1 %{/* Bison Grammar Parser -*- C -*-
3 Copyright (C) 2002-2013 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 (Loc, File)
45 static void version_check (location const *loc, char const *version);
47 static void gram_error (location const *, char const *);
49 /* A string that describes a char (e.g., 'a' -> "'a'"). */
50 static char const *char_name (char);
55 static int current_prec = 0;
56 static location current_lhs_location;
57 static named_ref *current_lhs_named_ref;
58 static symbol *current_lhs_symbol;
59 static symbol_class current_class = unknown_sym;
60 static uniqstr current_type = NULL;
62 /** Set the new current left-hand side symbol, possibly common
63 * to several right-hand side parts of rule.
67 current_lhs (symbol *sym, location loc, named_ref *ref)
69 current_lhs_symbol = sym;
70 current_lhs_location = loc;
71 /* In order to simplify memory management, named references for lhs
72 are always assigned by deep copy into the current symbol_list
73 node. This is because a single named-ref in the grammar may
74 result in several uses when the user factors lhs between several
75 rules using "|". Therefore free the parser's original copy. */
76 free (current_lhs_named_ref);
77 current_lhs_named_ref = ref;
80 #define YYTYPE_INT16 int_fast16_t
81 #define YYTYPE_INT8 int_fast8_t
82 #define YYTYPE_UINT16 uint_fast16_t
83 #define YYTYPE_UINT8 uint_fast8_t
86 %define api.prefix "gram_"
89 %define parse.error verbose
90 %define parse.lac full
98 /* Bison's grammar can initial empty locations, hence a default
99 location is needed. */
100 boundary_set (&@$.start, current_file, 1, 1);
101 boundary_set (&@$.end, current_file, 1, 1);
110 named_ref *named_ref;
114 unsigned char character;
117 /* Define the tokens together with their human representation. */
118 %token GRAM_EOF 0 "end of file"
119 %token STRING "string"
122 %token PERCENT_TOKEN "%token"
123 %token PERCENT_NTERM "%nterm"
125 %token PERCENT_TYPE "%type"
126 %token PERCENT_DESTRUCTOR "%destructor"
127 %token PERCENT_PRINTER "%printer"
129 %token PERCENT_LEFT "%left"
130 %token PERCENT_RIGHT "%right"
131 %token PERCENT_NONASSOC "%nonassoc"
132 %token PERCENT_PRECEDENCE "%precedence"
134 %token PERCENT_PREC "%prec"
135 %token PERCENT_DPREC "%dprec"
136 %token PERCENT_MERGE "%merge"
138 /*----------------------.
139 | Global Declarations. |
140 `----------------------*/
144 PERCENT_DEFAULT_PREC "%default-prec"
145 PERCENT_DEFINE "%define"
146 PERCENT_DEFINES "%defines"
147 PERCENT_ERROR_VERBOSE "%error-verbose"
148 PERCENT_EXPECT "%expect"
149 PERCENT_EXPECT_RR "%expect-rr"
150 PERCENT_FLAG "%<flag>"
151 PERCENT_FILE_PREFIX "%file-prefix"
152 PERCENT_GLR_PARSER "%glr-parser"
153 PERCENT_INITIAL_ACTION "%initial-action"
154 PERCENT_LANGUAGE "%language"
155 PERCENT_NAME_PREFIX "%name-prefix"
156 PERCENT_NO_DEFAULT_PREC "%no-default-prec"
157 PERCENT_NO_LINES "%no-lines"
158 PERCENT_NONDETERMINISTIC_PARSER
159 "%nondeterministic-parser"
160 PERCENT_OUTPUT "%output"
161 PERCENT_REQUIRE "%require"
162 PERCENT_SKELETON "%skeleton"
163 PERCENT_START "%start"
164 PERCENT_TOKEN_TABLE "%token-table"
165 PERCENT_VERBOSE "%verbose"
169 %token BRACED_CODE "{...}"
170 %token BRACED_PREDICATE "%?{...}"
171 %token BRACKETED_ID "[identifier]"
173 %token EPILOGUE "epilogue"
175 %token ID "identifier"
176 %token ID_COLON "identifier:"
177 %token PERCENT_PERCENT "%%"
179 %token PROLOGUE "%{...%}"
185 %type <character> CHAR
186 %printer { fputs (char_name ($$), yyo); } CHAR
188 /* braceless is not to be used for rule or symbol actions, as it
189 calls code_props_plain_init. */
190 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
191 %type <code> "{...}" "%?{...}"
192 %printer { fputs (quotearg_style (c_quoting_style, $$), yyo); }
194 %printer { fprintf (yyo, "{\n%s\n}", $$); }
195 braceless content.opt "{...}" "%{...%}" EPILOGUE
197 %type <uniqstr> BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable
198 %printer { fputs ($$, yyo); } <uniqstr>
199 %printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID
200 %printer { fprintf (yyo, "%s:", $$); } ID_COLON
201 %printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG
202 %printer { fprintf (yyo, "<%s>", $$); } TAG tag
205 %printer { fprintf (yyo, "%d", $$); } <integer>
207 %type <symbol> id id_colon string_as_id symbol symbol.prec
208 %printer { fprintf (yyo, "%s", $$->tag); } <symbol>
209 %printer { fprintf (yyo, "%s:", $$->tag); } id_colon
211 %type <assoc> precedence_declarator
212 %type <list> symbols.1 symbols.prec generic_symlist generic_symlist_item
213 %type <named_ref> named_ref.opt
226 param_parse = 1 << 1,
227 param_both = param_lex | param_parse
233 /** Add a lex-param and/or a parse-param.
235 * \param type where to push this formal argument.
236 * \param decl the formal argument. Destroyed.
237 * \param loc the location in the source.
239 static void add_param (param_type type, char *decl, location loc);
240 static param_type current_param = param_none;
246 %token <param> PERCENT_PARAM "%param";
251 #define CASE(In, Out) \
252 case param_ ## In: fputs ("%" #Out, stderr); break
253 CASE (lex, lex-param);
254 CASE (parse, parse-param);
257 case param_none: aver (false); break;
268 prologue_declarations "%%" grammar epilogue.opt
272 /*------------------------------------.
273 | Declarations: before the first %%. |
274 `------------------------------------*/
276 prologue_declarations:
278 | prologue_declarations prologue_declaration
281 prologue_declaration:
285 code_props plain_code;
286 code_props_plain_init (&plain_code, $1, @1);
287 code_props_translate_code (&plain_code);
288 gram_scanner_last_string_free ();
289 muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
290 plain_code.code, @1);
291 code_scanner_last_string_free ();
295 muscle_percent_define_ensure ($1, @1, true);
297 | "%define" variable content.opt
299 muscle_percent_define_insert ($2, @2, $3,
300 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
302 | "%defines" { defines_flag = true; }
306 spec_defines_file = xstrdup ($2);
310 muscle_percent_define_insert ("parse.error", @1, "verbose",
311 MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
313 | "%expect" INT { expected_sr_conflicts = $2; }
314 | "%expect-rr" INT { expected_rr_conflicts = $2; }
315 | "%file-prefix" STRING { spec_file_prefix = $2; }
318 nondeterministic_parser = true;
321 | "%initial-action" "{...}"
324 code_props_symbol_action_init (&action, $2, @2);
325 code_props_translate_code (&action);
326 gram_scanner_last_string_free ();
327 muscle_code_grow ("initial_action", action.code, @2);
328 code_scanner_last_string_free ();
330 | "%language" STRING { language_argmatch ($2, grammar_prio, @1); }
331 | "%name-prefix" STRING { spec_name_prefix = $2; }
332 | "%no-lines" { no_lines_flag = true; }
333 | "%nondeterministic-parser" { nondeterministic_parser = true; }
334 | "%output" STRING { spec_outfile = $2; }
335 | "%param" { current_param = $1; } params { current_param = param_none; }
336 | "%require" STRING { version_check (&@2, $2); }
339 char const *skeleton_user = $2;
340 if (strchr (skeleton_user, '/'))
342 size_t dir_length = strlen (current_file);
343 char *skeleton_build;
344 while (dir_length && current_file[dir_length - 1] != '/')
346 while (dir_length && current_file[dir_length - 1] == '/')
349 xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
352 memcpy (skeleton_build, current_file, dir_length);
353 skeleton_build[dir_length++] = '/';
355 strcpy (skeleton_build + dir_length, skeleton_user);
356 skeleton_user = uniqstr_new (skeleton_build);
357 free (skeleton_build);
359 skeleton_arg (skeleton_user, grammar_prio, @1);
361 | "%token-table" { token_table_flag = true; }
362 | "%verbose" { report_flag |= report_states; }
363 | "%yacc" { yacc_flag = true; }
364 | /*FIXME: Err? What is this horror doing here? */ ";"
368 params "{...}" { add_param (current_param, $2, @2); }
369 | "{...}" { add_param (current_param, $1, @1); }
373 /*----------------------.
