CGAL users discussion list

[Stefan Schirra <>]

 wrote:

> I think, I have an example of a potential bug in ch_melkman.
> The example should have returned 6 points instead of 5.
> I try with convex_hull_2 which returns the 6 points. (just uncomment the 
> comment to see it)
> I try with the gmpq kernel to find if there is a precision issue but the 
> results are the same.
> Do I miss something on the condition of use of ch_melkman ?

Are you sure that your polyline (polygon) is SIMPLE?

best regards

        Stefan