1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002-2010 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/>. */
24 #include "conflicts.h"
28 #include "muscle-tab.h"
29 #include "named-ref.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);
56 static int current_prec = 0;
57 static location current_lhs_location;
58 static named_ref *current_lhs_named_ref;
59 static symbol *current_lhs;
60 static symbol_class current_class = unknown_sym;
61 static uniqstr current_type = NULL;
63 #define YYTYPE_INT16 int_fast16_t
64 #define YYTYPE_INT8 int_fast8_t
65 #define YYTYPE_UINT16 uint_fast16_t
66 #define YYTYPE_UINT8 uint_fast8_t
74 %define parse.error "verbose"
80 /* Bison's grammar can initial empty locations, hence a default
81 location is needed. */
82 boundary_set (&@$.start, current_file, 1, 1);
83 boundary_set (&@$.end, current_file, 1, 1);
96 unsigned char character;
99 /* Define the tokens together with their human representation. */
100 %token GRAM_EOF 0 "end of file"
101 %token STRING "string"
104 %token PERCENT_TOKEN "%token"
105 %token PERCENT_NTERM "%nterm"
107 %token PERCENT_TYPE "%type"
108 %token PERCENT_DESTRUCTOR "%destructor"
109 %token PERCENT_PRINTER "%printer"
111 %token PERCENT_LEFT "%left"
112 %token PERCENT_RIGHT "%right"
113 %token PERCENT_NONASSOC "%nonassoc"
114 %token PERCENT_PRECEDENCE "%precedence"
116 %token PERCENT_PREC "%prec"
117 %token PERCENT_DPREC "%dprec"
118 %token PERCENT_MERGE "%merge"
121 /*----------------------.
122 | Global Declarations. |
123 `----------------------*/
127 PERCENT_DEFAULT_PREC "%default-prec"
128 PERCENT_DEFINE "%define"
129 PERCENT_DEFINES "%defines"
130 PERCENT_ERROR_VERBOSE "%error-verbose"
131 PERCENT_EXPECT "%expect"
132 PERCENT_EXPECT_RR "%expect-rr"
133 PERCENT_FLAG "%<flag>"
134 PERCENT_FILE_PREFIX "%file-prefix"
135 PERCENT_GLR_PARSER "%glr-parser"
136 PERCENT_INITIAL_ACTION "%initial-action"
137 PERCENT_LANGUAGE "%language"
138 PERCENT_NAME_PREFIX "%name-prefix"
139 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
140 PERCENT_NO_LINES "%no-lines"
141 PERCENT_NONDETERMINISTIC_PARSER
142 "%nondeterministic-parser"
143 PERCENT_OUTPUT "%output"
144 PERCENT_REQUIRE "%require"
145 PERCENT_SKELETON "%skeleton"
146 PERCENT_START "%start"
147 PERCENT_TOKEN_TABLE "%token-table"
148 PERCENT_VERBOSE "%verbose"
152 %token BRACED_CODE "{...}"
153 %token BRACKETED_ID "[identifier]"
155 %token EPILOGUE "epilogue"
157 %token ID "identifier"
158 %token ID_COLON "identifier:"
159 %token PERCENT_PERCENT "%%"
161 %token PROLOGUE "%{...%}"
167 %type <character> CHAR
168 %printer { fputs (char_name ($$), stderr); } CHAR
170 /* braceless is not to be used for rule or symbol actions, as it
171 calls code_props_plain_init. */
172 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
174 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
176 %printer { fprintf (stderr, "{\n%s\n}", $$); }
177 braceless content.opt "{...}" "%{...%}" EPILOGUE
179 %type <uniqstr> BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG variable
180 %printer { fputs ($$, stderr); } <uniqstr>
181 %printer { fprintf (stderr, "[%s]", $$); } BRACKETED_ID
182 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
183 %printer { fprintf (stderr, "%%%s", $$); } PERCENT_FLAG
184 %printer { fprintf (stderr, "<%s>", $$); } TAG
187 %printer { fprintf (stderr, "%d", $$); } <integer>
189 %type <symbol> id id_colon string_as_id symbol symbol.prec
190 %printer { fprintf (stderr, "%s", $$->tag); } <symbol>
191 %printer { fprintf (stderr, "%s:", $$->tag); } id_colon
193 %type <assoc> precedence_declarator
194 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
195 %type <named_ref> named_ref.opt
208 param_parse = 1 << 1,
209 param_both = param_lex | param_parse
215 /** Add a lex-param and/or a parse-param.
217 * \param type where to push this formal argument.
218 * \param decl the formal argument. Destroyed.
219 * \param loc the location in the source.
221 static void add_param (param_type type, char *decl, location loc);
222 static param_type current_param = param_none;
228 %token <param> PERCENT_PARAM "%param";
233 #define CASE(In, Out) \
234 case param_ ## In: fputs ("%" #Out, stderr); break
235 CASE(lex, lex-param);
236 CASE(parse, parse-param);
239 case param_none: aver (false); break;
250 prologue_declarations "%%" grammar epilogue.opt
254 /*------------------------------------.
255 | Declarations: before the first %%. |
256 `------------------------------------*/
258 prologue_declarations:
260 | prologue_declarations prologue_declaration
263 prologue_declaration:
267 code_props plain_code;
268 code_props_plain_init (&plain_code, $1, @1);
269 code_props_translate_code (&plain_code);
270 gram_scanner_last_string_free ();
271 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
272 plain_code.code, @1);
273 code_scanner_last_string_free ();
277 muscle_percent_define_ensure ($1, @1, true);
279 | "%define" variable content.opt
281 muscle_percent_define_insert ($2, @2, $3,
282 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
284 | "%defines" { defines_flag = true; }
288 spec_defines_file = xstrdup ($2);
292 muscle_percent_define_insert ("parse.error", @1, "verbose",
293 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
295 | "%expect" INT { expected_sr_conflicts = $2; }
296 | "%expect-rr" INT { expected_rr_conflicts = $2; }
297 | "%file-prefix" STRING { spec_file_prefix = $2; }
298 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
301 nondeterministic_parser = true;
304 | "%initial-action" "{...}"
307 code_props_symbol_action_init (&action, $2, @2);
308 code_props_translate_code (&action);
309 gram_scanner_last_string_free ();
310 muscle_code_grow ("initial_action", action.code, @2);
311 code_scanner_last_string_free ();
313 | "%language" STRING { language_argmatch ($2, grammar_prio, @1); }
314 | "%name-prefix" STRING { spec_name_prefix = $2; }
315 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
316 | "%no-lines" { no_lines_flag = true; }
317 | "%nondeterministic-parser" { nondeterministic_parser = true; }
318 | "%output" STRING { spec_outfile = $2; }
319 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
320 | "%param" { current_param = $1; } params { current_param = param_none; }
321 | "%require" STRING { version_check (&@2, $2); }
324 char const *skeleton_user = $2;
325 if (strchr (skeleton_user, '/'))
327 size_t dir_length = strlen (current_file);
328 char *skeleton_build;
329 while (dir_length && current_file[dir_length - 1] != '/')
331 while (dir_length && current_file[dir_length - 1] == '/')
334 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
337 strncpy (skeleton_build, current_file, dir_length);
338 skeleton_build[dir_length++] = '/';
340 strcpy (skeleton_build + dir_length, skeleton_user);
341 skeleton_user = uniqstr_new (skeleton_build);
342 free (skeleton_build);
344 skeleton_arg (skeleton_user, grammar_prio, @1);
346 | "%token-table" { token_table_flag = true; }
347 | "%verbose" { report_flag |= report_states; }
348 | "%yacc" { yacc_flag = true; }
349 | /*FIXME: Err? What is this horror doing here? */ ";"
353 params "{...}" { add_param (current_param, $2, @2); }
354 | "{...}" { add_param (current_param, $1, @1); }
358 /*----------------------.
359 | grammar_declaration. |
360 `----------------------*/
363 precedence_declaration
367 grammar_start_symbol_set ($2, @2);
369 | "%destructor" "{...}" generic_symlist
372 for (list = $3; list; list = list->next)
373 symbol_list_destructor_set (list, $2, @2);
374 symbol_list_free ($3);
376 | "%printer" "{...}" generic_symlist
379 for (list = $3; list; list = list->next)
380 symbol_list_printer_set (list, $2, @2);
381 symbol_list_free ($3);
389 default_prec = false;
393 /* Do not invoke muscle_percent_code_grow here since it invokes
394 muscle_user_name_list_grow. */
395 muscle_code_grow ("percent_code()", $2, @2);
396 code_scanner_last_string_free ();
398 | "%code" ID braceless
400 muscle_percent_code_grow ($2, @2, $3, @3);
401 code_scanner_last_string_free ();
410 %token PERCENT_UNION "%union";
414 | ID { muscle_code_grow ("union_name", $1, @1); }
418 "%union" union_name braceless
421 muscle_code_grow ("stype", $3, @3);
422 code_scanner_last_string_free ();
430 "%nterm" { current_class = nterm_sym; } symbol_defs.1
432 current_class = unknown_sym;
435 | "%token" { current_class = token_sym; } symbol_defs.1
437 current_class = unknown_sym;
440 | "%type" TAG symbols.1
444 for (list = $3; list; list = list->next)
445 symbol_type_set (list->content.sym, $2, @2);
446 symbol_list_free ($3);
450 precedence_declaration:
451 precedence_declarator tag.opt symbols.prec
455 for (list = $3; list; list = list->next)
457 symbol_type_set (list->content.sym, current_type, @2);
458 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
460 symbol_list_free ($3);
465 precedence_declarator:
466 "%left" { $$ = left_assoc; }
467 | "%right" { $$ = right_assoc; }
468 | "%nonassoc" { $$ = non_assoc; }
469 | "%precedence" { $$ = precedence_assoc; }
473 /* Nothing. */ { current_type = NULL; }
474 | TAG { current_type = $1; tag_seen = true; }
477 /* Just like symbols.1 but accept INT for the sake of POSIX. */
480 { $$ = symbol_list_sym_new ($1, @1); }
481 | symbols.prec symbol.prec
482 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
487 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
490 /* One or more symbols to be %typed. */
493 { $$ = symbol_list_sym_new ($1, @1); }
495 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
499 generic_symlist_item { $$ = $1; }
500 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
503 generic_symlist_item:
504 symbol { $$ = symbol_list_sym_new ($1, @1); }
505 | TAG { $$ = symbol_list_type_new ($1, @1); }
506 | "<*>" { $$ = symbol_list_default_tagged_new (@1); }
507 | "<>" { $$ = symbol_list_default_tagless_new (@1); }
510 /* One token definition. */
519 symbol_class_set ($1, current_class, @1, true);
520 symbol_type_set ($1, current_type, @1);
524 symbol_class_set ($1, current_class, @1, true);
525 symbol_type_set ($1, current_type, @1);
526 symbol_user_token_number_set ($1, $2, @2);
530 symbol_class_set ($1, current_class, @1, true);
531 symbol_type_set ($1, current_type, @1);
532 symbol_make_alias ($1, $2, @$);
534 | id INT string_as_id
536 symbol_class_set ($1, current_class, @1, true);
537 symbol_type_set ($1, current_type, @1);
538 symbol_user_token_number_set ($1, $2, @2);
539 symbol_make_alias ($1, $3, @$);
543 /* One or more symbol definitions. */
546 | symbol_defs.1 symbol_def
550 /*------------------------------------------.
551 | The grammar section: between the two %%. |
552 `------------------------------------------*/
555 rules_or_grammar_declaration
556 | grammar rules_or_grammar_declaration
559 /* As a Bison extension, one can use the grammar declarations in the
560 body of the grammar. */
561 rules_or_grammar_declaration:
563 | grammar_declaration ";"
571 id_colon named_ref.opt { current_lhs = $1; current_lhs_location = @1;
572 current_lhs_named_ref = $2; } rhses.1
576 rhs { grammar_current_rule_end (@1); }
577 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
583 { grammar_current_rule_begin (current_lhs, current_lhs_location,
584 current_lhs_named_ref); }
585 | rhs symbol named_ref.opt
586 { grammar_current_rule_symbol_append ($2, @2, $3); }
587 | rhs "{...}" named_ref.opt
588 { grammar_current_rule_action_append ($2, @2, $3); }
590 { grammar_current_rule_prec_set ($3, @3); }
592 { grammar_current_rule_dprec_set ($3, @3); }
594 { grammar_current_rule_merge_set ($3, @3); }
598 /* Nothing. */ { $$ = 0; }
600 BRACKETED_ID { $$ = named_ref_new($1, @1); }
604 /*---------------------------.
605 | variable and content.opt. |
606 `---------------------------*/
608 /* The STRING form of variable is deprecated and is not M4-friendly.
609 For example, M4 fails for `%define "[" "value"'. */
612 | STRING { $$ = uniqstr_new ($1); }
615 /* Some content or empty by default. */
617 /* Nothing. */ { $$ = ""; }
630 code_props plain_code;
631 $1[strlen ($1) - 1] = '\n';
632 code_props_plain_init (&plain_code, $1+1, @1);
633 code_props_translate_code (&plain_code);
634 gram_scanner_last_string_free ();
635 $$ = plain_code.code;
644 /* Identifiers are returned as uniqstr values by the scanner.
645 Depending on their use, we may need to make them genuine symbols. */
649 { $$ = symbol_from_uniqstr ($1, @1); }
652 $$ = symbol_get (char_name ($1), @1);
653 symbol_class_set ($$, token_sym, @1, false);
654 symbol_user_token_number_set ($$, $1, @1);
659 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
668 /* A string used as an ID: quote it. */
672 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
673 symbol_class_set ($$, token_sym, @1, false);
681 code_props plain_code;
682 code_props_plain_init (&plain_code, $2, @2);
683 code_props_translate_code (&plain_code);
684 gram_scanner_last_string_free ();
685 muscle_code_grow ("epilogue", plain_code.code, @2);
686 code_scanner_last_string_free ();
693 /* Return the location of the left-hand side of a rule whose
694 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
695 the right-hand side, and return an empty location equal to the end
696 boundary of RHS[0] if the right-hand side is empty. */
699 lloc_default (YYLTYPE const *rhs, int n)
704 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
705 The bug is fixed in 7.4.2m, but play it safe for now. */
706 loc.start = rhs[n].end;
707 loc.end = rhs[n].end;
709 /* Ignore empty nonterminals the start of the the right-hand side.
710 Do not bother to ignore them at the end of the right-hand side,
711 since empty nonterminals have the same end as their predecessors. */
712 for (i = 1; i <= n; i++)
713 if (! equal_boundaries (rhs[i].start, rhs[i].end))
715 loc.start = rhs[i].start;
724 add_param (param_type type, char *decl, location loc)
726 static char const alphanum[26 + 26 + 1 + 10] =
727 "abcdefghijklmnopqrstuvwxyz"
728 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
732 char const *name_start = NULL;
735 /* Stop on last actual character. */
736 for (p = decl; p[1]; p++)
738 || ! memchr (alphanum, p[-1], sizeof alphanum))
739 && memchr (alphanum, p[0], sizeof alphanum - 10))
742 /* Strip the surrounding '{' and '}', and any blanks just inside
744 while (*--p == ' ' || *p == '\t')
747 while (*++decl == ' ' || *decl == '\t')
752 complain_at (loc, _("missing identifier in parameter declaration"));
759 memchr (alphanum, name_start[name_len], sizeof alphanum);
763 name = xmalloc (name_len + 1);
764 memcpy (name, name_start, name_len);
765 name[name_len] = '\0';
766 if (type & param_lex)
767 muscle_pair_list_grow ("lex_param", decl, name);
768 if (type & param_parse)
769 muscle_pair_list_grow ("parse_param", decl, name);
773 gram_scanner_last_string_free ();
778 version_check (location const *loc, char const *version)
780 if (strverscmp (version, PACKAGE_VERSION) > 0)
782 complain_at (*loc, "require bison %s, but have %s",
783 version, PACKAGE_VERSION);
789 gram_error (location const *loc, char const *msg)
791 complain_at (*loc, "%s", msg);
795 token_name (int type)
797 return yytname[YYTRANSLATE (type)];
808 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
809 return quotearg_style (escape_quoting_style, buf);