You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
EPICS tests take 15s to run. Tango tests take 85s to run. This makes waiting for CI painful. I think a target of 20s for Tango tests is reasonable.
Some ideas:
Some tests take a long time because of long timeout or many repeats of the same code. See if these can be reduced.
All tango tests are currently forked, because this is apparently necessary, but when there is a module level fixture then we maybe we could get away with forking then running multiple tests? Test this theory by taking one of the parameterized tests and turning the parameters into a for loop inside the tests. If this is a good strategy then we could work out if it possible to extend pytest-forked to support forked fixtures, where every test that uses the fixture runs in the same fork.
Some tests may be duplicates and could be deleted
Acceptance Criteria
Tango tests take 20s to run
The text was updated successfully, but these errors were encountered:
Taking this to mean we can use the same echo_device for each test instead of generating a new one each time, looping over the ATTRIBUTES_SET inside test_tango_signal_rw takes the time from ~16s to under a second, so definitely worth trying to find a nicer way to handle the fixtures!!
EPICS tests take 15s to run. Tango tests take 85s to run. This makes waiting for CI painful. I think a target of 20s for Tango tests is reasonable.
Some ideas:
pytest-forked
to support forked fixtures, where every test that uses the fixture runs in the same fork.Acceptance Criteria
The text was updated successfully, but these errors were encountered: