Conversation
c880cde to
2c68c43
Compare
Unit Test ResultsSee test report for an extended history of previous test failures. This is useful for diagnosing flaky tests. 15 files ±0 15 suites ±0 6h 36m 59s ⏱️ - 4m 28s Results for commit 82fc870. ± Comparison against base commit 2963bf1. ♻️ This comment has been updated with latest results. |
2c68c43 to
9caefe4
Compare
jrbourbeau
left a comment
There was a problem hiding this comment.
Thanks @graingert. cc'ing @wence- into this PR to make sure we're coordinating with #7134, which looks very similar
d0ce45b to
fd74414
Compare
this is the code from #7134 but without the looprunner changes, because they're not strictly needed for 6.2 support - but they are desirable soon - maybe in the next release? |
Thanks @graingert for factoring this out, I am hoping to come back to #7134 next week. |
jrbourbeau
left a comment
There was a problem hiding this comment.
this is the code from #7134 but without the looprunner changes, because they're not strictly needed for 6.2 support - but they are desirable soon - maybe in the next release?
Thanks @graingert, that context is useful. I just saw multiple PRs, so wanted to make sure we were all on the same page.
| - tblib | ||
| - toolz | ||
| - tornado<6.2 | ||
| - tornado >=6.2 |
There was a problem hiding this comment.
This seems like the version restriction we want, given what's specified in requirements.txt
| - tornado >=6.2 | |
| - tornado >=6.0.3 |
There was a problem hiding this comment.
if the restriction doesn't require >=6.2 in the environment then it follows what's currently released for the distributed package
There was a problem hiding this comment.
I think the problem we're having here is that all python versions are now testing against 6.2 but we're not testing against 6.0.3 anymore, are we?
I'm fine with this for now if we add <6.2 to #7285
There was a problem hiding this comment.
I think the problem we're having here is that all python versions are now testing against 6.2 but we're not testing against 6.0.3 anymore, are we?
Yeah, that's what I was wanting to avoid.
I'm fine with this for now if we add <6.2 to #7285
Yep, I think the mindeps build will be the long term answer for testing against our lowest supported minimum version.
FWIW I didn't mean for this comment to be majorly blocking. @graingert if you and @fjetter are happy with things as they are, that's fine by me.
Co-Authored-By: Lawrence Mitchell <lmitchell@nvidia.com>
tornado 6.2 deprecates functionality that relies on deprecated asyncio calls (get_event_loop and friends); a few tests still use a pattern of implicitly creating a new loop (and test that appropriate deprecation warnings are uttered), so just catch these new tornado warnings for now, until the functionality is removed in distributed.
fd74414 to
82fc870
Compare
Closes #6669.
pre-commit run --all-files