<Swing Dev> [13] RFR 8194944:Regression automated test 'open/test/jdk/javax/swing/JInternalFrame/8145896/TestJInternalFrameMaximize.java' fails

[Ambarish Rapte]

Hi Tejpal,

Here are few comments,

Update copyright year to include 2019
Add bug id 8145896 to @bug
Add a new line before imports begin, line #33
point should be volatile, line #58 [Won’t be needed, if changes are made to address comment #7]
Add space after try, line #66
The empty lines #75, #151 can be removed.
In blockTillDisplayed(): If frame.getLocationOnScreen() throws an exception then it won't get caught.

So if frame.getLocationOnScreen() throws an exception, the test will fail, but instead the test should sleep and try again.

The blockTillDisplayed() method defined in test bug6427244.java looks more correct, you can use that method here.

 

Regards,

Ambarish

 

From: Prasanta Sadhukhan 
Sent: Thursday, July 25, 2019 3:43 PM
To: Tejpal Rebari <tejpal.rebari at oracle.com>
Cc: swing-dev at openjdk.java.net
Subject: Re: <Swing Dev> [13] RFR 8194944:Regression automated test 'open/test/jdk/javax/swing/JInternalFrame/8145896/TestJInternalFrameMaximize.java' fails

 

+1

Regards
Prasanta 

On 25-Jul-19 3:40 PM, Tejpal Rebari wrote:

Hi All,

I have verified that the test fails without the changes of the fix 8145896. 

 

Regards

Tejpal

On 25-Jul-2019, at 12:40 PM, Prasanta Sadhukhan <HYPERLINK "mailto:prasanta.sadhukhan at oracle.com"prasanta.sadhukhan at oracle.com> wrote:

 

OK. Last thing is, please confirm if the modified test still fails if we remove 8145896 fix [http://hg.openjdk.java.net/jdk9/jdk9/jdk/rev/ef8fd276efe1].

Regards
Prasanta

On 25-Jul-19 12:33 PM, Tejpal Rebari wrote:

Hi Prasanta,

Agreed that the two points should be different, so retained the point variable for executeTest().

 

Here is the latest changes. 

webrev : http://cr.openjdk.java.net/~trebari/swing/8194944/webrev2/

 

Thanks and regards

Tejpal





On 22-Jul-2019, at 4:08 PM, Prasanta Sadhukhan <HYPERLINK "mailto:prasanta.sadhukhan at oracle.com"prasanta.sadhukhan at oracle.com> wrote:

 

Looks better but I guess frame's locationOnScreen point would be different from Menu's centrepoint so we should not use same "point" instance variable for both. Guess local "point" variable for executeTest() should be retained.

Regards
Prasanta

On 19-Jul-19 1:11 PM, Tejpal Rebari wrote:

Hi All, 

Added the blockTillDisplayed() method instead of 1000ms delay.

 

here is the latest changes.

webrev : http://cr.openjdk.java.net/~pkbalakr/Tejpal/8194944/webrev1

 

Thanks and regards

Tejpal Rebari

 





On 08-Jul-2019, at 1:43 PM, Prasanta Sadhukhan <HYPERLINK "mailto:prasanta.sadhukhan at oracle.com"prasanta.sadhukhan at oracle.com> wrote:

 

Adding delay is indeterministic.  I guess it will be better if you use blockTillDisplayed() method which you can find in many of the testcase(s) like javax/swing/JComboBox/8182031/ComboPopupTest.java, javax/swing/JEditorPane/6917744/bug6917744.java

Regards
Prasanta 

On 08-Jul-19 1:02 PM, Tejpal Rebari wrote:

Hi All, 

Please review a test fix for the issue where the test fails on Mac  and Linux intermittently.

 

The test fails because of “component must be showing on the screen to determine its location” error.

Which is because the test tries to get the location of the component and the component is not displayed on the screen yet.

 

The test tries to get the location of menu in the executeTest method so By adding some delay before the executeTest call solves the problem.

 

Tested in our internal build and test system and test passes on all the three Platforms.

 

Bug             : https://bugs.openjdk.java.net/browse/JDK-8194944

webrev : http://cr.openjdk.java.net/~pkbalakr/Tejpal/8194944/webrev0

 

Thanks and regards

Tejpal 

 

 

 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.java.net/pipermail/swing-dev/attachments/20190801/935d6704/attachment.html>