src/muscle-tab.c

   1 /* Muscle table manager for Bison.
   2
   3    Copyright (C) 2001-2013 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 muscle_kind
  32 muscle_kind_new (char const *k)
  33 {
  34   if (STREQ (k, "code"))
  35     return muscle_code;
  36   else if (STREQ (k, "keyword"))
  37     return muscle_keyword;
  38   else if (STREQ (k, "string"))
  39     return muscle_string;
  40   aver (0);
  41 }
  42
  43 char const *
  44 muscle_kind_string (muscle_kind k)
  45 {
  46   switch (k)
  47     {
  48     case muscle_code:    return "code";
  49     case muscle_keyword: return "keyword";
  50     case muscle_string:  return "string";
  51     }
  52   aver (0);
  53 }
  54
  55
  56 /* A key-value pair, along with storage that can be reclaimed when
  57    this pair is no longer needed.  */
  58 typedef struct
  59 {
  60   char const *key;
  61   char const *value;
  62   char *storage;
  63   muscle_kind kind;
  64 } muscle_entry;
  65
  66
  67 /* The name of muscle for the %define variable VAR (corresponding to
  68    FIELD, if defined).  */
  69 static uniqstr
  70 muscle_name (char const *var, char const *field)
  71 {
  72   if (field)
  73     return UNIQSTR_CONCAT ("percent_define_", field, "(", var, ")");
  74   else
  75     return UNIQSTR_CONCAT ("percent_define(", var, ")");
  76 }
  77
  78 /* An obstack used to create some entries.  */
  79 struct obstack muscle_obstack;
  80
  81 /* Initial capacity of muscles hash table.  */
  82 #define HT_INITIAL_CAPACITY 257
  83
  84 static struct hash_table *muscle_table = NULL;
  85
  86 static bool
  87 hash_compare_muscles (void const *x, void const *y)
  88 {
  89   muscle_entry const *m1 = x;
  90   muscle_entry const *m2 = y;
  91   return STREQ (m1->key, m2->key);
  92 }
  93
  94 static size_t
  95 hash_muscle (const void *x, size_t tablesize)
  96 {
  97   muscle_entry const *m = x;
  98   return hash_string (m->key, tablesize);
  99 }
 100
 101 /* Create a fresh muscle name KEY, and insert in the hash table.  */
 102 static void *
 103 muscle_entry_new (char const *key)
 104 {
 105   muscle_entry *res = xmalloc (sizeof *res);
 106   res->key = key;
 107   res->value = NULL;
 108   res->storage = NULL;
 109   if (!hash_insert (muscle_table, res))
 110     xalloc_die ();
 111   return res;
 112 }
 113
 114 static void
 115 muscle_entry_free (void *entry)
 116 {
 117   muscle_entry *mentry = entry;
 118   free (mentry->storage);
 119   free (mentry);
 120 }
 121
 122 void
 123 muscle_init (void)
 124 {
 125   /* Initialize the muscle obstack.  */
 126   obstack_init (&muscle_obstack);
 127
 128   muscle_table = hash_initialize (HT_INITIAL_CAPACITY, NULL, hash_muscle,
 129                                   hash_compare_muscles, muscle_entry_free);
 130
 131   /* Version and input file.  */
 132   MUSCLE_INSERT_STRING ("version", VERSION);
 133 }
 134
 135
 136 void
 137 muscle_free (void)
 138 {
 139   hash_free (muscle_table);
 140   obstack_free (&muscle_obstack, NULL);
 141 }
 142
 143 /* Look for the muscle named KEY.  Return NULL if does not exist.  */
 144 static
 145 muscle_entry *
 146 muscle_lookup (char const *key)
 147 {
 148   muscle_entry probe;
 149   probe.key = key;
 150   return hash_lookup (muscle_table, &probe);
 151 }
 152
 153
 154 void
 155 muscle_insert (char const *key, char const *value)
 156 {
 157   muscle_entry *entry = muscle_lookup (key);
 158   if (entry)
 159     free (entry->storage);
 160   else
 161     /* First insertion in the hash. */
 162     entry = muscle_entry_new (key);
 163   entry->value = value;
 164   entry->storage = NULL;
 165 }
 166
 167
 168 /* Append VALUE to the current value of KEY.  If KEY did not already
 169    exist, create it.  Use MUSCLE_OBSTACK.  De-allocate the previously
 170    associated value.  Copy VALUE and SEPARATOR.  If VALUE does not end
 171    with TERMINATOR, append one.  */
 172
 173 static void
 174 muscle_grow (const char *key, const char *val,
 175              const char *separator, const char *terminator)
 176 {
 177   muscle_entry *entry = muscle_lookup (key);
 178   size_t vals = strlen (val);
 179   size_t terms = strlen (terminator);
 180
 181   if (entry)
 182     {
 183       obstack_sgrow (&muscle_obstack, entry->value);
 184       obstack_sgrow (&muscle_obstack, separator);
 185       free (entry->storage);
 186     }
 187   else
 188     entry = muscle_entry_new (key);
 189
 190   obstack_sgrow (&muscle_obstack, val);
 191
 192   if (terms <= vals
 193       && STRNEQ (val + vals - terms, terminator))
 194     obstack_sgrow (&muscle_obstack, terminator);
 195
 196   {
 197     char *new_val = obstack_finish0 (&muscle_obstack);
 198     entry->value = entry->storage = xstrdup (new_val);
 199     obstack_free (&muscle_obstack, new_val);
 200   }
 201 }
 202
 203 /*------------------------------------------------------------------.
 204 | Using muscle_grow, append a synchronization line for the location |
 205 | LOC to the current value of KEY.                                  |
 206 `------------------------------------------------------------------*/
 207
 208 static void
 209 muscle_syncline_grow (char const *key, location loc)
 210 {
 211   char *extension = NULL;
 212   obstack_printf (&muscle_obstack, "]b4_syncline(%d, ", loc.start.line);
 213   obstack_quote (&muscle_obstack,
 214                  quotearg_style (c_quoting_style, loc.start.file));
 215   obstack_sgrow (&muscle_obstack, ")[");
 216   extension = obstack_finish0 (&muscle_obstack);
 217   muscle_grow (key, extension, "", "");
 218   obstack_free (&muscle_obstack, extension);
 219 }
 220
 221 /*------------------------------------------------------------------.
 222 | Append VALUE to the current value of KEY, using muscle_grow.  But |
 223 | in addition, issue a synchronization line for the location LOC    |
 224 | using muscle_syncline_grow.                                       |
 225 `------------------------------------------------------------------*/
 226
 227 void
 228 muscle_code_grow (const char *key, const char *val, location loc)
 229 {
 230   muscle_syncline_grow (key, loc);
 231   muscle_grow (key, val, "\n", "\n");
 232 }
 233
 234
 235 void
 236 muscle_pair_list_grow (const char *muscle,
 237                        const char *a1, const char *a2)
 238 {
 239   char *pair;
 240   obstack_sgrow (&muscle_obstack, "[");
 241   obstack_quote (&muscle_obstack, a1);
 242   obstack_sgrow (&muscle_obstack, ", ");
 243   obstack_quote (&muscle_obstack, a2);
 244   obstack_sgrow (&muscle_obstack, "]");
 245   pair = obstack_finish0 (&muscle_obstack);
 246   muscle_grow (muscle, pair, ",\n", "");
 247   obstack_free (&muscle_obstack, pair);
 248 }
 249
 250
 251 char const *
 252 muscle_find_const (char const *key)
 253 {
 254   muscle_entry *entry = muscle_lookup (key);
 255   return entry ? entry->value : NULL;
 256 }
 257
 258
 259 char *
 260 muscle_find (char const *key)
 261 {
 262   muscle_entry *entry = muscle_lookup (key);
 263   if (entry)
 264     {
 265       aver (entry->value == entry->storage);
 266       return entry->storage;
 267     }
 268   return NULL;
 269 }
 270
 271
 272 /* In the format 'file_name:line.column', append BOUND to MUSCLE.  Use
 273    digraphs for special characters in the file name.  */
 274
 275 static void
 276 muscle_boundary_grow (char const *key, boundary bound)
 277 {
 278   char *extension;
 279   obstack_sgrow  (&muscle_obstack, "[[");
 280   obstack_escape (&muscle_obstack, bound.file);
 281   obstack_printf (&muscle_obstack, ":%d.%d]]", bound.line, bound.column);
 282   extension = obstack_finish0 (&muscle_obstack);
 283   muscle_grow (key, extension, "", "");
 284   obstack_free (&muscle_obstack, extension);
 285 }
 286
 287
 288 /* In the format '[[file_name:line.column]], [[file_name:line.column]]',
 289    append LOC to MUSCLE.  Use digraphs for special characters in each
 290    file name.  */
 291
 292 static void
