RFR: 8276058: Some swing test fails on specific CI macos system [v2]

[Phil Race]

On Thu, 28 Oct 2021 21:35:52 GMT, Sergey Bylokhov <serb at openjdk.org> wrote:

> > Sorry, not following .. do you mean you think the cursor is captured when it is enlarged ? Like somehow a software cursor is used that is painted in a different way so that it always show up ? But the screen captures in the bug report seem to show a small cursor and the above A11Y setting on the test system has Cursor Size: Normal (ie the very minimum of the slider)
> 
> I can confirm that the changing the size of the cursor on that setting page can change the visibility of the cursor on the screenshot. If the cursor is "small" then it is completely invisible, if the cursor size is changed above some threshold it became visible on the screenshot. So it might be a case when on that system the default cursor is bigger that the threshold.

Hmm .. well it is at the minimum right now. So throw that one in the bag of things that can go wrong .. but I'm not currently seeing it set to anything except absolute minimum,
But we could change it manually and see if that toggles whether the unmodified tests pass or fail .. 

The mechanism by which any of these things suddenly happen and then suddenly go away is really unclear to me ..
but at least it makes sense that these can persist across reboots

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

PR: https://git.openjdk.java.net/jdk/pull/6140