DOC: setting active axes position is ineffective#31314
DOC: setting active axes position is ineffective#31314rcomer wants to merge 1 commit intomatplotlib:mainfrom
Conversation
| For native Matplotlib `.Axes`, the active position is | ||
| determined by a combination of the original position and the | ||
| aspect ratio. Setting the active position directly is | ||
| therefore not effective. This option is retained as it may be | ||
| relevant for some third party `.Axes` subclasses. |
There was a problem hiding this comment.
Some similarity to https://matplotlib.org/stable/api/_as_gen/matplotlib.axes.Axes.apply_aspect.html#matplotlib.axes.Axes.apply_aspect and more specific explanation.
| For native Matplotlib `.Axes`, the active position is | |
| determined by a combination of the original position and the | |
| aspect ratio. Setting the active position directly is | |
| therefore not effective. This option is retained as it may be | |
| relevant for some third party `.Axes` subclasses. | |
| This parameter is considered internal. End users should not use it. | |
| For native Matplotlib `.Axes`, the active position is | |
| determined by a combination of the original position and the | |
| aspect ratio. Setting the active position only is only meaningful | |
| in the context of correct internal state, which is managed | |
| internally inside `.Axes`. This option is retained as it may be | |
| relevant for some third party `.Axes` subclasses. |
There was a problem hiding this comment.
Setting the active position only is only meaningful in the context of correct internal state, which is managed internally inside
.Axes.
I'm sorry but I'm not following this sentence. I'm happy with the rest.
There was a problem hiding this comment.
Agreed it's a bit convoluted. What I want to say is that position="active" requires does have an effect - as opposed to being ineffective. But you cannot set arbitrary values with position="active", they must be consistent with the internal state (specifically the position should be the one derived from the original position and aspect calculation, otherwise surprising things like other Axes overlapping or change through a updated aspect calculation may happen.) Simpler wording welcome.
There was a problem hiding this comment.
"Any active position set by the user will therefore be overwritten by internal handling"
?
PR summary
Closes #31254. I do not have any evidence that this matters for third party subclasses, but it is a possibility. Since we only pass the parameter down, I think continuing to support it indefinitely is likely less maintenance work than deprecating it.
AI Disclosure
None
PR checklist