 293 muscle_location_grow (char const *key, location loc)
 294 {
 295   muscle_boundary_grow (key, loc.start);
 296   muscle_grow (key, "", ", ", "");
 297   muscle_boundary_grow (key, loc.end);
 298 }
 299
 300 #define COMMON_DECODE(Value)                                    \
 301   case '$':                                                     \
 302     aver (*++(Value) == ']');                                   \
 303     aver (*++(Value) == '[');                                   \
 304     obstack_sgrow (&muscle_obstack, "$");                       \
 305     break;                                                      \
 306   case '@':                                                     \
 307     switch (*++(Value))                                         \
 308       {                                                         \
 309         case '@': obstack_sgrow (&muscle_obstack, "@" ); break; \
 310         case '{': obstack_sgrow (&muscle_obstack, "[" ); break; \
 311         case '}': obstack_sgrow (&muscle_obstack, "]" ); break; \
 312         default: aver (false); break;                           \
 313       }                                                         \
 314     break;                                                      \
 315   default:                                                      \
 316     obstack_1grow (&muscle_obstack, *(Value));                  \
 317     break;
 318
 319 /* Reverse of obstack_escape.  */
 320 static char *
 321 string_decode (char const *key)
 322 {
 323   char const *value = muscle_find_const (key);
 324   char *value_decoded;
 325   char *result;
 326
 327   if (!value)
 328     return NULL;
 329   do {
 330     switch (*value)
 331       {
 332         COMMON_DECODE (value)
 333         case '[':
 334         case ']':
 335           aver (false);
 336           break;
 337       }
 338   } while (*value++);
 339   value_decoded = obstack_finish (&muscle_obstack);
 340   result = xstrdup (value_decoded);
 341   obstack_free (&muscle_obstack, value_decoded);
 342   return result;
 343 }
 344
 345 /* Reverse of muscle_location_grow.  */
 346 static location
 347 location_decode (char const *value)
 348 {
 349   location loc;
 350   aver (value);
 351   aver (*value == '[');
 352   aver (*++value == '[');
 353   while (*++value)
 354     switch (*value)
 355       {
 356         COMMON_DECODE (value)
 357         case '[':
 358           aver (false);
 359           break;
 360         case ']':
 361           {
 362             char *boundary_str;
 363             aver (*++value == ']');
 364             boundary_str = obstack_finish0 (&muscle_obstack);
 365             switch (*++value)
 366               {
 367                 case ',':
 368                   boundary_set_from_string (&loc.start, boundary_str);
 369                   obstack_free (&muscle_obstack, boundary_str);
 370                   aver (*++value == ' ');
 371                   aver (*++value == '[');
 372                   aver (*++value == '[');
 373                   break;
 374                 case '\0':
 375                   boundary_set_from_string (&loc.end, boundary_str);
 376                   obstack_free (&muscle_obstack, boundary_str);
 377                   return loc;
 378                   break;
 379                 default:
 380                   aver (false);
 381                   break;
 382               }
 383           }
 384           break;
 385       }
 386   aver (false);
 387   return loc;
 388 }
 389
 390 void
 391 muscle_user_name_list_grow (char const *key, char const *user_name,
 392                             location loc)
 393 {
 394   muscle_grow (key, "[[[[", ",", "");
 395   muscle_grow (key, user_name, "", "");
 396   muscle_grow (key, "]], ", "", "");
 397   muscle_location_grow (key, loc);
 398   muscle_grow (key, "]]", "", "");
 399 }
 400
 401
 402 /** Return an allocated string that represents the %define directive
 403     that performs the assignment.
