Thread doubt (examlab question)

[Emanuele Ghe]

Hi,
this code comes from an exam lab question:

 class A implements Runnable{

     public static void doSleep(int y){
          try{
               Thread.sleep(y);
          }catch(Throwable e){
          }
     }

     public synchronized void run(){
          doSleep(500);
          System.out.println(Thread.currentThread().getName());
          doSleep(500);
     }

}


public class Multi{

     public static void main(String arguments[]){
          Thread rba=new Thread(new Thread(new A()));
          Thread rbb=new Thread(new Thread(new A()));
          rba.start();
          rbb.start();
     }

}   

The answer of this question is

  The result is unpredictable.  

In the explanation is stated that, even if the run() method is synchronized, there is no way to know
which thread will run first.

I think this explanation is misleading.
In my opinion the result is unpredictable because, since 2 threads are started with 2 different A objects, the synchronization is useless, because each object will acquire the lock on its A object independently from each other.

What do you think ?

p.s. I don't want to critic devaka work, is awesome. I just wanna be sure to understand everithing in the correct way.

Thanks.

[Javier Cortes]

You can indeed drop synchronized on the run method because, as you said, each Thread object is getting different locks when needed. Since synchronization was useless try and create a single object A and send it as target to both rba and rbb (don't synchronize either). You are at the same scenario here, sort of proof by contrapositive. Then, it all boils down to the fact that the JVM doesn't guarantee that the order in which the threads are started is going to be the order in which they'll become runnable.

[Emanuele Ghe]

Javier Cortes wrote:You can indeed drop synchronized on the run method because, as you said, each Thread object is getting different locks when needed. Since synchronization was useless try and create a single object A and send it as target to both rba and rbb (don't synchronize either). You are at the same scenario here, sort of proof by contrapositive. Then, it all boils down to the fact that the JVM doesn't guarantee that the order in which the threads are started is going to be the order in which they'll become runnable.

Thank you, you confirmed my thoughts.