Liu Hui <liuhui1...@gmail.com> writes: > OK, I have added a test to the patch.
While your test works on its own, it seems to break subsequent tests (the next test hangs). My guess is that it has something to do with the fact that most of the Python session tests share the same session, and ob-python is getting confused about which session to use on the next test. One possible fix might be to kill the Python session started by your new test, after it's finished running. (It is probably not the best design that the Python session tests all share the same session...)