New patch threadstate_join_5.patch adds more testing of is_alive().

An inelegance I don't care about (but someone might):  if join() is called with a timeout, and the Python part of the thread ends before the timeout expires (_stopped gets set), then a _non_-blocking attempt to acquire _tstate_lock is made, and join() returns regardless of the outcome.  So, with a timeout, it's possible for join() to return before the C part of the thread is done even if the timeout isn't yet exhausted.

That's unlikely, and I don't see that it makes any difference.  Anyone doing a join() with a timeout has to be aware that they have no idea whether the thread is done when join() returns, and do another join() or check is_alive() later.

I'd rather not complicate the code to wait longer for _tstate_lock in this case.