Bug#378753: tetex-bin: texshow sometimes segfaults
Helge Kreutzmann <debian@helgefjell.de> wrote:
> Hello Frank,
> I now booted into 2.6.10 (without grsec) and was unable to crash
> texshow. But I was also unable to launch texshow via the menu option
> (Help->TexShow), both in fvwm2 and windowmaker (my default here). Do
> you know if / where errors with programm invocation from X menues are
> logged?
I think it is ~/.xsession-errors. If I'm wrong, there's an other option
to get debugging info:
Create a file /etc/menu/texshow-debug with the content:
?package(local.texshow):needs="X11" \
section="Help" \
title="TeXshow-debug" \
command="sh -c /usr/bin/texshow > ~/texshow.log 2>&1"
Then run update-menus (as root) and invoke the new command from the
menu.
> So this indeed might be related to grsec. But since this is the first
> programm to encounter this, I strongly expect a problem in texshow or
> in an underlying perl component.
FYI, there was once an other Perl-TK segfault in tetex-bin,
http://bugs.debian.org/348513
> Does invocation from the Help->Texshow entry work for you? If yes,
> I'll have to find out how to debug this, I guess (since this might be
> a second issue).
Yes, it works for me, no problems at all. Err, no.
It works on sarge with XOrg and backported tetex-bin package. But I can
reproduce the error on sid with texlive-context; even without using the
menu:
frank@sid:~$ texshow
TeXShow-XML 0.2 beta / Taco Hoekwater 2004
interface : cont-en
status : searching for setup files
status : loading setups
status : loading 'cont-cz' took 0.059133 seconds
status : loading 'cont-de' took 0.058673 seconds
status : loading 'cont-it' took 0.057886 seconds
status : loading 'cont-nl' took 0.057762 seconds
status : loading 'cont-ro' took 0.05785 seconds
status : loading 'cont-en' took 0.057636 seconds
status : initializing display
*** glibc detected *** malloc(): memory corruption: 0x087e19d4 ***
/usr/bin/texshow: line 24: 8172 Aborted perl "$p" ${1+"$@"}
frank@sid:~$
The line number is different, although the scripts are identical up to
line 28 (where tetex-bin has a short patch added).
So I guess we need to switch on the debugging machinery used in #348513.
Regards, Frank
--
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX/TeXLive)
Reply to: