ENH XFAIL in common tests with estimator tags (v3)#16502
ENH XFAIL in common tests with estimator tags (v3)#16502glemaitre merged 18 commits intoscikit-learn:masterfrom
Conversation
Co-Authored-By: Joel Nothman <joel.nothman@gmail.com>
| if xfail_checks: | ||
| if check_name in xfail_checks: | ||
| msg = xfail_checks[check_name] | ||
| request.applymarker(pytest.mark.xfail(reason=msg)) |
There was a problem hiding this comment.
There is probably a way to make this work for contrib projects automatically with parametrize_with_checks, but for now I just want something that we can use in scikit-learn as the ability to XFAIL is blocking multiple PRs.
jnothman
left a comment
There was a problem hiding this comment.
I'm. much happier with this..
Co-Authored-By: Joel Nothman <joel.nothman@gmail.com>
|
Thanks @rth and @thomasjpfan. I think that we should look at the |
|
Thanks for the reviews! Looks like somebody broke master,
hope it's intermittent. If you have other comments @thomasjpfan let me know, we can address them in follow up PRs. |
A third version of #16306 and #16328
Puts the common checks to mark as xfail into the
_xfail_testestimator tag.Hopefully addresses concerns raised in the two previous implementations. In particular, this would make the workflow identical between scikit-learn and contrib projects.
cc @thomasjpfan @jnothman @glemaitre