C.06 (9) C.06 (13) C.06 (14)
22571 @emph{AI-0072 Task signalling using 'Terminated (0000-00-00)}
22572 @cindex AI-0072 (Ada 2012 feature)
22575 This AI clarifies that task signalling for reading @code{'Terminated} only
22576 occurs if the result is True. GNAT semantics has always been consistent with
22577 this notion of task signalling.
22580 RM References: 9.10 (6.1/1)
22583 @emph{AI-0108 Limited incomplete view and discriminants (0000-00-00)}
22584 @cindex AI-0108 (Ada 2012 feature)
22587 This AI confirms that an incomplete type from a limited view does not have
22588 discriminants. This has always been the case in GNAT.
22591 RM References: 10.01.01 (12.3/2)
22594 @emph{AI-0129 Limited views and incomplete types (0000-00-00)}
22595 @cindex AI-0129 (Ada 2012 feature)
22598 This AI clarifies the description of limited views: a limited view of a
22599 package includes only one view of a type that has an incomplete declaration
22600 and a full declaration (there is no possible ambiguity in a client package).
22601 This AI also fixes an omission: a nested package in the private part has no
22602 limited view. GNAT always implemented this correctly.
22605 RM References: 10.01.01 (12.2/2) 10.01.01 (12.3/2)
22610 @emph{AI-0077 Limited withs and scope of declarations (0000-00-00)}
22611 @cindex AI-0077 (Ada 2012 feature)
22614 This AI clarifies that a declaration does not include a context clause,
22615 and confirms that it is illegal to have a context in which both a limited
22616 and a nonlimited view of a package are accessible. Such double visibility
22617 was always rejected by GNAT.
22620 RM References: 10.01.02 (12/2) 10.01.02 (21/2) 10.01.02 (22/2)
22623 @emph{AI-0122 Private with and children of generics (0000-00-00)}
22624 @cindex AI-0122 (Ada 2012 feature)
22627 This AI clarifies the visibility of private children of generic units within
22628 instantiations of a parent. GNAT has always handled this correctly.
22631 RM References: 10.01.02 (12/2)
22636 @emph{AI-0040 Limited with clauses on descendant (0000-00-00)}
22637 @cindex AI-0040 (Ada 2012 feature)
22640 This AI confirms that a limited with clause in a child unit cannot name
22641 an ancestor of the unit. This has always been checked in GNAT.
22644 RM References: 10.01.02 (20/2)
22647 @emph{AI-0132 Placement of library unit pragmas (0000-00-00)}
22648 @cindex AI-0132 (Ada 2012 feature)
22651 This AI fills a gap in the description of library unit pragmas. The pragma
22652 clearly must apply to a library unit, even if it does not carry the name
22653 of the enclosing unit. GNAT has always enforced the required check.
22656 RM References: 10.01.05 (7)
22660 @emph{AI-0034 Categorization of limited views (0000-00-00)}
22661 @cindex AI-0034 (Ada 2012 feature)
22664 The RM makes certain limited with clauses illegal because of categorization
22665 considerations, when the corresponding normal with would be legal. This is
22666 not intended, and GNAT has always implemented the recommended behavior.
22669 RM References: 10.02.01 (11/1) 10.02.01 (17/2)
22673 @emph{AI-0035 Inconsistencies with Pure units (0000-00-00)}
22674 @cindex AI-0035 (Ada 2012 feature)
22677 This AI remedies some inconsistencies in the legality rules for Pure units.
22678 Derived access types are legal in a pure unit (on the assumption that the
22679 rule for a zero storage pool size has been enforced on the ancestor type).
22680 The rules are enforced in generic instances and in subunits. GNAT has always
22681 implemented the recommended behavior.
22684 RM References: 10.02.01 (15.1/2) 10.02.01 (15.4/2) 10.02.01 (15.5/2) 10.02.01 (17/2)
22688 @emph{AI-0219 Pure permissions and limited parameters (2010-05-25)}
22689 @cindex AI-0219 (Ada 2012 feature)
22692 This AI refines the rules for the cases with limited parameters which do not
22693 allow the implementations to omit ``redundant''. GNAT now properly conforms
22694 to the requirements of this binding interpretation.
22697 RM References: 10.02.01 (18/2)
22700 @emph{AI-0043 Rules about raising exceptions (0000-00-00)}
22701 @cindex AI-0043 (Ada 2012 feature)
22704 This AI covers various omissions in the RM regarding the raising of
22705 exceptions. GNAT has always implemented the intended semantics.
22708 RM References: 11.04.01 (10.1/2) 11 (2)
22712 @emph{AI-0200 Mismatches in formal package declarations (0000-00-00)}
22713 @cindex AI-0200 (Ada 2012 feature)
22716 This AI plugs a gap in the RM which appeared to allow some obviously intended
22717 illegal instantiations. GNAT has never allowed these instantiations.
22720 RM References: 12.07 (16)
22724 @emph{AI-0112 Detection of duplicate pragmas (2010-07-24)}
22725 @cindex AI-0112 (Ada 2012 feature)
22728 This AI concerns giving names to various representation aspects, but the
22729 practical effect is simply to make the use of duplicate
22730 @code{Atomic}[@code{_Components}],
22731 @code{Volatile}[@code{_Components}] and
22732 @code{Independent}[@code{_Components}] pragmas illegal, and GNAT
22733 now performs this required check.
22736 RM References: 13.01 (8)
22739 @emph{AI-0106 No representation pragmas on generic formals (0000-00-00)}
22740 @cindex AI-0106 (Ada 2012 feature)
22743 The RM appeared to allow representation pragmas on generic formal parameters,
22744 but this was not intended, and GNAT has never permitted this usage.
22747 RM References: 13.01 (9.1/1)
22751 @emph{AI-0012 Pack/Component_Size for aliased/atomic (2010-07-15)}
22752 @cindex AI-0012 (Ada 2012 feature)
22755 It is now illegal to give an inappropriate component size or a pragma
22756 @code{Pack} that attempts to change the component size in the case of atomic
22757 or aliased components. Previously GNAT ignored such an attempt with a
22761 RM References: 13.02 (6.1/2) 13.02 (7) C.06 (10) C.06 (11) C.06 (21)
22765 @emph{AI-0039 Stream attributes cannot be dynamic (0000-00-00)}
22766 @cindex AI-0039 (Ada 2012 feature)
22769 The RM permitted the use of dynamic expressions (such as @code{ptr.@b{all})}
22770 for stream attributes, but these were never useful and are now illegal. GNAT
22771 has always regarded such expressions as illegal.
22774 RM References: 13.03 (4) 13.03 (6) 13.13.02 (38/2)
22778 @emph{AI-0095 Address of intrinsic subprograms (0000-00-00)}
22779 @cindex AI-0095 (Ada 2012 feature)
22782 The prefix of @code{'Address} cannot statically denote a subprogram with
22783 convention @code{Intrinsic}. The use of the @code{Address} attribute raises
22784 @code{Program_Error} if the prefix denotes a subprogram with convention
22788 RM References: 13.03 (11/1)
22792 @emph{AI-0116 Alignment of class-wide objects (0000-00-00)}
22793 @cindex AI-0116 (Ada 2012 feature)
22796 This AI requires that the alignment of a class-wide object be no greater
22797 than the alignment of any type in the class. GNAT has always followed this
22801 RM References: 13.03 (29) 13.11 (16)
22805 @emph{AI-0146 Type invariants (2009-09-21)}
22806 @cindex AI-0146 (Ada 2012 feature)
22809 Type invariants may be specified for private types using the aspect notation.
22810 Aspect @code{Type_Invariant} may be specified for any private type,
22811 @code{Type_Invariant'Class} can
22812 only be specified for tagged types, and is inherited by any descendent of the
22813 tagged types. The invariant is a boolean expression that is tested for being
22814 true in the following situations: conversions to the private type, object
22815 declarations for the private type that are default initialized, and
22817 parameters and returned result on return from any primitive operation for
22818 the type that is visible to a client.
22819 GNAT defines the synonyms @code{Invariant} for @code{Type_Invariant} and
22820 @code{Invariant'Class} for @code{Type_Invariant'Class}.
22823 RM References: 13.03.03 (00)
22826 @emph{AI-0078 Relax Unchecked_Conversion alignment rules (0000-00-00)}
22827 @cindex AI-0078 (Ada 2012 feature)
22830 In Ada 2012, compilers are required to support unchecked conversion where the
22831 target alignment is a multiple of the source alignment. GNAT always supported
22832 this case (and indeed all cases of differing alignments, doing copies where
22833 required if the alignment was reduced).
