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