src/muscle-tab.c

   1 /* Muscle table manager for Bison.
   2
   3    Copyright (C) 2001-2012 Free Software Foundation, Inc.
   4
   5    This file is part of Bison, the GNU Compiler Compiler.
   6
   7    This program is free software: you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation, either version 3 of the License, or
  10    (at your option) any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  19
  20 #include <config.h>
  21 #include "system.h"
  22
  23 #include <hash.h>
  24
  25 #include "complain.h"
  26 #include "files.h"
  27 #include "getargs.h"
  28 #include "muscle-tab.h"
  29 #include "quote.h"
  30
  31 /* A key-value pair, along with storage that can be reclaimed when
  32    this pair is no longer needed.  */
  33 typedef struct
  34 {
  35   char const *key;
  36   char const *value;
  37   char *storage;
  38 } muscle_entry;
  39
  40 /* An obstack used to create some entries.  */
  41 struct obstack muscle_obstack;
  42
  43 /* Initial capacity of muscles hash table.  */
  44 #define HT_INITIAL_CAPACITY 257
  45
  46 static struct hash_table *muscle_table = NULL;
  47
  48 static bool
  49 hash_compare_muscles (void const *x, void const *y)
  50 {
  51   muscle_entry const *m1 = x;
  52   muscle_entry const *m2 = y;
  53   return STREQ (m1->key, m2->key);
  54 }
  55
  56 static size_t
  57 hash_muscle (const void *x, size_t tablesize)
  58 {
  59   muscle_entry const *m = x;
  60   return hash_string (m->key, tablesize);
  61 }
  62
  63 /*-----------------------------------------------------------------.
  64 | Create the MUSCLE_TABLE, and initialize it with default values.  |
  65 | Also set up the MUSCLE_OBSTACK.                                  |
  66 `-----------------------------------------------------------------*/
  67
  68 static void
  69 muscle_entry_free (void *entry)
  70 {
  71   muscle_entry *mentry = entry;
  72   free (mentry->storage);
  73   free (mentry);
  74 }
  75
  76 void
  77 muscle_init (void)
  78 {
  79   /* Initialize the muscle obstack.  */
  80   obstack_init (&muscle_obstack);
  81
  82   muscle_table = hash_initialize (HT_INITIAL_CAPACITY, NULL, hash_muscle,
  83                                   hash_compare_muscles, muscle_entry_free);
  84
  85   /* Version and input file.  */
  86   MUSCLE_INSERT_STRING ("version", VERSION);
  87 }
  88
  89
  90 /*------------------------------------------------------------.
  91 | Free all the memory consumed by the muscle machinery only.  |
  92 `------------------------------------------------------------*/
  93
  94 void
  95 muscle_free (void)
  96 {
  97   hash_free (muscle_table);
  98   obstack_free (&muscle_obstack, NULL);
  99 }
 100
 101
 102
 103 /*------------------------------------------------------------.
 104 | Insert (KEY, VALUE).  If KEY already existed, overwrite the |
 105 | previous value.                                             |
 106 `------------------------------------------------------------*/
 107
 108 void
 109 muscle_insert (char const *key, char const *value)
 110 {
 111   muscle_entry probe;
 112   muscle_entry *entry;
 113
 114   probe.key = key;
 115   entry = hash_lookup (muscle_table, &probe);
 116
 117   if (!entry)
 118     {
 119       /* First insertion in the hash. */
 120       entry = xmalloc (sizeof *entry);
 121       entry->key = key;
 122       if (!hash_insert (muscle_table, entry))
 123         xalloc_die ();
 124     }
 125   else
 126     free (entry->storage);
 127   entry->value = value;
 128   entry->storage = NULL;
 129 }
 130
 131
 132 /*-------------------------------------------------------------------.
