<AWT Dev> [9] Review request for 8176097 Window set location to a display with different DPI does not properly work

[Alexandr Scherbatiy]

Hello,

Could you review the updated fix:
   http://cr.openjdk.java.net/~alexsch/8176097/webrev.02

   - sysX and sysY values are kept the the same as the passed x and y 
values.

Thanks,
Alexandr.

On 3/3/2017 8:19 PM, Alexandr Scherbatiy wrote:
>
> Hello,
>
> Could you review the updated fix:
>   http://cr.openjdk.java.net/~alexsch/8176097/webrev.01
>
>   - sysX and sysY values are updated in the same way as x and y in the 
> WWindowPeer.setBounds(...) method.
>
> Thanks,
> Alexandr.
>
> On 3/2/2017 10:13 PM, Sergey Bylokhov wrote:
>> Hi, Alex.
>> Looks like after the fix the fields in sysX/Y and sysW/sysH will have 
>> incompatible values, the location will use the one screen to scale 
>> and the size will use another, Is it possible to store the 
>> location+size based on same screen?
>>
>>> Hello,
>>>
>>> Could you review the fix:
>>>   bug: https://bugs.openjdk.java.net/browse/JDK-8176097
>>>   webrev: http://cr.openjdk.java.net/~alexsch/8176097/webrev.00
>>>
>>>   Suppose the system has two displays with different DPI. A user can 
>>> create a window on the first screen
>>>   and set the windows bounds from the second screen.
>>>   The windows scales the given bounds from the user space (x, y, w, 
>>> h) to the device space (s1 * x, s1 * y, s1 * w, s1 * h),
>>>   preserves the location in the device space and updates the windows 
>>> size so it has the same size in the screen 2 user space.
>>>   (s1 * x, s1 * y, s1 * w, s1 * h) -> (s1 * x, s1 * y, [s1 * w] * s2 
>>> / s1, [s1 * h] * s2 / s1) = (s1 * x, s1 * y, s2 * w, s2 * h)
>>>   which is (s1 * x / s2, s1 * y / s2, w, h) in screen 2 user space.
>>>   This is necessary for the windows dragging from one display to 
>>> another.
>>>
>>>   The fix updates the window size during the size setting if it 
>>> detects that the window is located on a different screen with 
>>> different DPI
>>>   so the the second screen scale factor is properly taken into account.
>>>
>>>   The fix also updates the Robot.createCompatibleImage() method to 
>>> use the scale from the screen the the screenshot is taken.
>>>   See the issue JDK-8173972 createScreenCapture not working as 
>>> expected on multimonitor setup with different DPI scales.
>>>
>>> Thanks,
>>> Alexandr.
>>>
>