I'm not sure this is the correct fix? The renderer for the subfigure should be the same as the parent's renderer. I think we need to set this up as a property and have the getter go up the tree to the parent?

I'm not sure this is the correct fix?

No idea. What you says make sense though, so I'll give it a go.

On draw SubFigure redefines the _cachedRenderer. Will there ever be cases where it differs from the parent renderer?

Updated with properties. Not convinced that one need to set the renderer in draw, but guess it shouldn't hurt?

The subfigure really should not have a cached renderer, but subfigures have been abstracted to be like figures so that child artists don't know if they are in a figure or a subfigure. So if an artist does self.figure._cachedRenderer that should always point to the top-most renderer.

I think this is correct. Note that #22745 would have fixed this as well by calling figure._get_renderer().

Note that #22745 would have fixed this as well by calling figure._get_renderer().

Ahh, OK! Maybe better to try to get that in then? I have not really followed (nor fully understand) that discussion.

I guess that the test added here very well can be worthwhile keeping as it seems to exercise the code in a way that once was (currently is) broken.

I think we can have both fixes. #22745 isn't necessarily going to go in ;-)

I suspect this is an argument as to why we actually do need to cache the renderer (and should move it to a central canonical place) and my position that we should not cache it at all is wrong (or at least currently infeasible).

I restarted the failed windows job (it was a blit test in a subprocess I have not seen fail before 🤞🏻 it is transient).

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

2. Cherry pick the first parent branch of the this PR on top of the older branch:

git cherry-pick -x -m1 af835787ac405d6021d4ed11e57734873c401b9c

3. You will likely have some merge/cherry-pick conflict here, fix them and commit:

git commit -am 'Backport PR #23174: Make sure SubFigure has _cachedRenderer'

4. Push to a named branch:

git push YOURFORK v3.5.x:auto-backport-of-pr-23174-on-v3.5.x

5. Create a PR against branch v3.5.x, I would have named this PR:

"Backport PR #23174 on branch v3.5.x (Make sure SubFigure has _cachedRenderer)"

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.