java.util.TimerTask

[Andrew Haley <aph at redhat dot com>]

Erik Poupaert writes:
 > 
 > Hi
 > 
 > Does anybody know how java.util.Timer is supposed to work? I think I'm
 > doing something wrong, but I'm not really sure as to exactly what:
 > 
 > 		mTimer=new Timer();
 > 		mQueueTask=new QueueTask(); 	//inherits from TimerTask 
 > 						//and does something
 > 		mTimer.schedule(mQueueTask,0,pIntervalMillis);
 > 
 > The fact that a task has been scheduled to run, doesn't keep the
 > executable running. It exits anyway.
 > 
 > The JDK documentation, however, suggests that the executable should stay
 > alive.
 > 
 > http://java.sun.com/docs/books/tutorial/essential/threads/timer.html
 > 
 > "By default, a program keeps running as long as its timer threads are
 > running."
 > 
 > Is my interpretation correct that the code above should keep on running?

This:

import java.util.*;

  class QueueTask extends TimerTask
  {
    public void run() { }
  }

public class x
{

  public static void main (String s[])
  {
    
    Timer mTimer=new Timer();
    QueueTask mQueueTask=new QueueTask();
    //and does something
    mTimer.schedule(mQueueTask,0,100);
  }
}

works correctly AFAICS.

 > When I add the following workaround, the executable doesn't exit any
 > longer, but indeed stays alive:
 > 
 > //		while(true)
 > //		{
 > //			try
 > //			{
 > //				wait(10000);
 > //			}
 > //			catch(Exception e)
 > //			{
 > //			}
 > //		}
 > 
 > The only issue with this way of keeping the executable alive, is the
 > fact that it consumes exorbitant amounts of CPU. I don't know why but,
 > "wait(10000)" seems to be very expensive in terms of processor capacity.
 
What platform is this?  Windows uses WaitForMultipleObjects(), posix
uses pthread_cond_timedwait().  Neither should eat CPU.

Andrew.