CGAL users discussion list

[Bernhard Kornberger <>]

Am 2011-05-13 09:36, schrieb Marc Glisse:
> On Thu, 12 May 2011, Marc Glisse wrote:
>
> > Indeed, I can't see anything in Lazy.h to interpret the
> > std::vector<Approximate_kernel::Point_3> returned by the approximate
> > kernel to output a std::vector<Lazy_kernel::Point_3> (both wrapped in
> > an Object). And we had better hope that the order of the points in the
> > vector will be the same with the exact kernel (but then I guess the
> > same applies to segments and triangles and is safe enough as long as
> > they use the same deterministic code).
>
> Compiling with -DCGAL_DONT_USE_LAZY_KERNEL would work around this, at
> the price of a slowdown.
>
> The attached patch partially fixes things when the intersection type is
> vector<Point_2> (replace Point_2 by Point_3 in the patch for your case)
> but it forces an exact evaluation in this case because I was too lazy to
> finish it properly.


Thank you for your work! I go on holliday tomorrow but I will
try that as soon as possible.

Best
Bernhard