Are this information applicable also to harbour? ftp://ftp.alaska-software.com/documents/oop-6-rules.pdf
„Class“ is an abstract term. The Xbase++ runtime knows class object and instance objects. Class objects are created at the first time the class function is executed, subsequent class function execution returns the same class object. To „free“ a class object the function ClassDestroy() can be used. The Xbase++ runtime has a process-global hashtable used to associate class objects with class names. ClassDestroy() simply removes the name<->class object association. However the class object is not freed until the last instance object of this class is destroyed. :New() is the default constructor to create instance objects. Instance objects are always created by class objects. Therefore class objects are sometimes called „Object Factories“. You can have an unlimited amount instance objects of the same class. In the following the term object is used to describe instance objects. Note: Class objects which reference code or data of a DLL lead to sideeffects such as that the DLL is not unloadable until all objects of the class are destroyed and the class object itself has been destroyed. Until this precondition is met the DLL can not be unloaded. All class are dynamyc There is no difference between a class object created from a class function A() generated by the compiler and a class object created with CreateClass() by a Function A(). Runtime-classes can replace compile-time classes and vice versa. Any class can be replaced by another class with the same name. That‘s why classes are dynamic. The FREEZE attribute at CLASS declaration prohibits the implicit replacement of this class. FREEZE can be used to protect class implementations from being replaced implicitly. The FINAL attribute prohibits subclassing of this class. The behaviour of FINAL classes can therefore be no longer changed -It is FINAL Consider the code on the right side: Executing the method :Bark() on objects of Class A executes method :Print() at Class A. Executing the method :Bark() on objects of Class B executes the method :Print() implemented at Class B. This behaviour is called virtual override because each method implementation in a subclass does override a possible implementation of its superclasses - Overriding of METHOD is always a replacement method replacement totally overwrites the method of its base classes during execution. That is the code in the base class is never executed when instances of subclasses are manipulated. The Class B and the method :Print() is an example of method replacement. Refinement can be achieved as outlined in Class C. To do that, we must be able to execute the superclass implementation. The SUPER statement is used to refer to the direct superclass implementation. The SUPER statement executes the superclass implementation of the current method and passes all parameters passed to this method. SUPER:Print() can be used as an alternative to control method name and parameters passed. Note: You can write SELF:A: Consider the code on the right side: Executing the method :PrintA() on objects of Class B prints „Value from A“, while executing the method :PrintB() shows „Value from B“. This behaviour is called hiding in OO terms. It simple means that redefining a variable in a subclass hides the variable of its superclass. It therefore introduces the variable. CLASS A EXPORTED: VAR Lastname INLINE METHOD Init() ::LastName:= „ValuefromA“ RETURN INLINE METHOD PrintA() ? ::LastName RETURN ENDCLASS CLASS B FROM A EXPORTED: VAR Lastname INLINE METHOD Init() ::A:Init() ::LastName:= „ValuefromB“ RETURN INLINE METHOD PrintB() ? ::LastName RETURN ENDCLASS _______________________________________________ Harbour mailing list Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
