CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

On 01/31/2012 10:28 PM, Winnie Hellmann wrote:
> > Unfortunately this is not that simple because if you change the sweep
> > direction required functors are no longer the same.
> I can see your point, but shouldn't that be the business of the Traits
> class? For example one could define a Y_monotone_curve_2 in
> Arr_linear_traits_2, which would be exactly the same as
> X_monotone_curve_2 (... or am I wrong?) and then define the necessary
> functors for that class. In my understanding, Arr_linear_traits_2 allows
> for both: x- and y-monotone curves and therefore should be a model of
> ArrangementXMonotoneTraits_2 and (the non-existent)
> ArrangementYMonotoneTraits_2.
On the paper it is easy but the implementation of each functor_x
should be redone so that the functor_y use the same base
implementation. But I only cited one functor but there are more
(Compare_y_at_x,...). This means that for each functor of each Traits
(linear, circular, bezier,...) you need to do that job of
refactorization.


> By the way: Is there a reason, why Arr_linear_traits_2 is listed as a
> model for ArrangementXMonotoneTraits_2 [1], but
> ArrangementXMonotoneTraits_2 is not listed as a concept supported by
> Arr_linear_traits_2 [2]?
The documentation  is correct:
Arr_linear_traits_2 is said to model concept ArrangementTraits_2 which
refines ArrangementXMonotoneTraits_2.

Sebastien.


> Winnie Hellmann
>
> [1]
> http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Arrangement_on_surface_2_ref/Concept_ArrangementXMonotoneTraits_2.html
> [2]
> http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Arrangement_on_surface_2_ref/Class_Arr_linear_traits_2.html