Hi,
I was reading about the implementation of the Computational Geometry
and Algorithms Library (www.cgal.org) recently and one particular
thing struck me. They mentioned things like
template<class T>
class Calc_something
{
operator()(T &t1, T& t2) {...}
};
when used in a context like
...
Point p1(0, 0, 0);
Point p2(1, 0, 0)
Calc_something()(p1, p2);
are inlined in many cases by the compiler. Is this the case? Are
things like this fairly straightforward optimizations? I don't really
know compiler design at all, so I was thinking worst case scenario
where an object is build and its method called. I'm also curious what
generally happens when such objects have state stored in them?
thanks,
wes
