src/symtab.c

   1 /* Symbol table manager for Bison.
   2
   3    Copyright (C) 1984, 1989, 2000, 2001, 2002, 2004, 2005 Free Software
   4    Foundation, Inc.
   5
   6    This file is part of Bison, the GNU Compiler Compiler.
   7
   8    Bison is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 2, or (at your option)
  11    any later version.
  12
  13    Bison is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with Bison; see the file COPYING.  If not, write to
  20    the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  21    Boston, MA 02110-1301, USA.  */
  22
  23
  24 #include "system.h"
  25
  26 #include <hash.h>
  27 #include <quotearg.h>
  28
  29 #include "complain.h"
  30 #include "gram.h"
  31 #include "symtab.h"
  32
  33 /*------------------------.
  34 | Distinguished symbols.  |
  35 `------------------------*/
  36
  37 symbol *errtoken = NULL;
  38 symbol *undeftoken = NULL;
  39 symbol *endtoken = NULL;
  40 symbol *accept = NULL;
  41 symbol *startsymbol = NULL;
  42 location startsymbol_location;
  43
  44 /*---------------------------------.
  45 | Create a new symbol, named TAG.  |
  46 `---------------------------------*/
  47
  48 static symbol *
  49 symbol_new (uniqstr tag, location loc)
  50 {
  51   symbol *res = xmalloc (sizeof *res);
  52
  53   uniqstr_assert (tag);
  54   res->tag = tag;
  55   res->location = loc;
  56
  57   res->type_name = NULL;
  58   res->destructor = NULL;
  59   res->printer = NULL;
  60
  61   res->number = NUMBER_UNDEFINED;
  62   res->prec = 0;
  63   res->assoc = undef_assoc;
  64   res->user_token_number = USER_NUMBER_UNDEFINED;
  65
  66   res->alias = NULL;
  67   res->class = unknown_sym;
  68
  69   if (nsyms == SYMBOL_NUMBER_MAXIMUM)
  70     fatal (_("too many symbols in input grammar (limit is %d)"),
  71            SYMBOL_NUMBER_MAXIMUM);
  72   nsyms++;
  73   return res;
  74 }
  75
  76
  77 /*------------------------------------------------------------------.
  78 | Complain that S's WHAT is redeclared at SECOND, and was first set |
  79 | at FIRST.                                                         |
  80 `------------------------------------------------------------------*/
  81
  82 static void
  83 redeclaration (symbol* s, const char *what, location first, location second)
  84 {
  85   complain_at (second, _("%s redeclaration for %s"), what, s->tag);
  86   complain_at (first, _("first declaration"));
  87 }
  88
  89
  90 /*-----------------------------------------------------------------.
  91 | Set the TYPE_NAME associated with SYM.  Does nothing if passed 0 |
  92 | as TYPE_NAME.                                                    |
  93 `-----------------------------------------------------------------*/
  94
  95 void
  96 symbol_type_set (symbol *sym, uniqstr type_name, location loc)
  97 {
  98   if (type_name)
  99     {
 100       if (sym->type_name)
 101         redeclaration (sym, "%type", sym->type_location, loc);
 102       uniqstr_assert (type_name);
 103       sym->type_name = type_name;
 104       sym->type_location = loc;
 105     }
 106 }
 107
 108
 109 /*------------------------------------------------------------------.
 110 | Set the DESTRUCTOR associated with SYM.  Do nothing if passed 0.  |
 111 `------------------------------------------------------------------*/
 112
 113 void
 114 symbol_destructor_set (symbol *sym, char *destructor, location loc)
 115 {
 116   if (destructor)
 117     {
 118       if (sym->destructor)
 119         redeclaration (sym, "%destructor", sym->destructor_location, loc);
 120       sym->destructor = destructor;
 121       sym->destructor_location = loc;
 122     }
 123 }
 124
 125
 126 /*---------------------------------------------------------------.
 127 | Set the PRINTER associated with SYM.  Do nothing if passed 0.  |
 128 `---------------------------------------------------------------*/
 129
 130 void
 131 symbol_printer_set (symbol *sym, char *printer, location loc)
 132 {
 133   if (printer)
 134     {
 135       if (sym->printer)
 136         redeclaration (sym, "%printer", sym->destructor_location, loc);
 137       sym->printer = printer;
 138       sym->printer_location = loc;
 139     }
 140 }
 141
 142
 143 /*-----------------------------------------------------------------.
 144 | Set the PRECEDENCE associated with SYM.  Does nothing if invoked |
 145 | with UNDEF_ASSOC as ASSOC.                                       |
 146 `-----------------------------------------------------------------*/
 147
 148 void
 149 symbol_precedence_set (symbol *sym, int prec, assoc a, location loc)
 150 {
 151   if (a != undef_assoc)
 152     {
 153       if (sym->prec != 0)
 154         redeclaration (sym, assoc_to_string (a), sym->prec_location, loc);
 155       sym->prec = prec;
 156       sym->assoc = a;
 157       sym->prec_location = loc;
 158     }
 159
 160   /* Only terminals have a precedence. */
 161   symbol_class_set (sym, token_sym, loc);
 162 }
 163
 164
 165 /*------------------------------------.
 166 | Set the CLASS associated with SYM.  |
 167 `------------------------------------*/
 168
 169 void
 170 symbol_class_set (symbol *sym, symbol_class class, location loc)
 171 {
 172   if (sym->class != unknown_sym && sym->class != class)
 173     complain_at (loc, _("symbol %s redefined"), sym->tag);
 174
 175   if (class == nterm_sym && sym->class != nterm_sym)
 176     sym->number = nvars++;
 177   else if (class == token_sym && sym->number == NUMBER_UNDEFINED)
 178     sym->number = ntokens++;
 179
 180   sym->class = class;
 181 }
 182
 183
 184 /*------------------------------------------------.
 185 | Set the USER_TOKEN_NUMBER associated with SYM.  |
 186 `------------------------------------------------*/
 187
 188 void
 189 symbol_user_token_number_set (symbol *sym, int user_token_number, location loc)
 190 {
 191   if (sym->class != token_sym)
 192     abort ();
 193
 194   if (sym->user_token_number != USER_NUMBER_UNDEFINED
 195       && sym->user_token_number != user_token_number)
 196     complain_at (loc, _("redefining user token number of %s"), sym->tag);
 197
 198   sym->user_token_number = user_token_number;
 199   /* User defined $end token? */
 200   if (user_token_number == 0)
 201     {
 202       endtoken = sym;
 203       endtoken->number = 0;
 204       /* It is always mapped to 0, so it was already counted in
 205          NTOKENS.  */
 206       --ntokens;
 207     }
 208 }
 209
 210
 211 /*----------------------------------------------------------.
 212 | If SYM is not defined, report an error, and consider it a |
 213 | nonterminal.                                              |
 214 `----------------------------------------------------------*/
 215
 216 static inline bool
 217 symbol_check_defined (symbol *sym)
 218 {
 219   if (sym->class == unknown_sym)
 220     {
 221       complain_at
 222         (sym->location,
 223          _("symbol %s is used, but is not defined as a token and has no rules"),
 224          sym->tag);
 225       sym->class = nterm_sym;
 226       sym->number = nvars++;
 227     }
 228
 229   return true;
 230 }
 231
 232 static bool
 233 symbol_check_defined_processor (void *sym, void *null ATTRIBUTE_UNUSED)
 234 {
 235   return symbol_check_defined (sym);
 236 }
 237
 238
 239 /*------------------------------------------------------------------.
 240 | Declare the new symbol SYM.  Make it an alias of SYMVAL, and type |
 241 | SYMVAL with SYM's type.                                           |
 242 `------------------------------------------------------------------*/
 243
 244 void
 245 symbol_make_alias (symbol *sym, symbol *symval, location loc)
 246 {
 247   if (symval->alias)
 248     warn_at (loc, _("symbol `%s' used more than once as a literal string"),
 249              symval->tag);
 250   else if (sym->alias)
 251     warn_at (loc, _("symbol `%s' given more than one literal string"),
 252              sym->tag);
 253   else
 254     {
 255       symval->class = token_sym;
 256       symval->user_token_number = sym->user_token_number;
 257       sym->user_token_number = USER_NUMBER_ALIAS;
 258       symval->alias = sym;
 259       sym->alias = symval;
 260       /* sym and symval combined are only one symbol.  */
 261       nsyms--;
 262       ntokens--;
 263       if (ntokens != sym->number && ntokens != symval->number)
 264         abort ();
 265       sym->number = symval->number =
 266         (symval->number < sym->number) ? symval->number : sym->number;
 267       symbol_type_set (symval, sym->type_name, loc);
 268     }
 269 }
 270
 271
 272 /*---------------------------------------------------------.
 273 | Check that THIS, and its alias, have same precedence and |
 274 | associativity.                                           |
 275 `---------------------------------------------------------*/
 276
 277 static inline void
 278 symbol_check_alias_consistency (symbol *this)
 279 {
 280   symbol *alias = this;
 281   symbol *orig  = this->alias;
 282
 283   /* Check only those that _are_ the aliases.  */
 284   if (!(this->alias && this->user_token_number == USER_NUMBER_ALIAS))
 285     return;
 286
 287   if (orig->type_name != alias->type_name)
 288     {
 289       if (orig->type_name)
 290         symbol_type_set (alias, orig->type_name, orig->type_location);
 291       else
 292         symbol_type_set (orig, alias->type_name, alias->type_location);
 293     }
 294
 295
 296   if (orig->destructor || alias->destructor)
 297     {
 298       if (orig->destructor)
 299         symbol_destructor_set (alias, orig->destructor,
 300                                orig->destructor_location);
 301       else
 302         symbol_destructor_set (orig, alias->destructor,
 303                                alias->destructor_location);
 304     }
 305
 306   if (orig->printer || alias->printer)
 307     {
 308       if (orig->printer)
 309         symbol_printer_set (alias, orig->printer, orig->printer_location);
 310       else
 311         symbol_printer_set (orig, alias->printer, alias->printer_location);
 312     }
 313
 314   if (alias->prec || orig->prec)
 315     {
 316       if (orig->prec)
 317         symbol_precedence_set (alias, orig->prec, orig->assoc,
 318                                orig->prec_location);
 319       else
 320         symbol_precedence_set (orig, alias->prec, alias->assoc,
 321                                alias->prec_location);
 322     }
 323 }
 324
 325 static bool
 326 symbol_check_alias_consistency_processor (void *this,
 327                                           void *null ATTRIBUTE_UNUSED)
 328 {
 329   symbol_check_alias_consistency (this);
 330   return true;
 331 }
 332
 333
 334 /*-------------------------------------------------------------------.
 335 | Assign a symbol number, and write the definition of the token name |
 336 | into FDEFINES.  Put in SYMBOLS.                                    |
 337 `-------------------------------------------------------------------*/
 338
 339 static inline bool
 340 symbol_pack (symbol *this)
 341 {
 342   if (this->class == nterm_sym)
 343     {
 344       this->number += ntokens;
 345     }
 346   else if (this->alias)
 347     {
 348       /* This symbol and its alias are a single token defn.
 349          Allocate a tokno, and assign to both check agreement of
 350          prec and assoc fields and make both the same */
 351       if (this->number == NUMBER_UNDEFINED)
 352         {
 353           if (this == endtoken || this->alias == endtoken)
 354             this->number = this->alias->number = 0;
 355           else
 356             {
 357               if (this->alias->number == NUMBER_UNDEFINED)
 358                 abort ();
 359               this->number = this->alias->number;
 360             }
 361         }
 362       /* Do not do processing below for USER_NUMBER_ALIASes.  */
 363       if (this->user_token_number == USER_NUMBER_ALIAS)
 364         return true;
 365     }
 366   else /* this->class == token_sym */
 367     {
 368       if (this->number == NUMBER_UNDEFINED)
 369         abort ();
 370     }
 371
 372   symbols[this->number] = this;
 373   return true;
 374 }
 375
 376 static bool
 377 symbol_pack_processor (void *this, void *null ATTRIBUTE_UNUSED)
 378 {
 379   return symbol_pack (this);
 380 }
 381
 382
 383
 384
 385 /*--------------------------------------------------.