 133 | Append VALUE to the current value of KEY.  If KEY did not already  |
 134 | exist, create it.  Use MUSCLE_OBSTACK.  De-allocate the previously |
 135 | associated value.  Copy VALUE and SEPARATOR.                       |
 136 `-------------------------------------------------------------------*/
 137
 138 void
 139 muscle_grow (const char *key, const char *val, const char *separator)
 140 {
 141   muscle_entry probe;
 142   muscle_entry *entry = NULL;
 143
 144   probe.key = key;
 145   entry = hash_lookup (muscle_table, &probe);
 146
 147   if (!entry)
 148     {
 149       /* First insertion in the hash. */
 150       entry = xmalloc (sizeof *entry);
 151       entry->key = key;
 152       if (!hash_insert (muscle_table, entry))
 153         xalloc_die ();
 154       entry->value = entry->storage = xstrdup (val);
 155     }
 156   else
 157     {
 158       /* Grow the current value. */
 159       char *new_val;
 160       obstack_printf (&muscle_obstack, "%s%s%s", entry->value, separator, val);
 161       free (entry->storage);
 162       new_val = obstack_finish0 (&muscle_obstack);
 163       entry->value = entry->storage = xstrdup (new_val);
 164       obstack_free (&muscle_obstack, new_val);
 165     }
 166 }
 167
 168 /*------------------------------------------------------------------.
 169 | Using muscle_grow, append a synchronization line for the location |
 170 | LOC to the current value of KEY.                                  |
 171 `------------------------------------------------------------------*/
 172
 173 static void
 174 muscle_syncline_grow (char const *key, location loc)
 175 {
 176   char *extension = NULL;
 177   obstack_printf (&muscle_obstack, "]b4_syncline(%d, ", loc.start.line);
 178   obstack_quote (&muscle_obstack,
 179                  quotearg_style (c_quoting_style, loc.start.file));
 180   obstack_sgrow (&muscle_obstack, ")[");
 181   extension = obstack_finish0 (&muscle_obstack);
 182   muscle_grow (key, extension, "");
 183   obstack_free (&muscle_obstack, extension);
 184 }
 185
 186 /*------------------------------------------------------------------.
 187 | Append VALUE to the current value of KEY, using muscle_grow.  But |
 188 | in addition, issue a synchronization line for the location LOC    |
 189 | using muscle_syncline_grow.                                       |
 190 `------------------------------------------------------------------*/
 191
 192 void
 193 muscle_code_grow (const char *key, const char *val, location loc)
 194 {
 195   muscle_syncline_grow (key, loc);
 196   muscle_grow (key, val, "\n");
 197 }
 198
 199
 200 void muscle_pair_list_grow (const char *muscle,
 201                             const char *a1, const char *a2)
 202 {
 203   char *pair;
 204   obstack_sgrow (&muscle_obstack, "[");
 205   obstack_quote (&muscle_obstack, a1);
 206   obstack_sgrow (&muscle_obstack, ", ");
 207   obstack_quote (&muscle_obstack, a2);
 208   obstack_sgrow (&muscle_obstack, "]");
 209   pair = obstack_finish0 (&muscle_obstack);
 210   muscle_grow (muscle, pair, ",\n");
 211   obstack_free (&muscle_obstack, pair);
 212 }
 213
 214
 215 /*----------------------------------------------------------------------------.
 216 | Find the value of muscle KEY.  Unlike MUSCLE_FIND, this is always reliable  |
 217 | to determine whether KEY has a value.                                       |
 218 `----------------------------------------------------------------------------*/
 219
 220 char const *
 221 muscle_find_const (char const *key)
 222 {
 223   muscle_entry probe;
 224   muscle_entry *result = NULL;
 225
 226   probe.key = key;
 227   result = hash_lookup (muscle_table, &probe);
 228   if (result)
 229     return result->value;
 230   return NULL;
 231 }
 232
 233
 234 /*----------------------------------------------------------------------------.
