CGAL users discussion list

[Bernd Gaertner <>]

Dear Amir,

it's a known misconception in using CGAL. The kernel Cartesian<double> 
is inexact and can't in general guarantee correct results. Please see 
the FAQ item at http://www.cgal.org/FAQ.html#inexact_NT for more about 
this topic, and for ways of computing exactly with CGAL.

Best,
Bernd.


 wrote:
> I use this code:
>
> Result=CGAL::intersection(TwinPlane, NewPlane, AftPlane);       
> Point3D Target;                 
> bool isdo=assign(Target,Result);                        
> Number d1=CGAL::squared_distance(NewPlane,Target);
> Number d2=CGAL::squared_distance(AftPlane,Target);
> Number d3=CGAL::squared_distance(TwinPlane,Target);
>
> "Number" beign a cartesian<double> kernel number type, and (TwinPlane, NewPlane, 
> AftPlane) are all Plane_3<cartesian<double>> types. However, the distances d1,d2,d3 which are 
> supposed to be 0 (because the point Target is the intersection) are sometimes way off, and the Target 
> point is not correct. Moreover, when I change the order of the planes in the CGAL::intersection function, 
> the results sometimes differ. Is that a known bug?
>
> Amir.