<AWT Dev> [11][JDK-8197808] Test java/awt/Dialog/MakeWindowAlwaysOnTop/MakeWindowAlwaysOnTop.java fails on Windows

Semyon Sadetsky semyon.sadetsky at oracle.com
Wed Feb 28 17:00:20 UTC 2018

Hi Krishna,

Does the line

EventQueue.invokeAndWait(() -> { /* Empty */ });

is really necessary for test to pass?


On 02/27/2018 09:56 AM, Krishna Addepalli wrote:
> Hi Sergey/Phil,
> Thanks for pointing that out. I have created a new webrev: http://cr.openjdk.java.net/~kaddepalli/8197808/webrev01/
>   with the appropriate changes for ProblemList.txt as well.
> I have tested on Windows 7, Ubuntu 16.04, Ubuntu 17.10 and MacOS 10.13, and the test passed with/without my changes.
> As noted earlier, on Windows 10, the test passes after the fix.
> Thanks,
> Krishna
> -----Original Message-----
> From: Philip Race
> Sent: Tuesday, February 27, 2018 6:01 AM
> To: Sergey Bylokhov <sergey.bylokhov at oracle.com>
> Cc: Krishna Addepalli <krishna.addepalli at oracle.com>; awt-dev at openjdk.java.net
> Subject: Re: <AWT Dev> [11][JDK-8197808] Test java/awt/Dialog/MakeWindowAlwaysOnTop/MakeWindowAlwaysOnTop.java fails on Windows
> Yes .. even if bugs were added to the problem list in bulk, the expectation is that they'll be removed by the bug fix that fixes the failure.
> And if it still fails on some platforms then either (1) you aren't done with your fix, or (2) you need to find a different bug ID under which to now exclude the test for the platform(s) that still fail for what would normally be some completely un-related reason.
> Other than that the fix is fine.
> -phil.
> On 2/26/18, 2:16 PM, Sergey Bylokhov wrote:
>> Hi, Krishna.
>> This test was added to the problemlist, please check that it works on
>> other platforms and remove it from the problemlist.
>> On 25/02/2018 17:53, Krishna Addepalli wrote:
>>> Hi All,
>>> Please review a fix for JDK-8197808:
>>> https://bugs.openjdk.java.net/browse/JDK-8197808
>>> Webrev: http://cr.openjdk.java.net/~kaddepalli/8197808/webrev00/
>>> The problem was that by the time Robot tries to read the color on
>>> screen, the compositing of the window is not completed, and hence,
>>> instead of reading Blue color, it is reading a combination of Blue
>>> and Red colors. Increasing the delay by 100 ms made sure that the
>>> composition is complete and the color is being read correctly. Also,
>>> this happens only on Windows 10.
>>> Apart from this, the test was not cleaning up the windows when it
>>> throws exception, so fixed that, and also added all explicit imports.
>>> Thanks,
>>> Krishna

More information about the awt-dev mailing list