 235 | Find the value of muscle KEY.  Abort if muscle_insert was invoked more      |
 236 | recently than muscle_grow for KEY since muscle_find can't return a          |
 237 | char const *.                                                               |
 238 `----------------------------------------------------------------------------*/
 239
 240 char *
 241 muscle_find (char const *key)
 242 {
 243   muscle_entry probe;
 244   muscle_entry *result = NULL;
 245
 246   probe.key = key;
 247   result = hash_lookup (muscle_table, &probe);
 248   if (result)
 249     {
 250       aver (result->value == result->storage);
 251       return result->storage;
 252     }
 253   return NULL;
 254 }
 255
 256
 257 /* In the format `file_name:line.column', append BOUND to MUSCLE.  Use
 258    digraphs for special characters in the file name.  */
 259
 260 static void
 261 muscle_boundary_grow (char const *key, boundary bound)
 262 {
 263   char *extension;
 264   obstack_sgrow  (&muscle_obstack, "[[");
 265   obstack_escape (&muscle_obstack, bound.file);
 266   obstack_printf (&muscle_obstack, ":%d.%d]]", bound.line, bound.column);
 267   extension = obstack_finish0 (&muscle_obstack);
 268   muscle_grow (key, extension, "");
 269   obstack_free (&muscle_obstack, extension);
 270 }
 271
 272
 273 /* In the format `[[file_name:line.column]], [[file_name:line.column]]',
 274    append LOC to MUSCLE.  Use digraphs for special characters in each
 275    file name.  */
 276
 277 static void
 278 muscle_location_grow (char const *key, location loc)
 279 {
 280   muscle_boundary_grow (key, loc.start);
 281   muscle_grow (key, "", ", ");
 282   muscle_boundary_grow (key, loc.end);
 283 }
 284
 285 #define COMMON_DECODE(Value)                                    \
 286   case '$':                                                     \
 287     aver (*++(Value) == ']');                                   \
 288     aver (*++(Value) == '[');                                   \
 289     obstack_sgrow (&muscle_obstack, "$");                       \
 290     break;                                                      \
 291   case '@':                                                     \
 292     switch (*++(Value))                                         \
 293       {                                                         \
 294         case '@': obstack_sgrow (&muscle_obstack, "@" ); break; \
 295         case '{': obstack_sgrow (&muscle_obstack, "[" ); break; \
 296         case '}': obstack_sgrow (&muscle_obstack, "]" ); break; \
 297         default: aver (false); break;                           \
 298       }                                                         \
 299     break;                                                      \
 300   default:                                                      \
 301     obstack_1grow (&muscle_obstack, *(Value));                  \
 302     break;
 303
 304 /* Reverse of obstack_escape.  */
 305 static char *
 306 string_decode (char const *key)
 307 {
 308   char const *value = muscle_find_const (key);
 309   char *value_decoded;
 310   char *result;
 311
 312   if (!value)
 313     return NULL;
 314   do {
 315     switch (*value)
 316       {
 317         COMMON_DECODE (value)
 318         case '[':
 319         case ']':
 320           aver (false);
 321           break;
 322       }
 323   } while (*value++);
 324   value_decoded = obstack_finish (&muscle_obstack);
 325   result = xstrdup (value_decoded);
 326   obstack_free (&muscle_obstack, value_decoded);
 327   return result;
 328 }
 329
 330 /* Reverse of muscle_location_grow.  */
 331 static location
 332 location_decode (char const *key)
 333 {
 334   location loc;
 335   char const *value = muscle_find_const (key);
 336   aver (value);
 337   aver (*value == '[');
 338   aver (*++value == '[');
 339   while (*++value)
 340     switch (*value)
 341       {
 342         COMMON_DECODE (value)
 343         case '[':
 344           aver (false);
 345           break;
 346         case ']':
 347           {
 348             char *boundary_str;
 349             aver (*++value == ']');
 350             boundary_str = obstack_finish0 (&muscle_obstack);
 351             switch (*++value)
