InheritanceTests | Code for analyzing cpp inheritance | Compiler library

Regarding (+), why is the B instance's private field x accessible even after the cast to A?

Accessible to whom? In your example, B.Foo() is called, because B has overridden A.Foo() (which is virtual). Since the method being called is in fact in class B, the method uses the x field declared in B. That's the only field x that method can see.

The field x in A isn't visible to B anyway, since it's private. But if it were (i.e. if it were something other than private), the field x in B would hide it and the B implementation of Foo() would still use x from the B class, not from A.

Regarding (++), why when A's Foo method is executed does this.x not evaluate to 3?

In this example, the derived class C has not overridden the original method Foo() in A. Instead, it's declared a whole new method Foo(). But by casting the object back to A, the compiler no longer knows about the method declared in C. The only method it can see is A.Foo(), which is not overridden, and so the A implementation of Foo() is called.

And of course, that implementation can see only the x field declared in A, where the method itself is declared. Hence, you get 1 in the output instead of 3.