<AWT Dev> [14] RFR JDK-8234385:[TESTBUG] java/awt/EventQueue/6980209/bug6980209.java fails in linux nightly

[Sergey Bylokhov]

On 11/20/19 1:41 am, Prasanta Sadhukhan wrote:
> Hi Sergey,
> 
> I have ran for 50 iteration per platform with my change and mach5 is green (link is in JBS).

And it will be red w/o the fix?

> 
> Regards
> 
> Prasanta
> 
> On 19-Nov-19 11:38 PM, Sergey Bylokhov wrote:
>> Hi, Prasanta.
>>
>> I have checked the logs for this test on different platforms and it seems we have 3 issues here:
>>  - RealSync does not work properly and produce next exception:
>>     sun.awt.SunToolkit$OperationTimedOut: 20000
>>     at java.desktop/sun.awt.X11.XToolkit.syncNativeQueue(XToolkit.java:2645)
>>     at java.desktop/sun.awt.SunToolkit.realSync(SunToolkit.java:1470)
>>     at java.desktop/sun.awt.SunToolkit.realSync(SunToolkit.java:1402)
>>     at java.desktop/java.awt.Robot.waitForIdle(Robot.java:675)
>>     at bug6980209.main(bug6980209.java:72)
>>     at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
>>     at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
>>     at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
>>     at java.base/java.lang.reflect.Method.invoke(Method.java:564)
>>     at com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:127)
>>     at java.base/java.lang.Thread.run(Thread.java:830)
>>
>>  - Exception reported in JDK-8198615, "It seems the secondary loop will never end"
>>  - Timeout in the realSync(), is it possible that we need to increase the default timeout, or
>>    we have a bug in realSync()?
>>
>> BTW to check the intermittent test you need to run it more than 3 times(I guess 50 pre platform is fine).
>>
>>
>> On 11/19/19 1:42 am, Prasanta Sadhukhan wrote:
>>> Hi All,
>>>
>>> Please review a testbug fix for an issue seen intermittently in linux nightly.
>>>
>>> Issue is that the test times out after the test finishes without actually propagating the test results to jtreg harness. It seems like the JFrame which is created is called with EXIT_ON_CLOSE which will call System.exit(0) causing the vm to terminate and if the test is not run with othervm option, it might create some issue when run in a bunch, as it is done in nightly testing.
>>>
>>> Proposed fix is to remove the EXIT_ON_CLOSE option as we already have frame.dispose() call. mach5 run is green.
>>>
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8234385
>>>
>>> diff -r 9db62a092725 test/jdk/java/awt/EventQueue/6980209/bug6980209.java
>>> --- a/test/jdk/java/awt/EventQueue/6980209/bug6980209.java Thu Nov 14 10:59:45 2019 +0530
>>> +++ b/test/jdk/java/awt/EventQueue/6980209/bug6980209.java Tue Nov 19 15:06:03 2019 +0530
>>> @@ -63,7 +63,6 @@
>>>                   public void run() {
>>>                       frame = new JFrame();
>>>                       frame.setUndecorated(true);
>>> *- frame.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);*
>>>                       setup(frame);
>>>                   }
>>>               });
>>>
>>> Regards
>>>
>>> Prasanta
>>>
>>
>>


-- 
Best regards, Sergey.