CGAL users discussion list

[Peter Hachenberger <>]

Hi Stephen,

I somehow neglected the Extended_cartesian kernel for a while and
therefore it is not compiling at the current release. Usually nobody
uses the extended kernels, but recently somebody wanted to work with the
Extended_cartesian_kernel and I fixed it. So, if you want to use it, I
can send you 2-3 updated files that will fix the problems. 

Before you do that, I want to find out, whether you really need an
extended kernel. They are pretty slow and not often needed. In the
latest release there is an undocumented function that allows
intersection with a halfspace without constructing the halfspace
explicitly. If that is the only reason why you need halfspaces, then you
don't need the extended kernels.

Thanks for the note on the Extended_XXX_3 kernels. That's an error. In
the beginning I thought I need separate kernels for different
dimensions, but that is not true. These kernels do not exist. Use
Extended_homogeneous and Extended_cartesian.

Peter

On Tue, 2008-01-29 at 18:03 -0800, Stephen Wong wrote:
> Hi there,
> 
>  
> 
> Another quick question for you guys.  In Section 18.5.1 in the
> documentation where a Extended_homogeneous<Gmpz> kernel is used to
> demonstrate construction of Nef_polyhedron_3 using a half-plane
> definition, I tried the same with Extended_cartesian<Gmpq> but this
> did not work – compiler errors resulted.  This documentation is found
> here:
> 
>  
> 
> http://www.cgal.org/Manual/3.3/doc_html/cgal_manual/Nef_3/Chapter_main.html#Subsection_18.5.1
> 
>  
> 
> It is also note worthy to mention that in the reference to
> Nef_polyhedron_3, there is a reference to a Extended_homogeneous_3
> (presumably 3D version of the Extended homogeneous) and
> Extended_cartesian_3, but none of these two kernels seem to actually
> exist.  In the example of 18.5.1, the Extended_homogeneous (w/o “_3”)
>  is used to initialize a 3D plane in Nef_polyhedron_3 with success.
> What am I not doing correctly to obtain the same results with
> Extended_cartesian<Gmpq>?
> 
>  
> 
> Thanks,
> 
>  
> 
> Stephen
> 
>