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/>. */
24 #include "conflicts.h"
28 #include "muscle-tab.h"
29 #include "named-ref.h"
34 #include "scan-gram.h"
35 #include "scan-code.h"
38 #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
39 static YYLTYPE lloc_default (YYLTYPE const *, int);
41 #define YY_LOCATION_PRINT(File, Loc) \
42 location_print (File, Loc)
44 static void version_check (location const *loc, char const *version);
46 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
47 FIXME: depends on the undocumented availability of YYLLOC. */
49 #define yyerror(Msg) \
50 gram_error (&yylloc, Msg)
51 static void gram_error (location const *, char const *);
53 static char const *char_name (char);
58 static int current_prec = 0;
59 static location current_lhs_location;
60 static named_ref *current_lhs_named_ref;
61 static symbol *current_lhs_symbol;
62 static symbol_class current_class = unknown_sym;
63 static uniqstr current_type = NULL;
65 /** Set the new current left-hand side symbol, possibly common
66 * to several right-hand side parts of rule.
70 current_lhs(symbol *sym, location loc, named_ref *ref)
72 current_lhs_symbol = sym;
73 current_lhs_location = loc;
74 /* In order to simplify memory management, named references for lhs
75 are always assigned by deep copy into the current symbol_list
76 node. This is because a single named-ref in the grammar may
77 result in several uses when the user factors lhs between several
78 rules using "|". Therefore free the parser's original copy. */
79 free (current_lhs_named_ref);
80 current_lhs_named_ref = ref;
83 #define YYTYPE_INT16 int_fast16_t
84 #define YYTYPE_INT8 int_fast8_t
85 #define YYTYPE_UINT16 uint_fast16_t
86 #define YYTYPE_UINT8 uint_fast8_t
93 %define parse.error verbose
94 %define parse.lac full
101 /* Bison's grammar can initial empty locations, hence a default
102 location is needed. */
103 boundary_set (&@$.start, current_file, 1, 1);
104 boundary_set (&@$.end, current_file, 1, 1);
113 named_ref *named_ref;
117 unsigned char character;
120 /* Define the tokens together with their human representation. */
121 %token GRAM_EOF 0 "end of file"
122 %token STRING "string"
125 %token PERCENT_TOKEN "%token"
126 %token PERCENT_NTERM "%nterm"
128 %token PERCENT_TYPE "%type"
129 %token PERCENT_DESTRUCTOR "%destructor"
130 %token PERCENT_PRINTER "%printer"
132 %token PERCENT_LEFT "%left"
133 %token PERCENT_RIGHT "%right"
134 %token PERCENT_NONASSOC "%nonassoc"
135 %token PERCENT_PRECEDENCE "%precedence"
137 %token PERCENT_PREC "%prec"
138 %token PERCENT_DPREC "%dprec"
139 %token PERCENT_MERGE "%merge"
141 /*----------------------.
142 | Global Declarations. |
143 `----------------------*/
147 PERCENT_DEFAULT_PREC "%default-prec"
148 PERCENT_DEFINE "%define"
149 PERCENT_DEFINES "%defines"
150 PERCENT_ERROR_VERBOSE "%error-verbose"
151 PERCENT_EXPECT "%expect"
152 PERCENT_EXPECT_RR "%expect-rr"
153 PERCENT_FLAG "%<flag>"
154 PERCENT_FILE_PREFIX "%file-prefix"
155 PERCENT_GLR_PARSER "%glr-parser"
156 PERCENT_INITIAL_ACTION "%initial-action"
157 PERCENT_LANGUAGE "%language"
158 PERCENT_NAME_PREFIX "%name-prefix"
159 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
160 PERCENT_NO_LINES "%no-lines"
161 PERCENT_NONDETERMINISTIC_PARSER
162 "%nondeterministic-parser"
163 PERCENT_OUTPUT "%output"
164 PERCENT_REQUIRE "%require"
165 PERCENT_SKELETON "%skeleton"
166 PERCENT_START "%start"
167 PERCENT_TOKEN_TABLE "%token-table"
168 PERCENT_VERBOSE "%verbose"
172 %token BRACED_CODE "{...}"
173 %token BRACED_PREDICATE "%?{...}"
174 %token BRACKETED_ID "[identifier]"
176 %token EPILOGUE "epilogue"
178 %token ID "identifier"
179 %token ID_COLON "identifier:"
180 %token PERCENT_PERCENT "%%"
182 %token PROLOGUE "%{...%}"
188 %type <character> CHAR
189 %printer { fputs (char_name ($$), yyo); } CHAR
191 /* braceless is not to be used for rule or symbol actions, as it
192 calls code_props_plain_init. */
193 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
194 %type <code> "{...}" "%?{...}"
195 %printer { fputs (quotearg_style (c_quoting_style, $$), yyo); }
197 %printer { fprintf (yyo, "{\n%s\n}", $$); }
198 braceless content.opt "{...}" "%{...%}" EPILOGUE
200 %type <uniqstr> BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable
201 %printer { fputs ($$, yyo); } <uniqstr>
202 %printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID
203 %printer { fprintf (yyo, "%s:", $$); } ID_COLON
204 %printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG
205 %printer { fprintf (yyo, "<%s>", $$); } TAG tag
208 %printer { fprintf (yyo, "%d", $$); } <integer>
210 %type <symbol> id id_colon string_as_id symbol symbol.prec
211 %printer { fprintf (yyo, "%s", $$->tag); } <symbol>
212 %printer { fprintf (yyo, "%s:", $$->tag); } id_colon
214 %type <assoc> precedence_declarator
215 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
216 %type <named_ref> named_ref.opt
229 param_parse = 1 << 1,
230 param_both = param_lex | param_parse
236 /** Add a lex-param and/or a parse-param.
238 * \param type where to push this formal argument.
239 * \param decl the formal argument. Destroyed.
240 * \param loc the location in the source.
242 static void add_param (param_type type, char *decl, location loc);
243 static param_type current_param = param_none;
249 %token <param> PERCENT_PARAM "%param";
254 #define CASE(In, Out) \
255 case param_ ## In: fputs ("%" #Out, stderr); break
256 CASE(lex, lex-param);
257 CASE(parse, parse-param);
260 case param_none: aver (false); break;
271 prologue_declarations "%%" grammar epilogue.opt
275 /*------------------------------------.
276 | Declarations: before the first %%. |
277 `------------------------------------*/
279 prologue_declarations:
281 | prologue_declarations prologue_declaration
284 prologue_declaration:
288 code_props plain_code;
289 code_props_plain_init (&plain_code, $1, @1);
290 code_props_translate_code (&plain_code);
291 gram_scanner_last_string_free ();
292 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
293 plain_code.code, @1);
294 code_scanner_last_string_free ();
298 muscle_percent_define_ensure ($1, @1, true);
300 | "%define" variable content.opt
302 muscle_percent_define_insert ($2, @2, $3,
303 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
305 | "%defines" { defines_flag = true; }
309 spec_defines_file = xstrdup ($2);
313 muscle_percent_define_insert ("parse.error", @1, "verbose",
314 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
316 | "%expect" INT { expected_sr_conflicts = $2; }
317 | "%expect-rr" INT { expected_rr_conflicts = $2; }
318 | "%file-prefix" STRING { spec_file_prefix = $2; }
319 | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
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 | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */
337 | "%no-lines" { no_lines_flag = true; }
338 | "%nondeterministic-parser" { nondeterministic_parser = true; }
339 | "%output" STRING { spec_outfile = $2; }
340 | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
341 | "%param" { current_param = $1; } params { current_param = param_none; }
342 | "%require" STRING { version_check (&@2, $2); }
345 char const *skeleton_user = $2;
346 if (strchr (skeleton_user, '/'))
348 size_t dir_length = strlen (current_file);
349 char *skeleton_build;
350 while (dir_length && current_file[dir_length - 1] != '/')
352 while (dir_length && current_file[dir_length - 1] == '/')
355 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
358 memcpy (skeleton_build, current_file, dir_length);
359 skeleton_build[dir_length++] = '/';
361 strcpy (skeleton_build + dir_length, skeleton_user);
362 skeleton_user = uniqstr_new (skeleton_build);
363 free (skeleton_build);
365 skeleton_arg (skeleton_user, grammar_prio, @1);
367 | "%token-table" { token_table_flag = true; }
368 | "%verbose" { report_flag |= report_states; }
369 | "%yacc" { yacc_flag = true; }
370 | /*FIXME: Err? What is this horror doing here? */ ";"
374 params "{...}" { add_param (current_param, $2, @2); }
375 | "{...}" { add_param (current_param, $1, @1); }
379 /*----------------------.
