1------------------------------------------------------------------------------ 2-- -- 3-- GNAT COMPILER COMPONENTS -- 4-- -- 5-- S C A N S -- 6-- -- 7-- S p e c -- 8-- -- 9-- Copyright (C) 1992-2012, 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 3, 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. -- 17-- -- 18-- As a special exception under Section 7 of GPL version 3, you are granted -- 19-- additional permissions described in the GCC Runtime Library Exception, -- 20-- version 3.1, as published by the Free Software Foundation. -- 21-- -- 22-- You should have received a copy of the GNU General Public License and -- 23-- a copy of the GCC Runtime Library Exception along with this program; -- 24-- see the files COPYING3 and COPYING.RUNTIME respectively. If not, see -- 25-- <http://www.gnu.org/licenses/>. -- 26-- -- 27-- GNAT was originally developed by the GNAT team at New York University. -- 28-- Extensive contributions were provided by Ada Core Technologies Inc. -- 29-- -- 30------------------------------------------------------------------------------ 31 32with Namet; use Namet; 33with Types; use Types; 34with Uintp; use Uintp; 35with Urealp; use Urealp; 36 37package Scans is 38 39-- The scanner maintains a current state in the global variables defined 40-- in this package. The call to the Scan routine advances this state to 41-- the next token. The state is initialized by the call to one of the 42-- initialization routines in Sinput. 43 44 -- The following type is used to identify token types returned by Scan. 45 -- The class column in this table indicates the token classes which 46 -- apply to the token, as defined by subsequent subtype declarations. 47 48 -- Note: Namet.Is_Keyword_Name depends on the fact that the first entry in 49 -- this type declaration is *not* for a reserved word. For details on why 50 -- there is this requirement, see Initialize_Ada_Keywords below. 51 52 type Token_Type is ( 53 54 -- Token name Token type Class(es) 55 56 Tok_Integer_Literal, -- numeric lit Literal, Lit_Or_Name 57 58 Tok_Real_Literal, -- numeric lit Literal, Lit_Or_Name 59 60 Tok_String_Literal, -- string lit Literal. Lit_Or_Name 61 62 Tok_Char_Literal, -- char lit Name, Literal. Lit_Or_Name 63 64 Tok_Operator_Symbol, -- op symbol Name, Literal, Lit_Or_Name, Desig 65 66 Tok_Identifier, -- identifier Name, Lit_Or_Name, Desig 67 68 Tok_Double_Asterisk, -- ** 69 70 Tok_Ampersand, -- & Binary_Addop 71 Tok_Minus, -- - Binary_Addop, Unary_Addop 72 Tok_Plus, -- + Binary_Addop, Unary_Addop 73 74 Tok_Asterisk, -- * Mulop 75 Tok_Mod, -- MOD Mulop 76 Tok_Rem, -- REM Mulop 77 Tok_Slash, -- / Mulop 78 79 Tok_New, -- NEW 80 81 Tok_Abs, -- ABS 82 Tok_Others, -- OTHERS 83 Tok_Null, -- NULL 84 85 Tok_Dot, -- . Namext 86 Tok_Apostrophe, -- ' Namext 87 88 Tok_Left_Paren, -- ( Namext, Consk 89 90 Tok_Delta, -- DELTA Atkwd, Sterm, Consk 91 Tok_Digits, -- DIGITS Atkwd, Sterm, Consk 92 Tok_Range, -- RANGE Atkwd, Sterm, Consk 93 94 Tok_Right_Paren, -- ) Sterm 95 Tok_Comma, -- , Sterm 96 97 Tok_And, -- AND Logop, Sterm 98 Tok_Or, -- OR Logop, Sterm 99 Tok_Xor, -- XOR Logop, Sterm 100 101 Tok_Less, -- < Relop, Sterm 102 Tok_Equal, -- = Relop, Sterm 103 Tok_Greater, -- > Relop, Sterm 104 Tok_Not_Equal, -- /= Relop, Sterm 105 Tok_Greater_Equal, -- >= Relop, Sterm 106 Tok_Less_Equal, -- <= Relop, Sterm 107 108 Tok_In, -- IN Relop, Sterm 