Red Hat Bugzilla – Bug 237206
Clicking ok on bad cert crashes (see 220714)
Last modified: 2007-11-30 17:12:02 EST
Description of problem:
Version-Release number of selected component (if applicable):
Every time evolution starts
Steps to Reproduce:
1. Connect to a pop server with a self signed cert (may require to accounts on
the same server)
2. Click ok on both dialogs
It crashes of course.
It shouldn't crash.
What version-release number are you using?
I was afraid I had forgoten that. evolution-2.10.1-4.fc7.
Created attachment 154226 [details]
Trace back for problem
I'm hitting the same problem now.
I'm seeing a deadlock, not a crash. The traceback doesn't show a crash either.
Yeah, I'm seeing a deadlock as well.
Whatever it is, bugbuddy came up. It stalls (hangs) for about 1 minute before
bug buddy shows up for me.
Ah, maybe I wasn't waiting long enough then.
Bugbuddy never came up for me. I waited for a long time.
Should be fixed in evolution-2.10.1-17.fc7.
This turned out to be another game of "who owns the message and when should it
be freed?", just like bug #220714 and others like it were. I've had a patch
waiting in the wings for some time that makes Evolution's message passing use
reference counting to track ownership of the messages, rather than the
convoluted logic that's there now.
The patch fixes the crash but it's a bit experimental. If you guys could help
me test this over the next few days I'd appreciate it, since we're in the final
stretch for Fedora 7 now and I'm a bit nervous about putting this in at the 11th
Please post comments here if you see any regressions that you think might be
related to this.
This build fixes the bad cert issue for me. I'll continue using it and see if I
notice any other problems. Thanks!
Can you confirm that this is fixed with the updated evolution package Matthew
mentioned, available here:
I have changed my email accounts in such a way that I no longer see the bug. (It
was only hitting if I had two accounts on the same server or two servers with
bad certs... I no longer have that.)
I am sorry I cannot test further. If others say it fixes the problem, then that
is good enough I guess.
Closing based on comment #11 and my own testing.
Feel free to reopen this bug if you encounter these crashes again.