]> git.saurik.com Git - bison.git/blob - src/parse-gram.y
802347fa628ddaa32232c85a1085743d1f7dcbb0
[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 char const name_prefix[] = "percent_define_";
239 char *name = xmalloc (sizeof name_prefix + strlen ($2));
240 strcpy (name, name_prefix);
241 strcpy (name + sizeof name_prefix - 1, $2);
242 if (muscle_find_const (name))
243 warn_at (@2, _("%s `%s' redefined"), "%define variable", $2);
244 MUSCLE_INSERT_STRING (uniqstr_new (name), $3);
245 free (name);
246 muscle_grow_user_name_list ("user_percent_define_variables", $2, @2);
247 }
248 | "%defines" { defines_flag = true; }
249 | "%defines" STRING
250 {
251 defines_flag = true;
252 spec_defines_file = xstrdup ($2);
253 }
254 | "%error-verbose" { error_verbose = true; }
255 | "%expect" INT { expected_sr_conflicts = $2; }
256 | "%expect-rr" INT { expected_rr_conflicts = $2; }
257 | "%file-prefix" STRING { spec_file_prefix = $2; }
258 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
259 | "%glr-parser"
260 {
261 nondeterministic_parser = true;
262 glr_parser = true;
263 }
264 | "%initial-action" "{...}"
265 {
266 code_props action;
267 code_props_symbol_action_init (&action, $2, @2);
268 code_props_translate_code (&action);
269 gram_scanner_last_string_free ();
270 muscle_code_grow ("initial_action", action.code, @2);
271 code_scanner_last_string_free ();
272 }
273 | "%language" STRING { language_argmatch ($2, 1, &@1); }
274 | "%lex-param" "{...}" { add_param ("lex_param", $2, @2); }
275 | "%locations" { locations_flag = true; }
276 | "%name-prefix" STRING { spec_name_prefix = $2; }
277 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
278 | "%no-lines" { no_lines_flag = true; }
279 | "%nondeterministic-parser" { nondeterministic_parser = true; }
280 | "%output" STRING { spec_outfile = $2; }
281 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
282 | "%parse-param" "{...}" { add_param ("parse_param", $2, @2); }
283 | "%pure-parser" { pure_parser = true; }
284 | "%push-parser" { push_parser = true; pull_parser = false; }
285 | "%push-pull-parser" { push_parser = true; pull_parser = true; }
286 | "%require" STRING { version_check (&@2, $2); }
287 | "%skeleton" STRING { skeleton_arg ($2, 1, &@1); }
288 | "%token-table" { token_table_flag = true; }
289 | "%verbose" { report_flag = report_states; }
290 | "%yacc" { yacc_flag = true; }
291 | /*FIXME: Err? What is this horror doing here? */ ";"
292 ;
293
294 grammar_declaration:
295 precedence_declaration
296 | symbol_declaration
297 | "%start" symbol
298 {
299 grammar_start_symbol_set ($2, @2);
300 }
301 | "%destructor" "{...}" generic_symlist
302 {
303 symbol_list *list;
304 for (list = $3; list; list = list->next)
305 symbol_list_destructor_set (list, $2, @2);
306 symbol_list_free ($3);
307 }
308 | "%printer" "{...}" generic_symlist
309 {
310 symbol_list *list;
311 for (list = $3; list; list = list->next)
312 symbol_list_printer_set (list, $2, @2);
313 symbol_list_free ($3);
314 }
315 | "%default-prec"
316 {
317 default_prec = true;
318 }
319 | "%no-default-prec"
320 {
321 default_prec = false;
322 }
323 | "%code" braceless
324 {
325 muscle_code_grow ("percent_code", $2, @2);
326 code_scanner_last_string_free ();
327 }
328 | "%code" ID braceless
329 {
330 char const name_prefix[] = "percent_code_";
331 char *name = xmalloc (sizeof name_prefix + strlen ($2));
332 strcpy (name, name_prefix);
333 strcpy (name + sizeof name_prefix - 1, $2);
334 muscle_code_grow (uniqstr_new (name), $3, @3);
335 free (name);
336 code_scanner_last_string_free ();
337 muscle_grow_user_name_list ("user_percent_code_qualifiers", $2, @2);
338 }
339 ;
340
341
342 /*----------*
343 | %union. |
344 *----------*/
345
346 %token PERCENT_UNION "%union";
347
348 union_name:
349 /* Nothing. */ {}
350 | ID { muscle_code_grow ("union_name", $1, @1); }
351 ;
352
353 grammar_declaration:
354 "%union" union_name braceless
355 {
356 union_seen = true;
357 muscle_code_grow ("stype", $3, @3);
358 code_scanner_last_string_free ();
359 }
360 ;
361
362
363
364
365 symbol_declaration:
366 "%nterm" { current_class = nterm_sym; } symbol_defs.1
367 {
368 current_class = unknown_sym;
369 current_type = NULL;
370 }
371 | "%token" { current_class = token_sym; } symbol_defs.1
372 {
373 current_class = unknown_sym;
374 current_type = NULL;
375 }
376 | "%type" TYPE symbols.1
377 {
378 symbol_list *list;
379 tag_seen = true;
380 for (list = $3; list; list = list->next)
381 symbol_type_set (list->content.sym, $2, @2);
382 symbol_list_free ($3);
383 }
384 ;
385
386 precedence_declaration:
387 precedence_declarator type.opt symbols.1
388 {
389 symbol_list *list;
390 ++current_prec;
391 for (list = $3; list; list = list->next)
392 {
393 symbol_type_set (list->content.sym, current_type, @2);
394 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
395 }
396 symbol_list_free ($3);
397 current_type = NULL;
398 }
399 ;
400
401 precedence_declarator:
402 "%left" { $$ = left_assoc; }
403 | "%right" { $$ = right_assoc; }
404 | "%nonassoc" { $$ = non_assoc; }
405 ;
406
407 type.opt:
408 /* Nothing. */ { current_type = NULL; }
409 | TYPE { current_type = $1; tag_seen = true; }
410 ;
411
412 /* One or more symbols to be %typed. */
413 symbols.1:
414 symbol
415 { $$ = symbol_list_sym_new ($1, @1); }
416 | symbols.1 symbol
417 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
418 ;
419
420 generic_symlist:
421 generic_symlist_item { $$ = $1; }
422 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
423 ;
424
425 generic_symlist_item:
426 symbol { $$ = symbol_list_sym_new ($1, @1); }
427 | TYPE { $$ = symbol_list_type_new ($1, @1); }
428 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
429 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
430 ;
431
432 /* One token definition. */
433 symbol_def:
434 TYPE
435 {
436 current_type = $1;
437 tag_seen = true;
438 }
439 | id
440 {
441 symbol_class_set ($1, current_class, @1, true);
442 symbol_type_set ($1, current_type, @1);
443 }
444 | id INT
445 {
446 symbol_class_set ($1, current_class, @1, true);
447 symbol_type_set ($1, current_type, @1);
448 symbol_user_token_number_set ($1, $2, @2);
449 }
450 | id string_as_id
451 {
452 symbol_class_set ($1, current_class, @1, true);
453 symbol_type_set ($1, current_type, @1);
454 symbol_make_alias ($1, $2, @$);
455 }
456 | id INT string_as_id
457 {
458 symbol_class_set ($1, current_class, @1, true);
459 symbol_type_set ($1, current_type, @1);
460 symbol_user_token_number_set ($1, $2, @2);
461 symbol_make_alias ($1, $3, @$);
462 }
463 ;
464
465 /* One or more symbol definitions. */
466 symbol_defs.1:
467 symbol_def
468 | symbol_defs.1 symbol_def
469 ;
470
471
472 /*------------------------------------------.
473 | The grammar section: between the two %%. |
474 `------------------------------------------*/
475
476 grammar:
477 rules_or_grammar_declaration
478 | grammar rules_or_grammar_declaration
479 ;
480
481 /* As a Bison extension, one can use the grammar declarations in the
482 body of the grammar. */
483 rules_or_grammar_declaration:
484 rules
485 | grammar_declaration ";"
486 | error ";"
487 {
488 yyerrok;
489 }
490 ;
491
492 rules:
493 id_colon { current_lhs = $1; current_lhs_location = @1; } rhses.1
494 ;
495
496 rhses.1:
497 rhs { grammar_current_rule_end (@1); }
498 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
499 | rhses.1 ";"
500 ;
501
502 rhs:
503 /* Nothing. */
504 { grammar_current_rule_begin (current_lhs, current_lhs_location); }
505 | rhs symbol
506 { grammar_current_rule_symbol_append ($2, @2); }
507 | rhs "{...}"
508 { grammar_current_rule_action_append ($2, @2); }
509 | rhs "%prec" symbol
510 { grammar_current_rule_prec_set ($3, @3); }
511 | rhs "%dprec" INT
512 { grammar_current_rule_dprec_set ($3, @3); }
513 | rhs "%merge" TYPE
514 { grammar_current_rule_merge_set ($3, @3); }
515 ;
516
517
518 /*----------------------------*
519 | variable and content.opt. |
520 *---------------------------*/
521
522 variable:
523 ID
524 | STRING { $$ = uniqstr_new ($1); } /* deprecated and not M4-friendly */
525 ;
526
527 /* Some content or "1" by default. */
528 content.opt:
529 /* Nothing. */
530 {
531 static char one[] = "1";
532 $$ = one;
533 }
534 | STRING
535 ;
536
537
538 /*-------------*
539 | braceless. |
540 *-------------*/
541
542 braceless:
543 "{...}"
544 {
545 code_props plain_code;
546 $1[strlen ($1) - 1] = '\n';
547 code_props_plain_init (&plain_code, $1+1, @1);
548 code_props_translate_code (&plain_code);
549 gram_scanner_last_string_free ();
550 $$ = plain_code.code;
551 }
552 ;
553
554
555 /*---------------*
556 | Identifiers. |
557 *---------------*/
558
559 /* Identifiers are returned as uniqstr values by the scanner.
