CGAL users discussion list

[Bernd Gaertner <>]

 wrote:
> I can do:
>
> Point1  P1(2,3);
> Point2  P2(CGAL::to_double(P1.x()),CGAL::to_double(P1.y()));
>
> But I think that this is not very efficient. 

And it's not safe, since to_double is not exact on Quotient<double>.
You could either use simply double as NT (since it's a field number
type, there is no need to put a quotient on top), in which case the
above is safe (and works even without the to_double's). Or, if you
insist on NT = GAL::Quotient<double> for some reason, you could use

P2 ( CGAL::Gmpq(P1.x().numerator()) / CGAL::Gmpq(P1.x().denominator()),
     CGAL::Gmpq(P1.y().numerator()) / CGAL::Gmpq(P1.y().denominator()) )

Bernd.