On Mon, 7 Apr 2014, Dominique Leuenberger wrote:
Seeing the segfault in gnome-shell/libc, i think the best chance is to configure the system to create coredumps of crashes (i generally configure my factory ststem to write any dump to /cores) and once this happens, you can create backtraces based on the dump. Allows you to analyze an error after the crash actually happened, as opposed to running gs constantly under gdb, hoping to ever hit it.
If info is required for how to set this up, please let know... I have this documented somewhere, but it would do for a good blog post :)
I'm looking forward to your blog post :-), which is me saying I
have never set up something like this globally yet but it sounds
like a good idea.
Gerald
--
Dr. Gerald Pfeifer