The interpreter is not crashing in your case, so this is a mostly-unrelated problem.  (The part that is related is that it is triggered by module finalization.)  As Berker said, please open a new issue, but be warned that it may get closed as a being addressed by the issue for rewriting the python startup sequence (pep 432 I think, which I think Nick has resurrected), since this is a systemic problem.  On the other hand there might be a way to fix it in the datetime module if someone decides it is worth doing.