Matthew Seal

Looks like tests are failing now due to some test scaffolding including the input_path in the arguments.

So the issue is below the papermill level. This is saying that the kernel you are trying to run never replied to the intial ZMQ ping. What kernel are you...

Hi @rchui, The reason it doesn't do this by default is more a security problem than a capability. If you don't know the input source of your parameters you open...

That looks suspiciously like some issue with async loop configuration or modifications. Can you list the wider list of packages that are locally installed and if you have any init...

Oh good catch, that should be improved

So tuples are partially specific to Python. If you change the input to a list instead of a tuple you should get a direct translation that's accurate. That conversion could...

Are these test being run in parallel? https://github.com/jupyter/jupyter_client/issues/487 is a known race condition with port acquisition. I haven't had to time to implement the proposed handshake changes to remedy that...

In general nbclient is kept less complicated on this capability and exists as a pure manager of kernel object that can be easily extended to more sophisticated use cases (e.g....

That's definitely not an expected behavior. I take it your build chain with 2 cores is passing (to eliminate possibility of it being dependency related rather than hardware)? Are you...

test_async_parallel_notebooks has a known issue rooted in jupyter_client that is unresolved: https://github.com/jupyter/jupyter_client/issues/487. Basically there's a race condition that test are sometimes loosing for port aquisition.