22836 RM References: 13.09 (7)
22840 @emph{AI-0195 Invalid value handling is implementation defined (2010-07-03)}
22841 @cindex AI-0195 (Ada 2012 feature)
22844 The handling of invalid values is now designated to be implementation
22845 defined. This is a documentation change only, requiring Annex M in the GNAT
22846 Reference Manual to document this handling.
22847 In GNAT, checks for invalid values are made
22848 only when necessary to avoid erroneous behavior. Operations like assignments
22849 which cannot cause erroneous behavior ignore the possibility of invalid
22850 values and do not do a check. The date given above applies only to the
22851 documentation change, this behavior has always been implemented by GNAT.
22854 RM References: 13.09.01 (10)
22857 @emph{AI-0193 Alignment of allocators (2010-09-16)}
22858 @cindex AI-0193 (Ada 2012 feature)
22861 This AI introduces a new attribute @code{Max_Alignment_For_Allocation},
22862 analogous to @code{Max_Size_In_Storage_Elements}, but for alignment instead
22866 RM References: 13.11 (16) 13.11 (21) 13.11.01 (0) 13.11.01 (1)
22867 13.11.01 (2) 13.11.01 (3)
22871 @emph{AI-0177 Parameterized expressions (2010-07-10)}
22872 @cindex AI-0177 (Ada 2012 feature)
22875 The new Ada 2012 notion of parameterized expressions is implemented. The form
22878 @i{function specification} @b{is} (@i{expression})
22882 This is exactly equivalent to the
22883 corresponding function body that returns the expression, but it can appear
22884 in a package spec. Note that the expression must be parenthesized.
22887 RM References: 13.11.01 (3/2)
22890 @emph{AI-0033 Attach/Interrupt_Handler in generic (2010-07-24)}
22891 @cindex AI-0033 (Ada 2012 feature)
22894 Neither of these two pragmas may appear within a generic template, because
22895 the generic might be instantiated at other than the library level.
22898 RM References: 13.11.02 (16) C.03.01 (7/2) C.03.01 (8/2)
22902 @emph{AI-0161 Restriction No_Default_Stream_Attributes (2010-09-11)}
22903 @cindex AI-0161 (Ada 2012 feature)
22906 A new restriction @code{No_Default_Stream_Attributes} prevents the use of any
22907 of the default stream attributes for elementary types. If this restriction is
22908 in force, then it is necessary to provide explicit subprograms for any
22909 stream attributes used.
22912 RM References: 13.12.01 (4/2) 13.13.02 (40/2) 13.13.02 (52/2)
22915 @emph{AI-0194 Value of Stream_Size attribute (0000-00-00)}
22916 @cindex AI-0194 (Ada 2012 feature)
22919 The @code{Stream_Size} attribute returns the default number of bits in the
22920 stream representation of the given type.
22921 This value is not affected by the presence
22922 of stream subprogram attributes for the type. GNAT has always implemented
22923 this interpretation.
22926 RM References: 13.13.02 (1.2/2)
22929 @emph{AI-0109 Redundant check in S'Class'Input (0000-00-00)}
22930 @cindex AI-0109 (Ada 2012 feature)
22933 This AI is an editorial change only. It removes the need for a tag check
22934 that can never fail.
22937 RM References: 13.13.02 (34/2)
22940 @emph{AI-0007 Stream read and private scalar types (0000-00-00)}
22941 @cindex AI-0007 (Ada 2012 feature)
22944 The RM as written appeared to limit the possibilities of declaring read
22945 attribute procedures for private scalar types. This limitation was not
22946 intended, and has never been enforced by GNAT.
22949 RM References: 13.13.02 (50/2) 13.13.02 (51/2)
22953 @emph{AI-0065 Remote access types and external streaming (0000-00-00)}
22954 @cindex AI-0065 (Ada 2012 feature)
22957 This AI clarifies the fact that all remote access types support external
22958 streaming. This fixes an obvious oversight in the definition of the
22959 language, and GNAT always implemented the intended correct rules.
22962 RM References: 13.13.02 (52/2)
22965 @emph{AI-0019 Freezing of primitives for tagged types (0000-00-00)}
22966 @cindex AI-0019 (Ada 2012 feature)
22969 The RM suggests that primitive subprograms of a specific tagged type are
22970 frozen when the tagged type is frozen. This would be an incompatible change
22971 and is not intended. GNAT has never attempted this kind of freezing and its
22972 behavior is consistent with the recommendation of this AI.