 386 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 387 `--------------------------------------------------*/
 388
 389 static inline bool
 390 symbol_translation (symbol *this)
 391 {
 392   /* Non-terminal? */
 393   if (this->class == token_sym
 394       && this->user_token_number != USER_NUMBER_ALIAS)
 395     {
 396       /* A token which translation has already been set? */
 397       if (token_translations[this->user_token_number] != undeftoken->number)
 398         complain_at (this->location,
 399                      _("tokens %s and %s both assigned number %d"),
 400                      symbols[token_translations[this->user_token_number]]->tag,
 401                      this->tag, this->user_token_number);
 402
 403       token_translations[this->user_token_number] = this->number;
 404     }
 405
 406   return true;
 407 }
 408
 409 static bool
 410 symbol_translation_processor (void *this, void *null ATTRIBUTE_UNUSED)
 411 {
 412   return symbol_translation (this);
 413 }
 414
 415
 416 /*----------------------.
 417 | A symbol hash table.  |
 418 `----------------------*/
 419
 420 /* Initial capacity of symbols hash table.  */
 421 #define HT_INITIAL_CAPACITY 257
 422
 423 static struct hash_table *symbol_table = NULL;
 424
 425 static inline bool
 426 hash_compare_symbol (const symbol *m1, const symbol *m2)
 427 {
 428   /* Since tags are unique, we can compare the pointers themselves.  */
 429   return UNIQSTR_EQ (m1->tag, m2->tag);
 430 }
 431
 432 static bool
 433 hash_symbol_comparator (void const *m1, void const *m2)
 434 {
 435   return hash_compare_symbol (m1, m2);
 436 }
 437
 438 static inline size_t
 439 hash_symbol (const symbol *m, size_t tablesize)
 440 {
 441   /* Since tags are unique, we can hash the pointer itself.  */
 442   return ((uintptr_t) m->tag) % tablesize;
 443 }
 444
 445 static size_t
 446 hash_symbol_hasher (void const *m, size_t tablesize)
 447 {
 448   return hash_symbol (m, tablesize);
 449 }
 450
 451
 452 /*-------------------------------.
 453 | Create the symbol hash table.  |
 454 `-------------------------------*/
 455
 456 void
 457 symbols_new (void)
 458 {
 459   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 460                                   NULL,
 461                                   hash_symbol_hasher,
 462                                   hash_symbol_comparator,
 463                                   free);
 464 }
 465
 466
 467 /*----------------------------------------------------------------.
 468 | Find the symbol named KEY, and return it.  If it does not exist |
 469 | yet, create it.                                                 |
 470 `----------------------------------------------------------------*/
 471
 472 symbol *
 473 symbol_get (const char *key, location loc)
 474 {
 475   symbol probe;
 476   symbol *entry;
 477
 478   key = uniqstr_new (key);
 479   probe.tag = key;
 480   entry = hash_lookup (symbol_table, &probe);
 481
 482   if (!entry)
 483     {
 484       /* First insertion in the hash. */
 485       entry = symbol_new (key, loc);
 486       hash_insert (symbol_table, entry);
 487     }
 488   return entry;
 489 }
 490
 491
 492 /*------------------------------------------------------------------.
