1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free
4 Software Foundation, Inc.
6 This file is part of Bison, the GNU Compiler Compiler.
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 3 of the License, or
11 (at your option) any later version.
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.
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
25 #include "conflicts.h"
29 #include "muscle-tab.h"
30 #include "named-ref.h"
34 #include "scan-gram.h"
35 #include "scan-code.h"
37 #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
38 static YYLTYPE lloc_default (YYLTYPE const *, int);
40 #define YY_LOCATION_PRINT(File, Loc) \
41 location_print (File, Loc)
43 static void version_check (location const *loc, char const *version);
45 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
46 FIXME: depends on the undocumented availability of YYLLOC. */
48 #define yyerror(Msg) \
49 gram_error (&yylloc, Msg)
50 static void gram_error (location const *, char const *);
52 static char const *char_name (char);
54 /** Add a lex-param or a parse-param.
56 * \param type \a lex_param or \a parse_param
57 * \param decl the formal argument
58 * \param loc the location in the source.
60 static void add_param (char const *type, char *decl, location loc);
63 static symbol_class current_class = unknown_sym;
64 static uniqstr current_type = NULL;
65 static symbol *current_lhs;
66 static location current_lhs_location;
67 static named_ref *current_lhs_named_ref;
68 static int current_prec = 0;
70 #define YYTYPE_INT16 int_fast16_t
71 #define YYTYPE_INT8 int_fast8_t
72 #define YYTYPE_UINT16 uint_fast16_t
73 #define YYTYPE_UINT8 uint_fast8_t
87 /* Bison's grammar can initial empty locations, hence a default
88 location is needed. */
89 boundary_set (&@$.start, current_file, 1, 1);
90 boundary_set (&@$.end, current_file, 1, 1);
102 unsigned char character;
103 named_ref *named_ref;
106 /* Define the tokens together with their human representation. */
107 %token GRAM_EOF 0 "end of file"
108 %token STRING "string"
111 %token PERCENT_TOKEN "%token"
112 %token PERCENT_NTERM "%nterm"
114 %token PERCENT_TYPE "%type"
115 %token PERCENT_DESTRUCTOR "%destructor"
116 %token PERCENT_PRINTER "%printer"
118 %token PERCENT_LEFT "%left"
119 %token PERCENT_RIGHT "%right"
120 %token PERCENT_NONASSOC "%nonassoc"
122 %token PERCENT_PREC "%prec"
123 %token PERCENT_DPREC "%dprec"
124 %token PERCENT_MERGE "%merge"
127 /*----------------------.
128 | Global Declarations. |
129 `----------------------*/
133 PERCENT_DEBUG "%debug"
134 PERCENT_DEFAULT_PREC "%default-prec"
135 PERCENT_DEFINE "%define"
136 PERCENT_DEFINES "%defines"
137 PERCENT_ERROR_VERBOSE "%error-verbose"
138 PERCENT_EXPECT "%expect"
139 PERCENT_EXPECT_RR "%expect-rr"
140 PERCENT_FILE_PREFIX "%file-prefix"
141 PERCENT_GLR_PARSER "%glr-parser"
142 PERCENT_INITIAL_ACTION "%initial-action"
143 PERCENT_LANGUAGE "%language"
144 PERCENT_LEX_PARAM "%lex-param"
145 PERCENT_LOCATIONS "%locations"
146 PERCENT_NAME_PREFIX "%name-prefix"
147 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
148 PERCENT_NO_LINES "%no-lines"
149 PERCENT_NONDETERMINISTIC_PARSER
150 "%nondeterministic-parser"
151 PERCENT_OUTPUT "%output"
152 PERCENT_PARSE_PARAM "%parse-param"
153 PERCENT_PURE_PARSER "%pure-parser"
154 PERCENT_REQUIRE "%require"
155 PERCENT_SKELETON "%skeleton"
156 PERCENT_START "%start"
157 PERCENT_TOKEN_TABLE "%token-table"
158 PERCENT_VERBOSE "%verbose"
162 %token BRACED_CODE "{...}"
163 %token BRACKETED_ID "[identifier]"
165 %token EPILOGUE "epilogue"
167 %token ID "identifier"
168 %token ID_COLON "identifier:"
169 %token PERCENT_PERCENT "%%"
171 %token PROLOGUE "%{...%}"
174 %token TYPE_TAG_ANY "<*>"
175 %token TYPE_TAG_NONE "<>"
177 %type <character> CHAR
178 %printer { fputs (char_name ($$), stderr); } CHAR
180 /* braceless is not to be used for rule or symbol actions, as it
181 calls code_props_plain_init. */
182 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
