multibyte completion broken on some locales

[Jiro SEKIBA]

Configuration Information [Automatically generated, do not change]:
Machine: i386
OS: linux-gnu
Compiler: cc
Compilation CFLAGS:  -DPROGRAM='bash' -DCONF_HOSTTYPE='i386' 
-DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='i386-pc-linux-gnu' 
-DCONF_VENDOR='pc' -DSHELL -DHAVE_CONFIG_H  -I.  -I../src -I../src/include 
-I../src/lib
uname output: Linux harlequin 2.4.17 #1 Wed Feb 20 18:34:43 JST 2002 i686 
GNU/Linux
Machine Type: i386-pc-linux-gnu

Bash Version: 2.05b
Patch Level: 0
Release Status: release

Description:
        readline completion functionality automatically adds escape character
        just before the shell special characters.  However some codesets
        like SJIS, have shell speciall characters in byte sequences.

Repeat-By:
        $ unset LC_ALL
        $ export LANG=ja_JP.SJIS
        $ touch `printf "a\x89\5c"`
        $ ls a[TAB]

        then you'll see escape character '\' added

        0x895c in SJIS is U+5642 in Unicode

Fix:
diff -ru bash-2.05b.orig/lib/readline/complete.c 
bash-2.05b/lib/readline/complete.c
--- bash-2.05b.orig/lib/readline/complete.c     2002-05-08 04:39:32.000000000 
+0900
+++ bash-2.05b/lib/readline/complete.c  2003-07-03 14:52:37.000000000 +0900
@@ -663,7 +663,14 @@
         quote substrings for the completer.  Try to find the start
         of an unclosed quoted substring. */
       /* FOUND_QUOTE is set so we know what kind of quotes we found. */
-      for (scan = pass_next = 0; scan < end; scan++)
+      for (scan = pass_next = 0; scan < end;
+#if defined(HANDLE_MULTIBYTE)
+           scan = ((MB_CUR_MAX == 1 || rl_byte_oriented) ?
+                   (scan + 1) :
+                   _rl_find_next_mbchar(rl_line_buffer,scan,1,MB_FIND_ANY)))
+#else
+           scan++)
+#endif
        {
          if (pass_next)
            {
diff -ru bash-2.05b.orig/lib/readline/rldefs.h bash-2.05b/lib/readline/rldefs.h
--- bash-2.05b.orig/lib/readline/rldefs.h       2002-05-08 04:40:07.000000000 
+0900
+++ bash-2.05b/lib/readline/rldefs.h    2003-07-03 14:04:50.000000000 +0900
@@ -77,7 +77,7 @@
 extern int _rl_strnicmp PARAMS((char *, char *, int));
 #endif
 
-#if defined (HAVE_STRPBRK)
+#if defined (HAVE_STRPBRK) && ! defined(HANDLE_MULTIBYTE)
 #  define _rl_strpbrk(a,b)     strpbrk((a),(b))
 #else
 extern char *_rl_strpbrk PARAMS((const char *, const char *));
diff -ru bash-2.05b.orig/lib/readline/util.c bash-2.05b/lib/readline/util.c
--- bash-2.05b.orig/lib/readline/util.c 2002-03-14 06:45:43.000000000 +0900
+++ bash-2.05b/lib/readline/util.c      2003-07-03 14:04:50.000000000 +0900
@@ -223,7 +223,7 @@
   return ((char *)NULL);
 }
 
-#ifndef HAVE_STRPBRK
+#if defined(HAVE_STRPBRK) || defined(HANDLE_MULTIBYTE)
 /* Find the first occurrence in STRING1 of any character from STRING2.
    Return a pointer to the character in STRING1. */
 char *
@@ -250,7 +250,7 @@
        {
          v = _rl_get_char_len (string1, &ps);
          if (v > 1)
-           string += v - 1;    /* -1 to account for auto-increment in loop */
+           string1 += v - 1;   /* -1 to account for auto-increment in loop */
        }
 #endif
     }