]> git.saurik.com Git - bison.git/blob - src/parse-gram.y
* lib/.cvsignore: Add charset.alias.
[bison.git] / src / parse-gram.y
1 %{/* Bison Grammar Parser -*- C -*-
2
3 Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007 Free Software Foundation,
4 Inc.
5
6 This file is part of Bison, the GNU Compiler Compiler.
7
8 This program is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 2 of the License, or
11 (at your option) any later version.
12
13 This program is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
17
18 You should have received a copy of the GNU General Public License
19 along with this program; if not, write to the Free Software
20 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
21 02110-1301 USA
22 */
23
24 #include <config.h>
25 #include "system.h"
26
27 #include <strverscmp.h>
28
29 #include "complain.h"
30 #include "conflicts.h"
31 #include "files.h"
32 #include "getargs.h"
33 #include "gram.h"
34 #include "muscle_tab.h"
35 #include "quotearg.h"
36 #include "reader.h"
37 #include "symlist.h"
38 #include "scan-gram.h"
39 #include "scan-code.h"
40
41 #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
42 static YYLTYPE lloc_default (YYLTYPE const *, int);
43
44 #define YY_LOCATION_PRINT(File, Loc) \
45 location_print (File, Loc)
46
47 static void version_check (location const *loc, char const *version);
48
49 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
50 FIXME: depends on the undocumented availability of YYLLOC. */
51 #undef yyerror
52 #define yyerror(Msg) \
53 gram_error (&yylloc, Msg)
54 static void gram_error (location const *, char const *);
55
56 static char const *char_name (char);
57
58 /** Add a lex-param or a parse-param.
59 *
60 * \param type \a lex_param or \a parse_param
61 * \param decl the formal argument
62 * \param loc the location in the source.
63 */
64 static void add_param (char const *type, char *decl, location loc);
65
66
67 static symbol_class current_class = unknown_sym;
68 static uniqstr current_type = NULL;
69 static symbol *current_lhs;
70 static location current_lhs_location;
71 static int current_prec = 0;
72
73 #define YYTYPE_INT16 int_fast16_t
74 #define YYTYPE_INT8 int_fast8_t
75 #define YYTYPE_UINT16 uint_fast16_t
76 #define YYTYPE_UINT8 uint_fast8_t
77 %}
78
79 %debug
80 %verbose
81 %defines
82 %locations
83 %pure-parser
84 %error-verbose
85 %name-prefix="gram_"
86 %expect 0
87
88 %initial-action
89 {
90 /* Bison's grammar can initial empty locations, hence a default
91 location is needed. */
92 boundary_set (&@$.start, current_file, 1, 1);
93 boundary_set (&@$.end, current_file, 1, 1);
94 }
95
96 %union
97 {
98 symbol *symbol;
99 symbol_list *list;
100 int integer;
101 char const *chars;
102 char *code;
103 assoc assoc;
104 uniqstr uniqstr;
105 unsigned char character;
106 };
107
108 /* Define the tokens together with their human representation. */
109 %token GRAM_EOF 0 "end of file"
110 %token STRING "string"
111 %token INT "integer"
112
113 %token PERCENT_TOKEN "%token"
114 %token PERCENT_NTERM "%nterm"
115
116 %token PERCENT_TYPE "%type"
117 %token PERCENT_DESTRUCTOR "%destructor"
118 %token PERCENT_PRINTER "%printer"
119
120 %token PERCENT_LEFT "%left"
121 %token PERCENT_RIGHT "%right"
122 %token PERCENT_NONASSOC "%nonassoc"
123
124 %token PERCENT_PREC "%prec"
125 %token PERCENT_DPREC "%dprec"
126 %token PERCENT_MERGE "%merge"
127
128
129 /*----------------------.
