cc: @charris @ndawe @tpoole
I finally got some time to work on this. This might not be that important especially with the GSoC going on, but it might be good to have.

bump?

☔ The latest upstream changes (presumably #4619) made this pull request unmergeable. Please resolve the merge conflicts.

What is the status of this PR?

What is the status of this PR?

hmm, this never got followed up on.

The two types of weights are weird imho. Also in np.cov they're weird, explanation isn't clear and no example. It doesn't look too good. I'd be in favor of closing this, or otherwise just using normal weights.

@rgommers IIRC, there was quite a long discussion about the two weights.

maybe we can find it and document it properly then? it doesn't make too much sense right now imho .....

I can't find it with a quick search, but there was this, that tickles my memory.

The previous discussion can be found here: #4960

So was this abandoned since there is something missing in the PR or it just can be re-based and merged?

Considering the age and the fact that it requires quite a bit of rebasing: I am going to close the PR. I will add a note to gh-8581 that the PR is stale but could be a basis for bringing this up again.

Thanks for opening the initial PR!