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
94 %define parse.error verbose
95 %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; }
320 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
323 nondeterministic_parser = true;
326 | "%initial-action" "{...}"
329 code_props_symbol_action_init (&action, $2, @2);
330 code_props_translate_code (&action);
331 gram_scanner_last_string_free ();
332 muscle_code_grow ("initial_action", action.code, @2);
333 code_scanner_last_string_free ();
335 | "%language" STRING { language_argmatch ($2, grammar_prio, @1); }
336 | "%name-prefix" STRING { spec_name_prefix = $2; }
337 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
338 | "%no-lines" { no_lines_flag = true; }
339 | "%nondeterministic-parser" { nondeterministic_parser = true; }
340 | "%output" STRING { spec_outfile = $2; }
341 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
342 | "%param" { current_param = $1; } params { current_param = param_none; }
343 | "%require" STRING { version_check (&@2, $2); }
346 char const *skeleton_user = $2;
347 if (strchr (skeleton_user, '/'))
349 size_t dir_length = strlen (current_file);
350 char *skeleton_build;
351 while (dir_length && current_file[dir_length - 1] != '/')
353 while (dir_length && current_file[dir_length - 1] == '/')
356 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
359 memcpy (skeleton_build, current_file, dir_length);
360 skeleton_build[dir_length++] = '/';
362 strcpy (skeleton_build + dir_length, skeleton_user);
363 skeleton_user = uniqstr_new (skeleton_build);
364 free (skeleton_build);
366 skeleton_arg (skeleton_user, grammar_prio, @1);
368 | "%token-table" { token_table_flag = true; }
369 | "%verbose" { report_flag |= report_states; }
370 | "%yacc" { yacc_flag = true; }
371 | /*FIXME: Err? What is this horror doing here? */ ";"
375 params "{...}" { add_param (current_param, $2, @2); }
376 | "{...}" { add_param (current_param, $1, @1); }
380 /*----------------------.
381 | grammar_declaration. |
382 `----------------------*/
385 precedence_declaration
389 grammar_start_symbol_set ($2, @2);
391 | code_props_type "{...}" generic_symlist
394 code_props_symbol_action_init (&code, $2, @2);
395 code_props_translate_code (&code);
398 for (list = $3; list; list = list->next)
399 symbol_list_code_props_set (list, $1, &code);
400 symbol_list_free ($3);
409 default_prec = false;
413 /* Do not invoke muscle_percent_code_grow here since it invokes
414 muscle_user_name_list_grow. */
415 muscle_code_grow ("percent_code()", $2, @2);
416 code_scanner_last_string_free ();
418 | "%code" ID braceless
420 muscle_percent_code_grow ($2, @2, $3, @3);
421 code_scanner_last_string_free ();
425 %type <code_type> code_props_type;
426 %union {code_props_type code_type;};
427 %printer { fprintf (yyo, "%s", code_props_type_string ($$)); } <code_type>;
429 "%destructor" { $$ = destructor; }
430 | "%printer" { $$ = printer; }
437 %token PERCENT_UNION "%union";
441 | ID { muscle_code_grow ("union_name", $1, @1); }
445 "%union" union_name braceless
448 muscle_code_grow ("stype", $3, @3);
449 code_scanner_last_string_free ();
457 "%nterm" { current_class = nterm_sym; } symbol_defs.1
459 current_class = unknown_sym;
462 | "%token" { current_class = token_sym; } symbol_defs.1
464 current_class = unknown_sym;
467 | "%type" TAG symbols.1
471 for (list = $3; list; list = list->next)
472 symbol_type_set (list->content.sym, $2, @2);
473 symbol_list_free ($3);
477 precedence_declaration:
478 precedence_declarator tag.opt symbols.prec
482 for (list = $3; list; list = list->next)
484 symbol_type_set (list->content.sym, current_type, @2);
485 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
487 symbol_list_free ($3);
492 precedence_declarator:
493 "%left" { $$ = left_assoc; }
494 | "%right" { $$ = right_assoc; }
495 | "%nonassoc" { $$ = non_assoc; }
496 | "%precedence" { $$ = precedence_assoc; }
500 /* Nothing. */ { current_type = NULL; }
501 | TAG { current_type = $1; tag_seen = true; }
504 /* Just like symbols.1 but accept INT for the sake of POSIX. */
507 { $$ = symbol_list_sym_new ($1, @1); }
508 | symbols.prec symbol.prec
509 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
514 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
517 /* One or more symbols to be %typed. */
520 { $$ = symbol_list_sym_new ($1, @1); }
522 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
526 generic_symlist_item { $$ = $1; }
527 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
530 generic_symlist_item:
531 symbol { $$ = symbol_list_sym_new ($1, @1); }
532 | tag { $$ = symbol_list_type_new ($1, @1); }
537 | "<*>" { $$ = uniqstr_new ("*"); }
538 | "<>" { $$ = uniqstr_new (""); }
541 /* One token definition. */
550 symbol_class_set ($1, current_class, @1, true);
551 symbol_type_set ($1, current_type, @1);
555 symbol_class_set ($1, current_class, @1, true);
556 symbol_type_set ($1, current_type, @1);
557 symbol_user_token_number_set ($1, $2, @2);
561 symbol_class_set ($1, current_class, @1, true);
562 symbol_type_set ($1, current_type, @1);
563 symbol_make_alias ($1, $2, @$);
565 | id INT string_as_id
567 symbol_class_set ($1, current_class, @1, true);
568 symbol_type_set ($1, current_type, @1);
569 symbol_user_token_number_set ($1, $2, @2);
570 symbol_make_alias ($1, $3, @$);
574 /* One or more symbol definitions. */
577 | symbol_defs.1 symbol_def
581 /*------------------------------------------.
