Behdad Esfahbod
Behdad Esfahbod
I understand what's happening here: a *reverse* isomorphism is deemed a better match. I tried to only match isomorphisms of the same winding direction, which should fix this. However, I'm...
> I tried to only match isomorphisms of the same winding direction, which should fix this. However, I'm approximating winding direction by the sign of the area of the contour,...
Thanks Laurence. I might have to do something like that I'm afraid. I want to avoid trigonometry as much as I can, but I don't think it's possible in this...
This is way harder than I was hoping. I have a fix for this particular font, but breaks a lot of other fonts. Still investigating.
Just thinking out loud, this is what I think so far: - Say we're dealing with a HOI-style avar2. That is, there's `wght` and `wgt2`, the latter being completely driven...
@davelab6 We really need those avar2 test fonts for various use-cases...
cc @anthrotype
@justvanrossum
This kinda seems to work now, even though it's not complete. I need testers like @lorp
#1537 is a weird issue. It talks about `varLib.mutator` (not `varLib.instancer`); but then talks about partial instancing that `varLib.mutator` doesn't support. So I'm confused.