<AWT Dev> [7u8] Review request for 7177040: Deadlock between PostEventQueue.noEvents, EventQueue.isDispatchThread and SwingUtilities.invokeLater

[Oleg Pekhovskiy]

Hi Peter,

so obvious, thank you!

Oleg.

7/24/2012 3:42 PM, Peter Levart wrote:
> Hi Oleg,
>
> This is just cosmetics, but:
>
> SunToolkit:
> public synchronized boolean noEvents() {
>         return queueHead == null && !isFlushing;
>     }
>
> ... a thread calling noEvents could see occasional "spikes" of false 
> return even though there is no flushing being performed (when other 
> thread is calling flush on an empty PostEventQueue).
>
> Improved flush method would look like this:
>
>      public void flush() {
>          EventQueueItem tempQueue;
>          synchronized (this) {
>              tempQueue = queueHead;
>              queueHead = queueTail = null;
>              isFlushing =*/(tempQueue != null)/*;
>          }
>          try {
>              while (tempQueue != null) {
>                  eventQueue.postEvent(tempQueue.event);
>                  tempQueue = tempQueue.next;
>              }
>          }
>          finally {
>              isFlushing = false;
>          }
>      }
>
> Regards, Peter
>
> 2012/7/23 Oleg Pekhovskiy <oleg.pekhovskiy at oracle.com 
> <mailto:oleg.pekhovskiy at oracle.com>>
>
>     Hi!
>
>     Please review this back-port being already pushed to jdk8 but
>     deferred for 7u6.
>
>     Bug:
>     http://bugs.sun.com/view_bug.do?bug_id=7177040
>
>     Webrev:
>     http://cr.openjdk.java.net/~bagiras/7u8/7177040.1
>     <http://cr.openjdk.java.net/%7Ebagiras/7u8/7177040.1>
>
>     Review thread for 7u6:
>     http://mail.openjdk.java.net/pipermail/awt-dev/2012-July/003106.html
>
>     Reviewers 7u6 & 8:
>     Anthony Petrov, Anton Tarasov
>
>     Thanks,
>     Oleg
>
>