I think you're right. I did an analysis of the Grid3D version with ANY, and I think I found a bug or two there too.
I have committed revised versions in AbstractGrid2D and AbstractGrid3D (and copies in SparseGrid2D and SparseGrid3D). Please try stomping on both of them and see if you can find any more errors -- particularly in AbstractGrid3D, which is more complex.
On Sep 17, 2015, at 4:27 AM, Christian Meyer <[log in to unmask]> wrote:
> Thanks for the update. I think I found a bug now. A lookup with
> Grid2D.ANY should return at least the same locations as Grid2D.CENTER,
> as far as I understand. But this does not work if a value smaller
> sqrt(0.5) is passed as dist. ANY will return 0 locations then, but
> CENTER returns 1 (the origin), which is correct and should be reported
> by ANY as well.