LFPlugin: use sub-plugins to deselect during collection#6448
LFPlugin: use sub-plugins to deselect during collection#6448blueyed merged 1 commit intopytest-dev:masterfrom
Conversation
| "* 1 failed in *", | ||
| "collected 2 items", | ||
| "run-last-failure: 1 known failures not in selected tests", | ||
| "* 1 failed, 1 passed in *", |
There was a problem hiding this comment.
Not really sure, but this might show the fix for #5301.
Could use a more explicit test otherwise, of course.
|
Thanks, sorry for the delay. At first glance the approach seems to work, but also seems a bit more complicated than I would like, so before we discuss this PR further I would like for us to consider the alternative using mtime you proposed in #5301 (comment). 👍 |
|
@nicoddemus |
|
@nicoddemus |
|
Yes, have you? |
|
No. But I've guessed that an existing (changed) test would cover it already (#6448 (review)). Not sure though, and it might be good to review/extend it with regard to that. |
|
I suggest to make use of provided reproducible use cases when opening PRs claiming to fix a bug ;) Either way, the change seems fine so feel free to merge it. |
Fixes pytest-dev#5301. Refactor/steps: - use var - harden test_lastfailed_usecase - harden test_failedfirst_order - revisit last_failed_paths - harden test_lastfailed_with_known_failures_not_being_selected
It should only return `True` when something is to be ignored, not `False` otherwise typically. This caused e.g. bad interaction with the cacheprovider (before pytest-dev#6448).
It should only return `True` when something is to be ignored, not `False` otherwise typically. This caused e.g. bad interaction with the cacheprovider (before #6448).
Fixes #5301
Two subplugins are necessary due to limitations in pluggy (pytest-dev/pluggy#246).
TODO: