focus issue with popups

[Glen Schrader]

I was hoping that this commit
http://hg.openjdk.java.net/caciocavallo/ng/rev/d8cb05675875 would have
fixed my issue but it appears this test case is still failing.  Any
thoughts?

On Thu, Sep 27, 2012 at 11:19 AM, Glen Schrader <gschrader at gmail.com> wrote:
> realized I never included the mailing list, just forwarding this so I
> have a public link to put in our issue tracker
>
>
> ---------- Forwarded message ----------
> From: Glen Schrader <gschrader at gmail.com>
> Date: Tue, Sep 25, 2012 at 12:56 PM
> Subject: Re: focus issue with popups
> To: Roman Kennke <roman at kennke.org>
>
>
> On Mon, Sep 24, 2012 at 4:37 AM, Roman Kennke <roman at kennke.org> wrote:
>>
>> Hi Glen,
>>
>> Yes, the issues that you mentioned might be causing the problems. I will
>> look into it soon. I wonder why it's only sporadic, this seems to point
>> to some threading/EDT issue. Have you made sure that all your
>> interactions happen in the EDT? (see http://fest.codehaus.org/Writing
>> +EDT-safe+GUI+tests) In the meantime it would be great if you can come
>> up with a testcase.
>>
>> Roman
>>
>
> Yeah the sporadic nature sure does suggest EDT violations as we've hit
> those before and have had weird painting issues. I doubled checked the
> code, everything is accessed through fixtures which enforces things
> are run on the EDT, also we use the FestSwingJUnitTestCase class which
> installs the FailOnThreadViolationRepaintManager which catches a lot
> of EDT violations.  So I'm pretty confident the code is ok.
>
> I narrowed it down to a tooltip that is appearing when the test fails.
>  I think this is why it only fails some of the time, the tooltip
> doesn't always appear (not sure why that would be the case, there must
> be some timing difference somewhere).  I was able to come up with a
> test case (attached) which will thankfully fail consistently under the
> cacio runner.  This test will pass when I comment out the
> aforementioned call in CacioWindowPeer to the setFocusableWindowState
> in the constructor.  I don't know what other affects that will have
> though as this must have been put in for a reason.
>
> Let me know if I can help further.
>
> Thanks,
> Glen