At first, I noticed bug4490179.java didn't release mouse button 1, which could affect the following tests.
I started looking at the test and updated it to use `CountDownLatch` instead of delays. @azvegint handled releasing mouse button 1 in https://github.com/openjdk/jdk/pull/19339#discussion_r1609201116. The updated version first verifies `mouseButton3Released` event is received and ensures `actionPerformed` is not received at this time. Then mouse button is released; this should trigger two events: `mouseButton1Released` and `actionPerformed`. It is verified that these events are received. I kept debugging prints, they help understanding which actions lead to which events. The updated test reproduces the original, it is easily verified using Java 8; the updated test is stable in CI. The `javax/swing/JButton/bug4490179.java` test is also modified in https://github.com/openjdk/jdk/pull/20861, @mrserb, could you take a look? I was going to suggested factoring out the changes to this test from #20861, yet I modified it myself. I realised it when I started creating the PR. ------------- Commit messages: - 8341978: Improve JButton/bug4490179.java Changes: https://git.openjdk.org/jdk/pull/21474/files Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=21474&range=00 Issue: https://bugs.openjdk.org/browse/JDK-8341978 Stats: 88 lines in 1 file changed: 66 ins; 3 del; 19 mod Patch: https://git.openjdk.org/jdk/pull/21474.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/21474/head:pull/21474 PR: https://git.openjdk.org/jdk/pull/21474
