Tamer Higazi said:
As long as i know are constructors and destructors only provided in an
PUBLIC Area to give parameters for a class.
All constructors and destructors MUST be declarated "public" to get
access to it.
Nonsense. Just like other members of the class, constructors and
destructors can have access specifiers which limit the places where
objects of that class can be created or destroyed. "Private" access
specifier limits those places to either other member functions of
the same class or friends (classes or functions).
The private area are class variables and methods which can be called
from public methods only.
No. They can be just as well called from private member functions
or from functions declared as friends or from member functions of
classes declared as friends.
But you won't be able to call these methods
and variables outside the class in the "main" part.
Yes, that's the main idea.
Konstruktors and Destruktors are only good if you want to give special
values on the way to a class. Even you can overload them, by having
different numbers (kinds) of parameters.
Nonsense. Constructors are needed if you want to have the control
over the process of constructing of the object (opening files,
checking system conditions, etc.)
the destructor destroys the konstructor and frees the memorie.
This statement doesn't make sense.
you are NOT forced to create an Konstruktor and Destruktor. If you don't
declare them, then a Standard Kon- and destruc- tor will be called
Not "standard". "Default".
I make you an example:
class GoodTest
{
private:
int result;
public:
GoodTest(int a, int b, int c)
{
result = a * b *c;
}
~GoodTest()
{
};
void output
{
cout << "result is: " << result;
}
};
int main()
{
GoodTest Teste(3,4,5);
Teste.output();
return 0;
}
Along with the fact that this example doesn't compile due to errors,
it also doesn't demonstrate the use of _private_ constructors or
destructors as the OP asked.
Polymorphic objects. Constructed by a "factory method" or a friend
function, the objects reside in freestore. The user receives
a pointer to the object and uses it, then passes it to be destroyed
in another function.
class SomeBaseClass {
SomeBaseClass(); // private constructor
virtual ~SomeBaseClass(); // private destructor
public:
static SomeBaseClass* createMeOne(int); // loaded from a DLL
virtual void doWork();
static void deleteThis(SomeBaseClass*&);
};
int main() {
SomeBaseClass* p = SomeBaseClass::createMeOne(42);
p->doWork();
SomeBaseClass::deleteThis(p);
}
... somewhere in the library
class SomeDerivedClass : public SomeBaseClass {
public:
SomeDerivedClass(int);
private:
void doWork() {
doActualWork();
}
};
SomeBaseClass* SomeBaseClass::createMeOne(int what) {
return new SomeDerivedClass(what); /// <<<<<<<<<<<<<<<<< !!!
}
void SomeBaseClass::deleteThis(SomeBaseClass*& ptr) {
delete ptr;
ptr = 0;
}
Victor