RFR: 8307311: Timeouts on one macOS 12.6.1 host of two Swing JTableHeader tests

[Prasanta Sadhukhan]

On Fri, 19 May 2023 11:18:16 GMT, Alexander Zvegintsev <azvegint at openjdk.org> wrote:

>> Issue was bug6889007.java was causing a timeout issue on mac system.  Although bug6889007.java fail with small test group with "Wrong cursor type" error, it could not reproduce bug6884066.java 
>> The hang issue might be due to waitForIdle() being called in loop which will be calling sync() so replaced with delay..
>> 
>> Modified fix passed the suggested small test group `desktop_timeout ` several times in all platforms 
>> in addition to several iterations of the test run in standalone mode in all platforms..
>
> test/jdk/javax/swing/JTableHeader/6889007/bug6889007.java line 72:
> 
>> 70:                 frame.pack();
>> 71:                 frame.setLocationRelativeTo(null);
>> 72:                 frame.setAlwaysOnTop(true);
> 
> Why do we need to keep the window on top?

In windows primarily, I see sometimes the frame is rendered below some if there are other windows present ie remains hidden and mouse cursor is moved which looks odd, so I ensured the frame is always on top for the test

> test/jdk/javax/swing/JTableHeader/6889007/bug6889007.java line 87:
> 
>> 85:             for(int i = -shift; i < width + 2*shift; i++) {
>> 86:                 robot.mouseMove(x++, y);
>> 87:                 robot.delay(100);
> 
> It looks like we don't need this delay since there is `robot.setAutoDelay(100);` on line 56. A total delay of 200ms seems too much to me.

I dont think it's too much as compared to 500ms we have in some tests..Since I have tested in all platforms with this values without any problems observed, I will like to keep this value for now..

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

PR Review Comment: https://git.openjdk.org/jdk/pull/14026#discussion_r1198855883
PR Review Comment: https://git.openjdk.org/jdk/pull/14026#discussion_r1198855919