184 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
186 %printer { fprintf (stderr, "{\n%s\n}", $$); }
187 braceless content.opt "{...}" "%{...%}" EPILOGUE
189 %type <uniqstr> BRACKETED_ID ID ID_COLON TYPE variable
190 %printer { fputs ($$, stderr); } <uniqstr>
191 %printer { fprintf (stderr, "[%s]", $$); } BRACKETED_ID
192 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
193 %printer { fprintf (stderr, "<%s>", $$); } TYPE
196 %printer { fprintf (stderr, "%d", $$); } <integer>
198 %type <symbol> id id_colon string_as_id symbol symbol.prec
199 %printer { fprintf (stderr, "%s", $$->tag); } <symbol>
200 %printer { fprintf (stderr, "%s:", $$->tag); } id_colon
202 %type <assoc> precedence_declarator
203 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
204 %type <named_ref> named_ref.opt
209 prologue_declarations "%%" grammar epilogue.opt
213 /*------------------------------------.
214 | Declarations: before the first %%. |
215 `------------------------------------*/
217 prologue_declarations:
219 | prologue_declarations prologue_declaration
222 prologue_declaration:
226 code_props plain_code;
227 code_props_plain_init (&plain_code, $1, @1);
228 code_props_translate_code (&plain_code);
229 gram_scanner_last_string_free ();
230 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
231 plain_code.code, @1);
232 code_scanner_last_string_free ();
234 | "%debug" { debug_flag = true; }
235 | "%define" variable content.opt
237 muscle_percent_define_insert ($2, @2, $3,
238 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
240 | "%defines" { defines_flag = true; }
244 spec_defines_file = xstrdup ($2);
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 */
253 nondeterministic_parser = true;
256 | "%initial-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 ();
265 | "%language" STRING { language_argmatch ($2, grammar_prio, @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); }
277 /* %pure-parser is deprecated in favor of `%define api.pure', so use
278 `%define api.pure' in a backward-compatible manner here. First, don't
279 complain if %pure-parser is specified multiple times. */
280 if (!muscle_find_const ("percent_define(api.pure)"))
281 muscle_percent_define_insert ("api.pure", @1, "",
282 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
283 /* In all cases, use api.pure now so that the backend doesn't complain if
284 the skeleton ignores api.pure, but do warn now if there's a previous
285 conflicting definition from an actual %define. */
286 if (!muscle_percent_define_flag_if ("api.pure"))
287 muscle_percent_define_insert ("api.pure", @1, "",
288 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
290 | "%require" STRING { version_check (&@2, $2); }
293 char const *skeleton_user = $2;
294 if (strchr (skeleton_user, '/'))
296 size_t dir_length = strlen (current_file);
297 char *skeleton_build;
298 while (dir_length && current_file[dir_length - 1] != '/')
300 while (dir_length && current_file[dir_length - 1] == '/')
303 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
306 strncpy (skeleton_build, current_file, dir_length);
307 skeleton_build[dir_length++] = '/';
309 strcpy (skeleton_build + dir_length, skeleton_user);
310 skeleton_user = uniqstr_new (skeleton_build);
311 free (skeleton_build);
313 skeleton_arg (skeleton_user, grammar_prio, @1);
315 | "%token-table" { token_table_flag = true; }
316 | "%verbose" { report_flag |= report_states; }
317 | "%yacc" { yacc_flag = true; }
318 | /*FIXME: Err? What is this horror doing here? */ ";"
322 precedence_declaration
326 grammar_start_symbol_set ($2, @2);
328 | "%destructor" "{...}" generic_symlist
331 for (list = $3; list; list = list->next)
332 symbol_list_destructor_set (list, $2, @2);
333 symbol_list_free ($3);
335 | "%printer" "{...}" generic_symlist
338 for (list = $3; list; list = list->next)
339 symbol_list_printer_set (list, $2, @2);
340 symbol_list_free ($3);
348 default_prec = false;
352 /* Do not invoke muscle_percent_code_grow here since it invokes
