Closed lucas-belo closed 1 month ago
Note that you’ll need to resolve conflicts.
Trying to find the lowest version of the packages with which tests pass, with a binary search strategy among all the packages versions.
Merging #6272 (e91d63e) into master (95a70d3) will increase coverage by
0.10%
. The diff coverage isn/a
.
Waiting for CI, given we have to manually approve every time, might be too time-consuming. You can run tests locally with tox
, e.g. tox -e py
for regular tests with your system version of Python, tox -e extra-deps-pinned
and tox -e extra-deps
for the extra-deps environments, etc.
Hey @Gallaecio and @wRAR, all tests were successful for the last commit on my fork's CI, is there any difference? Maybe re-run the failed tests?
https://github.com/lucas-belo/scrapy/actions/runs/8211648105
3.8 tests fail because of #6286
Great work!
I think we can merge after a run of
pre-commit run --all
to fix the pre-commit CI job, and provided no other unexpected CI job failures.
@Gallaecio, Done!
https://github.com/lucas-belo/scrapy/actions/runs/8249523479
I think the issues in extra-deps-pinned may be real issues. We might need a higher version of some extra dep.
Of uvloop, it seems: https://github.com/MagicStack/uvloop/issues/126
I am gonna see it
With uvloop==0.12.0, all tests passed, but we got 4 erros:
With uvloop==0.14.0 we got all tests successful and no errors, I think this is the right one:
I have added pygments
to deps, and I squashed your commits because, although git diff scrapy/master
showed the right changes, the GitHub diff included 30 files*, making review online a bit harder.
* As to why it happened, I believe either because you merged a PR of mine to fix CI jobs, which I then squash-merged into the main branch, or due to the way you merged/rebased the main branch; doesn’t not really matter.
@lucas-belo Thank you for your amazing work here!
Resolves https://github.com/scrapy/scrapy/issues/6270