src/parse-gram.y

   1 %{/* Bison Grammar Parser                             -*- C -*-
   2
   3    Copyright (C) 2002, 2003, 2004 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., 59 Temple Place, Suite 330, Boston, MA
  20    02111-1307  USA
  21 */
  22
  23 #include "system.h"
  24
  25 #include "complain.h"
  26 #include "conflicts.h"
  27 #include "files.h"
  28 #include "getargs.h"
  29 #include "gram.h"
  30 #include "muscle_tab.h"
  31 #include "output.h"
  32 #include "reader.h"
  33 #include "symlist.h"
  34
  35 #define YYLLOC_DEFAULT(Current, Rhs, N)  (Current) = lloc_default (Rhs, N)
  36 static YYLTYPE lloc_default (YYLTYPE const *, int);
  37
  38 #define YY_LOCATION_PRINT(File, Loc) \
  39           location_print (File, Loc)
  40
  41 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
  42    FIXME: depends on the undocumented availability of YYLLOC.  */
  43 #undef  yyerror
  44 #define yyerror(Msg) \
  45         gram_error (&yylloc, Msg)
  46 static void gram_error (location const *, char const *);
  47
  48 #define YYPRINT(File, Type, Value) \
  49         print_token_value (File, Type, &Value)
  50 static void print_token_value (FILE *, int, YYSTYPE const *);
  51
  52 static void add_param (char const *, char *, location);
  53
  54 symbol_class current_class = unknown_sym;
  55 uniqstr current_type = 0;
  56 symbol *current_lhs;
  57 location current_lhs_location;
  58 assoc current_assoc;
  59 int current_prec = 0;
  60 %}
  61
  62 %debug
  63 %defines
  64 %locations
  65 %pure-parser
  66 %error-verbose
  67 %defines
  68 %name-prefix="gram_"
  69
  70 %initial-action
  71 {
  72   /* Bison's grammar can initial empty locations, hence a default
  73      location is needed. */
  74   @$.start.file   = @$.end.file   = current_file;
  75   @$.start.line   = @$.end.line   = 1;
  76   @$.start.column = @$.end.column = 0;
  77 }
  78
  79 /* Only NUMBERS have a value.  */
  80 %union
  81 {
  82   symbol *symbol;
  83   symbol_list *list;
  84   int integer;
  85   char *chars;
  86   assoc assoc;
  87   uniqstr uniqstr;
  88 };
  89
  90 /* Define the tokens together with their human representation.  */
  91 %token GRAM_EOF 0 "end of file"
  92 %token STRING     "string"
  93 %token INT        "integer"
  94
  95 %token PERCENT_TOKEN       "%token"
  96 %token PERCENT_NTERM       "%nterm"
  97
  98 %token PERCENT_TYPE        "%type"
  99 %token PERCENT_DESTRUCTOR  "%destructor {...}"
 100 %token PERCENT_PRINTER     "%printer {...}"
 101
 102 %token PERCENT_UNION       "%union {...}"
 103
 104 %token PERCENT_LEFT        "%left"
 105 %token PERCENT_RIGHT       "%right"
 106 %token PERCENT_NONASSOC    "%nonassoc"
 107
 108 %token PERCENT_PREC          "%prec"
 109 %token PERCENT_DPREC         "%dprec"
 110 %token PERCENT_MERGE         "%merge"
 111
 112
 113 /*----------------------.
 114 | Global Declarations.  |
 115 `----------------------*/
 116
 117 %token
 118   PERCENT_DEBUG           "%debug"
 119   PERCENT_DEFAULT_PREC    "%default-prec"
 120   PERCENT_DEFINE          "%define"
 121   PERCENT_DEFINES         "%defines"
 122   PERCENT_ERROR_VERBOSE   "%error-verbose"
 123   PERCENT_EXPECT          "%expect"
 124   PERCENT_EXPECT_RR       "%expect-rr"
 125   PERCENT_FILE_PREFIX     "%file-prefix"
 126   PERCENT_GLR_PARSER      "%glr-parser"
 127   PERCENT_INITIAL_ACTION  "%initial-action {...}"
 128   PERCENT_LEX_PARAM       "%lex-param {...}"
 129   PERCENT_LOCATIONS       "%locations"
 130   PERCENT_NAME_PREFIX     "%name-prefix"
 131   PERCENT_NO_DEFAULT_PREC "%no-default-prec"
 132   PERCENT_NO_LINES        "%no-lines"
 133   PERCENT_NONDETERMINISTIC_PARSER
 134                           "%nondeterministic-parser"
 135   PERCENT_OUTPUT          "%output"
 136   PERCENT_PARSE_PARAM     "%parse-param {...}"
 137   PERCENT_PURE_PARSER     "%pure-parser"
 138   PERCENT_SKELETON        "%skeleton"
 139   PERCENT_START           "%start"
 140   PERCENT_TOKEN_TABLE     "%token-table"
 141   PERCENT_VERBOSE         "%verbose"
 142   PERCENT_YACC            "%yacc"
 143 ;
 144
 145 %token TYPE            "type"
 146 %token EQUAL           "="
 147 %token SEMICOLON       ";"
 148 %token PIPE            "|"
 149 %token ID              "identifier"
 150 %token ID_COLON        "identifier:"
 151 %token PERCENT_PERCENT "%%"
 152 %token PROLOGUE        "%{...%}"
 153 %token EPILOGUE        "epilogue"
 154 %token BRACED_CODE     "{...}"
 155
 156
 157 %type <chars> STRING string_content
 158               "%destructor {...}"
 159               "%initial-action {...}"
 160               "%lex-param {...}"
 161               "%parse-param {...}"
 162               "%printer {...}"
 163               "%union {...}"
 164               BRACED_CODE action
 165               PROLOGUE EPILOGUE
 166 %type <uniqstr> TYPE
 167 %type <integer> INT
 168 %type <symbol> ID ID_COLON symbol string_as_id
 169 %type <assoc> precedence_declarator
 170 %type <list>  symbols.1
 171 %%
 172
 173 input:
 174   declarations "%%" grammar epilogue.opt
 175 ;
 176
 177
 178         /*------------------------------------.
