Eval ${#name[subscript]} incorrect when having multibyte characters.

[Wang Xin]

Configuration Information [Automatically generated, do not change]:
Machine: i686
OS: linux-gnu
Compiler: gcc
Compilation CFLAGS:  -DPROGRAM='bash' -DCONF_HOSTTYPE='i686' -DCONF_OSTYPE='linu
x-gnu' -DCONF_MACHTYPE='i686-pc-linux-gnu' -DCONF_VENDOR='pc' -DLOCALEDIR='/usr/
local/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H   -I.  -I. -I./incl
ude -I./lib   -g -O2
uname output: Linux shan 2.6.24.5-smp #2 SMP Wed Apr 30 13:41:38 CDT 2008 i686 I
ntel(R) Celeron(R) M processor         1.40GHz GenuineIntel GNU/Linux
Machine Type: i686-pc-linux-gnu

Bash Version: 3.2
Patch Level: 0
Release Status: release

Description:
    When there are multibyte characters in an element of array, the
    result of ${#name[subscript]} will be incorrect. It will be evaled
    to numbers of bytes, but not numbers of characters.

Repeat-By:
    This can be reproduced by:
    1. a[0]=你好
    2. echo ${#a[0]}

    There are only two chinese characters, but the result is 6.

Fix:
    Following patch may be helpful.

    --- subst.c    2008-07-06 15:47:14.000000000 +0800
    +++ bash-3.2/subst.c    2008-07-06 15:47:39.000000000 +0800
    @@ -4763,7 +4763,7 @@
       else
         t = (ind == 0) ? value_cell (var) : (char *)NULL;

    -  len = STRLEN (t);
    +  len = MB_STRLEN (t);
       return (len);
     }
     #endif /* ARRAY_VARS */