X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/66c209cf435b5bfc5cf982d8df9cc3ef05b4546b..3f21a394f481c829831592adcf3ea040f87805ee:/src/parse-gram.y?ds=sidebyside diff --git a/src/parse-gram.y b/src/parse-gram.y index e889eb5e..8b57ce80 100644 --- a/src/parse-gram.y +++ b/src/parse-gram.y @@ -1,4 +1,4 @@ -%{/* Bison Grammar Parser -*- C -*- +/* Bison Grammar Parser -*- C -*- Copyright (C) 2002-2013 Free Software Foundation, Inc. @@ -17,41 +17,30 @@ You should have received a copy of the GNU General Public License along with this program. If not, see . */ -#include -#include "system.h" - -#include "c-ctype.h" -#include "complain.h" -#include "conflicts.h" -#include "files.h" -#include "getargs.h" -#include "gram.h" -#include "muscle-tab.h" -#include "named-ref.h" -#include "quotearg.h" -#include "reader.h" -#include "symlist.h" -#include "symtab.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 (Loc, File) - -static void version_check (location const *loc, char const *version); - -static void gram_error (location const *, char const *); - -/* A string that describes a char (e.g., 'a' -> "'a'"). */ -static char const *char_name (char); -%} +%code requires +{ + #include "symlist.h" + #include "symtab.h" +} %code { + #include + #include "system.h" + + #include "c-ctype.h" + #include "complain.h" + #include "conflicts.h" + #include "files.h" + #include "getargs.h" + #include "gram.h" + #include "named-ref.h" + #include "quotearg.h" + #include "reader.h" + #include "scan-gram.h" + #include "scan-code.h" + #include "xmemdup0.h" + static int current_prec = 0; static location current_lhs_location; static named_ref *current_lhs_named_ref; @@ -62,20 +51,21 @@ static char const *char_name (char); /** 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; - } + static void current_lhs (symbol *sym, location loc, named_ref *ref); + + #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 (Loc, File) + + static void version_check (location const *loc, char const *version); + + static void gram_error (location const *, char const *); + + /* A string that describes a char (e.g., 'a' -> "'a'"). */ + static char const *char_name (char); #define YYTYPE_INT16 int_fast16_t #define YYTYPE_INT8 int_fast8_t @@ -83,7 +73,7 @@ static char const *char_name (char); #define YYTYPE_UINT8 uint_fast8_t } -%define api.prefix "gram_" +%define api.prefix {gram_} %define api.pure full %define locations %define parse.error verbose @@ -101,23 +91,9 @@ static char const *char_name (char); boundary_set (&@$.end, current_file, 1, 1); } -%union -{ - assoc assoc; - char *code; - char const *chars; - int integer; - named_ref *named_ref; - symbol *symbol; - symbol_list *list; - uniqstr uniqstr; - unsigned char character; -}; - /* Define the tokens together with their human representation. */ %token GRAM_EOF 0 "end of file" %token STRING "string" -%token INT "integer" %token PERCENT_TOKEN "%token" %token PERCENT_NTERM "%nterm" @@ -182,18 +158,25 @@ static char const *char_name (char); %token TAG_ANY "<*>" %token TAG_NONE "<>" +%union {unsigned char character;} %type CHAR %printer { fputs (char_name ($$), yyo); } CHAR /* braceless is not to be used for rule or symbol actions, as it calls code_props_plain_init. */ -%type STRING "%{...%}" EPILOGUE braceless content.opt +%union +{ + char *code; + char const *chars; +}; +%type STRING "%{...%}" EPILOGUE braceless %type "{...}" "%?{...}" %printer { fputs (quotearg_style (c_quoting_style, $$), yyo); } STRING %printer { fprintf (yyo, "{\n%s\n}", $$); } - braceless content.opt "{...}" "%{...%}" EPILOGUE + braceless "{...}" "%{...%}" EPILOGUE +%union {uniqstr uniqstr;} %type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable %printer { fputs ($$, yyo); } %printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID @@ -201,15 +184,22 @@ static char const *char_name (char); %printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG %printer { fprintf (yyo, "<%s>", $$); } TAG tag -%type INT +%union {int integer;}; +%token INT "integer" %printer { fprintf (yyo, "%d", $$); } +%union {symbol *symbol;} %type id id_colon string_as_id symbol symbol.prec %printer { fprintf (yyo, "%s", $$->tag); } %printer { fprintf (yyo, "%s:", $$->tag); } id_colon +%union {assoc assoc;}; %type precedence_declarator + +%union {symbol_list *list;} %type symbols.1 symbols.prec generic_symlist generic_symlist_item + +%union {named_ref *named_ref;} %type named_ref.opt /*---------. @@ -217,8 +207,6 @@ static char const *char_name (char); `---------*/ %code requires { -# ifndef PARAM_TYPE -# define PARAM_TYPE typedef enum { param_none = 0, @@ -226,7 +214,6 @@ static char const *char_name (char); param_parse = 1 << 1, param_both = param_lex | param_parse } param_type; -# endif }; %code { @@ -239,17 +226,14 @@ static char const *char_name (char); static void add_param (param_type type, char *decl, location loc); static param_type current_param = param_none; }; -%union -{ - param_type param; -} +%union {param_type param;} %token PERCENT_PARAM "%param"; %printer { switch ($$) { #define CASE(In, Out) \ - case param_ ## In: fputs ("%" #Out, stderr); break + case param_ ## In: fputs ("%" #Out, yyo); break CASE (lex, lex-param); CASE (parse, parse-param); CASE (both, param); @@ -274,7 +258,7 @@ input: `------------------------------------*/ prologue_declarations: - /* Nothing */ + %empty | prologue_declarations prologue_declaration ; @@ -294,9 +278,9 @@ prologue_declaration: { muscle_percent_define_ensure ($1, @1, true); } -| "%define" variable content.opt +| "%define" variable value { - muscle_percent_define_insert ($2, @2, $3, + muscle_percent_define_insert ($2, @2, $3.kind, $3.chars, MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); } | "%defines" { defines_flag = true; } @@ -307,7 +291,8 @@ prologue_declaration: } | "%error-verbose" { - muscle_percent_define_insert ("parse.error", @1, "verbose", + muscle_percent_define_insert ("parse.error", @1, muscle_keyword, + "verbose", MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE); } | "%expect" INT { expected_sr_conflicts = $2; } @@ -430,8 +415,8 @@ code_props_type: %token PERCENT_UNION "%union"; union_name: - /* Nothing. */ {} -| ID { muscle_code_grow ("union_name", $1, @1); } + %empty {} +| ID { muscle_code_grow ("union_name", $1, @1); } ; grammar_declaration: @@ -490,8 +475,8 @@ precedence_declarator: ; tag.opt: - /* Nothing. */ { current_type = NULL; } -| TAG { current_type = $1; tag_seen = true; } + %empty { current_type = NULL; } +| TAG { current_type = $1; tag_seen = true; } ; /* Just like symbols.1 but accept INT for the sake of POSIX. */ @@ -503,8 +488,17 @@ symbols.prec: ; symbol.prec: - symbol { $$ = $1; } -| symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); } + symbol + { + $$ = $1; + symbol_class_set ($1, token_sym, @1, false); + } +| symbol INT + { + $$ = $1; + symbol_user_token_number_set ($1, $2, @2); + symbol_class_set ($1, token_sym, @1, false); + } ; /* One or more symbols to be %typed. */ @@ -605,8 +599,9 @@ rhses.1: | rhses.1 ";" ; +%token PERCENT_EMPTY "%empty"; rhs: - /* Nothing. */ + %empty { grammar_current_rule_begin (current_lhs_symbol, current_lhs_location, current_lhs_named_ref); } | rhs symbol named_ref.opt @@ -615,6 +610,8 @@ rhs: { grammar_current_rule_action_append ($2, @2, $3, false); } | rhs "%?{...}" { grammar_current_rule_action_append ($2, @2, NULL, true); } +| rhs "%empty" + { grammar_current_rule_empty_set (@2); } | rhs "%prec" symbol { grammar_current_rule_prec_set ($3, @3); } | rhs "%dprec" INT @@ -624,27 +621,47 @@ rhs: ; named_ref.opt: - /* Nothing. */ { $$ = 0; } -| - BRACKETED_ID { $$ = named_ref_new ($1, @1); } + %empty { $$ = 0; } +| BRACKETED_ID { $$ = named_ref_new($1, @1); } ; -/*---------------------------. -| variable and content.opt. | -`---------------------------*/ +/*---------------------. +| variable and value. | +`---------------------*/ /* The STRING form of variable is deprecated and is not M4-friendly. - For example, M4 fails for `%define "[" "value"'. */ + For example, M4 fails for '%define "[" "value"'. */ variable: ID | STRING { $$ = uniqstr_new ($1); } ; /* Some content or empty by default. */ -content.opt: - /* Nothing. */ { $$ = ""; } -| ID { $$ = $1; } -| STRING +%code requires {#include "muscle-tab.h"}; +%union +{ + struct + { + char const *chars; + muscle_kind kind; + } value; +}; +%type value; +%printer +{ + switch ($$.kind) + { + case muscle_code: fprintf (yyo, "{%s}", $$.chars); break; + case muscle_keyword: fprintf (yyo, "%s", $$.chars); break; + case muscle_string: fprintf (yyo, "\"%s\"", $$.chars); break; + } +} ; + +value: + %empty { $$.kind = muscle_keyword; $$.chars = ""; } +| ID { $$.kind = muscle_keyword; $$.chars = $1; } +| STRING { $$.kind = muscle_string; $$.chars = $1; } +| braceless { $$.kind = muscle_code; $$.chars = $1; } ; @@ -656,7 +673,7 @@ braceless: "{...}" { code_props plain_code; - $1[strlen ($1) - 1] = '\n'; + $1[strlen ($1) - 1] = '\0'; code_props_plain_init (&plain_code, $1+1, @1); code_props_translate_code (&plain_code); gram_scanner_last_string_free (); @@ -703,7 +720,7 @@ string_as_id: ; epilogue.opt: - /* Nothing. */ + %empty | "%%" EPILOGUE { code_props plain_code; @@ -828,3 +845,18 @@ char_name (char c) return quotearg_style (escape_quoting_style, buf); } } + +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; +}