<Swing Dev> RFR: 8211999: Window positioning bugs due to overlapping GraphicsDevice bounds (Windows/HiDPI)

[Sergey Bylokhov]

On Thu, 29 Oct 2020 17:55:36 GMT, Alexander Zuev <kizune at openjdk.org> wrote:

>> @prrace Can you review this?
>
> I'm trying to test these changes but i have weird problems with mouse event coordinate translation. For example - i have two displays side by side with some vertical shift and with different scale factor. After this fix if the window spans across multiple monitors when i move cursor over the part on the secondary monitor related to the main window position mouse events got sent to the wrong component. Look at the attached image - by the position of the tooltip you can see that the mouse is hovering over the menu bar area - namely over "Options" menu, yet the tooltip is from the JProgressBar demo and if you look at the toolbar you will see JProgressBar demo button highlighted as if mouse is on top of it. And if i click instead of selecting the menu the button got pressed. Not sure why it happens but seems like it was not happened before the fix.
> ![events](https://user-images.githubusercontent.com/69642324/97612555-cf808680-19d4-11eb-9707-9eeb22b196d0.PNG)

You found this bug https://bugs.openjdk.java.net/browse/JDK-8249164 the case when the window is split across two screens with different DPI are not properly supported before and after the fix, in both cases, there are values which break it. But I need to confirm that, please provide exact display resolution/java pipeline/and scales.
BTW this is the case why thr JUMP TO machinery was added to the "awt_Window.cpp"

-------------

PR: https://git.openjdk.java.net/jdk/pull/375