22975 RM References: 13.14 (2) 13.14 (3/1) 13.14 (8.1/1) 13.14 (10) 13.14 (14) 13.14 (15.1/2)
22978 @emph{AI-0017 Freezing and incomplete types (0000-00-00)}
22979 @cindex AI-0017 (Ada 2012 feature)
22982 So-called ``Taft-amendment types'' (i.e., types that are completed in package
22983 bodies) are not frozen by the occurrence of bodies in the
22984 enclosing declarative part. GNAT always implemented this properly.
22987 RM References: 13.14 (3/1)
22991 @emph{AI-0060 Extended definition of remote access types (0000-00-00)}
22992 @cindex AI-0060 (Ada 2012 feature)
22995 This AI extends the definition of remote access types to include access
22996 to limited, synchronized, protected or task class-wide interface types.
22997 GNAT already implemented this extension.
23000 RM References: A (4) E.02.02 (9/1) E.02.02 (9.2/1) E.02.02 (14/2) E.02.02 (18)
23003 @emph{AI-0114 Classification of letters (0000-00-00)}
23004 @cindex AI-0114 (Ada 2012 feature)
23007 The code points 170 (@code{FEMININE ORDINAL INDICATOR}),
23008 181 (@code{MICRO SIGN}), and
23009 186 (@code{MASCULINE ORDINAL INDICATOR}) are technically considered
23010 lower case letters by Unicode.
23011 However, they are not allowed in identifiers, and they
23012 return @code{False} to @code{Ada.Characters.Handling.Is_Letter/Is_Lower}.
23013 This behavior is consistent with that defined in Ada 95.
23016 RM References: A.03.02 (59) A.04.06 (7)
23020 @emph{AI-0185 Ada.Wide_[Wide_]Characters.Handling (2010-07-06)}
23021 @cindex AI-0185 (Ada 2012 feature)
23024 Two new packages @code{Ada.Wide_[Wide_]Characters.Handling} provide
23025 classification functions for @code{Wide_Character} and
23026 @code{Wide_Wide_Character}, as well as providing
23027 case folding routines for @code{Wide_[Wide_]Character} and
23028 @code{Wide_[Wide_]String}.
23031 RM References: A.03.05 (0) A.03.06 (0)
23035 @emph{AI-0031 Add From parameter to Find_Token (2010-07-25)}
23036 @cindex AI-0031 (Ada 2012 feature)
23039 A new version of @code{Find_Token} is added to all relevant string packages,
23040 with an extra parameter @code{From}. Instead of starting at the first
23041 character of the string, the search for a matching Token starts at the
23042 character indexed by the value of @code{From}.
23043 These procedures are available in all versions of Ada
23044 but if used in versions earlier than Ada 2012 they will generate a warning
23045 that an Ada 2012 subprogram is being used.
23048 RM References: A.04.03 (16) A.04.03 (67) A.04.03 (68/1) A.04.04 (51)
23053 @emph{AI-0056 Index on null string returns zero (0000-00-00)}
23054 @cindex AI-0056 (Ada 2012 feature)
23057 The wording in the Ada 2005 RM implied an incompatible handling of the
23058 @code{Index} functions, resulting in raising an exception instead of
23059 returning zero in some situations.
23060 This was not intended and has been corrected.
23061 GNAT always returned zero, and is thus consistent with this AI.
23064 RM References: A.04.03 (56.2/2) A.04.03 (58.5/2)
23068 @emph{AI-0137 String encoding package (2010-03-25)}
23069 @cindex AI-0137 (Ada 2012 feature)
23072 The packages @code{Ada.Strings.UTF_Encoding}, together with its child
23073 packages, @code{Conversions}, @code{Strings}, @code{Wide_Strings},
23074 and @code{Wide_Wide_Strings} have been
23075 implemented. These packages (whose documentation can be found in the spec
23076 files @file{a-stuten.ads}, @file{a-suenco.ads}, @file{a-suenst.ads},
23077 @file{a-suewst.ads}, @file{a-suezst.ads}) allow encoding and decoding of
23078 @code{String}, @code{Wide_String}, and @code{Wide_Wide_String}
23079 values using UTF coding schemes (including UTF-8, UTF-16LE, UTF-16BE, and
23080 UTF-16), as well as conversions between the different UTF encodings. With
23081 the exception of @code{Wide_Wide_Strings}, these packages are available in
23082 Ada 95 and Ada 2005 mode as well as Ada 2012 mode.
23083 The @code{Wide_Wide_Strings package}
23084 is available in Ada 2005 mode as well as Ada 2012 mode (but not in Ada 95
23085 mode since it uses @code{Wide_Wide_Character}).
