Sorry, something went wrong.
There was a problem hiding this comment.
Looks good. Any reason you didn't mark this for backporting to 3.10?
Sorry, something went wrong.
|
Looks good. Any reason you didn't mark this for backporting to 3.10? No real reason not to -- for all these other coverage improvements linked to #94808, we've just been backporting to 3.11. |
Sorry, something went wrong.
|
Thanks @mdboom for the PR, and @JelleZijlstra for merging it 🌮🎉.. I'm working now to backport this PR to: 3.10, 3.11. |
Sorry, something went wrong.
|
Sorry, @mdboom and @JelleZijlstra, I could not cleanly backport this to 3.11 due to a conflict. |
Sorry, something went wrong.
|
Sorry @mdboom and @JelleZijlstra, I had trouble checking out the 3.10 backport branch. |
Sorry, something went wrong.
|
Thanks! I think it's good to backport new tests to the bugfix branches so we can be more confident in any future bugfixes. But testing on 3.11 is definitely more important. Would you mind doing the manual backports? Feel free to skip 3.10 if you don't think it's worth it. |
Sorry, something went wrong.
When both are provided, tp_ass_subscript takes precedence over tp_ass_item. Since bytesarray provides both, the existing test_setitem tests for bytesarray were not testing bytesarray_setitem, but bytesarray_ass_subscript. This is mostly fine, since Python code has to jump through some hoops to even call it, but a third-party library using PySequence_SetItem could potentially run into this uncovered case.