Bug 261061
Summary: | Evolution crashes don't trigger bug-buddy | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Christopher Beland <beland> |
Component: | evolution | Assignee: | Matthew Barnes <mbarnes> |
Status: | CLOSED CURRENTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 7 | CC: | mcepl |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | All | ||
Whiteboard: | |||
Fixed In Version: | evolution-2.12.3-1.fc8 | Doc Type: | Bug Fix |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2008-01-29 19:18:44 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Christopher Beland
2007-08-28 17:25:34 UTC
If either of these crashes are reproducible, could you try running Evolution from a terminal and see if any messages get printed when Evolution crashes? Sometimes, when programs do really nasty things like freeing the same memory twice, glibc will abort the program before Bug Buddy gets triggered. Do other programs trigger bug-buddy when they abort? Perhaps try "kill -SEGV <pid-of-some-program>". Interestingly, I *can* get Evolution to trigger bug-buddy if I kill -SEGV it. I could not reproduce these crashes, but I will capture Evolution terminal output until it crashes again. Yeah, the Bug Buddy trigger works by listening for UNIX kill signals, SEGV being the most common (trying to access an invalid memory address). I guess glibc aborts by some other means, if that is indeed the cause. I experienced another crash today which did not trigger bug-buddy. I was replying to one message while forwarding another, and I edited my preferences to change the forwarding style to "quoted" from the default. The following was captured in .xsession-errors, though without timestamps, it's difficult to say whether or not they were printed around the time of the crash. (evolution:2433): e-data-server-DEBUG: Loading categories from "/home/beland/.evolution/categories.xml" (evolution:2433): e-data-server-DEBUG: Loaded 29 categories (evolution:2433): e-utils-WARNING **: Cannot resolve symbol 'org_gnome_new_mail_config' in plugin '/usr/lib/evolution/2.10/plugins/liborg-gnome-new-mail-notify.so' (not exported?) (evolution:2433): GLib-GObject-WARNING **: gsignal.c:1669: signal `source_selected' is invalid for instance `0x9cab370' (evolution:2433): GLib-GObject-WARNING **: gsignal.c:1669: signal `source_selected' is invalid for instance `0x9cab490' ***MEMORY-ERROR***: evolution[2433]: GSlice: assertion failed: sinfo->n_allocated > 0 Those warnings look like an upstream bug I fixed recently [1]. I don't think I've pushed the patch for it to Fedora 7 yet; I'll make it a point to do that. But I'm still not sure why Bug Buddy is not opening for you, unless the GSlice code is bypassing the Bug Buddy trigger like glibc tends to do. My only suggestion at this point is to try running Evolution directly from GDB and get a backtrace that way: $ gdb /usr/bin/evolution ... (gdb) run ... ... (wait for crash) ... (gdb) thread apply all backtrace In the meantime I'll push the aforementioned bugfix to Fedora 7 and hope that at least reduces the amount of crashes you're seeing. [1] http://bugzilla.gnome.org/show_bug.cgi?id=447591 Cool, thanks. I just experienced another crash while changing from viewing an IMAP folder to viewing a local one by clicking on the folder tree. Unfortunately, even using the gdb method you outlined above, it did not result in a stack trace. This is the last bit of GDB output: [Thread -1311253616 (LWP 5331) exited] [Thread -1254634608 (LWP 5333) exited] [Thread -1332233328 (LWP 5332) exited] [New Thread -1332233328 (LWP 5334)] [New Thread -1254634608 (LWP 5335)] [Thread -1332233328 (LWP 5334) exited] ***MEMORY-ERROR***: evolution[4391]: GSlice: assertion failed: sinfo->n_allocated > 0 [Switching to Thread -1332233328 (LWP 5334)] Cannot remove breakpoints because program is no longer writable. It might be running in another process. Further execution is probably impossible. 0x0065f7e1 in __nptl_death_event () from /lib/libpthread.so.0 ptrace: No such process. (gdb) thread apply all backtrace Couldn't get registers: No such process. At this point, process 4391 is a zombie. I pushed evolution-2.10.3-4.fc7 to Updates Testing with a fix for the "signal is invalid" warnings you're seeing. Should be available within a day or so. Hmm, I've enabled [updates-testing] in /etc/yum.repos.d/fedora-updates-testing.repo and got a ton of updates as a result, but after running "yum update" just now, evolution-2.10.3-2.fc7 is still installed. Looks like it got pushed to Updates Testing today at 15:01 EST. If you're using a mirror, it might take a day for the mirror to sync up with the latest updates. Maybe try again tomorrow? Yay, it came through today. Did the update fix the problem? I see Bug Buddy pop up all too often. Reporter, could you please reply to the previous question? If you won't reply in one month, I will have to close this bug as INSUFFICIENT_DATA. Thank you. Yes, I've since had Evolution crash in Fedora 8 and Bug Buddy did indeed handle it. Closing as CURRENTRELEASE. |