X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/22b081deb7ab1d0ec7a279e5213fa7fdc2c85207..9402b623533a83d8aade70988622ad7e5f063117:/src/parse-gram.y
diff --git a/src/parse-gram.y b/src/parse-gram.y
index c932ba54..9a266251 100644
--- a/src/parse-gram.y
+++ b/src/parse-gram.y
@@ -1,6 +1,6 @@
-%{/* Bison Grammar Parser -*- C -*-
+/* Bison Grammar Parser -*- C -*-
- Copyright (C) 2002-2012 Free Software Foundation, Inc.
+ Copyright (C) 2002-2013 Free Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
@@ -17,43 +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 "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 "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)
-
-static void version_check (location const *loc, char const *version);
-
-/* Request detailed syntax error messages, and pass them to GRAM_ERROR.
- FIXME: depends on the undocumented availability of YYLLOC. */
-#undef yyerror
-#define yyerror(Msg) \
- gram_error (&yylloc, Msg)
-static void gram_error (location const *, char const *);
-
-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;
@@ -64,20 +51,37 @@ 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)
+
+ /* Strip initial '{' and final '}' (must be first and last characters).
+ Return the result. */
+ static char *strip_braces (char *code);
+
+ /* Convert CODE by calling code_props_plain_init if PLAIN, otherwise
+ code_props_symbol_action_init. Call
+ gram_scanner_last_string_free to release the latest string from
+ the scanner (should be CODE). */
+ static char const *translate_code (char *code, location loc, bool plain);
+
+ /* Convert CODE by calling code_props_plain_init after having
+ stripped the first and last characters (expected to be '{', and
+ '}'). Call gram_scanner_last_string_free to release the latest
+ string from the scanner (should be CODE). */
+ static char const *translate_code_braceless (char *code, location loc);
+
+ 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
@@ -85,15 +89,15 @@ static char const *char_name (char);
#define YYTYPE_UINT8 uint_fast8_t
}
-%debug
-%verbose
-%defines
-%locations
-%pure-parser
-%define parse.error "verbose"
+%define api.prefix {gram_}
+%define api.pure full
+%define locations
+%define parse.error verbose
%define parse.lac full
-%name-prefix="gram_"
+%define parse.trace
+%defines
%expect 0
+%verbose
%initial-action
{
@@ -103,23 +107,9 @@ 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"
@@ -184,34 +174,39 @@ static char const *char_name (char);
%token TAG_ANY "<*>"
%token TAG_NONE "<>"
+%union {unsigned char character;}
%type CHAR
-%printer { fputs (char_name ($$), stderr); } CHAR
-
-/* braceless is not to be used for rule or symbol actions, as it
- calls code_props_plain_init. */
-%type STRING "%{...%}" EPILOGUE braceless content.opt
-%type "{...}" "%?{...}"
-%printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
- STRING
-%printer { fprintf (stderr, "{\n%s\n}", $$); }
- braceless content.opt "{...}" "%{...%}" EPILOGUE
-
-%type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG variable
-%printer { fputs ($$, 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", $$); }
-
+%printer { fputs (char_name ($$), yyo); } CHAR
+
+%union {char *code;};
+%type "{...}" "%?{...}" "%{...%}" EPILOGUE STRING
+%printer { fputs (quotearg_style (c_quoting_style, $$), yyo); } STRING
+%printer { fprintf (yyo, "{\n%s\n}", $$); }
+
+%union {uniqstr uniqstr;}
+%type BRACKETED_ID ID ID_COLON PERCENT_FLAG TAG tag variable
+%printer { fputs ($$, yyo); }
+%printer { fprintf (yyo, "[%s]", $$); } BRACKETED_ID
+%printer { fprintf (yyo, "%s:", $$); } ID_COLON
+%printer { fprintf (yyo, "%%%s", $$); } PERCENT_FLAG
+%printer { fprintf (yyo, "<%s>", $$); } TAG tag
+
+%union {int integer;};
+%token INT "integer"
+%printer { fprintf (yyo, "%d", $$); }
+
+%union {symbol *symbol;}
%type id id_colon string_as_id symbol symbol.prec
-%printer { fprintf (stderr, "%s", $$->tag); }
-%printer { fprintf (stderr, "%s:", $$->tag); } id_colon
+%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
/*---------.
