Configuration Information [Automatically generated, do not change]:
Machine: i386
OS: freebsd5.5
Compiler: gcc
Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='i386'
-DCONF_OSTYPE='freebsd5.5' -DCONF_MACHTYPE='i386-unknown-freebsd5.5'
-DCONF_VENDOR='unknown' -DLOCALEDIR='/home/svdb/bashtemp/share/locale'
-DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib
-I./lib/intl -I/home/svdb/src/bash-3.1/lib/intl -g -O0
uname output: FreeBSD frog.stack.nl 5.5-RELEASE-p1 FreeBSD 5.5-RELEASE-p1 #6:
Mon Jun 12 18:16:16 CEST 2006 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/FROG
i386
Machine Type: i386-unknown-freebsd5.5
Bash Version: 3.1
Patch Level: 0
Release Status: release
Description:
When the LANG environment variable is set to an UTF-8 encoding,
giving a command containing a vertical tab character will cause
bash to race and eventually crash.
This does not happen on all systems; I suspect this has something
to do with the locale definitions. It worked on a FreeBSD 5.5 box.
On closer investigation it appears the following is happening:
parse.y line 2559:
while ((character = shell_getc (1)) != EOF && whitespace
(character))
will stop at the vertical tab; whitespace() returns false for it.
parse.y line 3594:
if MBTEST(shellbreak (character))
will see the vertical tab as a word break character; shellbreak()
returns true for it. The result is that an empty string is taken
as a word and no progress is made. And as for each found word
memory is allocated,
(parse.y line 3672 and onwards:
the_word = (WORD_DESC *)xmalloc (sizeof (WORD_DESC));
the_word->word = (char *)xmalloc (1 + token_index);
strcpy (the_word->word, token);
)
eventually the process will run out of memory, and it will crash.
Repeat-By:
On a FreeBSD system:
env -i `which bash` --norc -c 'export LANG=en_GB.UTF-8; eval
`printf \\\v`'
_______________________________________________
Bug-bash mailing list
Bug-bash@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-bash
