src/symtab.c

   1 /* Symbol table manager for Bison.
   2
   3    Copyright (C) 1984, 1989, 2000, 2001, 2002, 2004 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., 59 Temple Place - Suite 330,
  21    Boston, MA 02111-1307, 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   nsyms++;
  70   return res;
  71 }
  72
  73
  74 /*------------------------------------------------------------------.
  75 | Complain that S's WHAT is redeclared at SECOND, and was first set |
  76 | at FIRST.                                                         |
  77 `------------------------------------------------------------------*/
  78
  79 static void
  80 redeclaration (symbol* s, const char *what, location first, location second)
  81 {
  82   complain_at (second, _("%s redeclaration for %s"), what, s->tag);
  83   complain_at (first, _("first declaration"));
  84 }
  85
  86
  87 /*-----------------------------------------------------------------.
  88 | Set the TYPE_NAME associated with SYM.  Does nothing if passed 0 |
  89 | as TYPE_NAME.                                                    |
  90 `-----------------------------------------------------------------*/
  91
  92 void
  93 symbol_type_set (symbol *sym, uniqstr type_name, location loc)
  94 {
  95   if (type_name)
  96     {
  97       if (sym->type_name)
  98         redeclaration (sym, "%type", sym->type_location, loc);
  99       uniqstr_assert (type_name);
 100       sym->type_name = type_name;
 101       sym->type_location = loc;
 102     }
 103 }
 104
 105
 106 /*------------------------------------------------------------------.
 107 | Set the DESTRUCTOR associated with SYM.  Do nothing if passed 0.  |
 108 `------------------------------------------------------------------*/
 109
 110 void
 111 symbol_destructor_set (symbol *sym, char *destructor, location loc)
 112 {
 113   if (destructor)
 114     {
 115       if (sym->destructor)
 116         redeclaration (sym, "%destructor", sym->destructor_location, loc);
 117       sym->destructor = destructor;
 118       sym->destructor_location = loc;
 119     }
 120 }
 121
 122
 123 /*---------------------------------------------------------------.
 124 | Set the PRINTER associated with SYM.  Do nothing if passed 0.  |
 125 `---------------------------------------------------------------*/
 126
 127 void
 128 symbol_printer_set (symbol *sym, char *printer, location loc)
 129 {
 130   if (printer)
 131     {
 132       if (sym->printer)
 133         redeclaration (sym, "%printer", sym->destructor_location, loc);
 134       sym->printer = printer;
 135       sym->printer_location = loc;
 136     }
 137 }
 138
 139
 140 /*-----------------------------------------------------------------.
 141 | Set the PRECEDENCE associated with SYM.  Does nothing if invoked |
 142 | with UNDEF_ASSOC as ASSOC.                                       |
 143 `-----------------------------------------------------------------*/
 144
 145 void
 146 symbol_precedence_set (symbol *sym, int prec, assoc a, location loc)
 147 {
 148   if (a != undef_assoc)
 149     {
 150       if (sym->prec != 0)
 151         redeclaration (sym, assoc_to_string (a), sym->prec_location, loc);
 152       sym->prec = prec;
 153       sym->assoc = a;
 154       sym->prec_location = loc;
 155     }
 156
 157   /* Only terminals have a precedence. */
 158   symbol_class_set (sym, token_sym, loc);
 159 }
 160
 161
 162 /*------------------------------------.
 163 | Set the CLASS associated with SYM.  |
 164 `------------------------------------*/
 165
 166 void
 167 symbol_class_set (symbol *sym, symbol_class class, location loc)
 168 {
 169   if (sym->class != unknown_sym && sym->class != class)
 170     complain_at (loc, _("symbol %s redefined"), sym->tag);
 171
 172   if (class == nterm_sym && sym->class != nterm_sym)
 173     sym->number = nvars++;
 174   else if (class == token_sym && sym->number == NUMBER_UNDEFINED)
 175     sym->number = ntokens++;
 176
 177   sym->class = class;
 178 }
 179
 180
 181 /*------------------------------------------------.
 182 | Set the USER_TOKEN_NUMBER associated with SYM.  |
 183 `------------------------------------------------*/
 184
 185 void
 186 symbol_user_token_number_set (symbol *sym, int user_token_number, location loc)
 187 {
 188   if (sym->class != token_sym)
 189     abort ();
 190
 191   if (sym->user_token_number != USER_NUMBER_UNDEFINED
 192       && sym->user_token_number != user_token_number)
 193     complain_at (loc, _("redefining user token number of %s"), sym->tag);
 194
 195   sym->user_token_number = user_token_number;
 196   /* User defined $end token? */
 197   if (user_token_number == 0)
 198     {
 199       endtoken = sym;
 200       endtoken->number = 0;
 201       /* It is always mapped to 0, so it was already counted in
 202          NTOKENS.  */
 203       --ntokens;
 204     }
 205 }
 206
 207
 208 /*----------------------------------------------------------.
