Skip to content

Conversation

@ngoldbaum
Copy link
Contributor

See #28611 (comment) for context.

Copy link
Contributor

@greglucas greglucas left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I tested this out on a mac and this does get rid of the GIL warning and everything works as expected.

Also as an FYI, I tested this with multiple windows trying to resize/zoom/pan between them and all of that state seemed good to me.

I also ran the threading example from here: #5675 (comment), updating data in background threads and that also played nicely with resizing/interactiveness and the free-threaded build. So I think we are at least OK with reasonable expectations for now. It will be good to get people hammering this a bit more and more exotic use-cases to be tested out.

edit: I also just added constrained_layout to test our interaction with kiwisolver's free threaded capability and that works nicely too.

@greglucas greglucas merged commit bfbd0b5 into matplotlib:main Mar 27, 2025
40 of 41 checks passed
@QuLogic QuLogic added this to the v3.11.0 milestone Mar 27, 2025
@ngoldbaum
Copy link
Contributor Author

@greglucas just noticing that I'm hitting the issue fixed by this on a fresh Python 3.14.0t environment. And indeed, the latest release is 3.10.7, so this fix isn't included. Does it maybe make sense to backport?

@tacaswell
Copy link
Member

This does not feel like a back-port type thing to me, but could be talked into it.

@greglucas
Copy link
Contributor

I would generally agree that this seems potentially odd to put in a bugfix release. On the other hand, this is pretty niche still and I'm not sure guarantees us to anything as it is pretty heavily opt-in at this point still.

It has been 10 months since our last release https://github.com/matplotlib/matplotlib/releases/tag/v3.10.0, so I think people in the community are getting excited about updates and new features and wondering when they will be released, so I do understand the desire to get the features people have added out in some kind of release. I'm not sure if another bugfix release will happen again before the next version's release though, so maybe a backport is moot here 🤷

@greglucas
Copy link
Contributor

Actually, I do think this is valid for a backport because it was a bug that we forgot to mark this specific module in the original PR adding this declaration, so in a sense this is fixing that oversight.
#28819

@greglucas greglucas modified the milestones: v3.11.0, v3.10.8 Oct 16, 2025
@greglucas
Copy link
Contributor

@meeseeksdev backport to v3.10.x

meeseeksmachine pushed a commit to meeseeksmachine/matplotlib that referenced this pull request Oct 16, 2025
QuLogic added a commit that referenced this pull request Oct 30, 2025
…tension (#30656)

Co-authored-by: Greg Lucas <greg.m.lucas@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants