Andrew Hunter

Yes, but it's tricky and I've avoided commenting because I have a really strong reaction to what rustfmt does: I really hate it, but that's no good as the basis...

It's definitely an intended use for coordinates to represent vectors: they're used for normals and tangents already and have things like the dot product defined for them. (It's also definitely...

Ah, you probably want to use `path_remove_overlapped_points()`, which doesn't have an entry in `PathCombine`, so I probably need to add one. 'Remove interior points' in this case classifies any point...

Hmm: this looks more like a bug when detecting where two curves intersect, or possibly incorrect edge ordering with overlapping curves. The filled-in 'g' in your example looks like it...

Thanks. A bit of editing and I get this as a representation in `flo_curves` own datatypes (for future reference): ```Rust let paths = vec![ BezierPathBuilder::::start(Coord2(-14.354116, 66.98954)) .curve_to( (Coord2(-8.825491, 59.061977), Coord2(2.0829022,...

I've put this demo code on a branch: https://github.com/Logicalshift/flo_curves/blob/issue-14/demos/remove_interior_points/src/main.rs so you can check my results for yourself. I'm still investigating, but I've found a few things. If I remove the...

I updated the test to use the types from MFEKmath with the same results so that seems to eliminate that as a possible cause.

I've updated https://github.com/Logicalshift/flo_curves/blob/issue-14/demos/remove_interior_points/src/main.rs to use glifparser: still seems to work for me at least - but this does make it easier for me to try other .glif files. I've looked...

I haven't managed to get a reproduction yet, but I have found an issue that could potentially affect the way this glyph is processed: if a ray crosses at a...

I've finished the fix for the issues I've found now, so this is back on the main v0.6 branch. There was another problem that could occur particularly if paths were...