Re: i18n for yacc backend messages

[Paul Eggert]

Jan Nieuwenhuizen <address@hidden> writes:

> Ok, see patch below.

Thanks, but your patch was incomplete, as it didn't contain any fixes
to lib/quotearg.c.

I installed this patch instead.  It doesn't require any changes to
quotearg (a shared module) so it's better in that respect.

Quoting is a pain to get right, I'm afraid.  I wouldn't be surprised
at all if the code is still buggy after this patch.

2005-04-15  Paul Eggert  <address@hidden>

        * src/parse-gram.y: Include quotearg.h.
        (string_as_id): Quote $1 before using it as a key, since the
        lexer no longer quotes it for us.
        (string_content): Don't strip quotes, since lexer no longer
        quotes it for us.
        * src/scan-gram.l: Include quotearg.h.
        ("\""): Omit quote.
        ("'"<SC_ESCAPED_CHARACTER>): Quote symbol before using it as
        a key, since the rest of the lexer doesn't quote it.
        * src/symtab.c (symbol_get): Don't quote symbol; caller does it now.
        * tests/regression.at (Token definitions): Check for backslashes
        in token strings.

Index: src/parse-gram.y
===================================================================
RCS file: /cvsroot/bison/bison/src/parse-gram.y,v
retrieving revision 1.52
diff -p -u -r1.52 parse-gram.y
--- src/parse-gram.y    6 Oct 2004 15:52:24 -0000       1.52
+++ src/parse-gram.y    16 Apr 2005 06:14:42 -0000
@@ -1,6 +1,6 @@
 %{/* Bison Grammar Parser                             -*- C -*-
 
-   Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
+   Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
 
    This file is part of Bison, the GNU Compiler Compiler.
 
@@ -29,6 +29,7 @@
 #include "gram.h"
 #include "muscle_tab.h"
 #include "output.h"
+#include "quotearg.h"
 #include "reader.h"
 #include "symlist.h"
 
@@ -403,25 +404,23 @@ symbol:
 
 action:
   BRACED_CODE
-   { $$ = $1; }
+    { $$ = $1; }
 ;
 
-/* A string used as an ID: we have to keep the quotes. */
+/* A string used as an ID: quote it.  */
 string_as_id:
   STRING
     {
-      $$ = symbol_get ($1, @1);
+      $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
       symbol_class_set ($$, token_sym, @1);
     }
 ;
 
-/* A string used for its contents.  Strip the quotes. */
+/* A string used for its contents.  Don't quote it.  */
 string_content:
   STRING
-    {
-      $$ = $1 + 1;
-      $$[strlen ($$) - 1] = '\0';
-    };
+    { $$ = $1; }
+;
 
 
 epilogue.opt:
Index: src/scan-gram.l
===================================================================
RCS file: /cvsroot/bison/bison/src/scan-gram.l,v
retrieving revision 1.71
diff -p -u -r1.71 scan-gram.l
--- src/scan-gram.l     7 Mar 2005 06:41:39 -0000       1.71
+++ src/scan-gram.l     16 Apr 2005 06:14:42 -0000
@@ -34,6 +34,7 @@
 #include "files.h"
 #include "getargs.h"
 #include "gram.h"
+#include "quotearg.h"
 #include "reader.h"
 #include "uniqstr.h"
 
@@ -249,7 +250,7 @@ splice       (\\[ \f\t\v]*\n)*
   "'"        STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_CHARACTER;
 
   /* Strings. */
-  "\""       STRING_GROW; token_start = loc->start; BEGIN SC_ESCAPED_STRING;
+  "\""       token_start = loc->start; BEGIN SC_ESCAPED_STRING;
 
   /* Prologue. */
   "%{"        code_start = loc->start; BEGIN SC_PROLOGUE;
@@ -360,7 +361,6 @@ splice       (\\[ \f\t\v]*\n)*
 <SC_ESCAPED_STRING>
 {
   "\"" {
-    STRING_GROW;
     STRING_FINISH;
     loc->start = token_start;
     val->chars = last_string;
@@ -384,7 +384,9 @@ splice       (\\[ \f\t\v]*\n)*
     STRING_GROW;
     STRING_FINISH;
     loc->start = token_start;
-    val->symbol = symbol_get (last_string, *loc);
+    val->symbol = symbol_get (quotearg_style (escape_quoting_style,
+                                             last_string),
+                             *loc);
     symbol_class_set (val->symbol, token_sym, *loc);
     last_string_1 = last_string[1];
     symbol_user_token_number_set (val->symbol, last_string_1, *loc);
Index: src/symtab.c
===================================================================
RCS file: /cvsroot/bison/bison/src/symtab.c,v
retrieving revision 1.60
diff -p -u -r1.60 symtab.c
--- src/symtab.c        12 Apr 2005 22:38:20 -0000      1.60
+++ src/symtab.c        16 Apr 2005 06:14:42 -0000
@@ -475,8 +475,7 @@ symbol_get (const char *key, location lo
   symbol probe;
   symbol *entry;
 
-  /* Keep the symbol in a printable form.  */
-  key = uniqstr_new (quotearg_style (escape_quoting_style, key));
+  key = uniqstr_new (key);
   probe.tag = key;
   entry = hash_lookup (symbol_table, &probe);
 
Index: tests/regression.at
===================================================================
RCS file: /cvsroot/bison/bison/tests/regression.at,v
retrieving revision 1.88
diff -p -u -r1.88 regression.at
--- tests/regression.at 26 Mar 2005 01:12:59 -0000      1.88
+++ tests/regression.at 16 Apr 2005 06:14:42 -0000
@@ -314,20 +314,44 @@ AT_SETUP([Token definitions])
 # Bison managed, when fed with `%token 'f' "f"' to #define 'f'!
 AT_DATA_GRAMMAR([input.y],
 [%{
-void yyerror (const char *s);
+#include <stdio.h>
+void yyerror (char const *s);
 int yylex (void);
 %}
-[%token MYEOF 0 "end of file"
+[%error-verbose
+%token MYEOF 0 "end of file"
 %token 'a' "a"
 %token B_TOKEN "b"
 %token C_TOKEN 'c'
 %token 'd' D_TOKEN
+%token SPECIAL "\\\'\?\"\a\b\f\n\r\t\v\001\377\x001\x0000ff"
+%%
+exp: "a" "\\\'\?\"\a\b\f\n\r\t\v\001\377\x001\x0000ff";
 %%
-exp: "a";
+void
+yyerror (char const *s)
+{
+  fprintf (stderr, "%s\n", s);
+}
+
+int
+yylex (void)
+{
+  return SPECIAL;
+}
+
+int
+main (void)
+{
+  return yyparse ();
+}
 ]])
 
 AT_CHECK([bison -o input.c input.y])
-AT_COMPILE([input.o], [-c input.c])
+AT_COMPILE([input])
+AT_PARSER_CHECK([./input], 1, [],
+[syntax error, unexpected "\\'?\"\a\b\f\n\r\t\v\001\377\001\377", expecting "a"
+])
 AT_CLEANUP