On Thu, 23 May 2024 08:08:17 GMT, Alexander Zuev <kiz...@openjdk.org> wrote:
>> Ok. >> Is it possible to add a test to verify this behavior? > > The test would be manual and it is hard to even describe what is the result > of the test should look or sound like since we do not provide any string > messages, VoiceOver does it and we can not tell ahead how it will be put. And > for multilivel trees i am just running the SwingSet2 JTree demo and see that > VoiceOver correctly handles full/partial collapse and expand - that after > that there are no anomalies like VO output does not correspond to where > cursor actually positioned. Ok. I was thinking of something like when nodes are collapsed or expanded then can we log some output like Children are valid or invalid. Otherwise I have tested the fix with swingset2 demo and the test mentioned in [JDK-8320664](https://bugs.openjdk.org/browse/JDK-8320664) which behaves fine with the fix. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/19255#discussion_r1611480453
