On Mon, Nov 25, 2024 at 5:09 PM mauro russo <ing.russomauro@gmail.com> wrote:

> I agree, these examples seem to be wrong, Binding `int&&`
> to an rvalue is better than binding `const int&` to an rvalue,
> but then for the other argument we have that binding
> `const int&` to an rvalue is better than the ellipsis.

Should I open an issue with PR for that ?

I think this issue has to be reviewed by CWG. If you can propose a better (and correct) example then please do. But you can also submit the issue without it.

> If `T` is `int&`, then `const T&` and `T&&` are both `int&`.

Sure, I agree... in the sense that this would be
true if deduction would lead to int&.
However, since we agree above that deduction
fails, the statement about #1 and #2 is
consequently wrong, isn't it ?

I think it would be ideal to amend the example so that we still deduce to `A<int&>`, but obviously if we end up not doing that, then we may not end up with an error upon instantiation.

Brian Bi