…f intersection is at an endpoint, though.

…n boolean values. Lots of assumptions, e.g. at most one intersection point. Known bug: problem if arc intersects segment at arc endpoint. Why is u=nan?

…ines.

…ontours are drawn. Need to fix subsequent contours though.

…tour intersection now checks roughly before doing too many calculations.

…ete yet!

…still visible, and coinciding contours seems to fail.

… with cedilla case.

…ture, made a vertex for each contour.

…n be used to add dotted edges soon.

…ch other.

…e contour relationship graph.

…t contains other contours passing through it.

…ter merge, though. Requires bipartite graphs.

…ral code cleanup.

… not correct for line segments, and does not use this stored value.

…phy_sdf, but still not perfect at corners.

… compartmentalized in glyphy_sdf somewhat, but a seaming bug is back, and vector sometimes points backwards.

…-sdf. Still have seams. :-(

…hen scaling.