Random test failures during CI

[mvdwetering]

In CI there are regularly random test failures like the one below.
It seems to be timing related and an attempt was made to "fix" it with a sleep, but stil randomly fails.

This should be fixed properly. Not sure how though. Kind of want to wait for the disconnect callback to be called with a timeout, but can't figure out how to do it.

It is also weird that it takes > 0.5 seconds for it to be called after calling close() on the serial mock. Maybe there is something going on there also.

2024-11-09T14:13:48.6356054Z         mock_serial.close()
2024-11-09T14:13:48.6356643Z         time.sleep(SHORT_DELAY)
2024-11-09T14:13:48.6357183Z     
2024-11-09T14:13:48.6357818Z >       assert disconnect_callback.call_count == 1
2024-11-09T14:13:48.6358576Z E       AssertionError: assert 0 == 1
2024-11-09T14:13:48.6359530Z E        +  where 0 = <MagicMock id='140220867950992'>.call_count

Attached full logs from a failed run because I don't know how long the logs will be kept. Failure was in the test_wtih_pytest step for Python 3.11.
logs_30695975446.zip