1------------------------------------------------------------------------------ 2-- -- 3-- GNAT COMPILER COMPONENTS -- 4-- -- 5-- T Y P E S -- 6-- -- 7-- S p e c -- 8-- -- 9-- Copyright (C) 1992-2003 Free Software Foundation, Inc. -- 10-- -- 11-- GNAT is free software; you can redistribute it and/or modify it under -- 12-- terms of the GNU General Public License as published by the Free Soft- -- 13-- ware Foundation; either version 2, or (at your option) any later ver- -- 14-- sion. GNAT is distributed in the hope that it will be useful, but WITH- -- 15-- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY -- 16-- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License -- 17-- for more details. You should have received a copy of the GNU General -- 18-- Public License distributed with GNAT; see file COPYING. If not, write -- 19-- to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, -- 20-- MA 02111-1307, USA. -- 21-- -- 22-- As a special exception, if other files instantiate generics from this -- 23-- unit, or you link this unit with other files to produce an executable, -- 24-- this unit does not by itself cause the resulting executable to be -- 25-- covered by the GNU General Public License. This exception does not -- 26-- however invalidate any other reasons why the executable file might be -- 27-- covered by the GNU Public License. -- 28-- -- 29-- GNAT was originally developed by the GNAT team at New York University. -- 30-- Extensive contributions were provided by Ada Core Technologies Inc. -- 31-- -- 32------------------------------------------------------------------------------ 33 34with Unchecked_Deallocation; 35 36package Types is 37pragma Preelaborate (Types); 38 39-- This package contains host independent type definitions which are used 40-- in more than one unit in the compiler. They are gathered here for easy 41-- reference, though in some cases the full description is found in the 42-- relevant module which implements the definition. The main reason that 43-- they are not in their "natural" specs is that this would cause a lot of 44-- inter-spec dependencies, and in particular some awkward circular 45-- dependencies would have to be dealt with. 46 47-- WARNING: There is a C version of this package. Any changes to this 48-- source file must be properly reflected in the C header file a-types.h 49 50-- Note: the declarations in this package reflect an expectation that the 51-- host machine has an efficient integer base type with a range at least 52-- 32 bits 2s-complement. If there are any machines for which this is not 53-- a correct assumption, a significant number of changes will be required! 54 55 ------------------------------- 56 -- General Use Integer Types -- 57 ------------------------------- 58 59 type Int is range -2 ** 31 .. +2 ** 31 - 1; 60 -- Signed 32-bit integer 61 62 type Dint is range -2 ** 63 .. +2 ** 63 - 1; 63 -- Double length (64-bit) integer 64 65 subtype Nat is Int range 0 .. Int'Last; 66 -- Non-negative Int values 67 68 subtype Pos is Int range 1 .. Int'Last; 69 -- Positive Int values 70 71 type Word is mod 2 ** 32; 72 -- Unsigned 32-bit integer 73 74 type Short is range -32768 .. +32767; 75 for Short'Size use 16; 76 -- 16-bit signed integer 77 78 type Byte is mod 2 ** 8; 79 for Byte'Size use 8; 80 -- 8-bit unsigned integer 81 82 type size_t is mod 2 ** Standard'Address_Size; 83 -- Memory size value, for use in calls to C routines 84 85 -------------------------------------- 86 -- 8-Bit Character and String Types -- 87 -------------------------------------- 88 89 -- We use Standard.Character and Standard.String freely, since we are 90 -- compiling ourselves, and we properly implement the required 8-bit 91 -- character code as required in Ada 95. This section defines a few 92 -- general use constants and subtypes. 93 94 EOF : constant Character := ASCII.SUB; 95 -- The character SUB (16#1A#) is used in DOS and other systems derived 96 -- from DOS (OS/2, NT etc) to signal the end of a text file. Internally 97 -- all source files are ended by an EOF character, even on Unix systems. 98 -- An EOF character acts as the end of file only as the last character 99 -- of a source buffer, in any other position, it is treated as a blank 100 -- if it appears between tokens, and as an illegal character otherwise. 101 -- This makes life easier dealing with files that originated from DOS, 102 -- including concatenated files with interspersed EOF characters. 103 104 subtype Graphic_Character is Character range ' ' .. '~'; 105 -- Graphic characters, as defined in ARM 106 107 subtype Line_Terminator is Character range ASCII.LF .. ASCII.CR; 108 -- Line terminator characters (LF, VT, FF, CR) 109 110 subtype Upper_Half_Character is 111 Character range Character'Val (16#80#) .. Character'Val (16#FF#); 112 -- Characters with the upper bit set 113 114 type Character_Ptr is access all Character; 115 type String_Ptr is access all String; 116 -- Standard character and string pointers 117 118 procedure Free is new Unchecked_Deallocation (String, String_Ptr); 119 -- Procedure for freeing dynamically allocated String values 120 121 subtype Word_Hex_String is String (1 .. 8); 122 -- Type used to represent Word value as 8 hex digits, with lower case 123 -- letters for the alphabetic cases. 124 125 function Get_Hex_String (W : Word) return Word_Hex_String; 126 -- Convert word value to 8-character hex string 127 128 ----------------------------------------- 129 -- Types Used for Text Buffer Handling -- 130 ----------------------------------------- 131 132 -- We can't use type String for text buffers, since we must use the 133 -- standard 32-bit integer as an index value, since we count on all 134 -- index values being the same size. 135 136 type Text_Ptr is new Int; 137 -- Type used for subscripts in text buffer 138 139 type Text_Buffer is array (Text_Ptr range <>) of Character; 140 -- Text buffer used to hold source file or library information file 141 142 type Text_Buffer_Ptr is access all Text_Buffer; 143 -- Text buffers for input files are allocated dynamically and this type 144 -- is used to reference these text buffers. 145 146 procedure Free is new Unchecked_Deallocation (Text_Buffer, Text_Buffer_Ptr); 147 -- Procedure for freeing dynamically allocated text buffers 148 149 ------------------------------------------ 150 -- Types Used for Source Input Handling -- 151 ------------------------------------------ 152 153 type Logical_Line_Number is range 0 .. Int'Last; 154 for Logical_Line_Number'Size use 32; 155 -- Line number type, used for storing logical line numbers (i.e. line 156 -- numbers that include effects of any Source_Reference pragmas in the 157 -- source file). The value zero indicates a line containing a source 158 -- reference pragma. 159 160 No_Line_Number : constant Logical_Line_Number := 0; 161 -- Special value used to indicate no line number 162 163 type Physical_Line_Number is range 1 .. Int'Last; 164 for Physical_Line_Number'Size use 32; 165 -- Line number type, used for storing physical line numbers (i.e. 166 -- line numbers in the physical file being compiled, unaffected by 167 -- the presence of source reference pragmas. 168 169 type Column_Number is range 0 .. 32767; 170 for Column_Number'Size use 16; 171 -- Column number (assume that 2**15 is large enough, see declaration of 172 -- Hostparm.Max_Line_Length, and also processing for -gnatyM in Stylesw) 173 174 No_Column_Number : constant Column_Number := 0; 175 -- Special value used to indicate no column number 176 177 subtype Source_Buffer is Text_Buffer; 178 -- Type used to store text of a source file . The buffer for the main 179 -- source (the source specified on the command line) has a lower bound 180 -- starting at zero. Subsequent subsidiary sources have lower bounds 181 -- which are one greater than the previous upper bound. 182 183 subtype Big_Source_Buffer is Text_Buffer (0 .. Text_Ptr'Last); 184 -- This is a virtual type used as the designated type of the access 185 -- type Source_Buffer_Ptr, see Osint.Read_Source_File for details. 186 187 type Source_Buffer_Ptr is access all Big_Source_Buffer; 188 -- Pointer to source buffer. We use virtual origin addressing for 189 -- source buffers, with thin pointers. The pointer points to a virtual 190 -- instance of type Big_Source_Buffer, where the actual type is in fact 191 -- of type Source_Buffer. The address is adjusted so that the virtual 192 -- origin addressing works correctly. See Osint.Read_Source_Buffer for 193 -- further details. 194 195 subtype Source_Ptr is Text_Ptr; 196 -- Type used to represent a source location, which is a subscript of a 197 -- character in the source buffer. As noted above, diffferent source 198 -- buffers have different ranges, so it is possible to tell from a 199 -- Source_Ptr value which source it refers to. Note that negative numbers 200 -- are allowed to accommodate the following special values. 201 202 No_Location : constant Source_Ptr := -1; 203 -- Value used to indicate no source position set in a node. A test for 204 -- a Source_Ptr value being >= No_Location is the apporoved way to test 205 -- for a standard value that does not include No_Location or any of the 206 -- following special definitions. 207 208 Standard_Location : constant Source_Ptr := -2; 209 -- Used for all nodes in the representation of package Standard other 210 -- than nodes representing the contents of Standard.ASCII. Note that 211 -- testing for <= Standard_Location tests for both Standard_Location 212 -- and for Standard_ASCII_Location. 213 214 Standard_ASCII_Location : constant Source_Ptr := -3; 215 -- Used for all nodes in the presentation of package Standard.ASCII 216 217 System_Location : constant Source_Ptr := -4; 218 -- Used to identify locations of pragmas scanned by Targparm, where we 219 -- know the location is in System, but we don't know exactly what line. 220 221 First_Source_Ptr : constant Source_Ptr := 0; 222 -- Starting source pointer index value for first source program 223 224 ------------------------------------- 225 -- Range Definitions for Tree Data -- 226 ------------------------------------- 227 228 -- The tree has fields that can hold any of the following types: 229 230 -- Pointers to other tree nodes (type Node_Id) 231 -- List pointers (type List_Id) 232 -- Element list pointers (type Elist_Id) 233 -- Names (type Name_Id) 234 -- Strings (type String_Id) 235 -- Universal integers (type Uint) 236 -- Universal reals (type Ureal) 237 -- Character codes (type Char_Code stored with a bias) 238 239 -- In most contexts, the strongly typed interface determines which of 240 -- these types is present. However, there are some situations (involving 241 -- untyped traversals of the tree), where it is convenient to be easily 242 -- able to distinguish these values. The underlying representation in all 243 -- cases is an integer type Union_Id, and we ensure that the range of 244 -- the various possible values for each of the above types is disjoint 245 -- so that this distinction is possible. 246 247 type Union_Id is new Int; 248 -- The type in the tree for a union of possible ID values 249 250 -- Note: it is also helpful for debugging purposes to make these ranges 251 -- distinct. If a bug leads to misidentification of a value, then it will 252 -- typically result in an out of range value and a Constraint_Error. 253 254 List_Low_Bound : constant := -100_000_000; 255 -- The List_Id values are subscripts into an array of list headers which 256 -- has List_Low_Bound as its lower bound. This value is chosen so that all 257 -- List_Id values are negative, and the value zero is in the range of both 258 -- List_Id and Node_Id values (see further description below). 259 260 List_High_Bound : constant := 0; 261 -- Maximum List_Id subscript value. This allows up to 100 million list 262 -- Id values, which is in practice infinite, and there is no need to 263 -- check the range. The range overlaps the node range by one element 264 -- (with value zero), which is used both for the Empty node, and for 265 -- indicating no list. The fact that the same value is used is convenient 266 -- because it means that the default value of Empty applies to both nodes 267 -- and lists, and also is more efficient to test for. 268 269 Node_Low_Bound : constant := 0; 270 -- The tree Id values start at zero, because we use zero for Empty (to 271 -- allow a zero test for Empty). Actual tree node subscripts start at 0 272 -- since Empty is a legitimate node value. 273 274 Node_High_Bound : constant := 099_999_999; 275 -- Maximum number of nodes that can be allocated is 100 million, which 276 -- is in practice infinite, and there is no need to check the range. 277 278 Elist_Low_Bound : constant := 100_000_000; 279 -- The Elist_Id values are subscripts into an array of elist headers which 280 -- has Elist_Low_Bound as its lower bound. 281 282 Elist_High_Bound : constant := 199_999_999; 283 -- Maximum Elist_Id subscript value. This allows up to 100 million Elists, 284 -- which is in practice infinite and there is no need to check the range. 285 286 Elmt_Low_Bound : constant := 200_000_000; 287 -- Low bound of element Id values. The use of these values is internal to 288 -- the Elists package, but the definition of the range is included here 289 -- since it must be disjoint from other Id values. The Elmt_Id values are 290 -- subscripts into an array of list elements which has this as lower bound. 291 292 Elmt_High_Bound : constant := 299_999_999; 293 -- Upper bound of Elmt_Id values. This allows up to 100 million element 294 -- list members, which is in practice infinite (no range check needed). 295 296 Names_Low_Bound : constant := 300_000_000; 297 -- Low bound for name Id values 298 299 Names_High_Bound : constant := 399_999_999; 300 -- Maximum number of names that can be allocated is 100 million, which is 301 -- in practice infinite and there is no need to check the range. 302 303 Strings_Low_Bound : constant := 400_000_000; 304 -- Low bound for string Id values 305 306 Strings_High_Bound : constant := 499_999_999; 307 -- Maximum number of strings that can be allocated is 100 million, which 308 -- is in practice infinite and there is no need to check the range. 309 310 Ureal_Low_Bound : constant := 500_000_000; 311 -- Low bound for Ureal values. 312 313 Ureal_High_Bound : constant := 599_999_999; 314 -- Maximum number of Ureal values stored is 100_000_000 which is in 315 -- practice infinite so that no check is required. 316 317 Uint_Low_Bound : constant := 600_000_000; 318 -- Low bound for Uint values. 319 320 Uint_Table_Start : constant := 2_000_000_000; 321 -- Location where table entries for universal integers start (see 322 -- Uintp spec for details of the representation of Uint values). 323 324 Uint_High_Bound : constant := 2_099_999_999; 325 -- The range of Uint values is very large, since a substantial part 326 -- of this range is used to store direct values, see Uintp for details. 327 328 Char_Code_Bias : constant := 2_100_000_000; 329 -- A bias value added to character code values stored in the tree which 330 -- ensures that they have different values from any of the above types. 331 332 -- The following subtype definitions are used to provide convenient names 333 -- for membership tests on Int values to see what data type range they 334 -- lie in. Such tests appear only in the lowest level packages. 335 336 subtype List_Range is Union_Id 337 range List_Low_Bound .. List_High_Bound; 338 339 subtype Node_Range is Union_Id 340 range Node_Low_Bound .. Node_High_Bound; 341 342 subtype Elist_Range is Union_Id 343 range Elist_Low_Bound .. Elist_High_Bound; 344 345 subtype Elmt_Range is Union_Id 346 range Elmt_Low_Bound .. Elmt_High_Bound; 347 348 subtype Names_Range is Union_Id 349 range Names_Low_Bound .. Names_High_Bound; 350 351 subtype Strings_Range is Union_Id 352 range Strings_Low_Bound .. Strings_High_Bound; 353 354 subtype Uint_Range is Union_Id 355 range Uint_Low_Bound .. Uint_High_Bound; 356 357 subtype Ureal_Range is Union_Id 358 range Ureal_Low_Bound .. Ureal_High_Bound; 359 360 subtype Char_Code_Range is Union_Id 361 range Char_Code_Bias .. Char_Code_Bias + 2**16 - 1; 362 363 ----------------------------- 364 -- Types for Namet Package -- 365 ----------------------------- 366 367 -- Name_Id values are used to identify entries in the names table. Except 368 -- for the special values No_Name, and Error_Name, they are subscript 369 -- values for the Names table defined in package Namet. 370 371 -- Note that with only a few exceptions, which are clearly documented, the 372 -- type Name_Id should be regarded as a private type. In particular it is 373 -- never appropriate to perform arithmetic operations using this type. 374 375 type Name_Id is range Names_Low_Bound .. Names_High_Bound; 376 for Name_Id'Size use 32; 377 -- Type used to identify entries in the names table 378 379 No_Name : constant Name_Id := Names_Low_Bound; 380 -- The special Name_Id value No_Name is used in the parser to indicate 381 -- a situation where no name is present (e.g. on a loop or block). 382 383 Error_Name : constant Name_Id := Names_Low_Bound + 1; 384 -- The special Name_Id value Error_Name is used in the parser to 385 -- indicate that some kind of error was encountered in scanning out 386 -- the relevant name, so it does not have a representable label. 387 388 subtype Error_Name_Or_No_Name is Name_Id range No_Name .. Error_Name; 389 -- Used to test for either error name or no name 390 391 First_Name_Id : constant Name_Id := Names_Low_Bound + 2; 392 -- Subscript of first entry in names table 393 394 ---------------------------- 395 -- Types for Atree Package -- 396 ---------------------------- 397 398 -- Node_Id values are used to identify nodes in the tree. They are 399 -- subscripts into the Node table declared in package Tree. Note that 400 -- the special values Empty and Error are subscripts into this table, 401 -- See package Atree for further details. 402 403 type Node_Id is range Node_Low_Bound .. Node_High_Bound; 404 -- Type used to identify nodes in the tree 405 406 subtype Entity_Id is Node_Id; 407 -- A synonym for node types, used in the entity package to refer to 408 -- nodes that are entities (i.e. nodes with an Nkind of N_Defining_xxx) 409 -- All such nodes are extended nodes and these are the only extended 410 -- nodes, so that in practice entity and extended nodes are synonymous. 411 412 subtype Node_Or_Entity_Id is Node_Id; 413 -- A synonym for node types, used in cases where a given value may be used 414 -- to represent either a node or an entity. We like to minimize such uses 415 -- for obvious reasons of logical type consistency, but where such uses 416 -- occur, they should be documented by use of this type. 417 418 Empty : constant Node_Id := Node_Low_Bound; 419 -- Used to indicate null node. A node is actually allocated with this 420 -- Id value, so that Nkind (Empty) = N_Empty. Note that Node_Low_Bound 421 -- is zero, so Empty = No_List = zero. 422 423 Empty_List_Or_Node : constant := 0; 424 -- This constant is used in situations (e.g. initializing empty fields) 425 -- where the value set will be used to represent either an empty node 426 -- or a non-existent list, depending on the context. 427 428 Error : constant Node_Id := Node_Low_Bound + 1; 429 -- Used to indicate that there was an error in the source program. A node 430 -- is actually allocated at this address, so that Nkind (Error) = N_Error. 431 432 Empty_Or_Error : constant Node_Id := Error; 433 -- Since Empty and Error are the first two Node_Id values, the test for 434 -- N <= Empty_Or_Error tests to see if N is Empty or Error. This definition 435 -- provides convenient self-documentation for such tests. 436 437 First_Node_Id : constant Node_Id := Node_Low_Bound; 438 -- Subscript of first allocated node. Note that Empty and Error are both 439 -- allocated nodes, whose Nkind fields can be accessed without error. 440 441 ------------------------------ 442 -- Types for Nlists Package -- 443 ------------------------------ 444 445 -- List_Id values are used to identify node lists in the tree. They are 446 -- subscripts into the Lists table declared in package Tree. Note that 447 -- the special value Error_List is a subscript in this table, but the 448 -- value No_List is *not* a valid subscript, and any attempt to apply 449 -- list operations to No_List will cause a (detected) error. 450 451 type List_Id is range List_Low_Bound .. List_High_Bound; 452 -- Type used to identify a node list 453 454 No_List : constant List_Id := List_High_Bound; 455 -- Used to indicate absence of a list. Note that the value is zero, which 456 -- is the same as Empty, which is helpful in intializing nodes where a 457 -- value of zero can represent either an empty node or an empty list. 458 459 Error_List : constant List_Id := List_Low_Bound; 460 -- Used to indicate that there was an error in the source program in a 461 -- context which would normally require a list. This node appears to be 462 -- an empty list to the list operations (a null list is actually allocated 463 -- which has this Id value). 464 465 First_List_Id : constant List_Id := Error_List; 466 -- Subscript of first allocated list header 467 468 ------------------------------ 469 -- Types for Elists Package -- 470 ------------------------------ 471 472 -- Element list Id values are used to identify element lists stored in 473 -- the tree (see package Tree for further details). They are formed by 474 -- adding a bias (Element_List_Bias) to subscript values in the same 475 -- array that is used for node list headers. 476 477 type Elist_Id is range Elist_Low_Bound .. Elist_High_Bound; 478 -- Type used to identify an element list (Elist header table subscript) 479 480 No_Elist : constant Elist_Id := Elist_Low_Bound; 481 -- Used to indicate absense of an element list. Note that this is not 482 -- an actual Elist header, so element list operations on this value 483 -- are not valid. 