353 muscle_user_name_list_grow. */
354 muscle_code_grow ("percent_code()", $2, @2);
355 code_scanner_last_string_free ();
357 | "%code" ID braceless
359 muscle_percent_code_grow ($2, @2, $3, @3);
360 code_scanner_last_string_free ();
369 %token PERCENT_UNION "%union";
373 | ID { muscle_code_grow ("union_name", $1, @1); }
377 "%union" union_name braceless
380 muscle_code_grow ("stype", $3, @3);
381 code_scanner_last_string_free ();
389 "%nterm" { current_class = nterm_sym; } symbol_defs.1
391 current_class = unknown_sym;
394 | "%token" { current_class = token_sym; } symbol_defs.1
396 current_class = unknown_sym;
399 | "%type" TYPE symbols.1
403 for (list = $3; list; list = list->next)
404 symbol_type_set (list->content.sym, $2, @2);
405 symbol_list_free ($3);
409 precedence_declaration:
410 precedence_declarator type.opt symbols.prec
414 for (list = $3; list; list = list->next)
416 symbol_type_set (list->content.sym, current_type, @2);
417 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
419 symbol_list_free ($3);
424 precedence_declarator:
425 "%left" { $$ = left_assoc; }
426 | "%right" { $$ = right_assoc; }
427 | "%nonassoc" { $$ = non_assoc; }
431 /* Nothing. */ { current_type = NULL; }
432 | TYPE { current_type = $1; tag_seen = true; }
435 /* Just like symbols.1 but accept INT for the sake of POSIX. */
438 { $$ = symbol_list_sym_new ($1, @1); }
439 | symbols.prec symbol.prec
440 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
445 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
448 /* One or more symbols to be %typed. */
451 { $$ = symbol_list_sym_new ($1, @1); }
453 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
457 generic_symlist_item { $$ = $1; }
458 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
461 generic_symlist_item:
462 symbol { $$ = symbol_list_sym_new ($1, @1); }
463 | TYPE { $$ = symbol_list_type_new ($1, @1); }
464 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
465 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
468 /* One token definition. */
477 symbol_class_set ($1, current_class, @1, true);
478 symbol_type_set ($1, current_type, @1);
482 symbol_class_set ($1, current_class, @1, true);
483 symbol_type_set ($1, current_type, @1);
484 symbol_user_token_number_set ($1, $2, @2);
488 symbol_class_set ($1, current_class, @1, true);
489 symbol_type_set ($1, current_type, @1);
490 symbol_make_alias ($1, $2, @$);
492 | id INT string_as_id
494 symbol_class_set ($1, current_class, @1, true);
495 symbol_type_set ($1, current_type, @1);
496 symbol_user_token_number_set ($1, $2, @2);
497 symbol_make_alias ($1, $3, @$);
501 /* One or more symbol definitions. */
504 | symbol_defs.1 symbol_def
508 /*------------------------------------------.
509 | The grammar section: between the two %%. |
510 `------------------------------------------*/
513 rules_or_grammar_declaration
514 | grammar rules_or_grammar_declaration
517 /* As a Bison extension, one can use the grammar declarations in the
518 body of the grammar. */
519 rules_or_grammar_declaration:
521 | grammar_declaration ";"
529 id_colon named_ref.opt { current_lhs = $1; current_lhs_location = @1;
530 current_lhs_named_ref = $2; } rhses.1
534 rhs { grammar_current_rule_end (@1); }
535 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
541 { grammar_current_rule_begin (current_lhs, current_lhs_location,
542 current_lhs_named_ref); }
543 | rhs symbol named_ref.opt
544 { grammar_current_rule_symbol_append ($2, @2, $3); }
545 | rhs "{...}" named_ref.opt
546 { grammar_current_rule_action_append ($2, @2, $3); }
548 { grammar_current_rule_prec_set ($3, @3); }
550 { grammar_current_rule_dprec_set ($3, @3); }
552 { grammar_current_rule_merge_set ($3, @3); }
556 /* Nothing. */ { $$ = 0; }
558 BRACKETED_ID { $$ = named_ref_new($1, @1); }
562 /*----------------------------*
563 | variable and content.opt. |
564 *---------------------------*/
566 /* The STRING form of variable is deprecated and is not M4-friendly.
567 For example, M4 fails for `%define "[" "value"'. */
570 | STRING { $$ = uniqstr_new ($1); }
573 /* Some content or empty by default. */
575 /* Nothing. */ { $$ = ""; }
588 code_props plain_code;
589 $1[strlen ($1) - 1] = '\n';
590 code_props_plain_init (&plain_code, $1+1, @1);
591 code_props_translate_code (&plain_code);
592 gram_scanner_last_string_free ();
593 $$ = plain_code.code;
602 /* Identifiers are returned as uniqstr values by the scanner.
603 Depending on their use, we may need to make them genuine symbols. */
607 { $$ = symbol_from_uniqstr ($1, @1); }
610 $$ = symbol_get (char_name ($1), @1);
611 symbol_class_set ($$, token_sym, @1, false);
612 symbol_user_token_number_set ($$, $1, @1);
617 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
626 /* A string used as an ID: quote it. */
630 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
631 symbol_class_set ($$, token_sym, @1, false);
639 code_props plain_code;
640 code_props_plain_init (&plain_code, $2, @2);
641 code_props_translate_code (&plain_code);
642 gram_scanner_last_string_free ();
643 muscle_code_grow ("epilogue", plain_code.code, @2);
644 code_scanner_last_string_free ();
651 /* Return the location of the left-hand side of a rule whose
652 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
653 the right-hand side, and return an empty location equal to the end
654 boundary of RHS[0] if the right-hand side is empty. */
657 lloc_default (YYLTYPE const *rhs, int n)
662 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
663 The bug is fixed in 7.4.2m, but play it safe for now. */
664 loc.start = rhs[n].end;
665 loc.end = rhs[n].end;
667 /* Ignore empty nonterminals the start of the the right-hand side.
668 Do not bother to ignore them at the end of the right-hand side,
669 since empty nonterminals have the same end as their predecessors. */
670 for (i = 1; i <= n; i++)
671 if (! equal_boundaries (rhs[i].start, rhs[i].end))
673 loc.start = rhs[i].start;
681 /* Add a lex-param or a parse-param (depending on TYPE) with
682 declaration DECL and location LOC. */
685 add_param (char const *type, char *decl, location loc)
687 static char const alphanum[26 + 26 + 1 + 10] =
688 "abcdefghijklmnopqrstuvwxyz"
689 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
692 char const *name_start = NULL;
695 /* Stop on last actual character. */
696 for (p = decl; p[1]; p++)
698 || ! memchr (alphanum, p[-1], sizeof alphanum))
699 && memchr (alphanum, p[0], sizeof alphanum - 10))
702 /* Strip the surrounding '{' and '}', and any blanks just inside
704 while (*--p == ' ' || *p == '\t')
707 while (*++decl == ' ' || *decl == '\t')
711 complain_at (loc, _("missing identifier in parameter declaration"));
718 memchr (alphanum, name_start[name_len], sizeof alphanum);
722 name = xmalloc (name_len + 1);
723 memcpy (name, name_start, name_len);
724 name[name_len] = '\0';
725 muscle_pair_list_grow (type, decl, name);
729 gram_scanner_last_string_free ();
734 version_check (location const *loc, char const *version)
736 if (strverscmp (version, PACKAGE_VERSION) > 0)
738 complain_at (*loc, "require bison %s, but have %s",
739 version, PACKAGE_VERSION);
745 gram_error (location const *loc, char const *msg)
747 complain_at (*loc, "%s", msg);
751 token_name (int type)
753 return yytname[YYTRANSLATE (type)];
764 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
765 return quotearg_style (escape_quoting_style, buf);