FIX: do not cache canvas object #13606
Merged
+12
−17
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Matplotlib may change the canvas object on a figure. It is better to ask the figure what its current canvas is rather than cache the one it has the first time you see it.
larsoner
approved these changes
Jan 22, 2026
Member
larsoner
left a comment
larsoner
left a comment
There was a problem hiding this comment.
LGTM marking for merge-when-green, thanks for the thorough investigation @tacaswell !
Comment on lines
-328
to
-330
| # https://github.com/matplotlib/matplotlib/issues/30575 | ||
| monkeypatch.setattr(mne.viz.utils, "_BLIT_KWARGS", dict(useblit=False)) | ||
| monkeypatch.setattr(mne.viz._mpl_figure, "_BLIT_KWARGS", dict(useblit=False)) |
Member
There was a problem hiding this comment.
@tacaswell I pushed a commit to remove the workaround...
| .. _Teon Brooks: https://github.com/teonbrooks | ||
| .. _Tharupahan Jayawardana: https://github.com/tharu-jwd | ||
| .. _Thomas Binns: https://github.com/tsbinns | ||
| .. _Thomas Caswell: https://tacaswell.github.io |
Member
There was a problem hiding this comment.
... and add a changelog entry for you, pulled this URL from your GitHub profile hopefully it's good enough!
Contributor
Author
There was a problem hiding this comment.
Thank you!
Sorry for not following your contribution guidelines. We debugged this from the traceback you posted on the mpl issue during our weekly dev call and I squeezed in getting this PR open between the next set of meetings, some other work, and dinnertime!
Member
There was a problem hiding this comment.
No problem, happy to push a quick commit to get this over the finish line! You did the hard work of debugging, much appreciated
|
🎉 Congrats on merging your first pull request! 🥳 Looking forward to seeing more from you in the future! 💪 |
This was referenced Jan 23, 2026
larsoner
added a commit
to teonbrooks/mne-python
that referenced
this pull request
Jan 30, 2026
* upstream/main: (23 commits) DOC: Add jupyterlite idea to roadmap (mne-tools#13620) MAINT: Use f-strings in test_import_nesting.py (mne-tools#13551) Improve docs for raw.to_data_frame (mne-tools#13590) Sensitivity map doc improved (mne-tools#13578) [pre-commit.ci] pre-commit autoupdate (mne-tools#13612) MAINT: Fix for latest SciPy (mne-tools#13613) Fix pre-commit call in SPEC0 action [ci skip] (mne-tools#13609) MAINT: Add mne-denoise to CI dependencies (mne-tools#13607) FIX: do not cache canvas object (mne-tools#13606) FIX: Set calibration plot axes to screen resolution if available (mne-tools#13558) Refactoring eyetracking.py (mne-tools#13602) [pre-commit.ci] pre-commit autoupdate (mne-tools#13601) Follow up PR to PR - mne-tools#13596 (mne-tools#13599) FIX: Sphinx (mne-tools#13600) Doc improvement - Examples using <some-method> section quirk fix (mne-tools#13596) Add more information to eSSS in examples and docsstring (mne-tools#13591) np.fix -> np.trunc (deprecation) (mne-tools#13594) Make mne.sys_info() work with powershell 7+ (mne-tools#13593) BUG: Fix minor bug with T1 check (mne-tools#13588) [pre-commit.ci] pre-commit autoupdate (mne-tools#13587) ...
larsoner
added a commit
to contsili/mne-python
that referenced
this pull request
Jan 30, 2026
* upstream/main: (67 commits) DOC: Add jupyterlite idea to roadmap (mne-tools#13620) MAINT: Use f-strings in test_import_nesting.py (mne-tools#13551) Improve docs for raw.to_data_frame (mne-tools#13590) Sensitivity map doc improved (mne-tools#13578) [pre-commit.ci] pre-commit autoupdate (mne-tools#13612) MAINT: Fix for latest SciPy (mne-tools#13613) Fix pre-commit call in SPEC0 action [ci skip] (mne-tools#13609) MAINT: Add mne-denoise to CI dependencies (mne-tools#13607) FIX: do not cache canvas object (mne-tools#13606) FIX: Set calibration plot axes to screen resolution if available (mne-tools#13558) Refactoring eyetracking.py (mne-tools#13602) [pre-commit.ci] pre-commit autoupdate (mne-tools#13601) Follow up PR to PR - mne-tools#13596 (mne-tools#13599) FIX: Sphinx (mne-tools#13600) Doc improvement - Examples using <some-method> section quirk fix (mne-tools#13596) Add more information to eSSS in examples and docsstring (mne-tools#13591) np.fix -> np.trunc (deprecation) (mne-tools#13594) Make mne.sys_info() work with powershell 7+ (mne-tools#13593) BUG: Fix minor bug with T1 check (mne-tools#13588) [pre-commit.ci] pre-commit autoupdate (mne-tools#13587) ...
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Matplotlib may change the canvas object on a figure. It is better to
ask the figure what its current canvas is rather than cache the one it has the
first time you see it.
Reference issue (if any)
This may address the new issues idenitfied in
matplotlib/matplotlib#30575
What does this implement/fix?
In mpl 3.11 we are being better about cleaning up GUI state on Figures when
they are closed. As part of this, when closed the canvas is replaced with a
base non-operative canvas. We think that because mne is caching the initial
GUI/Agg canvas the tear down callbacks are driving the figure to a chimera state
that is the source of the problems.
Additional information