libc-2.10.1 causes previously well-behaved program to abort
I have a small gtk-based program which puts up buttonbars on my desktop. I have used it in both Crux and Debian and have never had any problems with it. But since upgrading to Crux-2.6, I have had intermittent aborts. The traceback (attached) suggests that the problem is inside libc, specifically in malloc. Once the program has aborted, it aborts each time it is restarted. But weirdly I was able to break this cycle simply by altering the order in which buttons were created, via the program's config file. For example, this version - (Files) </home/data/icons/file-manager.png> {gksu-not|sudo -S emelfm2} (Terminal) </home/data/icons/root-term.png> {gksu-not|sudo -S xterm -font 10x20} (Editor) </usr/share/pixmaps/vim32x32.png> {gksu-not|sudo -S gvim} (Other apps) </home/hazel/icons/applications-other.png> {bbrun -w} (Reboot) </home/data/icons/reboot.png> {gksu-not|sudo -S /sbin/shutdown -r now} (Shutdown) </home/data/icons/gnome-logout.png> {gksu-not|sudo -S /sbin/shutdown -h now} caused the program to abort yesterday (though not on previous days), but this one - (Files) </home/data/icons/file-manager.png> {gksu-not|sudo -S emelfm2} (Editor) </usr/share/pixmaps/vim32x32.png> {gksu-not|sudo -S gvim} (Terminal) </home/data/icons/root-term.png> {gksu-not|sudo -S xterm -font 10x20} (Other apps) </home/hazel/icons/applications-other.png> {bbrun -w} (Reboot) </home/data/icons/reboot.png> {gksu-not|sudo -S /sbin/shutdown -r now} (Shutdown) </home/data/icons/gnome-logout.png> {gksu-not|sudo -S /sbin/shutdown -h now} allowed it to run again. Is this a bug or am I doing something wrong? -- Hazel Russman <hazel_russman@yahoo.co.uk>
I've also seen strange behaviour with the new glibc. The most noticeable one are random segfaults when building packages using fakeroot. I don't get them when building as root, so I thought it was a bug of fakeroot, but now that I read your message I'm not sure anymore. Has anybody else had similar experiences? Alan
On Thu, Sep 17, 2009 at 09:15:54AM +0900, Alan wrote:
I've also seen strange behaviour with the new glibc.
The most noticeable one are random segfaults when building packages using fakeroot.
I don't get them when building as root, so I thought it was a bug of fakeroot, but now that I read your message I'm not sure anymore.
Has anybody else had similar experiences?
Alan
Sorry, no. I've run 2.10.1 for a couple of months now without issues. i've even built a few iso's without trouble. are you guys sure the toolchain is ok and that there are no broken library deps? are there any specific apps (gtk, qt or whatever) that segfaults? -- /Fredrik Rinnestam
On Thursday 17 September 2009 10:31:41 Fredrik Rinnestam wrote:
On Thu, Sep 17, 2009 at 09:15:54AM +0900, Alan wrote:
I've also seen strange behaviour with the new glibc.
The most noticeable one are random segfaults when building packages using fakeroot.
I don't get them when building as root, so I thought it was a bug of fakeroot, but now that I read your message I'm not sure anymore.
Has anybody else had similar experiences?
Alan
Sorry, no. I've run 2.10.1 for a couple of months now without issues. i've even built a few iso's without trouble. are you guys sure the toolchain is ok and that there are no broken library deps? are there any specific apps (gtk, qt or whatever) that segfaults?
Neither have I. How did you upgrade?
El 17/09/09 Lucas Hazel escribió:
On Thursday 17 September 2009 10:31:41 Fredrik Rinnestam wrote:
On Thu, Sep 17, 2009 at 09:15:54AM +0900, Alan wrote:
I've also seen strange behaviour with the new glibc.
The most noticeable one are random segfaults when building packages using fakeroot.
I don't get them when building as root, so I thought it was a bug of fakeroot, but now that I read your message I'm not sure anymore.
Has anybody else had similar experiences?
Alan
Sorry, no. I've run 2.10.1 for a couple of months now without issues. i've even built a few iso's without trouble. are you guys sure the toolchain is ok and that there are no broken library deps? are there any specific apps (gtk, qt or whatever) that segfaults?
Neither have I. How did you upgrade?
After 11 months of not touching this machine I just updated all the packages in core from the 2.6 ISO, then rebuilt most of the stuff I installed from the ISO, and updated the kernel to 2.6.31. I'm still working on this, but I haven't had the time. Alan
Hazel Russman [2009-09-16 15:43]:
I have a small gtk-based program which puts up buttonbars on my desktop. I have used it in both Crux and Debian and have never had any problems with it. But since upgrading to Crux-2.6, I have had intermittent aborts. The traceback (attached) suggests that the problem is inside libc, specifically in malloc.
Once the program has aborted, it aborts each time it is restarted. But weirdly I was able to break this cycle simply by altering the order in which buttons were created, via the program's config file. For example, this version -
[...]
Is this a bug or am I doing something wrong?
I'm guessing that the program does something wrong, and that the error was tolerated by earlier glibc versions. Try running the program in valgrind; that should give you more clues on where the heap corruption happens resp what causes it. Regards, Tilman -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?
participants (5)
-
Alan -
Fredrik Rinnestam -
Hazel Russman -
Lucas Hazel -
Tilman Sauerbeck