Hans Moritz Günther

There seems to be broad agreement by reviewers that this is the right direction to take except for the small comments by @mhvk. @MridulS : Do you have time to...

For the first point and second point, I suggest to leave everything at the backend default (cycling for matplotlib) and implement #251. If I don't want to cycle, I can...

Given that I extensively use `fake_pha` for Arcus and Lynx work, I have some thoughts on what @DougBurke lists as "to think about", but I'll get to that after #1734...

@DougBurke Is this ready for final review to make it into the next release or you planning any more changes? I'm sorry if you've just been waiting for me, I...

It's going to help that I generally review "per PR" and not "per commit", so I might skip some intermediate steps.

I'll leave it to @wmclaugh to decide if a rebase is needed to remove the merge commit. Most projects don't like merge commits like this one in their history, but...

I’m opposed to the “comply with third-party licensing” part suggested in https://github.com/pyOpenSci/software-peer-review/issues/201#issuecomment-2822458607. A common case I struggle with is when a plug-in is considered a “single combined program” under the...

That's better. Still, can we just go with "are clearly listed"? It's hard enough to find reviewers and anything that sounds as if we require them to be experts in...

Effectively, what you are doing here is to define a protocol. `Callback` is one function that implements that protocol, `InfinitePotential` is the other. They are not derived in a hierarchy,...

> Maybe we should add bokeh to at least one of our conda everything-and-the-kitchen-sink tests (but not the macOS one)? Sure, why not? It depends a bit on how much...