<AWT Dev> RFR: 8144074: [PIT] Crash calling Toolkit.getScreenSize() on Windows.

[Sergey Bylokhov]

On 26.11.15 17:38, Alexander Scherbatiy wrote:
>
>    There is a solution to use the default configuration width/height for
> the WToolkit.getScreenWidth()/Height() methods:
>      http://cr.openjdk.java.net/~alexsch/8144074/webrev.00/

This version looks fine to me.

>
>    Or we can just revert changes introduced by the JDK-JDK-8073320 to
> fix the crash only and
>     address the getScreenWidth()/Height() scaling to a new issue for the
> further investigation:
>      http://cr.openjdk.java.net/~alexsch/8144074/webrev-revert-00
>
>
>   Thanks,
>   Alexandr.
>
> On 11/26/2015 12:00 AM, Phil Race wrote:
>> No problems in that case, ie this code :-
>> import java.awt.*;
>>
>> public class GW {
>>
>>
>>     public static void main(String s[]) {
>>         GraphicsEnvironment ge =
>> GraphicsEnvironment.getLocalGraphicsEnvironment
>> ();
>>         ge.getScreenDevices();
>>         java.awt.Toolkit.getDefaultToolkit().getScreenSize();
>>
>>     }
>>
>> }
>>
>> Also I did basic checking that SwingSet still  comes up fine.
>>
>> -phil.
>>
>> On 11/25/2015 12:52 PM, Jim Graham wrote:
>>> I don't see anything wrong with this and it fixes the crash so it
>>> must be better.
>>>
>>> The only thing I'd suggest is to verify that the new code doesn't
>>> cause problems in the case where the devices were already
>>> initialized.  If that's OK too, then I think this is a good fix.
>>>
>>> But hopefully someone in the AWT team can confirm that UpdateDevices
>>> is safe to call there...
>>>
>>> One issue to track for maybe fixing up later is if this bypasses the
>>> handling of the -Dsun.java2d.uiScale property...?
>>>
>>>         ...jim
>>>
>>> On 11/25/15 11:39 AM, Phil Race wrote:
>>>> This resolves a crash on WIndows startup due to changes in the hidpi
>>>> support
>>>> that requires the graphics devices be initialised first in order to get
>>>> screen dimensions.
>>>>
>>>> http://cr.openjdk.java.net/~prr/8144074/
>>>> https://bugs.openjdk.java.net/browse/JDK-8144074
>>>>
>>>> -phil.
>>
>


-- 
Best regards, Sergey.