This repository has been archived by the owner on Sep 2, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 5
Investigate slowness of unit tests #1220
Comments
I think this is due to logging, on my machine:
This fits with the issue being whenever there is an exception as this is when the debug logs get flushed. I suggest we turn off logging by default in the tests and only turn on for the tests that are testing logging logic? |
Solution is to:
|
Will take ~0.5 days |
Interestingly this didn't give us much of a speed up. I guess enough debug logs are produced in general that the buffer becomes reasonably full part way through the tests anyway |
DominicOram
added a commit
that referenced
this issue
Mar 13, 2024
…ytest flag that didn't seem to do anything
DominicOram
added a commit
that referenced
this issue
Mar 13, 2024
DominicOram
added a commit
that referenced
this issue
Mar 13, 2024
olliesilvester
pushed a commit
to olliesilvester/mx-bluesky
that referenced
this issue
Aug 23, 2024
…also remove debug-logging pytest flag that didn't seem to do anything
olliesilvester
pushed a commit
to olliesilvester/mx-bluesky
that referenced
this issue
Aug 23, 2024
olliesilvester
pushed a commit
to olliesilvester/mx-bluesky
that referenced
this issue
Aug 23, 2024
…me test with no changes each time
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Following on from #1215
There are still a number of unit tests that seem to be slowing down unit test execution. This seems to be related to exceptions that are raised during successful execution of a test, and these seem to cause a delay. The reason why these exceptions cause the delay is unknown.
e.g. in
test_given_no_tip_found_ever_when_get_tip_into_view_then_smargon_moved_positive_and_exception_thrown()
a number of exceptions similar to the following are thrown and they each cause ~0.7s of delay:It would be good if we could speed up the slow tests by eliminating the exceptions, ideally we would also want to:
In investigating the latter, one possible avenue may be to temporarily re-instate the exceptions in the optimise_attenuation test and run the profiler against it (may need to tweak the timeout so that the exceptions happen before the test completes), possibly that may make the cause more obvious.
The text was updated successfully, but these errors were encountered: