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