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 PRINTER 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   free (this->tag);
 204   free (this);
 205 }
 206
 207
 208 /*-----------------------------------------------------------.
 209 | If THIS is not defined, report an error, and consider it a |
 210 | nonterminal.                                               |
 211 `-----------------------------------------------------------*/
 212
 213 static bool
 214 symbol_check_defined (symbol_t *this)
 215 {
 216   if (this->class == unknown_sym)
 217     {
 218       complain_at
 219         (this->location,
 220          _("symbol %s is used, but is not defined as a token and has no rules"),
 221          this->tag);
 222       this->class = nterm_sym;
 223       this->number = nvars++;
 224     }
 225
 226   return true;
 227 }
 228
 229
 230 /*-------------------------------------------------------------------.
 231 | Declare the new SYMBOL.  Make it an alias of SYMVAL, and type them |
 232 | with TYPENAME.                                                     |
 233 `-------------------------------------------------------------------*/
 234
 235 void
 236 symbol_make_alias (symbol_t *symbol, symbol_t *symval, location_t loc)
 237 {
 238   if (symval->alias)
 239     warn_at (loc, _("symbol `%s' used more than once as a literal string"),
 240           symval->tag);
 241   else if (symbol->alias)
 242     warn_at (loc, _("symbol `%s' given more than one literal string"),
 243           symbol->tag);
 244   else
 245     {
 246       symval->class = token_sym;
 247       symval->user_token_number = symbol->user_token_number;
 248       symbol->user_token_number = USER_NUMBER_ALIAS;
 249       symval->alias = symbol;
 250       symbol->alias = symval;
 251       /* symbol and symval combined are only one symbol */
 252       nsyms--;
 253       ntokens--;
 254       assert (ntokens == symbol->number || ntokens == symval->number);
 255       symbol->number = symval->number =
 256         (symval->number < symbol->number) ? symval->number : symbol->number;
 257     }
 258 }
 259
 260
 261 /*---------------------------------------------------------.
 262 | Check that THIS, and its alias, have same precedence and |
 263 | associativity.                                           |
 264 `---------------------------------------------------------*/
 265
 266 static bool
 267 symbol_check_alias_consistence (symbol_t *this)
 268 {
 269   /* Check only those who _are_ the aliases. */
 270   if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
 271     {
 272       if (this->prec != this->alias->prec)
 273         {
 274           if (this->prec != 0 && this->alias->prec != 0)
 275             complain_at (this->alias->location,
 276                          _("conflicting precedences for %s and %s"),
 277                          this->tag, this->alias->tag);
 278           if (this->prec != 0)
 279             this->alias->prec = this->prec;
 280           else
 281             this->prec = this->alias->prec;
 282         }
 283
 284       if (this->assoc != this->alias->assoc)
 285         {
 286           if (this->assoc != undef_assoc && this->alias->assoc != undef_assoc)
 287             complain_at (this->alias->location,
 288                          _("conflicting associativities for %s (%s) and %s (%s)"),
 289                          this->tag, assoc_to_string (this->assoc),
 290                          this->alias->tag, assoc_to_string (this->alias->assoc));
 291           if (this->assoc != undef_assoc)
 292             this->alias->assoc = this->assoc;
 293           else
 294             this->assoc = this->alias->assoc;
 295         }
 296     }
 297   return true;
 298 }
 299
 300
 301 /*-------------------------------------------------------------------.
 302 | Assign a symbol number, and write the definition of the token name |
 303 | into FDEFINES.  Put in SYMBOLS.                                    |
 304 `-------------------------------------------------------------------*/
 305
 306 static bool
 307 symbol_pack (symbol_t *this)
 308 {
 309   if (this->class == nterm_sym)
 310     {
 311       this->number += ntokens;
 312     }
 313   else if (this->alias)
 314     {
 315       /* This symbol and its alias are a single token defn.
 316          Allocate a tokno, and assign to both check agreement of
 317          prec and assoc fields and make both the same */
 318       if (this->number == NUMBER_UNDEFINED)
 319         {
 320           if (this == endtoken || this->alias == endtoken)
 321             this->number = this->alias->number = 0;
 322           else
 323             {
 324               assert (this->alias->number != NUMBER_UNDEFINED);
 325               this->number = this->alias->number;
 326             }
 327         }
 328       /* Do not do processing below for USER_NUMBER_ALIASes.  */
 329       if (this->user_token_number == USER_NUMBER_ALIAS)
 330         return true;
 331     }
 332   else /* this->class == token_sym */
 333     {
 334       assert (this->number != NUMBER_UNDEFINED);
 335     }
 336
 337   symbols[this->number] = this;
 338   return true;
 339 }
 340
 341
 342
 343
 344 /*--------------------------------------------------.
 345 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 346 `--------------------------------------------------*/
 347
 348 static bool
 349 symbol_translation (symbol_t *this)
 350 {
 351   /* Non-terminal? */
 352   if (this->class == token_sym
 353       && this->user_token_number != USER_NUMBER_ALIAS)
 354     {
 355       /* A token which translation has already been set? */
 356       if (token_translations[this->user_token_number] != undeftoken->number)
 357         complain_at (this->location,
 358                      _("tokens %s and %s both assigned number %d"),
 359                      symbols[token_translations[this->user_token_number]]->tag,
 360                      this->tag, this->user_token_number);
 361
 362       token_translations[this->user_token_number] = this->number;
 363     }
 364
 365   return true;
 366 }
 367
 368
 369 /*----------------------.
 370 | A symbol hash table.  |
 371 `----------------------*/
 372
 373 /* Initial capacity of symbols hash table.  */
 374 #define HT_INITIAL_CAPACITY 257
 375
 376 static struct hash_table *symbol_table = NULL;
 377
 378 static bool
 379 hash_compare_symbol_t (const symbol_t *m1, const symbol_t *m2)
 380 {
 381   return strcmp (m1->tag, m2->tag) == 0;
 382 }
 383
 384 static unsigned int
 385 hash_symbol_t (const symbol_t *m, unsigned int tablesize)
 386 {
 387   return hash_string (m->tag, tablesize);
 388 }
 389
 390
 391 /*-------------------------------.
 392 | Create the symbol hash table.  |
 393 `-------------------------------*/
 394
 395 void
 396 symbols_new (void)
 397 {
 398   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 399                                   NULL,
 400                                   (Hash_hasher) hash_symbol_t,
 401                                   (Hash_comparator) hash_compare_symbol_t,
 402                                   (Hash_data_freer) symbol_free);
 403 }
 404
 405
 406 /*----------------------------------------------------------------.
 407 | Find the symbol named KEY, and return it.  If it does not exist |
 408 | yet, create it.                                                 |
 409 `----------------------------------------------------------------*/
 410
 411 symbol_t *
 412 symbol_get (const char *key, location_t location)
 413 {
 414   symbol_t probe;
 415   symbol_t *entry;
 416
 417   /* Keep the symbol in a printable form.  */
 418   key = quotearg_style (escape_quoting_style, key);
 419   *(char const **) &probe.tag = key;
 420   entry = hash_lookup (symbol_table, &probe);
 421
 422   if (!entry)
 423     {
 424       /* First insertion in the hash. */
 425       entry = symbol_new (key, location);
 426       hash_insert (symbol_table, entry);
 427     }
 428   return entry;
 429 }
 430
 431
 432 /*------------------------------------------------------------------.
 433 | Generate a dummy nonterminal, whose name cannot conflict with the |
 434 | user's names.                                                     |
 435 `------------------------------------------------------------------*/
 436
 437 symbol_t *
 438 dummy_symbol_get (location_t location)
 439 {
 440   /* Incremented for each generated symbol.  */
 441   static int dummy_count = 0;
 442   static char buf[256];
 443
 444   symbol_t *sym;
 445
 446   sprintf (buf, "@%d", ++dummy_count);
 447   sym = symbol_get (buf, location);
 448   sym->class = nterm_sym;
 449   sym->number = nvars++;
 450   return sym;
 451 }
 452
 453
 454 /*-------------------.
 455 | Free the symbols.  |
 456 `-------------------*/
 457
 458 void
 459 symbols_free (void)
 460 {
 461   hash_free (symbol_table);
 462   free (symbols);
 463 }
 464
 465
 466 /*---------------------------------------------------------------.
 467 | Look for undefined symbols, report an error, and consider them |
 468 | terminals.                                                     |
 469 `---------------------------------------------------------------*/
 470
 471 void
 472 symbols_do (symbol_processor processor, void *processor_data)
 473 {
 474   hash_do_for_each (symbol_table,
 475                     (Hash_processor) processor,
 476                     processor_data);
 477 }
 478
 479
 480 /*--------------------------------------------------------------.
 481 | Check that all the symbols are defined.  Report any undefined |
 482 | symbols and consider them nonterminals.                       |
 483 `--------------------------------------------------------------*/
 484
 485 void
 486 symbols_check_defined (void)
 487 {
 488   symbols_do (symbol_check_defined, NULL);
 489 }
 490
 491 /*------------------------------------------------------------------.
 492 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 493 | number.                                                           |
 494 `------------------------------------------------------------------*/
 495
 496 static void
 497 symbols_token_translations_init (void)
 498 {
 499   bool num_256_available_p = true;
 500   int i;
 501
 502   /* Find the highest user token number, and whether 256, the POSIX
 503      preferred user token number for the error token, is used.  */
 504   max_user_token_number = 0;
 505   for (i = 0; i < ntokens; ++i)
 506     {
 507       symbol_t *this = symbols[i];
 508       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 509         {
 510           if (this->user_token_number > max_user_token_number)
 511             max_user_token_number = this->user_token_number;
 512           if (this->user_token_number == 256)
 513             num_256_available_p = false;
 514         }
 515     }
 516
 517   /* If 256 is not used, assign it to error, to follow POSIX.  */
 518   if (num_256_available_p
 519       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 520     errtoken->user_token_number = 256;
 521
 522   /* Set the missing user numbers. */
 523   if (max_user_token_number < 256)
 524     max_user_token_number = 256;
 525
 526   for (i = 0; i < ntokens; ++i)
 527     {
 528       symbol_t *this = symbols[i];
 529       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 530         this->user_token_number = ++max_user_token_number;
 531       if (this->user_token_number > max_user_token_number)
 532         max_user_token_number = this->user_token_number;
 533     }
 534
 535   token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
 536
 537   /* Initialize all entries for literal tokens to 2, the internal
 538      token number for $undefined, which represents all invalid inputs.
 539      */
 540   for (i = 0; i < max_user_token_number + 1; i++)
 541     token_translations[i] = undeftoken->number;
 542   symbols_do (symbol_translation, NULL);
 543 }
 544
 545
 546 /*----------------------------------------------------------------.
 547 | Assign symbol numbers, and write definition of token names into |
 548 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 549 `----------------------------------------------------------------*/
 550
 551 void
 552 symbols_pack (void)
 553 {
 554   symbols = XCALLOC (symbol_t *, nsyms);
 555
 556   symbols_do (symbol_check_alias_consistence, NULL);
 557   symbols_do (symbol_pack, NULL);
 558
 559   symbols_token_translations_init ();
 560
 561   if (startsymbol->class == unknown_sym)
 562     fatal_at (startsymbol_location,
 563               _("the start symbol %s is undefined"),
 564               startsymbol->tag);
 565   else if (startsymbol->class == token_sym)
 566     fatal_at (startsymbol_location,
 567               _("the start symbol %s is a token"),
 568               startsymbol->tag);
 569 }