🚀 go-pugleaf

In article <4fa17f$qk4@netlab.cs.rpi.edu>, Paul.Henshaw@jrc.it wrote:

> Though Roger's argument is (IMHO) valid in general, I have a specific example
> where I feel I am justified in keeping virtual methods private:
> 
> class ActiveObject {
> public:
>         ActiveObject(void* data=NULL, options_t=Normal);
>         virtual ~ActiveObject();
          ...
> private:
>         virtual void* main(void* data)=0;
> }
> 
> class ActiveSocketReader : public ActiveObject {
> public:
         ...
>         }
> 
> private:
>         virtual void* main(void* data);
>         
> }

[portion of example elided]

> Since main should be called only by ActiveObject::ActivObject() main should
> be private - however the derived class must be able to provide its own main
> function in order to be useful.

It is not necessarily the case that main should be called only by the
ActiveObject base class.  Consider a SpecialActiveSocketReader class
derived from ActiveSocketReader.  An instance of the
SpecialActiveSocketReader is just like an instance of the
ActiveSocketReader class, except that it performs some additional
processing in the main member function.  The first thing that the
SpecialActiveSocketReader::main function should do is to call the
ActiveSocketReader::main function.  It cannot do so, however, because it
was declared private instead of protected.

class SpecialActiveSocketReader : public ActiveSocketReader
   {
   public:
      ...
   private:
      virtual void* main(void* data);
   };

void * SpecialActiveSocketReader::main(void *data)
   {
   // Perform default processing.
   ActiveSocketReader::main(data);    // <--- Compiler error:  main is private.

   // Perform "special" extra processing.
   }

By declaring main private, you are making it impossible to derive from
ActiveSocketReader in this manner.


Roger

      [ Articles to moderate: mailto:c++-submit@netlab.cs.rpi.edu ]
      [  Read the C++ FAQ: http://www.connobj.com/cpp/cppfaq.htm  ]
      [  Moderation policy: http://www.connobj.com/cpp/guide.htm  ]
      [      Comments? mailto:c++-request@netlab.cs.rpi.edu       ]