CGAL users discussion list

[Efraim Fogel <>]

T.vanLankveld wrote:
> As I already said, this does fix the problem for that case, but makes the
> rest of my code much too slow. The example given is heavily cut down from my
> code to make as minimal example.
>
> My full code includes calculating Delaunay triangulations for two point sets
> from which the polygons are extracted, and a number of tests on these
> polygons, including some that require boolean set operations.
>
> When using double coordinates, the program runs in 30 seconds, while it
> takes roughly a day when using either Quotient or Gmpq. For this reason,
> using exact construction is not an option.
>   
As has been said and explained many times in this list and in the
documentation, if you want robust code, you must use an exact number
type. If you use an inexact number type all bets are off. However you
can use a filtered kernel. So, have you tried the
Exact_predicates_exact_kernel kernel?
> --
> View this message in context: 
> http://cgal-discuss.949826.n4.nabble.com/Boolean-set-operation-assertion-tp3396262p3396809.html
> Sent from the cgal-discuss mailing list archive at Nabble.com.
>   
-- 
  ____  _        ____             _
 /_____/_) o    /__________  __  //
(____ (   (    (    (_/ (_/-(-'_(/
                        _/