Re: [boost] Preview 2 of a template Geometry Library

From: Michael Fawcett (michael.fawcett_at_[hidden])
Date: 2008-03-26 18:12:26

On Wed, Mar 26, 2008 at 4:36 PM, John Femiani <JOHN.FEMIANI_at_[hidden]> wrote:
>
>  IMHO the core need from a point library is the concepts, not the models,
>  because there are so many different well-tuned geometry libraries that
>  exist, and the idea is that boost _algorithms_ work on any "point",
>  rather than that boost have a great point class (those are easy).

I agree, what do you think of this extremely simple starting point?
The names definitely need work.

template <typename T>
struct CompileTimePointConcept {
        void constraints() {
                typedef typename at_c<typename T::element_seq, 0>::type x_type;
                typedef typename at_c<typename T::element_seq, 1>::type y_type;

                x_type &x = get<0>(value);
                y_type &y = get<1>(value);
        }

private:
        T value;
};
template <typename T, int N>
struct CompileTimePointConcept<T[N]> {
        void constraints() {
                T &x = get<0>(value);
                T &y = get<1>(value);
        }

private:
        T value[N];
};

template <typename T>
struct RuntimeIndexablePointConcept {
        void constraints() {
                typedef typename at_c<typename T::element_seq, 0>::type x_type;
                typedef typename at_c<typename T::element_seq, 1>::type y_type;

                BOOST_MPL_ASSERT((is_same<x_type, y_type>));

                int i = 0;
                x_type &x = value[i];
                y_type &y = value[i + 1];
        }

private:
        T value;
};
template <typename T, int N>
struct RuntimeIndexablePointConcept<T[N]> {
        void constraints() {
                int i = 0;
                T &x = value[i];
                T &y = value[i + 1];
        }

private:
        T value[N];
};

As you can see, one behaves very much like a tuple, the other like an
array.  T::element_seq could be an mpl::vector or maybe Fusion could
help adapt legacy structs easier (I've not played with Fusion yet).

What else would a PointConcept need?  "Point" almost feels like a
misnomer since it implies much more than most algorithms require...

--Michael Fawcett