Created attachment 3686 backtrace of the crash Orignal bug report at https://bugzilla.redhat.com/show_bug.cgi?id=706677 abrt version: 2.0.1 architecture: x86_64 cmdline: /usr/lib64/xfce4/notifyd/xfce4-notifyd component: xfce4-notifyd crash_function: handle_error executable: /usr/lib64/xfce4/notifyd/xfce4-notifyd kernel: 2.6.38.6-27.fc15.x86_64 os_release: Fedora release 15 (Lovelock) package: xfce4-notifyd-0.2.1-2.fc15 rating: 4 reason: Process /usr/lib64/xfce4/notifyd/xfce4-notifyd was killed by signal 6 (SIGABRT) time: Sat May 21 23:57:41 2011 uid: 500 username: rave comment: Doing a Download in firefox. The download was finished and FF was about to display the notification. The crash only randomly. Some more info's: I am using compiz and GNOME's notification-deamon is uninstalled. More info (memory map, list of DSOs, environemt variables) is available in the downstream bug report.
I had a similar report with 0.1.1, see https://bugzilla.redhat.com/show_bug.cgi?id=679792
There seems to be an analysis of the same bug at http://docs.openmoko.org/trac/ticket/863. It looks like an inherent race with gdk_screen_get_window_stack and gdk_window_get_type_hint. Here's an attempted patch based on the one from the that ticket: http://ricky.fedorapeople.org/xfce/0001-Fix-race-with-window-becoming-invalid.patch It's hasn't been tested at all unfortunately, as I just switched to XFCE and don't have a development environment setup for it yet :-)
Hi, any update on this? Does the patch look acceptable?
Hi, from me is the original bug report on redhat bugzilla. If i know which packet i have to patch i will try to do that. Have i to patch xfce4-notifyd ?
Thanks for digging this, the patch looks fine! I committed it to the git master branch (commit 37d571aee335f263d05a117cdf9051aa66b154ad), thanks for your work.
(In reply to comment #4) > Have i to patch xfce4-notifyd ? Yes, but I already did that. Please test the Fedora update and provide feedback at https://admin.fedoraproject.org/updates/xfce4-notifyd-0.2.1-3.fc15