/* Bison Grammar Parser -*- C -*-
- Copyright (C) 2002 Free Software Foundation, Inc.
+
+ Copyright (C) 2002, 2003 Free Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
%defines
%locations
%pure-parser
-// %error-verbose
+%error-verbose
%defines
%name-prefix="gram_"
#include "reader.h"
#include "symlist.h"
-/* Produce verbose syntax errors. */
-#define YYERROR_VERBOSE 1
-
#define YYLLOC_DEFAULT(Current, Rhs, N) (Current) = lloc_default (Rhs, N)
static YYLTYPE lloc_default (YYLTYPE const *, int);
int current_prec = 0;
%}
+%initial-action
+{
+ /* Bison's grammar can initial empty locations, hence a default
+ location is needed. */
+ @$.start.file = @$.end.file = current_file;
+ @$.start.line = @$.end.line = 1;
+ @$.start.column = @$.end.column = 0;
+}
/* Only NUMBERS have a value. */
%union
`----------------------*/
%token
- PERCENT_DEBUG "%debug"
- PERCENT_DEFINE "%define"
- PERCENT_DEFINES "%defines"
- PERCENT_ERROR_VERBOSE "%error-verbose"
- PERCENT_EXPECT "%expect"
- PERCENT_FILE_PREFIX "%file-prefix"
- PERCENT_GLR_PARSER "%glr-parser"
- PERCENT_LEX_PARAM "%lex-param {...}"
- PERCENT_LOCATIONS "%locations"
- PERCENT_NAME_PREFIX "%name-prefix"
- PERCENT_NO_LINES "%no-lines"
- PERCENT_OUTPUT "%output"
- PERCENT_PARSE_PARAM "%parse-param {...}"
- PERCENT_PURE_PARSER "%pure-parser"
- PERCENT_SKELETON "%skeleton"
- PERCENT_START "%start"
- PERCENT_TOKEN_TABLE "%token-table"
- PERCENT_VERBOSE "%verbose"
- PERCENT_YACC "%yacc"
+ PERCENT_DEBUG "%debug"
+ PERCENT_DEFAULT_PREC "%default-prec"
+ PERCENT_DEFINE "%define"
+ PERCENT_DEFINES "%defines"
+ PERCENT_ERROR_VERBOSE "%error-verbose"
+ PERCENT_EXPECT "%expect"
+ PERCENT_FILE_PREFIX "%file-prefix"
+ PERCENT_GLR_PARSER "%glr-parser"
+ PERCENT_INITIAL_ACTION "%initial-action {...}"
+ PERCENT_LEX_PARAM "%lex-param {...}"
+ PERCENT_LOCATIONS "%locations"
+ PERCENT_NAME_PREFIX "%name-prefix"
+ PERCENT_NO_DEFAULT_PREC "%no-default-prec"
+ PERCENT_NO_LINES "%no-lines"
+ PERCENT_NONDETERMINISTIC_PARSER
+ "%nondeterministic-parser"
+ PERCENT_OUTPUT "%output"
+ PERCENT_PARSE_PARAM "%parse-param {...}"
+ PERCENT_PURE_PARSER "%pure-parser"
+ PERCENT_SKELETON "%skeleton"
+ PERCENT_START "%start"
+ PERCENT_TOKEN_TABLE "%token-table"
+ PERCENT_VERBOSE "%verbose"
+ PERCENT_YACC "%yacc"
;
%token TYPE "type"
%type <chars> STRING string_content
"%destructor {...}"
+ "%initial-action {...}"
"%lex-param {...}"
"%parse-param {...}"
"%printer {...}"
declaration:
grammar_declaration
| PROLOGUE { prologue_augment ($1, @1); }
-| "%debug" { debug_flag = 1; }
+| "%debug" { debug_flag = true; }
| "%define" string_content string_content { muscle_insert ($2, $3); }
-| "%defines" { defines_flag = 1; }
-| "%error-verbose" { error_verbose = 1; }
+| "%defines" { defines_flag = true; }
+| "%error-verbose" { error_verbose = true; }
| "%expect" INT { expected_conflicts = $2; }
| "%file-prefix" "=" string_content { spec_file_prefix = $3; }
-| "%glr-parser" { glr_parser = 1; }
+| "%glr-parser"
+ {
+ nondeterministic_parser = true;
+ glr_parser = true;
+ }
+| "%initial-action {...}"
+ {
+ muscle_code_grow ("initial_action", $1, @1);
+ }
| "%lex-param {...}" { add_param ("lex_param", $1, @1); }
-| "%locations" { locations_flag = 1; }
+| "%locations" { locations_flag = true; }
| "%name-prefix" "=" string_content { spec_name_prefix = $3; }
-| "%no-lines" { no_lines_flag = 1; }
+| "%no-lines" { no_lines_flag = true; }
+| "%nondeterministic-parser" { nondeterministic_parser = true; }
| "%output" "=" string_content { spec_outfile = $3; }
-| "%parse-param {...}" { add_param ("parse_param", $1, @1); }
-| "%pure-parser" { pure_parser = 1; }
+| "%parse-param {...}" { add_param ("parse_param", $1, @1); }
+| "%pure-parser" { pure_parser = true; }
| "%skeleton" string_content { skeleton = $2; }
-| "%token-table" { token_table_flag = 1; }
+| "%token-table" { token_table_flag = true; }
| "%verbose" { report_flag = report_states; }
-| "%yacc" { yacc_flag = 1; }
-| ";"
+| "%yacc" { yacc_flag = true; }
+| /*FIXME: Err? What is this horror doing here? */ ";"
;
grammar_declaration:
}
| "%union {...}"
{
- typed = 1;
+ typed = true;
MUSCLE_INSERT_INT ("stype_line", @1.start.line);
muscle_insert ("stype", $1);
}
symbol_printer_set (list->sym, $1, list->location);
symbol_list_free ($2);
}
+| "%default-prec"
+ {
+ default_prec = true;
+ }
+| "%no-default-prec"
+ {
+ default_prec = false;
+ }
;
symbol_declaration:
body of the grammar. */
rules_or_grammar_declaration:
rules
-| grammar_declaration
+| grammar_declaration ";"
{
if (yacc_flag)
complain_at (@$, _("POSIX forbids declarations in the grammar"));
{
yyerrok;
}
-| ";"
;
rules:
rhses.1:
rhs { grammar_rule_end (@1); }
| rhses.1 "|" rhs { grammar_rule_end (@3); }
+| rhses.1 ";"
;
rhs:
/* Nothing. */
| "%%" EPILOGUE
{
- epilogue_augment ($2, @2);
+ muscle_code_grow ("epilogue", $2, @2);
scanner_last_string_free ();
}
;
lloc_default (YYLTYPE const *rhs, int n)
{
int i;
- YYLTYPE r;
- r.start = r.end = rhs[n].end;
+ YYLTYPE loc;
+ loc.start = loc.end = rhs[n].end;
/* Ignore empty nonterminals the start of the the right-hand side.
Do not bother to ignore them at the end of the right-hand side,
for (i = 1; i <= n; i++)
if (! equal_boundaries (rhs[i].start, rhs[i].end))
{
- r.start = rhs[i].start;
+ loc.start = rhs[i].start;
break;
}
- return r;
+ return loc;
}