On Fri, 11 Oct 2024 19:35:51 GMT, Damon Nguyen <dngu...@openjdk.org> wrote:
> Intermittent failure of awt test in windows only. When re-testing locally, > the drag button wasn't actually drawn in the test. I made p static, added > stability delays, and moved robot out of the EDT (as well as caught a > duplicate robot instantiation). There's a merging issue that also missed > adding this test to the problem list for MacOS, so it's been added here. test/jdk/java/awt/dnd/DnDRemoveFocusOwnerCrashTest.java line 86: > 84: frame.setVisible(true); > 85: > 86: p = dragSourceButton.getLocationOnScreen(); Getting the location just after `setVisible` on EDT may give an inaccurate result(especially if the tested system is slow, and decided to move the window somewhere else). I suggest to wait as usual before getting the location robot.waitForIdle(); robot.delay(500); // or FRAME_ACTIVATION_TIMEOUT ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/21477#discussion_r1797366961