Uh oh!
There was an error while loading. Please reload this page.
gh-113937 Fix failures in type cache tests due to re-running - #113953
gh-113937 Fix failures in type cache tests due to re-running#113953Fidget-Spinner merged 2 commits into
Conversation
Fidget-Spinner
left a comment
There was a problem hiding this comment.
Seems right. Thanks for the quick fix! We usually don't need a news entry for tests unless its significant or went out in a released Python version.
lazorchakp
commented
Jan 11, 2024
Sounds good. Hopefully this is the only issue here. I'm considering waiting until tomorrow to perform the backports just in case another bug comes up - I've read through https://devguide.python.org/testing/run-write-tests/ and think I have everything covered, but I'm still somewhat pessimistic |
Fidget-Spinner
commented
Jan 12, 2024
I could test with buildbots to see if it would pass on main, but main buildbots seem to be currently failing so I won't do that. |
|
Fidget-Spinner
commented
Jan 12, 2024
I'll investigate later. |
Fidget-Spinner
commented
Jan 12, 2024
Just to clarify: this PR fixed the buildbots. The other PR #113937 was the one that introduced the leaks. |
lazorchakp
commented
Jan 12, 2024
From what I can tell, after this PR there are no longer any issues related to If everything still looks good next week, I'll backport #113558 + this PR to 3.11 and 3.12. |
This addresses test failures in
test_type_cachearising from re-running the same tests multiple times. There are no source code changes. New behavior in test_type_cache:To test these changes, I ran the following tests:
./python.exe -m test test_type_cache./python.exe -m test test_type_cache --forever(killed manually after 5 minutes)./python.exe -m test test_type_cache -R 3:3I specifically did not add a news entry, but will happily do so if anyone thinks it's appropriate.
test_type_cachefails with--foreverargument #113937