CGAL users discussion list

[Benoît Presles <>]

Hello everyone,

I tried the program on windows XP (32 bits) and I get the same problem: 
I have got 2 connected components instead of 1.

Do I do something wrong ?
Is there a workaround to this problem ?

Thank you for your help,
Benoît


Le 02/08/2010 16:40, Benoît Presles a écrit :
> Dear Sebastien,
>
> I have just tried what you suggested me and I can still observe the
> problem.
> The program is the same as the one I sent you, I only changed the
> include "#include
> <CGAL/Exact_predicates_inexact_constructions_kernel.h>" to "#include
> <CGAL/Exact_predicates_exact_constructions_kernel.h>" and "typedef
> CGAL::Exact_predicates_inexact_constructions_kernel K" to "typedef
> CGAL::Exact_predicates_exact_constructions_kernel K".
>
> Thank you for your help,
> Best Regards,
> Benoît
>
>
> Le 30/07/2010 14:54, Sebastien Loriot (GeometryFactory) a écrit :
> > Can you please try with Exact_predicates_exact_constructions_kernel and
> > tell us whether you still observe the problem?
> >
> > (this may be an floating point arithmetic problem)
> >
> > Thanks.
> >
> > S.
> >
> > Benoît Presles wrote:
> > > Hello Everybody,
> > >
> > > I use the latest version of CGAL and I think a bug could be present in
> > > the function "find_optimal_alpha" of the class "Alpha_shape_2".
> > > - First, I use the function "find_optimal_alpha" to find the best
> > > alpha to get 1 connected component.
> > > - Second, I set this alpha.
> > > - Third, I use the function "number_of_solid_components" to get the
> > > number of solid components of A.
> > >
> > > Sometimes, the program gives me 2 connected components instead of 1.
> > >
> > > Attached, you can fine the c++ code (basically the same as in the
> > > example) and a list of points for which the program doesn't work as
> > > expected.
> > >
> > >
> > > Thank you for your help,
> > > Best Regards,
> > > Ben_P