 404
 405     @param assignment "VAR", or "VAR=VAL".
 406     @param value      default value if VAL \a assignment has no '='.
 407
 408     For instance:
 409     "foo", NULL      => "%define foo"
 410     "foo", "baz"     => "%define foo baz"
 411     "foo=bar", NULL  => "%define foo bar"
 412     "foo=bar", "baz" => "%define foo bar"
 413     "foo=", NULL     => "%define foo"
 414     "foo=", "baz"    => "%define foo"
 415  */
 416
 417 static
 418 char *
 419 define_directive (char const *assignment, char const *value)
 420 {
 421   char *eq = strchr (assignment, '=');
 422   char const *fmt = !eq && value && *value ? "%%define %s %s" : "%%define %s";
 423   char *res = xmalloc (strlen (fmt) + strlen (assignment)
 424                        + (value ? strlen (value) : 0));
 425   sprintf (res, fmt, assignment, value);
 426   eq = strchr (res, '=');
 427   if (eq)
 428     *eq = eq[1] ? ' ' : '\0';
 429   return res;
 430 }
 431
 432 /** If the \a variable name is obsolete, return the name to use,
 433  * otherwise \a variable.  If the \a value is obsolete, update it too.
 434  *
 435  * Allocates the returned value.  */
 436 static
 437 char *
 438 muscle_percent_variable_update (char const *variable, location variable_loc,
 439                                 char const **value)
 440 {
 441   typedef struct
 442   {
 443     const char *obsolete;
 444     const char *updated;
 445   } conversion_type;
 446   const conversion_type conversion[] =
 447     {
 448       { "api.push_pull", "api.push-pull", },
 449       { "api.tokens.prefix", "api.token.prefix", },
 450       { "lex_symbol", "api.token.constructor", },
 451       { "location_type", "api.location.type", },
 452       { "lr.default-reductions", "lr.default-reduction", },
 453       { "lr.keep-unreachable-states", "lr.keep-unreachable-state", },
 454       { "lr.keep_unreachable_states", "lr.keep-unreachable-state", },
 455       { "namespace", "api.namespace", },
 456       { "stype", "api.value.type", },
 457       { "variant=",     "api.value.type=variant", },
 458       { "variant=true", "api.value.type=variant", },
 459       { NULL, NULL, }
 460     };
 461   conversion_type const *c;
 462   for (c = conversion; c->obsolete; ++c)
 463     {
 464       char const *eq = strchr (c->obsolete, '=');
 465       if (eq
 466           ? (!strncmp (c->obsolete, variable, eq - c->obsolete)
 467              && STREQ (eq + 1, *value))
 468           : STREQ (c->obsolete, variable))
 469         {
 470           char *old = define_directive (c->obsolete, *value);
 471           char *upd = define_directive (c->updated, *value);
 472           deprecated_directive (&variable_loc, old, upd);
 473           free (old);
 474           free (upd);
 475           char *res = xstrdup (c->updated);
 476           {
 477             char *eq2 = strchr (res, '=');
 478             if (eq2)
 479               {
 480                 *eq2 = '\0';
 481                 *value = eq2 + 1;
 482               }
 483           }
 484           return res;
 485         }
 486     }
 487   return xstrdup (variable);
 488 }
 489
 490 void
 491 muscle_percent_define_insert (char const *var, location variable_loc,
 492                               muscle_kind kind,
 493                               char const *value,
 494                               muscle_percent_define_how how)
 495 {
 496   /* Backward compatibility.  */
 497   char *variable = muscle_percent_variable_update (var, variable_loc, &value);
 498   uniqstr name = muscle_name (variable, NULL);
 499   uniqstr loc_name = muscle_name (variable, "loc");
 500   uniqstr syncline_name = muscle_name (variable, "syncline");
 501   uniqstr how_name = muscle_name (variable, "how");
 502   uniqstr kind_name = muscle_name (variable, "kind");
 503
 504   /* Command-line options are processed before the grammar file.  */
 505   if (how == MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE
 506       && muscle_find_const (name))
 507     {
 508       muscle_percent_define_how how_old = atoi (muscle_find_const (how_name));
 509       unsigned i = 0;
 510       if (how_old == MUSCLE_PERCENT_DEFINE_F)
 511         goto end;
 512       complain_indent (&variable_loc, complaint, &i,
 513                        _("%%define variable %s redefined"),
 514                        quote (variable));
 515       i += SUB_INDENT;
 516       location loc = muscle_percent_define_get_loc (variable);
 517       complain_indent (&loc, complaint, &i, _("previous definition"));
 518     }
 519
 520   MUSCLE_INSERT_STRING (name, value);
 521   muscle_insert (loc_name, "");
 522   muscle_location_grow (loc_name, variable_loc);
 523   muscle_insert (syncline_name, "");
 524   muscle_syncline_grow (syncline_name, variable_loc);
 525   muscle_user_name_list_grow ("percent_define_user_variables", variable,
 526                               variable_loc);
 527   MUSCLE_INSERT_INT (how_name, how);
 528   MUSCLE_INSERT_STRING (kind_name, muscle_kind_string (kind));
 529  end:
 530   free (variable);
 531 }
 532
 533 /* This is used for backward compatibility, e.g., "%define api.pure"
 534    supersedes "%pure-parser".  */
 535 void
 536 muscle_percent_define_ensure (char const *variable, location loc,
 537                               bool value)
 538 {
 539   uniqstr name = muscle_name (variable, NULL);
 540   char const *val = value ? "" : "false";
 541
 542   /* Don't complain is VARIABLE is already defined, but be sure to set
 543      its value to VAL.  */
 544   if (!muscle_find_const (name)
 545       || muscle_percent_define_flag_if (variable) != value)
 546     muscle_percent_define_insert (variable, loc, muscle_keyword, val,
 547                                   MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
 548 }
 549
 550 /* Mark %define VARIABLE as used.  */
 551 static void
 552 muscle_percent_define_use (char const *variable)
 553 {
 554   muscle_insert (muscle_name (variable, "bison_variables"), "");
 555 }
 556
 557 /* The value of %define variable VARIABLE (corresponding to FIELD, if
 558    defined).  Do not register as used, but diagnose unset variables.  */
 559
 560 char const *
 561 muscle_percent_define_get_raw (char const *variable, char const *field)
 562 {
 563   uniqstr name = muscle_name (variable, field);
 564   char const *res = muscle_find_const (name);
 565   if (!res)
 566     complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 567               "muscle_percent_define_get_raw", quote (variable));
 568   return res;
 569 }
 570
 571 char *
 572 muscle_percent_define_get (char const *variable)
 573 {
 574   uniqstr name = muscle_name (variable, NULL);
 575   char *value = string_decode (name);
 576   if (!value)
 577     value = xstrdup ("");
 578   muscle_percent_define_use (variable);
 579   return value;
 580 }
 581
 582 location
 583 muscle_percent_define_get_loc (char const *variable)
 584 {
 585   return location_decode (muscle_percent_define_get_raw (variable, "loc"));
 586 }
 587
 588 char const *
 589 muscle_percent_define_get_syncline (char const *variable)
 590 {
 591   return muscle_percent_define_get_raw (variable, "syncline");
 592 }
 593
 594 bool
 595 muscle_percent_define_ifdef (char const *variable)
 596 {
 597   if (muscle_find_const (muscle_name (variable, NULL)))
 598     {
 599       muscle_percent_define_use (variable);
 600       return true;
 601     }
 602   else
 603     return false;
 604 }
 605
 606 bool
 607 muscle_percent_define_flag_if (char const *variable)
 608 {
 609   uniqstr invalid_boolean_name = muscle_name (variable, "invalid_boolean");
 610   bool result = false;
 611
 612   if (muscle_percent_define_ifdef (variable))
 613     {
 614       char *value = muscle_percent_define_get (variable);
 615       if (value[0] == '\0' || STREQ (value, "true"))
 616         result = true;
 617       else if (STREQ (value, "false"))
