Make specialization for all unsigned integer types.

This commit is contained in:
Atgeirr Flø Rasmussen 2016-06-16 08:47:21 +02:00
parent 291b13c2d4
commit 89cb2680a1

View File

@ -579,25 +579,26 @@ private:
namespace detail
{
/// \brief Computes the maximum of the absolute values of two values.
template<typename T>
template<typename T, typename Enable = void>
struct MaxAbsFunctor
{
typedef T result_type;
result_type operator()(const T& t1, const T& t2)
using result_type = T;
result_type operator()(const T& t1,
const T& t2)
{
return std::max(std::abs(t1),std::abs(t2));
return std::max(std::abs(t1), std::abs(t2));
}
};
/// \brief Specialization to avoid ambiguous abs() for unsigned integers.
template <>
struct MaxAbsFunctor<std::size_t>
// Specialization for unsigned integers. They need their own
// version since abs(x) is ambiguous (as well as somewhat
// meaningless).
template<typename T>
struct MaxAbsFunctor<T, typename std::enable_if<std::is_unsigned<T>::value>::type>
{
typedef std::size_t result_type;
result_type operator()(const std::size_t& t1, const std::size_t& t2)
using result_type = T;
result_type operator()(const T& t1,
const T& t2)
{
return std::max(t1, t2);
}