I think I can fix those. Mostly related to the multi-threaded unit tests where an assert outcome depends on certain things happening in a parallel thread. So there's always a chance that on a slow machine those things don't happen very quickly.
Will take a look hopefully later tonight. Andrus
