You're probably right about it being a libzephyr bug. On the other hand it's not uncommon for protocol libraries to have timeouts in them, and I think pidgin could probably be handling this one a little better.
In particular in zephyr_close, pidgin goes out of its way to reset the network timeouts in order to re-try the closing. So it ends up going something like this: * zephyr_login () tries to connect, times out on ZInitialize () * zephyr_disconnect () tries to disconnect, resets the timeouts, and then times out again trying to clear subscriptions that have never been set It'd be better if the zephyr_disconnect () code were to only try to shut down state that it was actually able to initialize, thus avoiding the second timeout. But I agree that it's a relatively minor problem, and that the real problem seems to be that libzephyr is hanging when the zhm isn't running. Ari Pollak wrote: > On Tue, 2007-11-13 at 23:03 -0600, Klee Dienes wrote: >> The periodic UI hangs started happening a few months ago, and I >> suspect are unrelated to the X crashes I was having earlier today. >> They seem to be a fairly straightforward block in libzephyr trying >> to cancel out the zephyr subscriptions, even though the reason it's >> closing the connection is that it wasn't able to connect to the >> zephyr server in the first place > > That seems like a libzephyr bug, no? > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
