CGAL users discussion list

[Ben Supnik <>]

Hi Y'all,

I am using a kernel based on

Simple_cartesian<lazy_exact_nt<Quotient<MP_Float> > >

I make an exact-predicates triangulation where every finite vertex is 
inside an axis-aligned bounding box with integer corners.

Here's the part I don't understand:

- If I compare every vertex to this bounding box (promoting the bounding 
box from double to my kernel numeric type) every point is within or on 
the box.

- If I convert my coordinates to_double, those doubles sometimes are 
outside the box by rounding error.

What's particularly weird is that
- CGAL tells me that the points that "go outside" the box are EXACTLY on 
the AABB edge and
- Since the AABB edge is "integral" (e.g. -117.0, 34.0 -> -116.0, 35.0) 
they should be exactly representable in a IEEE double.

I see from the manual that to_double is an approximation...what is the 
right way to get a conversion back to double that will be exact when 
exactness is possible?

cheers
ben


--
Scenery Home Page: http://scenery.x-plane.com/
Scenery blog: http://xplanescenery.blogspot.com/
Plugin SDK: http://www.xsquawkbox.net/xpsdk/
X-Plane Wiki: http://wiki.x-plane.com/
Scenery mailing list: 

Developer mailing list: