Convert .format() to f-strings #6035
Conversation
ruff 0.3.1 (which we will get with the next tasks container refresh) starts complaining about some .format() usages.
There was a problem hiding this comment.
This particular failure is extremely well-timed and I'd like to avoid wasting the opportunity it presents.
How do you feel about:
- separate PR introducing the
.cockpit-ci/containerfile to pin us back to the old version (making things green again) - a lock-step single commit to update to the new version while fixing the issues here?
Just to prove the concept, basically...
The thing is, I guess this wouldn't actually work properly currently, since the failing test is a workflow. We'd need to up our game a bit there.
If you'd prefer to take the "omg main is red we need to fix this now" tack, I'm not going to insist.
|
@allisonkarlitskaya you probably meant that commit for cockpit-project/cockpit#20149 ? that might also fail with a new tasks container, but it wouldn't help us here indeed. And in cockpit it's the "tox" workflow, which isn't pinnned. A bit "meh".. But I'll think about it at some later time (this day has been long enough) |
|
@allisonkarlitskaya TBH I'd like to wait with introducing .cockpit-ci/container until we have an |
|
@allisonkarlitskaya : c-machines is a perfect example to start that. It fails with ruff 0.3.1, so how about we test-drive it there, and use that as a first opportunity to automate container updates? |
There was a problem hiding this comment.
@allisonkarlitskaya : c-machines is a perfect example to start that. It fails with ruff 0.3.1, so how about we test-drive it there, and use that as a first opportunity to automate container updates?
Deal.
ruff 0.3.1 (which we will get with the next tasks container refresh) starts complaining about some .format() usages.
See this test run, which was running against a locally (in the workflow) refreshed tasks container.
Our integration test already exercises PRs, image refresh etc., but I still triggered two tests for good measure.