X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/a7706735d143e3d7ab255e162499584e5337ba72..775673717fc9b7e5174418f7420ba21040a8e675:/src/parse-gram.y diff --git a/src/parse-gram.y b/src/parse-gram.y index 683a05ff..532349bb 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. @@ -38,7 +37,7 @@ 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 +45,7 @@ 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); @@ -54,12 +53,30 @@ static char const *char_name (char); %code { - static symbol_class current_class = unknown_sym; - static uniqstr current_type = NULL; - static symbol *current_lhs; + static int current_prec = 0; static location current_lhs_location; static named_ref *current_lhs_named_ref; - static int current_prec = 0; + 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 @@ -73,6 +90,7 @@ static char const *char_name (char); %locations %pure-parser %define parse.error "verbose" +%define parse.lac full %name-prefix="gram_" %expect 0 @@ -118,7 +136,6 @@ static char const *char_name (char); %token PERCENT_DPREC "%dprec" %token PERCENT_MERGE "%merge" - /*----------------------. | Global Declarations. | `----------------------*/ @@ -151,6 +168,8 @@ static char const *char_name (char); ; %token BRACED_CODE "{...}" +%token BRACED_PREDICATE "%?{...}" +%token BRACKETED_ID "[identifier]" %token CHAR "char" %token EPILOGUE "epilogue" %token EQUAL "=" @@ -163,7 +182,6 @@ static char const *char_name (char); %token TAG "" %token TAG_ANY "<*>" %token TAG_NONE "<>" -%token BRACKETED_ID "[identifier]" %type CHAR %printer { fputs (char_name ($$), stderr); } CHAR @@ -171,28 +189,29 @@ static char const *char_name (char); /* braceless is not to be used for rule or symbol actions, as it calls code_props_plain_init. */ %type STRING "%{...%}" EPILOGUE braceless content.opt -%type "{...}" +%type "{...}" "%?{...}" %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); } - STRING + STRING %printer { fprintf (stderr, "{\n%s\n}", $$); } - braceless content.opt "{...}" "%{...%}" EPILOGUE + braceless content.opt "{...}" "%{...%}" EPILOGUE -%type TAG ID ID_COLON BRACKETED_ID PERCENT_FLAG variable -%type named_ref.opt -%printer { fputs ($$, stderr); } ID variable +%type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG variable +%printer { fputs ($$, stderr); } +%printer { fprintf (stderr, "[%s]", $$); } BRACKETED_ID %printer { fprintf (stderr, "%s:", $$); } ID_COLON %printer { fprintf (stderr, "%%%s", $$); } PERCENT_FLAG %printer { fprintf (stderr, "<%s>", $$); } TAG %type INT -%printer { fprintf (stderr, "%d", $$); } INT +%printer { fprintf (stderr, "%d", $$); } -%type id id_colon symbol symbol.prec string_as_id -%printer { fprintf (stderr, "%s", $$->tag); } id symbol string_as_id +%type id id_colon string_as_id symbol symbol.prec +%printer { fprintf (stderr, "%s", $$->tag); } %printer { fprintf (stderr, "%s:", $$->tag); } id_colon %type precedence_declarator %type symbols.1 symbols.prec generic_symlist generic_symlist_item +%type named_ref.opt /*---------. | %param. | @@ -203,6 +222,7 @@ static char const *char_name (char); # define PARAM_TYPE typedef enum { + param_none = 0, param_lex = 1 << 0, param_parse = 1 << 1, param_both = param_lex | param_parse @@ -218,6 +238,7 @@ static char const *char_name (char); * \param loc the location in the source. */ static void add_param (param_type type, char *decl, location loc); + static param_type current_param = param_none; }; %union { @@ -229,15 +250,19 @@ static char const *char_name (char); switch ($$) { #define CASE(In, Out) \ - case param_ ## In: fputs ("%" #Out, stderr); break - + case param_ ## In: fputs ("%" #Out, stderr); break CASE(lex, lex-param); CASE(parse, parse-param); CASE(both, param); - } #undef CASE + case param_none: aver (false); break; + } } ; + + /*==========\ + | Grammar. | + \==========*/ %% input: @@ -245,9 +270,9 @@ input: ; - /*------------------------------------. - | Declarations: before the first %%. | - `------------------------------------*/ + /*------------------------------------. + | Declarations: before the first %%. | + `------------------------------------*/ prologue_declarations: /* Nothing */ @@ -287,7 +312,7 @@ 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" @@ -304,19 +329,19 @@ 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" "{...