Just changing the code to not block during TearDown causes problems in the test suite because it loses the synchronous nature. We use a hand-crafted mock/stub, and issue a request and then a Kill, and that request should trigger 2 calls. But since our Kill() *should* kill things in flight, it wasn't a great test to begin with.
Just changing the code to not block during TearDown causes problems in the test suite because it loses the synchronous nature. We use a hand-crafted mock/stub, and issue a request and then a Kill, and that request should trigger 2 calls. But since our Kill() *should* kill things in flight, it wasn't a great test to begin with.