On Thu, 18 Sep 2025 09:46:16 GMT, Thiago Milczarek Sayao <[email protected]> wrote:
>>> I've been going through the review and tests here, I still have some more >>> code to review but I did notice there are some test failures on Windows: >> >> I've tracked down one issue with these tests. Some of them try to verify >> where the top stage is by sampling from a fixed location in screen >> coordinates, typically calling `getColor(100, 100)`. Depending on the >> platform and stage style this might sample a pixel from the Label in the >> middle of the top stage's scene. This happens on Windows when the top >> stage's style is EXTENDED. This doesn't happen on macOS because the top >> stage gets shifted down to avoid the menu bar (the test tries to position >> the top stage at y = 0 but it ends up at y = 25 instead). > >> > I've been going through the review and tests here, I still have some more >> > code to review but I did notice there are some test failures on Windows: >> >> I've tracked down one issue with these tests. Some of them try to verify >> where the top stage is by sampling from a fixed location in screen >> coordinates, typically calling `getColor(100, 100)`. Depending on the >> platform and stage style this might sample a pixel from the Label in the >> middle of the top stage's scene. This happens on Windows when the top >> stage's style is EXTENDED. This doesn't happen on macOS because the top >> stage gets shifted down to avoid the menu bar (the test tries to position >> the top stage at y = 0 but it ends up at y = 25 instead). > > Moved the label to the bottom right. Another cause of this "shifting" > behavior are desktop panels. @tsayao Please merge the latest master to resolve the (trivial) merge conflict with #1924 ------------- PR Comment: https://git.openjdk.org/jfx/pull/1789#issuecomment-3407691079