 618         result = false;
 619       else if (!muscle_find_const (invalid_boolean_name))
 620         {
 621           muscle_insert (invalid_boolean_name, "");
 622           location loc = muscle_percent_define_get_loc (variable);
 623           complain (&loc, complaint,
 624                     _("invalid value for %%define Boolean variable %s"),
 625                     quote (variable));
 626         }
 627       free (value);
 628     }
 629   else
 630     complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 631               "muscle_percent_define_flag", quote (variable));
 632
 633   return result;
 634 }
 635
 636 void
 637 muscle_percent_define_default (char const *variable, char const *value)
 638 {
 639   uniqstr name = muscle_name (variable, NULL);
 640   if (!muscle_find_const (name))
 641     {
 642       MUSCLE_INSERT_STRING (name, value);
 643       {
 644         uniqstr loc_name = muscle_name (variable, "loc");
 645         location loc;
 646         loc.start.file = loc.end.file = "<default value>";
 647         loc.start.line = loc.end.line = -1;
 648         loc.start.column = loc.end.column = -1;
 649         muscle_insert (loc_name, "");
 650         muscle_location_grow (loc_name, loc);
 651       }
 652       muscle_insert (muscle_name (variable, "syncline"), "");
 653     }
 654 }
 655
 656 void
 657 muscle_percent_define_check_values (char const * const *values)
 658 {
 659   for (; *values; ++values)
 660     {
 661       char const * const *variablep = values;
 662       uniqstr name = muscle_name (*variablep, NULL);
 663       char *value = string_decode (name);
 664       if (value)
 665         {
 666           for (++values; *values; ++values)
 667             {
 668               if (STREQ (value, *values))
 669                 break;
 670             }
 671           if (!*values)
 672             {
 673               unsigned i = 0;
 674               location loc = muscle_percent_define_get_loc (*variablep);
 675               complain_indent (&loc, complaint, &i,
 676                                _("invalid value for %%define variable %s: %s"),
 677                                quote (*variablep), quote_n (1, value));
 678               i += SUB_INDENT;
 679               for (values = variablep + 1; *values; ++values)
 680                 complain_indent (&loc, complaint | no_caret | silent, &i,
 681                                  _("accepted value: %s"), quote (*values));
 682             }
 683           else
 684             {
 685               while (*values)
 686                 ++values;
 687             }
 688           free (value);
 689         }
 690       else
 691         complain (NULL, fatal, _("%s: undefined %%define variable %s"),
 692                   "muscle_percent_define_check_values", quote (*variablep));
 693     }
 694 }
 695
 696 void
 697 muscle_percent_code_grow (char const *qualifier, location qualifier_loc,
 698                           char const *code, location code_loc)
 699 {
 700   char const *name = UNIQSTR_CONCAT ("percent_code(", qualifier, ")");
 701   muscle_code_grow (name, code, code_loc);
 702   muscle_user_name_list_grow ("percent_code_user_qualifiers", qualifier,
 703                                qualifier_loc);
 704 }
 705
 706
 707 /*------------------------------------------------.
 708 | Output the definition of ENTRY as a m4_define.  |
 709 `------------------------------------------------*/
 710
 711 static inline bool
 712 muscle_m4_output (muscle_entry *entry, FILE *out)
 713 {
 714   fprintf (out,
 715            "m4_define([b4_%s],\n"
 716            "[[%s]])\n\n\n", entry->key, entry->value);
 717   return true;
 718 }
 719
 720 static bool
 721 muscle_m4_output_processor (void *entry, void *out)
 722 {
 723   return muscle_m4_output (entry, out);
 724 }
 725
 726
 727 void
 728 muscles_m4_output (FILE *out)
 729 {
 730   hash_do_for_each (muscle_table, muscle_m4_output_processor, out);
 731 }