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 "hash.h"
  24 #include "complain.h"
  25 #include "symtab.h"
  26 #include "gram.h"
  27
  28 /*------------------------.
  29 | Distinguished symbols.  |
  30 `------------------------*/
  31
  32 symbol_t *errtoken = NULL;
  33 symbol_t *undeftoken = NULL;
  34 symbol_t *eoftoken = NULL;
  35 symbol_t *axiom = NULL;
  36 symbol_t *startsymbol = NULL;
  37
  38 /*---------------------------------.
  39 | Create a new symbol, named TAG.  |
  40 `---------------------------------*/
  41
  42 static symbol_t *
  43 symbol_new (const char *tag)
  44 {
  45   symbol_t *res = XMALLOC (symbol_t, 1);
  46
  47   res->tag = xstrdup (tag);
  48   res->type_name = NULL;
  49   res->number = NUMBER_UNDEFINED;
  50   res->prec = 0;
  51   res->assoc = right_assoc;
  52   res->user_token_number = USER_NUMBER_UNDEFINED;
  53   res->alias = NULL;
  54   res->class = unknown_sym;
  55
  56   nsyms++;
  57   return res;
  58 }
  59
  60
  61 /*------------------------------------------------------------------.
  62 | Set the TYPE_NAME associated to SYMBOL.  Does nothing if passed 0 |
  63 | as TYPE_NAME.                                                     |
  64 `------------------------------------------------------------------*/
  65
  66 void
  67 symbol_type_set (symbol_t *symbol, char *type_name)
  68 {
  69   if (type_name)
  70     {
  71       if (symbol->type_name)
  72         complain (_("type redeclaration for %s"), symbol->tag);
  73       symbol->type_name = type_name;
  74     }
  75 }
  76
  77
  78 /*------------------------------------------------------------------.
  79 | Set the PRECEDENCE associated to SYMBOL.  Does nothing if invoked |
  80 | with UNDEF_ASSOC as ASSOC.                                        |
  81 `------------------------------------------------------------------*/
  82
  83 void
  84 symbol_precedence_set (symbol_t *symbol,
  85                        int prec, associativity assoc)
  86 {
  87   if (assoc != undef_assoc)
  88     {
  89       if (symbol->prec != 0)
  90         complain (_("redefining precedence of %s"), symbol->tag);
  91       symbol->prec = prec;
  92       symbol->assoc = assoc;
  93     }
  94
  95   /* Only terminals have a precedence. */
  96   symbol_class_set (symbol, token_sym);
  97 }
  98
  99
 100 /*-------------------------------------.
 101 | Set the CLASS associated to SYMBOL.  |
 102 `-------------------------------------*/
 103
 104 void
 105 symbol_class_set (symbol_t *symbol, symbol_class class)
 106 {
 107   if (symbol->class != unknown_sym && symbol->class != class)
 108     complain (_("symbol %s redefined"), symbol->tag);
 109
 110   if (class == nterm_sym && symbol->class != nterm_sym)
 111     symbol->number = nvars++;
 112   else if (class == token_sym && symbol->number == NUMBER_UNDEFINED)
 113     symbol->number = ntokens++;
 114
 115   symbol->class = class;
 116 }
 117
 118
 119 /*-------------------------------------------------.
 120 | Set the USER_TOKEN_NUMBER associated to SYMBOL.  |
 121 `-------------------------------------------------*/
 122
 123 void
 124 symbol_user_token_number_set (symbol_t *symbol, int user_token_number)
 125 {
 126   assert (symbol->class == token_sym);
 127
 128   if (symbol->user_token_number != USER_NUMBER_UNDEFINED
 129       && symbol->user_token_number != user_token_number)
 130     complain (_("redefining user token number of %s"), symbol->tag);
 131
 132   symbol->user_token_number = user_token_number;
 133   /* User defined EOF token? */
 134   if (user_token_number == 0)
 135     {
 136       eoftoken = symbol;
 137       eoftoken->number = 0;
 138       /* It is always mapped to 0, so it was already counted in
 139          NTOKENS.  */
 140       --ntokens;
 141     }
 142 }
 143
 144
 145 /*------------.
 146 | Free THIS.  |
 147 `------------*/
 148
 149 static void
 150 symbol_free (symbol_t *this)
 151 {
 152 #if 0
 153   /* This causes crashes because one string can appear more
 154      than once.  */
 155   XFREE (this->type_name);
 156 #endif
 157   XFREE (this->tag);
 158   XFREE (this);
 159 }
 160
 161
 162 /*-----------------------------------------------------------.
 163 | If THIS is not defined, report an error, and consider it a |
 164 | nonterminal.                                               |
 165 `-----------------------------------------------------------*/
 166
 167 static bool
 168 symbol_check_defined (symbol_t *this)
 169 {
 170   if (this->class == unknown_sym)
 171     {
 172       complain
 173         (_("symbol %s is used, but is not defined as a token and has no rules"),
 174          this->tag);
 175       this->class = nterm_sym;
 176       this->number = nvars++;
 177     }
 178
 179   return TRUE;
 180 }
 181
 182
 183 /*-------------------------------------------------------------------.
 184 | Declare the new SYMBOL.  Make it an alias of SYMVAL, and type them |
 185 | with TYPENAME.                                                     |
 186 `-------------------------------------------------------------------*/
 187
 188 void
 189 symbol_make_alias (symbol_t *symbol, symbol_t *symval)
 190 {
 191   if (symval->alias)
 192     warn (_("symbol `%s' used more than once as a literal string"),
 193           symval->tag);
 194   else if (symbol->alias)
 195     warn (_("symbol `%s' given more than one literal string"),
 196           symbol->tag);
 197   else
 198     {
 199       symval->class = token_sym;
 200       symval->user_token_number = symbol->user_token_number;
 201       symbol->user_token_number = USER_NUMBER_ALIAS;
 202       symval->alias = symbol;
 203       symbol->alias = symval;
 204       /* symbol and symval combined are only one symbol */
 205       nsyms--;
 206       ntokens--;
 207       assert (ntokens == symbol->number || ntokens == symval->number);
 208       symbol->number = symval->number =
 209         (symval->number < symbol->number) ? symval->number : symbol->number;
 210     }
 211 }
 212
 213
 214 /*---------------------------------------------------------.
 215 | Check that THIS, and its alias, have same precedence and |
 216 | associativity.                                           |
 217 `---------------------------------------------------------*/
 218
 219 static bool
 220 symbol_check_alias_consistence (symbol_t *this)
 221 {
 222   /* Check only those who _are_ the aliases. */
 223   if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
 224     {
 225       if (this->prec != this->alias->prec)
 226         {
 227           if (this->prec != 0 && this->alias->prec != 0)
 228             complain (_("conflicting precedences for %s and %s"),
 229                       this->tag, this->alias->tag);
 230           if (this->prec != 0)
 231             this->alias->prec = this->prec;
 232           else
 233             this->prec = this->alias->prec;
 234         }
 235
 236       if (this->assoc != this->alias->assoc)
 237         {
 238           /* FIXME: For some reason (probably the S/R => keep the S),
 239              the right assoc is chosen has the ``not set''.  This is
 240              not nice, fix this!  */
 241           if (this->assoc != right_assoc
 242               && this->alias->assoc != right_assoc)
 243             complain (_("conflicting associativities for %s and %s"),
 244                       this->tag, this->alias->tag);
 245           if (this->assoc != 0)
 246             this->alias->assoc = this->assoc;
 247           else
 248             this->assoc = this->alias->assoc;
 249         }
 250     }
 251   return TRUE;
 252 }
 253
 254
 255 /*-------------------------------------------------------------------.
 256 | Assign a symbol number, and write the definition of the token name |
 257 | into FDEFINES.  Put in SYMBOLS.                                    |
 258 `-------------------------------------------------------------------*/
 259
 260 static bool
 261 symbol_pack (symbol_t *this)
 262 {
 263   if (this->class == nterm_sym)
 264     {
 265       this->number += ntokens;
 266     }
 267   else if (this->alias)
 268     {
 269       /* This symbol and its alias are a single token defn.
 270          Allocate a tokno, and assign to both check agreement of
 271          prec and assoc fields and make both the same */
 272       if (this->number == NUMBER_UNDEFINED)
 273         {
 274           if (this == eoftoken || this->alias == eoftoken)
 275             this->number = this->alias->number = 0;
 276           else
 277             {
 278               assert (this->alias->number != NUMBER_UNDEFINED);
 279               this->number = this->alias->number;
 280             }
 281         }
 282       /* Do not do processing below for USER_NUMBER_ALIASs.  */
 283       if (this->user_token_number == USER_NUMBER_ALIAS)
 284         return TRUE;
 285     }
 286   else /* this->class == token_sym */
 287     {
 288       assert (this->number != NUMBER_UNDEFINED);
 289     }
 290
 291   symbols[this->number] = this;
 292   return TRUE;
 293 }
 294
 295
 296
 297
 298 /*--------------------------------------------------.
 299 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 300 `--------------------------------------------------*/
 301
 302 static bool
 303 symbol_translation (symbol_t *this)
 304 {
 305   /* Non-terminal? */
 306   if (this->class == token_sym
 307       && this->user_token_number != USER_NUMBER_ALIAS)
 308     {
 309       /* A token which translation has already been set? */
 310       if (token_translations[this->user_token_number] != undeftoken->number)
 311         complain (_("tokens %s and %s both assigned number %d"),
 312                   symbols[token_translations[this->user_token_number]]->tag,
 313                   this->tag, this->user_token_number);
 314
 315       token_translations[this->user_token_number] = this->number;
 316     }
 317
 318   return TRUE;
 319 }
 320
 321
 322 /*----------------------.
 323 | A symbol hash table.  |
 324 `----------------------*/
 325
 326 /* Initial capacity of symbols hash table.  */
 327 #define HT_INITIAL_CAPACITY 257
 328
 329 static struct hash_table *symbol_table = NULL;
 330
 331 static bool
 332 hash_compare_symbol_t (const symbol_t *m1, const symbol_t *m2)
 333 {
 334   return strcmp (m1->tag, m2->tag) ? FALSE : TRUE;
 335 }
 336
 337 static unsigned int
 338 hash_symbol_t (const symbol_t *m, unsigned int tablesize)
 339 {
 340   return hash_string (m->tag, tablesize);
 341 }
 342
 343
 344 /*-------------------------------.
 345 | Create the symbol hash table.  |
 346 `-------------------------------*/
 347
 348 void
 349 symbols_new (void)
 350 {
 351   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 352                                   NULL,
 353                                   (Hash_hasher) hash_symbol_t,
 354                                   (Hash_comparator) hash_compare_symbol_t,
 355                                   (Hash_data_freer) symbol_free);
 356 }
 357
 358
 359 /*----------------------------------------------------------------.
 360 | Find the symbol named KEY, and return it.  If it does not exist |
 361 | yet, create it.                                                 |
 362 `----------------------------------------------------------------*/
 363
 364 symbol_t *
 365 getsym (const char *key)
 366 {
 367   symbol_t probe;
 368   symbol_t *entry;
 369
 370   (const char *) probe.tag = key;
 371   entry = hash_lookup (symbol_table, &probe);
 372
 373   if (!entry)
 374     {
 375       /* First insertion in the hash. */
 376       entry = symbol_new (key);
 377       hash_insert (symbol_table, entry);
 378     }
 379   return entry;
 380 }
 381
 382
 383 /*-------------------.
 384 | Free the symbols.  |
 385 `-------------------*/
 386
 387 void
 388 symbols_free (void)
 389 {
 390   hash_free (symbol_table);
 391 }
 392
 393
 394 /*---------------------------------------------------------------.
 395 | Look for undefined symbols, report an error, and consider them |
 396 | terminals.                                                     |
 397 `---------------------------------------------------------------*/
 398
 399 void
 400 symbols_do (symbol_processor processor, void *processor_data)
 401 {
 402   hash_do_for_each (symbol_table,
 403                     (Hash_processor) processor,
 404                     processor_data);
 405 }
 406
 407
 408 /*--------------------------------------------------------------.
 409 | Check that all the symbols are defined.  Report any undefined |
 410 | symbols and consider them nonterminals.                       |
 411 `--------------------------------------------------------------*/
 412
 413 void
 414 symbols_check_defined (void)
 415 {
 416   symbols_do (symbol_check_defined, NULL);
 417 }
 418
 419 /*------------------------------------------------------------------.
 420 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 421 | number.                                                           |
 422 `------------------------------------------------------------------*/
 423
 424 static void
 425 symbols_token_translations_init (void)
 426 {
 427   int num_256_available_p = TRUE;
 428   int i;
 429
 430   /* Find the highest user token number, and whether 256, the POSIX
 431      preferred user token number for the error token, is used.  */
 432   max_user_token_number = 0;
 433   for (i = 0; i < ntokens; ++i)
 434     {
 435       symbol_t *this = symbols[i];
 436       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 437         {
 438           if (this->user_token_number > max_user_token_number)
 439             max_user_token_number = this->user_token_number;
 440           if (this->user_token_number == 256)
 441             num_256_available_p = FALSE;
 442         }
 443     }
 444
 445   /* If 256 is not used, assign it to error, to follow POSIX.  */
 446   if (num_256_available_p
 447       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 448     errtoken->user_token_number = 256;
 449
 450   /* Set the missing user numbers. */
 451   if (max_user_token_number < 256)
 452     max_user_token_number = 256;
 453
 454   for (i = 0; i < ntokens; ++i)
 455     {
 456       symbol_t *this = symbols[i];
 457       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 458         this->user_token_number = ++max_user_token_number;
 459       if (this->user_token_number > max_user_token_number)
 460         max_user_token_number = this->user_token_number;
 461     }
 462
 463   token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
 464
 465   /* Initialize all entries for literal tokens to 2, the internal
 466      token number for $undefined., which represents all invalid
 467      inputs.  */
 468   for (i = 0; i < max_user_token_number + 1; i++)
 469     token_translations[i] = undeftoken->number;
 470   symbols_do (symbol_translation, NULL);
 471 }
 472
 473
 474 /*----------------------------------------------------------------.
 475 | Assign symbol numbers, and write definition of token names into |
 476 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 477 `----------------------------------------------------------------*/
 478
 479 void
 480 symbols_pack (void)
 481 {
 482   symbols = XCALLOC (symbol_t *, nsyms);
 483
 484   symbols_do (symbol_check_alias_consistence, NULL);
 485   symbols_do (symbol_pack, NULL);
 486
 487   symbols_token_translations_init ();
 488
 489   if (startsymbol->class == unknown_sym)
 490     fatal (_("the start symbol %s is undefined"), startsymbol->tag);
 491   else if (startsymbol->class == token_sym)
 492     fatal (_("the start symbol %s is a token"), startsymbol->tag);
 493 }