I've pushed two commits for openvoronoi - one for the vertex positioner, where the equidistance
constraint should help in situations where the t-filter is not able to discard solutions on the wrong side of s3.
This happens in case the t-value does not vary sufficiently (e.g. due to S1/S2 being (nearly) parallel edges).
The infinite loop was something I observed after doing a cross-compile using mingw -
float behaviour can be quite esoteric at times :-) See the accompanying comment.
Hi Anders,
I've pushed two commits for openvoronoi - one for the vertex positioner, where the equidistance constraint should help in situations where the t-filter is not able to discard solutions on the wrong side of s3. This happens in case the t-value does not vary sufficiently (e.g. due to S1/S2 being (nearly) parallel edges).
The infinite loop was something I observed after doing a cross-compile using mingw - float behaviour can be quite esoteric at times :-) See the accompanying comment.