23088 RM References: A.04.11
23091 @emph{AI-0038 Minor errors in Text_IO (0000-00-00)}
23092 @cindex AI-0038 (Ada 2012 feature)
23095 These are minor errors in the description on three points. The intent on
23096 all these points has always been clear, and GNAT has always implemented the
23097 correct intended semantics.
23100 RM References: A.10.05 (37) A.10.07 (8/1) A.10.07 (10) A.10.07 (12) A.10.08 (10) A.10.08 (24)
23103 @emph{AI-0044 Restrictions on container instantiations (0000-00-00)}
23104 @cindex AI-0044 (Ada 2012 feature)
23107 This AI places restrictions on allowed instantiations of generic containers.
23108 These restrictions are not checked by the compiler, so there is nothing to
23109 change in the implementation. This affects only the RM documentation.
23112 RM References: A.18 (4/2) A.18.02 (231/2) A.18.03 (145/2) A.18.06 (56/2) A.18.08 (66/2) A.18.09 (79/2) A.18.26 (5/2) A.18.26 (9/2)
23115 @emph{AI-0127 Adding Locale Capabilities (2010-09-29)}
23116 @cindex AI-0127 (Ada 2012 feature)
23119 This package provides an interface for identifying the current locale.
23122 RM References: A.19 A.19.01 A.19.02 A.19.03 A.19.05 A.19.06
23123 A.19.07 A.19.08 A.19.09 A.19.10 A.19.11 A.19.12 A.19.13
23128 @emph{AI-0002 Export C with unconstrained arrays (0000-00-00)}
23129 @cindex AI-0002 (Ada 2012 feature)
23132 The compiler is not required to support exporting an Ada subprogram with
23133 convention C if there are parameters or a return type of an unconstrained
23134 array type (such as @code{String}). GNAT allows such declarations but
23135 generates warnings. It is possible, but complicated, to write the
23136 corresponding C code and certainly such code would be specific to GNAT and
23140 RM References: B.01 (17) B.03 (62) B.03 (71.1/2)
23144 @emph{AI-0216 No_Task_Hierarchy forbids local tasks (0000-00-00)}
23145 @cindex AI05-0216 (Ada 2012 feature)
23148 It is clearly the intention that @code{No_Task_Hierarchy} is intended to
23149 forbid tasks declared locally within subprograms, or functions returning task
23150 objects, and that is the implementation that GNAT has always provided.
23151 However the language in the RM was not sufficiently clear on this point.
23152 Thus this is a documentation change in the RM only.
23155 RM References: D.07 (3/3)
23158 @emph{AI-0211 No_Relative_Delays forbids Set_Handler use (2010-07-09)}
23159 @cindex AI-0211 (Ada 2012 feature)
23162 The restriction @code{No_Relative_Delays} forbids any calls to the subprogram
23163 @code{Ada.Real_Time.Timing_Events.Set_Handler}.
23166 RM References: D.07 (5) D.07 (10/2) D.07 (10.4/2) D.07 (10.7/2)
23169 @emph{AI-0190 pragma Default_Storage_Pool (2010-09-15)}
23170 @cindex AI-0190 (Ada 2012 feature)
23173 This AI introduces a new pragma @code{Default_Storage_Pool}, which can be
23174 used to control storage pools globally.
23175 In particular, you can force every access
23176 type that is used for allocation (@b{new}) to have an explicit storage pool,
23177 or you can declare a pool globally to be used for all access types that lack
23181 RM References: D.07 (8)
23184 @emph{AI-0189 No_Allocators_After_Elaboration (2010-01-23)}
23185 @cindex AI-0189 (Ada 2012 feature)
23188 This AI introduces a new restriction @code{No_Allocators_After_Elaboration},
23189 which says that no dynamic allocation will occur once elaboration is
23191 In general this requires a run-time check, which is not required, and which
23192 GNAT does not attempt. But the static cases of allocators in a task body or
23193 in the body of the main program are detected and flagged at compile or bind
23197 RM References: D.07 (19.1/2) H.04 (23.3/2)
23200 @emph{AI-0171 Pragma CPU and Ravenscar Profile (2010-09-24)}
23201 @cindex AI-0171 (Ada 2012 feature)
23204 A new package @code{System.Multiprocessors} is added, together with the
23205 definition of pragma @code{CPU} for controlling task affinity. A new no
23206 dependence restriction, on @code{System.Multiprocessors.Dispatching_Domains},
23207 is added to the Ravenscar profile.