 352               {
 353                 case ',':
 354                   boundary_set_from_string (&loc.start, boundary_str);
 355                   obstack_free (&muscle_obstack, boundary_str);
 356                   aver (*++value == ' ');
 357                   aver (*++value == '[');
 358                   aver (*++value == '[');
 359                   break;
 360                 case '\0':
 361                   boundary_set_from_string (&loc.end, boundary_str);
 362                   obstack_free (&muscle_obstack, boundary_str);
 363                   return loc;
 364                   break;
 365                 default:
 366                   aver (false);
 367                   break;
 368               }
 369           }
 370           break;
 371       }
 372   aver (false);
 373   return loc;
 374 }
 375
 376 void
 377 muscle_user_name_list_grow (char const *key, char const *user_name,
 378                             location loc)
 379 {
 380   muscle_grow (key, "[[[[", ",");
 381   muscle_grow (key, user_name, "");
 382   muscle_grow (key, "]], ", "");
 383   muscle_location_grow (key, loc);
 384   muscle_grow (key, "]]", "");
 385 }
 386
 387 static
 388 char *
 389 define_directive (char const *assignment, char const *value)
 390 {
 391   char const *fmt = value && *value ? "%%define %s %s" : "%%define %s";
 392   char *res = xmalloc (strlen (fmt) + strlen (assignment)
 393                        + (value ? strlen (value) : 0));
 394   sprintf (res, fmt, assignment, value);
 395   return res;
 396 }
 397
 398 /** If the \a variable name is obsolete, return the name to use,
 399  * otherwise \a variable. */
 400 static
 401 char const *
 402 muscle_percent_variable_update (char const *variable, location variable_loc,
 403                                 char const **value)
 404 {
 405   typedef struct
 406   {
 407     const char *obsolete;
 408     const char *updated;
 409   } conversion_type;
 410   const conversion_type conversion[] =
 411     {
 412       { "api.push_pull", "api.push-pull", },
 413       { "api.tokens.prefix", "api.token.prefix", },
 414       { "lex_symbol", "api.token.constructor", },
 415       { "location_type", "api.location.type", },
 416       { "lr.default-reductions", "lr.default-reduction", },
 417       { "lr.keep-unreachable-states", "lr.keep-unreachable-state", },
 418       { "lr.keep_unreachable_states", "lr.keep-unreachable-state", },
 419       { "namespace", "api.namespace", },
 420       { NULL, NULL, }
 421     };
 422   char const *res = variable;
 423   conversion_type const *c;
 424   for (c = conversion; c->obsolete; ++c)
 425     if (STREQ (c->obsolete, variable))
 426       {
 427         char *old = define_directive (c->obsolete, *value);
 428         char *upd = define_directive (c->updated, *value);
 429         deprecated_directive (&variable_loc, old, upd);
 430         free (old);
 431         free (upd);
 432         res = c->updated;
 433         break;
 434     }
 435   return res;
 436 }
 437
 438 void
 439 muscle_percent_define_insert (char const *var, location variable_loc,
 440                               char const *value,
 441                               muscle_percent_define_how how)
 442 {
 443   /* Backward compatibility.  */
 444   char *variable = muscle_percent_variable_update (var, variable_loc, &value);
 445   char const *name = UNIQSTR_CONCAT ("percent_define(", variable, ")");
 446   char const *loc_name = UNIQSTR_CONCAT ("percent_define_loc(", variable, ")");
 447   char const *syncline_name =
 448     UNIQSTR_CONCAT ("percent_define_syncline(", variable, ")");
 449   char const *how_name = UNIQSTR_CONCAT ("percent_define_how(", variable, ")");
 450
 451   /* Command-line options are processed before the grammar file.  */
 452   if (how == MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE
 453       && muscle_find_const (name))
 454     {
 455       muscle_percent_define_how how_old = atoi (muscle_find_const (how_name));
 456       unsigned i = 0;
 457       if (how_old == MUSCLE_PERCENT_DEFINE_F)
 458         return;
 459       complain_indent (&variable_loc, complaint, &i,
 460                        _("%%define variable %s redefined"),
