On Tue, 23 Sep 2025 18:24:52 GMT, Phil Race <[email protected]> wrote:
>> This test would fail with Xrender. But it did odd things like sleep in the >> paint method. >> When I fixed that it started to pass every time. >> It also had a problem listing for macOS supposedly for edge pixels. >> I didn't see that myself but I changed it to not sample the edges. > > Phil Race has updated the pull request incrementally with one additional > commit since the last revision: > > 8221451 It seems this patch bypasses the main purpose of the test. When something is rendered in the paint method to the graphics of a window/frame, it should become visible on the screen in the current paint_frame. After the patch, it is unclear which repaint operation is actually being tested. It is possible that the test initially renders garbage and only later refreshes the frame to display correct data. ------------- PR Comment: https://git.openjdk.org/jdk/pull/27400#issuecomment-3331105431
