CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

On 05/07/2012 06:01 PM, antoine wrote:
> Hello,
> I'm a new CGAL user, and am especially interested in Nef_Polyhedrons_2 :
>
> I'm actually using an Apollonius graph, and convert it into a nef polyhedron
> (N1). This seems to be working fine.
> Then, I need to generate another nef polyhedron (N2), based on a given
> skeleton and its width. The generation of this one seems to be working fine
> too.
> Both of them are well imported inside CGAL's Nef_Polyhedron_2 demo.
> However, when I try N1.join(N2) (or any other operation), the program
> crashes ("xxx.exe stopped working" and windows trying to find a solution).
>
> I have tried using different types :
>
>
> 1: typedef CGAL::Gmpz
> CGAL_RT;
>      typedef CGAL::Filtered_extended_homogeneous<CGAL_RT>
> CGAL_ExtendedKernel;
> 2: typedef CGAL::Filtered_extended_homogeneous<int>
> CGAL_ExtendedKernel;
> 3: typedef CGAL::Exact_predicates_inexact_constructions_kernel
> CGAL_Kernel ;
>      typedef Extended_cartesian<CGAL_Kernel::FT>
> CGAL_ExtendedKernel;

The Kernel 1 should be fine.
The Kernel 2 is dangerous as you can get overflow.
The Kernel 3 is not correct, it has inexact constructions.

I tried opening the file provided and get assertion on reading.
Can you provide a minimal example program showing the problem?
(reading + union + crash)

Sebastien.

> with :
> typedef CGAL::Nef_polyhedron_2<CGAL_ExtendedKernel>
> CGAL_Polyhedron;
>
> The first one crashes as soon as I create a new
> CGAL_Polyhedron(CGAL_Polyhedron::EMPTY).
> The second and third ones crash during the join operation (based on what I
> read here, it may be due to a precision issue).
>
> I have no idea on why the program crashes with gmpz stuff, and would really
> apreciate if someone could help me on that, or to make the other methods
> work.
>
> I've enclosed the resulting files too, so if there is something wrong with
> them, please tell me (The first one is the result of the export of the
> Apollonius graph, I.E. 7 regions. The other one is the result of the
> skeleton "enlargement" : Basically, the contours of a set of roads). They
> should both be openable (and usable) with the Nef_Polyhedron_2 demo
> (http://www.cgal.org/demo/4.0/nef_2.zip).
>
> Thanks in advance,
> Antoine.
>
> http://cgal-discuss.949826.n4.nabble.com/file/n4615185/ApoloniusPolyhedron.cgal
> ApoloniusPolyhedron.cgal
> http://cgal-discuss.949826.n4.nabble.com/file/n4615185/SkeletonPolyhedron.cgal
> SkeletonPolyhedron.cgal
>
> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Crash-with-Nef-Polyhedrons-tp4615185.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.