Resolve 'text ignores rotational part of transformation' (#698)

[immaxchen]

PR Summary

Considering the text rotation don't get transformed may be a desired behavior
(such as text for annotation to some data point). Here made it as an option to
control whether or not the text rotation get transformed.

• original behavior described in Text ignores rotational part of transformation #698:

text = plt.text(0, 0, 'hello', size=50, transform=R + ax.transData)

• behavior in this patch:

text = plt.text(0, 0, 'hello', size=50, transform=R + ax.transData, 
                rotation_mode='anchor', rotation_transformed=True)

PR Checklist

• Has Pytest style unit tests
• Code is Flake 8 compliant
• New features are documented, with examples if plot related
• Documentation is sphinx and numpydoc compliant
• Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
• Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

[anntzer]

Do we really want to expose a control knob as to whether rotation affects text? It would seem(??) like we generally do want it to affect text. Perhaps just have that for a transition period and warn if the transform has nonzero rotation but that rotation was getting ignored and won't be ignored after the transition?
(Not really sure of the best strategy here.)

[timhoffm]

Do we really want to expose a control knob as to whether rotation affects text? It would seem(??) like we generally do want it to affect text.

Not sure how it works internally, but wouldn't that rotate tick labels on a polar plot?

[anntzer]

Ah, fair point.

[anntzer]

Does this need a whatsnew note?

[immaxchen]

Sure, no problem.

[QuLogic]

This needs a rebase to resolve conflicts.

[tacaswell]

@immaxchen I took the liberty of rebasing and force-pushing for you, I hope you do not mind.

[immaxchen]

sure & thank you for the help.

[anntzer]

To what extent would it be possible to make this also improve the situation with #17005?

[immaxchen]

#17005 can be easily fixed this way, not sure if you like it though.

[tacaswell]

This seems like it has type instability?

[immaxchen]

I fixed the inconsistent type, but don't know why CI keeps failing.. i think i didn't do anything about docs in the new commit
(edit) it turns out that a rebase is needed to pass CI ..

[immaxchen]

@tacaswell resolved, please have a look

[dopplershift]

I'm not sure why Travis didn't report here, but it's green.