[dcl.spec.auto.general] Call to conversion function whose conversion-function-id contains a placeholder CWG2493

struct A{
  operator auto(){
     return 0;
  }
};
int main(){
   A a;
   a.operator auto(); // #1
   a.operator int(); //  #2
}

In this example, GCC only accepts the manner of #2 while Clang only accepts #1. However, The usage of #1 seems to be ill-formed as per [dcl.spec.auto#6]. For operator int, it's not the id-expression of the conversion function whose name should have been operator auto.

As per [dcl.spec.auto#13], that is:

Redeclarations or specializations of a function or function template with a declared return type that uses a placeholder type shall also use that placeholder, not a deduced type.

So, it seems that operator auto might be the unique name to nominate the conversion function(As a qualified-id that appears in the declarator-id when defining the entity outside the enclosing class definition).

So, the issue falls out which is the name of the conversion function？ Literally, the unique name should be operator auto, however, when using the name as an id-expression of a class member access expression, such usage is forbidden by [dcl.spec.auto]. The standard does not specify what should do here.

Start by reading [dcl.spec.auto#6] and [dcl.spec.auto#13] in the C++ draft, then compare the GCC and Clang behavior described for the two calls. Determine how the standard names and permits a conversion function whose conversion-function-id contains a placeholder; done means the wording resolves the ambiguity between the two forms.