On 9/16/2013 4:09 PM, Anthony Petrov wrote:
Hi Alexander,
The fix looks good overall. A couple of minor issues:
src/macosx/native/sun/awt/AWTWindow.m
369 if(!self.isEnabled){
Please add missing spaces.
377 "checkBlockingAndOrder", "()Z");
378 JNFCallVoidMethod(env, platformWindow,
jm_checkBlockingAndOrder);
This is actually a Boolean method, not a Void one. Please use the
correct JNFCall* function.
Sorry. I missed your comments before the push.
I have created new issue 8025438 [macosx] right JNFCall* method
should be used in JDK-8008728 fix
As for the test, how does it fail? I see that all exceptions get
thrown on the test thread, not the main thread. Does this cause the
jtreg harness to consider the test failed or not?
Yes. The jtreg fails the test without the fix.
The modalDialog.setVisible(true) call blocks the thread. May be
there is a better way to test a modal dialog behaviour.
Thanks,
Alexandr.
--
best regards,
Anthony
On 09/05/13 18:59, Alexander Scherbatiy wrote:
Hello,
Could you review the fix:
bug: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=8008728
webrev: http://cr.openjdk.java.net/~alexsch/8008728/webrev.00
Suppose there are 2 frames and one modal dialog which has the first
frame as a parent.
When we click on the second window the canBecomeMainWindow method
returns NO (the second window
is disabled because of the modal dialog) and it does not receive the
windowDidBecomeMain event.
But Mac OS X system places the second window over the modal dialog.
The fix checks the right order of windows in the canBecomeMainWindow
method.
Thanks,
Alexandr.
