Tim Hoffmann

The only thing I would keep is „Inserting matplotlib figures“. That’s specific to us. Everything else is standard Sphinx and I don’t see a reason why we should bother to...

> On balance, I am still leaning towards `contains` should not check visibility and leave that decision/check to the caller. Agreed.

To me the question is: What is the best way forward? Should we make '.' unfilled, or just change the docs and move '.' to the filled markers in the...

> Presumably there was a reason to have two distinct circular markers at some point The existence of the circle *and* point marker is surely motivated by [MATLAB having both](https://de.mathworks.com/help/matlab/creating_plots/specify-line-and-marker-appearance-in-plots.html)....

> Is this just not a case of the (relatively new) Marker reference being incorrect, and "point" should have been in the filled marker group? Agreed. We should declare this...

The reduced flexibility would indeed be an issue and to be discussed. I think the quick and simple fix here is to propagate the kwargs. An IndicateInsetConnector can be a...

Yes. The task here is to make sure that suitable keyword arguments like `edgecolor`are applied to the rectangle and the connection patch.

https://github.com/matplotlib/matplotlib/issues/23424#issuecomment-1185529156 is exactly the place to look at.

Simply passing all **kwargs will not do, you have to carefully think through which parameters to pass: - There may be kwargs accepted by `Rectangle`, that are not valid for...

Depending on which one is easier to specify you can go with - blacklisting: ``patch_kwargs = {key: val for key, val in kwargs.items() if key not in blacklist}`` - whitelisting:...