https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92171
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution|--- |INVALID
--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
https://wg21.link/p2280 was accepted as a defect report against all versions of
C++.
and that makes this code valid.
ref.value is now a constexpr even though ref is not one after the paper.
(In reply to Marek Polacek from comment #4)
> I think this is another example:
>
> template<int N> struct A {
> static constexpr int size() { return N; }
> };
>
> template<typename T> void f1(T t) {
> constexpr int k = t.size();
> }
>
> template<typename T> void f2(const T &t) {
> constexpr int k = t.size(); // error
> }
>
> template<typename T> void f3(const T &t) {
> constexpr int k = T::size();
> }
>
> void g(A<3> a)
> {
> f1(a);
> f2(a);
> f3(a);
> }
This full example is also fully valid after this paper too.