109 Tok_Not, -- NOT Relop, Sterm 110 111 Tok_Box, -- <> Relop, Eterm, Sterm 112 Tok_Colon_Equal, -- := Eterm, Sterm 113 Tok_Colon, -- : Eterm, Sterm 114 Tok_Greater_Greater, -- >> Eterm, Sterm 115 116 Tok_Abstract, -- ABSTRACT Eterm, Sterm 117 Tok_Access, -- ACCESS Eterm, Sterm 118 Tok_Aliased, -- ALIASED Eterm, Sterm 119 Tok_All, -- ALL Eterm, Sterm 120 Tok_Array, -- ARRAY Eterm, Sterm 121 Tok_At, -- AT Eterm, Sterm 122 Tok_Body, -- BODY Eterm, Sterm 123 Tok_Constant, -- CONSTANT Eterm, Sterm 124 Tok_Do, -- DO Eterm, Sterm 125 Tok_Is, -- IS Eterm, Sterm 126 Tok_Interface, -- INTERFACE Eterm, Sterm 127 Tok_Limited, -- LIMITED Eterm, Sterm 128 Tok_Of, -- OF Eterm, Sterm 129 Tok_Out, -- OUT Eterm, Sterm 130 Tok_Record, -- RECORD Eterm, Sterm 131 Tok_Renames, -- RENAMES Eterm, Sterm 132 Tok_Reverse, -- REVERSE Eterm, Sterm 133 Tok_Some, -- SOME Eterm, Sterm 134 Tok_Tagged, -- TAGGED Eterm, Sterm 135 Tok_Then, -- THEN Eterm, Sterm 136 137 Tok_Less_Less, -- << Eterm, Sterm, After_SM 138 139 Tok_Abort, -- ABORT Eterm, Sterm, After_SM 140 Tok_Accept, -- ACCEPT Eterm, Sterm, After_SM 141 Tok_Case, -- CASE Eterm, Sterm, After_SM 142 Tok_Delay, -- DELAY Eterm, Sterm, After_SM 143 Tok_Else, -- ELSE Eterm, Sterm, After_SM 144 Tok_Elsif, -- ELSIF Eterm, Sterm, After_SM 145 Tok_End, -- END Eterm, Sterm, After_SM 146 Tok_Exception, -- EXCEPTION Eterm, Sterm, After_SM 147 Tok_Exit, -- EXIT Eterm, Sterm, After_SM 148 Tok_Goto, -- GOTO Eterm, Sterm, After_SM 149 Tok_If, -- IF Eterm, Sterm, After_SM 150 Tok_Pragma, -- PRAGMA Eterm, Sterm, After_SM 151 Tok_Raise, -- RAISE Eterm, Sterm, After_SM 152 Tok_Requeue, -- REQUEUE Eterm, Sterm, After_SM 153 Tok_Return, -- RETURN Eterm, Sterm, After_SM 154 Tok_Select, -- SELECT Eterm, Sterm, After_SM 155 Tok_Terminate, -- TERMINATE Eterm, Sterm, After_SM 156 Tok_Until, -- UNTIL Eterm, Sterm, After_SM 157 Tok_When, -- WHEN Eterm, Sterm, After_SM 158 159 Tok_Begin, -- BEGIN Eterm, Sterm, After_SM, Labeled_Stmt 160 Tok_Declare, -- DECLARE Eterm, Sterm, After_SM, Labeled_Stmt 161 Tok_For, -- FOR Eterm, Sterm, After_SM, Labeled_Stmt 162 Tok_Loop, -- LOOP Eterm, Sterm, After_SM, Labeled_Stmt 163 Tok_While, -- WHILE Eterm, Sterm, After_SM, Labeled_Stmt 164 165 Tok_Entry, -- ENTRY Eterm, Sterm, Declk, Deckn, After_SM 166 Tok_Protected, -- PROTECTED Eterm, Sterm, Declk, Deckn, After_SM 167 Tok_Task, -- TASK Eterm, Sterm, Declk, Deckn, After_SM 168 Tok_Type, -- TYPE Eterm, Sterm, Declk, Deckn, After_SM 169 Tok_Subtype, -- SUBTYPE Eterm, Sterm, Declk, Deckn, After_SM 170 Tok_Overriding, -- OVERRIDING Eterm, Sterm, Declk, Declk, After_SM 171 Tok_Synchronized, -- SYNCHRONIZED Eterm, Sterm, Declk, Deckn, After_SM 172 Tok_Use, -- USE Eterm, Sterm, Declk, Deckn, After_SM 173 174 Tok_Function, -- FUNCTION Eterm, Sterm, Cunit, Declk, After_SM 175 Tok_Generic, -- GENERIC Eterm, Sterm, Cunit, Declk, After_SM 176 Tok_Package, -- PACKAGE Eterm, Sterm, Cunit, Declk, After_SM 177 Tok_Procedure, -- PROCEDURE Eterm, Sterm, Cunit, Declk, After_SM 178 179 Tok_Private, -- PRIVATE Eterm, Sterm, Cunit, After_SM 180 Tok_With, -- WITH Eterm, Sterm, Cunit, After_SM 181 Tok_Separate, -- SEPARATE Eterm, Sterm, Cunit, After_SM 182 183 Tok_EOF, -- End of file Eterm, Sterm, Cterm, After_SM 184 185 Tok_Semicolon, -- ; Eterm, Sterm, Cterm 186 187 Tok_Arrow, -- => Sterm, Cterm, Chtok 188 189 Tok_Vertical_Bar, -- | Cterm, Sterm, Chtok 190 191 Tok_Dot_Dot, -- .. Sterm, Chtok 192 193 Tok_Project, 194 Tok_Extends, 195 Tok_External, 196 Tok_External_As_List, 197 -- These four entries represent keywords for the project file language 198 -- and can be returned only in the case of scanning project files. 199 200 Tok_Comment, 201 -- This entry is used when scanning project files (where it represents 202 -- an entire comment), and in preprocessing with the -C switch set 203 -- (where it represents just the "--" of a comment). For the project 204 -- file case, the text of the comment is stored in Comment_Id. 205 206 Tok_End_Of_Line, 207 -- Represents an end of line. Not used during normal compilation scans 208 -- where end of line is ignored. Active for preprocessor scanning and 209 -- also when scanning project files (where it is needed because of ???) 210 211 Tok_Special, 212 -- Used only in preprocessor scanning (to represent one of the 213 -- characters '#', '$', '?', '@', '`', '\', '^', '~', or '_'. The 214 -- character value itself is stored in Scans.Special_Character. 215 216 Tok_SPARK_Hide, 217 -- HIDE directive in SPARK 218 219 No_Token); 220 -- No_Token is used for initializing Token values to indicate that 221 -- no value has been set yet. 222 223 -- Note: in the RM, operator symbol is a special case of string literal. 224 -- We distinguish at the lexical level in this compiler, since there are 225 -- many syntactic situations in which only an operator symbol is allowed. 226 227 -- The following subtype declarations group the token types into classes. 228 -- These are used for class tests in the parser. 229 230 subtype Token_Class_Numeric_Literal is 231 Token_Type range Tok_Integer_Literal .. Tok_Real_Literal; 232 -- Numeric literal 233 234 subtype Token_Class_Literal is 235 Token_Type range Tok_Integer_Literal .. Tok_Operator_Symbol; 236 -- Literal 237 238 subtype Token_Class_Lit_Or_Name is 239 Token_Type range Tok_Integer_Literal .. Tok_Identifier; 240 241 subtype Token_Class_Binary_Addop is 242 Token_Type range Tok_Ampersand .. Tok_Plus; 243 -- Binary adding operator (& + -) 244 245 subtype Token_Class_Unary_Addop is 246 Token_Type range Tok_Minus .. Tok_Plus; 247 -- Unary adding operator (+ -) 248 249 subtype Token_Class_Mulop is 250 Token_Type range Tok_Asterisk .. Tok_Slash; 251 -- Multiplying operator 252 253 subtype Token_Class_Logop is 254 Token_Type range Tok_And .. Tok_Xor; 255 -- Logical operator (and, or, xor) 256 257 subtype Token_Class_Relop is 258 Token_Type range Tok_Less .. Tok_Box; 259 -- Relational operator (= /= < <= > >= not, in plus <> to catch misuse 260 -- of Pascal style not equal operator). 261 262 subtype Token_Class_Name is 263 Token_Type range Tok_Char_Literal .. Tok_Identifier; 264 -- First token of name (4.1), 265 -- (identifier, char literal, operator symbol) 266 267 subtype Token_Class_Desig is 268 Token_Type range Tok_Operator_Symbol .. Tok_Identifier; 269 -- Token which can be a Designator (identifier, operator symbol) 270 271 subtype Token_Class_Namext is 272 Token_Type range Tok_Dot .. Tok_Left_Paren; 273 -- Name extension tokens. These are tokens which can appear immediately 274 -- after a name to extend it recursively (period, quote, left paren) 275 276 subtype Token_Class_Consk is 277 Token_Type range Tok_Left_Paren .. Tok_Range; 278 -- Keywords which can start constraint 279 -- (left paren, delta, digits, range) 280 281 subtype Token_Class_Eterm is 282 Token_Type range Tok_Colon_Equal .. Tok_Semicolon; 283 -- Expression terminators. These tokens can never appear within a simple 284 -- expression. This is used for error recovery purposes (if we encounter 285 -- an error in an expression, we simply scan to the next Eterm token). 