Closed krassowski closed 3 months ago
From the macOS failure it looks like your workaround could also be applied to test_room_concurrent_initialization
?
The Windows failure is caused by the jupyter_server_fileid
manager which does some cleanup on garbage collection, which I don't think is a good practice.
Well, down to one failure:
I guess we can try to force jupyter_server_fileid
to delete during the test too?
But the failure means that your workaround didn't work, right?
Not quite sure. Could it have been an interaction with jupyter_server_fileid
? It is all passing now?
I think it's flaky, we should rerun tests to check.
I will close and reopen to see if we get all green again (of note Tests / Test Minimum Versions (pull_request)
looks stuck right now).
Addresses issue discussed in https://github.com/jupyterlab/jupyter-collaboration/issues/252 to help bring the CI green again.
The Windows tests are still failing though - are these indicative or Windows-specific bugs?