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 *eoftoken = NULL;
  36 symbol_t *axiom = 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
  55   res->number = NUMBER_UNDEFINED;
  56   res->prec = 0;
  57   res->assoc = right_assoc;
  58   res->user_token_number = USER_NUMBER_UNDEFINED;
  59   res->alias = NULL;
  60   res->class = unknown_sym;
  61
  62   nsyms++;
  63   return res;
  64 }
  65
  66
  67 /*-----------------------------------------------------------------.
  68 | Return the tag of this SYMBOL in a printable form.  Warning: use |
  69 | the first QUOTEARG slot: 0.                                      |
  70 `-----------------------------------------------------------------*/
  71
  72 const char *
  73 symbol_tag_get (symbol_t *symbol)
  74 {
  75   return quotearg_style (escape_quoting_style, symbol->tag);
  76 }
  77
  78
  79 /*------------------------------------------------------------.
  80 | Return the tag of this SYMBOL in a printable form.  Use the |
  81 | QUOTEARG slot number N.                                     |
  82 `------------------------------------------------------------*/
  83
  84 const char *
  85 symbol_tag_get_n (symbol_t *symbol, int n)
  86 {
  87   return quotearg_n_style (n, escape_quoting_style, symbol->tag);
  88 }
  89
  90
  91 /*-------------------------------.
  92 | Print the tag of this SYMBOL.  |
  93 `-------------------------------*/
  94
  95 void
  96 symbol_tag_print (symbol_t *symbol, FILE *out)
  97 {
  98   fputs (symbol_tag_get (symbol), out);
  99 }
 100
 101
 102 /*------------------------------------------------------------------.
 103 | Set the TYPE_NAME associated to SYMBOL.  Does nothing if passed 0 |
 104 | as TYPE_NAME.                                                     |
 105 `------------------------------------------------------------------*/
 106
 107 void
 108 symbol_type_set (symbol_t *symbol, location_t location, char *type_name)
 109 {
 110   if (type_name)
 111     {
 112       if (symbol->type_name)
 113         complain_at (location,
 114                      _("type redeclaration for %s"), symbol_tag_get (symbol));
 115       symbol->type_name = type_name;
 116     }
 117 }
 118
 119
 120 /*-------------------------------------------------------------------.
 121 | Set the DESTRUCTOR associated to SYMBOL.  Do nothing if passed 0.  |
 122 `-------------------------------------------------------------------*/
 123
 124 void
 125 symbol_destructor_set (symbol_t *symbol, location_t location, char *destructor)
 126 {
 127   if (destructor)
 128     {
 129       if (symbol->destructor)
 130         complain_at (location,
 131                      _("%s redeclaration for %s"),
 132                      "%destructor", symbol_tag_get (symbol));
 133       symbol->destructor = destructor;
 134       symbol->destructor_location = location;
 135     }
 136 }
 137
 138
 139 /*----------------------------------------------------------------.
 140 | Set the PRITNER associated to SYMBOL.  Do nothing if passed 0.  |
 141 `----------------------------------------------------------------*/
 142
 143 void
 144 symbol_printer_set (symbol_t *symbol, char *printer, location_t location)
 145 {
 146   if (printer)
 147     {
 148       if (symbol->printer)
 149         complain_at (location,
 150                      _("%s redeclaration for %s"),
 151                      "%printer", symbol_tag_get (symbol));
 152       symbol->printer = printer;
 153       symbol->printer_location = location;
 154     }
 155 }
 156
 157
 158 /*------------------------------------------------------------------.
 159 | Set the PRECEDENCE associated to SYMBOL.  Does nothing if invoked |
 160 | with UNDEF_ASSOC as ASSOC.                                        |
 161 `------------------------------------------------------------------*/
 162
 163 void
 164 symbol_precedence_set (symbol_t *symbol, location_t location,
 165                        int prec, associativity assoc)
 166 {
 167   if (assoc != undef_assoc)
 168     {
 169       if (symbol->prec != 0)
 170         complain_at (location,
 171                      _("redefining precedence of %s"),
 172                      symbol_tag_get (symbol));
 173       symbol->prec = prec;
 174       symbol->assoc = assoc;
 175     }
 176
 177   /* Only terminals have a precedence. */
 178   symbol_class_set (symbol, token_sym);
 179 }
 180
 181
 182 /*-------------------------------------.
 183 | Set the CLASS associated to SYMBOL.  |
 184 `-------------------------------------*/
 185
 186 void
 187 symbol_class_set (symbol_t *symbol, symbol_class class)
 188 {
 189   if (symbol->class != unknown_sym && symbol->class != class)
 190     complain (_("symbol %s redefined"), symbol_tag_get (symbol));
 191
 192   if (class == nterm_sym && symbol->class != nterm_sym)
 193     symbol->number = nvars++;
 194   else if (class == token_sym && symbol->number == NUMBER_UNDEFINED)
 195     symbol->number = ntokens++;
 196
 197   symbol->class = class;
 198 }
 199
 200
 201 /*-------------------------------------------------.
 202 | Set the USER_TOKEN_NUMBER associated to SYMBOL.  |
 203 `-------------------------------------------------*/
 204
 205 void
 206 symbol_user_token_number_set (symbol_t *symbol, int user_token_number)
 207 {
 208   assert (symbol->class == token_sym);
 209
 210   if (symbol->user_token_number != USER_NUMBER_UNDEFINED
 211       && symbol->user_token_number != user_token_number)
 212     complain (_("redefining user token number of %s"),
 213               symbol_tag_get (symbol));
 214
 215   symbol->user_token_number = user_token_number;
 216   /* User defined EOF token? */
 217   if (user_token_number == 0)
 218     {
 219       eoftoken = symbol;
 220       eoftoken->number = 0;
 221       /* It is always mapped to 0, so it was already counted in
 222          NTOKENS.  */
 223       --ntokens;
 224     }
 225 }
 226
 227
 228 /*------------.
 229 | Free THIS.  |
 230 `------------*/
 231
 232 static void
 233 symbol_free (symbol_t *this)
 234 {
 235 #if 0
 236   /* This causes crashes because one string can appear more
 237      than once.  */
 238   XFREE (this->type_name);
 239 #endif
 240   XFREE (this->tag);
 241   XFREE (this);
 242 }
 243
 244
 245 /*-----------------------------------------------------------.
 246 | If THIS is not defined, report an error, and consider it a |
 247 | nonterminal.                                               |
 248 `-----------------------------------------------------------*/
 249
 250 static bool
 251 symbol_check_defined (symbol_t *this)
 252 {
 253   if (this->class == unknown_sym)
 254     {
 255       complain_at
 256         (this->location,
 257          _("symbol %s is used, but is not defined as a token and has no rules"),
 258          symbol_tag_get (this));
 259       this->class = nterm_sym;
 260       this->number = nvars++;
 261     }
 262
 263   return TRUE;
 264 }
 265
 266
 267 /*-------------------------------------------------------------------.
 268 | Declare the new SYMBOL.  Make it an alias of SYMVAL, and type them |
 269 | with TYPENAME.                                                     |
 270 `-------------------------------------------------------------------*/
 271
 272 void
 273 symbol_make_alias (symbol_t *symbol, symbol_t *symval)
 274 {
 275   if (symval->alias)
 276     warn (_("symbol `%s' used more than once as a literal string"),
 277           symbol_tag_get (symval));
 278   else if (symbol->alias)
 279     warn (_("symbol `%s' given more than one literal string"),
 280           symbol_tag_get (symbol));
 281   else
 282     {
 283       symval->class = token_sym;
 284       symval->user_token_number = symbol->user_token_number;
 285       symbol->user_token_number = USER_NUMBER_ALIAS;
 286       symval->alias = symbol;
 287       symbol->alias = symval;
 288       /* symbol and symval combined are only one symbol */
 289       nsyms--;
 290       ntokens--;
 291       assert (ntokens == symbol->number || ntokens == symval->number);
 292       symbol->number = symval->number =
 293         (symval->number < symbol->number) ? symval->number : symbol->number;
 294     }
 295 }
 296
 297
 298 /*---------------------------------------------------------.
 299 | Check that THIS, and its alias, have same precedence and |
 300 | associativity.                                           |
 301 `---------------------------------------------------------*/
 302
 303 static bool
 304 symbol_check_alias_consistence (symbol_t *this)
 305 {
 306   /* Check only those who _are_ the aliases. */
 307   if (this->alias && this->user_token_number == USER_NUMBER_ALIAS)
 308     {
 309       if (this->prec != this->alias->prec)
 310         {
 311           if (this->prec != 0 && this->alias->prec != 0)
 312             complain (_("conflicting precedences for %s and %s"),
 313                       symbol_tag_get (this), symbol_tag_get (this->alias));
 314           if (this->prec != 0)
 315             this->alias->prec = this->prec;
 316           else
 317             this->prec = this->alias->prec;
 318         }
 319
 320       if (this->assoc != this->alias->assoc)
 321         {
 322           /* FIXME: For some reason (probably the S/R => keep the S),
 323              the right assoc is chosen has the ``not set''.  This is
 324              not nice, fix this!  */
 325           if (this->assoc != right_assoc
 326               && this->alias->assoc != right_assoc)
 327             complain (_("conflicting associativities for %s and %s"),
 328                       symbol_tag_get (this), symbol_tag_get (this->alias));
 329           if (this->assoc != 0)
 330             this->alias->assoc = this->assoc;
 331           else
 332             this->assoc = this->alias->assoc;
 333         }
 334     }
 335   return TRUE;
 336 }
 337
 338
 339 /*-------------------------------------------------------------------.
 340 | Assign a symbol number, and write the definition of the token name |
 341 | into FDEFINES.  Put in SYMBOLS.                                    |
 342 `-------------------------------------------------------------------*/
 343
 344 static bool
 345 symbol_pack (symbol_t *this)
 346 {
 347   if (this->class == nterm_sym)
 348     {
 349       this->number += ntokens;
 350     }
 351   else if (this->alias)
 352     {
 353       /* This symbol and its alias are a single token defn.
 354          Allocate a tokno, and assign to both check agreement of
 355          prec and assoc fields and make both the same */
 356       if (this->number == NUMBER_UNDEFINED)
 357         {
 358           if (this == eoftoken || this->alias == eoftoken)
 359             this->number = this->alias->number = 0;
 360           else
 361             {
 362               assert (this->alias->number != NUMBER_UNDEFINED);
 363               this->number = this->alias->number;
 364             }
 365         }
 366       /* Do not do processing below for USER_NUMBER_ALIASs.  */
 367       if (this->user_token_number == USER_NUMBER_ALIAS)
 368         return TRUE;
 369     }
 370   else /* this->class == token_sym */
 371     {
 372       assert (this->number != NUMBER_UNDEFINED);
 373     }
 374
 375   symbols[this->number] = this;
 376   return TRUE;
 377 }
 378
 379
 380
 381
 382 /*--------------------------------------------------.
 383 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
 384 `--------------------------------------------------*/
 385
 386 static bool
 387 symbol_translation (symbol_t *this)
 388 {
 389   /* Non-terminal? */
 390   if (this->class == token_sym
 391       && this->user_token_number != USER_NUMBER_ALIAS)
 392     {
 393       /* A token which translation has already been set? */
 394       if (token_translations[this->user_token_number] != undeftoken->number)
 395         complain (_("tokens %s and %s both assigned number %d"),
 396                   symbol_tag_get (symbols[token_translations[this->user_token_number]]),
 397                   symbol_tag_get (this), this->user_token_number);
 398
 399       token_translations[this->user_token_number] = this->number;
 400     }
 401
 402   return TRUE;
 403 }
 404
 405
 406 /*----------------------.
 407 | A symbol hash table.  |
 408 `----------------------*/
 409
 410 /* Initial capacity of symbols hash table.  */
 411 #define HT_INITIAL_CAPACITY 257
 412
 413 static struct hash_table *symbol_table = NULL;
 414
 415 static bool
 416 hash_compare_symbol_t (const symbol_t *m1, const symbol_t *m2)
 417 {
 418   return strcmp (m1->tag, m2->tag) ? FALSE : TRUE;
 419 }
 420
 421 static unsigned int
 422 hash_symbol_t (const symbol_t *m, unsigned int tablesize)
 423 {
 424   return hash_string (m->tag, tablesize);
 425 }
 426
 427
 428 /*-------------------------------.
 429 | Create the symbol hash table.  |
 430 `-------------------------------*/
 431
 432 void
 433 symbols_new (void)
 434 {
 435   symbol_table = hash_initialize (HT_INITIAL_CAPACITY,
 436                                   NULL,
 437                                   (Hash_hasher) hash_symbol_t,
 438                                   (Hash_comparator) hash_compare_symbol_t,
 439                                   (Hash_data_freer) symbol_free);
 440 }
 441
 442
 443 /*----------------------------------------------------------------.