374 | grammar_declaration. |
375 `----------------------*/
378 precedence_declaration
382 grammar_start_symbol_set ($2, @2);
384 | code_props_type "{...}" generic_symlist
387 code_props_symbol_action_init (&code, $2, @2);
388 code_props_translate_code (&code);
391 for (list = $3; list; list = list->next)
392 symbol_list_code_props_set (list, $1, &code);
393 symbol_list_free ($3);
402 default_prec = false;
406 /* Do not invoke muscle_percent_code_grow here since it invokes
407 muscle_user_name_list_grow. */
408 muscle_code_grow ("percent_code()", $2, @2);
409 code_scanner_last_string_free ();
411 | "%code" ID braceless
413 muscle_percent_code_grow ($2, @2, $3, @3);
414 code_scanner_last_string_free ();
418 %type <code_type> code_props_type;
419 %union {code_props_type code_type;};
420 %printer { fprintf (yyo, "%s", code_props_type_string ($$)); } <code_type>;
422 "%destructor" { $$ = destructor; }
423 | "%printer" { $$ = printer; }
430 %token PERCENT_UNION "%union";
434 | ID { muscle_code_grow ("union_name", $1, @1); }
438 "%union" union_name braceless
441 muscle_code_grow ("union_members", $3, @3);
442 code_scanner_last_string_free ();
450 "%nterm" { current_class = nterm_sym; } symbol_defs.1
452 current_class = unknown_sym;
455 | "%token" { current_class = token_sym; } symbol_defs.1
457 current_class = unknown_sym;
460 | "%type" TAG symbols.1
464 for (list = $3; list; list = list->next)
465 symbol_type_set (list->content.sym, $2, @2);
466 symbol_list_free ($3);
470 precedence_declaration:
471 precedence_declarator tag.opt symbols.prec
475 for (list = $3; list; list = list->next)
477 symbol_type_set (list->content.sym, current_type, @2);
478 symbol_precedence_set (list->content.sym, current_prec, $1, @1);
480 symbol_list_free ($3);
485 precedence_declarator:
486 "%left" { $$ = left_assoc; }
487 | "%right" { $$ = right_assoc; }
488 | "%nonassoc" { $$ = non_assoc; }
489 | "%precedence" { $$ = precedence_assoc; }
493 %empty { current_type = NULL; }
494 | TAG { current_type = $1; tag_seen = true; }
497 /* Just like symbols.1 but accept INT for the sake of POSIX. */
500 { $$ = symbol_list_sym_new ($1, @1); }
501 | symbols.prec symbol.prec
502 { $$ = symbol_list_append ($1, symbol_list_sym_new ($2, @2)); }
507 | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
510 /* One or more symbols to be %typed. */
513 { $$ = symbol_list_sym_new ($1, @1); }
515 { $$ = symbol_list_append ($1, symbol_list_sym_new ($2, @2)); }
519 generic_symlist_item { $$ = $1; }
520 | generic_symlist generic_symlist_item { $$ = symbol_list_append ($1, $2); }
523 generic_symlist_item:
524 symbol { $$ = symbol_list_sym_new ($1, @1); }
525 | tag { $$ = symbol_list_type_new ($1, @1); }
530 | "<*>" { $$ = uniqstr_new ("*"); }
531 | "<>" { $$ = uniqstr_new (""); }
534 /* One token definition. */
543 symbol_class_set ($1, current_class, @1, true);
544 symbol_type_set ($1, current_type, @1);
548 symbol_class_set ($1, current_class, @1, true);
549 symbol_type_set ($1, current_type, @1);
550 symbol_user_token_number_set ($1, $2, @2);
554 symbol_class_set ($1, current_class, @1, true);
555 symbol_type_set ($1, current_type, @1);
556 symbol_make_alias ($1, $2, @$);
558 | id INT string_as_id
560 symbol_class_set ($1, current_class, @1, true);
561 symbol_type_set ($1, current_type, @1);
562 symbol_user_token_number_set ($1, $2, @2);
563 symbol_make_alias ($1, $3, @$);
567 /* One or more symbol definitions. */
570 | symbol_defs.1 symbol_def
574 /*------------------------------------------.
575 | The grammar section: between the two %%. |
576 `------------------------------------------*/
579 rules_or_grammar_declaration
580 | grammar rules_or_grammar_declaration
583 /* As a Bison extension, one can use the grammar declarations in the
584 body of the grammar. */
585 rules_or_grammar_declaration:
587 | grammar_declaration ";"
595 id_colon named_ref.opt { current_lhs ($1, @1, $2); } rhses.1
597 /* Free the current lhs. */
598 current_lhs (0, @1, 0);
603 rhs { grammar_current_rule_end (@1); }
604 | rhses.1 "|" rhs { grammar_current_rule_end (@3); }
608 %token PERCENT_EMPTY "%empty";
611 { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location,
612 current_lhs_named_ref); }
613 | rhs symbol named_ref.opt
614 { grammar_current_rule_symbol_append ($2, @2, $3); }
615 | rhs "{...}" named_ref.opt
616 { grammar_current_rule_action_append ($2, @2, $3, false); }
618 { grammar_current_rule_action_append ($2, @2, NULL, true); }
620 { grammar_current_rule_empty_set (@2); }
622 { grammar_current_rule_prec_set ($3, @3); }
624 { grammar_current_rule_dprec_set ($3, @3); }
626 { grammar_current_rule_merge_set ($3, @3); }
631 | BRACKETED_ID { $$ = named_ref_new($1, @1); }
634 /*---------------------------.
635 | variable and content.opt. |
636 `---------------------------*/
638 /* The STRING form of variable is deprecated and is not M4-friendly.
639 For example, M4 fails for '%define "[" "value"'. */
642 | STRING { $$ = uniqstr_new ($1); }
645 /* Some content or empty by default. */
649 | STRING { $$ = $1; }
660 code_props plain_code;
661 $1[strlen ($1) - 1] = '\n';
662 code_props_plain_init (&plain_code, $1+1, @1);
663 code_props_translate_code (&plain_code);
664 gram_scanner_last_string_free ();
665 $$ = plain_code.code;
674 /* Identifiers are returned as uniqstr values by the scanner.
675 Depending on their use, we may need to make them genuine symbols. */
679 { $$ = symbol_from_uniqstr ($1, @1); }
682 $$ = symbol_get (char_name ($1), @1);
683 symbol_class_set ($$, token_sym, @1, false);
684 symbol_user_token_number_set ($$, $1, @1);
689 ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
698 /* A string used as an ID: quote it. */
702 $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
703 symbol_class_set ($$, token_sym, @1, false);
711 code_props plain_code;
712 code_props_plain_init (&plain_code, $2, @2);
713 code_props_translate_code (&plain_code);
714 gram_scanner_last_string_free ();
715 muscle_code_grow ("epilogue", plain_code.code, @2);
716 code_scanner_last_string_free ();
722 /* Return the location of the left-hand side of a rule whose
723 right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
724 the right-hand side, and return an empty location equal to the end
725 boundary of RHS[0] if the right-hand side is empty. */
728 lloc_default (YYLTYPE const *rhs, int n)
733 /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
734 The bug is fixed in 7.4.2m, but play it safe for now. */
735 loc.start = rhs[n].end;
736 loc.end = rhs[n].end;
738 /* Ignore empty nonterminals the start of the right-hand side.
739 Do not bother to ignore them at the end of the right-hand side,
740 since empty nonterminals have the same end as their predecessors. */
741 for (i = 1; i <= n; i++)
742 if (! equal_boundaries (rhs[i].start, rhs[i].end))
744 loc.start = rhs[i].start;
753 add_param (param_type type, char *decl, location loc)
755 static char const alphanum[26 + 26 + 1 + 10] =
756 "abcdefghijklmnopqrstuvwxyz"
757 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
761 char const *name_start = NULL;
764 /* Stop on last actual character. */
765 for (p = decl; p[1]; p++)
767 || ! memchr (alphanum, p[-1], sizeof alphanum))
768 && memchr (alphanum, p[0], sizeof alphanum - 10))
771 /* Strip the surrounding '{' and '}', and any blanks just inside
774 while (c_isspace ((unsigned char) *p))
778 while (c_isspace ((unsigned char) *decl))
783 complain (&loc, complaint, _("missing identifier in parameter declaration"));
786 char *name = xmemdup0 (name_start, strspn (name_start, alphanum));
787 if (type & param_lex)
788 muscle_pair_list_grow ("lex_param", decl, name);
789 if (type & param_parse)
790 muscle_pair_list_grow ("parse_param", decl, name);
794 gram_scanner_last_string_free ();
799 version_check (location const *loc, char const *version)
801 if (strverscmp (version, PACKAGE_VERSION) > 0)
803 complain (loc, complaint, "require bison %s, but have %s",
804 version, PACKAGE_VERSION);
810 gram_error (location const *loc, char const *msg)
812 complain (loc, complaint, "%s", msg);
816 token_name (int type)
818 return yytname[YYTRANSLATE (type)];
829 buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
830 return quotearg_style (escape_quoting_style, buf);