Kevin Rushforth

As with the earlier PR, this will need a lot of testing and careful review. @Maran23 I'd be interested in your thoughts on this approach. /reviewers 2

This looks like a safe, targeted fix for the problem. It will need to be tested on all platforms with a full run of all system tests.

I get the following internal assertion failure when running the test on Linux: ``` NestedEventLoopTest > testCanExitAndThenEnterNewLoop FAILED java.lang.AssertionError: Internal inconsistency - wrong EventLoop at javafx.graphics@23-ea/com.sun.glass.ui.EventLoop.enter(EventLoop.java:108) at javafx.graphics@23-ea/com.sun.javafx.tk.quantum.QuantumToolkit.enterNestedEventLoop(QuantumToolkit.java:656) at javafx.graphics@23-ea/javafx.application.Platform.enterNestedEventLoop(Platform.java:310)...

Reviewers: @tiainen @kevinrushforth /reviewers 2

See [JDK-8273754](https://bugs.openjdk.org/browse/JDK-8273754). It affected JavaFX 17, and was fixed in 17.0.1 (and in a 17.0.0.1 patch).

Overall I quite like this enhancement. I'll focus my initial review on the public API. I'll then provide comments on the behavior of the new policies. Later I'll review the...

Behavioral issues: All the policies work much better than the existing (buggy) `CONSTRAINED_RESIZE_POLICY`. I did notice a few oddities, though. More manual testing is needed by a few folks. 1....

I reworded the title of the JBS Enhancement slightly to reflect that we are adding new policies (not just providing a single alternative to the existing policy). Can you update...

> > 1. The divider being dragged doesn't track the cursor. All policies other than CONSTRAINED_RESIZE_POLICY_ALL_COLUMNS should be able to track (the Swing implementation does). > > Could you describe...

The API changes to make the class abstract look good.