Thanks for your feedbacks.
I have updated my changes and uploaded to webrev at
http://cr.openjdk.java.net/~dxu/8016592/webrev.01/. I have tested the
fix with the jck tests and regression tests. And the results look good.
Thanks!
-Dan
On 06/17/2013 01:03 AM, Daniel Fuchs wrote:
Hi Dan,
1. To be on the safe side I'd suggest to use:
return Objects.hashCode(listener); rather than
plain return listener.hashCode().
(I see that
removeNotificationListener(...) does not check for null).
2. I wouldn't have changed WildcardListenerInfo.equals();
If you insist in changing it then I think the proper
implementation would be:
317 if (!(o instanceof ListenerInfo))
318 return false;
319 return ((ListenerInfo)o).listener == this.listener;
which would have the advantage of being more straightforward, and
the disadvantage of not making programatically obvious that:
W.equals(L) == L.equals(W) always
(where W is WLI and L is a plain LI)
Please make sure to run jck for javax/management and
java/lang/management in addition to unit tests.
best regards,
-- daniel
On 6/14/13 9:49 PM, Dan Xu wrote:
Hi,
I have a simple fix to clean up the new javac overrides warnings in
NotificationBroadcasterSupport.java file. Please help review it. Thanks!
webrev: http://cr.openjdk.java.net/~dxu/8016592/webrev.00/
<http://cr.openjdk.java.net/%7Edxu/8016592/webrev.00/>
bug: http://bugs.sun.com/view_bug.do?bug_id=8016592
-Dan
