1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008 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"
119 %token PERCENT_PREC "%prec"
120 %token PERCENT_DPREC "%dprec"
121 %token PERCENT_MERGE "%merge"
124 /*----------------------.
125 | Global Declarations. |
126 `----------------------*/
130 PERCENT_DEBUG "%debug"
131 PERCENT_DEFAULT_PREC "%default-prec"
132 PERCENT_DEFINE "%define"
133 PERCENT_DEFINES "%defines"
134 PERCENT_ERROR_VERBOSE "%error-verbose"
135 PERCENT_EXPECT "%expect"
136 PERCENT_EXPECT_RR "%expect-rr"
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_PURE_PARSER "%pure-parser"
151 PERCENT_REQUIRE "%require"
152 PERCENT_SKELETON "%skeleton"
153 PERCENT_START "%start"
154 PERCENT_TOKEN_TABLE "%token-table"
155 PERCENT_VERBOSE "%verbose"
159 %token BRACED_CODE "{...}"
161 %token EPILOGUE "epilogue"
163 %token ID "identifier"
164 %token ID_COLON "identifier:"
165 %token PERCENT_PERCENT "%%"
167 %token PROLOGUE "%{...%}"
170 %token TYPE_TAG_ANY "<*>"
171 %token TYPE_TAG_NONE "<>"
173 %type <character> CHAR
174 %printer { fputs (char_name ($$), stderr); } CHAR
176 /* braceless is not to be used for rule or symbol actions, as it
177 calls code_props_plain_init. */
178 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
180 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
182 %printer { fprintf (stderr, "{\n%s\n}", $$); }
183 braceless content.opt "{...}" "%{...%}" EPILOGUE
185 %type <uniqstr> TYPE ID ID_COLON variable
186 %printer { fprintf (stderr, "<%s>", $$); } TYPE
187 %printer { fputs ($$, stderr); } ID variable
188 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
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 ();
227 | "%debug" { debug_flag = true; }
228 | "%define" variable content.opt
230 muscle_percent_define_insert ($2, @2, $3);
232 | "%defines" { defines_flag = true; }
236 spec_defines_file = xstrdup ($2);
238 | "%error-verbose" { error_verbose = true; }
239 | "%expect" INT { expected_sr_conflicts = $2; }
240 | "%expect-rr" INT { expected_rr_conflicts = $2; }
241 | "%file-prefix" STRING { spec_file_prefix = $2; }
242 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
245 nondeterministic_parser = true;
248 | "%initial-action" "{...}"
251 code_props_symbol_action_init (&action, $2, @2);
252 code_props_translate_code (&action);
253 gram_scanner_last_string_free ();
254 muscle_code_grow ("initial_action", action.code, @2);
255 code_scanner_last_string_free ();
257 | "%language" STRING { language_argmatch ($2, 1, @1); }
258 | "%lex-param" "{...}" { add_param ("lex_param", $2, @2); }
259 | "%locations" { locations_flag = true; }
260 | "%name-prefix" STRING { spec_name_prefix = $2; }
261 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
262 | "%no-lines" { no_lines_flag = true; }
263 | "%nondeterministic-parser" { nondeterministic_parser = true; }
264 | "%output" STRING { spec_outfile = $2; }
265 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
266 | "%parse-param" "{...}" { add_param ("parse_param", $2, @2); }
269 /* %pure-parser is deprecated in favor of `%define api.pure', so use
270 `%define api.pure' in a backward-compatible manner here. First, don't
271 complain if %pure-parser is specified multiple times. */
272 if (!muscle_find_const ("percent_define(api.pure)"))
273 muscle_percent_define_insert ("api.pure", @1, "");
274 /* In all cases, use api.pure now so that the backend doesn't complain if
275 the skeleton ignores api.pure, but do warn now if there's a previous
276 conflicting definition from an actual %define. */
277 if (!muscle_percent_define_flag_if ("api.pure"))
278 muscle_percent_define_insert ("api.pure", @1, "");
280 | "%require" STRING { version_check (&@2, $2); }
283 char const *skeleton_user = $2;
284 if (strchr (skeleton_user, '/'))
286 size_t dir_length = strlen (current_file);
287 char *skeleton_build;
