CI: temporary fix to the CI #32011
CI: temporary fix to the CI #32011
Conversation
|
OK, i guess a very 'quick and dirty' fix for the benefit of the doc sprint. can you do an IMMEDIATE follow-up PR with a skipif or xfail for np version only with the issue as a reason |
and undoes the commenting out |
I have made it "xfail" for now, what would be a good "skipif"? |
|
I don't this the xfail will work, since some environments pass. happy to keep this PR with the commenting out to get CI green and do this on another PR. |
|
Sorry for the "force-push" I tend to fight with git a lot. |
|
ping @simonjayhawkins Numpydev is green. |
| @@ -384,8 +384,11 @@ def test_td_div_nan(self, nan): | |||
| result = td / nan | |||
| assert result is NaT | |||
|
|
|||
| result = td // nan | |||
| assert result is NaT | |||
| # TODO: Don't leave commented, this is just a temporary fix for | |||
There was a problem hiding this comment.
make sure u open an issue
an xfail with a version test is the right way to handle this
i guess ok for right now
There was a problem hiding this comment.
@MomIsBestFriend the issue will be to revert this PR, which is different. Since I assume that another PR will follow this immediately for that purpose, the issue may not be necessary.
OK since ci has a backlog, will wait a little longer for full green to be sure. |
|
will NOT be backporting this as it is a very temporary fix (duration of doc sprint only) and merging doc PRs |
I completely agree, the doc sprint run is the sole purpose of why this PR exists |
|
i'm going to merge this shortly as it seems that Linux py37_locale has hung and restarting is pointless given the ci backlog. any objections @jreback |
|
@simonjayhawkins go for it |
|
@datapythonista I leave it to you to weight up the benefits of merging PRs to master vs reducing the burden on ci for the benefit of the sprint. I think OK to merge PRs to master without getting contributors to rebase to reduce the CI burden further |
|
@MomIsBestFriend do you want to take the lead on the follow-up PR? if not, can you raise an issue as @jreback suggested. |
|
ah almost forgot.... Thanks @MomIsBestFriend. |
|
@datapythonista as an aside. I had the same issue with the CI at the Man Hackathon. after just a few hours into the sprint, ci was taking more than 2 hours to complete. |
|
@MomIsBestFriend np-dev still failing on master! |
cannot reproduce locally. I assume that I am not setting up the environment/build correctly. maybe skip/xfail those as well? |
ref #31992
This is just a temporary fix to the CI.
Until someone smarter than me fix it :)