@@ -219,8 +214,6 @@ static char const *char_name (char);
`---------*/
%code requires
{
-# ifndef PARAM_TYPE
-# define PARAM_TYPE
typedef enum
{
param_none = 0,
@@ -228,7 +221,6 @@ static char const *char_name (char);
param_parse = 1 << 1,
param_both = param_lex | param_parse
} param_type;
-# endif
};
%code
{
@@ -241,20 +233,17 @@ 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(lex, lex-param);
- CASE(parse, parse-param);
- CASE(both, param);
+ case param_ ## In: fputs ("%" #Out, yyo); break
+ CASE (lex, lex-param);
+ CASE (parse, parse-param);
+ CASE (both, param);
#undef CASE
case param_none: aver (false); break;
}
@@ -276,7 +265,7 @@ input:
`------------------------------------*/
prologue_declarations:
- /* Nothing */
+ %empty
| prologue_declarations prologue_declaration
;
@@ -284,21 +273,17 @@ prologue_declaration:
grammar_declaration
| "%{...%}"
{
- code_props plain_code;
- code_props_plain_init (&plain_code, $1, @1);
- code_props_translate_code (&plain_code);
- gram_scanner_last_string_free ();
muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
- plain_code.code, @1);
+ translate_code ($1, @1, true), @1);
code_scanner_last_string_free ();
}
| "%"
{
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; }
@@ -309,13 +294,13 @@ 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; }
| "%expect-rr" INT { expected_rr_conflicts = $2; }
| "%file-prefix" STRING { spec_file_prefix = $2; }
-| "%file-prefix" "=" STRING { spec_file_prefix = $3; } /* deprecated */
| "%glr-parser"
{
nondeterministic_parser = true;
@@ -323,20 +308,14 @@ prologue_declaration:
}
| "%initial-action" "{...}"
{
- code_props action;
- code_props_symbol_action_init (&action, $2, @2);
- code_props_translate_code (&action);
- gram_scanner_last_string_free ();
- muscle_code_grow ("initial_action", action.code, @2);
+ muscle_code_grow ("initial_action", translate_code ($2, @2, false), @2);
code_scanner_last_string_free ();
}
| "%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; }
| "%output" STRING { spec_outfile = $2; }
-| "%output" "=" STRING { spec_outfile = $3; } /* deprecated */
| "%param" { current_param = $1; } params { current_param = param_none; }
| "%require" STRING { version_check (&@2, $2); }
| "%skeleton" STRING
@@ -386,19 +365,17 @@ grammar_declaration:
{
grammar_start_symbol_set ($2, @2);
}
-| "%destructor" "{...}" generic_symlist
+| code_props_type "{...}" generic_symlist
{
- symbol_list *list;
- for (list = $3; list; list = list->next)
- 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_code_props_set (list, printer, @2, $2);
- symbol_list_free ($3);
+ code_props code;
+ code_props_symbol_action_init (&code, $2, @2);
+ code_props_translate_code (&code);
+ {
+ symbol_list *list;
+ for (list = $3; list; list = list->next)
+ symbol_list_code_props_set (list, $1, &code);
+ symbol_list_free ($3);
+ }
}
| "%default-prec"
{
@@ -408,20 +385,28 @@ grammar_declaration:
{
default_prec = false;
}
-| "%code" braceless
+| "%code" "{...}"
{
/* Do not invoke muscle_percent_code_grow here since it invokes
muscle_user_name_list_grow. */
- muscle_code_grow ("percent_code()", $2, @2);
+ muscle_code_grow ("percent_code()",
+ translate_code_braceless ($2, @2), @2);
code_scanner_last_string_free ();
}
-| "%code" ID braceless
+| "%code" ID "{...}"
{
- muscle_percent_code_grow ($2, @2, $3, @3);
+ muscle_percent_code_grow ($2, @2, translate_code_braceless ($3, @3), @3);
code_scanner_last_string_free ();
}
;
+%type code_props_type;
+%union {code_props_type code_type;};
+%printer { fprintf (yyo, "%s", code_props_type_string ($$)); } ;
+code_props_type:
+ "%destructor" { $$ = destructor; }
+| "%printer" { $$ = printer; }
+;
/*---------.
| %union. |
@@ -430,15 +415,15 @@ grammar_declaration:
%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:
- "%union" union_name braceless
+ "%union" union_name "{...}"
{
union_seen = true;
- muscle_code_grow ("stype", $3, @3);
+ muscle_code_grow ("union_members", translate_code_braceless ($3, @3), @3);
code_scanner_last_string_free ();
}
;
@@ -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. */
@@ -499,46 +484,59 @@ symbols.prec:
symbol.prec
{ $$ = symbol_list_sym_new ($1, @1); }
| symbols.prec symbol.prec
- { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
+ { $$ = symbol_list_append ($1, symbol_list_sym_new ($2, @2)); }
;
symbol.prec:
- symbol { $$ = $1; }
- | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
- ;
+ symbol
+ {
+ $$ = $1;
+ symbol_class_set ($1, token_sym, @1, false);
+ }
+| symbol INT
+ {
+ $$ = $1;
+ symbol_user_token_number_set ($1, $2, @2);
+ symbol_class_set ($1, token_sym, @1, false);
+ }
+;
/* One or more symbols to be %typed. */
symbols.1:
symbol
{ $$ = symbol_list_sym_new ($1, @1); }
| symbols.1 symbol
- { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
+ { $$ = symbol_list_append ($1, symbol_list_sym_new ($2, @2)); }
;
generic_symlist:
generic_symlist_item { $$ = $1; }
-| generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
+| generic_symlist generic_symlist_item { $$ = symbol_list_append ($1, $2); }
;
generic_symlist_item:
symbol { $$ = symbol_list_sym_new ($1, @1); }
-| TAG { $$ = symbol_list_type_new ($1, @1); }
-| "<*>" { $$ = symbol_list_default_tagged_new (@1); }
-| "<>" { $$ = symbol_list_default_tagless_new (@1); }
+| tag { $$ = symbol_list_type_new ($1, @1); }
+;
+
+tag:
+ TAG
+| "<*>" { $$ = uniqstr_new ("*"); }
+| "<>" { $$ = uniqstr_new (""); }
;
/* One token definition. */
symbol_def:
TAG
- {
- current_type = $1;
- tag_seen = true;
- }
+ {
+ current_type = $1;
+ tag_seen = true;
+ }
| id
- {
- symbol_class_set ($1, current_class, @1, true);
- symbol_type_set ($1, current_type, @1);
- }
+ {
+ symbol_class_set ($1, current_class, @1, true);
+ symbol_type_set ($1, current_type, @1);
+ }
| id INT
{
symbol_class_set ($1, current_class, @1, true);
@@ -601,8 +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
@@ -611,6 +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
@@ -620,44 +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 { $$ = $1; }
-| STRING
-;
-
-
-/*------------.
-| braceless. |
-`------------*/
-
-braceless:
- "{...}"
+%code requires {#include "muscle-tab.h"};
+%union
+{
+ struct
+ {
+ char const *chars;
+ muscle_kind kind;
+ } value;
+};
+%type value;
+%printer
+{
+ switch ($$.kind)
{
- code_props plain_code;
- $1[strlen ($1) - 1] = '\n';
- code_props_plain_init (&plain_code, $1+1, @1);
- code_props_translate_code (&plain_code);
- gram_scanner_last_string_free ();
- $$ = plain_code.code;
+ case muscle_code: fprintf (yyo, "{%s}", $$.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; }
+| "{...}" { $$.kind = muscle_code; $$.chars = strip_braces ($1); }
;
@@ -699,21 +703,16 @@ string_as_id:
;
epilogue.opt:
- /* Nothing. */
+ %empty
| "%%" EPILOGUE
{
- code_props plain_code;
- code_props_plain_init (&plain_code, $2, @2);
- code_props_translate_code (&plain_code);
- gram_scanner_last_string_free ();
- muscle_code_grow ("epilogue", plain_code.code, @2);
+ muscle_code_grow ("epilogue", translate_code ($2, @2, true), @2);
code_scanner_last_string_free ();
}
;
%%
-
/* Return the location of the left-hand side of a rule whose
right-hand side is RHS[1] ... RHS[N]. Ignore empty nonterminals in
the right-hand side, and return an empty location equal to the end
@@ -743,6 +742,33 @@ lloc_default (YYLTYPE const *rhs, int n)
return loc;
}
+static
+char *strip_braces (char *code)
+{
+ code[strlen(code) - 1] = 0;
+ return code + 1;
+}
+
+static
+char const *
+translate_code (char *code, location loc, bool plain)
+{
+ code_props plain_code;
+ if (plain)
+ code_props_plain_init (&plain_code, code, loc);
+ else
+ code_props_symbol_action_init (&plain_code, code, loc);
+ code_props_translate_code (&plain_code);
+ gram_scanner_last_string_free ();
+ return plain_code.code;
+}
+
+static
+char const *
+translate_code_braceless (char *code, location loc)
+{
+ return translate_code (strip_braces (code), loc, true);
+}
static void
add_param (param_type type, char *decl, location loc)
@@ -765,15 +791,17 @@ add_param (param_type type, char *decl, location loc)
/* Strip the surrounding '{' and '}', and any blanks just inside
the braces. */
- while (*--p == ' ' || *p == '\t')
- continue;
+ --p;
+ while (c_isspace ((unsigned char) *p))
+ --p;
p[1] = '\0';
- while (*++decl == ' ' || *decl == '\t')
- continue;
+ ++decl;
+ while (c_isspace ((unsigned char) *decl))
+ ++decl;
}
if (! name_start)
- complain_at (loc, _("missing identifier in parameter declaration"));
+ complain (&loc, complaint, _("missing identifier in parameter declaration"));
else
{
char *name = xmemdup0 (name_start, strspn (name_start, alphanum));
@@ -793,8 +821,8 @@ 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);
+ complain (loc, complaint, "require bison %s, but have %s",
+ version, PACKAGE_VERSION);
exit (EX_MISMATCH);
}
}
@@ -802,7 +830,7 @@ version_check (location const *loc, char const *version)
static void
gram_error (location const *loc, char const *msg)
{
- complain_at (*loc, "%s", msg);
+ complain (loc, complaint, "%s", msg);
}
char const *
@@ -823,3 +851,18 @@ 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;
+}