 444 | Find the symbol named KEY, and return it.  If it does not exist |
 445 | yet, create it.                                                 |
 446 `----------------------------------------------------------------*/
 447
 448 symbol_t *
 449 getsym (const char *key, location_t location)
 450 {
 451   symbol_t probe;
 452   symbol_t *entry;
 453
 454   (const char *) probe.tag = key;
 455   entry = hash_lookup (symbol_table, &probe);
 456
 457   if (!entry)
 458     {
 459       /* First insertion in the hash. */
 460       entry = symbol_new (key, location);
 461       hash_insert (symbol_table, entry);
 462     }
 463   return entry;
 464 }
 465
 466
 467 /*-------------------.
 468 | Free the symbols.  |
 469 `-------------------*/
 470
 471 void
 472 symbols_free (void)
 473 {
 474   hash_free (symbol_table);
 475 }
 476
 477
 478 /*---------------------------------------------------------------.
 479 | Look for undefined symbols, report an error, and consider them |
 480 | terminals.                                                     |
 481 `---------------------------------------------------------------*/
 482
 483 void
 484 symbols_do (symbol_processor processor, void *processor_data)
 485 {
 486   hash_do_for_each (symbol_table,
 487                     (Hash_processor) processor,
 488                     processor_data);
 489 }
 490
 491
 492 /*--------------------------------------------------------------.
 493 | Check that all the symbols are defined.  Report any undefined |
 494 | symbols and consider them nonterminals.                       |
 495 `--------------------------------------------------------------*/
 496
 497 void
 498 symbols_check_defined (void)
 499 {
 500   symbols_do (symbol_check_defined, NULL);
 501 }
 502
 503 /*------------------------------------------------------------------.
 504 | Set TOKEN_TRANSLATIONS.  Check that no two symbols share the same |
 505 | number.                                                           |
 506 `------------------------------------------------------------------*/
 507
 508 static void
 509 symbols_token_translations_init (void)
 510 {
 511   int num_256_available_p = TRUE;
 512   int i;
 513
 514   /* Find the highest user token number, and whether 256, the POSIX
 515      preferred user token number for the error token, is used.  */
 516   max_user_token_number = 0;
 517   for (i = 0; i < ntokens; ++i)
 518     {
 519       symbol_t *this = symbols[i];
 520       if (this->user_token_number != USER_NUMBER_UNDEFINED)
 521         {
 522           if (this->user_token_number > max_user_token_number)
 523             max_user_token_number = this->user_token_number;
 524           if (this->user_token_number == 256)
 525             num_256_available_p = FALSE;
 526         }
 527     }
 528
 529   /* If 256 is not used, assign it to error, to follow POSIX.  */
 530   if (num_256_available_p
 531       && errtoken->user_token_number == USER_NUMBER_UNDEFINED)
 532     errtoken->user_token_number = 256;
 533
 534   /* Set the missing user numbers. */
 535   if (max_user_token_number < 256)
 536     max_user_token_number = 256;
 537
 538   for (i = 0; i < ntokens; ++i)
 539     {
 540       symbol_t *this = symbols[i];
 541       if (this->user_token_number == USER_NUMBER_UNDEFINED)
 542         this->user_token_number = ++max_user_token_number;
 543       if (this->user_token_number > max_user_token_number)
 544         max_user_token_number = this->user_token_number;
 545     }
 546
 547   token_translations = XCALLOC (symbol_number_t, max_user_token_number + 1);
 548
 549   /* Initialize all entries for literal tokens to 2, the internal
 550      token number for $undefined., which represents all invalid
 551      inputs.  */
 552   for (i = 0; i < max_user_token_number + 1; i++)
 553     token_translations[i] = undeftoken->number;
 554   symbols_do (symbol_translation, NULL);
 555 }
 556
 557
 558 /*----------------------------------------------------------------.
 559 | Assign symbol numbers, and write definition of token names into |
 560 | FDEFINES.  Set up vectors SYMBOL_TABLE, TAGS of symbols.        |
 561 `----------------------------------------------------------------*/
 562
 563 void
 564 symbols_pack (void)
 565 {
 566   symbols = XCALLOC (symbol_t *, nsyms);
 567
 568   symbols_do (symbol_check_alias_consistence, NULL);
 569   symbols_do (symbol_pack, NULL);
 570
 571   symbols_token_translations_init ();
 572
 573   if (startsymbol->class == unknown_sym)
 574     fatal_at (startsymbol_location,
 575               _("the start symbol %s is undefined"),
 576               symbol_tag_get (startsymbol));
 577   else if (startsymbol->class == token_sym)
 578     fatal_at (startsymbol_location,
 579               _("the start symbol %s is a token"),
 580               symbol_tag_get (startsymbol));
 581 }