Negative winding broken

[seandepagnier]

I use the glu tesselator for computing intersections and subtractions of regions. When I switched to tess2 it does not work for subtraction anymore and apparently the negative winding does not work anymore. Maybe this was an "optimization" for the case of triangles but not correct for contours.

Also, it seems to be slower than the original glu for simple polygons without overlapping contours or inner contours.

[memononen]

Do you have a repro case for the failing subtraction? It should work.

Also, do you have a specific case for the simple polygon that is slower, and how much it is slower?

[seandepagnier]

This example shows the difference from libtess and libtess2 for the same operation. It is simply a single negative contour!

I will work on an example where it is slower.

tesscomparison.tar.xz.zip