MischaMegens2

Ready for review again.

Anything else needed for a merge? (Does that python 3.9 on macos-12 need to be revived?) (Extra 0.13% code coverage?) (More reviewers?)

Added a test to improve coverage, and rebased. Ready for merge (I don't think the remaining failing checks have anything to do with the PR).

ping @scottshambaugh Could you take a look at it, or ask someone else for additional review (timhoffm?)

Ping @timhoffm Dear Tim, would you be able to take a look and provide a second review?

@timhoffm: Do I understand correctly then that - the `view_init()` suggestion is fine in your opinion, but - the changing of the order in the examples you think is a...

> The _only_ thing we should do, is change our internal examples to use kwargs, to make the users aware of the meaning and order of the parameters. Here, I...

(The remaining CircleCI warnings are: ``` /home/circleci/project/lib/mpl_toolkits/mplot3d/axes3d.py:docstring of mpl_toolkits.mplot3d.axes3d.Axes3D.view_init:23: WARNING: py:obj reference target not found: azim, elev /home/circleci/project/doc/users/next_whats_new/view_angles_keyword_arguments.rst:4: WARNING: py:obj reference target not found: elev, azim ``` Not quite sure...

@timhoffm: Should be good to go, please take a look. (I left the 'Resolve conversation' up to you.)

> I argee with @scottshambaugh. Let's prioritize internal consistency within Matplotlib. How about: I revise it to just clarifying the situation in the documentation?