23210 RM References: D.13.01 (4/2) D.16
23214 @emph{AI-0210 Correct Timing_Events metric (0000-00-00)}
23215 @cindex AI-0210 (Ada 2012 feature)
23218 This is a documentation only issue regarding wording of metric requirements,
23219 that does not affect the implementation of the compiler.
23222 RM References: D.15 (24/2)
23226 @emph{AI-0206 Remote types packages and preelaborate (2010-07-24)}
23227 @cindex AI-0206 (Ada 2012 feature)
23230 Remote types packages are now allowed to depend on preelaborated packages.
23231 This was formerly considered illegal.
23234 RM References: E.02.02 (6)
23239 @emph{AI-0152 Restriction No_Anonymous_Allocators (2010-09-08)}
23240 @cindex AI-0152 (Ada 2012 feature)
23243 Restriction @code{No_Anonymous_Allocators} prevents the use of allocators
23244 where the type of the returned value is an anonymous access type.
23247 RM References: H.04 (8/1)
23251 @node Obsolescent Features
23252 @chapter Obsolescent Features
23255 This chapter describes features that are provided by GNAT, but are
23256 considered obsolescent since there are preferred ways of achieving
23257 the same effect. These features are provided solely for historical
23258 compatibility purposes.
23261 * pragma No_Run_Time::
23262 * pragma Ravenscar::
23263 * pragma Restricted_Run_Time::
23264 * pragma Task_Info::
23265 * System.Task_Info (s-tasinf.ads)::
23268 @node pragma No_Run_Time
23269 @section pragma No_Run_Time
23271 The pragma @code{No_Run_Time} is used to achieve an affect similar
23272 to the use of the "Zero Foot Print" configurable run time, but without
23273 requiring a specially configured run time. The result of using this
23274 pragma, which must be used for all units in a partition, is to restrict
23275 the use of any language features requiring run-time support code. The
23276 preferred usage is to use an appropriately configured run-time that
23277 includes just those features that are to be made accessible.
23279 @node pragma Ravenscar
23280 @section pragma Ravenscar
23282 The pragma @code{Ravenscar} has exactly the same effect as pragma
23283 @code{Profile (Ravenscar)}. The latter usage is preferred since it
23284 is part of the new Ada 2005 standard.
23286 @node pragma Restricted_Run_Time
23287 @section pragma Restricted_Run_Time
23289 The pragma @code{Restricted_Run_Time} has exactly the same effect as
23290 pragma @code{Profile (Restricted)}. The latter usage is
23291 preferred since the Ada 2005 pragma @code{Profile} is intended for
23292 this kind of implementation dependent addition.
23294 @node pragma Task_Info
23295 @section pragma Task_Info
23297 The functionality provided by pragma @code{Task_Info} is now part of the
23298 Ada language. The @code{CPU} aspect and the package
23299 @code{System.Multiprocessors} offer a less system-dependent way to specify
23300 task affinity or to query the number of processsors.
23305 @smallexample @c ada
23306 pragma Task_Info (EXPRESSION);
23310 This pragma appears within a task definition (like pragma
23311 @code{Priority}) and applies to the task in which it appears. The
23312 argument must be of type @code{System.Task_Info.Task_Info_Type}.
23313 The @code{Task_Info} pragma provides system dependent control over
23314 aspects of tasking implementation, for example, the ability to map
23315 tasks to specific processors. For details on the facilities available
23316 for the version of GNAT that you are using, see the documentation
23317 in the spec of package System.Task_Info in the runtime
23320 @node System.Task_Info (s-tasinf.ads)
23321 @section package System.Task_Info (@file{s-tasinf.ads})
23324 This package provides target dependent functionality that is used
23325 to support the @code{Task_Info} pragma. The predefined Ada package
23326 @code{System.Multiprocessors} and the @code{CPU} aspect now provide a
23327 standard replacement for GNAT's @code{Task_Info} functionality.
23330 @c GNU Free Documentation License
23332 @node Concept Index