volatile... is this not correct way?

[Ramakrishna Nalla]

please go throgh my code:

 
class A{
volatile int i=10;// requsting compiler not to optimize(inline) code for variable i. 
}
 
class B extends A{
int i=20;
}
 
class Test extends B{
public static void main(String args[]){
A a=new A();
B b=new B();
a=b;
System.out.println(a.i);//at compile time=a.i and at runtime b.i 
}
}
 

Output: 10

Please clarify me about 'volatile'...
Is my comments correct...
Why output came 10 rather than expected 20.

Waiting....for your comments

[Srinivasa Raghavan]

Because value of member variables are decided at compile time ( not overrided ),
even if you remove the volatile key word you should get the same output.
[ July 15, 2005: Message edited by: Srinivasa Raghavan ]

[narendra darlanka]

hi,
i think it has nothing to do with volatile. accessing instance variable is not like accessing instance methods.so no dynamic binding in here
a.i gets the variable from the class of reference variable a not of the object actually held in it.

[Ramakrishna Nalla]

Thanks Srinivasa and Narendhra for your quick reply....

you are all mean..."volatile only effects in Thread concept...I mean the variable which participate in thread..only has volatile effect...?

Is it possible to give an example code for volatile without using thread concept...? :roll:

[Ramakrishna Nalla]

I am waiting for your comments... All you mean
"There is no example code for volatile without using thread concept"...

let me know please...
[ July 16, 2005: Message edited by: Ramakrishna Nalla ]

[Ramakrishna Nalla]

I feel bad with this type of response...

[Devender Thareja]

Originally posted by Ramakrishna Nalla:
Thanks Srinivasa and Narendhra for your quick reply....

you are all mean..."volatile only effects in Thread concept...I mean the variable which participate in thread..only has volatile effect...?

Is it possible to give an example code for volatile without using thread concept...? :roll:

As you mentioned there is no effect of volatile if you are not running multiple threads. So I am not sure what example you are looking for?

[Gokul Somasundaram]

I just feel, that the problem with the above code is that the subclass redeclares the variable already declared by superclass. The Runtime switching happens fine, If we remove it.
Please correct me if i am wrong.

Thanks,
Gokul.

Code:

Working fine:
class A{
int i=30;
}

class B extends A{
int j;
}

class Test extends B{
public static void main(String args[]){
A a;
B b1 = new B();
b1.i = 10;
a=b1;
System.out.println(a.i);
B b2 = new B();
b2.i = 20;
a=b2;
System.out.println(a.i);
}
}

Not Working Fine:

class A{
int i=30;
}

class B extends A{
int i;
int j;
}

class Test extends B{
public static void main(String args[]){
A a;
B b1 = new B();
b1.i = 10;
a=b1;
System.out.println(a.i);
B b2 = new B();
b2.i = 20;
a=b2;
System.out.println(a.i);//at compile time=a.i and at runtime b.i
}
}

Thanks,
Gokul.

[Tom Richards]

I don't believe this has anything to do with the keyword volatile. It is due to the "true" type of a class reference.

Example:

public class A {
int i = 10;
}

public class B extends A {
int i = 99;
}

public class C {
public static void main(String[] args) {
A a = new B();
System.out.println("a.i = " + a.i); // "a.i = 10"
System.out.println("a.i = " + ((B)a).i); // "a.i = 99"

}
}

The variable 'a' true type is "A", even though the reference has been set to B. If you cast the a variable to B, then you can get at the variables in b.