1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software
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"
33 #include "scan-gram.h"
34 #include "scan-code.h"
36 #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
37 static YYLTYPE lloc_default (YYLTYPE const *, int);
39 #define YY_LOCATION_PRINT(File, Loc) \
40 location_print (File, Loc)
42 static void version_check (location const *loc, char const *version);
44 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
45 FIXME: depends on the undocumented availability of YYLLOC. */
47 #define yyerror(Msg) \
48 gram_error (&yylloc, Msg)
49 static void gram_error (location const *, char const *);
51 static char const *char_name (char);
53 /** Add a lex-param or a parse-param.
55 * \param type \a lex_param or \a parse_param
56 * \param decl the formal argument
57 * \param loc the location in the source.
59 static void add_param (char const *type, char *decl, location loc);
62 static symbol_class current_class = unknown_sym;
63 static uniqstr current_type = NULL;
64 static symbol *current_lhs;
65 static location current_lhs_location;
66 static int current_prec = 0;
68 #define YYTYPE_INT16 int_fast16_t
69 #define YYTYPE_INT8 int_fast8_t
70 #define YYTYPE_UINT16 uint_fast16_t
71 #define YYTYPE_UINT8 uint_fast8_t
85 /* Bison's grammar can initial empty locations, hence a default
86 location is needed. */
87 boundary_set (&@$.start, current_file, 1, 1);
88 boundary_set (&@$.end, current_file, 1, 1);
100 unsigned char character;
103 /* Define the tokens together with their human representation. */
104 %token GRAM_EOF 0 "end of file"
105 %token STRING "string"
108 %token PERCENT_TOKEN "%token"
109 %token PERCENT_NTERM "%nterm"
111 %token PERCENT_TYPE "%type"
112 %token PERCENT_DESTRUCTOR "%destructor"
113 %token PERCENT_PRINTER "%printer"
115 %token PERCENT_LEFT "%left"
116 %token PERCENT_RIGHT "%right"
117 %token PERCENT_NONASSOC "%nonassoc"
118 %token PERCENT_PRECEDENCE "%precedence"
120 %token PERCENT_PREC "%prec"
121 %token PERCENT_DPREC "%dprec"
122 %token PERCENT_MERGE "%merge"
125 /*----------------------.
126 | Global Declarations. |
127 `----------------------*/
131 PERCENT_DEFAULT_PREC "%default-prec"
132 PERCENT_DEFINE "%define"
133 PERCENT_DEFINES "%defines"
134 PERCENT_EXPECT "%expect"
135 PERCENT_EXPECT_RR "%expect-rr"
136 PERCENT_FLAG "%<flag>"
137 PERCENT_FILE_PREFIX "%file-prefix"
138 PERCENT_GLR_PARSER "%glr-parser"
139 PERCENT_INITIAL_ACTION "%initial-action"
140 PERCENT_LANGUAGE "%language"
141 PERCENT_LEX_PARAM "%lex-param"
142 PERCENT_LOCATIONS "%locations"
143 PERCENT_NAME_PREFIX "%name-prefix"
144 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
145 PERCENT_NO_LINES "%no-lines"
146 PERCENT_NONDETERMINISTIC_PARSER
147 "%nondeterministic-parser"
148 PERCENT_OUTPUT "%output"
149 PERCENT_PARSE_PARAM "%parse-param"
150 PERCENT_REQUIRE "%require"
151 PERCENT_SKELETON "%skeleton"
152 PERCENT_START "%start"
153 PERCENT_TOKEN_TABLE "%token-table"
154 PERCENT_VERBOSE "%verbose"
158 %token BRACED_CODE "{...}"
160 %token EPILOGUE "epilogue"
162 %token ID "identifier"
163 %token ID_COLON "identifier:"
164 %token PERCENT_PERCENT "%%"
166 %token PROLOGUE "%{...%}"
172 %type <character> CHAR
173 %printer { fputs (char_name ($$), stderr); } CHAR
175 /* braceless is not to be used for rule or symbol actions, as it
176 calls code_props_plain_init. */
177 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
179 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
181 %printer { fprintf (stderr, "{\n%s\n}", $$); }
182 braceless content.opt "{...}" "%{...%}" EPILOGUE
184 %type <uniqstr> TAG ID ID_COLON PERCENT_FLAG variable
185 %printer { fputs ($$, stderr); } ID variable
186 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
187 %printer { fprintf (stderr, "%%%s", $$); } PERCENT_FLAG
188 %printer { fprintf (stderr, "<%s>", $$); } TAG
191 %printer { fprintf (stderr, "%d", $$); } INT
193 %type <symbol> id id_colon symbol symbol.prec string_as_id
194 %printer { fprintf (stderr, "%s", $$->tag); } id symbol string_as_id
195 %printer { fprintf (stderr, "%s:", $$->tag); } id_colon
197 %type <assoc> precedence_declarator
198 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
202 prologue_declarations "%%" grammar epilogue.opt
206 /*------------------------------------.
207 | Declarations: before the first %%. |
208 `------------------------------------*/
210 prologue_declarations:
212 | prologue_declarations prologue_declaration
215 prologue_declaration:
219 code_props plain_code;
220 code_props_plain_init (&plain_code, $1, @1);
221 code_props_translate_code (&plain_code);
222 gram_scanner_last_string_free ();
223 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
224 plain_code.code, @1);
225 code_scanner_last_string_free ();
229 muscle_percent_define_ensure ($1, @1, true);
231 | "%define" variable content.opt
233 muscle_percent_define_insert ($2, @2, $3);
235 | "%defines" { defines_flag = true; }
239 spec_defines_file = xstrdup ($2);
241 | "%expect" INT { expected_sr_conflicts = $2; }
242 | "%expect-rr" INT { expected_rr_conflicts = $2; }
243 | "%file-prefix" STRING { spec_file_prefix = $2; }
244 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
247 nondeterministic_parser = true;
250 | "%initial-action" "{...}"
253 code_props_symbol_action_init (&action, $2, @2);
254 code_props_translate_code (&action);
255 gram_scanner_last_string_free ();
256 muscle_code_grow ("initial_action", action.code, @2);
257 code_scanner_last_string_free ();
259 | "%language" STRING { language_argmatch ($2, grammar_prio, @1); }
260 | "%lex-param" "{...}" { add_param ("lex_param", $2, @2); }
261 | "%locations" { locations_flag = true; }
262 | "%name-prefix" STRING { spec_name_prefix = $2; }
263 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
264 | "%no-lines" { no_lines_flag = true; }
265 | "%nondeterministic-parser" { nondeterministic_parser = true; }
266 | "%output" STRING { spec_outfile = $2; }
267 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
268 | "%parse-param" "{...}" { add_param ("parse_param", $2, @2); }
269 | "%require" STRING { version_check (&@2, $2); }
272 char const *skeleton_user = $2;
273 if (strchr (skeleton_user, '/'))
275 size_t dir_length = strlen (current_file);
276 char *skeleton_build;
277 while (dir_length && current_file[dir_length - 1] != '/')
279 while (dir_length && current_file[dir_length - 1] == '/')
282 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
285 strncpy (skeleton_build, current_file, dir_length);
286 skeleton_build[dir_length++] = '/';
288 strcpy (skeleton_build + dir_length, skeleton_user);
289 skeleton_user = uniqstr_new (skeleton_build);
290 free (skeleton_build);
292 skeleton_arg (skeleton_user, grammar_prio, @1);
294 | "%token-table" { token_table_flag = true; }
295 | "%verbose" { report_flag |= report_states; }
296 | "%yacc" { yacc_flag = true; }
297 | /*FIXME: Err? What is this horror doing here? */ ";"
301 precedence_declaration
305 grammar_start_symbol_set ($2, @2);
307 | "%destructor" "{...}" generic_symlist
310 for (list = $3; list; list = list->next)
311 symbol_list_destructor_set (list, $2, @2);
312 symbol_list_free ($3);
314 | "%printer" "{...}" generic_symlist
317 for (list = $3; list; list = list->next)
318 symbol_list_printer_set (list, $2, @2);
319 symbol_list_free ($3);
327 default_prec = false;
331 /* Do not invoke muscle_percent_code_grow here since it invokes
332 muscle_user_name_list_grow. */
333 muscle_code_grow ("percent_code()", $2, @2);
334 code_scanner_last_string_free ();
336 | "%code" ID braceless
338 muscle_percent_code_grow ($2, @2, $3, @3);
339 code_scanner_last_string_free ();
348 %token PERCENT_UNION "%union";
352 | ID { muscle_code_grow ("union_name", $1, @1); }
356 "%union" union_name braceless
359 muscle_code_grow ("stype", $3, @3);
360 code_scanner_last_string_free ();
368 "%nterm" { current_class = nterm_sym; } symbol_defs.1
370 current_class = unknown_sym;
373 | "%token" { current_class = token_sym; } symbol_defs.1
375 current_class = unknown_sym;
378 | "%type" TAG symbols.1
382 for (list = $3; list; list = list->next)
383 symbol_type_set (list->content.sym, $2, @2);
384 symbol_list_free ($3);
388 precedence_declaration:
389 precedence_declarator tag.opt symbols.prec
393 for (list = $3; list; list = list->next)
395 symbol_type_set (list->content.sym, current_type, @2);
396 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
398 symbol_list_free ($3);
403 precedence_declarator:
404 "%left" { $$ = left_assoc; }
405 | "%right" { $$ = right_assoc; }
406 | "%nonassoc" { $$ = non_assoc; }
407 | "%precedence" { $$ = precedence_assoc; }
411 /* Nothing. */ { current_type = NULL; }
412 | TAG { current_type = $1; tag_seen = true; }
415 /* Just like symbols.1 but accept INT for the sake of POSIX. */
418 { $$ = symbol_list_sym_new ($1, @1); }
419 | symbols.prec symbol.prec
420 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
425 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
428 /* One or more symbols to be %typed. */
431 { $$ = symbol_list_sym_new ($1, @1); }
433 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
437 generic_symlist_item { $$ = $1; }
438 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
441 generic_symlist_item:
442 symbol { $$ = symbol_list_sym_new ($1, @1); }
443 | TAG { $$ = symbol_list_type_new ($1, @1); }
444 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
445 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
448 /* One token definition. */
457 symbol_class_set ($1, current_class, @1, true);
458 symbol_type_set ($1, current_type, @1);
462 symbol_class_set ($1, current_class, @1, true);
463 symbol_type_set ($1, current_type, @1);
464 symbol_user_token_number_set ($1, $2, @2);
468 symbol_class_set ($1, current_class, @1, true);
469 symbol_type_set ($1, current_type, @1);
470 symbol_make_alias ($1, $2, @$);
472 | id INT string_as_id
474 symbol_class_set ($1, current_class, @1, true);
475 symbol_type_set ($1, current_type, @1);
476 symbol_user_token_number_set ($1, $2, @2);
477 symbol_make_alias ($1, $3, @$);
481 /* One or more symbol definitions. */
484 | symbol_defs.1 symbol_def
488 /*------------------------------------------.
489 | The grammar section: between the two %%. |
490 `------------------------------------------*/
493 rules_or_grammar_declaration
494 | grammar rules_or_grammar_declaration
497 /* As a Bison extension, one can use the grammar declarations in the
498 body of the grammar. */
499 rules_or_grammar_declaration:
501 | grammar_declaration ";"
509 id_colon { current_lhs = $1; current_lhs_location = @1; } rhses.1
513 rhs { grammar_current_rule_end (@1); }
514 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
520 { grammar_current_rule_begin (current_lhs, current_lhs_location); }
522 { grammar_current_rule_symbol_append ($2, @2); }
524 { grammar_current_rule_action_append ($2, @2); }
526 { grammar_current_rule_prec_set ($3, @3); }
528 { grammar_current_rule_dprec_set ($3, @3); }
530 { grammar_current_rule_merge_set ($3, @3); }
534 /*---------------------------.
535 | variable and content.opt. |
536 `---------------------------*/
540 | STRING { $$ = uniqstr_new ($1); } /* deprecated and not M4-friendly */
543 /* Some content or empty by default. */
560 code_props plain_code;
561 $1[strlen ($1) - 1] = '\n';
562 code_props_plain_init (&plain_code, $1+1, @1);
563 code_props_translate_code (&plain_code);
564 gram_scanner_last_string_free ();
565 $$ = plain_code.code;
574 /* Identifiers are returned as uniqstr values by the scanner.
575 Depending on their use, we may need to make them genuine symbols. */
579 { $$ = symbol_from_uniqstr ($1, @1); }
582 $$ = symbol_get (char_name ($1), @1);
583 symbol_class_set ($$, token_sym, @1, false);
584 symbol_user_token_number_set ($$, $1, @1);
589 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
598 /* A string used as an ID: quote it. */
602 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
603 symbol_class_set ($$, token_sym, @1, false);
611 code_props plain_code;
612 code_props_plain_init (&plain_code, $2, @2);
613 code_props_translate_code (&plain_code);
614 gram_scanner_last_string_free ();
615 muscle_code_grow ("epilogue", plain_code.code, @2);
616 code_scanner_last_string_free ();
623 /* Return the location of the left-hand side of a rule whose
624 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
625 the right-hand side, and return an empty location equal to the end
626 boundary of RHS[0] if the right-hand side is empty. */
629 lloc_default (YYLTYPE const *rhs, int n)
634 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
635 The bug is fixed in 7.4.2m, but play it safe for now. */
636 loc.start = rhs[n].end;
637 loc.end = rhs[n].end;
639 /* Ignore empty nonterminals the start of the the right-hand side.
640 Do not bother to ignore them at the end of the right-hand side,
641 since empty nonterminals have the same end as their predecessors. */
642 for (i = 1; i <= n; i++)
643 if (! equal_boundaries (rhs[i].start, rhs[i].end))
645 loc.start = rhs[i].start;
653 /* Add a lex-param or a parse-param (depending on TYPE) with
654 declaration DECL and location LOC. */
657 add_param (char const *type, char *decl, location loc)
659 static char const alphanum[26 + 26 + 1 + 10] =
660 "abcdefghijklmnopqrstuvwxyz"
661 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
664 char const *name_start = NULL;
667 /* Stop on last actual character. */
668 for (p = decl; p[1]; p++)
670 || ! memchr (alphanum, p[-1], sizeof alphanum))
671 && memchr (alphanum, p[0], sizeof alphanum - 10))
674 /* Strip the surrounding '{' and '}', and any blanks just inside
676 while (*--p == ' ' || *p == '\t')
679 while (*++decl == ' ' || *decl == '\t')
683 complain_at (loc, _("missing identifier in parameter declaration"));
690 memchr (alphanum, name_start[name_len], sizeof alphanum);
694 name = xmalloc (name_len + 1);
695 memcpy (name, name_start, name_len);
696 name[name_len] = '\0';
697 muscle_pair_list_grow (type, decl, name);
701 gram_scanner_last_string_free ();
706 version_check (location const *loc, char const *version)
708 if (strverscmp (version, PACKAGE_VERSION) > 0)
710 complain_at (*loc, "require bison %s, but have %s",
711 version, PACKAGE_VERSION);
717 gram_error (location const *loc, char const *msg)
719 complain_at (*loc, "%s", msg);
723 token_name (int type)
725 return yytname[YYTRANSLATE (type)];
736 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
737 return quotearg_style (escape_quoting_style, buf);