CGAL users discussion list

 wrote:
> This doesn't work as the circular kernel hides some functors from the linear kernel. 
> For instance, it defines CircularKernel::HasOn_2, which overrides LinearKernel::HasOn_2.
> Both functors answer the same question for different argument types.
> However, the CircularKernel::HasOn_2 cannot answer queries on linear kernel 
> objects.
> This yields errors when trying to use the circular kernel where a linear kernel is 
> expected (as in the example above).

Dear Michiel,

Does this yield errors...? I would say it should not: The Circular 
Kernel (CK) is supposed to _extend_ the the Linear Kernel (LK), that's 
the whole point of the design. Functors of the CK derive from functors 
of the LK (if they also exist there), so that they actually don't hide them.

But well, I just had a very quick look at the code. CK::Has_on does not 
derive from LK::Has_on, whereas Compare_..., Intersect_2 and so on are 
all deriving from their linear counterparts.

So, my first quick answer is that it's a bug, I apparently forgot the 
derivation for Has_on, sorry about this. I would like to fix this before 
the bug fix release, but I am afraid I won't have time to (maybe 
somebody else...?)

Best,
        Monique Teillaud