c++templatesgeneric-programmingidioms

How to express "the minimum integral type larger than T"?

Suppose I have an integer type T (signed or unsigned). I want to refer (at compile time) to the smallest integer type (signed or unsigned) which can hold, say, std::numeric_limits<T>::max() plus 1 (in the non-overflowing sense, I mean).

What's a nice generic way of doing that?

Solution

  • For unsigned types, this does the trick:

    template <typename T>
constexpr unsigned size_in_bits() { return  sizeof(T) * CHAR_BIT; }

template <typename T>
using least_larger_uint_t = 
    typename boost::uint_t<size_in_bits<T>() + 1>::least;

    And if we want something to work for any integer type:

    template <typename T, int NumBits>
using boost_integer_type = 
    typename std::conditional<
        std::is_unsigned<T>::value,
        boost::uint_t<NumBits>, 
        boost::int_t<NumBits>
    >::type;

template <typename T>
constexpr unsigned size_in_bits() { return sizeof(T) * CHAR_BIT; }

template <typename T>
using least_larger_integral_t = 
    typename boost_integer_type<T, size_in_bits<T>() + 1>::least;

    See the documentation for Boost.Integer for details on int_t<N> and uint_t<N>.