Nothing super major to report, I am just curious about this point:
On 10/20/2018 at 11:45 AM, Dirk Gregorius said:GJK computes the closest points between two *disjoint* convex shapes.
Given that you felt the need to bold this sentence and asterisk the word disjoint, am I to infer that the distance version of this algorithm is not reliable for determining intersections as well? I was messing around and noticing that indeed it appears in certain circumstances to report a distance ever so slightly greater than zero while intersecting. I assume this is likely due to floating point error, but it would make me feel more comfortable knowing that for sure. If it is the case that it isn't reliable for that, I can just use the bastardized version for checking that prior to a full-on distance check. I just want to be sure I'm not overlooking something.