inheritance

[cyril vidal]

Dear all,
I ask myself about some of the rules governing the choice of variables or methods in inheritance.
1�) Is it correct to say that the same rules govern the choice of instance variables and static methods during process of subclassing?
I mean the choice of both of them depends on the type of the reference at compile time?
And there's another rule governing the choice of instance method, telling that this must be chosen depending on the type of the object referenced at runtime
So, the following code:

class A {
  String s1 = "A.s1";
  String s2 = "A.s2";
  static void m(){System.out.println("A.m()");}
}
class B extends A {
  String s1 = "B.s1";
  static void m(){System.out.println("B.m()");}
  public static void main(String args[]) {
    B x = new B();
    A y = x;
    System.out.println(x.s1+" "+x.s2+" "+y.s1+" "+y.s2);
    x.m();  y.m();
  }
}
  

implies the following output:
B.s1 A.s2 A.s1 A.s2
B.m()
A.m()
x.s1 and x.m() call the instance variable and static method of B, whereas y.s1 and y.m() refere to to the instance variable and static method of A. It is the same logic in both cases.
Am I right here?
2�)question extracted from Dan's exam:

class G {
  String s1 = "G.s1";
  void printS1(){System.out.print("G.printS1," + s1);}
  G() { printS1();}
}
class H extends G {
  String s1 = "H.s1";
  void printS1(){System.out.print("H.printS1," + s1);}
  public static void main(String[] args) {
    H h = new H();
  }
}
  

Response: H.printS1,null
Explanations:

The default constructor of H calls the constructor of G. The constructor of G calls the overridden method printS1 causing the invocation of H.printS1. The instance variable H.s1 hides G.s1. Although G.s1 has already been initialized, the initialization of the instance variables of H won't begin until the initialization and construction of G is complete.

I understand H.printS1, but definitly not the null.
Could someone give me light about this?
Thanks in advance for your response,
Cyril.

[Joel McNary]

As far as your first question, you are correct (assuming that I've read it right).
The second question breaks down into a matter of timelines:
Request new H() (Zeroes memory so have blank object)
- Call default constructor H()
--- call super constuctor G()
--- initialize G.s1 = "G.s1"
--- call this.printS1()
----- through polymorphism, method goes to H class's printS1().
----- H class's s1 has not yet been initialized, so h.s1 = null. Print (H.prints1,null);
----- return from method
--- return from super constructor.
- initialize h.s1 = "H.s1"
- return from default constructor
Remember, the call to the super class's constructor occurs before anything else in the constructor--even the initialization of the default variables.

[cyril vidal]

thanks Joel for your very good and clear explanations.
I've a
Cyril.

[cyril vidal]

... good grasp of this topic now.
Thanks a lot.

[cyril vidal]

Is it correct to say that the same rules govern the choice of instance variables and static methods during process of subclassing?
I mean the choice of both of them depends on the type of the reference at compile time?

I could have added to this rule the case of overloaded methods, that are chosen based on the referenced type too.
Example extracted from Dan'x mock exam:

class A {
  void m1(A a) {System.out.print("A");}
}
class B extends A {
  void m1(B b) {System.out.print("B");}
}
class C extends B {
  void m1(C c) {System.out.print("C");}
}
class D {
  public static void main(String[] args) {
    A c1 = new C();
    B c2 = new C();
    C c3 = new C();
    [B]C[/B] c4 = new C();
    c4.m1(c1);
    c4.m1(c2);
    c4.m1(c3);
  }
}
  

Output: ABC because reference c4 refers to an object of type C that has three overloaded implementations of method m1.
But just changin the referenced type of c4 as follows:

 
class A {
  void m1(A a) {System.out.print("A");}
}
class B extends A {
  void m1(B b) {System.out.print("B");}
}
class C extends B {
  void m1(C c) {System.out.print("C");}
}
class D {
  public static void main(String[] args) {
    A c1 = new C();
    B c2 = new C();
    C c3 = new C();
    [B]D[/B] c4 = new C();
    c4.m1(c1);
    c4.m1(c2);
    c4.m1(c3);
  }
}
 

gives AAA as output. Because c4 now just has access to A.m1() method.

[Lalitha Chandran]

Hi cyril,
This thread was very informative. But I think there is some typo in the last example you gave. Class D doesn't has m1 method. There is a compile time error in this program.
Could you please let use know the correction.
Thanks
Lalitha

[Anupam Sinha]

There are a lot of errors in the second program posted by Cyril.

[Darren McLeod]

change
D c4 = new C();
to:
A c4 = new C();
and you will get the AAA result he describes. Cyril posted another thread describing this effect.
[ July 17, 2003: Message edited by: Darren McLeod ]

[Lalitha Chandran]

Thanks Darren...

[cyril vidal]

Oh, I'm sorry, I did'nt connect for a long time and couldn't answer to the question of Lalitha.
Of course, Darren is right.
You juste have to change the type of c4:
A c4 = new C();
and you get the correct output: AAA
I'm sorry for this mistake,
Cyril.