From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030703 Description of problem: This bug seems quite similar to the RH8.0 bug: 76366 If evolution is running and the user just logs out, an oafd process remains running and if they then log back in and try to open evolution it complains that it can't find the configuration database and won't come up. (Sorry I'm not more specific about the error. I'm reporting this for a nontechnical third party and am not at the machine.) This may or may not happen every time, but is at least quite frequent. After seeing bug 76366, I oaf-slay'd it, had him log back in, bring up evolution, minimize it, and log out. He logged back in the problem was reproduced. Killed the oafd process and it worked again. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Open evolution. 2.minimize (may or may not be necessary) 3.Log out. 4.Log back in. 5.Try to open evolution. Actual Results: Message indicating that it can't find (open?) the configuration database. Expected Results: Evolution comes up successfully. Additional info: This is a week old "everything" install updated to all the latest packages from up2date. The only thing that is not stock is the vanilla 2.4.21 kernel compiled with the RH9 provided i686 .config file with the addition of the linux-abi patch which the customer requires for a report writer they use. No other patches, Redhat or otherwise are applied. However, I'm quite certain that I observed this behavior when I was running RH9 completely stock. I still have a RH9 partition and will try to reproduce here.
This is a "happens sometimes" sort of things and is an oaf bug, not evolution's doing.
From what I gather a few similar problems have been previously reported and the bug filed upsteam with bugzilla.gnome.org and seems inactive there. It seems readily reproducable on the machine in question, which I will have physical access to tomorrow. (Hard to reproduce via VNC.) Anything I need to check and report while I'm there? Also, I guess a workaround for now would be to put an oafslay somewhere where it executes as the user before the desktop comes up. So would that be in /etc/X11/gdm/PreSession/Default? I never can remember that whole set of chain of events that lives under /etc/X11...
I think this is related more to Bug 102554. If I logout, re-login - I can reproduce your problem. However, if I "evolution --force-shutdown" or use "killev", the problem goes away.
May be better in evolution 1.4. If not best course is probably a bugzilla.gnome.org bug vs. component bonobo-activation (oaf is fairly old/unmaintained at this point). Or alternately bugzilla.ximian.com evolution bug.