Uh oh!
There was an error while loading. Please reload this page.
gh-140138: NOGIL: add atomic owner refcount to PyInterpreterState; finalize on last tstate - #140224
gh-140138: NOGIL: add atomic owner refcount to PyInterpreterState; finalize on last tstate#140224ashm-dev wants to merge 11 commits into
Conversation
ashm-dev
commented
Oct 16, 2025
Changes made on top of the previous fix from PR |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
ZeroIntensity
left a comment
There was a problem hiding this comment.
Thanks, this PR looks good if we decide to fix this, but I'm worried about whether this is even a bug in the first place. See my comment on the issue. Please let me know if I missed something in my analysis.
Move the `_owners` field to be after `finalizing` and add it back to the struct. It was removed in a previous commit and is needed.
Fixes a typo where the atomic variable `_owners` was incorrectly named and should have been `owners`.
Rename functions to be more descriptive of their actions.
Move the cleanup logic into PyInterpreterState_Delete and remove the now-unused helper function.
The field name was inconsistent with other usages.
ashm-dev
commented
Oct 19, 2025
@ZeroIntensity this PR necessary or is it better to close it? |
ZeroIntensity
commented
Oct 19, 2025
I don't think this PR is necessary. |
Uh oh!
There was an error while loading. Please reload this page.