On Wed, 16 Nov 2022 05:33:26 GMT, Prasanta Sadhukhan <[email protected]> 
wrote:

>> We will not throw NPE if robot is null. Because we have null check for robot 
>> when we want to use it.
>> This test was originally written to check for crash after producing 
>> artifacts, thats why i think we continue to drawImage onto resizing frame 
>> even if Robot fails. So its better if we keep it that way.
>> 
>> Also initially i used mouseMove() before we do screen capture, but this was 
>> causing considerably lag in drawing on my local mac machine. So i have kept 
>> it in initialization logic and i can see that mouse moves to (0, 0) and 
>> stays there until the test is completed. Next operation we do after this 
>> mouseMove is screenCapture. So keeping autodelay here or later should not 
>> matter.
>> 
>> I have moved whole logic under try block and disposing frame under finally 
>> block.
>
> SInce some coordinates are changed, can you please confirm the test can be 
> used still to confirm JDK-6664068 meaning it fails before the fix and pass 
> with it?

I have changed the frame/image coordinates to use bounds from screen 
coordinates instead of using constants (128,128) for BufferedImage and 
(512,512) for VolatileImage, it doesn't change the way the test works.

-------------

PR: https://git.openjdk.org/jdk/pull/11158

Reply via email to