src/symtab.c

   1 /* Symbol table manager for Bison.
   2
   3    Copyright (C) 1984, 1989, 2000, 2001, 2002, 2004, 2005, 2006 Free
   4    Software 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 #include <config.h>
  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   res->declared = false;
  69
  70   if (nsyms == SYMBOL_NUMBER_MAXIMUM)
  71     fatal (_("too many symbols in input grammar (limit is %d)"),
  72            SYMBOL_NUMBER_MAXIMUM);
  73   nsyms++;
  74   return res;
  75 }
  76
  77
  78 /*-----------------.
  79 | Print a symbol.  |
  80 `-----------------*/
  81
  82 #define SYMBOL_ATTR_PRINT(Attr)                         \
  83   if (s->Attr)                                          \
  84     fprintf (f, " %s { %s }", #Attr, s->Attr)
  85
  86 void
  87 symbol_print (symbol *s, FILE *f)
  88 {
  89   if (s)
  90     {
  91       fprintf (f, "\"%s\"", s->tag);
  92       SYMBOL_ATTR_PRINT (type_name);
  93       SYMBOL_ATTR_PRINT (destructor);
  94       SYMBOL_ATTR_PRINT (printer);
  95     }
  96   else
  97     fprintf (f, "<NULL>");
  98 }
  99
 100 #undef SYMBOL_ATTR_PRINT
 101
 102 /*------------------------------------------------------------------.
 103 | Complain that S's WHAT is redeclared at SECOND, and was first set |
 104 | at FIRST.                                                         |
 105 `------------------------------------------------------------------*/
 106
 107 static void
 108 redeclaration (symbol* s, const char *what, location first, location second)
 109 {
 110   complain_at (second, _("%s redeclaration for %s"), what, s->tag);
 111   complain_at (first, _("first declaration"));
 112 }
 113
 114
 115 /*-----------------------------------------------------------------.
 116 | Set the TYPE_NAME associated with SYM.  Does nothing if passed 0 |
 117 | as TYPE_NAME.                                                    |
 118 `-----------------------------------------------------------------*/
 119
 120 void
 121 symbol_type_set (symbol *sym, uniqstr type_name, location loc)
 122 {
 123   if (type_name)
 124     {
 125       if (sym->type_name)
 126         redeclaration (sym, "%type", sym->type_location, loc);
 127       uniqstr_assert (type_name);
 128       sym->type_name = type_name;
 129       sym->type_location = loc;
 130     }
 131 }
 132
 133
 134 /*------------------------------------------------------------------.
 135 | Set the DESTRUCTOR associated with SYM.  Do nothing if passed 0.  |
 136 `------------------------------------------------------------------*/
 137
 138 void
 139 symbol_destructor_set (symbol *sym, const char *destructor, location loc)
 140 {
 141   if (destructor)
 142     {
 143       if (sym->destructor)
 144         redeclaration (sym, "%destructor", sym->destructor_location, loc);
 145       sym->destructor = destructor;
 146       sym->destructor_location = loc;
 147     }
 148 }
 149
 150
 151 /*---------------------------------------------------------------.
 152 | Set the PRINTER associated with SYM.  Do nothing if passed 0.  |
 153 `---------------------------------------------------------------*/
 154
 155 void
 156 symbol_printer_set (symbol *sym, const char *printer, location loc)
 157 {
 158   if (printer)
 159     {
 160       if (sym->printer)
 161         redeclaration (sym, "%printer", sym->destructor_location, loc);
 162       sym->printer = printer;
 163       sym->printer_location = loc;
 164     }
 165 }
 166
 167
 168 /*-----------------------------------------------------------------.
 169 | Set the PRECEDENCE associated with SYM.  Does nothing if invoked |
 170 | with UNDEF_ASSOC as ASSOC.                                       |
 171 `-----------------------------------------------------------------*/
 172
 173 void
 174 symbol_precedence_set (symbol *sym, int prec, assoc a, location loc)
 175 {
 176   if (a != undef_assoc)
 177     {
 178       if (sym->prec != 0)
 179         redeclaration (sym, assoc_to_string (a), sym->prec_location, loc);
 180       sym->prec = prec;
 181       sym->assoc = a;
 182       sym->prec_location = loc;
 183     }
 184
 185   /* Only terminals have a precedence. */
 186   symbol_class_set (sym, token_sym, loc, false);
 187 }
 188
 189
 190 /*------------------------------------.
 191 | Set the CLASS associated with SYM.  |
 192 `------------------------------------*/
 193
 194 void
 195 symbol_class_set (symbol *sym, symbol_class class, location loc, bool declaring)
 196 {
 197   if (sym->class != unknown_sym && sym->class != class)
 198     {
 199       complain_at (loc, _("symbol %s redefined"), sym->tag);
 200       sym->declared = false;
 201     }
 202
 203   if (class == nterm_sym && sym->class != nterm_sym)
 204     sym->number = nvars++;
 205   else if (class == token_sym && sym->number == NUMBER_UNDEFINED)
 206     sym->number = ntokens++;
 207
 208   sym->class = class;
 209
 210   if (declaring)
 211     {
 212       if (sym->declared)
 213         warn_at (loc, _("symbol %s redeclared"), sym->tag);
 214       sym->declared = true;
 215     }
 216 }
 217
 218
 219 /*------------------------------------------------.
 220 | Set the USER_TOKEN_NUMBER associated with SYM.  |
 221 `------------------------------------------------*/
 222
 223 void
 224 symbol_user_token_number_set (symbol *sym, int user_token_number, location loc)
 225 {
 226   if (sym->class != token_sym)
 227     abort ();
 228
 229   if (sym->user_token_number != USER_NUMBER_UNDEFINED
 230       && sym->user_token_number != user_token_number)
 231     complain_at (loc, _("redefining user token number of %s"), sym->tag);
 232
 233   sym->user_token_number = user_token_number;
 234   /* User defined $end token? */
 235   if (user_token_number == 0)
 236     {
 237       endtoken = sym;
 238       endtoken->number = 0;
 239       /* It is always mapped to 0, so it was already counted in
 240          NTOKENS.  */
 241       --ntokens;
 242     }
 243 }
 244
 245
 246 /*----------------------------------------------------------.
 247 | If SYM is not defined, report an error, and consider it a |
 248 | nonterminal.                                              |
 249 `----------------------------------------------------------*/
 250
 251 static inline bool
 252 symbol_check_defined (symbol *sym)
 253 {
 254   if (sym->class == unknown_sym)
 255     {
 256       complain_at
 257         (sym->location,
 258          _("symbol %s is used, but is not defined as a token and has no rules"),
 259          sym->tag);
 260       sym->class = nterm_sym;
 261       sym->number = nvars++;
 262     }
 263
 264   return true;
 265 }
 266
 267 static bool
 268 symbol_check_defined_processor (void *sym, void *null ATTRIBUTE_UNUSED)
 269 {
 270   return symbol_check_defined (sym);
 271 }
 272
 273
 274 /*------------------------------------------------------------------.
 275 | Declare the new symbol SYM.  Make it an alias of SYMVAL, and type |
 276 | SYMVAL with SYM's type.                                           |
 277 `------------------------------------------------------------------*/
 278
 279 void
 280 symbol_make_alias (symbol *sym, symbol *symval, location loc)
 281 {
 282   if (symval->alias)
 283     warn_at (loc, _("symbol `%s' used more than once as a literal string"),
 284              symval->tag);
 285   else if (sym->alias)
 286     warn_at (loc, _("symbol `%s' given more than one literal string"),
 287              sym->tag);
 288   else
 289     {
 290       symval->class = token_sym;
 291       symval->user_token_number = sym->user_token_number;
 292       sym->user_token_number = USER_NUMBER_ALIAS;
 293       symval->alias = sym;
 294       sym->alias = symval;
 295       /* sym and symval combined are only one symbol.  */
 296       nsyms--;
 297       ntokens--;
 298       if (ntokens != sym->number && ntokens != symval->number)
 299         abort ();
 300       sym->number = symval->number =
 301         (symval->number < sym->number) ? symval->number : sym->number;
 302       symbol_type_set (symval, sym->type_name, loc);
 303     }
 304 }
 305
 306
 307 /*---------------------------------------------------------.
