CGAL users discussion list

["Jens K. Becker" <>]

I just repeat a first answer that I got here (since I forgot to change the
subject it is now in a different thread):

Please start a new thread (with a new title), when you want to talk about
something different from the thread subject!

The answer came from Laurent Rineau:

That is a FAQ. You use a kernel with non-exact predicates, and that can lead
to crashes. See http://www.cgal.org/FAQ.html#inexact_NT



I tried different Kernels types (including
Exact_predicates_inexact_constructions_kernel) but the problem remains.

I am currently thinking of mixing things (Homogeneous<Gmpz> for triangulation
related stuff, Cartesian<double> for the rest). Is that a good idea? I dont
want to use Homogeneous<Gmpz> for everything because I need to do a lot of
calculations (sqrt, cos, acos, areas etc.) and there seems to be a general
problem with me and Gmpz, we just don't get along very well....

Jens

Quoting "Jens K. Becker" 
<>:

> Hi all,
>
>
> Sorry for this double posting, I forgot to change the subject...
>
>
> I have a problem with polyhedra. I have a single polyhedra which I 
> put together
> from points from a Delaunay-Triangulation. This leaves me with a 
> polyhedra that
> only has triangles as facets. I need to have a polyhedra where all incident
> facets that are in the same plane are merged. I had a code (see below) that
> does that and it worked nicely using a homogeneous kernel with Gmpz. 
> Now I have
> changed the kernel to be Cartesian<double> and the function does not work
> anymore. It crashes when I try to delete a halfedge (therewith merging two
> facets), the error-message is:
>
> CGAL error: precondition violation!
> Expr: circulator_size(h->opposite()->vertex_begin()) >= size_type(3)
> File: /home/jkbecker/Code/eKolos/src/eKolos/core/aPolyhedron_3.h
> Line: 1191
> Explanation:
> terminate called after throwing an instance of 'CGAL::Precondition_exception'
> what():  CGAL ERROR: precondition violation!
> Expr: circulator_size(h->opposite()->vertex_begin()) >= size_type(3)
> File: /home/jkbecker/Code/eKolos/src/eKolos/core/aPolyhedron_3.h
> Line: 1191
> Aborted
>
> If I run the program a couple of times (using the same polyhedra), it always
> crashes at a different polyhedron, not always at the same one. However, the
> ones it does process are not done correctly. I suspect a problem with 
> the plane
> equation and the number-type and/or kernel-type...
>
> The code first compares the plane equation of twofacets. If it is the 
> same, they
> are in the same plane and therefore should be merged. I then cycle 
> through the
> halfedges of one of the facets to see which one I need to remove to merge the
> facets. Any ideas?
>
> Thanks for your help,
>
> Jens Becker
>
>
> Here is the code of the function. It just gets a pointer to the polyhedron.:
>
> // FIXME: This loops through the polygon a lot, that can probably be 
> done much
> faster. But then who cares, as long as it is only done to get the initial
> structure...
> void TopoCheck::DeleteAllInternals(Polyhedron *p)
> {
>        //do this for all facets of the polyhedron
>        Facet_iterator
> f1=p->facets_begin(),f2=p->facets_begin(),fend=p->facets_end();
>        Halfedge_facet_circulator hfc;
>        bool a;
>        while(f1!=p->facets_end())
>        {
>                while(f2!=p->facets_end())
>                {
>                        a=false;
>                        //if the points of both facets lie in the same plane,
> the facets lie in the
> same plane. In that case, we dan join both facets
>                       
> //if(Plane(f1->halfedge()->vertex()->point(),f1->halfedge()->next()->vertex()->point(),f1->halfedge()->prev()->vertex()->point())==Plane(f2->halfedge()->vertex()->point(),f2->halfedge()->next()->vertex()->point(),f2->halfedge()->prev()->vertex()->point()))
>                        if(f1->plane()==f2->plane())
>                        {
>                                //to join them, I have to find the 
> halfedge that
> separates the two faces.
>                                if(f2->halfedge()->opposite()->facet()==f1)
>                                {
>                                        p->join_facet(f2->halfedge());
>                                        a=true;
>                                }
>                                else
> if(f2->halfedge()->next()->opposite()->facet()==f1)
>                                {
>                                        p->join_facet(f1->halfedge()->next());
>                                        a=true;
>                                }
>                                else
> if(f2->halfedge()->prev()->opposite()->facet()==f1)
>                                {
>                                        p->join_facet(f2->halfedge()->prev());
>                                        a=true;
>                                }
>                        }
>                        if(a)
>                        {
>                                f2=p->facets_begin();
>                                f1=p->facets_begin();
>                        }
>                        else
>                                f2++;
>                }
>                f1++;
>                f2=p->facets_begin();
>        }
> }
>
> --
> Dr. J.K. Becker
> University of Tuebingen - Institute for Geoscience
> Sigwartst. 10 - 72076 Tuebingen (Germany)
> Tel.: ++49 7071 29 73139   Fax: +49 7071 5059
> web: http://www.jkbecker.de
>
> --
> Dr. J.K. Becker
> University of Tuebingen - Institute for Geoscience
> Sigwartst. 10 - 72076 Tuebingen (Germany)
> Tel.: ++49 7071 29 73139   Fax: +49 7071 5059
> web: http://www.jkbecker.de
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss