-
-
Notifications
You must be signed in to change notification settings - Fork 7.9k
Fix axis inversion with loglocator and logitlocator. #14624
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.
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
Conversation
177164d
to
eec8dba
Compare
release critical: a similar fix already made it to 3.1 in #14623. |
eec8dba
to
df9c152
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Technically correct. The argument for making nonsingular
order the values sounds reasonable, but I don't have a complete overview of the consequences.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
modulo CI passing on rebase.
Anyone can merge on clean CI
df9c152
to
0e94bb9
Compare
0e94bb9
to
9bc8097
Compare
In matplotlib#14624 we reverted changing the order of the limits to match the input order. Doing the same here for consistency.
In matplotlib#14624 we reverted changing the order of the limits to match the input order. Doing the same here for consistency. Sort the limits before comparing. We want to make sure that it works with the values reversed, but do not expect non-singular to preserve the order.
PR Summary
Reverts a minor part of #13593 to provide a different fix to #14615 and #14162
I chose to make nonsingular always order its arguments (and thus just change the behavior in the base class, which was only introduced in 3.1, rather than change the behavior of the subclass, which was only introduced in #13593 and not released yet) as that seems more likely to be respected by third-party implementations. On the other hand in #14643 I realized we can't really trust (for now) whether third-party implementations (or our own) do this ordering, so redoing a sort after-the-fact is likely the most robust solution...
The 3.1.x version is at #14623; this PR should not be backported.
PR Checklist