Some questions about implementation of nullptr in Effective C++ by Scott Meyers, Second Edition?

Question

const // It is a const object...
class nullptr_t 
{
  public:
    template<class T>
    inline operator T*() const // convertible to any type of null non-member pointer...
    { return 0; }

    template<class C, class T>
    inline operator T C::*() const   // or any type of null member pointer...
    { return 0; }

  private:
    void operator&() const;  // Can't take address of nullptr

} nullptr = {};

1. operator T*() const and operator T C::*() const are already defined in class, so it can inline automatically. So, why add inline again?
2. why void operator&() const;, notvoid operator&() = delete?
3. what does nullptr = {}; mean?

Answer 1

1. Adding the extra inline is a matter of style. That specifier has to do with linkage and ODR violations, not actual inlining. All inline member function definitions and template member functions in general have that specifier implicitly. I assume Scott Meyers added it there for pedagogical reasons.

2. Effective C++ was written for C++03 originally. There was no = delete back then. Declaring but not defining a function was all you could do back then.

3. This is aggregate initialization. This implementation of nullptr_t can be initialized like that even in C++03. It creates the value of nullptr. Since nullptr_t doesn't have a user provided default c'tor, it's required.