Andrei Zavatski

Hmm, found discrepancy in behaviour. While `Circle` changes it's shape on `Scale`, `FastCircle` is always preserving the pill shape and I'm not sure which one is more expected to the...

> Hmm, found discrepancy in behaviour Okay, this one is easily fixable. But also: while I can see how this component is too niche to include with the framework and...

> I will once again say that I feel uncomfortable with changes like this where I would hope the overhead of FFC isn't this large. > > Where is the...

Seems like my lower bound isn't low enough. In that case we can increase thickness only for 4k. I'm not sure if I have any other ideas otherwise..

And I'm not against that, but I'm not sure if there's a sane way to get DPI info in this context. Probably should've stated that earlier.

I've stumbled upon a similar (same?) issue with autosized `FillFlowContainer`s and it indeed looks like a precision issue somewhere. In this case parent's autosize depends on children positions and their...

Okay, I don't think I can do any better for now. That said, results are still pretty insane compared to what we had just a few months ago, and it's...

> @EVAST9919 did you happen to benchmark against existing code under similar scenarios? Good idea to compare overall time of arbitrary path creation, will do that

> @EVAST9919 did you happen to benchmark against existing code under similar scenarios? I've updated the pr description to include that.

With more and more changes in https://github.com/ppy/osu-framework/pull/6630 it would be easier to make a new pr instead of fixing all the conflicts, so I'll mark this one as a draft...