X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/b143f4048fe39451ec739d956b4ca1f4fd8e694d..423ca9b50e4321c078cc6fe7903b10ed4907d269:/src/parse-gram.y diff --git a/src/parse-gram.y b/src/parse-gram.y index 4d48d7ad..c932ba54 100644 --- a/src/parse-gram.y +++ b/src/parse-gram.y @@ -1,7 +1,6 @@ %{/* Bison Grammar Parser -*- C -*- - Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software - Foundation, Inc. + Copyright (C) 2002-2012 Free Software Foundation, Inc. This file is part of Bison, the GNU Compiler Compiler. @@ -33,12 +32,13 @@ #include "symlist.h" #include "scan-gram.h" #include "scan-code.h" +#include "xmemdup0.h" #define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N) static YYLTYPE lloc_default (YYLTYPE const *, int); #define YY_LOCATION_PRINT(File, Loc) \ - location_print (File, Loc) + location_print (File, Loc) static void version_check (location const *loc, char const *version); @@ -46,7 +46,7 @@ static void version_check (location const *loc, char const *version); FIXME: depends on the undocumented availability of YYLLOC. */ #undef yyerror #define yyerror(Msg) \ - gram_error (&yylloc, Msg) + gram_error (&yylloc, Msg) static void gram_error (location const *, char const *); static char const *char_name (char); @@ -57,10 +57,28 @@ static char const *char_name (char); static int current_prec = 0; static location current_lhs_location; static named_ref *current_lhs_named_ref; - static symbol *current_lhs; + static symbol *current_lhs_symbol; static symbol_class current_class = unknown_sym; static uniqstr current_type = NULL; + /** Set the new current left-hand side symbol, possibly common + * to several right-hand side parts of rule. + */ + static + void + current_lhs(symbol *sym, location loc, named_ref *ref) + { + current_lhs_symbol = sym; + current_lhs_location = loc; + /* In order to simplify memory management, named references for lhs + are always assigned by deep copy into the current symbol_list + node. This is because a single named-ref in the grammar may + result in several uses when the user factors lhs between several + rules using "|". Therefore free the parser's original copy. */ + free (current_lhs_named_ref); + current_lhs_named_ref = ref; + } + #define YYTYPE_INT16 int_fast16_t #define YYTYPE_INT8 int_fast8_t #define YYTYPE_UINT16 uint_fast16_t @@ -73,6 +91,7 @@ static char const *char_name (char); %locations %pure-parser %define parse.error "verbose" +%define parse.lac full %name-prefix="gram_" %expect 0 @@ -118,7 +137,6 @@ static char const *char_name (char); %token PERCENT_DPREC "%dprec" %token PERCENT_MERGE "%merge" - /*----------------------. | Global Declarations. | `----------------------*/ @@ -151,6 +169,7 @@ static char const *char_name (char); ; %token BRACED_CODE "{...}" +%token BRACED_PREDICATE "%?{...}" %token BRACKETED_ID "[identifier]" %token CHAR "char" %token EPILOGUE "epilogue" @@ -171,11 +190,11 @@ static char const *char_name (char); /* braceless is not to be used for rule or symbol actions, as it calls code_props_plain_init. */ %type STRING "%{...%}" EPILOGUE braceless content.opt -%type "{...}" +%type "{...}" "%?{...}" %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); } - STRING + STRING %printer { fprintf (stderr, "{\n%s\n}", $$); } - braceless content.opt "{...}" "%{...%}" EPILOGUE + braceless content.opt "{...}" "%{...%}" EPILOGUE %type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG variable %printer { fputs ($$, stderr); } @@ -252,9 +271,9 @@ input: ; - /*------------------------------------. - | Declarations: before the first %%. | - `------------------------------------*/ + /*------------------------------------. + | Declarations: before the first %%. | + `------------------------------------*/ prologue_declarations: /* Nothing */ @@ -294,7 +313,7 @@ prologue_declaration: MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); } | "%expect" INT { expected_sr_conflicts = $2; } -| "%expect-rr" INT { expected_rr_conflicts = $2; } +| "%expect-rr" INT { expected_rr_conflicts = $2; } | "%file-prefix" STRING { spec_file_prefix = $2; } | "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */ | "%glr-parser" @@ -311,11 +330,11 @@ prologue_declaration: muscle_code_grow ("initial_action", action.code, @2); code_scanner_last_string_free (); } -| "%language" STRING { language_argmatch ($2, grammar_prio, @1); } +| "%language" STRING { language_argmatch ($2, grammar_prio, @1); } | "%name-prefix" STRING { spec_name_prefix = $2; } | "%name-prefix" "=" STRING { spec_name_prefix = $3; } /* deprecated */ | "%no-lines" { no_lines_flag = true; } -| "%nondeterministic-parser" { nondeterministic_parser = true; } +| "%nondeterministic-parser" { nondeterministic_parser = true; } | "%output" STRING { spec_outfile = $2; } | "%output" "=" STRING { spec_outfile = $3; } /* deprecated */ | "%param" { current_param = $1; } params { current_param = param_none; } @@ -335,7 +354,7 @@ prologue_declaration: xmalloc (dir_length + 1 + strlen (skeleton_user) + 1); if (dir_length > 0) { - strncpy (skeleton_build, current_file, dir_length); + memcpy (skeleton_build, current_file, dir_length); skeleton_build[dir_length++] = '/'; } strcpy (skeleton_build + dir_length, skeleton_user); @@ -371,14 +390,14 @@ grammar_declaration: { symbol_list *list; for (list = $3; list; list = list->next) - symbol_list_destructor_set (list, $2, @2); + symbol_list_code_props_set (list, destructor, @2, $2); symbol_list_free ($3); } | "%printer" "{...}" generic_symlist { symbol_list *list; for (list = $3; list; list = list->next) - symbol_list_printer_set (list, $2, @2); + symbol_list_code_props_set (list, printer, @2, $2); symbol_list_free ($3); } | "%default-prec" @@ -443,7 +462,7 @@ symbol_declaration: symbol_list *list; tag_seen = true; for (list = $3; list; list = list->next) - symbol_type_set (list->content.sym, $2, @2); + symbol_type_set (list->content.sym, $2, @2); symbol_list_free ($3); } ; @@ -454,10 +473,10 @@ precedence_declaration: symbol_list *list; ++current_prec; for (list = $3; list; list = list->next) - { - symbol_type_set (list->content.sym, current_type, @2); - symbol_precedence_set (list->content.sym, current_prec, $1, @1); - } + { + symbol_type_set (list->content.sym, current_type, @2); + symbol_precedence_set (list->content.sym, current_prec, $1, @1); + } symbol_list_free ($3); current_type = NULL; } @@ -548,9 +567,9 @@ symbol_defs.1: ; - /*------------------------------------------. - | The grammar section: between the two %%. | - `------------------------------------------*/ + /*------------------------------------------. + | The grammar section: between the two %%. | + `------------------------------------------*/ grammar: rules_or_grammar_declaration @@ -569,8 +588,11 @@ rules_or_grammar_declaration: ; rules: - id_colon named_ref.opt { current_lhs = $1; current_lhs_location = @1; - current_lhs_named_ref = $2; } rhses.1 + id_colon named_ref.opt { current_lhs ($1, @1, $2); } rhses.1 + { + /* Free the current lhs. */ + current_lhs (0, @1, 0); + } ; rhses.1: @@ -581,12 +603,14 @@ rhses.1: rhs: /* Nothing. */ - { grammar_current_rule_begin (current_lhs, current_lhs_location, - current_lhs_named_ref); } + { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location, + current_lhs_named_ref); } | rhs symbol named_ref.opt { grammar_current_rule_symbol_append ($2, @2, $3); } | rhs "{...}" named_ref.opt - { grammar_current_rule_action_append ($2, @2, $3); } + { grammar_current_rule_action_append ($2, @2, $3, false); } +| rhs "%?{...}" + { grammar_current_rule_action_append ($2, @2, NULL, true); } | rhs "%prec" symbol { grammar_current_rule_prec_set ($3, @3); } | rhs "%dprec" INT @@ -601,7 +625,6 @@ named_ref.opt: BRACKETED_ID { $$ = named_ref_new($1, @1); } ; - /*---------------------------. | variable and content.opt. | `---------------------------*/ @@ -707,14 +730,14 @@ lloc_default (YYLTYPE const *rhs, int n) loc.start = rhs[n].end; loc.end = rhs[n].end; - /* Ignore empty nonterminals the start of the the right-hand side. + /* Ignore empty nonterminals the start of the right-hand side. Do not bother to ignore them at the end of the right-hand side, since empty nonterminals have the same end as their predecessors. */ for (i = 1; i <= n; i++) if (! equal_boundaries (rhs[i].start, rhs[i].end)) { - loc.start = rhs[i].start; - break; + loc.start = rhs[i].start; + break; } return loc; @@ -753,17 +776,7 @@ add_param (param_type type, char *decl, location loc) complain_at (loc, _("missing identifier in parameter declaration")); else { - char *name; - size_t name_len; - - for (name_len = 1; - memchr (alphanum, name_start[name_len], sizeof alphanum); - name_len++) - continue; - - name = xmalloc (name_len + 1); - memcpy (name, name_start, name_len); - name[name_len] = '\0'; + char *name = xmemdup0 (name_start, strspn (name_start, alphanum)); if (type & param_lex) muscle_pair_list_grow ("lex_param", decl, name); if (type & param_parse) @@ -781,8 +794,8 @@ version_check (location const *loc, char const *version) if (strverscmp (version, PACKAGE_VERSION) > 0) { complain_at (*loc, "require bison %s, but have %s", - version, PACKAGE_VERSION); - exit (63); + version, PACKAGE_VERSION); + exit (EX_MISMATCH); } }