CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

What you sent confirm my first guess, you have a degenerate triangle.
It is the only way you can have a plane with (0,0,0) normal vector.

Could you make a check on the polyhedron before building the AABB-tree?



for (Polyhedron::Facet_iterator fit=polyhedron.facets_begin();
                                fit!=polyhedron.facets_end();
                                ++fit)
{
  if ( CGAL::orientation(
         fit->halfedge()->vertex()->point(),
         fit->halfedge()->next()->vertex()->point(),
         fit->halfedge()->opposite()->vertex()->point() )
       == CGAL::COLLINEAR )
  { std::cout << "degenerate triangle\n"; }
}

Sebastien.

On 06/06/2013 10:35 PM, schrodingersnewcat wrote:
> The four stacks are from four process running independently on four different
> nodes.  They have no shared memory.  They don't even pass messages at this
> point in the program.
>
> Here is a backtrace from a run where I have removed MPI.  single.trace
> <http://cgal-discuss.949826.n4.nabble.com/file/n4657660/single.trace>
>
> Sorry I didn't get back sooner, I didn't see your response until today.
>
>
>
> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Divide-by-Zero-in-AABB-tree-tp4657026p4657660.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.