On Wed, 2005-05-25 at 17:00 -0400, John Dennis wrote:
> On Wed, 2005-05-25 at 14:46 -0500, Gustavo A. Lozano wrote:
> > My evolution is generating a lot of crashes (currently I am up 2 date to
> > everything in the yum development tree)... Before posting the bugzilla,
> > I want to know about other experiences out there...
I see a lot of evolution crashes as well as a pernicious "denial of
workstation" ;-) I'll explain in a moment.
It is not unusual for the evolution GUI to spontaneously abort, when it
does it leaves the other evolution processes running. This happens once
or twice a day and appears to be completely random. Interestingly, if it
aborts while I'm composing mail not a single key stroke is lost when the
draft is recovered so I'm assuming it has some graceful opportunity to
save its state on the way down and is not an absolutely hard crash.
Please can you try to get a backtrace into Bugzilla. I suspect a crash
in gtkhtml3, so having the debuginfo for that package (as well as
Evolution) would be useful.
You should be getting backtraces via bug-buddy. If not, and if it's
happening that often, please can you start evolution under gdb and try
to get a backtrace.
The other problem I was seeing on a regular basis was all the resources
of the machine being consumed such that I could not get to an alternate
console, run ps or top, the machine was dragged to its knees. I'm pretty
sure evolution in concert with spamd (filter incoming mail checkbox) was
the culprit, but given how hard it was to do anything when the problem
occurred I only have anecdotal evidence. I was often forced to resort to
using the reset button. I suspect it was some type of fork bombing with
spamd where it got caught in tight loop launching processes that would
immediately die, but that's only speculation. I have not seen this
problem since turning off the filter option on incoming mail.
I've talked to David Malcom about this and he encouraged me to file a
bugzilla, but it so hard to collect any information or see a pattern
that I wasn't sure I saw the value in a bug report that simply asserted
"evolution is doing something bad" ;-)
I went ahead and filed this here:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=158807