src/symtab.c

   1 /* Symbol table manager for Bison,
   2    Copyright (C) 1984, 1989, 2000, 2001, 2002 Free Software Foundation, Inc.
   3
   4    This file is part of Bison, the GNU Compiler Compiler.
   5
   6    Bison is free software; you can redistribute it and/or modify
   7    it under the terms of the GNU General Public License as published by
   8    the Free Software Foundation; either version 2, or (at your option)
   9    any later version.
  10
  11    Bison is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14    GNU General Public License for more details.
  15
  16    You should have received a copy of the GNU General Public License
  17    along with Bison; see the file COPYING.  If not, write to
  18    the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19    Boston, MA 02111-1307, USA.  */
  20
  21
  22 #include "system.h"
  23 #include "quotearg.h"
  24 #include "hash.h"
  25 #include "complain.h"
  26 #include "symtab.h"
  27 #include "gram.h"
  28
  29 /*------------------------.
  30 | Distinguished symbols.  |
  31 `------------------------*/
  32
  33 symbol_t *errtoken = NULL;
  34 symbol_t *undeftoken = NULL;
  35 symbol_t *endtoken = NULL;
  36 symbol_t *accept = NULL;
  37 symbol_t *startsymbol = NULL;
  38 location_t startsymbol_location;
  39
  40 /*---------------------------------.
  41 | Create a new symbol, named TAG.  |
  42 `---------------------------------*/
  43
  44 static symbol_t *
  45 symbol_new (const char *tag, location_t location)
  46 {
  47   symbol_t *res = XMALLOC (symbol_t, 1);
  48
  49   res->tag = xstrdup (tag);
  50   res->location = location;
  51
  52   res->type_name = NULL;
  53   res->destructor = NULL;
  54   res->printer = NULL;
  55
  56   res->number = NUMBER_UNDEFINED;
  57   res->prec = 0;
  58   res->assoc = undef_assoc;
  59   res->user_token_number = USER_NUMBER_UNDEFINED;
  60
  61   res->alias = NULL;
  62   res->class = unknown_sym;
  63
  64   nsyms++;
  65   return res;
  66 }
  67
  68
  69 /*------------------------------------------------------------------.
  70 | Set the TYPE_NAME associated to SYMBOL.  Does nothing if passed 0 |
  71 | as TYPE_NAME.                                                     |
  72 `------------------------------------------------------------------*/
  73
  74 void
  75 symbol_type_set (symbol_t *symbol, char *type_name, location_t location)
  76 {
  77   if (type_name)
  78     {
  79       if (symbol->type_name)
  80         complain_at (location,
  81                      _("type redeclaration for %s"), symbol->tag);
  82       symbol->type_name = type_name;
  83     }
  84 }
  85
  86
  87 /*-------------------------------------------------------------------.
  88 | Set the DESTRUCTOR associated to SYMBOL.  Do nothing if passed 0.  |
  89 `-------------------------------------------------------------------*/
  90
  91 void
  92 symbol_destructor_set (symbol_t *symbol, char *destructor, location_t location)
  93 {
  94   if (destructor)
  95     {
  96       if (symbol->destructor)
  97         complain_at (location,
  98                      _("%s redeclaration for %s"),
  99                      "%destructor", symbol->tag);
 100       symbol->destructor = destructor;
 101       symbol->destructor_location = location;
 102     }
 103 }
 104
 105
 106 /*----------------------------------------------------------------.
 107 | Set the PRITNER associated to SYMBOL.  Do nothing if passed 0.  |
 108 `----------------------------------------------------------------*/
 109
 110 void
 111 symbol_printer_set (symbol_t *symbol, char *printer, location_t location)
 112 {
 113   if (printer)
 114     {
 115       if (symbol->printer)
 116         complain_at (location,
 117                      _("%s redeclaration for %s"),
 118                      "%printer", symbol->tag);
 119       symbol->printer = printer;
 120       symbol->printer_location = location;
 121     }
 122 }
 123
 124
 125 /*------------------------------------------------------------------.
 126 | Set the PRECEDENCE associated to SYMBOL.  Does nothing if invoked |
 127 | with UNDEF_ASSOC as ASSOC.                                        |
 128 `------------------------------------------------------------------*/
 129
 130 void
 131 symbol_precedence_set (symbol_t *symbol,
 132                        int prec, assoc_t assoc, location_t location)
 133 {
 134   if (assoc != undef_assoc)
 135     {
 136       if (symbol->prec != 0)
 137         complain_at (location,
 138                      _("redefining precedence of %s"),
 139                      symbol->tag);
 140       symbol->prec = prec;
 141       symbol->assoc = assoc;
 142     }
 143
 144   /* Only terminals have a precedence. */
 145   symbol_class_set (symbol, token_sym, location);
 146 }
 147
 148
 149 /*-------------------------------------.
 150 | Set the CLASS associated to SYMBOL.  |
 151 `-------------------------------------*/
 152
 153 void
 154 symbol_class_set (symbol_t *symbol, symbol_class class, location_t location)
 155 {
 156   if (symbol->class != unknown_sym && symbol->class != class)
 157     complain_at (location, _("symbol %s redefined"), symbol->tag);
 158
 159   if (class == nterm_sym && symbol->class != nterm_sym)
 160     symbol->number = nvars++;
 161   else if (class == token_sym && symbol->number == NUMBER_UNDEFINED)
 162     symbol->number = ntokens++;
 163
 164   symbol->class = class;
 165 }
 166
 167
 168 /*-------------------------------------------------.
 169 | Set the USER_TOKEN_NUMBER associated to SYMBOL.  |
 170 `-------------------------------------------------*/
 171
 172 void
 173 symbol_user_token_number_set (symbol_t *symbol,
 174                               int user_token_number, location_t location)
 175 {
 176   assert (symbol->class == token_sym);
 177
 178   if (symbol->user_token_number != USER_NUMBER_UNDEFINED
 179       && symbol->user_token_number != user_token_number)
 180     complain_at (location, _("redefining user token number of %s"),
 181                  symbol->tag);
 182
 183   symbol->user_token_number = user_token_number;
 184   /* User defined $end token? */
 185   if (user_token_number == 0)
 186     {
 187       endtoken = symbol;
 188       endtoken->number = 0;
 189       /* It is always mapped to 0, so it was already counted in
 190          NTOKENS.  */
 191       --ntokens;
 192     }
 193 }
 194
 195
 196 /*------------.
 197 | Free THIS.  |
 198 `------------*/
 199
 200 static void
 201 symbol_free (symbol_t *this)
 202 {
 203 #if 0
 204   /* This causes crashes because one string can appear more
 205      than once.  */
 206   XFREE (this->type_name);
 207 #endif
 208   XFREE (this->tag);
 209   XFREE (this);
 210 }
 211
 212
 213 /*-----------------------------------------------------------.
 214 | If THIS is not defined, report an error, and consider it a |
 215 | nonterminal.                                               |
 216 `-----------------------------------------------------------*/
 217
 218 static bool
 219 symbol_check_defined (symbol_t *this)
 220 {
 221   if (this->class == unknown_sym)
 222     {
 223       complain_at
 224         (this->location,
 225          _("symbol %s is used, but is not defined as a token and has no rules"),
 226          this->tag);
 227       this->class = nterm_sym;
 228       this->number = nvars++;
 229     }
 230
 231   return true;
 232 }
 233
 234
 235 /*-------------------------------------------------------------------.
 236 | Declare the new SYMBOL.  Make it an alias of SYMVAL, and type them |
 237 | with TYPENAME.                                                     |
 238 `-------------------------------------------------------------------*/
 239
 240 void
 241 symbol_make_alias (symbol_t *symbol, symbol_t *symval, location_t loc)
 242 {
 243   if (symval->alias)
 244     warn_at (loc, _("symbol `%s' used more than once as a literal string"),
 245           symval->tag);
 246   else if (symbol->alias)
 247     warn_at (loc, _("symbol `%s' given more than one literal string"),
 248           symbol->tag);
 249   else
 250     {
 251       symval->class = token_sym;
 252       symval->user_token_number = symbol->user_token_number;
 253       symbol->user_token_number = USER_NUMBER_ALIAS;
 254       symval->alias = symbol;
 255       symbol->alias = symval;
 256       /* symbol and symval combined are only one symbol */
 257       nsyms--;
 258       ntokens--;
 259       assert (ntokens == symbol->number || ntokens == symval->number);
 260       symbol->number = symval->number =
 261         (symval->number < symbol->number) ? symval->number : symbol->number;
 262     }
 263 }
 264
 265
 266 /*---------------------------------------------------------.
 267 | Check that THIS, and its alias, have same precedence and |
 268 | associativity.                                           |
 269 `---------------------------------------------------------*/
 270
 271 static bool
 272 symbol_check_alias_consistence (symbol_t *this)
 273 {
 274   /* Check only those who _are_ the aliases. */
 275   if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
 276     {
 277       if (this->prec != this->alias->prec)
 278         {
 279           if (this->prec != 0 && this->alias->prec != 0)
 280             complain_at (this->alias->location,
 281                          _("conflicting precedences for %s and %s"),
 282                          this->tag, this->alias->tag);
 283           if (this->prec != 0)
 284             this->alias->prec = this->prec;
 285           else
 286             this->prec = this->alias->prec;
 287         }
 288
 289       if (this->assoc != this->alias->assoc)
 290         {
 291           if (this->assoc != undef_assoc && this->alias->assoc != undef_assoc)
 292             complain_at (this->alias->location,
 293                          _("conflicting associativities for %s (%s) and %s (%s)"),
 294                          this->tag, assoc_to_string (this->assoc),
 295                          this->alias->tag, assoc_to_string (this->alias->assoc));
 296           if (this->assoc != undef_assoc)
 297             this->alias->assoc = this->assoc;
 298           else
 299             this->assoc = this->alias->assoc;
 300         }
 301     }
 302   return true;
 303 }
 304
 305
 306 /*-------------------------------------------------------------------.
 307 | Assign a symbol number, and write the definition of the token name |
 308 | into FDEFINES.  Put in SYMBOLS.                                    |
 309 `-------------------------------------------------------------------*/
 310
 311 static bool
 312 symbol_pack (symbol_t *this)
 313 {
 314   if (this->class == nterm_sym)
 315     {
 316       this->number += ntokens;
 317     }
 318   else if (this->alias)
 319     {
 320       /* This symbol and its alias are a single token defn.
 321          Allocate a tokno, and assign to both check agreement of
 322          prec and assoc fields and make both the same */
 323       if (this->number == NUMBER_UNDEFINED)
 324         {
 325           if (this == endtoken || this->alias == endtoken)
 326             this->number = this->alias->number = 0;
 327           else
 328             {
 329               assert (this->alias->number != NUMBER_UNDEFINED);
 330               this->number = this->alias->number;
 331             }
 332         }
 333       /* Do not do processing below for USER_NUMBER_ALIASs.  */
 334       if (this->user_token_number == USER_NUMBER_ALIAS)
 335         return true;
 336     }
 337   else /* this->class == token_sym */
 338     {
 339       assert (this->number != NUMBER_UNDEFINED);
 340     }
 341
 342   symbols[this->number] = this;
 343   return true;
 344 }
 345
 346
 347
 348
 349 /*--------------------------------------------------.
 350 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 351 `--------------------------------------------------*/
 352
 353 static bool
 354 symbol_translation (symbol_t *this)
 355 {
 356   /* Non-terminal? */
 357   if (this->class == token_sym
 358       && this->user_token_number != USER_NUMBER_ALIAS)
 359     {
 360       /* A token which translation has already been set? */
 361       if (token_translations[this->user_token_number] != undeftoken->number)
 362         complain_at (this->location,
 363                      _("tokens %s and %s both assigned number %d"),
 364                      symbols[token_translations[this->user_token_number]]->tag,
 365                      this->tag, this->user_token_number);
 366
 367       token_translations[this->user_token_number] = this->number;
 368     }
 369
 370   return true;
 371 }
 372
 373
 374 /*----------------------.
 375 | A symbol hash table.  |
 376 `----------------------*/
 377
 378 /* Initial capacity of symbols hash table.  */
 379 #define HT_INITIAL_CAPACITY 257
 380
 381 static struct hash_table *symbol_table = NULL;
 382
 383 static bool
 384 hash_compare_symbol_t (const symbol_t *m1, const symbol_t *m2)
 385 {
 386   return strcmp (m1->tag, m2->tag) == 0;
 387 }
 388
 389 static unsigned int
 390 hash_symbol_t (const symbol_t *m, unsigned int tablesize)
 391 {
 392   return hash_string (m->tag, tablesize);
 393 }
 394
 395
 396 /*-------------------------------.
 397 | Create the symbol hash table.  |
 398 `-------------------------------*/
 399
 400 void
 401 symbols_new (void)
 402 {
 403   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 404                                   NULL,
 405                                   (Hash_hasher) hash_symbol_t,
 406                                   (Hash_comparator) hash_compare_symbol_t,
 407                                   (Hash_data_freer) symbol_free);
 408 }
 409
 410
 411 /*----------------------------------------------------------------.
 412 | Find the symbol named KEY, and return it.  If it does not exist |
 413 | yet, create it.                                                 |
 414 `----------------------------------------------------------------*/
 415
 416 symbol_t *
 417 symbol_get (const char *key, location_t location)
 418 {
 419   symbol_t probe;
 420   symbol_t *entry;
 421
 422   /* Keep the symbol in a printable form.  */
 423   key = quotearg_style (escape_quoting_style, key);
 424   *(char const **) &probe.tag = key;
 425   entry = hash_lookup (symbol_table, &probe);
 426
 427   if (!entry)
 428     {
 429       /* First insertion in the hash. */
 430       entry = symbol_new (key, location);
 431       hash_insert (symbol_table, entry);
 432     }
 433   return entry;
 434 }
 435
 436
 437 /*------------------------------------------------------------------.
 438 | Generate a dummy nonterminal, whose name cannot conflict with the |
 439 | user's names.                                                     |
 440 `------------------------------------------------------------------*/
 441
 442 symbol_t *
 443 dummy_symbol_get (location_t location)
 444 {
 445   /* Incremented for each generated symbol.  */
 446   static int dummy_count = 0;
 447   static char buf[256];
 448
 449   symbol_t *sym;
 450
 451   sprintf (buf, "@%d", ++dummy_count);
 452   sym = symbol_get (buf, location);
 453   sym->class = nterm_sym;
 454   sym->number = nvars++;
 455   return sym;
 456 }
 457
 458
 459 /*-------------------.
 460 | Free the symbols.  |
 461 `-------------------*/
 462
 463 void
 464 symbols_free (void)
 465 {
 466   hash_free (symbol_table);
 467   free (symbols);
 468 }
 469
 470
 471 /*---------------------------------------------------------------.
 472 | Look for undefined symbols, report an error, and consider them |
 473 | terminals.                                                     |
 474 `---------------------------------------------------------------*/
 475
 476 void
 477 symbols_do (symbol_processor processor, void *processor_data)
 478 {
 479   hash_do_for_each (symbol_table,
 480                     (Hash_processor) processor,
 481                     processor_data);
 482 }
 483
 484
 485 /*--------------------------------------------------------------.
 486 | Check that all the symbols are defined.  Report any undefined |
 487 | symbols and consider them nonterminals.                       |
 488 `--------------------------------------------------------------*/
 489
 490 void
 491 symbols_check_defined (void)
 492 {
 493   symbols_do (symbol_check_defined, NULL);
 494 }
 495
 496 /*------------------------------------------------------------------.
 497 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 498 | number.                                                           |
 499 `------------------------------------------------------------------*/
 500
 501 static void
 502 symbols_token_translations_init (void)
 503 {
 504   bool num_256_available_p = true;
 505   int i;
 506
 507   /* Find the highest user token number, and whether 256, the POSIX
 508      preferred user token number for the error token, is used.  */
 509   max_user_token_number = 0;
 510   for (i = 0; i < ntokens; ++i)
 511     {
 512       symbol_t *this = symbols[i];
 513       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 514         {
 515           if (this->user_token_number > max_user_token_number)
 516             max_user_token_number = this->user_token_number;
 517           if (this->user_token_number == 256)
 518             num_256_available_p = false;
 519         }
 520     }
 521
 522   /* If 256 is not used, assign it to error, to follow POSIX.  */
 523   if (num_256_available_p
 524       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 525     errtoken->user_token_number = 256;
 526
 527   /* Set the missing user numbers. */
 528   if (max_user_token_number < 256)
 529     max_user_token_number = 256;
 530
 531   for (i = 0; i < ntokens; ++i)
 532     {
 533       symbol_t *this = symbols[i];
 534       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 535         this->user_token_number = ++max_user_token_number;
 536       if (this->user_token_number > max_user_token_number)
 537         max_user_token_number = this->user_token_number;
 538     }
 539
 540   token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
 541
 542   /* Initialize all entries for literal tokens to 2, the internal
 543      token number for $undefined, which represents all invalid inputs.
 544      */
 545   for (i = 0; i < max_user_token_number + 1; i++)
 546     token_translations[i] = undeftoken->number;
 547   symbols_do (symbol_translation, NULL);
 548 }
 549
 550
 551 /*----------------------------------------------------------------.
 552 | Assign symbol numbers, and write definition of token names into |
 553 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 554 `----------------------------------------------------------------*/
 555
 556 void
 557 symbols_pack (void)
 558 {
 559   symbols = XCALLOC (symbol_t *, nsyms);
 560
 561   symbols_do (symbol_check_alias_consistence, NULL);
 562   symbols_do (symbol_pack, NULL);
 563
 564   symbols_token_translations_init ();
 565
 566   if (startsymbol->class == unknown_sym)
 567     fatal_at (startsymbol_location,
 568               _("the start symbol %s is undefined"),
 569               startsymbol->tag);
 570   else if (startsymbol->class == token_sym)
 571     fatal_at (startsymbol_location,
 572               _("the start symbol %s is a token"),
 573               startsymbol->tag);
 574 }