286 287 subtype Token_Class_Sterm is 288 Token_Type range Tok_Delta .. Tok_Dot_Dot; 289 -- Simple_Expression terminators. A Simple_Expression must be followed 290 -- by a token in this class, or an error message is issued complaining 291 -- about a missing binary operator. 292 293 subtype Token_Class_Atkwd is 294 Token_Type range Tok_Delta .. Tok_Range; 295 -- Attribute keywords. This class includes keywords which can be used 296 -- as an Attribute_Designator, namely DELTA, DIGITS and RANGE 297 298 subtype Token_Class_Cterm is 299 Token_Type range Tok_EOF .. Tok_Vertical_Bar; 300 -- Choice terminators. These tokens terminate a choice. This is used for 301 -- error recovery purposes (if we encounter an error in a Choice, we 302 -- simply scan to the next Cterm token). 303 304 subtype Token_Class_Chtok is 305 Token_Type range Tok_Arrow .. Tok_Dot_Dot; 306 -- Choice tokens. These tokens signal a choice when used in an Aggregate 307 308 subtype Token_Class_Cunit is 309 Token_Type range Tok_Function .. Tok_Separate; 310 -- Tokens which can begin a compilation unit 311 312 subtype Token_Class_Declk is 313 Token_Type range Tok_Entry .. Tok_Procedure; 314 -- Keywords which start a declaration 315 316 subtype Token_Class_Deckn is 317 Token_Type range Tok_Entry .. Tok_Use; 318 -- Keywords which start a declaration but can't start a compilation unit 319 320 subtype Token_Class_After_SM is 321 Token_Type range Tok_Less_Less .. Tok_EOF; 322 -- Tokens which always, or almost always, appear after a semicolon. Used 323 -- in the Resync_Past_Semicolon routine to avoid gobbling up stuff when 324 -- a semicolon is missing. Of significance only for error recovery. 325 326 subtype Token_Class_Labeled_Stmt is 327 Token_Type range Tok_Begin .. Tok_While; 328 -- Tokens which start labeled statements 329 330 type Token_Flag_Array is array (Token_Type) of Boolean; 331 Is_Reserved_Keyword : constant Token_Flag_Array := 332 Token_Flag_Array' 333 (Tok_Mod .. Tok_Rem => True, 334 Tok_New .. Tok_Null => True, 335 Tok_Delta .. Tok_Range => True, 336 Tok_And .. Tok_Xor => True, 337 Tok_In .. Tok_Not => True, 338 Tok_Abstract .. Tok_Then => True, 339 Tok_Abort .. Tok_Separate => True, 340 others => False); 341 -- Flag array used to test for reserved word 342 343 procedure Initialize_Ada_Keywords; 344 -- Set up Token_Type values in Names table entries for Ada reserved 345 -- words. This ignores Ada_Version; Ada_Version is taken into account in 346 -- Snames.Is_Keyword_Name. 347 348 -------------------------- 349 -- Scan State Variables -- 350 -------------------------- 351 352 -- Note: these variables can only be referenced during the parsing of a 353 -- file. Reference to any of them from Sem or the expander is wrong. 354 355 -- These variables are initialized as required by Scn.Initialize_Scanner, 356 -- and should not be referenced before such a call. However, there are 357 -- situations in which these variables are saved and restored, and this 358 -- may happen before the first Initialize_Scanner call, resulting in the 359 -- assignment of invalid values. To avoid this, and allow building with 360 -- the -gnatVa switch, we initialize some variables to known valid values. 361 362 Scan_Ptr : Source_Ptr := No_Location; -- init for -gnatVa 363 -- Current scan pointer location. After a call to Scan, this points 364 -- just past the end of the token just scanned. 365 366 Token : Token_Type := No_Token; -- init for -gnatVa 367 -- Type of current token 368 369 Token_Ptr : Source_Ptr := No_Location; -- init for -gnatVa 370 -- Pointer to first character of current token 371 372 Current_Line_Start : Source_Ptr := No_Location; -- init for -gnatVa 373 -- Pointer to first character of line containing current token 374 375 Start_Column : Column_Number := No_Column_Number; -- init for -gnatVa 376 -- Starting column number (zero origin) of the first non-blank character 377 -- on the line containing the current token. This is used for error 378 -- recovery circuits which depend on looking at the column line up. 379 380 Type_Token_Location : Source_Ptr := No_Location; -- init for -gnatVa 381 -- Within a type declaration, gives the location of the TYPE keyword that 382 -- opened the type declaration. Used in checking the end column of a record 383 -- declaration, which can line up either with the TYPE keyword, or with the 384 -- start of the line containing the RECORD keyword. 385 386 Checksum : Word := 0; -- init for -gnatVa 387 -- Used to accumulate a CRC representing the tokens in the source 388 -- file being compiled. This CRC includes only program tokens, and 389 -- excludes comments. 390 391 First_Non_Blank_Location : Source_Ptr := No_Location; -- init for -gnatVa 392 -- Location of first non-blank character on the line containing the 393 -- current token (i.e. the location of the character whose column number 394 -- is stored in Start_Column). 395 396 Token_Node : Node_Id := Empty; 397 -- Node table Id for the current token. This is set only if the current 398 -- token is one for which the scanner constructs a node (i.e. it is an 399 -- identifier, operator symbol, or literal. For other token types, 400 -- Token_Node is undefined. 401 402 Token_Name : Name_Id := No_Name; 403 -- For identifiers, this is set to the Name_Id of the identifier scanned. 404 -- For all other tokens, Token_Name is set to Error_Name. Note that it 405 -- would be possible for the caller to extract this information from 406 -- Token_Node. We set Token_Name separately for two reasons. First it 407 -- allows a quicker test for a specific identifier. Second, it allows 408 -- a version of the parser to be built that does not build tree nodes, 409 -- usable as a syntax checker. 410 411 Prev_Token : Token_Type := No_Token; 412 -- Type of previous token 413 414 Prev_Token_Ptr : Source_Ptr; 415 -- Pointer to first character of previous token 416 417 Version_To_Be_Found : Boolean; 418 -- This flag is True if the scanner is still looking for an RCS version 419 -- number in a comment. Normally it is initialized to False so that this 420 -- circuit is not activated. If the -dv switch is set, then this flag is 421 -- initialized to True, and then reset when the version number is found. 422 -- We do things this way to minimize the impact on comment scanning. 423 424 Character_Code : Char_Code; 425 -- Valid only when Token is Tok_Char_Literal. Contains the value of the 426 -- scanned literal. 427 428 Real_Literal_Value : Ureal; 429 -- Valid only when Token is Tok_Real_Literal, contains the value of the 430 -- scanned literal. 431 432 Int_Literal_Value : Uint; 433 -- Valid only when Token = Tok_Integer_Literal, contains the value of the 434 -- scanned literal. 435 436 Based_Literal_Uses_Colon : Boolean; 437 -- Valid only when Token = Tok_Integer_Literal or Tok_Real_Literal. Set 438 -- True only for the case of a based literal using ':' instead of '#'. 439 440 String_Literal_Id : String_Id; 441 -- Valid only when Token = Tok_String_Literal or Tok_Operator_Symbol. 442 -- Contains the Id for currently scanned string value. 443 444 Wide_Character_Found : Boolean := False; 445 -- Valid only when Token = Tok_String_Literal. Set True if wide character 446 -- found (i.e. a character that does not fit in Character, but fits in 447 -- Wide_Wide_Character). 448 449 Wide_Wide_Character_Found : Boolean := False; 450 -- Valid only when Token = Tok_String_Literal. Set True if wide wide 451 -- character found (i.e. a character that does not fit in Character or 452 -- Wide_Character). 453 454 Special_Character : Character; 455 -- Valid only when Token = Tok_Special. Returns one of the characters 456 -- '#', '$', '?', '@', '`', '\', '^', '~', or '_'. 457 -- 458 -- Why only this set? What about wide characters??? 459 460 Comment_Id : Name_Id := No_Name; 461 -- Valid only when Token = Tok_Comment. Store the string that follows 462 -- the "--" of a comment when scanning project files. 463 -- 464 -- Is it really right for this to be a Name rather than a String, what 465 -- about the case of Wide_Wide_Characters??? 466 467 Inside_If_Expression : Nat := 0; 468 -- This is a counter that is set non-zero while scanning out an if 469 -- expression (incremented on entry, decremented on exit). It is used to 470 -- disconnect format checks that normally apply to keywords THEN, ELSE etc. 471 472 -------------------------------------------------------- 473 -- Procedures for Saving and Restoring the Scan State -- 474 -------------------------------------------------------- 475 476 -- The following procedures can be used to save and restore the entire 477 -- scan state. They are used in cases where it is necessary to backup 478 -- the scan during the parse. 479 480 type Saved_Scan_State is private; 481 -- Used for saving and restoring the scan state 482 483 procedure Save_Scan_State (Saved_State : out Saved_Scan_State); 484 pragma Inline (Save_Scan_State); 485 -- Saves the current scan state for possible later restoration. Note that 486 -- there is no harm in saving the state and then never restoring it. 487 488 procedure Restore_Scan_State (Saved_State : Saved_Scan_State); 489 pragma Inline (Restore_Scan_State); 490 -- Restores a scan state saved by a call to Save_Scan_State. 491 -- The saved scan state must refer to the current source file. 492 493private 494 type Saved_Scan_State is record 495 Save_Scan_Ptr : Source_Ptr; 496 Save_Token : Token_Type; 497 Save_Token_Ptr : Source_Ptr; 498 Save_Current_Line_Start : Source_Ptr; 499 Save_Start_Column : Column_Number; 500 Save_Checksum : Word; 501 Save_First_Non_Blank_Location : Source_Ptr; 502 Save_Token_Node : Node_Id; 503 Save_Token_Name : Name_Id; 504 Save_Prev_Token : Token_Type; 505 Save_Prev_Token_Ptr : Source_Ptr; 506 end record; 507 508end Scans; 509