288 while (dir_length && current_file[dir_length - 1] != '/')
290 while (dir_length && current_file[dir_length - 1] == '/')
293 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
296 strncpy (skeleton_build, current_file, dir_length);
297 skeleton_build[dir_length++] = '/';
299 strcpy (skeleton_build + dir_length, skeleton_user);
300 skeleton_user = uniqstr_new (skeleton_build);
301 free (skeleton_build);
303 skeleton_arg (skeleton_user, 1, @1);
305 | "%token-table" { token_table_flag = true; }
306 | "%verbose" { report_flag |= report_states; }
307 | "%yacc" { yacc_flag = true; }
308 | /*FIXME: Err? What is this horror doing here? */ ";"
312 precedence_declaration
316 grammar_start_symbol_set ($2, @2);
318 | "%destructor" "{...}" generic_symlist
321 for (list = $3; list; list = list->next)
322 symbol_list_destructor_set (list, $2, @2);
323 symbol_list_free ($3);
325 | "%printer" "{...}" generic_symlist
328 for (list = $3; list; list = list->next)
329 symbol_list_printer_set (list, $2, @2);
330 symbol_list_free ($3);
338 default_prec = false;
342 /* Do not invoke muscle_percent_code_grow here since it invokes
343 muscle_user_name_list_grow. */
344 muscle_code_grow ("percent_code()", $2, @2);
345 code_scanner_last_string_free ();
347 | "%code" ID braceless
349 muscle_percent_code_grow ($2, @2, $3, @3);
350 code_scanner_last_string_free ();
359 %token PERCENT_UNION "%union";
363 | ID { muscle_code_grow ("union_name", $1, @1); }
367 "%union" union_name braceless
370 muscle_code_grow ("stype", $3, @3);
371 code_scanner_last_string_free ();
379 "%nterm" { current_class = nterm_sym; } symbol_defs.1
381 current_class = unknown_sym;
384 | "%token" { current_class = token_sym; } symbol_defs.1
386 current_class = unknown_sym;
389 | "%type" TYPE symbols.1
393 for (list = $3; list; list = list->next)
394 symbol_type_set (list->content.sym, $2, @2);
395 symbol_list_free ($3);
399 precedence_declaration:
400 precedence_declarator type.opt symbols.prec
404 for (list = $3; list; list = list->next)
406 symbol_type_set (list->content.sym, current_type, @2);
407 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
409 symbol_list_free ($3);
414 precedence_declarator:
415 "%left" { $$ = left_assoc; }
416 | "%right" { $$ = right_assoc; }
417 | "%nonassoc" { $$ = non_assoc; }
421 /* Nothing. */ { current_type = NULL; }
422 | TYPE { current_type = $1; tag_seen = true; }
425 /* Just like symbols.1 but accept INT for the sake of POSIX. */
428 { $$ = symbol_list_sym_new ($1, @1); }
429 | symbols.prec symbol.prec
430 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
435 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
438 /* One or more symbols to be %typed. */
441 { $$ = symbol_list_sym_new ($1, @1); }
443 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
447 generic_symlist_item { $$ = $1; }
448 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
451 generic_symlist_item:
452 symbol { $$ = symbol_list_sym_new ($1, @1); }
453 | TYPE { $$ = symbol_list_type_new ($1, @1); }
454 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
455 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
458 /* One token definition. */
467 symbol_class_set ($1, current_class, @1, true);
468 symbol_type_set ($1, current_type, @1);
472 symbol_class_set ($1, current_class, @1, true);
473 symbol_type_set ($1, current_type, @1);
474 symbol_user_token_number_set ($1, $2, @2);
478 symbol_class_set ($1, current_class, @1, true);
479 symbol_type_set ($1, current_type, @1);
480 symbol_make_alias ($1, $2, @$);
482 | id INT string_as_id
484 symbol_class_set ($1, current_class, @1, true);
485 symbol_type_set ($1, current_type, @1);
486 symbol_user_token_number_set ($1, $2, @2);
487 symbol_make_alias ($1, $3, @$);
491 /* One or more symbol definitions. */
494 | symbol_defs.1 symbol_def
498 /*------------------------------------------.
499 | The grammar section: between the two %%. |
500 `------------------------------------------*/
503 rules_or_grammar_declaration
504 | grammar rules_or_grammar_declaration
507 /* As a Bison extension, one can use the grammar declarations in the
508 body of the grammar. */
509 rules_or_grammar_declaration:
511 | grammar_declaration ";"
519 id_colon { current_lhs = $1; current_lhs_location = @1; } rhses.1
523 rhs { grammar_current_rule_end (@1); }
524 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
530 { grammar_current_rule_begin (current_lhs, current_lhs_location); }
532 { grammar_current_rule_symbol_append ($2, @2); }
534 { grammar_current_rule_action_append ($2, @2); }
536 { grammar_current_rule_prec_set ($3, @3); }
538 { grammar_current_rule_dprec_set ($3, @3); }
540 { grammar_current_rule_merge_set ($3, @3); }
544 /*---------------------------.
545 | variable and content.opt. |
546 `---------------------------*/
550 | STRING { $$ = uniqstr_new ($1); } /* deprecated and not M4-friendly */
553 /* Some content or empty by default. */
570 code_props plain_code;
571 $1[strlen ($1) - 1] = '\n';
572 code_props_plain_init (&plain_code, $1+1, @1);
573 code_props_translate_code (&plain_code);
574 gram_scanner_last_string_free ();
575 $$ = plain_code.code;
584 /* Identifiers are returned as uniqstr values by the scanner.
585 Depending on their use, we may need to make them genuine symbols. */
589 { $$ = symbol_from_uniqstr ($1, @1); }
592 $$ = symbol_get (char_name ($1), @1);
593 symbol_class_set ($$, token_sym, @1, false);
594 symbol_user_token_number_set ($$, $1, @1);
599 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
608 /* A string used as an ID: quote it. */
612 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
613 symbol_class_set ($$, token_sym, @1, false);
621 code_props plain_code;
622 code_props_plain_init (&plain_code, $2, @2);
623 code_props_translate_code (&plain_code);
624 gram_scanner_last_string_free ();
625 muscle_code_grow ("epilogue", plain_code.code, @2);
626 code_scanner_last_string_free ();
633 /* Return the location of the left-hand side of a rule whose
634 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
635 the right-hand side, and return an empty location equal to the end
636 boundary of RHS[0] if the right-hand side is empty. */
639 lloc_default (YYLTYPE const *rhs, int n)
644 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
645 The bug is fixed in 7.4.2m, but play it safe for now. */
646 loc.start = rhs[n].end;
647 loc.end = rhs[n].end;
649 /* Ignore empty nonterminals the start of the the right-hand side.
650 Do not bother to ignore them at the end of the right-hand side,
651 since empty nonterminals have the same end as their predecessors. */
652 for (i = 1; i <= n; i++)
653 if (! equal_boundaries (rhs[i].start, rhs[i].end))
655 loc.start = rhs[i].start;
663 /* Add a lex-param or a parse-param (depending on TYPE) with
664 declaration DECL and location LOC. */
667 add_param (char const *type, char *decl, location loc)
669 static char const alphanum[26 + 26 + 1 + 10] =
670 "abcdefghijklmnopqrstuvwxyz"
671 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
674 char const *name_start = NULL;
677 /* Stop on last actual character. */
678 for (p = decl; p[1]; p++)
680 || ! memchr (alphanum, p[-1], sizeof alphanum))
681 && memchr (alphanum, p[0], sizeof alphanum - 10))
684 /* Strip the surrounding '{' and '}', and any blanks just inside
686 while (*--p == ' ' || *p == '\t')
689 while (*++decl == ' ' || *decl == '\t')
693 complain_at (loc, _("missing identifier in parameter declaration"));
700 memchr (alphanum, name_start[name_len], sizeof alphanum);
704 name = xmalloc (name_len + 1);
705 memcpy (name, name_start, name_len);
706 name[name_len] = '\0';
707 muscle_pair_list_grow (type, decl, name);
711 gram_scanner_last_string_free ();
716 version_check (location const *loc, char const *version)
718 if (strverscmp (version, PACKAGE_VERSION) > 0)
720 complain_at (*loc, "require bison %s, but have %s",
721 version, PACKAGE_VERSION);
727 gram_error (location const *loc, char const *msg)
729 complain_at (*loc, "%s", msg);
733 token_name (int type)
735 return yytname[YYTRANSLATE (type)];
746 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
747 return quotearg_style (escape_quoting_style, buf);