<AWT Dev> [9] Review request for 8013116: Robot moves mouse to point which differs from set in mouseMove on Unity shell

[Oleg Pekhovskiy]

Hi Sergey,

I tested on Windows, Mac OS X and Ubuntu. As a result I found
calculation error in CRobot.mouseEvent native method.
It was fixed.
I also updated Javadoc for Robot(GraphicsDevice) constructor
(CCC request will be done).

Please review the next version of fix:
http://cr.openjdk.java.net/~bagiras/8013116.2

I built JDK8 with my fix and ran JCK runtime test suite for 3 platforms
for 'api.java_applet', 'api.java_awt' and 'api.javax_swing' branches.
There were no differences for Mac OS X and Windows and one test fixed on 
Ubuntu
(api/java_awt/Container/getMousePosition_NotHeadless).

Thanks,
Oleg

On 01/14/2014 09:42 PM, Sergey Bylokhov wrote:
> On 14.01.2014 21:14, Oleg Pekhovskiy wrote:
>> Hi all,
>>
>> please review the fix
>> http://cr.openjdk.java.net/~bagiras/8013116.1/
>> for
>> https://bugs.openjdk.java.net/browse/JDK-8013116
>>
>> java.awt.Robot.gdLoc field was removed to make Robot's point of 
>> origin be the same as in setters and getters from
>> GraphicsDevice (GraphicsConfiguration), MouseInfo (PointerInfo) and 
>> Frame.
>>
>> This makes sense especially on Ubuntu with Xinerama extension where 
>> all screens have common coordinate space.
> Did it work on other platfroms before the fix(windows, osx)? If yes, why?
>>
>> Two helper methods were added to 
>> 'test/java/awt/regtesthelpers/Util.java'.
>> They were used in regression test for this fix. These methods could 
>> be useful for asynchronous window managers.
>>
>> Thanks,
>> Oleg
>>
>
>