 179         | Declarations: before the first %%.  |
 180         `------------------------------------*/
 181
 182 declarations:
 183   /* Nothing */
 184 | declarations declaration
 185 ;
 186
 187 declaration:
 188   grammar_declaration
 189 | PROLOGUE                                 { prologue_augment ($1, @1); }
 190 | "%debug"                                 { debug_flag = true; }
 191 | "%define" string_content string_content  { muscle_insert ($2, $3); }
 192 | "%defines"                               { defines_flag = true; }
 193 | "%error-verbose"                         { error_verbose = true; }
 194 | "%expect" INT                            { expected_sr_conflicts = $2; }
 195 | "%expect-rr" INT                         { expected_rr_conflicts = $2; }
 196 | "%file-prefix" "=" string_content        { spec_file_prefix = $3; }
 197 | "%glr-parser"
 198   {
 199     nondeterministic_parser = true;
 200     glr_parser = true;
 201   }
 202 | "%initial-action {...}"
 203   {
 204     muscle_code_grow ("initial_action", $1, @1);
 205   }
 206 | "%lex-param {...}"                       { add_param ("lex_param", $1, @1); }
 207 | "%locations"                             { locations_flag = true; }
 208 | "%name-prefix" "=" string_content        { spec_name_prefix = $3; }
 209 | "%no-lines"                              { no_lines_flag = true; }
 210 | "%nondeterministic-parser"               { nondeterministic_parser = true; }
 211 | "%output" "=" string_content             { spec_outfile = $3; }
 212 | "%parse-param {...}"                     { add_param ("parse_param", $1, @1); }
 213 | "%pure-parser"                           { pure_parser = true; }
 214 | "%skeleton" string_content               { skeleton = $2; }
 215 | "%token-table"                           { token_table_flag = true; }
 216 | "%verbose"                               { report_flag = report_states; }
 217 | "%yacc"                                  { yacc_flag = true; }
 218 | /*FIXME: Err?  What is this horror doing here? */ ";"
 219 ;
 220
 221 grammar_declaration:
 222   precedence_declaration
 223 | symbol_declaration
 224 | "%start" symbol
 225     {
 226       grammar_start_symbol_set ($2, @2);
 227     }
 228 | "%union {...}"
 229     {
 230       typed = true;
 231       MUSCLE_INSERT_INT ("stype_line", @1.start.line);
 232       muscle_insert ("stype", $1);
 233     }
 234 | "%destructor {...}" symbols.1
 235     {
 236       symbol_list *list;
 237       for (list = $2; list; list = list->next)
 238         symbol_destructor_set (list->sym, $1, @1);
 239       symbol_list_free ($2);
 240     }
 241 | "%printer {...}" symbols.1
 242     {
 243       symbol_list *list;
 244       for (list = $2; list; list = list->next)
 245         symbol_printer_set (list->sym, $1, list->location);
 246       symbol_list_free ($2);
 247     }
 248 | "%default-prec"
 249     {
 250       default_prec = true;
 251     }
 252 | "%no-default-prec"
 253     {
 254       default_prec = false;
 255     }
 256 ;
 257
 258 symbol_declaration:
 259   "%nterm" { current_class = nterm_sym; } symbol_defs.1
 260     {
 261       current_class = unknown_sym;
 262       current_type = NULL;
 263     }
 264 | "%token" { current_class = token_sym; } symbol_defs.1
 265     {
 266       current_class = unknown_sym;
 267       current_type = NULL;
 268     }
 269 | "%type" TYPE symbols.1
 270     {
 271       symbol_list *list;
 272       for (list = $3; list; list = list->next)
 273         symbol_type_set (list->sym, $2, @2);
 274       symbol_list_free ($3);
 275     }
 276 ;
 277
 278 precedence_declaration:
 279   precedence_declarator type.opt symbols.1
 280     {
 281       symbol_list *list;
 282       ++current_prec;
 283       for (list = $3; list; list = list->next)
 284         {
 285           symbol_type_set (list->sym, current_type, @2);
 286           symbol_precedence_set (list->sym, current_prec, $1, @1);
 287         }
 288       symbol_list_free ($3);
 289       current_type = NULL;
 290     }
 291 ;
 292
 293 precedence_declarator:
 294   "%left"     { $$ = left_assoc; }
 295 | "%right"    { $$ = right_assoc; }
 296 | "%nonassoc" { $$ = non_assoc; }
 297 ;
 298
 299 type.opt:
 300   /* Nothing. */ { current_type = NULL; }
 301 | TYPE           { current_type = $1; }
 302 ;
 303
 304 /* One or more nonterminals to be %typed. */
 305
 306 symbols.1:
 307   symbol            { $$ = symbol_list_new ($1, @1); }
 308 | symbols.1 symbol  { $$ = symbol_list_prepend ($1, $2, @2); }
 309 ;
 310
 311 /* One token definition.  */
 312 symbol_def:
 313   TYPE
 314      {
 315        current_type = $1;
 316      }
 317 | ID
 318      {
 319        symbol_class_set ($1, current_class, @1);
 320        symbol_type_set ($1, current_type, @1);
 321      }
 322 | ID INT
 323     {
 324       symbol_class_set ($1, current_class, @1);
 325       symbol_type_set ($1, current_type, @1);
 326       symbol_user_token_number_set ($1, $2, @2);
 327     }
 328 | ID string_as_id
 329     {
 330       symbol_class_set ($1, current_class, @1);
 331       symbol_type_set ($1, current_type, @1);
 332       symbol_make_alias ($1, $2, @$);
 333     }
 334 | ID INT string_as_id
 335     {
 336       symbol_class_set ($1, current_class, @1);
 337       symbol_type_set ($1, current_type, @1);
 338       symbol_user_token_number_set ($1, $2, @2);
 339       symbol_make_alias ($1, $3, @$);
 340     }
 341 ;
 342
 343 /* One or more symbol definitions. */
 344 symbol_defs.1:
 345   symbol_def
 346 | symbol_defs.1 symbol_def
 347 ;
 348
 349
 350         /*------------------------------------------.
 351         | The grammar section: between the two %%.  |
 352         `------------------------------------------*/
 353
 354 grammar:
 355   rules_or_grammar_declaration
 356 | grammar rules_or_grammar_declaration
 357 ;
 358
 359 /* As a Bison extension, one can use the grammar declarations in the
 360    body of the grammar.  */
 361 rules_or_grammar_declaration:
 362   rules
 363 | grammar_declaration ";"
 364     {
 365       if (yacc_flag)
 366         complain_at (@$, _("POSIX forbids declarations in the grammar"));
 367     }
 368 | error ";"
 369     {
 370       yyerrok;
 371     }
 372 ;
 373
 374 rules:
 375   ID_COLON { current_lhs = $1; current_lhs_location = @1; } rhses.1
 376 ;
 377
 378 rhses.1:
 379   rhs                { grammar_rule_end (@1); }
 380 | rhses.1 "|" rhs    { grammar_rule_end (@3); }
 381 | rhses.1 ";"
 382 ;
 383
 384 rhs:
 385   /* Nothing.  */
 386     { grammar_rule_begin (current_lhs, current_lhs_location); }
 387 | rhs symbol
 388     { grammar_current_rule_symbol_append ($2, @2); }
 389 | rhs action
 390     { grammar_current_rule_action_append ($2, @2); }
 391 | rhs "%prec" symbol
 392     { grammar_current_rule_prec_set ($3, @3); }
 393 | rhs "%dprec" INT
 394     { grammar_current_rule_dprec_set ($3, @3); }
 395 | rhs "%merge" TYPE
 396     { grammar_current_rule_merge_set ($3, @3); }
 397 ;
 398
 399 symbol:
 400   ID              { $$ = $1; }
 401 | string_as_id    { $$ = $1; }
 402 ;
 403
 404 action:
 405   BRACED_CODE
 406    { $$ = $1; }
 407 ;
 408
 409 /* A string used as an ID: we have to keep the quotes. */
 410 string_as_id:
 411   STRING
 412     {
 413       $$ = symbol_get ($1, @1);
 414       symbol_class_set ($$, token_sym, @1);
 415     }
 416 ;
 417
 418 /* A string used for its contents.  Strip the quotes. */
 419 string_content:
 420   STRING
 421     {
 422       $$ = $1 + 1;
 423       $$[strlen ($$) - 1] = '\0';
 424     };
 425
 426
 427 epilogue.opt:
 428   /* Nothing.  */
 429 | "%%" EPILOGUE
 430     {
 431       muscle_code_grow ("epilogue", $2, @2);
 432       scanner_last_string_free ();
 433     }
 434 ;
 435
 436 %%
 437
 438
 439 /* Return the location of the left-hand side of a rule whose
 440    right-hand side is RHS[1] ... RHS[N].  Ignore empty nonterminals in
 441    the right-hand side, and return an empty location equal to the end
 442    boundary of RHS[0] if the right-hand side is empty.  */
 443
 444 static YYLTYPE
 445 lloc_default (YYLTYPE const *rhs, int n)
 446 {
 447   int i;
 448   YYLTYPE loc;
 449
 450   /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
 451      The bug is fixed in 7.4.2m, but play it safe for now.  */
 452   loc.start = rhs[n].end;
 453   loc.end = rhs[n].end;
 454
 455   /* Ignore empty nonterminals the start of the the right-hand side.
 456      Do not bother to ignore them at the end of the right-hand side,
 457      since empty nonterminals have the same end as their predecessors.  */
 458   for (i = 1; i <= n; i++)
 459     if (! equal_boundaries (rhs[i].start, rhs[i].end))
 460       {
 461         loc.start = rhs[i].start;
 462         break;
 463       }
 464
 465   return loc;
 466 }
 467
 468
 469 /* Add a lex-param or a parse-param (depending on TYPE) with
 470    declaration DECL and location LOC.  */
 471
 472 static void
 473 add_param (char const *type, char *decl, location loc)
 474 {
 475   static char const alphanum[] =
 476     "0123456789"
 477     "abcdefghijklmnopqrstuvwxyz"
 478     "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 479     "_";
 480   static char const blank[] = " \t";
 481   char const *alpha = alphanum + 10;
 482   char const *name_start = NULL;
 483   char *p;
 484
 485   for (p = decl; *p; p++)
 486     if ((p == decl || ! strchr (alphanum, p[-1])) && strchr (alpha, p[0]))
 487       name_start = p;
 488
 489   /* Strip the surrounding '{' and '}'.  */
 490   decl++;
 491   *--p = '\0';
 492
 493   /* Strip surrounding white spaces.  */
 494   while (strchr (blank, *decl))
 495     ++decl;
 496   while (strchr (blank, p[-1]))
 497     *--p = '\0';
 498
 499   if (! name_start)
 500     complain_at (loc, _("missing identifier in parameter declaration"));
 501   else
 502     {
 503       char *name;
 504       size_t name_len;
 505
 506       for (name_len = 1;
 507            name_start[name_len] && strchr (alphanum, name_start[name_len]);
 508            name_len++)
 509         continue;
 510
 511       name = xmalloc (name_len + 1);
 512       memcpy (name, name_start, name_len);
 513       name[name_len] = '\0';
 514       muscle_pair_list_grow (type, decl, name);
 515       free (name);
 516     }
 517
 518   scanner_last_string_free ();
 519 }
 520
 521 /*----------------------------------------------------.
 522 | When debugging the parser, display tokens' values.  |
 523 `----------------------------------------------------*/
 524
 525 static void
 526 print_token_value (FILE *file, int type, YYSTYPE const *value)
 527 {
 528   fputc (' ', file);
 529   switch (type)
 530     {
 531     case ID:
 532       fprintf (file, " = %s", value->symbol->tag);
 533       break;
 534
 535     case INT:
 536       fprintf (file, " = %d", value->integer);
 537       break;
 538
 539     case STRING:
 540       fprintf (file, " = \"%s\"", value->chars);
 541       break;
 542
 543     case TYPE:
 544       fprintf (file, " = <%s>", value->uniqstr);
 545       break;
 546
 547     case BRACED_CODE:
 548     case PERCENT_DESTRUCTOR:
 549     case PERCENT_LEX_PARAM:
 550     case PERCENT_PARSE_PARAM:
 551     case PERCENT_PRINTER:
 552     case PERCENT_UNION:
 553     case PROLOGUE:
 554     case EPILOGUE:
 555       fprintf (file, " = {{ %s }}", value->chars);
 556       break;
 557
 558     default:
 559       fprintf (file, "unknown token type");
 560       break;
 561     }
 562 }
 563
 564 static void
 565 gram_error (location const *loc, char const *msg)
 566 {
 567   complain_at (*loc, "%s", msg);
 568 }
 569
 570 char const *
 571 token_name (int type)
 572 {
 573   return yytname[type];
 574 }