RFR: 6829250: Reg test: java/awt/Toolkit/ScreenInsetsTest/ScreenInsetsTest.java fails in Windows

[Alexander Zuev]

On Sat, 14 May 2022 00:53:34 GMT, Sergey Bylokhov <serb at openjdk.org> wrote:

> That a mistake to consider that as a correct behavior even based on the amount of bugs we have. The MAXIMIZED_BOTH property was considered as a "maximize" action, and should allow such behavior for the custom application when the decorations are rendered by the app. You can check all that bugs and see how many people report it and then accept, for example this one: https://bugs.openjdk.java.net/browse/JDK-4976497 which shows that even Swing is broken if setDefaultLookAndFeelDecorated is used.
> 
> Note that this test is not for the borderless window which should cover the full screen but specifically for the MAXIMIZED_BOTH action which expected to maximize the window in the same way as for decorated frame. And I think this is how it works on other platforms since the test passed there. This is why this test created 20 years ago was not changed, it was expected that we fix the JDK.
> 
> I suggest to rollback this and apply the fix to JDK.

You probably meant https://bugs.openjdk.java.net/browse/JDK-4737788 because 4976497 is closed as a duplicate of it.

And i still think it is debatable if this behavior constitutes a bug by itself. As pointed out in 2004:

"I think, there is no problem in covering taskbar after
maximizing for undecorated frames. For example, Internet Explorer covers
taskbar after pressing F11. I think, this is a normal behaviour.
But may be we should modify documentation to describe this."

So this is at beast a questionable feature that might not be implemented ever. I will not roll back this test change but i will ad the comment to the JDK-4737788 pointing out that change in behavior can warrant revisiting of this test should this bug be addressed at any point in time. There, i just did that.

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

PR: https://git.openjdk.java.net/jdk/pull/8314