rileychall
commented
2 years ago This appears to also be relevant for cases at very small scales, where rounding or floating point precision can accidentally exclude points at the edges. Adding a small buffer to the calculated bounding box should suffice.
The bounding box and related functionality, such as clipping extents, does not correctly handle cases where the minimum and maximum coordinates in an axis are the same. e.g. A 2D problem in the xy plane has z coordinates that span (0.0, 0.0).