484 485 First_Elist_Id : constant Elist_Id := No_Elist + 1; 486 -- Subscript of first allocated Elist header. 487 488 -- Element Id values are used to identify individual elements of an 489 -- element list (see package Elists for further details). 490 491 type Elmt_Id is range Elmt_Low_Bound .. Elmt_High_Bound; 492 -- Type used to identify an element list 493 494 No_Elmt : constant Elmt_Id := Elmt_Low_Bound; 495 -- Used to represent empty element 496 497 First_Elmt_Id : constant Elmt_Id := No_Elmt + 1; 498 -- Subscript of first allocated Elmt table entry 499 500 ------------------------------- 501 -- Types for Stringt Package -- 502 ------------------------------- 503 504 -- String_Id values are used to identify entries in the strings table. 505 -- They are subscripts into the strings table defined in package Strings. 506 507 -- Note that with only a few exceptions, which are clearly documented, the 508 -- type String_Id should be regarded as a private type. In particular it is 509 -- never appropriate to perform arithmetic operations using this type. 510 511 type String_Id is range Strings_Low_Bound .. Strings_High_Bound; 512 -- Type used to identify entries in the strings table 513 514 No_String : constant String_Id := Strings_Low_Bound; 515 -- Used to indicate missing string Id. Note that the value zero is used 516 -- to indicate a missing data value for all the Int types in this section. 517 518 First_String_Id : constant String_Id := No_String + 1; 519 -- First subscript allocated in string table 520 521 ------------------------- 522 -- Character Code Type -- 523 ------------------------- 524 525 -- The type Char is used for character data internally in the compiler, 526 -- but character codes in the source are represented by the Char_Code 527 -- type. Each character literal in the source is interpreted as being one 528 -- of the 2**16 possible Wide_Character codes, and a unique integer value 529 -- is assigned, corresponding to the POS value in the Wide_Character type. 530 -- String literals are similarly interpreted as a sequence of such codes. 531 532 -- Note: when character code values are stored in the tree, they are stored 533 -- by adding a bias value (Char_Code_Bias) that results in values that can 534 -- be distinguished from other types of values stored in the tree. 535 536 type Char_Code is mod 2 ** 16; 537 for Char_Code'Size use 16; 538 539 function Get_Char_Code (C : Character) return Char_Code; 540 pragma Inline (Get_Char_Code); 541 -- Function to obtain internal character code from source character. For 542 -- the moment, the internal character code is simply the Pos value of the 543 -- input source character, but we provide this interface for possible 544 -- later support of alternative character sets. 545 546 function In_Character_Range (C : Char_Code) return Boolean; 547 pragma Inline (In_Character_Range); 548 -- Determines if the given character code is in range of type Character, 549 -- and if so, returns True. If not, returns False. 550 551 function Get_Character (C : Char_Code) return Character; 552 pragma Inline (Get_Character); 553 -- For a character C that is in character range (see above function), this 554 -- function returns the corresponding Character value. It is an error to 555 -- call Get_Character if C is not in character range 556 557 --------------------------------------- 558 -- Types used for Library Management -- 559 --------------------------------------- 560 561 type Unit_Number_Type is new Int; 562 -- Unit number. The main source is unit 0, and subsidiary sources have 563 -- non-zero numbers starting with 1. Unit numbers are used to index the 564 -- file table in Lib. 565 566 Main_Unit : constant Unit_Number_Type := 0; 567 -- Unit number value for main unit 568 569 No_Unit : constant Unit_Number_Type := -1; 570 -- Special value used to signal no unit 571 572 type Source_File_Index is new Int range -1 .. Int'Last; 573 -- Type used to index the source file table (see package Sinput) 574 575 Internal_Source_File : constant Source_File_Index := 576 Source_File_Index'First; 577 -- Value used to indicate the buffer for the source-code-like strings 578 -- internally created withing the compiler (see package Sinput) 579 580 No_Source_File : constant Source_File_Index := 0; 581 -- Value used to indicate no source file present 582 583 subtype File_Name_Type is Name_Id; 584 -- File names are stored in the names table and this synonym is used to 585 -- indicate that a Name_Id value is being used to hold a simple file 586 -- name (which does not include any directory information). 587 588 No_File : constant File_Name_Type := File_Name_Type (No_Name); 589 -- Constant used to indicate no file found 590 591 subtype Unit_Name_Type is Name_Id; 592 -- Unit names are stored in the names table and this synonym is used to 593 -- indicate that a Name_Id value is being used to hold a unit name. 594 595 ----------------------------------- 596 -- Representation of Time Stamps -- 597 ----------------------------------- 598 599 -- All compiled units are marked with a time stamp which is derived from 600 -- the source file (we assume that the host system has the concept of a 601 -- file time stamp which is modified when a file is modified). These 602 -- time stamps are used to ensure consistency of the set of units that 603 -- constitutes a library. Time stamps are 12 character strings with 604 -- with the following format: 605 606 -- YYYYMMDDHHMMSS 607 608 -- YYYY year 609 -- MM month (2 digits 01-12) 610 -- DD day (2 digits 01-31) 611 -- HH hour (2 digits 00-23) 612 -- MM minutes (2 digits 00-59) 613 -- SS seconds (2 digits 00-59) 614 615 -- In the case of Unix systems (and other systems which keep the time in 616 -- GMT), the time stamp is the GMT time of the file, not the local time. 617 -- This solves problems in using libraries across networks with clients 618 -- spread across multiple time-zones. 619 620 Time_Stamp_Length : constant := 14; 621 -- Length of time stamp value 622 623 subtype Time_Stamp_Index is Natural range 1 .. Time_Stamp_Length; 624 type Time_Stamp_Type is new String (Time_Stamp_Index); 625 -- Type used to represent time stamp 626 627 Empty_Time_Stamp : constant Time_Stamp_Type := (others => ' '); 628 -- Type used to represent an empty or missing time stamp. Looks less 629 -- than any real time stamp if two time stamps are compared. Note that 630 -- although this is not a private type, clients should not rely on the 631 -- exact way in which this string is represented, and instead should 632 -- use the subprograms below. 633 634 Dummy_Time_Stamp : constant Time_Stamp_Type := (others => '0'); 635 -- This is used for dummy time stamp values used in the D lines for 636 -- non-existant files, and is intended to be an impossible value. 637 638 function "=" (Left, Right : Time_Stamp_Type) return Boolean; 639 function "<=" (Left, Right : Time_Stamp_Type) return Boolean; 640 function ">=" (Left, Right : Time_Stamp_Type) return Boolean; 641 function "<" (Left, Right : Time_Stamp_Type) return Boolean; 642 function ">" (Left, Right : Time_Stamp_Type) return Boolean; 643 -- Comparison functions on time stamps. Note that two time stamps 644 -- are defined as being equal if they have the same day/month/year 645 -- and the hour/minutes/seconds values are within 2 seconds of one 646 -- another. This deals with rounding effects in library file time 647 -- stamps caused by copying operations during installation. We have 648 -- particularly noticed that WinNT seems susceptible to such changes. 649 -- Note: the Empty_Time_Stamp value looks equal to itself, and less 650 -- than any non-empty time stamp value. 651 652 procedure Split_Time_Stamp 653 (TS : Time_Stamp_Type; 654 Year : out Nat; 655 Month : out Nat; 656 Day : out Nat; 657 Hour : out Nat; 658 Minutes : out Nat; 659 Seconds : out Nat); 660 -- Given a time stamp, decompose it into its components 661 662 procedure Make_Time_Stamp 663 (Year : Nat; 664 Month : Nat; 665 Day : Nat; 666 Hour : Nat; 667 Minutes : Nat; 668 Seconds : Nat; 669 TS : out Time_Stamp_Type); 670 -- Given the components of a time stamp, initialize the value 671 672 ----------------------------------------------- 673 -- Types used for Pragma Suppress Management -- 674 ----------------------------------------------- 675 676 type Check_Id is ( 677 Access_Check, 678 Accessibility_Check, 679 Discriminant_Check, 680 Division_Check, 681 Elaboration_Check, 682 Index_Check, 683 Length_Check, 684 Overflow_Check, 685 Range_Check, 686 Storage_Check, 687 Tag_Check, 688 All_Checks); 689 690 -- The following record contains an entry for each recognized check name 691 -- for pragma Suppress. It is used to represent current settings of scope 692 -- based suppress actions from pragma Suppress or command line settings. 693 694 type Suppress_Array is 695 array (Check_Id range Access_Check .. Tag_Check) of Boolean; 696 pragma Pack (Suppress_Array); 697 698 -- To add a new check type to GNAT, the following steps are required: 699 700 -- 1. Add an entry to Snames spec and body for the new name 701 -- 2. Add an entry to the definition of Check_Id above 702 -- 3. Add a new function to Checks to handle the new check test 703 -- 4. Add a new Do_xxx_Check flag to Sinfo (if required) 704 -- 5. Add appropriate checks for the new test 705 706 ----------------------------------- 707 -- Global Exception Declarations -- 708 ----------------------------------- 709 710 -- This section contains declarations of exceptions that are used 711 -- throughout the compiler. 712 713 Unrecoverable_Error : exception; 714 -- This exception is raised to immediately terminate the compilation 715 -- of the current source program. Used in situations where things are 716 -- bad enough that it doesn't seem worth continuing (e.g. max errors 717 -- reached, or a required file is not found). Also raised when the 718 -- compiler finds itself in trouble after an error (see Comperr). 719 720 --------------------------------- 721 -- Parameter Mechanism Control -- 722 --------------------------------- 723 724 -- Function and parameter entities have a field that records the 725 -- passing mechanism. See specification of Sem_Mech for full details. 726 -- The following subtype is used to represent values of this type: 727 728 subtype Mechanism_Type is Int range -10 .. Int'Last; 729 -- Type used to represent a mechanism value. This is a subtype rather 730 -- than a type to avoid some annoying processing problems with certain 731 -- routines in Einfo (processing them to create the corresponding C). 732 733 ------------------------------ 734 -- Run-Time Exception Codes -- 735 ------------------------------ 736 737 -- When the code generator generates a run-time exception, it provides 738 -- a reason code which is one of the following. This reason code is used 739 -- to select the appropriate run-time routine to be called, determining 740 -- both the exception to be raised, and the message text to be added. 741 742 -- The prefix CE/PE/SE indicates the exception to be raised 743 -- CE = Constraint_Error 744 -- PE = Program_Error 745 -- SE = Storage_Error 746 747 -- The remaining part of the name indicates the message text to be added, 748 -- where all letters are lower case, and underscores are converted to 749 -- spaces (for example CE_Invalid_Data adds the text "invalid data"). 750 751 -- To add a new code, you need to do the following: 752 753 -- 1. Modify the type and subtype declarations below appropriately, 754 -- keeping things in alphabetical order. 755 756 -- 2. Modify the corresponding definitions in types.h, including 757 -- the definition of last_reason_code. 758 759 -- 3. Add a new routine in Ada.Exceptions with the appropriate call 760 -- and static string constant 761 762 type RT_Exception_Code is ( 763 CE_Access_Check_Failed, 764 CE_Access_Parameter_Is_Null, 765 CE_Discriminant_Check_Failed, 766 CE_Divide_By_Zero, 767 CE_Explicit_Raise, 768 CE_Index_Check_Failed, 769 CE_Invalid_Data, 770 CE_Length_Check_Failed, 771 CE_Overflow_Check_Failed, 772 CE_Partition_Check_Failed, 773 CE_Range_Check_Failed, 774 CE_Tag_Check_Failed, 775 776 PE_Access_Before_Elaboration, 777 PE_Accessibility_Check_Failed, 778 PE_All_Guards_Closed, 779 PE_Duplicated_Entry_Address, 780 PE_Explicit_Raise, 781 PE_Finalize_Raised_Exception, 782 PE_Misaligned_Address_Value, 783 PE_Missing_Return, 784 PE_Overlaid_Controlled_Object, 785 PE_Potentially_Blocking_Operation, 786 PE_Stubbed_Subprogram_Called, 787 PE_Unchecked_Union_Restriction, 788 789 SE_Empty_Storage_Pool, 790 SE_Explicit_Raise, 791 SE_Infinite_Recursion, 792 SE_Object_Too_Large, 793 SE_Restriction_Violation); 794 795 subtype RT_CE_Exceptions is RT_Exception_Code range 796 CE_Access_Check_Failed .. 797 CE_Tag_Check_Failed; 798 799 subtype RT_PE_Exceptions is RT_Exception_Code range 800 PE_Access_Before_Elaboration .. 801 PE_Unchecked_Union_Restriction; 802 803 subtype RT_SE_Exceptions is RT_Exception_Code range 804 SE_Empty_Storage_Pool .. 805 SE_Restriction_Violation; 806 807end Types; 808