[Bug libstdc++/93923] [10 Regression] std::is_copy_constructible raises compilation error

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93923

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> ---
You can make the program valid by ensuring you do not try to do the
is_constructible when performing overload resolution for copy initialization of
B:

#include <type_traits>

class A;

template<typename T>
struct  B {
    template<typename... Args>
      static constexpr bool not_copy_constructor_candidate
        = sizeof...(Args) != 1 || (std::is_same_v<B, Args> && ...);

    template<
        class... Args,
        std::enable_if_t<not_copy_constructor_candidate<Args...>, int> = 0,
        std::enable_if_t<std::is_constructible<T, Args...>::value, int> = 0 >
    B(Args && ... args) {}
};

struct A {
    A(const B<A>&) {}
};

int main()
{
    static_assert(std::is_copy_constructible<A>::value, "");
}


I think GCC is correct to reject the original code.