Thanks to visit codestin.com
Credit goes to github.com

Skip to content

FIX: maybe improve renderer dance #22732

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Apr 4, 2022
Merged

FIX: maybe improve renderer dance #22732

merged 1 commit into from
Apr 4, 2022

Conversation

jklymak
Copy link
Member

@jklymak jklymak commented Mar 30, 2022 •
edited
Loading

PR Summary

Closes #22673

tight_layout and constrained_layout were both calling backend_bases._get_renderer. However, that was probably too low-level, and seems to cause canvas-dpi related flakiness as in #22673, in particular some users were seeing changed dpi because savefig.format was set to a vector backend.

This uses the more robust algorithm in _tight_layout.get_renderer. In particular if fig.canvas already exists and has a renderer, use that because it means the user has already (perhaps interactively) created a canvas. This stops prematurely falling back to a default canvas indicated by savefig.format. This probably mostly worked for folks because the default format is png which is enough like on-screen display that things didn't break.

PR Checklist

Tests and Styling

  • Has pytest style unit tests (and pytest passes).
  • Is Flake 8 compliant (install flake8-docstrings and run flake8 --docstring-convention=all).

Documentation

  • New features are documented, with examples if plot related.
  • New features have an entry in doc/users/next_whats_new/ (follow instructions in README.rst there).
  • API changes documented in doc/api/next_api_changes/ (follow instructions in README.rst there).
  • Documentation is sphinx and numpydoc compliant (the docs should build without error).

@jklymak jklymak mentioned this pull request Mar 30, 2022
Closed
@jklymak jklymak added this to the v3.5.2 milestone Mar 30, 2022
@jklymak jklymak added topic: geometry manager LayoutEngine, Constrained layout, Tight layout Release critical For bugs that make the library unusable (segfaults, incorrect plots, etc) and major regressions. labels Mar 30, 2022
@jklymak
Copy link
Member Author

jklymak commented Mar 30, 2022

BTW, I have no real way to test this, other than the existing tests, which pass....

@jklymak jklymak marked this pull request as ready for review March 30, 2022 15:34
@anntzer
Copy link
Contributor

anntzer commented Apr 4, 2022

I agree that this is better; likely this "use the cached renderer" logic should always be used even in backend_bases._get_renderer as long as print_method is None (i.e. we're not in a context of savefig with a known format)? That can be fixed lated, though.

@tacaswell tacaswell linked an issue Apr 4, 2022 that may be closed by this pull request
[Bug]: EPS savefig messed up by 'figure.autolayout' rcParam on 3.5.0 #21930
Closed
tacaswell
tacaswell approved these changes Apr 4, 2022
View reviewed changes
Copy link
Member

@tacaswell tacaswell left a comment

Choose a reason for hiding this comment

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

Seems a bit odd that everything is routed through tight_layout, but this closes at least 2 issues.

@tacaswell tacaswell merged commit 9e0eb54 into matplotlib:main Apr 4, 2022
@lumberbot-app
Copy link

lumberbot-app bot commented Apr 4, 2022

Owee, I'm MrMeeseeks, Look at me.

There seem to be a conflict, please backport manually. Here are approximate instructions:

  1. Checkout backport branch and update it.
git checkout v3.5.x
git pull
  1. Cherry pick the first parent branch of the this PR on top of the older branch:
git cherry-pick -x -m1 9e0eb54fd98bb804f51a74101db88164870a2458
  1. You will likely have some merge/cherry-pick conflict here, fix them and commit:
git commit -am 'Backport PR #22732: FIX: maybe improve renderer dance'
  1. Push to a named branch:
git push YOURFORK v3.5.x:auto-backport-of-pr-22732-on-v3.5.x
  1. Create a PR against branch v3.5.x, I would have named this PR:

"Backport PR #22732 on branch v3.5.x (FIX: maybe improve renderer dance)"

And apply the correct labels and milestones.

Congratulations — you did some good work! Hopefully your backport PR will be tested by the continuous integration and merged soon!

Remember to remove the Still Needs Manual Backport label once the PR gets merged.

If these instructions are inaccurate, feel free to suggest an improvement.

@anntzer
Copy link
Contributor

anntzer commented Apr 4, 2022

Seems a bit odd that everything is routed through tight_layout

As noted above, probably that logic should be pulled out of TL and moved into backend_bases.

@tacaswell
Copy link
Member

This backport is odd because a bunch of stuff has moved around.....I'm doing it and will make sure the two linked issues still work.

tacaswell added a commit to tacaswell/matplotlib that referenced this pull request Apr 4, 2022
@tacaswell
Backport PR matplotlib#22732: FIX: maybe improve renderer dance
c8c3767 
Merge pull request matplotlib#22732 from jklymak/fix-renderer-dance

FIX: improve cached renderer dance
(cherry picked from commit 9e0eb54)

The changes had to be moved to figure.py due to a major refactoring that will
be released in 3.6 that is already on main.
@tacaswell tacaswell mentioned this pull request Apr 4, 2022
Merged
@jklymak jklymak deleted the fix-renderer-dance branch April 5, 2022 07:22
@jklymak
Copy link
Member Author

jklymak commented Apr 5, 2022

As noted above, probably that logic should be pulled out of TL and moved into backend_bases.

#22745 Moves the logic to Figure. The Figure holds the cache, so I'm not clear what we gain from putting the logic in the backends?

tacaswell added a commit that referenced this pull request Apr 5, 2022
@tacaswell
Merge pull request #22783 from tacaswell/auto-backport-of-pr-22732-on…
0734a70 
…-v3.5.x

Backport PR #22732: FIX: maybe improve renderer dance
@anntzer
Copy link
Contributor

anntzer commented Apr 5, 2022

It doesn't really matter to me whether figure or b_bases hold that logic (well, in my view figure is more just for artists and b_bases for canvas/renderer related stuff, but that's no big deal), mostly the point is that having that in t_l doesn't make much sense.

@tacaswell tacaswell mentioned this pull request Apr 13, 2022
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@tacaswell tacaswell tacaswell approved these changes

@anntzer anntzer anntzer approved these changes

Assignees
No one assigned
Labels
Release critical For bugs that make the library unusable (segfaults, incorrect plots, etc) and major regressions. topic: geometry manager LayoutEngine, Constrained layout, Tight layout
Projects
None yet
Milestone
v3.5.2
Development

Successfully merging this pull request may close these issues.

[Bug]: tight_layout (version 3.5+) [Bug]: EPS savefig messed up by 'figure.autolayout' rcParam on 3.5.0
4 participants
@jklymak @anntzer @tacaswell @QuLogic