}" { add_param ($1, $2, @2); } +| "%param" { current_param = $1; } params { current_param = param_none; } | "%require" STRING { version_check (&@2, $2); } | "%skeleton" STRING { char const *skeleton_user = $2; - if (strchr (skeleton_user, '/')) + if (mbschr (skeleton_user, '/')) { size_t dir_length = strlen (current_file); char *skeleton_build; @@ -343,6 +368,11 @@ prologue_declaration: | /*FIXME: Err? What is this horror doing here? */ ";" ; +params: + params "{...}" { add_param (current_param, $2, @2); } +| "{...}" { add_param (current_param, $1, @1); } +; + /*----------------------. | grammar_declaration. | @@ -359,14 +389,14 @@ grammar_declaration: { symbol_list *list; for (list = $3; list; list = list->next) - symbol_list_destructor_set (list, $2, @2); + symbol_list_destructor_set (list, $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_printer_set (list, $2, @2); symbol_list_free ($3); } | "%default-prec" @@ -431,7 +461,7 @@ 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); } ; @@ -442,10 +472,10 @@ 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; } @@ -536,9 +566,9 @@ symbol_defs.1: ; - /*------------------------------------------. - | The grammar section: between the two %%. | - `------------------------------------------*/ + /*------------------------------------------. + | The grammar section: between the two %%. | + `------------------------------------------*/ grammar: rules_or_grammar_declaration @@ -557,8 +587,11 @@ 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: @@ -569,12 +602,14 @@ 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 @@ -589,7 +624,6 @@ named_ref.opt: BRACKETED_ID { $$ = named_ref_new($1, @1); } ; - /*---------------------------. | variable and content.opt. | `---------------------------*/ @@ -695,14 +729,14 @@ 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; @@ -717,23 +751,25 @@ add_param (param_type type, char *decl, location loc) "ABCDEFGHIJKLMNOPQRSTUVWXYZ" "_" "0123456789"; + char const *name_start = NULL; - char *p; - - /* Stop on last actual character. */ - for (p = decl; p[1]; p++) - if ((p == decl - || ! memchr (alphanum, p[-1], sizeof alphanum)) - && memchr (alphanum, p[0], sizeof alphanum - 10)) - name_start = p; - - /* Strip the surrounding '{' and '}', and any blanks just inside - the braces. */ - while (*--p == ' ' || *p == '\t') - continue; - p[1] = '\0'; - while (*++decl == ' ' || *decl == '\t') - continue; + { + char *p; + /* Stop on last actual character. */ + for (p = decl; p[1]; p++) + if ((p == decl + || ! memchr (alphanum, p[-1], sizeof alphanum)) + && memchr (alphanum, p[0], sizeof alphanum - 10)) + name_start = p; + + /* Strip the surrounding '{' and '}', and any blanks just inside + the braces. */ + while (*--p == ' ' || *p == '\t') + continue; + p[1] = '\0'; + while (*++decl == ' ' || *decl == '\t') + continue; + } if (! name_start) complain_at (loc, _("missing identifier in parameter declaration")); @@ -743,9 +779,9 @@ add_param (param_type type, char *decl, location loc) size_t name_len; for (name_len = 1; - memchr (alphanum, name_start[name_len], sizeof alphanum); - name_len++) - continue; + memchr (alphanum, name_start[name_len], sizeof alphanum); + name_len++) + continue; name = xmalloc (name_len + 1); memcpy (name, name_start, name_len); @@ -767,8 +803,8 @@ 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); } }