-
Notifications
You must be signed in to change notification settings - Fork 88
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Mark more tests as slow #1230
Comments
functional_duration.txt I have made a compilation (not full) of the durations for the tests in each test_folder. The question is what is your take on the duration bound above which we should mark a test as slow before starting to look at your suggestions @keileg? |
Thank you for the overview. If I understand the outputs correctly, it looks like all tests by itself are actually reasonably slow... |
@jwboth , so the output only provides all test durations in a descending order. |
Could you make sure the regression test actually runs, please? I suspect it's quite slow. |
Thanks for the overview, @mikeljordan. Where to put the cutoff is not clear, we should discuss in the developer group at some point. |
Attached is the durations for the regression test. |
Thanks for the regression results, @mikeljordan. They confirmed my suspicion. Regarding cut-off: Isn't the even more important question which tests can safely be silenced? |
Running the test suite on GH actions now taskes about 11 minutes. This is markedly longer than what was the case, say, a year ago, and it is a real drag on the workflow for PRs.
To speed things up, we should revisit the test suite and mark more tests as being slow, and thus not be part of the standard checks run as part of a PR. This must be done carefully to avoid weakening the quality checks applied to a PR. A place to start is to look for tests that are generously parametrized, and costly, and see if these by default can be run with fewer parameter cominations and preserve the integrity of the test suite.
The text was updated successfully, but these errors were encountered: