1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002-2012 Free Software Foundation, Inc.
5 This file is part of Bison, the GNU Compiler Compiler.
7 This program is free software: you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation, either version 3 of the License, or
10 (at your option) any later version.
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 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"
35 #include "scan-gram.h"
36 #include "scan-code.h"
39 #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
40 static YYLTYPE lloc_default (YYLTYPE const *, int);
42 #define YY_LOCATION_PRINT(File, Loc) \
43 location_print (File, Loc)
45 static void version_check (location const *loc, char const *version);
47 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
48 FIXME: depends on the undocumented availability of YYLLOC. */
50 #define yyerror(Msg) \
51 gram_error (&yylloc, Msg)
52 static void gram_error (location const *, char const *);
54 static char const *char_name (char);
59 static int current_prec = 0;
60 static location current_lhs_location;
61 static named_ref *current_lhs_named_ref;
62 static symbol *current_lhs_symbol;
63 static symbol_class current_class = unknown_sym;
64 static uniqstr current_type = NULL;
66 /** Set the new current left-hand side symbol, possibly common
67 * to several right-hand side parts of rule.
71 current_lhs(symbol *sym, location loc, named_ref *ref)
73 current_lhs_symbol = sym;
74 current_lhs_location = loc;
75 /* In order to simplify memory management, named references for lhs
76 are always assigned by deep copy into the current symbol_list
77 node. This is because a single named-ref in the grammar may
78 result in several uses when the user factors lhs between several
79 rules using "|". Therefore free the parser's original copy. */
80 free (current_lhs_named_ref);
81 current_lhs_named_ref = ref;
84 #define YYTYPE_INT16 int_fast16_t
85 #define YYTYPE_INT8 int_fast8_t
86 #define YYTYPE_UINT16 uint_fast16_t
87 #define YYTYPE_UINT8 uint_fast8_t
90 %define api.prefix "gram_"
93 %define parse.error verbose
94 %define parse.lac full
102 /* Bison's grammar can initial empty locations, hence a default
103 location is needed. */
104 boundary_set (&@$.start, current_file, 1, 1);
105 boundary_set (&@$.end, current_file, 1, 1);
114 named_ref *named_ref;
118 unsigned char character;
121 /* Define the tokens together with their human representation. */
122 %token GRAM_EOF 0 "end of file"
123 %token STRING "string"
126 %token PERCENT_TOKEN "%token"
127 %token PERCENT_NTERM "%nterm"
129 %token PERCENT_TYPE "%type"
130 %token PERCENT_DESTRUCTOR "%destructor"
131 %token PERCENT_PRINTER "%printer"
133 %token PERCENT_LEFT "%left"
134 %token PERCENT_RIGHT "%right"
135 %token PERCENT_NONASSOC "%nonassoc"
136 %token PERCENT_PRECEDENCE "%precedence"
138 %token PERCENT_PREC "%prec"
139 %token PERCENT_DPREC "%dprec"
140 %token PERCENT_MERGE "%merge"
142 /*----------------------.
143 | Global Declarations. |
144 `----------------------*/
148 PERCENT_DEFAULT_PREC "%default-prec"
149 PERCENT_DEFINE "%define"
150 PERCENT_DEFINES "%defines"
151 PERCENT_ERROR_VERBOSE "%error-verbose"
152 PERCENT_EXPECT "%expect"
153 PERCENT_EXPECT_RR "%expect-rr"
154 PERCENT_FLAG "%<flag>"
155 PERCENT_FILE_PREFIX "%file-prefix"
156 PERCENT_GLR_PARSER "%glr-parser"
157 PERCENT_INITIAL_ACTION "%initial-action"
158 PERCENT_LANGUAGE "%language"
159 PERCENT_NAME_PREFIX "%name-prefix"
160 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
161 PERCENT_NO_LINES "%no-lines"
162 PERCENT_NONDETERMINISTIC_PARSER
163 "%nondeterministic-parser"
164 PERCENT_OUTPUT "%output"
165 PERCENT_REQUIRE "%require"
166 PERCENT_SKELETON "%skeleton"
167 PERCENT_START "%start"
168 PERCENT_TOKEN_TABLE "%token-table"
169 PERCENT_VERBOSE "%verbose"
173 %token BRACED_CODE "{...}"
174 %token BRACED_PREDICATE "%?{...}"
175 %token BRACKETED_ID "[identifier]"
177 %token EPILOGUE "epilogue"
179 %token ID "identifier"
180 %token ID_COLON "identifier:"
181 %token PERCENT_PERCENT "%%"
183 %token PROLOGUE "%{...%}"
189 %type <character> CHAR
190 %printer { fputs (char_name ($$), yyo); } CHAR
192 /* braceless is not to be used for rule or symbol actions, as it
193 calls code_props_plain_init. */
194 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
195 %type <code> "{...}" "%?{...}"
196 %printer { fputs (quotearg_style (c_quoting_style, $$), yyo); }
198 %printer { fprintf (yyo, "{\n%s\n}", $$); }
199 braceless content.opt "{...}" "%{...%}" EPILOGUE
201 %type <uniqstr> BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable
202 %printer { fputs ($$, yyo); } <uniqstr>
203 %printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID
204 %printer { fprintf (yyo, "%s:", $$); } ID_COLON
205 %printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG
206 %printer { fprintf (yyo, "<%s>", $$); } TAG tag
209 %printer { fprintf (yyo, "%d", $$); } <integer>
211 %type <symbol> id id_colon string_as_id symbol symbol.prec
212 %printer { fprintf (yyo, "%s", $$->tag); } <symbol>
213 %printer { fprintf (yyo, "%s:", $$->tag); } id_colon
215 %type <assoc> precedence_declarator
216 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
217 %type <named_ref> named_ref.opt
230 param_parse = 1 << 1,
231 param_both = param_lex | param_parse
237 /** Add a lex-param and/or a parse-param.
239 * \param type where to push this formal argument.
240 * \param decl the formal argument. Destroyed.
241 * \param loc the location in the source.
243 static void add_param (param_type type, char *decl, location loc);
244 static param_type current_param = param_none;
250 %token <param> PERCENT_PARAM "%param";
255 #define CASE(In, Out) \
256 case param_ ## In: fputs ("%" #Out, stderr); break
257 CASE(lex, lex-param);
258 CASE(parse, parse-param);
261 case param_none: aver (false); break;
272 prologue_declarations "%%" grammar epilogue.opt
276 /*------------------------------------.
277 | Declarations: before the first %%. |
278 `------------------------------------*/
280 prologue_declarations:
282 | prologue_declarations prologue_declaration
285 prologue_declaration:
289 code_props plain_code;
290 code_props_plain_init (&plain_code, $1, @1);
291 code_props_translate_code (&plain_code);
292 gram_scanner_last_string_free ();
293 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
294 plain_code.code, @1);
295 code_scanner_last_string_free ();
299 muscle_percent_define_ensure ($1, @1, true);
301 | "%define" variable content.opt
303 muscle_percent_define_insert ($2, @2, $3,
304 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
306 | "%defines" { defines_flag = true; }
310 spec_defines_file = xstrdup ($2);
314 muscle_percent_define_insert ("parse.error", @1, "verbose",
315 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
317 | "%expect" INT { expected_sr_conflicts = $2; }
318 | "%expect-rr" INT { expected_rr_conflicts = $2; }
319 | "%file-prefix" STRING { spec_file_prefix = $2; }
322 nondeterministic_parser = true;
325 | "%initial-action" "{...}"
328 code_props_symbol_action_init (&action, $2, @2);
329 code_props_translate_code (&action);
330 gram_scanner_last_string_free ();
331 muscle_code_grow ("initial_action", action.code, @2);
332 code_scanner_last_string_free ();
334 | "%language" STRING { language_argmatch ($2, grammar_prio, @1); }
335 | "%name-prefix" STRING { spec_name_prefix = $2; }
336 | "%no-lines" { no_lines_flag = true; }
337 | "%nondeterministic-parser" { nondeterministic_parser = true; }
338 | "%output" STRING { spec_outfile = $2; }
339 | "%param" { current_param = $1; } params { current_param = param_none; }
340 | "%require" STRING { version_check (&@2, $2); }
343 char const *skeleton_user = $2;
344 if (strchr (skeleton_user, '/'))
346 size_t dir_length = strlen (current_file);
347 char *skeleton_build;
348 while (dir_length && current_file[dir_length - 1] != '/')
350 while (dir_length && current_file[dir_length - 1] == '/')
353 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
356 memcpy (skeleton_build, current_file, dir_length);
357 skeleton_build[dir_length++] = '/';
359 strcpy (skeleton_build + dir_length, skeleton_user);
360 skeleton_user = uniqstr_new (skeleton_build);
361 free (skeleton_build);
363 skeleton_arg (skeleton_user, grammar_prio, @1);
365 | "%token-table" { token_table_flag = true; }
366 | "%verbose" { report_flag |= report_states; }
367 | "%yacc" { yacc_flag = true; }
368 | /*FIXME: Err? What is this horror doing here? */ ";"
372 params "{...}" { add_param (current_param, $2, @2); }
373 | "{...}" { add_param (current_param, $1, @1); }
377 /*----------------------.
