Bug#766788: libreoffice-writer: Crashes with "stack smashing detected"
forwarded 766788 https://bugs.freedesktop.org/show_bug.cgi?id=78906
reassign 766788 libicu52
thanks
Hi,
On Tue, Oct 28, 2014 at 10:34:27AM +0100, Michal Sojka wrote:
> On Po, říj 27 2014, Rene Engelhard wrote:
> > On Mon, Oct 27, 2014 at 11:24:43AM +0100, Michal Sojka wrote:
> >> >> I can reproduce this in both unstable and testing
> >> >> (1:4.3.3~rc2~git20141011-1). I cannot reproduce this in the version
> >> >
> >> > And why are you then not marking it as such?
> >>
> >> How can I do that next time? https://www.debian.org/Bugs/Reporting does
> >> not mention how to mark multiple version.
> >
> > You add 1:4.3.3~rc2~git20141011-1 in Version: and the BTS then knows
> > it also affect 1:4.3.3~rc2-1 (see [1])
> >
> >> >> from libreoffice.org (LibreOffice_4.3.2_Linux_x86-64_deb.tar.gz).
> >> >
> >> > And with 4.3.3 rc1? (Or rc2 which would be in the next days)
> >> > You right now compare a 4.3.2 with a -between-4.3.3-rc1-and-rc2
> >> > or 4.3.3 rc2 ;)
> >> >
> >> >> After the crash the following information appears on the terminal:
> >> >>
> >> >> *** stack smashing detected ***: /usr/lib/libreoffice/program/soffice.bin terminated
> >> >> ======= Backtrace: =========
> >> >> /lib/x86_64-linux-gnu/libc.so.6(+0x72faf)[0x7fdd44a1ffaf]
> >> >> /lib/x86_64-linux-gnu/libc.so.6(__fortify_fail+0x37)[0x7fdd44aa30a7]
> >> >> /lib/x86_64-linux-gnu/libc.so.6(__fortify_fail+0x0)[0x7fdd44aa3070]
> >> >
> >> > But given it runs into the fortify functions it probably won't appear
> >> > in 4.3.3 rc1 upstream until it's a real crash also there; upstream doesn't
> >> > use those hardening flags.
> >>
> >> I was able to reproduce this in my own build of libreoffice. Any hint
> >
> > But probably without hardening or with?
>
> With. I run
>
> eval $(dpkg-buildflags --export=sh)
>
> which corresponds to
>
> export CFLAGS="-g -O2 -fstack-protector-strong -Wformat -Werror=format-security"
> export CPPFLAGS="-D_FORTIFY_SOURCE=2"
> export CXXFLAGS="-g -O2 -fstack-protector-strong -Wformat -Werror=format-security"
> export FCFLAGS="-g -O2 -fstack-protector-strong"
> export FFLAGS="-g -O2 -fstack-protector-strong"
> export GCJFLAGS="-g -O2 -fstack-protector-strong"
> export LDFLAGS="-Wl,-z,relro"
> export OBJCFLAGS="-g -O2 -fstack-protector-strong -Wformat -Werror=format-security"
> export OBJCXXFLAGS="-g -O2 -fstack-protector-strong -Wformat -Werror=format-security"
>
> before compiling.
>
> > Same backtrace or something else?
>
> Probably the same, but with more information:
I see.
I got a mail today from upstream saying
--- snip ---
Date: Sun, 30 Nov 2014 20:40:25 +0000
From: Caolán McNamara <caolanm@redhat.com> To: rene@debian.org Subject: fix for https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=766788
http://cgit.freedesktop.org/libreoffice/core/commit/?id=cced2f6b866877c39afe86f4af18622cb4022bef fixes https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=766788
C.
--- snip ---
so it actually seems to be a ICU bug...
Regards,
Rene
Reply to: