Closed martica closed 5 years ago
There are still a couple unexpected successes that failed locally for me, and one test failure in the 3.6 run, but looking at the error, it seems a candidate for flakiness as it depends on set ordering.
Everything that fails is now marked as expected.
These changes look great to help use get to a state where the expected tests are passing. If we get this merged in before #772, then I can remove the Python 3.4 expected failures since we are dropping support for that version.
Cool. I'm looking forward to rebasing the rest of my changes on this and seeing everything green. :)
Is there anything else that needs to be done to merge this? Brutus made some style comments that I corrected. All but the first two are marked as outdated, but even the first two have been corrected, so I'm not sure why it isn't updating to outdated.
@martica Thanks! :heart:
This attempts to mark all the outstanding test failures as expected to allow other changes to be submitted with proper test coverage and avoid continued tests from rotting.
PR Checklist: