@jensmaurer @opensdh

enum E {e};
struct X {
  using E::e;
};
int f(X x) {return x.e;}

The declaration set of the lookup for the member-qualified name e in scope X is {E::e} where E::e is not a member of X. I also think the subconscious intent of defining the subobject set, in this case, should be { X } where the using-declaration was found in X. So, I think the wording might be changed to

The lookup set for N in C, called S(N,C), consists of two component sets: the declaration set, a set of declarations that introduces the name N; and the subobject set, a set of subobjects where declarations that introduce N were found(possible using-declarations).

@opensdh, the suggested approach in the fix sounds reasonable. What do you think?

Sure. It's a very minor point—we don't say or care what they're members of—but the using of an enumerator definitely stretches it. I'd simplify by saying "… where such declarations were found (…).".

I'd simplify by saying "… where such declarations were found (…).".

I have to say the declarations in the declaration set are not necessary to be the declarations that were found. As the above example exposed, the found declaration is the using-declaration but the recorded declaration in the declaration set is the replacing declaration for the using-declaration.

In normal cases, the declarations in the declaration set are exact the declarations that were found by lookup. However, using-declaration is special here, since it has a special specification:

If any such declaration is a using-declarator whose terminal name ([expr.prim.id.unqual]) is not dependent ([temp.dep.type]), it is replaced by the declarations named by the using-declarator ([namespace.udecl]).

Because of the replacement, they are not the same declarations.