I'm having trouble telling where and what the change is supposed to be. Can you give more detail, pref with a link to https://circleci.com/gh/matplotlib/matplotlib/10434/artifacts/0//home/circleci/project/doc/build/html/index.html

I don't think that it is rendered in the docs at the moment but it probably should be in for example axes._base._AxesBase.__init__. (which will be the docstring in axes.Axes with my other PR) or in for example Figure.add_subplot.

The difference is that artist.kwdoc(matplotlib.axes.Axes) with rcParams['docstring.hardcopy]=True now renders to lines like
':meth:`yticklabels <matplotlib.axes.Axes.set_yticklabels>` List[str]'
which results in correct links to the documentation compared to the lines before:
':meth:`yticklabels <matplotlib.axes._base._AxesBase.set_yticklabels>` List[str]'
which don't link to a place in the documentation.

Some test code

import matplotlib
import matplotlib.artist as artist
import matplotlib.axes as axes

matplotlib.rcParams['docstring.hardcopy']=True
st=artist.kwdoc(axes.Axes)
print(st)

matplotlib.rcParams['docstring.hardcopy']=False             

I added some docstrings

Many links in the kwargs list in don't work. This PR fix those links.

I see now that the kwarg list was actually included in Figure.gca, new, old.

Could someone review this? This pr make the links in the kwarg lists in for example work.

I've checked against master and did not see any change in the links. Which ones exactly do not work on master?

axes

For example adjustable in the kwarglist should be a link and this pr fix that.

The same kwarglist with this pr
axes

But what has happend with the rendering of the types in the parameter list? Why are they bold?

The bold parameter list is a regression from #11495. I'll detail what's going on and fix it in a separate PR. You don't have to worry about that here.

Thanks @fredrik-1 !

(quick note--@timhoffm I don't know if you are aware that for documentation pull request, only one core dev is required to approve the PR before merging)

@NelleV I think that’s true for bare edits of the docs. This was an actual code change for how the docs are handled.

Yeah, good point.