Close database might hang waiting for no active replicators or live queries

I have found this issue as testDeleteWithActiveReplicationsAndURLEndpointListener was hanged on Jenkins. I couldn't reproduce the issue locally but after reviewing the code, I think the problem can be described as below.

In CBLDatabase's removeActiveStoppable: method, _closeCondition was notified without locking. So there could be a race condition with when _closeCondition is waiting for the broadcard signals and cause _closeCondition to wait forever.