Hi,

I have found myself implementing a lot of tuple access protocols lately,
and while tuple_size and get()[1] can, in general, be handled by just
one 'overload', I have not found an easy way for tuple_element. I thus
implemented a small helper locally, so I can say:

    template <size_t I>
    struct tuple_element<I, MyClass> : select<I, std::string, double,
int> {};

iow: select<I, Ts...> picks the I'th T (zero-based) from pack Ts...

Possible implementation:

    template <size_t I, typename Head, typename...Tail>
    struct select : select<I-1, Tail> {};
    template <typename Head, typename...Tail>
    struct select<0, Head, Tail...> { using type = Head; };

I haven't found anything existing on this mailing-list or in C++17, so
I'm wondering whether I should propose it.

I'm a bit concerned about the name. 'select' is used for the Posix
function already, and 'switch' is a keyword. 'selection' doesn't really
work, either. Or something entirely different like pick<I>::from<Ts...>?

What do you think?

Thanks,
Marc

[1] The four get() overloads can be generated with perfect forwarding
and constexpr-if:

    class QPoint {
       int x, y;
       // ...
       template <size_t I, typename Point>
         requires std::is_same_v<std::decay_t<Point>, QPoint>>
       friend constexpr decltype(auto) get(Point &&p) noexcept {
           if constexpr (I == 0) {
               return (std::forward<Point>(p).x);
           else if constexpr (I == 1) {
               return (std::forward<Point>(p).y);
       }
    };

(at least I have not found a way this breaks on my use-cases, but I also
admit that I have not gone to the length of proving it correct to
myself).



--
You received this message because you are subscribed to the Google Groups "ISO C++ Standard - Future Proposals" group.
To unsubscribe from this group and stop receiving emails from it, send an email to std-proposals+unsubscribe@isocpp.org.
To post to this group, send email to std-proposals@isocpp.org.
To view this discussion on the web visit https://groups.google.com/a/isocpp.org/d/msgid/std-proposals/851fb43e541107ecafde9e00efe3bf64%40kdab.com.

.

On 2017-10-24 10:49, Marc Mutz wrote:
> I'm a bit concerned about the name. 'select' is used for the Posix
> function already, and 'switch' is a keyword. 'selection' doesn't
> really work, either. Or something entirely different like
> pick<I>::from<Ts...>?

Boost.MPL uses at<>, but that doesn't work on pack expansions directly:

     mpl::at<mpl::vector<Ts...>, I>

One could also just use

     tuple_element<I, MyClass> : tuple_element<I, std::tuple<int,
std::string, double>> {};

but there's always the nagging feeling that some of tuple's magic will
change the result here...

--
You received this message because you are subscribed to the Google Groups "ISO C++ Standard - Future Proposals" group.
To unsubscribe from this group and stop receiving emails from it, send an email to std-proposals+unsubscribe@isocpp.org.
To post to this group, send email to std-proposals@isocpp.org.
To view this discussion on the web visit https://groups.google.com/a/isocpp.org/d/msgid/std-proposals/31ecee6e705799e1df3cd2170011b13b%40kdab.com.

.