560 Depending on their use, we may need to make them genuine symbols. */
561
562 id:
563 ID
564 { $$ = symbol_from_uniqstr ($1, @1); }
565 | CHAR
566 {
567 $$ = symbol_get (char_name ($1), @1);
568 symbol_class_set ($$, token_sym, @1, false);
569 symbol_user_token_number_set ($$, $1, @1);
570 }
571 ;
572
573 id_colon:
574 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
575 ;
576
577
578 symbol:
579 id
580 | string_as_id
581 ;
582
583 /* A string used as an ID: quote it. */
584 string_as_id:
585 STRING
586 {
587 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
588 symbol_class_set ($$, token_sym, @1, false);
589 }
590 ;
591
592 epilogue.opt:
593 /* Nothing. */
594 | "%%" EPILOGUE
595 {
596 code_props plain_code;
597 code_props_plain_init (&plain_code, $2, @2);
598 code_props_translate_code (&plain_code);
599 gram_scanner_last_string_free ();
600 muscle_code_grow ("epilogue", plain_code.code, @2);
601 code_scanner_last_string_free ();
602 }
603 ;
604
605 %%
606
607
608 /* Return the location of the left-hand side of a rule whose
609 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
610 the right-hand side, and return an empty location equal to the end
611 boundary of RHS[0] if the right-hand side is empty. */
612
613 static YYLTYPE
614 lloc_default (YYLTYPE const *rhs, int n)
615 {
616 int i;
617 YYLTYPE loc;
618
619 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
620 The bug is fixed in 7.4.2m, but play it safe for now. */
621 loc.start = rhs[n].end;
622 loc.end = rhs[n].end;
623
624 /* Ignore empty nonterminals the start of the the right-hand side.
625 Do not bother to ignore them at the end of the right-hand side,
626 since empty nonterminals have the same end as their predecessors. */
627 for (i = 1; i <= n; i++)
628 if (! equal_boundaries (rhs[i].start, rhs[i].end))
629 {
630 loc.start = rhs[i].start;
631 break;
632 }
633
634 return loc;
635 }
636
637
638 /* Add a lex-param or a parse-param (depending on TYPE) with
639 declaration DECL and location LOC. */
640
641 static void
642 add_param (char const *type, char *decl, location loc)
643 {
644 static char const alphanum[26 + 26 + 1 + 10] =
645 "abcdefghijklmnopqrstuvwxyz"
646 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
647 "_"
648 "0123456789";
649 char const *name_start = NULL;
650 char *p;
651
652 /* Stop on last actual character. */
653 for (p = decl; p[1]; p++)
654 if ((p == decl
655 || ! memchr (alphanum, p[-1], sizeof alphanum))
656 && memchr (alphanum, p[0], sizeof alphanum - 10))
657 name_start = p;
658
659 /* Strip the surrounding '{' and '}', and any blanks just inside
660 the braces. */
661 while (*--p == ' ' || *p == '\t')
662 continue;
663 p[1] = '\0';
664 while (*++decl == ' ' || *decl == '\t')
665 continue;
666
667 if (! name_start)
668 complain_at (loc, _("missing identifier in parameter declaration"));
669 else
670 {
671 char *name;
672 size_t name_len;
673
674 for (name_len = 1;
675 memchr (alphanum, name_start[name_len], sizeof alphanum);
676 name_len++)
677 continue;
678
679 name = xmalloc (name_len + 1);
680 memcpy (name, name_start, name_len);
681 name[name_len] = '\0';
682 muscle_pair_list_grow (type, decl, name);
683 free (name);
684 }
685
686 gram_scanner_last_string_free ();
687 }
688
689
690 static void
691 version_check (location const *loc, char const *version)
692 {
693 if (strverscmp (version, PACKAGE_VERSION) > 0)
694 {
695 complain_at (*loc, "require bison %s, but have %s",
696 version, PACKAGE_VERSION);
697 exit (63);
698 }
699 }
700
701 static void
702 gram_error (location const *loc, char const *msg)
703 {
704 complain_at (*loc, "%s", msg);
705 }
706
707 char const *
708 token_name (int type)
709 {
710 return yytname[YYTRANSLATE (type)];
711 }
712
713 static char const *
714 char_name (char c)
715 {
716 if (c == '\'')
717 return "'\\''";
718 else
719 {
720 char buf[4];
721 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
722 return quotearg_style (escape_quoting_style, buf);
723 }
724 }