 493 | Generate a dummy nonterminal, whose name cannot conflict with the |
 494 | user's names.                                                     |
 495 `------------------------------------------------------------------*/
 496
 497 symbol *
 498 dummy_symbol_get (location loc)
 499 {
 500   /* Incremented for each generated symbol.  */
 501   static int dummy_count = 0;
 502   static char buf[256];
 503
 504   symbol *sym;
 505
 506   sprintf (buf, "@%d", ++dummy_count);
 507   sym = symbol_get (buf, loc);
 508   sym->class = nterm_sym;
 509   sym->number = nvars++;
 510   return sym;
 511 }
 512
 513
 514 /*-------------------.
 515 | Free the symbols.  |
 516 `-------------------*/
 517
 518 void
 519 symbols_free (void)
 520 {
 521   hash_free (symbol_table);
 522   free (symbols);
 523 }
 524
 525
 526 /*---------------------------------------------------------------.
 527 | Look for undefined symbols, report an error, and consider them |
 528 | terminals.                                                     |
 529 `---------------------------------------------------------------*/
 530
 531 static void
 532 symbols_do (Hash_processor processor, void *processor_data)
 533 {
 534   hash_do_for_each (symbol_table, processor, processor_data);
 535 }
 536
 537
 538 /*--------------------------------------------------------------.
 539 | Check that all the symbols are defined.  Report any undefined |
 540 | symbols and consider them nonterminals.                       |
 541 `--------------------------------------------------------------*/
 542
 543 void
 544 symbols_check_defined (void)
 545 {
 546   symbols_do (symbol_check_defined_processor, NULL);
 547 }
 548
 549 /*------------------------------------------------------------------.
 550 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 551 | number.                                                           |
 552 `------------------------------------------------------------------*/
 553
 554 static void
 555 symbols_token_translations_init (void)
 556 {
 557   bool num_256_available_p = true;
 558   int i;
 559
 560   /* Find the highest user token number, and whether 256, the POSIX
 561      preferred user token number for the error token, is used.  */
 562   max_user_token_number = 0;
 563   for (i = 0; i < ntokens; ++i)
 564     {
 565       symbol *this = symbols[i];
 566       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 567         {
 568           if (this->user_token_number > max_user_token_number)
 569             max_user_token_number = this->user_token_number;
 570           if (this->user_token_number == 256)
 571             num_256_available_p = false;
 572         }
 573     }
 574
 575   /* If 256 is not used, assign it to error, to follow POSIX.  */
 576   if (num_256_available_p
 577       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 578     errtoken->user_token_number = 256;
 579
 580   /* Set the missing user numbers. */
 581   if (max_user_token_number < 256)
 582     max_user_token_number = 256;
 583
 584   for (i = 0; i < ntokens; ++i)
 585     {
 586       symbol *this = symbols[i];
 587       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 588         this->user_token_number = ++max_user_token_number;
 589       if (this->user_token_number > max_user_token_number)
 590         max_user_token_number = this->user_token_number;
 591     }
 592
 593   token_translations = xnmalloc (max_user_token_number + 1,
 594                                  sizeof *token_translations);
 595
 596   /* Initialize all entries for literal tokens to 2, the internal
 597      token number for $undefined, which represents all invalid inputs.
 598      */
 599   for (i = 0; i < max_user_token_number + 1; i++)
 600     token_translations[i] = undeftoken->number;
 601   symbols_do (symbol_translation_processor, NULL);
 602 }
 603
 604
 605 /*----------------------------------------------------------------.
 606 | Assign symbol numbers, and write definition of token names into |
 607 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 608 `----------------------------------------------------------------*/
 609
 610 void
 611 symbols_pack (void)
 612 {
 613   symbols = xcalloc (nsyms, sizeof *symbols);
 614
 615   symbols_do (symbol_check_alias_consistency_processor, NULL);
 616   symbols_do (symbol_pack_processor, NULL);
 617
 618   symbols_token_translations_init ();
 619
 620   if (startsymbol->class == unknown_sym)
 621     fatal_at (startsymbol_location,
 622               _("the start symbol %s is undefined"),
 623               startsymbol->tag);
 624   else if (startsymbol->class == token_sym)
 625     fatal_at (startsymbol_location,
 626               _("the start symbol %s is a token"),
 627               startsymbol->tag);
 628 }