On Thu, 21 Mar 2024 01:16:03 GMT, Harshitha Onkar <hon...@openjdk.org> wrote:
>> I think we should update the instructions. >> >> A high resolution mouse produces events like this: >> >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(164,90),absolute(1135,365),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=0,preciseWheelRotation=0.25] >> on panel0 >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(164,90),absolute(1135,365),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=0,preciseWheelRotation=0.25] >> on panel0 >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(164,90),absolute(1135,365),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=0,preciseWheelRotation=0.25] >> on panel0 >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(133,46),absolute(1104,321),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=1,preciseWheelRotation=1.0] >> on panel0 >> >> >> Thus, `wheelRotation=0` while `preciseWheelRotation=0.25`. When it reaches, >> `preciseWheelRotation=1.0`, the event contains **`wheelRotation=1`**. >> >> A regular mouse with mouse-wheel notches produces the following events: >> >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(90,115),absolute(1061,390),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=1,preciseWheelRotation=1.0] >> on panel0 >> java.awt.event.MouseWheelEvent[MOUSE_WHEEL,(176,76),absolute(1147,351),button=0,clickCount=0,scrollType=WHEEL_UNIT_SCROLL,scrollAmount=3,wheelRotation=1,preciseWheelRotation=1.0] >> on panel0 >> >> >> As you can see, `scrollType=WHEEL_UNIT_SCROLL` is the same in both cases. >> The only difference is that high-resolution mouse produces events where >> **`preciseWheelRotation < 1.0`**. >> >> This means that you should also track the value of `preciseWheelRotation`. >> Otherwise, the test will pass when a regular mouse is used. >> >> You may show a warning to the user if you detect, let's say, 5 events with >> `wheelRotation=1` but none of the events had `wheelRotation=0` and >> `preciseWheelRotation < 1.0`. > > @aivanov-jdk Since preciseWheelRotation amount depends on user's movement of > the high-res mouse and there might be a need to look through logs for few > events, it is better to convert this test from semi-automated to manual. Does > this sound good? > Updated the PR with the latest changes. @honkar-jdk I think the test could still remain semi-automatic. Can the test verify the conditions? I see no reason why it can't… yet it requires more code. It can be postponed to a later time, if necessary. I don't think we should make the tester read the log and verify the conditions which can be verified automatically. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/18312#discussion_r1533668635