380 | grammar_declaration. |
381 `----------------------*/
384 precedence_declaration
388 grammar_start_symbol_set ($2, @2);
390 | code_props_type "{...}" generic_symlist
393 code_props_symbol_action_init (&code, $2, @2);
394 code_props_translate_code (&code);
397 for (list = $3; list; list = list->next)
398 symbol_list_code_props_set (list, $1, &code);
399 symbol_list_free ($3);
408 default_prec = false;
412 /* Do not invoke muscle_percent_code_grow here since it invokes
413 muscle_user_name_list_grow. */
414 muscle_code_grow ("percent_code()", $2, @2);
415 code_scanner_last_string_free ();
417 | "%code" ID braceless
419 muscle_percent_code_grow ($2, @2, $3, @3);
420 code_scanner_last_string_free ();
424 %type <code_type> code_props_type;
425 %union {code_props_type code_type;};
426 %printer { fprintf (yyo, "%s", code_props_type_string ($$)); } <code_type>;
428 "%destructor" { $$ = destructor; }
429 | "%printer" { $$ = printer; }
436 %token PERCENT_UNION "%union";
440 | ID { muscle_code_grow ("union_name", $1, @1); }
444 "%union" union_name braceless
447 muscle_code_grow ("stype", $3, @3);
448 code_scanner_last_string_free ();
456 "%nterm" { current_class = nterm_sym; } symbol_defs.1
458 current_class = unknown_sym;
461 | "%token" { current_class = token_sym; } symbol_defs.1
463 current_class = unknown_sym;
466 | "%type" TAG symbols.1
470 for (list = $3; list; list = list->next)
471 symbol_type_set (list->content.sym, $2, @2);
472 symbol_list_free ($3);
476 precedence_declaration:
477 precedence_declarator tag.opt symbols.prec
481 for (list = $3; list; list = list->next)
483 symbol_type_set (list->content.sym, current_type, @2);
484 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
486 symbol_list_free ($3);
491 precedence_declarator:
492 "%left" { $$ = left_assoc; }
493 | "%right" { $$ = right_assoc; }
494 | "%nonassoc" { $$ = non_assoc; }
495 | "%precedence" { $$ = precedence_assoc; }
499 /* Nothing. */ { current_type = NULL; }
500 | TAG { current_type = $1; tag_seen = true; }
503 /* Just like symbols.1 but accept INT for the sake of POSIX. */
506 { $$ = symbol_list_sym_new ($1, @1); }
507 | symbols.prec symbol.prec
508 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
513 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
516 /* One or more symbols to be %typed. */
519 { $$ = symbol_list_sym_new ($1, @1); }
521 { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
525 generic_symlist_item { $$ = $1; }
526 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
529 generic_symlist_item:
530 symbol { $$ = symbol_list_sym_new ($1, @1); }
531 | tag { $$ = symbol_list_type_new ($1, @1); }
536 | "<*>" { $$ = uniqstr_new ("*"); }
537 | "<>" { $$ = uniqstr_new (""); }
540 /* One token definition. */
549 symbol_class_set ($1, current_class, @1, true);
550 symbol_type_set ($1, current_type, @1);
554 symbol_class_set ($1, current_class, @1, true);
555 symbol_type_set ($1, current_type, @1);
556 symbol_user_token_number_set ($1, $2, @2);
560 symbol_class_set ($1, current_class, @1, true);
561 symbol_type_set ($1, current_type, @1);
562 symbol_make_alias ($1, $2, @$);
564 | id INT string_as_id
566 symbol_class_set ($1, current_class, @1, true);
567 symbol_type_set ($1, current_type, @1);
568 symbol_user_token_number_set ($1, $2, @2);
569 symbol_make_alias ($1, $3, @$);
573 /* One or more symbol definitions. */
576 | symbol_defs.1 symbol_def
580 /*------------------------------------------.
