The wordings of the paragraph [temp.dep.type#1] miss some examples

template<class T> struct A {
  typedef int M;
  struct B {
    typedef void M;
    struct C;
  };
};

template<class T> struct A<T>::B::C : A<T> {
  M m;                          // OK, A<T>​::​M
};

For this example, the base class A<T> refers to a current instantiation name, hence it's not a dependent base class, hence name lookup will find name M in A<T>. However, the rule defines what is a current instantiation name are as the following:

A name refers to the current instantiation if it is

in the definition of a primary class template or a member of a primary class template, the name of the class template followed by the template argument list of the primary template (as described below) enclosed in <> (or an equivalent template alias specialization),

However, in the above example, we define the class C which is not the member of class template A. Firstly, these's no any rule in the standard describes such a member of a nested class is also considered as a member of a class of which the nested class is a member. Secondly, If we consider such a member (class C) is a member of the outmost enclosing class, then it would change the meaning of far too many other rules about members(discussed in this topic of std-discussion ). IIUC, the intent of [temp.dep.type#1] want to cover that any name of these enclosing classes(enclosing class templates) is the current instantiation name(include the template-id whose template-name is the name of these enclosing classes).

Start with paragraph [temp.dep.type#1] and the nested-class C++ example in the issue. Read the linked std-discussion topic to understand the concern about members of enclosing class templates and current instantiation names. Done means the standard wording and examples clearly resolve whether this case is covered.