 461                        quote (variable));
 462       i += SUB_INDENT;
 463       location loc = muscle_percent_define_get_loc (variable);
 464       complain_indent (&loc, complaint, &i, _("previous definition"));
 465     }
 466
 467   MUSCLE_INSERT_STRING (name, value);
 468   muscle_insert (loc_name, "");
 469   muscle_location_grow (loc_name, variable_loc);
 470   muscle_insert (syncline_name, "");
 471   muscle_syncline_grow (syncline_name, variable_loc);
 472   muscle_user_name_list_grow ("percent_define_user_variables", variable,
 473                               variable_loc);
 474   MUSCLE_INSERT_INT (how_name, how);
 475 }
 476
 477 /* This is used for backward compatibility, e.g., "%define api.pure"
 478    supersedes "%pure-parser".  */
 479 void
 480 muscle_percent_define_ensure (char const *variable, location loc,
 481                               bool value)
 482 {
 483   char const *val = value ? "" : "false";
 484   char const *name = UNIQSTR_CONCAT ("percent_define(", variable, ")");
 485
 486   /* %pure-parser is deprecated in favor of `%define api.pure', so use
 487      `%define api.pure' in a backward-compatible manner here.  First,
 488      don't complain if %pure-parser is specified multiple times.  */
 489   if (!muscle_find_const (name))
 490     muscle_percent_define_insert (variable, loc, val,
 491                                   MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
 492   /* In all cases, use api.pure now so that the backend doesn't complain if
 493      the skeleton ignores api.pure, but do warn now if there's a previous
 494      conflicting definition from an actual %define.  */
 495   if (muscle_percent_define_flag_if (variable) != value)
 496     muscle_percent_define_insert (variable, loc, val,
 497                                   MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
 498 }
 499
 500 char *
 501 muscle_percent_define_get (char const *variable)
 502 {
 503   char const *name = UNIQSTR_CONCAT ("percent_define(", variable, ")");
 504   char const *usage_name =
 505     UNIQSTR_CONCAT ("percent_define_bison_variables(", variable, ")");
 506   char *value = string_decode (name);
 507   if (!value)
 508     value = xstrdup ("");
 509
 510   muscle_insert (usage_name, "");
 511   return value;
 512 }
 513
 514 location
 515 muscle_percent_define_get_loc (char const *variable)
 516 {
 517   char const *loc_name = UNIQSTR_CONCAT ("percent_define_loc(", variable, ")");
 518   if (!muscle_find_const (loc_name))
 519     complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 520               "muscle_percent_define_get_loc", quote (variable));
 521   return location_decode (loc_name);
 522 }
 523
 524 char const *
 525 muscle_percent_define_get_syncline (char const *variable)
 526 {
 527   char const *syncline_name =
 528     UNIQSTR_CONCAT ("percent_define_syncline(", variable, ")");
 529   char const *syncline = muscle_find_const (syncline_name);
 530   if (!syncline)
 531     complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 532               "muscle_percent_define_get_syncline", quote (variable));
 533   return syncline;
 534 }
 535
 536 bool
 537 muscle_percent_define_ifdef (char const *variable)
 538 {
 539   char const *name = UNIQSTR_CONCAT ("percent_define(", variable, ")");
 540   char const *usage_name =
 541     UNIQSTR_CONCAT ("percent_define_bison_variables(", variable, ")");
 542   char const *value = muscle_find_const (name);
 543   if (value)
 544     {
 545       muscle_insert (usage_name, "");
 546       return true;
 547     }
 548
 549   return false;
 550 }
 551
 552 bool
 553 muscle_percent_define_flag_if (char const *variable)
 554 {
 555   char const *invalid_boolean_name =
 556     UNIQSTR_CONCAT ("percent_define_invalid_boolean(", variable, ")");
 557   bool result = false;
 558
 559   if (muscle_percent_define_ifdef (variable))
 560     {
 561       char *value = muscle_percent_define_get (variable);
 562       if (value[0] == '\0' || STREQ (value, "true"))
 563         result = true;
 564       else if (STREQ (value, "false"))
 565         result = false;
 566       else if (!muscle_find_const (invalid_boolean_name))
 567         {
 568           muscle_insert (invalid_boolean_name, "");
