is_optional would be more efficient as a variable template

[jwakely]

https://github.com/beman-project/Optional26/blob/bfeb9080c0d9c501965ae857a1decf7ed74686f5/include/Beman/Optional26/optional.hpp#L237-L242

There seems to be no reason to prefer a class template to a variable template, it just makes it slower to compile. Why decay_t not remove_cvref_t? Or why do that at all? All uses of is_optional already use either remove_cvref or decay (which should be remove_cvref in all cases), except for the use in optional<T&>::and_then which is probably incorrect.

template <class T>
inline constexpr bool is_optional = false;
template <class T>
inline constexpr bool is_optional<optional<T>> = true;

[steve-downey]

Testing the static_asserts moved up the priority list, too. It was ad-hoc, and therefore wrong.