friend class in TC++ book

[mickey]

hello,
i don't understand:

class Shape {
	//class Type;
	//friend class Shape::Type;
	int _number;
	double _x, _y;
 
	class Type {
		Shape* _parent;
	public:		
		Type(Shape* sh) : _parent(sh)  { 
			double d = _parent->compute();	
			double d2 = _parent->_x;		//it compiles....	
		}	
	};	
public:
	Type _type;
	Shape(double x = 0, double y = 0) : _type(this), _x(x), _y(y) {  }		
	double compute() { 	
		return _x * _y;  }
};

class Shape {
	//class Type;
	//friend class Shape::Type;
	int _number;
	double _x, _y;

	class Type {
		Shape* _parent;
	public:		
		Type(Shape* sh) : _parent(sh)  { 
			double d = _parent->compute();	
			double d2 = _parent->_x;		//it compiles....	
		}	
	};	
public:
	Type _type;
	Shape(double x = 0, double y = 0) : _type(this), _x(x), _y(y) {  }		
	double compute() { 	
		return _x * _y;  }
};

To copy to clipboard, switch view to plain text mode 

if i comment those lines, it complies anyway...

[wysota]

Could you prepare a minimal compilable example of the above?

[mickey]

Could you prepare a minimal compilable example of the above?

the example is compilable;
Thw question was: how come TC++ vol2 write those 2 lined commented in the post #7 (if they are unuseful) ???

[jacek]

how come TC++ vol2 write those 2 lined commented in the post #7 (if they are unuseful) ???

Not everything what's written in a book is correct.

[wysota]

I'd say this code in the book might be an explicit declaration of an implicit feature of the standard. Especially that we know some compilers tend to create their own "standards".

[jacek]

double d2 = _parent->_x; //it compiles....

The standard says:

A nested class is a member and as such has the same access rights as any other member.

Which means that nested classes behave as if they were friends of nesting class.