I believethis pull request is what caused this regression.

Would love to see this simple patch backported tov2.2.x as well 👍

@charlesruan@pbregener is this fix live? If so, how can I take advantage of the changes?

This MR is not merged, yet, so it's not "live" in the sense that it is part of any release or even current master. I guess the 5 commits should be squashed into one with a sensible commit message and then some maintainer should review and eventually merge it.

There is also this regression on polar plots:#10882
Are they related?

PRs need two reviews. This looks reasonable, but it would be nice to know what PR the regression came in from and whether this was just an oversight or a design decision.

@jklymak This goes back to#9068 and thereforev2.1.0. Also see the original issue for this:#10105

Thanks. ping@QuLogic who is the local polar expert.

If this really is a regression, it should probably be milestoned for v2.2.3

Its a 2.1 regression. Not sure on the policy there.

It is a regression and as mentioned above, I would love to see this inv2.2.x. Should be really easy to backport, too.

Not sure about how matplotlib usually handles this, but I guess it would be nice to squash the five commits for this fairly trivial change into one@charlesruan

@tacaswell said:

I propose the following criteria for backporting PRs to the 2.2.x branch. We always will backport

• critical bug fixes (segfault, failure to import, things that the user can not work around)
• fixes for regressions against 2.0 or 2.1

This would seem to fit.

Commits have been squashed

ping@jklymak and especially@QuLogic for an update on this PR?

@charlesruan thanks so much for this PR and your patience with the process!