On Monday 08 Apr 2013 11:24:16 Jonathan Wakely wrote:
> On 8 April 2013 11:09, scalpel4k wrote:
> > Hallo experts,
> >
> > I have a little problem with access specifiers for class inheritance, which can be demonstrated in the following example.
>
> This is a general C++ question, not specific to GCC, so this isn't
> really the right place to ask it.
Thanks Jonathan for answering anyway ....
>
>
> >>>> EXAMPLE >>>
> >
> > class A {
> > protected:
> > int i = 0;
> > };
> >
> > class B : private A {
> > private:
> > int getValue() {
> > return i;
> > }
> > };
> >
> > class C : private A, public B {
> > private:
> > int getValue() {
> > return i;
> > }
> > };
> >
> > <<< EXAMPLE <<<
> >
> > This will eventually cause a compiler error, because allegedly in class C parameter i is ambiguous.
> >
> > The standard clearly says (clause 11.2) that in class B all members of class A are privat to B, so they should not be visible to class C.
>
> Access control in C++ does not affect visibility.
Am I reading something else in section 11.2 ???
[...] If a class is declared to be a base class for another class using the private access specifier, the public and protected members of the base
class are accessible as private members of the derived class. [....]
>
> > The problem seems to come from clause 10.2 in the specification that says that name resolution runs before(!) visibilty rules are applied.
>
> Yes, that's how C++ works.
>
>
> > I would say my example is slightly similar to the example given in 11.6, but I assume the virtual keyword makes some difference here. Am I right? Otherwise, I wonder why this should be ok, since f() would also be resolved twice.
>
> The virtual keywod means there is only one base class of type W, so
> there is no ambiguity.
>
> > Can anybody explain why I can have the same symbol when it's declared private in the class scope, but I can't if I make it private through inheritance? This is an ambiguity I don't understand.
>
> I don't understand what you mean by "have the same symbol when it's
> declared private in the class scope"
>
> If you mean that it works if you define C::i then it's because name
> lookup stops when it finds C::i and doesn't look in the base classes,
> so doesn't find the two ambiguous definitions of A::i caused by two A
> bases.