581 | The grammar section: between the two %%. |
582 `------------------------------------------*/
585 rules_or_grammar_declaration
586 | grammar rules_or_grammar_declaration
589 /* As a Bison extension, one can use the grammar declarations in the
590 body of the grammar. */
591 rules_or_grammar_declaration:
593 | grammar_declaration ";"
601 id_colon named_ref.opt { current_lhs ($1, @1, $2); } rhses.1
603 /* Free the current lhs. */
604 current_lhs (0, @1, 0);
609 rhs { grammar_current_rule_end (@1); }
610 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
616 { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location,
617 current_lhs_named_ref); }
618 | rhs symbol named_ref.opt
619 { grammar_current_rule_symbol_append ($2, @2, $3); }
620 | rhs "{...}" named_ref.opt
621 { grammar_current_rule_action_append ($2, @2, $3, false); }
623 { grammar_current_rule_action_append ($2, @2, NULL, true); }
625 { grammar_current_rule_prec_set ($3, @3); }
627 { grammar_current_rule_dprec_set ($3, @3); }
629 { grammar_current_rule_merge_set ($3, @3); }
633 /* Nothing. */ { $$ = 0; }
635 BRACKETED_ID { $$ = named_ref_new($1, @1); }
638 /*---------------------------.
639 | variable and content.opt. |
640 `---------------------------*/
642 /* The STRING form of variable is deprecated and is not M4-friendly.
643 For example, M4 fails for `%define "[" "value"'. */
646 | STRING { $$ = uniqstr_new ($1); }
649 /* Some content or empty by default. */
651 /* Nothing. */ { $$ = ""; }
664 code_props plain_code;
665 $1[strlen ($1) - 1] = '\n';
666 code_props_plain_init (&plain_code, $1+1, @1);
667 code_props_translate_code (&plain_code);
668 gram_scanner_last_string_free ();
669 $$ = plain_code.code;
678 /* Identifiers are returned as uniqstr values by the scanner.
679 Depending on their use, we may need to make them genuine symbols. */
683 { $$ = symbol_from_uniqstr ($1, @1); }
686 $$ = symbol_get (char_name ($1), @1);
687 symbol_class_set ($$, token_sym, @1, false);
688 symbol_user_token_number_set ($$, $1, @1);
693 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
702 /* A string used as an ID: quote it. */
706 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
707 symbol_class_set ($$, token_sym, @1, false);
715 code_props plain_code;
716 code_props_plain_init (&plain_code, $2, @2);
717 code_props_translate_code (&plain_code);
718 gram_scanner_last_string_free ();
719 muscle_code_grow ("epilogue", plain_code.code, @2);
720 code_scanner_last_string_free ();
727 /* Return the location of the left-hand side of a rule whose
728 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
729 the right-hand side, and return an empty location equal to the end
730 boundary of RHS[0] if the right-hand side is empty. */
733 lloc_default (YYLTYPE const *rhs, int n)
738 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
739 The bug is fixed in 7.4.2m, but play it safe for now. */
740 loc.start = rhs[n].end;
741 loc.end = rhs[n].end;
743 /* Ignore empty nonterminals the start of the right-hand side.
744 Do not bother to ignore them at the end of the right-hand side,
745 since empty nonterminals have the same end as their predecessors. */
746 for (i = 1; i <= n; i++)
747 if (! equal_boundaries (rhs[i].start, rhs[i].end))
749 loc.start = rhs[i].start;
758 add_param (param_type type, char *decl, location loc)
760 static char const alphanum[26 + 26 + 1 + 10] =
761 "abcdefghijklmnopqrstuvwxyz"
762 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
766 char const *name_start = NULL;
769 /* Stop on last actual character. */
770 for (p = decl; p[1]; p++)
772 || ! memchr (alphanum, p[-1], sizeof alphanum))
773 && memchr (alphanum, p[0], sizeof alphanum - 10))
776 /* Strip the surrounding '{' and '}', and any blanks just inside
779 while (isspace ((unsigned char) *p))
783 while (isspace ((unsigned char) *decl))
788 complain_at (loc, complaint,
789 _("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_at (*loc, complaint, "require bison %s, but have %s",
810 version, PACKAGE_VERSION);
816 gram_error (location const *loc, char const *msg)
818 complain_at (*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);