src/parse-gram.y

   1 %{/* Bison Grammar Parser                             -*- C -*-
   2
   3    Copyright (C) 2002, 2003, 2004, 2005, 2006 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 2 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, write to the Free Software
  19    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  20    02110-1301  USA
  21 */
  22
  23 #include <config.h>
  24 #include "system.h"
  25
  26 #include "complain.h"
  27 #include "conflicts.h"
  28 #include "files.h"
  29 #include "getargs.h"
  30 #include "gram.h"
  31 #include "muscle_tab.h"
  32 #include "output.h"
  33 #include "quotearg.h"
  34 #include "reader.h"
  35 #include "symlist.h"
  36 #include "strverscmp.h"
  37
  38 #define YYLLOC_DEFAULT(Current, Rhs, N)  (Current) = lloc_default (Rhs, N)
  39 static YYLTYPE lloc_default (YYLTYPE const *, int);
  40
  41 #define YY_LOCATION_PRINT(File, Loc) \
  42           location_print (File, Loc)
  43
  44 static void version_check (location const *loc, char const *version);
  45
  46 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
  47    FIXME: depends on the undocumented availability of YYLLOC.  */
  48 #undef  yyerror
  49 #define yyerror(Msg) \
  50         gram_error (&yylloc, Msg)
  51 static void gram_error (location const *, char const *);
  52
  53 static void add_param (char const *, char *, location);
  54
  55 static symbol_class current_class = unknown_sym;
  56 static uniqstr current_type = 0;
  57 static symbol *current_lhs;
  58 static location current_lhs_location;
  59 static int current_prec = 0;
  60 %}
  61
  62 %debug
  63 %verbose
  64 %defines
  65 %locations
  66 %pure-parser
  67 %error-verbose
  68 %defines
  69 %name-prefix="gram_"
  70
  71 %initial-action
  72 {
  73   /* Bison's grammar can initial empty locations, hence a default
  74      location is needed. */
  75   @$.start.file   = @$.end.file   = current_file;
  76   @$.start.line   = @$.end.line   = 1;
  77   @$.start.column = @$.end.column = 0;
  78 }
  79
  80 /* Only NUMBERS have a value.  */
  81 %union
  82 {
  83   symbol *symbol;
  84   symbol_list *list;
  85   int integer;
  86   char *chars;
  87   assoc assoc;
  88   uniqstr uniqstr;
  89 };
  90
  91 /* Define the tokens together with their human representation.  */
  92 %token GRAM_EOF 0 "end of file"
  93 %token STRING     "string"
  94 %token INT        "integer"
  95
  96 %token PERCENT_TOKEN       "%token"
  97 %token PERCENT_NTERM       "%nterm"
  98
  99 %token PERCENT_TYPE        "%type"
 100 %token PERCENT_DESTRUCTOR  "%destructor {...}"
 101 %token PERCENT_PRINTER     "%printer {...}"
 102
 103 %token PERCENT_UNION       "%union {...}"
 104
 105 %token PERCENT_LEFT        "%left"
 106 %token PERCENT_RIGHT       "%right"
 107 %token PERCENT_NONASSOC    "%nonassoc"
 108
 109 %token PERCENT_PREC          "%prec"
 110 %token PERCENT_DPREC         "%dprec"
 111 %token PERCENT_MERGE         "%merge"
 112
 113
 114 /*----------------------.
 115 | Global Declarations.  |
 116 `----------------------*/
 117
 118 %token
 119   PERCENT_DEBUG           "%debug"
 120   PERCENT_DEFAULT_PREC    "%default-prec"
 121   PERCENT_DEFINE          "%define"
 122   PERCENT_DEFINES         "%defines"
 123   PERCENT_ERROR_VERBOSE   "%error-verbose"
 124   PERCENT_EXPECT          "%expect"
 125   PERCENT_EXPECT_RR       "%expect-rr"
 126   PERCENT_FILE_PREFIX     "%file-prefix"
 127   PERCENT_GLR_PARSER      "%glr-parser"
 128   PERCENT_INITIAL_ACTION  "%initial-action {...}"
 129   PERCENT_LEX_PARAM       "%lex-param {...}"
 130   PERCENT_LOCATIONS       "%locations"
 131   PERCENT_NAME_PREFIX     "%name-prefix"
 132   PERCENT_NO_DEFAULT_PREC "%no-default-prec"
 133   PERCENT_NO_LINES        "%no-lines"
 134   PERCENT_NONDETERMINISTIC_PARSER
 135                           "%nondeterministic-parser"
 136   PERCENT_OUTPUT          "%output"
 137   PERCENT_PARSE_PARAM     "%parse-param {...}"
 138   PERCENT_PURE_PARSER     "%pure-parser"
 139   PERCENT_REQUIRE         "%require"
 140   PERCENT_SKELETON        "%skeleton"
 141   PERCENT_START           "%start"
 142   PERCENT_TOKEN_TABLE     "%token-table"
 143   PERCENT_VERBOSE         "%verbose"
 144   PERCENT_YACC            "%yacc"
 145 ;
 146
 147 %token TYPE            "type"
 148 %token EQUAL           "="
 149 %token SEMICOLON       ";"
 150 %token PIPE            "|"
 151 %token ID              "identifier"
 152 %token ID_COLON        "identifier:"
 153 %token PERCENT_PERCENT "%%"
 154 %token PROLOGUE        "%{...%}"
 155 %token EPILOGUE        "epilogue"
 156 %token BRACED_CODE     "{...}"
 157
 158
 159 %type <chars> STRING string_content
 160               "%destructor {...}"
 161               "%initial-action {...}"
 162               "%lex-param {...}"
 163               "%parse-param {...}"
 164               "%printer {...}"
 165               "%union {...}"
 166               BRACED_CODE action
 167               PROLOGUE EPILOGUE
 168 %printer { fprintf (stderr, "\"%s\"", $$); }
 169               STRING string_content
 170 %printer { fprintf (stderr, "{\n%s\n}", $$); }
 171               "%destructor {...}"
 172               "%initial-action {...}"
 173               "%lex-param {...}"
 174               "%parse-param {...}"
 175               "%printer {...}"
 176               "%union {...}"
 177               BRACED_CODE action
 178               PROLOGUE EPILOGUE
 179 %type <uniqstr> TYPE
 180 %printer { fprintf (stderr, "<%s>", $$); } TYPE
 181 %type <integer> INT
 182 %printer { fprintf (stderr, "%d", $$); } INT
 183 %type <symbol> ID symbol string_as_id
 184 %printer { fprintf (stderr, "%s", $$->tag); } ID symbol string_as_id
 185 %type <symbol> ID_COLON
 186 %printer { fprintf (stderr, "%s:", $$->tag); } ID_COLON
 187 %type <assoc> precedence_declarator
 188 %type <list>  symbols.1
 189 %%
 190
 191 input:
 192   declarations "%%" grammar epilogue.opt
 193 ;
 194
 195
 196         /*------------------------------------.