 569           location loc = muscle_percent_define_get_loc (variable);
 570           complain (&loc, complaint,
 571                     _("invalid value for %%define Boolean variable %s"),
 572                     quote (variable));
 573         }
 574       free (value);
 575     }
 576   else
 577     complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 578               "muscle_percent_define_flag", quote (variable));
 579
 580   return result;
 581 }
 582
 583 void
 584 muscle_percent_define_default (char const *variable, char const *value)
 585 {
 586   char const *name = UNIQSTR_CONCAT ("percent_define(", variable, ")");
 587   char const *loc_name = UNIQSTR_CONCAT ("percent_define_loc(", variable, ")");
 588   char const *syncline_name =
 589     UNIQSTR_CONCAT ("percent_define_syncline(", variable, ")");
 590   if (!muscle_find_const (name))
 591     {
 592       location loc;
 593       MUSCLE_INSERT_STRING (name, value);
 594       loc.start.file = loc.end.file = "<default value>";
 595       loc.start.line = loc.end.line = -1;
 596       loc.start.column = loc.end.column = -1;
 597       muscle_insert (loc_name, "");
 598       muscle_location_grow (loc_name, loc);
 599       muscle_insert (syncline_name, "");
 600     }
 601 }
 602
 603 void
 604 muscle_percent_define_check_values (char const * const *values)
 605 {
 606   for (; *values; ++values)
 607     {
 608       char const * const *variablep = values;
 609       char const *name = UNIQSTR_CONCAT ("percent_define(", *variablep, ")");
 610       char *value = string_decode (name);
 611       if (value)
 612         {
 613           for (++values; *values; ++values)
 614             {
 615               if (STREQ (value, *values))
 616                 break;
 617             }
 618           if (!*values)
 619             {
 620               unsigned i = 0;
 621               location loc = muscle_percent_define_get_loc (*variablep);
 622               complain_indent (&loc, complaint, &i,
 623                                _("invalid value for %%define variable %s: %s"),
 624                                quote (*variablep), quote_n (1, value));
 625               i += SUB_INDENT;
 626               for (values = variablep + 1; *values; ++values)
 627                 complain_indent (&loc, complaint | no_caret, &i,
 628                                  _("accepted value: %s"), quote (*values));
 629             }
 630           else
 631             {
 632               while (*values)
 633                 ++values;
 634             }
 635           free (value);
 636         }
 637       else
 638         complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 639                   "muscle_percent_define_check_values", quote (*variablep));
 640     }
 641 }
 642
 643 void
 644 muscle_percent_code_grow (char const *qualifier, location qualifier_loc,
 645                           char const *code, location code_loc)
 646 {
 647   char const *name = UNIQSTR_CONCAT ("percent_code(", qualifier, ")");
 648   muscle_code_grow (name, code, code_loc);
 649   muscle_user_name_list_grow ("percent_code_user_qualifiers", qualifier,
 650                                qualifier_loc);
 651 }
 652
 653
 654 /*------------------------------------------------.
 655 | Output the definition of ENTRY as a m4_define.  |
 656 `------------------------------------------------*/
 657
 658 static inline bool
 659 muscle_m4_output (muscle_entry *entry, FILE *out)
 660 {
 661   fprintf (out,
 662            "m4_define([b4_%s],\n"
 663            "[[%s]])\n\n\n", entry->key, entry->value);
 664   return true;
 665 }
 666
 667 static bool
 668 muscle_m4_output_processor (void *entry, void *out)
 669 {
 670   return muscle_m4_output (entry, out);
 671 }
 672
 673
 674 /*----------------------------------------------------------------.
 675 | Output the definition of all the current muscles into a list of |
 676 | m4_defines.                                                     |
 677 `----------------------------------------------------------------*/
 678
 679 void
 680 muscles_m4_output (FILE *out)
 681 {
 682   hash_do_for_each (muscle_table, muscle_m4_output_processor, out);
 683 }