Hi, Anthony.
I suggest to fix it via separate CR because of:
- Shared code is not always ready to the situation, when the exception
is thrown with locked
handler.(SunDropTargetContextPeer.postDropTargetEvent()).
- The same approach is used on windows too.
(WToolkitThreadBlockedHandler.enter())
- It would be good to have straightforward backports, when possible.
On 30.04.2013 17:02, Anthony Petrov wrote:
Hi Petr,
src/macosx/classes/sun/lwawt/macosx/CToolkitThreadBlockedHandler.java
45 unlock();
46 LWCToolkit.doAWTRunLoop(awtRunLoopMediator, processEvents);
47 lock();
I suggest to use the try/finally pattern for the unlock()/lock() calls
here to ensure the method always returns with the lock acquired.
Perhaps the same changes need to be applied to JDK 8, too.
The fix looks good otherwise.
--
best regards,
Anthony
On 04/29/2013 05:56 PM, Petr Pchelko wrote:
Hello, AWT Team.
Please, review the back-port of the fix for JDK-8006634 to the JDK7u.
I am sending it as a separate review request because the code changes
are different in the patches.
The bug:
http://bugs.sun.com/view_bug.do?bug_id=8006634
The webrev for JDK7-u:
http://cr.openjdk.java.net/~pchelko/8006634/webrev.7u/
The original webrev for JDK-8
http://cr.openjdk.java.net/~pchelko/8006634/webrev.02/
The mailing thread:
http://mail.openjdk.java.net/pipermail/awt-dev/2013-February/004335.html
The idea of the fix stays the same. However the code changes are
different, because it is affected by the fix for 7171163 made by
Alexander Zuev. His fix was different for JDK8 and JDK7. (see
LWCToolkit.m, LWCToolkit.java) However, after those fix the
differences will be eliminated.
With best regards. Petr.
--
Best regards, Sergey.