 308 | Check that THIS, and its alias, have same precedence and |
 309 | associativity.                                           |
 310 `---------------------------------------------------------*/
 311
 312 static inline void
 313 symbol_check_alias_consistency (symbol *this)
 314 {
 315   symbol *alias = this;
 316   symbol *orig  = this->alias;
 317
 318   /* Check only those that _are_ the aliases.  */
 319   if (!(this->alias && this->user_token_number == USER_NUMBER_ALIAS))
 320     return;
 321
 322   if (orig->type_name != alias->type_name)
 323     {
 324       if (orig->type_name)
 325         symbol_type_set (alias, orig->type_name, orig->type_location);
 326       else
 327         symbol_type_set (orig, alias->type_name, alias->type_location);
 328     }
 329
 330
 331   if (orig->destructor || alias->destructor)
 332     {
 333       if (orig->destructor)
 334         symbol_destructor_set (alias, orig->destructor,
 335                                orig->destructor_location);
 336       else
 337         symbol_destructor_set (orig, alias->destructor,
 338                                alias->destructor_location);
 339     }
 340
 341   if (orig->printer || alias->printer)
 342     {
 343       if (orig->printer)
 344         symbol_printer_set (alias, orig->printer, orig->printer_location);
 345       else
 346         symbol_printer_set (orig, alias->printer, alias->printer_location);
 347     }
 348
 349   if (alias->prec || orig->prec)
 350     {
 351       if (orig->prec)
 352         symbol_precedence_set (alias, orig->prec, orig->assoc,
 353                                orig->prec_location);
 354       else
 355         symbol_precedence_set (orig, alias->prec, alias->assoc,
 356                                alias->prec_location);
 357     }
 358 }
 359
 360 static bool
 361 symbol_check_alias_consistency_processor (void *this,
 362                                           void *null ATTRIBUTE_UNUSED)
 363 {
 364   symbol_check_alias_consistency (this);
 365   return true;
 366 }
 367
 368
 369 /*-------------------------------------------------------------------.
 370 | Assign a symbol number, and write the definition of the token name |
 371 | into FDEFINES.  Put in SYMBOLS.                                    |
 372 `-------------------------------------------------------------------*/
 373
 374 static inline bool
 375 symbol_pack (symbol *this)
 376 {
 377   if (this->class == nterm_sym)
 378     {
 379       this->number += ntokens;
 380     }
 381   else if (this->alias)
 382     {
 383       /* This symbol and its alias are a single token defn.
 384          Allocate a tokno, and assign to both check agreement of
 385          prec and assoc fields and make both the same */
 386       if (this->number == NUMBER_UNDEFINED)
 387         {
 388           if (this == endtoken || this->alias == endtoken)
 389             this->number = this->alias->number = 0;
 390           else
 391             {
 392               if (this->alias->number == NUMBER_UNDEFINED)
 393                 abort ();
 394               this->number = this->alias->number;
 395             }
 396         }
 397       /* Do not do processing below for USER_NUMBER_ALIASes.  */
 398       if (this->user_token_number == USER_NUMBER_ALIAS)
 399         return true;
 400     }
 401   else /* this->class == token_sym */
 402     {
 403       if (this->number == NUMBER_UNDEFINED)
 404         abort ();
 405     }
 406
 407   symbols[this->number] = this;
 408   return true;
 409 }
 410
 411 static bool
 412 symbol_pack_processor (void *this, void *null ATTRIBUTE_UNUSED)
 413 {
 414   return symbol_pack (this);
 415 }
 416
 417
 418
 419
 420 /*--------------------------------------------------.
 421 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 422 `--------------------------------------------------*/
 423
 424 static inline bool
 425 symbol_translation (symbol *this)
 426 {
 427   /* Non-terminal? */
 428   if (this->class == token_sym
 429       && this->user_token_number != USER_NUMBER_ALIAS)
 430     {
 431       /* A token which translation has already been set? */
 432       if (token_translations[this->user_token_number] != undeftoken->number)
 433         complain_at (this->location,
 434                      _("tokens %s and %s both assigned number %d"),
 435                      symbols[token_translations[this->user_token_number]]->tag,
 436                      this->tag, this->user_token_number);
 437
 438       token_translations[this->user_token_number] = this->number;
 439     }
 440
 441   return true;
 442 }
 443
 444 static bool
 445 symbol_translation_processor (void *this, void *null ATTRIBUTE_UNUSED)
 446 {
 447   return symbol_translation (this);
 448 }
 449
 450
 451 /*----------------------.
 452 | A symbol hash table.  |
 453 `----------------------*/
 454
 455 /* Initial capacity of symbols hash table.  */
 456 #define HT_INITIAL_CAPACITY 257
 457
 458 static struct hash_table *symbol_table = NULL;
 459
 460 static inline bool
 461 hash_compare_symbol (const symbol *m1, const symbol *m2)
 462 {
 463   /* Since tags are unique, we can compare the pointers themselves.  */
 464   return UNIQSTR_EQ (m1->tag, m2->tag);
 465 }
 466
 467 static bool
 468 hash_symbol_comparator (void const *m1, void const *m2)
 469 {
 470   return hash_compare_symbol (m1, m2);
 471 }
 472
 473 static inline size_t
 474 hash_symbol (const symbol *m, size_t tablesize)
 475 {
 476   /* Since tags are unique, we can hash the pointer itself.  */
 477   return ((uintptr_t) m->tag) % tablesize;
 478 }
 479
 480 static size_t
 481 hash_symbol_hasher (void const *m, size_t tablesize)
 482 {
 483   return hash_symbol (m, tablesize);
 484 }
 485
 486
 487 /*-------------------------------.
 488 | Create the symbol hash table.  |
 489 `-------------------------------*/
 490
 491 void
 492 symbols_new (void)
 493 {
 494   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 495                                   NULL,
 496                                   hash_symbol_hasher,
 497                                   hash_symbol_comparator,
 498                                   free);
 499 }
 500
 501
 502 /*----------------------------------------------------------------.
 503 | Find the symbol named KEY, and return it.  If it does not exist |
 504 | yet, create it.                                                 |
 505 `----------------------------------------------------------------*/
 506
 507 symbol *
 508 symbol_get (const char *key, location loc)
 509 {
 510   symbol probe;
 511   symbol *entry;
 512
 513   key = uniqstr_new (key);
 514   probe.tag = key;
 515   entry = hash_lookup (symbol_table, &probe);
 516
 517   if (!entry)
 518     {
 519       /* First insertion in the hash. */
 520       entry = symbol_new (key, loc);
 521       hash_insert (symbol_table, entry);
 522     }
 523   return entry;
 524 }
 525
 526
 527 /*------------------------------------------------------------------.
 528 | Generate a dummy nonterminal, whose name cannot conflict with the |
 529 | user's names.                                                     |
 530 `------------------------------------------------------------------*/
 531
 532 symbol *
 533 dummy_symbol_get (location loc)
 534 {
 535   /* Incremented for each generated symbol.  */
 536   static int dummy_count = 0;
 537   static char buf[256];
 538
 539   symbol *sym;
 540
 541   sprintf (buf, "@%d", ++dummy_count);
 542   sym = symbol_get (buf, loc);
 543   sym->class = nterm_sym;
 544   sym->number = nvars++;
 545   return sym;
 546 }
 547
 548
 549 /*-------------------.
 550 | Free the symbols.  |
 551 `-------------------*/
 552
 553 void
 554 symbols_free (void)
 555 {
 556   hash_free (symbol_table);
 557   free (symbols);
 558 }
 559
 560
 561 /*---------------------------------------------------------------.
 562 | Look for undefined symbols, report an error, and consider them |
 563 | terminals.                                                     |
 564 `---------------------------------------------------------------*/
 565
 566 static void
 567 symbols_do (Hash_processor processor, void *processor_data)
 568 {
 569   hash_do_for_each (symbol_table, processor, processor_data);
 570 }
 571
 572
 573 /*--------------------------------------------------------------.
 574 | Check that all the symbols are defined.  Report any undefined |
 575 | symbols and consider them nonterminals.                       |
 576 `--------------------------------------------------------------*/
 577
 578 void
 579 symbols_check_defined (void)
 580 {
 581   symbols_do (symbol_check_defined_processor, NULL);
 582 }
 583
 584 /*------------------------------------------------------------------.
 585 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 586 | number.                                                           |
 587 `------------------------------------------------------------------*/
 588
 589 static void
 590 symbols_token_translations_init (void)
 591 {
 592   bool num_256_available_p = true;
 593   int i;
 594
 595   /* Find the highest user token number, and whether 256, the POSIX
 596      preferred user token number for the error token, is used.  */
 597   max_user_token_number = 0;
 598   for (i = 0; i < ntokens; ++i)
 599     {
 600       symbol *this = symbols[i];
 601       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 602         {
 603           if (this->user_token_number > max_user_token_number)
 604             max_user_token_number = this->user_token_number;
 605           if (this->user_token_number == 256)
 606             num_256_available_p = false;
 607         }
 608     }
 609
 610   /* If 256 is not used, assign it to error, to follow POSIX.  */
 611   if (num_256_available_p
 612       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 613     errtoken->user_token_number = 256;
 614
 615   /* Set the missing user numbers. */
 616   if (max_user_token_number < 256)
 617     max_user_token_number = 256;
 618
 619   for (i = 0; i < ntokens; ++i)
 620     {
 621       symbol *this = symbols[i];
 622       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 623         this->user_token_number = ++max_user_token_number;
 624       if (this->user_token_number > max_user_token_number)
 625         max_user_token_number = this->user_token_number;
 626     }
 627
 628   token_translations = xnmalloc (max_user_token_number + 1,
 629                                  sizeof *token_translations);
 630
 631   /* Initialize all entries for literal tokens to 2, the internal
 632      token number for $undefined, which represents all invalid inputs.
 633      */
 634   for (i = 0; i < max_user_token_number + 1; i++)
 635     token_translations[i] = undeftoken->number;
 636   symbols_do (symbol_translation_processor, NULL);
 637 }
 638
 639
 640 /*----------------------------------------------------------------.
 641 | Assign symbol numbers, and write definition of token names into |
 642 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 643 `----------------------------------------------------------------*/
 644
 645 void
 646 symbols_pack (void)
 647 {
 648   symbols = xcalloc (nsyms, sizeof *symbols);
 649
 650   symbols_do (symbol_check_alias_consistency_processor, NULL);
 651   symbols_do (symbol_pack_processor, NULL);
 652
 653   symbols_token_translations_init ();
 654
 655   if (startsymbol->class == unknown_sym)
 656     fatal_at (startsymbol_location,
 657               _("the start symbol %s is undefined"),
 658               startsymbol->tag);
 659   else if (startsymbol->class == token_sym)
 660     fatal_at (startsymbol_location,
 661               _("the start symbol %s is a token"),
 662               startsymbol->tag);
 663 }