<Swing Dev> RFR: 8198400: Test javax/swing/SwingUtilities/6797139/bug6797139.java fails in mach5

[Sergey Bylokhov]

On Sun, 18 Oct 2020 15:55:13 GMT, Prasanta Sadhukhan <psadhukhan at openjdk.org> wrote:

>> if you run it locally with -Djava.awt.headless=true what happens ?
>> And in the bug report I see
>> Caused by: java.lang.NullPointerException
>> at java.desktop/javax.swing.plaf.basic.LazyActionMap.installLazyActionMap(LazyActionMap.java:64)
>> 
>> So installUI failed which is why there's no componentUI but why did we get this NPE that started it ?
>
> Running -Djava.awt.headless=true passes locally.
> I tried running on mach5 in headless configuration also and it passes even on running several iterations. Link in JBS.
> It seems last time it gets NPE when executing UIManager.getLookAndFeelDefaults().put(defaultsKey, map); which indicate
> getLookAndFeelDefaults() returns null. But there are many instances in jdk swing code where no null check is done for
> this getLookAndFeelDefaults() call and it is working fine.

I think is that the root cause is in the stack trace "SameVMRunnable". It means that the test was run in the samevm
mode and was affected by the other tests. I suggest finding such traces in the JBS and de-problem list all of them.
Similar issue: https://github.com/openjdk/jdk/pull/715/files. The bug was fixed by the adding client tests to the
"othervm.dirs" in the TEST.ROOT

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

PR: https://git.openjdk.java.net/jdk/pull/719