> > is the deadlock still there? > > We are still testing with the fix of Nico, the deadlock seems to be gone. > To be sure, we have to test a few days more. Just for the record. I don't think moving the time is the solution it's better now (in the sync block) but the DijkstraaSemaphore will never get locked for more then a few CPU cycles.
I don't thing RvM solved the bug because I think I did. but changes have been made to the oracle driver by RvM that might have solved your bug(I don' know). Anyway. I agree that we must start behaving "seriousely" but community has already proved it selve and the mailing list /irc have been very good tools to communicate between developers. so maby it a good time to create other "channels" for people interesed in MMBase or people using MMBase (in the way they use a web-server or something else) greets, see you at the dev meeting?
