CGAL users discussion list

["S K BOSE" <>]

Dear Marc,

Thanks, It works even with CGAL::Gmpq, after converting types in normal
calculation.

With regards,

Bose

-----Original Message-----
From: Marc Glisse 
[mailto:]
 
Sent: Monday, October 04, 2010 2:46 PM
To: 

Subject: RE: [cgal-discuss] Operator Overloading problem for Vector in Gmpzf
Kernel

Wow, you seem to be out of luck.

Nef_* looks very limited in the kernels it accepts (basically only those 
where the number type is a quotient), and because the 
Algebraic_structure_traits doesn't provide a default trivial gcd for field 
types, you can't even provide a Fraction_trait for CORE::Expr and pretend 
it is a quotient (or use a CGAL::Quotient<CORE::Expr>).

typedef CGAL::Simple_cartesian<CGAL::Quotient<CGAL::Gmpz> > Kernel;
may compile, though the square root will be far from accurate.

An Sqrt_extension of Gmpq looks promizing, but it also seems to be missing 
required pieces.

You may be able to convert back and forth between types, say:
n = n / Kernel::FT(CGAL::sqrt(CGAL::to_double(n.squared_length())));
if you don't mind the lack of accuracy.

On Mon, 4 Oct 2010, S K BOSE wrote:

> Dear All,
>
> Here is my small extracted code. I want to use both Polyhedron_3 and Nef
> Polyhedra.
>
> #include <CGAL/Exact_predicates_exact_constructions_kernel_with_sqrt.h>
>
> #include <CGAL/Polyhedron_3.h>
> #include <CGAL/Nef_polyhedron_3.h>
>
> typedef CGAL::Exact_predicates_exact_constructions_kernel_with_sqrt
Kernel;
> typedef CGAL::Polyhedron_3<Kernel> Polyhedron;
> typedef CGAL::Nef_polyhedron_3<Kernel> Nef_polyhedron;
> typedef Kernel::Point_3   Point_3;
> typedef Polyhedron::Facet_iterator Facet_iterator;
> typedef Polyhedron::Halfedge_around_facet_circulator HF_circulator;
> typedef Kernel::Vector_3           Vector;
>
>
>
> int main() {
>
>  Point_3 p( 1.0, 0.0, 0.0);
>  Point_3 q( 0.0, 1.0, 0.0);
>  Point_3 r( 0.0, 0.0, 1.0);
>  Point_3 s( 0.0, 0.0, 0.0);
>
>  Polyhedron P;
>  P.make_tetrahedron( p, q, r, s);
>  Nef_polyhedron N(P);
>  //...........do some boolean operation
>  if(N.is_simple())
>  N.convert_to_Polyhedron(P);
>  Facet_iterator f;
>  for(f = P.facets_begin(); f != P.facets_end();f++)
>  {
>      HF_circulator he = f->facet_begin();
>      HF_circulator end = he;
>      CGAL_For_all(he,end)
>       {
>          const Point_3& prev = he->prev()->vertex()->point();
>          const Point_3& curr = he->vertex()->point();
>          const Point_3& next = he->next()->vertex()->point();
>          //Vector n = CGAL::cross_product(next-curr,prev-curr);
>          //  n = n / CGAL::sqrt(n.squared_length());
>          Vector n = CGAL::unit_normal(prev,curr,next);
>        }
>  }
> }
>
>
> a) If I use "Exact_predicates_exact_constructions_kernel_with_sqrt", it
> gives error in decomposer(h.b(),num,denom) from Nef_S2/Normalizing.h like
> 1>G:\cgal\CGAL-3.7-beta1\include\CGAL/Nef_S2/Normalizing.h(306) : error
> C2064: term does not evaluate to a function taking 3 arguments
> 1>G:\cgal\CGAL-3.7-beta1\include\CGAL/Nef_S2/Normalizing.h(307) : error
> C2676: binary '*=' : 'CGAL::Null_tag' does not define this operator or a
> conversion to a type acceptable to the predefined operator
> ....................
> ...................
>
> b) If I use "Exact_predicates_exact_constructions_kernel", it gives the
same
> old problem as I was getting with homogeneous
>
> G:\cgal\CGAL-3.7-beta1\include\CGAL/Kernel/function_objects.h(1878) :
error
> C2678: binary '/' : no operator found which takes a left-hand operand of
> type 'CGAL::Vector_3<R_>' (or there is no acceptable conversion) -----
>
> c) If I use "Exact_predicates_inexact_constructions_kernel", it gives
>
> G:\cgal\CGAL-3.7-beta1\include\CGAL/Nef_3/Bounding_box_3.h(89) : error
> C2039: 'lo' : is not a member of 'CGAL::Box_intersection_d::Box_d<NT_,N>'
>
> In this context what should be my kernel. My compiler is Microsoft Visual
> Studio 2005
>
> With regards,
>
> Bose
>
> Original Message-----
> From: Marc Glisse 
> [mailto:]
> Sent: Tuesday, September 21, 2010 7:01 PM
> To: 
> 
> Subject: RE: [cgal-discuss] Operator Overloading problem for Vector in
Gmpzf
> Kernel
>
> On Tue, 21 Sep 2010, S K BOSE wrote:
>
>> Thanks for reply. As Marc mentioned it is the problem for homogeneous
>> Vector, But if I use Exact_predicates_exact_constructions_kernel (which I
>> feel it is Cartesian) it gives same problem. Same problem for Gmpq kernel
>> also (as what Bernd suggested).
>
> If you really need sqrt, you want either an approximate kernel
> (Exact_predicates_inexact_constructions_kernel may work) or an exact
> kernel that can handle it
> (Exact_predicates_exact_constructions_kernel_with_sqrt).

-- 
You are currently subscribed to cgal-discuss.
To unsubscribe or access the archives, go to
https://lists-sop.inria.fr/wws/info/cgal-discuss