On Wed, 9 Mar 2022 11:02:08 GMT, Alexey Ivanov <[email protected]> wrote:
>>> @aivanov-jdk Thanks for re-running the test. Where does this failure come
>>> from - Linux or Windows? The saved image's size is correct (400x300 px),
>>> but I can't explain the location of the painted area inside it and why it
>>> is so small...
>>
>> It's Linux, Ubuntu 20.10. Neither can I.
>>
>> I checked the history and the test failed on different hosts, so it could be
>> an intermittent bug in JDK itself.
>>
>> ~~Shouldn't the image be larger than 400×300? The size of the frame is
>> 400×300 in user's space, the test is running with uiScaleX=3 and uiScaleY=2,
>> therefore the size of the frame should be 1200×600. In some cases, such a
>> size may not fit on the screen.~~ I see these `sun.java2d.win.uiScale{X,Y}`
>> affect Windows only. On Ubuntu, the test runs according to the system
>> settings.
>
>> Shouldn't the image be larger than 400×300? The size of the frame is 400×300
>> in user's space, the test is running with uiScaleX=3 and uiScaleY=2,
>> therefore the size of the frame should be 1200×600. In some cases, such a
>> size may not fit on the screen.
>
> This is what happened on Windows:
> 
>
> The frame didn't fit on the screen. According to the screenshot of the entire
> screen, the resolution on the host was 1024×768. It can't fit the frame which
> is larger than screen 1200×600, especially when it is positioned at 83*3=248,
> 97*2=194.
>
> This failure has nothing to do with
> [JDK-8280861](https://bugs.openjdk.java.net/browse/JDK-8280861), it's just a
> coincidence.
@aivanov-jdk I added image-saving code to the Linux test
`ScreenCaptureGtkTest.java`. Could you please re-run it?
As before, it passes my test runs on Ubuntu 20.04.
-------------
PR: https://git.openjdk.java.net/jdk/pull/7613