 209 | If SYM is not defined, report an error, and consider it a |
 210 | nonterminal.                                              |
 211 `----------------------------------------------------------*/
 212
 213 static inline bool
 214 symbol_check_defined (symbol *sym)
 215 {
 216   if (sym->class == unknown_sym)
 217     {
 218       complain_at
 219         (sym->location,
 220          _("symbol %s is used, but is not defined as a token and has no rules"),
 221          sym->tag);
 222       sym->class = nterm_sym;
 223       sym->number = nvars++;
 224     }
 225
 226   return true;
 227 }
 228
 229 static bool
 230 symbol_check_defined_processor (void *sym, void *null ATTRIBUTE_UNUSED)
 231 {
 232   return symbol_check_defined (sym);
 233 }
 234
 235
 236 /*------------------------------------------------------------------.
 237 | Declare the new symbol SYM.  Make it an alias of SYMVAL, and type |
 238 | SYMVAL with SYM's type.                                           |
 239 `------------------------------------------------------------------*/
 240
 241 void
 242 symbol_make_alias (symbol *sym, symbol *symval, location loc)
 243 {
 244   if (symval->alias)
 245     warn_at (loc, _("symbol `%s' used more than once as a literal string"),
 246              symval->tag);
 247   else if (sym->alias)
 248     warn_at (loc, _("symbol `%s' given more than one literal string"),
 249              sym->tag);
 250   else
 251     {
 252       symval->class = token_sym;
 253       symval->user_token_number = sym->user_token_number;
 254       sym->user_token_number = USER_NUMBER_ALIAS;
 255       symval->alias = sym;
 256       sym->alias = symval;
 257       /* sym and symval combined are only one symbol.  */
 258       nsyms--;
 259       ntokens--;
 260       if (ntokens != sym->number && ntokens != symval->number)
 261         abort ();
 262       sym->number = symval->number =
 263         (symval->number < sym->number) ? symval->number : sym->number;
 264     }
 265 }
 266
 267
 268 /*---------------------------------------------------------.
 269 | Check that THIS, and its alias, have same precedence and |
 270 | associativity.                                           |
 271 `---------------------------------------------------------*/
 272
 273 static inline void
 274 symbol_check_alias_consistency (symbol *this)
 275 {
 276   symbol *alias = this;
 277   symbol *orig  = this->alias;
 278
 279   /* Check only those that _are_ the aliases.  */
 280   if (!(this->alias && this->user_token_number == USER_NUMBER_ALIAS))
 281     return;
 282
 283   if (orig->type_name || alias->type_name)
 284     {
 285       if (orig->type_name)
 286         symbol_type_set (alias, orig->type_name, orig->type_location);
 287       else
 288         symbol_type_set (orig, alias->type_name, alias->type_location);
 289     }
 290
 291
 292   if (orig->destructor || alias->destructor)
 293     {
 294       if (orig->destructor)
 295         symbol_destructor_set (alias, orig->destructor,
 296                                orig->destructor_location);
 297       else
 298         symbol_destructor_set (orig, alias->destructor,
 299                                alias->destructor_location);
 300     }
 301
 302   if (orig->printer || alias->printer)
 303     {
 304       if (orig->printer)
 305         symbol_printer_set (alias, orig->printer, orig->printer_location);
 306       else
 307         symbol_printer_set (orig, alias->printer, alias->printer_location);
 308     }
 309
 310   if (alias->prec || orig->prec)
 311     {
 312       if (orig->prec)
 313         symbol_precedence_set (alias, orig->prec, orig->assoc,
 314                                orig->prec_location);
 315       else
 316         symbol_precedence_set (orig, alias->prec, alias->assoc,
 317                                alias->prec_location);
 318     }
 319 }
 320
 321 static bool
 322 symbol_check_alias_consistency_processor (void *this,
 323                                           void *null ATTRIBUTE_UNUSED)
 324 {
 325   symbol_check_alias_consistency (this);
 326   return true;
 327 }
 328
 329
 330 /*-------------------------------------------------------------------.
 331 | Assign a symbol number, and write the definition of the token name |
 332 | into FDEFINES.  Put in SYMBOLS.                                    |
 333 `-------------------------------------------------------------------*/
 334
 335 static inline bool
 336 symbol_pack (symbol *this)
 337 {
 338   if (this->class == nterm_sym)
 339     {
 340       this->number += ntokens;
 341     }
 342   else if (this->alias)
 343     {
 344       /* This symbol and its alias are a single token defn.
 345          Allocate a tokno, and assign to both check agreement of
 346          prec and assoc fields and make both the same */
 347       if (this->number == NUMBER_UNDEFINED)
 348         {
 349           if (this == endtoken || this->alias == endtoken)
 350             this->number = this->alias->number = 0;
 351           else
 352             {
 353               if (this->alias->number == NUMBER_UNDEFINED)
 354                 abort ();
 355               this->number = this->alias->number;
 356             }
 357         }
 358       /* Do not do processing below for USER_NUMBER_ALIASes.  */
 359       if (this->user_token_number == USER_NUMBER_ALIAS)
 360         return true;
 361     }
 362   else /* this->class == token_sym */
 363     {
 364       if (this->number == NUMBER_UNDEFINED)
 365         abort ();
 366     }
 367
 368   symbols[this->number] = this;
 369   return true;
 370 }
 371
 372 static bool
 373 symbol_pack_processor (void *this, void *null ATTRIBUTE_UNUSED)
 374 {
 375   return symbol_pack (this);
 376 }
 377
 378
 379
 380
 381 /*--------------------------------------------------.
 382 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 383 `--------------------------------------------------*/
 384
 385 static inline bool
 386 symbol_translation (symbol *this)
 387 {
 388   /* Non-terminal? */
 389   if (this->class == token_sym
 390       && this->user_token_number != USER_NUMBER_ALIAS)
 391     {
 392       /* A token which translation has already been set? */
 393       if (token_translations[this->user_token_number] != undeftoken->number)
 394         complain_at (this->location,
 395                      _("tokens %s and %s both assigned number %d"),
 396                      symbols[token_translations[this->user_token_number]]->tag,
 397                      this->tag, this->user_token_number);
 398
 399       token_translations[this->user_token_number] = this->number;
 400     }
 401
 402   return true;
 403 }
 404
 405 static bool
 406 symbol_translation_processor (void *this, void *null ATTRIBUTE_UNUSED)
 407 {
 408   return symbol_translation (this);
 409 }
 410
 411
 412 /*----------------------.
 413 | A symbol hash table.  |
 414 `----------------------*/
 415
 416 /* Initial capacity of symbols hash table.  */
 417 #define HT_INITIAL_CAPACITY 257
 418
 419 static struct hash_table *symbol_table = NULL;
 420
 421 static inline bool
 422 hash_compare_symbol (const symbol *m1, const symbol *m2)
 423 {
 424   /* Since tags are unique, we can compare the pointers themselves.  */
 425   return UNIQSTR_EQ (m1->tag, m2->tag);
 426 }
 427
 428 static bool
 429 hash_symbol_comparator (void const *m1, void const *m2)
 430 {
 431   return hash_compare_symbol (m1, m2);
 432 }
 433
 434 static inline size_t
 435 hash_symbol (const symbol *m, size_t tablesize)
 436 {
 437   /* Since tags are unique, we can hash the pointer itself.  */
 438   return ((uintptr_t) m->tag) % tablesize;
 439 }
 440
 441 static size_t
 442 hash_symbol_hasher (void const *m, size_t tablesize)
 443 {
 444   return hash_symbol (m, tablesize);
 445 }
 446
 447
 448 /*-------------------------------.
 449 | Create the symbol hash table.  |
 450 `-------------------------------*/
 451
 452 void
 453 symbols_new (void)
 454 {
 455   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 456                                   NULL,
 457                                   hash_symbol_hasher,
 458                                   hash_symbol_comparator,
 459                                   free);
 460 }
 461
 462
 463 /*----------------------------------------------------------------.
 464 | Find the symbol named KEY, and return it.  If it does not exist |
 465 | yet, create it.                                                 |
 466 `----------------------------------------------------------------*/
 467
 468 symbol *
 469 symbol_get (const char *key, location loc)
 470 {
 471   symbol probe;
 472   symbol *entry;
 473
 474   /* Keep the symbol in a printable form.  */
 475   key = uniqstr_new (quotearg_style (escape_quoting_style, key));
 476   probe.tag = key;
 477   entry = hash_lookup (symbol_table, &probe);
 478
 479   if (!entry)
 480     {
 481       /* First insertion in the hash. */
 482       entry = symbol_new (key, loc);
 483       hash_insert (symbol_table, entry);
 484     }
 485   return entry;
 486 }
 487
 488
 489 /*------------------------------------------------------------------.
 490 | Generate a dummy nonterminal, whose name cannot conflict with the |
 491 | user's names.                                                     |
 492 `------------------------------------------------------------------*/
 493
 494 symbol *
 495 dummy_symbol_get (location loc)
 496 {
 497   /* Incremented for each generated symbol.  */
 498   static int dummy_count = 0;
 499   static char buf[256];
 500
 501   symbol *sym;
 502
 503   sprintf (buf, "@%d", ++dummy_count);
 504   sym = symbol_get (buf, loc);
 505   sym->class = nterm_sym;
 506   sym->number = nvars++;
 507   return sym;
 508 }
 509
 510
 511 /*-------------------.
 512 | Free the symbols.  |
 513 `-------------------*/
 514
 515 void
 516 symbols_free (void)
 517 {
 518   hash_free (symbol_table);
 519   free (symbols);
 520 }
 521
 522
 523 /*---------------------------------------------------------------.
 524 | Look for undefined symbols, report an error, and consider them |
 525 | terminals.                                                     |
 526 `---------------------------------------------------------------*/
 527
 528 static void
 529 symbols_do (Hash_processor processor, void *processor_data)
 530 {
 531   hash_do_for_each (symbol_table, processor, processor_data);
 532 }
 533
 534
 535 /*--------------------------------------------------------------.
 536 | Check that all the symbols are defined.  Report any undefined |
 537 | symbols and consider them nonterminals.                       |
 538 `--------------------------------------------------------------*/
 539
 540 void
 541 symbols_check_defined (void)
 542 {
 543   symbols_do (symbol_check_defined_processor, NULL);
 544 }
 545
 546 /*------------------------------------------------------------------.
 547 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 548 | number.                                                           |
 549 `------------------------------------------------------------------*/
 550
 551 static void
 552 symbols_token_translations_init (void)
 553 {
 554   bool num_256_available_p = true;
 555   int i;
 556
 557   /* Find the highest user token number, and whether 256, the POSIX
 558      preferred user token number for the error token, is used.  */
 559   max_user_token_number = 0;
 560   for (i = 0; i < ntokens; ++i)
 561     {
 562       symbol *this = symbols[i];
 563       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 564         {
 565           if (this->user_token_number > max_user_token_number)
 566             max_user_token_number = this->user_token_number;
 567           if (this->user_token_number == 256)
 568             num_256_available_p = false;
 569         }
 570     }
 571
 572   /* If 256 is not used, assign it to error, to follow POSIX.  */
 573   if (num_256_available_p
 574       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 575     errtoken->user_token_number = 256;
 576
 577   /* Set the missing user numbers. */
 578   if (max_user_token_number < 256)
 579     max_user_token_number = 256;
 580
 581   for (i = 0; i < ntokens; ++i)
 582     {
 583       symbol *this = symbols[i];
 584       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 585         this->user_token_number = ++max_user_token_number;
 586       if (this->user_token_number > max_user_token_number)
 587         max_user_token_number = this->user_token_number;
 588     }
 589
 590   token_translations = xnmalloc (max_user_token_number + 1,
 591                                  sizeof *token_translations);
 592
 593   /* Initialize all entries for literal tokens to 2, the internal
 594      token number for $undefined, which represents all invalid inputs.
 595      */
 596   for (i = 0; i < max_user_token_number + 1; i++)
 597     token_translations[i] = undeftoken->number;
 598   symbols_do (symbol_translation_processor, NULL);
 599 }
 600
 601
 602 /*----------------------------------------------------------------.
 603 | Assign symbol numbers, and write definition of token names into |
 604 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 605 `----------------------------------------------------------------*/
 606
 607 void
 608 symbols_pack (void)
 609 {
 610   symbols = xcalloc (nsyms, sizeof *symbols);
 611
 612   symbols_do (symbol_check_alias_consistency_processor, NULL);
 613   symbols_do (symbol_pack_processor, NULL);
 614
 615   symbols_token_translations_init ();
 616
 617   if (startsymbol->class == unknown_sym)
 618     fatal_at (startsymbol_location,
 619               _("the start symbol %s is undefined"),
 620               startsymbol->tag);
 621   else if (startsymbol->class == token_sym)
 622     fatal_at (startsymbol_location,
 623               _("the start symbol %s is a token"),
 624               startsymbol->tag);
 625 }