Do not run the Test CI for PRs marked as draft#57
Do not run the Test CI for PRs marked as draft#57nrueh wants to merge 4 commits intopotassco:masterfrom
Conversation
|
@rkaminsk @MaxOstrowski @sthiele @susuhahnml |
There was a problem hiding this comment.
And what is the issue with running the tests for drafts? I would argue that we should keep the tests running at all time. If tests temporarily fail, they can be marked as expected to fail in the code.
From my point of view the CI can also run for drafts.
|
In my opinion it can be a annoying when I am pushing some code that I know already is not passing the tests and it feels like a lot of work to manually disable them temporarily. For example, for encoding or fact format changes with clintest. If you still want to run the tests, you can run them locally with little effort. |
|
The case you are describing should be rather niche. When refactoring, it is best to keep the tests running at all time. Like this we catch bugs early. If you want to skip the CI, I would suggest to tag the commit: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/skip-workflow-runs. |
2 participants
This is a suggestion to reduce the number of times the Test CI is run.
In my opinion in early stages of a PR it's not necessary to run the CI for every push.
I did some googling to get the correct code but maybe there are still ways to improve this.
I also used this to updated some dependencies and increase the default Python versions for testing but I am open to other suggestions. Not sure what the (dis-)advantages of each version currently are...