CGAL users discussion list

[Sylvain Pion <>]

Mathieu Brédif wrote:
> Hello,
>
> Has anyone been able to reproduce my problem?
> if I understood correctly, the kernel I use is exact :
>
> typedef CGAL::Gmpz RT;
> typedef CGAL::Homogeneous_d<RT> Kernel;
> typedef CGAL::Delaunay_d<Kernel> Delaunay_d;
>
> Thus the inifinite recursion looks like a bug or a prerequisite that I
> did not checked. For instance, does the "Delaunay_d" class handle
> degenerate inputs such as aligned, coplanar or cospherical points?

AFAIK, it's supposed to do so, yes.

> Also, apart from this problem, and using the current implementation,
> what are reasonable numbers of input points in 4D delaunay
> triangulation?

The current Delaunay_d is not the most efficient that can be.
We (mostly Samuel Hornus) are working on a new implementation
for d-dimensional triangulations which is more efficient.
You can check his paper at SoCG this year for benchmark numbers.

Hopefully the new code will be part of CGAL-3.6.

--
Sylvain Pion
INRIA Sophia-Antipolis
Geometrica Project-Team
CGAL, http://cgal.org/