 197         | Declarations: before the first %%.  |
 198         `------------------------------------*/
 199
 200 declarations:
 201   /* Nothing */
 202 | declarations declaration
 203 ;
 204
 205 declaration:
 206   grammar_declaration
 207 | PROLOGUE                                 { prologue_augment ($1, @1); }
 208 | "%debug"                                 { debug_flag = true; }
 209 | "%define" string_content                 { muscle_insert ($2, "1"); }
 210 | "%define" string_content string_content  { muscle_insert ($2, $3); }
 211 | "%defines"                               { defines_flag = true; }
 212 | "%error-verbose"                         { error_verbose = true; }
 213 | "%expect" INT                            { expected_sr_conflicts = $2; }
 214 | "%expect-rr" INT                         { expected_rr_conflicts = $2; }
 215 | "%file-prefix" "=" string_content        { spec_file_prefix = $3; }
 216 | "%glr-parser"
 217   {
 218     nondeterministic_parser = true;
 219     glr_parser = true;
 220   }
 221 | "%initial-action {...}"
 222   {
 223     muscle_code_grow ("initial_action", $1, @1);
 224   }
 225 | "%lex-param {...}"                       { add_param ("lex_param", $1, @1); }
 226 | "%locations"                             { locations_flag = true; }
 227 | "%name-prefix" "=" string_content        { spec_name_prefix = $3; }
 228 | "%no-lines"                              { no_lines_flag = true; }
 229 | "%nondeterministic-parser"               { nondeterministic_parser = true; }
 230 | "%output" "=" string_content             { spec_outfile = $3; }
 231 | "%parse-param {...}"                     { add_param ("parse_param", $1, @1); }
 232 | "%pure-parser"                           { pure_parser = true; }
 233 | "%require" string_content                { version_check (&@2, $2); }
 234 | "%skeleton" string_content               { skeleton = $2; }
 235 | "%token-table"                           { token_table_flag = true; }
 236 | "%verbose"                               { report_flag = report_states; }
 237 | "%yacc"                                  { yacc_flag = true; }
 238 | /*FIXME: Err?  What is this horror doing here? */ ";"
 239 ;
 240
 241 grammar_declaration:
 242   precedence_declaration
 243 | symbol_declaration
 244 | "%start" symbol
 245     {
 246       grammar_start_symbol_set ($2, @2);
 247     }
 248 | "%union {...}"
 249     {
 250       char const *body = $1;
 251
 252       if (typed)
 253         {
 254           /* Concatenate the union bodies, turning the first one's
 255              trailing '}' into '\n', and omitting the second one's '{'.  */
 256           char *code = muscle_find ("stype");
 257           code[strlen (code) - 1] = '\n';
 258           body++;
 259         }
 260
 261       typed = true;
 262       muscle_code_grow ("stype", body, @1);
 263     }
 264 | "%destructor {...}" symbols.1
 265     {
 266       symbol_list *list;
 267       for (list = $2; list; list = list->next)
 268         symbol_destructor_set (list->sym, $1, @1);
 269       symbol_list_free ($2);
 270     }
 271 | "%printer {...}" symbols.1
 272     {
 273       symbol_list *list;
 274       for (list = $2; list; list = list->next)
 275         symbol_printer_set (list->sym, $1, list->location);
 276       symbol_list_free ($2);
 277     }
 278 | "%default-prec"
 279     {
 280       default_prec = true;
 281     }
 282 | "%no-default-prec"
 283     {
 284       default_prec = false;
 285     }
 286 ;
 287
 288 symbol_declaration:
 289   "%nterm" { current_class = nterm_sym; } symbol_defs.1
 290     {
 291       current_class = unknown_sym;
 292       current_type = NULL;
 293     }
 294 | "%token" { current_class = token_sym; } symbol_defs.1
 295     {
 296       current_class = unknown_sym;
 297       current_type = NULL;
 298     }
 299 | "%type" TYPE symbols.1
 300     {
 301       symbol_list *list;
 302       for (list = $3; list; list = list->next)
 303         symbol_type_set (list->sym, $2, @2);
 304       symbol_list_free ($3);
 305     }
 306 ;
 307
 308 precedence_declaration:
 309   precedence_declarator type.opt symbols.1
 310     {
 311       symbol_list *list;
 312       ++current_prec;
 313       for (list = $3; list; list = list->next)
 314         {
 315           symbol_type_set (list->sym, current_type, @2);
 316           symbol_precedence_set (list->sym, current_prec, $1, @1);
 317         }
 318       symbol_list_free ($3);
 319       current_type = NULL;
 320     }
 321 ;
 322
 323 precedence_declarator:
 324   "%left"     { $$ = left_assoc; }
 325 | "%right"    { $$ = right_assoc; }
 326 | "%nonassoc" { $$ = non_assoc; }
 327 ;
 328
 329 type.opt:
 330   /* Nothing. */ { current_type = NULL; }
 331 | TYPE           { current_type = $1; }
 332 ;
 333
 334 /* One or more nonterminals to be %typed. */
 335
 336 symbols.1:
 337   symbol            { $$ = symbol_list_new ($1, @1); }
 338 | symbols.1 symbol  { $$ = symbol_list_prepend ($1, $2, @2); }
 339 ;
 340
 341 /* One token definition.  */
 342 symbol_def:
 343   TYPE
 344      {
 345        current_type = $1;
 346      }
 347 | ID
 348      {
 349        symbol_class_set ($1, current_class, @1, true);
 350        symbol_type_set ($1, current_type, @1);
 351      }
 352 | ID INT
 353     {
 354       symbol_class_set ($1, current_class, @1, true);
 355       symbol_type_set ($1, current_type, @1);
 356       symbol_user_token_number_set ($1, $2, @2);
 357     }
 358 | ID string_as_id
 359     {
 360       symbol_class_set ($1, current_class, @1, true);
 361       symbol_type_set ($1, current_type, @1);
 362       symbol_make_alias ($1, $2, @$);
 363     }
 364 | ID INT string_as_id
 365     {
 366       symbol_class_set ($1, current_class, @1, true);
 367       symbol_type_set ($1, current_type, @1);
 368       symbol_user_token_number_set ($1, $2, @2);
 369       symbol_make_alias ($1, $3, @$);
 370     }
 371 ;
 372
 373 /* One or more symbol definitions. */
 374 symbol_defs.1:
 375   symbol_def
 376 | symbol_defs.1 symbol_def
 377 ;
 378
 379
 380         /*------------------------------------------.
 381         | The grammar section: between the two %%.  |
 382         `------------------------------------------*/
 383
 384 grammar:
 385   rules_or_grammar_declaration
 386 | grammar rules_or_grammar_declaration
 387 ;
 388
 389 /* As a Bison extension, one can use the grammar declarations in the
 390    body of the grammar.  */
 391 rules_or_grammar_declaration:
 392   rules
 393 | grammar_declaration ";"
 394     {
 395       if (yacc_flag)
 396         complain_at (@$, _("POSIX forbids declarations in the grammar"));
 397     }
 398 | error ";"
 399     {
 400       yyerrok;
 401     }
 402 ;
 403
 404 rules:
 405   ID_COLON { current_lhs = $1; current_lhs_location = @1; } rhses.1
 406 ;
 407
 408 rhses.1:
 409   rhs                { grammar_current_rule_end (@1); }
 410 | rhses.1 "|" rhs    { grammar_current_rule_end (@3); }
 411 | rhses.1 ";"
 412 ;
 413
 414 rhs:
 415   /* Nothing.  */
 416     { grammar_current_rule_begin (current_lhs, current_lhs_location); }
 417 | rhs symbol
 418     { grammar_current_rule_symbol_append ($2, @2); }
 419 | rhs action
 420     { grammar_current_rule_action_append ($2, @2); }
 421 | rhs "%prec" symbol
 422     { grammar_current_rule_prec_set ($3, @3); }
 423 | rhs "%dprec" INT
 424     { grammar_current_rule_dprec_set ($3, @3); }
 425 | rhs "%merge" TYPE
 426     { grammar_current_rule_merge_set ($3, @3); }
 427 ;
 428
 429 symbol:
 430   ID              { $$ = $1; }
 431 | string_as_id    { $$ = $1; }
 432 ;
 433
 434 action:
 435   BRACED_CODE
 436     { $$ = $1; }
 437 ;
 438
 439 /* A string used as an ID: quote it.  */
 440 string_as_id:
 441   STRING
 442     {
 443       $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
 444       symbol_class_set ($$, token_sym, @1, false);
 445     }
 446 ;
 447
 448 /* A string used for its contents.  Don't quote it.  */
 449 string_content:
 450   STRING
 451     { $$ = $1; }
 452 ;
 453
 454
 455 epilogue.opt:
 456   /* Nothing.  */
 457 | "%%" EPILOGUE
 458     {
 459       muscle_code_grow ("epilogue", $2, @2);
 460       scanner_last_string_free ();
 461     }
 462 ;
 463
 464 %%
 465
 466
 467 /* Return the location of the left-hand side of a rule whose
 468    right-hand side is RHS[1] ... RHS[N].  Ignore empty nonterminals in
 469    the right-hand side, and return an empty location equal to the end
 470    boundary of RHS[0] if the right-hand side is empty.  */
 471
 472 static YYLTYPE
 473 lloc_default (YYLTYPE const *rhs, int n)
 474 {
 475   int i;
 476   YYLTYPE loc;
 477
 478   /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
 479      The bug is fixed in 7.4.2m, but play it safe for now.  */
 480   loc.start = rhs[n].end;
 481   loc.end = rhs[n].end;
 482
 483   /* Ignore empty nonterminals the start of the the right-hand side.
 484      Do not bother to ignore them at the end of the right-hand side,
 485      since empty nonterminals have the same end as their predecessors.  */
 486   for (i = 1; i <= n; i++)
 487     if (! equal_boundaries (rhs[i].start, rhs[i].end))
 488       {
 489         loc.start = rhs[i].start;
 490         break;
 491       }
 492
 493   return loc;
 494 }
 495
 496
 497 /* Add a lex-param or a parse-param (depending on TYPE) with
 498    declaration DECL and location LOC.  */
 499
 500 static void
 501 add_param (char const *type, char *decl, location loc)
 502 {
 503   static char const alphanum[26 + 26 + 1 + 10] =
 504     "abcdefghijklmnopqrstuvwxyz"
 505     "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 506     "_"
 507     "0123456789";
 508   char const *name_start = NULL;
 509   char *p;
 510
 511   /* Stop on last actual character.  */
 512   for (p = decl; p[1]; p++)
 513     if ((p == decl
 514          || ! memchr (alphanum, p[-1], sizeof alphanum))
 515         && memchr (alphanum, p[0], sizeof alphanum - 10))
 516       name_start = p;
 517
 518   /* Strip the surrounding '{' and '}', and any blanks just inside
 519      the braces.  */
 520   while (*--p == ' ' || *p == '\t')
 521     continue;
 522   p[1] = '\0';
 523   while (*++decl == ' ' || *decl == '\t')
 524     continue;
 525
 526   if (! name_start)
 527     complain_at (loc, _("missing identifier in parameter declaration"));
 528   else
 529     {
 530       char *name;
 531       size_t name_len;
 532
 533       for (name_len = 1;
 534            memchr (alphanum, name_start[name_len], sizeof alphanum);
 535            name_len++)
 536         continue;
 537
 538       name = xmalloc (name_len + 1);
 539       memcpy (name, name_start, name_len);
 540       name[name_len] = '\0';
 541       muscle_pair_list_grow (type, decl, name);
 542       free (name);
 543     }
 544
 545   scanner_last_string_free ();
 546 }
 547
 548 static void
 549 version_check (location const *loc, char const *version)
 550 {
 551   if (strverscmp (version, PACKAGE_VERSION) > 0)
 552     {
 553       complain_at (*loc, "require bison %s, but have %s",
 554                    version, PACKAGE_VERSION);
 555       exit (63);
 556     }
 557 }
 558
 559 static void
 560 gram_error (location const *loc, char const *msg)
 561 {
 562   complain_at (*loc, "%s", msg);
 563 }
 564
 565 char const *
 566 token_name (int type)
 567 {
 568   return yytname[YYTRANSLATE (type)];
 569 }