On Wed, 27 Oct 2021 00:52:10 GMT, Alexander Zuev <kiz...@openjdk.org> wrote:
> The problem here is that due to the approximation of the coordinates when > converting > to and from screen cordinate system on screen with 125% magnification > two things can happen: > 1. New mouse position calculated can be 149/149 which is the same as the > starting position > and system can refuse to generate new mouse event which causes false positive > of the test > since second mouse motion listener was never called; > 2. Even if event is generated the resulting reported coordinates can differ > by 1 pixel > in any of the directions; > > To solve it i moved away the target point to make sure that move event always > generated and > comparing expected and actual coordinates with approximation of +/- 1 pixel. I have tried to run this test on macOS as a standalone code, and it hangs. The second frame is never disposed and still visible on the screen, not sure why. ------------- PR: https://git.openjdk.java.net/jdk/pull/6128
