uni20::check::UlpOrderedReal Concept Reference

Real scalar with an implemented ordering of adjacent representable values. More...

#include <floating_eq.hpp>

Concept definition

template<typename T>
concept uni20::check::UlpOrderedReal = 
    IeeeBinaryReal<T> ||
    (std::same_as<std::remove_cvref_t<T>, long double> && uni20::numeric_limits<long double>::is_iec559 &&
     uni20::numeric_limits<long double>::radix == 2 && uni20::numeric_limits<long double>::digits() == 64 &&
     uni20::numeric_limits<long double>::min_exponent == -16381 &&
     uni20::numeric_limits<long double>::max_exponent == 16384 &&
     uni20::numeric_limits<long double>::denorm_min() > 0 &&
     uni20::numeric_limits<long double>::denorm_min() < uni20::numeric_limits<long double>::min())

Detailed Description

Real scalar with an implemented ordering of adjacent representable values.

Supports IEEE binary interchange formats and native extended long double, independently of whether the optional float80 alias is enabled. Unlike IeeeBinaryReal, this does not require a padding-free object representation.