optional<const in_place_t> should be ill-formed

[jwakely]

https://github.com/beman-project/Optional26/blob/bfeb9080c0d9c501965ae857a1decf7ed74686f5/include/Beman/Optional26/optional.hpp#L297-L299

This only rejects in_place_t not cv in_place_t as required by [optional.optional.general] p3.

For the second constraint, why decay_t instead of remove_cv_t?

Do you really want these constraints in the first place? Using static_assert to enforce these rules is likely to support better diagnostics. Also, constraining the primary template with a requires-clause means that these constraints will be checked for the optional<T&> specialization when that is added. That won't be wrong because T& will never match in_place_t or nullopt_t but it's just unnecessary compile-time overhead to check constraints that will never match.

Finally, why is this requires-clause using redundant parentheses, like it's a function call? That style seems to be used inconsistently throughout the file. Edit: sorry, was looking at the wrong file when I wrote that! The parens in requires-clauses in the Beman/Optional26 code are fine.

[steve-downey]

The lack of cv is probably me glossing over that in
T shall be a type other than cv in_place_t or cv nullopt_t that meets the Cpp17Destructible requirements (Table 35).

Should be able to move these to a static_assert in the definition of the template, also improving diagnostics instead of removing them from the overload set / making them not exist.