On Wed, 16 Sep 2026 12:16:43 GMT, Jayathirth D V <[email protected]> wrote:

> On macOS this test fails all the time because of rounded corners but we check 
> for pixels even outside the rounder corners.
> This test can also fail intermittently for edge pixels on HiDPI displays or 
> because robot screen capture downsamples from absolute fractional scaled data 
> on platforms like windows.
> 
> We are now using insets to avoid rounded corner issue, test is run with 
> uiScale = 1 and we are excluding some edge pixels. Test is also updated to 
> dump an image on failure for better debugging purpose.
> 
> On Linux we actually have blitting issue when we have source clipping and for 
> that i have created a separate issue 
> [JDK-8392471](https://bugs.openjdk.org/browse/JDK-8392471). Test is also 
> verified in CI.
> 
> 
> 
> ---------
> - [x] I confirm that I make this contribution in accordance with the [OpenJDK 
> Interim AI Policy](https://openjdk.org/legal/ai).

test/jdk/sun/java2d/SunGraphics2D/SourceClippingBlitTest/SourceClippingBlitTest.java
 line 30:

> 28:  * @summary Tests that source is clipped correctly when blitting
> 29:  * different types of images to the screen
> 30:  * @run main/othervm -Dsun.java2d.uiScale=1 SourceClippingBlitTest

Why do you need this parameter? I understand that it can cause some issues for 
some tests, but you know, we cannot change all the tests here and there to 
exclude the hidpi/transform code path which is default on most of the systems. 
If the problem is in the downlscaling caused by the robot api, it might be 
better to re-implement that part of the test via 
robot.createMultiResolutionScreenCapture?

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

PR Review Comment: https://git.openjdk.org/jdk/pull/32904#discussion_r4031159468

Reply via email to