130 | Global Declarations. |
131 `----------------------*/
132
133 %token
134 PERCENT_CODE "%code"
135 PERCENT_DEBUG "%debug"
136 PERCENT_DEFAULT_PREC "%default-prec"
137 PERCENT_DEFINE "%define"
138 PERCENT_DEFINES "%defines"
139 PERCENT_ERROR_VERBOSE "%error-verbose"
140 PERCENT_EXPECT "%expect"
141 PERCENT_EXPECT_RR "%expect-rr"
142 PERCENT_FILE_PREFIX "%file-prefix"
143 PERCENT_GLR_PARSER "%glr-parser"
144 PERCENT_INITIAL_ACTION "%initial-action"
145 PERCENT_LANGUAGE "%language"
146 PERCENT_LEX_PARAM "%lex-param"
147 PERCENT_LOCATIONS "%locations"
148 PERCENT_NAME_PREFIX "%name-prefix"
149 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
150 PERCENT_NO_LINES "%no-lines"
151 PERCENT_NONDETERMINISTIC_PARSER
152 "%nondeterministic-parser"
153 PERCENT_OUTPUT "%output"
154 PERCENT_PARSE_PARAM "%parse-param"
155 PERCENT_PURE_PARSER "%pure-parser"
156 PERCENT_PUSH_PARSER "%push-parser"
157 PERCENT_PUSH_PULL_PARSER
158 "%push-pull-parser"
159 PERCENT_REQUIRE "%require"
160 PERCENT_SKELETON "%skeleton"
161 PERCENT_START "%start"
162 PERCENT_TOKEN_TABLE "%token-table"
163 PERCENT_VERBOSE "%verbose"
164 PERCENT_YACC "%yacc"
165 ;
166
167 %token BRACED_CODE "{...}"
168 %token CHAR "char"
169 %token EPILOGUE "epilogue"
170 %token EQUAL "="
171 %token ID "identifier"
172 %token ID_COLON "identifier:"
173 %token PERCENT_PERCENT "%%"
174 %token PIPE "|"
175 %token PROLOGUE "%{...%}"
176 %token SEMICOLON ";"
177 %token TYPE "type"
178 %token TYPE_TAG_ANY "<*>"
179 %token TYPE_TAG_NONE "<>"
180
181 %type <character> CHAR
182 %printer { fputs (char_name ($$), stderr); } CHAR
183
184 /* braceless is not to be used for rule or symbol actions, as it
185 calls code_props_plain_init. */
186 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
187 %type <code> "{...}"
188 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
189 STRING
190 %printer { fprintf (stderr, "{\n%s\n}", $$); }
191 braceless content.opt "{...}" "%{...%}" EPILOGUE
192
193 %type <uniqstr> TYPE ID ID_COLON variable
194 %printer { fprintf (stderr, "<%s>", $$); } TYPE
195 %printer { fputs ($$, stderr); } ID variable
196 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
197
198 %type <integer> INT
199 %printer { fprintf (stderr, "%d", $$); } INT
200
201 %type <symbol> id id_colon symbol string_as_id
202 %printer { fprintf (stderr, "%s", $$->tag); } id symbol string_as_id
203 %printer { fprintf (stderr, "%s:", $$->tag); } id_colon
204
205 %type <assoc> precedence_declarator
206 %type <list> symbols.1 generic_symlist generic_symlist_item
207 %%
208
209 input:
210 prologue_declarations "%%" grammar epilogue.opt
211 ;
212
213
214 /*------------------------------------.
215 | Declarations: before the first %%. |
216 `------------------------------------*/
217
218 prologue_declarations:
219 /* Nothing */
220 | prologue_declarations prologue_declaration
221 ;
222
223 prologue_declaration:
224 grammar_declaration
225 | "%{...%}"
226 {
227 code_props plain_code;
228 code_props_plain_init (&plain_code, $1, @1);
229 code_props_translate_code (&plain_code);
230 gram_scanner_last_string_free ();
231 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
232 plain_code.code, @1);
233 code_scanner_last_string_free ();
234 }
235 | "%debug" { debug_flag = true; }
236 | "%define" variable content.opt
237 {
238 muscle_percent_define_insert ($2, @2, $3);
239 }
240 | "%defines" { defines_flag = true; }
241 | "%defines" STRING
242 {
243 defines_flag = true;
244 spec_defines_file = xstrdup ($2);
245 }
246 | "%error-verbose" { error_verbose = true; }
247 | "%expect" INT { expected_sr_conflicts = $2; }
248 | "%expect-rr" INT { expected_rr_conflicts = $2; }
249 | "%file-prefix" STRING { spec_file_prefix = $2; }
250 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
251 | "%glr-parser"
252 {
253 nondeterministic_parser = true;
254 glr_parser = true;
255 }
256 | "%initial-action" "{...}"
257 {
258 code_props action;
259 code_props_symbol_action_init (&action, $2, @2);
260 code_props_translate_code (&action);
261 gram_scanner_last_string_free ();
262 muscle_code_grow ("initial_action", action.code, @2);
263 code_scanner_last_string_free ();
264 }
265 | "%language" STRING { language_argmatch ($2, 1, &@1); }
266 | "%lex-param" "{...}" { add_param ("lex_param", $2, @2); }
267 | "%locations" { locations_flag = true; }
268 | "%name-prefix" STRING { spec_name_prefix = $2; }
269 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
270 | "%no-lines" { no_lines_flag = true; }
271 | "%nondeterministic-parser" { nondeterministic_parser = true; }
272 | "%output" STRING { spec_outfile = $2; }
273 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
274 | "%parse-param" "{...}" { add_param ("parse_param", $2, @2); }
275 | "%pure-parser" { pure_parser = true; }
276 | "%push-parser" { push_parser = true; pull_parser = false; }
277 | "%push-pull-parser" { push_parser = true; pull_parser = true; }
278 | "%require" STRING { version_check (&@2, $2); }
279 | "%skeleton" STRING
280 {
281 char const *skeleton_user = $2;
282 if (strchr (skeleton_user, '/'))
283 {
284 size_t dir_length = strlen (current_file);
285 char *skeleton_build;
286 while (dir_length && current_file[dir_length - 1] != '/')
287 --dir_length;
288 while (dir_length && current_file[dir_length - 1] == '/')
289 --dir_length;
290 skeleton_build =
291 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
292 if (dir_length > 0)
293 {
294 strncpy (skeleton_build, current_file, dir_length);
295 skeleton_build[dir_length++] = '/';
296 }
297 strcpy (skeleton_build + dir_length, skeleton_user);
298 skeleton_user = uniqstr_new (skeleton_build);
299 free (skeleton_build);
300 }
301 skeleton_arg (skeleton_user, 1, &@1);
302 }
303 | "%token-table" { token_table_flag = true; }
304 | "%verbose" { report_flag = report_states; }
305 | "%yacc" { yacc_flag = true; }
306 | /*FIXME: Err? What is this horror doing here? */ ";"
307 ;
308
309 grammar_declaration:
310 precedence_declaration
311 | symbol_declaration
312 | "%start" symbol
313 {
314 grammar_start_symbol_set ($2, @2);
315 }
316 | "%destructor" "{...}" generic_symlist
317 {
318 symbol_list *list;
319 for (list = $3; list; list = list->next)
320 symbol_list_destructor_set (list, $2, @2);
321 symbol_list_free ($3);
322 }
323 | "%printer" "{...}" generic_symlist
324 {
325 symbol_list *list;
326 for (list = $3; list; list = list->next)
327 symbol_list_printer_set (list, $2, @2);
328 symbol_list_free ($3);
329 }
330 | "%default-prec"
331 {
332 default_prec = true;
333 }
334 | "%no-default-prec"
335 {
336 default_prec = false;
337 }
338 | "%code" braceless
339 {
340 /* Do not invoke muscle_percent_code_grow here since it invokes
341 muscle_user_name_list_grow. */
342 muscle_code_grow ("percent_code()", $2, @2);
343 code_scanner_last_string_free ();
344 }
345 | "%code" ID braceless
346 {
347 muscle_percent_code_grow ($2, @2, $3, @3);
348 code_scanner_last_string_free ();
349 }
350 ;
351
352
353 /*----------*
354 | %union. |
355 *----------*/
356
357 %token PERCENT_UNION "%union";
358
359 union_name:
360 /* Nothing. */ {}
361 | ID { muscle_code_grow ("union_name", $1, @1); }
362 ;
363
364 grammar_declaration:
365 "%union" union_name braceless
366 {
367 union_seen = true;
368 muscle_code_grow ("stype", $3, @3);
369 code_scanner_last_string_free ();
370 }
371 ;
372
373
374
375
376 symbol_declaration:
377 "%nterm" { current_class = nterm_sym; } symbol_defs.1
378 {
379 current_class = unknown_sym;
380 current_type = NULL;
381 }
382 | "%token" { current_class = token_sym; } symbol_defs.1
383 {
384 current_class = unknown_sym;
385 current_type = NULL;
386 }
387 | "%type" TYPE symbols.1
388 {
389 symbol_list *list;
390 tag_seen = true;
391 for (list = $3; list; list = list->next)
392 symbol_type_set (list->content.sym, $2, @2);
393 symbol_list_free ($3);
394 }
395 ;
396
397 precedence_declaration:
398 precedence_declarator type.opt symbols.1
399 {
400 symbol_list *list;
401 ++current_prec;
402 for (list = $3; list; list = list->next)
403 {
404 symbol_type_set (list->content.sym, current_type, @2);
405 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
406 }
407 symbol_list_free ($3);
408 current_type = NULL;
409 }
410 ;
411
412 precedence_declarator:
413 "%left" { $$ = left_assoc; }
414 | "%right" { $$ = right_assoc; }
415 | "%nonassoc" { $$ = non_assoc; }
416 ;
417
418 type.opt:
419 /* Nothing. */ { current_type = NULL; }
420 | TYPE { current_type = $1; tag_seen = true; }
421 ;
422
423 /* One or more symbols to be %typed. */
424 symbols.1:
425 symbol
426 { $$ = symbol_list_sym_new ($1, @1); }
427 | symbols.1 symbol
428 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
429 ;
430
431 generic_symlist:
432 generic_symlist_item { $$ = $1; }
433 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
434 ;
435
436 generic_symlist_item:
437 symbol { $$ = symbol_list_sym_new ($1, @1); }
438 | TYPE { $$ = symbol_list_type_new ($1, @1); }
439 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
440 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
441 ;
442
443 /* One token definition. */
444 symbol_def:
445 TYPE
446 {
447 current_type = $1;
448 tag_seen = true;
449 }
450 | id
451 {
452 symbol_class_set ($1, current_class, @1, true);
453 symbol_type_set ($1, current_type, @1);
454 }
455 | id INT
456 {
457 symbol_class_set ($1, current_class, @1, true);
458 symbol_type_set ($1, current_type, @1);
459 symbol_user_token_number_set ($1, $2, @2);
460 }
461 | id string_as_id
462 {
463 symbol_class_set ($1, current_class, @1, true);
464 symbol_type_set ($1, current_type, @1);
465 symbol_make_alias ($1, $2, @$);
466 }
467 | id INT string_as_id
468 {
469 symbol_class_set ($1, current_class, @1, true);
470 symbol_type_set ($1, current_type, @1);
471 symbol_user_token_number_set ($1, $2, @2);
472 symbol_make_alias ($1, $3, @$);
473 }
474 ;
475
476 /* One or more symbol definitions. */
477 symbol_defs.1:
478 symbol_def
479 | symbol_defs.1 symbol_def
480 ;
481
482
483 /*------------------------------------------.
484 | The grammar section: between the two %%. |
485 `------------------------------------------*/
486
487 grammar:
488 rules_or_grammar_declaration
489 | grammar rules_or_grammar_declaration
490 ;
491
492 /* As a Bison extension, one can use the grammar declarations in the
493 body of the grammar. */
494 rules_or_grammar_declaration:
495 rules
496 | grammar_declaration ";"
497 | error ";"
498 {
499 yyerrok;
500 }
501 ;
502
503 rules:
504 id_colon { current_lhs = $1; current_lhs_location = @1; } rhses.1
505 ;
506
507 rhses.1:
508 rhs { grammar_current_rule_end (@1); }
509 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
510 | rhses.1 ";"
511 ;
512
513 rhs:
514 /* Nothing. */
515 { grammar_current_rule_begin (current_lhs, current_lhs_location); }
516 | rhs symbol
517 { grammar_current_rule_symbol_append ($2, @2); }
518 | rhs "{...}"
519 { grammar_current_rule_action_append ($2, @2); }
520 | rhs "%prec" symbol
521 { grammar_current_rule_prec_set ($3, @3); }
522 | rhs "%dprec" INT
523 { grammar_current_rule_dprec_set ($3, @3); }
524 | rhs "%merge" TYPE
525 { grammar_current_rule_merge_set ($3, @3); }
526 ;
527
528
529 /*----------------------------*
530 | variable and content.opt. |
531 *---------------------------*/
532
533 variable:
534 ID
535 | STRING { $$ = uniqstr_new ($1); } /* deprecated and not M4-friendly */
536 ;
537
538 /* Some content or empty by default. */
539 content.opt:
540 /* Nothing. */
541 {
542 $$ = "";
543 }
544 | STRING
545 ;
546
547
548 /*-------------*
549 | braceless. |
550 *-------------*/
551
552 braceless:
553 "{...}"
554 {
555 code_props plain_code;
556 $1[strlen ($1) - 1] = '\n';
557 code_props_plain_init (&plain_code, $1+1, @1);
558 code_props_translate_code (&plain_code);
559 gram_scanner_last_string_free ();
560 $$ = plain_code.code;
561 }
562 ;
563
564
565 /*---------------*
566 | Identifiers. |
567 *---------------*/
568
569 /* Identifiers are returned as uniqstr values by the scanner.
570 Depending on their use, we may need to make them genuine symbols. */
571
572 id:
573 ID
574 { $$ = symbol_from_uniqstr ($1, @1); }
575 | CHAR
576 {
577 $$ = symbol_get (char_name ($1), @1);
578 symbol_class_set ($$, token_sym, @1, false);
579 symbol_user_token_number_set ($$, $1, @1);
580 }
581 ;
582
583 id_colon:
584 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
585 ;
586
587
588 symbol:
589 id
590 | string_as_id
591 ;
592
593 /* A string used as an ID: quote it. */
594 string_as_id:
595 STRING
596 {
597 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
598 symbol_class_set ($$, token_sym, @1, false);
599 }
600 ;
601
602 epilogue.opt:
603 /* Nothing. */
604 | "%%" EPILOGUE
605 {
606 code_props plain_code;
607 code_props_plain_init (&plain_code, $2, @2);
608 code_props_translate_code (&plain_code);
609 gram_scanner_last_string_free ();
610 muscle_code_grow ("epilogue", plain_code.code, @2);
611 code_scanner_last_string_free ();
612 }
613 ;
614
615 %%
616
617
618 /* Return the location of the left-hand side of a rule whose
619 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
620 the right-hand side, and return an empty location equal to the end
621 boundary of RHS[0] if the right-hand side is empty. */
622
623 static YYLTYPE
624 lloc_default (YYLTYPE const *rhs, int n)
625 {
626 int i;
627 YYLTYPE loc;
628
629 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
630 The bug is fixed in 7.4.2m, but play it safe for now. */
631 loc.start = rhs[n].end;
632 loc.end = rhs[n].end;
633
634 /* Ignore empty nonterminals the start of the the right-hand side.
635 Do not bother to ignore them at the end of the right-hand side,
636 since empty nonterminals have the same end as their predecessors. */
637 for (i = 1; i <= n; i++)
638 if (! equal_boundaries (rhs[i].start, rhs[i].end))
639 {
640 loc.start = rhs[i].start;
641 break;
642 }
643
644 return loc;
645 }
646
647
648 /* Add a lex-param or a parse-param (depending on TYPE) with
649 declaration DECL and location LOC. */
650
651 static void
652 add_param (char const *type, char *decl, location loc)
653 {
654 static char const alphanum[26 + 26 + 1 + 10] =
655 "abcdefghijklmnopqrstuvwxyz"
656 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
657 "_"
658 "0123456789";
659 char const *name_start = NULL;
660 char *p;
661
662 /* Stop on last actual character. */
663 for (p = decl; p[1]; p++)
664 if ((p == decl
665 || ! memchr (alphanum, p[-1], sizeof alphanum))
666 && memchr (alphanum, p[0], sizeof alphanum - 10))
667 name_start = p;
668
669 /* Strip the surrounding '{' and '}', and any blanks just inside
670 the braces. */
671 while (*--p == ' ' || *p == '\t')
672 continue;
673 p[1] = '\0';
674 while (*++decl == ' ' || *decl == '\t')
675 continue;
676
677 if (! name_start)
678 complain_at (loc, _("missing identifier in parameter declaration"));
679 else
680 {
681 char *name;
682 size_t name_len;
683
684 for (name_len = 1;
685 memchr (alphanum, name_start[name_len], sizeof alphanum);
686 name_len++)
687 continue;
688
689 name = xmalloc (name_len + 1);
690 memcpy (name, name_start, name_len);
691 name[name_len] = '\0';
692 muscle_pair_list_grow (type, decl, name);
693 free (name);
694 }
695
696 gram_scanner_last_string_free ();
697 }
698
699
700 static void
701 version_check (location const *loc, char const *version)
702 {
703 if (strverscmp (version, PACKAGE_VERSION) > 0)
704 {
705 complain_at (*loc, "require bison %s, but have %s",
706 version, PACKAGE_VERSION);
707 exit (63);
708 }
709 }
710
711 static void
712 gram_error (location const *loc, char const *msg)
713 {
714 complain_at (*loc, "%s", msg);
715 }
716
717 char const *
718 token_name (int type)
719 {
720 return yytname[YYTRANSLATE (type)];
721 }
722
723 static char const *
724 char_name (char c)
725 {
726 if (c == '\'')
727 return "'\\''";
728 else
729 {
730 char buf[4];
731 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
732 return quotearg_style (escape_quoting_style, buf);
733 }
734 }