--- Begin Message ---
- To: 366469-done@bugs.debian.org
- Subject: Re: Bug#366469: grep segfaults
- From: Aurelien Jarno <aurelien@aurel32.net>
- Date: Mon, 23 Apr 2007 08:43:11 +0200
- Message-id: <20070423064310.GA15680@amd64.aurel32.net>
- In-reply-to: <20060509132730.GB6321@andromeda>
- References: <ea11805f0605081542ucb9342v7651c0e74afdf28d@mail.gmail.com> <20060509132730.GB6321@andromeda>
Version: 2.5-1
On Tue, May 09, 2006 at 09:27:31AM -0400, Justin Pryzby wrote:
> tag 366469 confirmed
> reassign 366469 grep,libc6
> retitle 366469 grep -i segfaults matching a binary file in zh_CN.GBK locale
> thanks
>
> On Mon, May 08, 2006 at 03:42:38PM -0700, Xin Liu wrote:
> > Package: grep
> > Version: 2.5.1.ds2-4
> >
> > When the locale is zh_CN.GBK, "grep -i" would produce segment fault
> > with binary files. For instance:
> > ***
> > lx@nightingale:~/tmp/tmp$ grep -i readline /bin/bash
> > Segmentation fault
> > ***
> I can confirm this; thanks for the report. It looks like it might be
> a glibc bug:
>
> #0 0xb7eb2c1f in *__GI_memcpy (dstpp=0x8075b48, srcpp=0xbfb7ba24,
> len=4294967295) at ../sysdeps/generic/memcpy.c:55
> #1 0xb7ee3aab in build_wcs_upper_buffer (pstr=0xbfb7bbc8)
> at regex_internal.c:424
> #2 0xb7ee3ee3 in re_string_reconstruct (pstr=0xbfb7bbc8, idx=163,
> eflags=<value optimized out>) at regex_internal.c:726
> #3 0xb7eea1cc in re_search_internal (preg=0x8075278, string=0x807e1c5 "",
> length=<value optimized out>, start=134699848, range=349, stop=349,
> nmatch=1, pmatch=0x8075f78, eflags=0) at regexec.c:797
> #4 0xb7eeb2af in re_search_stub (bufp=0x8075278, string=0x807e1c5 "",
> length=349, start=0, range=349, stop=-1, regs=0x8075298, ret_len=0)
> at regexec.c:454
> #5 0xb7eeb7c3 in __re_search (bufp=0xffffffff,
> string=0x8075b48 "ػ��\204������?� length=1073738376, start=0,
> range=0, regs=0x0) at regexec.c:318
> #6 0x08058666 in EGexecute ()
> #7 0x0804ab02 in grepbuf ()
> #8 0x0804b4b0 in grepfile ()
> #9 0x0804d0b4 in main ()
>
> Justin
>
>
I am able to reproduce this bug with glibc 2.3.6, but not with glibc
2.5, so I guess this bug is fixed.
--
.''`. Aurelien Jarno | GPG: 1024D/F1BCDB73
: :' : Debian developer | Electrical Engineer
`. `' aurel32@debian.org | aurelien@aurel32.net
`- people.debian.org/~aurel32 | www.aurel32.net
--- End Message ---