Pending Kernels and Windows Unit tests

[Zsailer]

Jupyter Server's unit tests involving pending kernels are failing on Windows in CI—I haven't tested locally.

It might take me some time to fix the issue, especially since I don't have a Windows machine. In the meantime, I mentioned in today's Jupyter Server meeting that we temporarily skip these tests in CI.

I'm opening this issue to track/ensure that we don't forget to fix this issue.

[Zsailer]

@vidartf, not urgent, but would you mind trying to run Jupyter Server's unit tests with the latest release of jupyter_client and seeing if the tests hang when they reach the pending kernels tests?

[vidartf]

Hard to tell, as my entire terminal window closes after running a few tests. I'll try to see if I can narrow down which test(s) are causing this to happen.

[vidartf]

The terminal window closing problem seem unrelated/spurious. Some observations on the tests (need to finish for today):

• Some sessions tests create a lot of these warnings:

  Task was destroyed but it is pending!
  task: <Task pending name='Task-559' coro=<WebSocketProtocol13._receive_frame_loop() running at C:\cleanconda\envs\server-dev\lib\site-packages\tornado\websocket.py:1116> wait_for=<Future pending cb=[IOLoop.add_future.<locals>.<lambda>() at C:\cleanconda\envs\server-dev\lib\site-packages\tornado\ioloop.py:688, <TaskWakeupMethWrapper object at 0x0000026A2FDEDD60>()]> cb=[IOLoop.add_future.<locals>.<lambda>() at C:\cleanconda\envs\server-dev\lib\site-packages\tornado\ioloop.py:688]>
  

• All the nbconvert tests fail with ValueError: No template sub-directory with name 'lab' found in the following paths:. Even though I did install pip install jupyter-server[test].
• tests.test_terminal.test_culling always fails.
• The services.kernels.test_api tests have a tendency to hang the pytest process so it needs to be force killed. I assume these are the tests that you were having issues with. I will do some more runs tomorrow to see if I can narrow things a bit.

[vidartf]

Resuming work here, I now see a bunch of these errors in test_authorizer:

_________________ ERROR at setup of test_authorized_requests[True-GET-/api/kernelspecs/{kernelspec}-None] _________________

cls = <class '_pytest.runner.CallInfo'>, func = <function call_runtest_hook.<locals>.<lambda> at 0x0000021B0490BE20>
when = 'setup', reraise = (<class '_pytest.outcomes.Exit'>, <class 'KeyboardInterrupt'>)

    @classmethod
    def from_call(
        cls,
        func: "Callable[[], TResult]",
        when: "Literal['collect', 'setup', 'call', 'teardown']",
        reraise: Optional[
            Union[Type[BaseException], Tuple[Type[BaseException], ...]]
        ] = None,
    ) -> "CallInfo[TResult]":
        excinfo = None
        start = timing.time()
        precise_start = timing.perf_counter()
        try:
>           result: Optional[TResult] = func()

<env-root>\lib\site-packages\_pytest\runner.py:311:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
<env-root>\lib\site-packages\_pytest\runner.py:255: in <lambda>
    lambda: ihook(item=item, **kwds), when=when, reraise=reraise
<env-root>\lib\site-packages\pluggy\_hooks.py:265: in __call__
    return self._hookexec(self.name, self.get_hookimpls(), kwargs, firstresult)
<env-root>\lib\site-packages\pluggy\_manager.py:80: in _hookexec
    return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
<env-root>\lib\site-packages\_pytest\unraisableexception.py:83: in pytest_runtest_setup
    yield from unraisable_exception_runtest_hook()
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

    def unraisable_exception_runtest_hook() -> Generator[None, None, None]:
        with catch_unraisable_exception() as cm:
            yield
            if cm.unraisable:
                if cm.unraisable.err_msg is not None:
                    err_msg = cm.unraisable.err_msg
                else:
                    err_msg = "Exception ignored in"
                msg = f"{err_msg}: {cm.unraisable.object!r}\n\n"
                msg += "".join(
                    traceback.format_exception(
                        cm.unraisable.exc_type,
                        cm.unraisable.exc_value,
                        cm.unraisable.exc_traceback,
                    )
                )
>               warnings.warn(pytest.PytestUnraisableExceptionWarning(msg))
E               pytest.PytestUnraisableExceptionWarning: Exception ignored in: <function Context.__del__ at 0x0000021B015280D0>
E
E               Traceback (most recent call last):
E                 File "<env-root>\lib\site-packages\zmq\sugar\context.py", line 65, in __del__
E                   warnings.warn(
E               ResourceWarning: unclosed context <zmq.asyncio.Context() at 0x21b0492ed40>

<env-root>\lib\site-packages\_pytest\unraisableexception.py:78: PytestUnraisableExceptionWarning

I believe this is telling us two things:

• Calling warnings.warn in __del__ is not really ok, as depending on the warning filter it might be an exception (and in my case they seem to be causing exceptions).
• When the ZMQ context is garbage collected it is complaining about not having been closed. I'm not sure if this is expected during test runs?

[blink1073]

Yes, the warning is in pyzmq itself, and is a result of us not properly closing contexts in some cases. Hopefully this will all be resolved as part of ongoing asyncio refactoring, e.g. https://github.com/jupyter/jupyter_client/issues/755

[blink1073]

The only tests we're skipping on Windows currently are related to files or unix sockets, closing as fixed.