Repository navigation
Factor out handling of missing spines in alignment calculations. - #28313
timhoffm merged 1 commit into
Conversation
Sorry, something went wrong.
This comment was marked as outdated.
timhoffm
left a comment
There was a problem hiding this comment.
Take or leave my comments. You can self-merge.
Sorry, something went wrong.
| # that have been set by `fig.align_xlabels()` | ||
| bboxes, bboxes2 = self._get_tick_boxes_siblings(renderer=renderer) | ||
|
|
||
| ax = self.axes |
There was a problem hiding this comment.
This is only extracted to make l.2409 / l.2414 shorter. I would not do that because it separates definition and usage and thereby results in more state and thus cognitive load.
I slightly prefer to either leave this in place self.axes.spines.get("bottom", self.axes) - or if that's reads too cumbersome, define it right before usage in each if branch.
Sorry, something went wrong.
There was a problem hiding this comment.
I think the repetition of self.axes reads a bit cumbersome here, but sure, I put it back. Ditto below.
Sorry, something went wrong.
| # get bounding boxes for this axis and any siblings | ||
| # that have been set by `fig.align_ylabels()` | ||
| bboxes, bboxes2 = self._get_tick_boxes_siblings(renderer=renderer) | ||
| ax = self.axes |
There was a problem hiding this comment.
As above.
Sorry, something went wrong.
... by using spines.get. Note that for an Axes, get_window_extent (with or without the renderer argument) is always equal to ax.bbox; the refactoring also makes that symmetry clearer.
... by using spines.get.
Note that for an Axes, get_window_extent (with or without the renderer argument) is always equal to ax.bbox; the refactoring also makes that symmetry clearer.
Noted while looking at #28300.
PR summary
PR checklist