Repository navigation
Ensure we do not fail when CI leap second file is expired. - #9908
Conversation
pllim
left a comment
pllim
left a comment
There was a problem hiding this comment.
I was just being baffled by this failure. Thanks for fixing it so quickly!
pllim
added
the
zzz 💤 merge-when-ci-passes
|
Now if only it solved the failures... |
|
I think the |
|
Yes, I think the test I do is not exactly the same as what is used in the real code... Will push a correction shortly... |
9c079ec to
08cf61c
Compare
|
Turns out our built-in leap-second file was expired, which was the real underlying source of the errors. I now updated that as well, and added a test that will make it clearer if that is the real problem. |
| ls = iers.LeapSeconds.open() | ||
| assert ls.meta['data_url'] == iers.IERS_LEAP_SECOND_FILE | ||
| # The built-in file should be kept up to date. | ||
| # TODO: would be nice to have automatic PRs for this! |
There was a problem hiding this comment.
Yes, but now every half year we'll get a reminder... (more seriously, will add this to #1697)
There was a problem hiding this comment.
and I see that issue ends with me promising to open a PR to the release procedure... 🤦♂️
There was a problem hiding this comment.
release procedure
Not sure how I feel about it being part of the release procedure. However it certainly can be a "last before release" type of reminder, similarly to the bundled libraries. See #9018
There was a problem hiding this comment.
maybe we can make this a cron only test (or simply just flag with remote_data thus allowed to fail on PRs), so as not to introduce false positives while the update PR is not merged?
There was a problem hiding this comment.
👍 to remote_data with a comment to explain why. But it will have to be a follow-up PR.
There was a problem hiding this comment.
OK, I'll make a quick PR that marks this remote_data - agreed that we don't really want to have regular PRs start failing with it!
|
|
08cf61c to
3769006
Compare
|
Grr, small error still.... |
|
Just to be sure, are any of the numpy-dev job failures related to this? |
|
One of those is |
Ensure we do not fail when CI leap second file is expired.
3 participants
We are getting failures because travis does not quite keep its distribution up to date enough for leap second tests. So, this PR changes the test such that on CI if the leap second file is expired, the test is simply skipped. But if it is not expired, it checks that it is used.
cc @bsipocz, as it is something that should make your life that little bit easier!