CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

On 02/28/2012 12:00 PM, Zohar wrote:
> Another emphasis: The alternative that cgal offers is to use exact_exact
> kernel. Intersecting a few thousands of triangles, which takes less than a
> second with double, takes more than a minute with an exact kernel. I don't
> know Sebastien, it doesn't sound to me a question of robustness anymore.
I suggest to read the references given in the FAQ about inexact NT,
and in particular this one:
http://www.mpi-inf.mpg.de/departments/d1/ClassroomExamples/

As Mark suggested, if you have a small amount of triangles intersecting, 
call do_intersect with Exact_predicates_inexact_constructions_kernel and 
only call the intersect function with 
Exact_predicates_exact_constructions_kernel.
This strategy is working well and will make your code robust.

The Cartesian_converter is your friend for that:
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Kernel_23_ref/Class_Cartesian_converter.html

Sebastien.


> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Intersecting-with-Cartesian-kerenl-tp4425392p4427756.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.