P0340R3 SFINAE-Friendly underlying_type

[StephanTLavavej]

P0340R3 SFINAE-Friendly underlying_type

[SuperWig]

Is this something that can be applied to all modes or does it need _HAS_CXX20?
And should it have a comment about the change e.g. quoting the standard "Mandates: T is not an incomplete enumeration type."

[CaseyCarter]

Unconditional implementation seems fine: this defines behavior that was previously undefined, so an implementation of C++20 underlying_type is also a conforming implementation of C++17 underlying_type.

e.g. quoting the standard "Mandates: T is not an incomplete enumeration type."

There's not much to do; we dispatch directly to a compiler intrinsic, and the compiler generally issues the diagnostic for an incomplete type. (Although to my understanding C1XX currently treats all enumeration types as complete: https://godbolt.org/z/VkcENq.)

[StephanTLavavej]

I agree that this one should be unconditional.

[SuperWig]

There's not much to do; we dispatch directly to a compiler intrinsic, and the compiler generally issues the diagnostic for an incomplete type. (Although to my understanding C1XX currently treats all enumeration types as complete: https://godbolt.org/z/VkcENq.)

The reason I ask is because there's comments for common_reference of a similar manner

// N4810 [meta.trans.other]/5.3.5: "Otherwise, there shall be no member type."

Though if the part I initially quoted isn't going to be true then 🤷‍♂
Edit: I believe I misunderstood.