CGAL users discussion list

["Eivind LM" <>]

Hello Fernando,

> Hello Eivind,
>
> > Speaking of Vertex_handle; am I the only one to mix up '.' and '->'  
> > with Vertex_handle?
>  >
> > [snip]
>  >
> > Using '*' and '->' at the same time on a variable (which is not a  
> > pointer type) looks "wrong" in my eyes
>  >
> How about:
>
>     std::list< Some_iterator > objects ;
>
> Or in the upcoming version of C++
>
>    std::list< std::shared_ptr<someobj> > objects ;
>
> ?
>
> With those you also need '->' instead of '.'

Ok, I have neither had to deal with lists of iterators, nor used the  
upcoming version of C++, but I understand that this is more common than  
what I am used to.

> > Can I generally assume that all types that end with "_handle" in CGAL  
> > implements an "operator->"?
> Yes.

Great! :)

> Handles *refers* to other objects, just like iterators and pointers, so  
> they provide indirect access via '->'
>
> It is idiomatic to expect the left side of '->' to possibly refer to a  
> non-existing object (say a past-the-end iterator or a null pointer),  
> while that would be quite surprising for the left side of a '.'.
>
> A default constructed handle, for instance, does not refer to any object  
> at all:
>
> void use_vertex ( Vertex_handle v )
> {
>    Vertex_handle null ;
>    if ( v != null )
>    {
>      // 'v' corresponds to a valid vertex so go ahead and use it.
>    }
> }

It all makes a lot of sense. Thanks for explaining.

If I had been able to find all this information by searching for  
"Vertex_handle" in the documentation index, then I couldn't have been  
happier. I guess I could have searched for "Handle" instead, then I would  
at least have found the link that Andreas just sent to the list. I'll try  
to be more creative next time I search the index :)

However, letting Vertex_handle have it's own page in the documentation  
makes sense to me, since it's such a basic part of the triangulation  
library.

Thanks again,
Eivind