378 | grammar_declaration. |
379 `----------------------*/
382 precedence_declaration
386 grammar_start_symbol_set ($2, @2);
388 | code_props_type "{...}" generic_symlist
391 code_props_symbol_action_init (&code, $2, @2);
392 code_props_translate_code (&code);
395 for (list = $3; list; list = list->next)
396 symbol_list_code_props_set (list, $1, &code);
397 symbol_list_free ($3);
406 default_prec = false;
410 /* Do not invoke muscle_percent_code_grow here since it invokes
411 muscle_user_name_list_grow. */
412 muscle_code_grow ("percent_code()", $2, @2);
413 code_scanner_last_string_free ();
415 | "%code" ID braceless
417 muscle_percent_code_grow ($2, @2, $3, @3);
418 code_scanner_last_string_free ();
422 %type <code_type> code_props_type;
423 %union {code_props_type code_type;};
424 %printer { fprintf (yyo, "%s", code_props_type_string ($$)); } <code_type>;
426 "%destructor" { $$ = destructor; }
427 | "%printer" { $$ = printer; }
434 %token PERCENT_UNION "%union";
438 | ID { muscle_code_grow ("union_name", $1, @1); }
442 "%union" union_name braceless
445 muscle_code_grow ("stype", $3, @3);
446 code_scanner_last_string_free ();
454 "%nterm" { current_class = nterm_sym; } symbol_defs.1
456 current_class = unknown_sym;
459 | "%token" { current_class = token_sym; } symbol_defs.1
461 current_class = unknown_sym;
464 | "%type" TAG symbols.1
468 for (list = $3; list; list = list->next)
469 symbol_type_set (list->content.sym, $2, @2);
470 symbol_list_free ($3);
474 precedence_declaration:
475 precedence_declarator tag.opt symbols.prec
479 for (list = $3; list; list = list->next)
481 symbol_type_set (list->content.sym, current_type, @2);
482 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
484 symbol_list_free ($3);
489 precedence_declarator:
490 "%left" { $$ = left_assoc; }
491 | "%right" { $$ = right_assoc; }
492 | "%nonassoc" { $$ = non_assoc; }
493 | "%precedence" { $$ = precedence_assoc; }
497 /* Nothing. */ { current_type = NULL; }
498 | TAG { current_type = $1; tag_seen = true; }
501 /* Just like symbols.1 but accept INT for the sake of POSIX. */
504 { $$ = symbol_list_sym_new ($1, @1); }
505 | symbols.prec symbol.prec
506 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
511 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
514 /* One or more symbols to be %typed. */
517 { $$ = symbol_list_sym_new ($1, @1); }
519 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
523 generic_symlist_item { $$ = $1; }
524 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
527 generic_symlist_item:
528 symbol { $$ = symbol_list_sym_new ($1, @1); }
529 | tag { $$ = symbol_list_type_new ($1, @1); }
534 | "<*>" { $$ = uniqstr_new ("*"); }
535 | "<>" { $$ = uniqstr_new (""); }
538 /* One token definition. */
547 symbol_class_set ($1, current_class, @1, true);
548 symbol_type_set ($1, current_type, @1);
552 symbol_class_set ($1, current_class, @1, true);
553 symbol_type_set ($1, current_type, @1);
554 symbol_user_token_number_set ($1, $2, @2);
558 symbol_class_set ($1, current_class, @1, true);
559 symbol_type_set ($1, current_type, @1);
560 symbol_make_alias ($1, $2, @$);
562 | id INT string_as_id
564 symbol_class_set ($1, current_class, @1, true);
565 symbol_type_set ($1, current_type, @1);
566 symbol_user_token_number_set ($1, $2, @2);
567 symbol_make_alias ($1, $3, @$);
571 /* One or more symbol definitions. */
574 | symbol_defs.1 symbol_def
578 /*------------------------------------------.
579 | The grammar section: between the two %%. |
580 `------------------------------------------*/
583 rules_or_grammar_declaration
584 | grammar rules_or_grammar_declaration
587 /* As a Bison extension, one can use the grammar declarations in the
588 body of the grammar. */
589 rules_or_grammar_declaration:
591 | grammar_declaration ";"
599 id_colon named_ref.opt { current_lhs ($1, @1, $2); } rhses.1
601 /* Free the current lhs. */
602 current_lhs (0, @1, 0);
607 rhs { grammar_current_rule_end (@1); }
608 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
614 { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location,
615 current_lhs_named_ref); }
616 | rhs symbol named_ref.opt
617 { grammar_current_rule_symbol_append ($2, @2, $3); }
618 | rhs "{...}" named_ref.opt
619 { grammar_current_rule_action_append ($2, @2, $3, false); }
621 { grammar_current_rule_action_append ($2, @2, NULL, true); }
623 { grammar_current_rule_prec_set ($3, @3); }
625 { grammar_current_rule_dprec_set ($3, @3); }
627 { grammar_current_rule_merge_set ($3, @3); }
631 /* Nothing. */ { $$ = 0; }
633 BRACKETED_ID { $$ = named_ref_new($1, @1); }
636 /*---------------------------.
637 | variable and content.opt. |
638 `---------------------------*/
640 /* The STRING form of variable is deprecated and is not M4-friendly.
641 For example, M4 fails for `%define "[" "value"'. */
644 | STRING { $$ = uniqstr_new ($1); }
647 /* Some content or empty by default. */
649 /* Nothing. */ { $$ = ""; }
662 code_props plain_code;
663 $1[strlen ($1) - 1] = '\n';
664 code_props_plain_init (&plain_code, $1+1, @1);
665 code_props_translate_code (&plain_code);
666 gram_scanner_last_string_free ();
667 $$ = plain_code.code;
676 /* Identifiers are returned as uniqstr values by the scanner.
677 Depending on their use, we may need to make them genuine symbols. */
681 { $$ = symbol_from_uniqstr ($1, @1); }
684 $$ = symbol_get (char_name ($1), @1);
685 symbol_class_set ($$, token_sym, @1, false);
686 symbol_user_token_number_set ($$, $1, @1);
691 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
700 /* A string used as an ID: quote it. */
704 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
705 symbol_class_set ($$, token_sym, @1, false);
713 code_props plain_code;
714 code_props_plain_init (&plain_code, $2, @2);
715 code_props_translate_code (&plain_code);
716 gram_scanner_last_string_free ();
717 muscle_code_grow ("epilogue", plain_code.code, @2);
718 code_scanner_last_string_free ();
724 /* Return the location of the left-hand side of a rule whose
725 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
726 the right-hand side, and return an empty location equal to the end
727 boundary of RHS[0] if the right-hand side is empty. */
730 lloc_default (YYLTYPE const *rhs, int n)
735 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
736 The bug is fixed in 7.4.2m, but play it safe for now. */
737 loc.start = rhs[n].end;
738 loc.end = rhs[n].end;
740 /* Ignore empty nonterminals the start of the right-hand side.
741 Do not bother to ignore them at the end of the right-hand side,
742 since empty nonterminals have the same end as their predecessors. */
743 for (i = 1; i <= n; i++)
744 if (! equal_boundaries (rhs[i].start, rhs[i].end))
746 loc.start = rhs[i].start;
755 add_param (param_type type, char *decl, location loc)
757 static char const alphanum[26 + 26 + 1 + 10] =
758 "abcdefghijklmnopqrstuvwxyz"
759 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
763 char const *name_start = NULL;
766 /* Stop on last actual character. */
767 for (p = decl; p[1]; p++)
769 || ! memchr (alphanum, p[-1], sizeof alphanum))
770 && memchr (alphanum, p[0], sizeof alphanum - 10))
773 /* Strip the surrounding '{' and '}', and any blanks just inside
776 while (c_isspace ((unsigned char) *p))
780 while (c_isspace ((unsigned char) *decl))
785 complain (&loc, complaint, _("missing identifier in parameter declaration"));
788 char *name = xmemdup0 (name_start, strspn (name_start, alphanum));
789 if (type & param_lex)
790 muscle_pair_list_grow ("lex_param", decl, name);
791 if (type & param_parse)
792 muscle_pair_list_grow ("parse_param", decl, name);
796 gram_scanner_last_string_free ();
801 version_check (location const *loc, char const *version)
803 if (strverscmp (version, PACKAGE_VERSION) > 0)
805 complain (loc, complaint, "require bison %s, but have %s",
806 version, PACKAGE_VERSION);
812 gram_error (location const *loc, char const *msg)
814 complain (loc, complaint, "%s", msg);
818 token_name (int type)
820 return yytname[YYTRANSLATE (type)];
831 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
832 return quotearg_style (escape_quoting_style, buf);