582 | The grammar section: between the two %%. |
583 `------------------------------------------*/
586 rules_or_grammar_declaration
587 | grammar rules_or_grammar_declaration
590 /* As a Bison extension, one can use the grammar declarations in the
591 body of the grammar. */
592 rules_or_grammar_declaration:
594 | grammar_declaration ";"
602 id_colon named_ref.opt { current_lhs ($1, @1, $2); } rhses.1
604 /* Free the current lhs. */
605 current_lhs (0, @1, 0);
610 rhs { grammar_current_rule_end (@1); }
611 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
617 { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location,
618 current_lhs_named_ref); }
619 | rhs symbol named_ref.opt
620 { grammar_current_rule_symbol_append ($2, @2, $3); }
621 | rhs "{...}" named_ref.opt
622 { grammar_current_rule_action_append ($2, @2, $3, false); }
624 { grammar_current_rule_action_append ($2, @2, NULL, true); }
626 { grammar_current_rule_prec_set ($3, @3); }
628 { grammar_current_rule_dprec_set ($3, @3); }
630 { grammar_current_rule_merge_set ($3, @3); }
634 /* Nothing. */ { $$ = 0; }
636 BRACKETED_ID { $$ = named_ref_new($1, @1); }
639 /*---------------------------.
640 | variable and content.opt. |
641 `---------------------------*/
643 /* The STRING form of variable is deprecated and is not M4-friendly.
644 For example, M4 fails for `%define "[" "value"'. */
647 | STRING { $$ = uniqstr_new ($1); }
650 /* Some content or empty by default. */
652 /* Nothing. */ { $$ = ""; }
665 code_props plain_code;
666 $1[strlen ($1) - 1] = '\n';
667 code_props_plain_init (&plain_code, $1+1, @1);
668 code_props_translate_code (&plain_code);
669 gram_scanner_last_string_free ();
670 $$ = plain_code.code;
679 /* Identifiers are returned as uniqstr values by the scanner.
680 Depending on their use, we may need to make them genuine symbols. */
684 { $$ = symbol_from_uniqstr ($1, @1); }
687 $$ = symbol_get (char_name ($1), @1);
688 symbol_class_set ($$, token_sym, @1, false);
689 symbol_user_token_number_set ($$, $1, @1);
694 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
703 /* A string used as an ID: quote it. */
707 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
708 symbol_class_set ($$, token_sym, @1, false);
716 code_props plain_code;
717 code_props_plain_init (&plain_code, $2, @2);
718 code_props_translate_code (&plain_code);
719 gram_scanner_last_string_free ();
720 muscle_code_grow ("epilogue", plain_code.code, @2);
721 code_scanner_last_string_free ();
728 /* Return the location of the left-hand side of a rule whose
729 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
730 the right-hand side, and return an empty location equal to the end
731 boundary of RHS[0] if the right-hand side is empty. */
734 lloc_default (YYLTYPE const *rhs, int n)
739 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
740 The bug is fixed in 7.4.2m, but play it safe for now. */
741 loc.start = rhs[n].end;
742 loc.end = rhs[n].end;
744 /* Ignore empty nonterminals the start of the right-hand side.
745 Do not bother to ignore them at the end of the right-hand side,
746 since empty nonterminals have the same end as their predecessors. */
747 for (i = 1; i <= n; i++)
748 if (! equal_boundaries (rhs[i].start, rhs[i].end))
750 loc.start = rhs[i].start;
759 add_param (param_type type, char *decl, location loc)
761 static char const alphanum[26 + 26 + 1 + 10] =
762 "abcdefghijklmnopqrstuvwxyz"
763 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
767 char const *name_start = NULL;
770 /* Stop on last actual character. */
771 for (p = decl; p[1]; p++)
773 || ! memchr (alphanum, p[-1], sizeof alphanum))
774 && memchr (alphanum, p[0], sizeof alphanum - 10))
777 /* Strip the surrounding '{' and '}', and any blanks just inside
780 while (c_isspace ((unsigned char) *p))
784 while (c_isspace ((unsigned char) *decl))
789 complain (&loc, complaint, _("missing identifier in parameter declaration"));
792 char *name = xmemdup0 (name_start, strspn (name_start, alphanum));
793 if (type & param_lex)
794 muscle_pair_list_grow ("lex_param", decl, name);
795 if (type & param_parse)
796 muscle_pair_list_grow ("parse_param", decl, name);
800 gram_scanner_last_string_free ();
805 version_check (location const *loc, char const *version)
807 if (strverscmp (version, PACKAGE_VERSION) > 0)
809 complain (loc, complaint, "require bison %s, but have %s",
810 version, PACKAGE_VERSION);
816 gram_error (location const *loc, char const *msg)
818 complain (loc, complaint, "%s", msg);
822 token_name (int type)
824 return yytname[YYTRANSLATE (type)];
835 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
836 return quotearg_style (escape_quoting_style, buf);