Alan Conway wrote:
I have seen other python tests hang, and more rarely client_test. I suspect it coincides with the new epoll IO layer but I'm not sure of that.
I suspect it could also be the 'execution layer'; if the client gets out of step with the broker on its understanding of the implicit command ids then it could end up waiting indefinitely for completion.
I've made a minor change to the python code that will timeout after a reasonable wait in this case (will checkin later this morning). That will at least provide a clearer indication of the point of failure. The cause will likely be some race condition on the broker.
