<AWT Dev> Review request for 8048171: Test java/awt/Mixing/AWT_Mixing/OpaqueOverlappingChoice.java fails on solaris, linux but passes on windows
Semyon Sadetsky
semyon.sadetsky at oracle.com
Thu Nov 5 12:54:21 UTC 2015
Hi Ambarish,
I have double-checked the fix, but again I got it failed for jdk8 and
jdk9-b88 on Ubuntu 14.04 LTS. See what I get:
Testing Choice
The test failed: Opacity test mismatchs
The test failed: Opacity test mismatchs
The test is over, hit Ctl-C to stop Java VM
The test is over, hit Ctl-C to stop Java VM
STDERR:
do not skipTestingEmbeddedFrame
Expected: 010000101
Observed: 001000000001001
java.lang.RuntimeException: Opacity test mismatchs
What OS version do you use for testing?
Since you found a new bug in Choice component it would be perfect to
have it JIRA.
--Semyon
On 11/3/2015 4:07 PM, Ambarish Rapte wrote:
>
> Hi Semyon,
>
> The issue with Choice which I mentioned, is my observation for the
> test failure. There is no bug reported.
>
> Regarding the test failure, because mouseClicked() was not getting
> called at all.
>
> Without the fix, the test was failing with below log,
>
> Expected: 010000101
>
> Observed:
>
> java.lang.RuntimeException: Opacity test mismatchs
>
> Reason: mouseClicked() was not getting called at all for Choice, hence
> ‘testSeq’ variable was not getting updated.
>
> But As per your observation, mouseClicked or mousePressed is getting
> called twice.
>
> There is a difference in behavior for event.
>
> The cause of this difference in behavior should be identified before
> proceeding further.
>
> Please provide any inputs if possible.
>
> Many Thanks,
>
> Ambarish
>
> *From:*Semyon Sadetsky
> *Sent:* Tuesday, November 03, 2015 3:18 PM
> *To:* Ambarish Rapte; Prasanta Sadhukhan; awt-dev at openjdk.java.net
> *Subject:* Re: Review request for 8048171: Test
> java/awt/Mixing/AWT_Mixing/OpaqueOverlappingChoice.java fails on
> solaris, linux but passes on windows
>
> Hi Ambarish,
>
> Thank you for the clarification. But on my Ubuntu 14.04 test fails
> before and after the fix due to the same reason: mouse click or press
> is fired twice. Could you send me link to the Choice issue you've
> mentioned?
>
> On 11/2/2015 5:31 PM, Ambarish Rapte wrote:
>
> Hi Semyon,
>
> OpaqueOverlapping :: performTest() function performs the test by
> generating mouse click event using,
>
> mouse press & mouse release events.
>
> But choice receives only mouse Press and mouse release event &
> does not receive the mouse clicked event.
>
> The test already does an adjustment for this, please refer line 148.
>
> There is one more possibility that the test fails, If JTwork is
> not cleared before executing fix.
>
> Please try as,
>
> 1.Run original test using jtreg => Test FAILs
>
> 2.Make the fix change.
>
> 3.Delete the JTwork folder
>
> 4.Run the test again using jtreg => Test PASSes.
>
> If we skip the step 3, the test would fail.
>
> Request you to try as above.
>
> Many Thanks,
>
> Ambarish
>
> *From:*Semyon Sadetsky
> *Sent:* Monday, November 02, 2015 2:50 PM
> *To:* Ambarish Rapte; Prasanta Sadhukhan; awt-dev at openjdk.java.net
> <mailto:awt-dev at openjdk.java.net>
> *Subject:* Re: Review request for 8048171: Test
> java/awt/Mixing/AWT_Mixing/OpaqueOverlappingChoice.java fails on
> solaris, linux but passes on windows
>
> Hi Ambarish,
>
> But I did not touch the mouse.
> Usually mouse events come in the next sequence MOUSE_PRESSED,
> MOUSE_RELEASED,MOUSE_CLICKED. Why do you think that
> mouseClicked()->mousePressed() replacement should change something
> in the test?
>
> --Semyon
>
> On 11/2/2015 8:51 AM, Ambarish Rapte wrote:
>
> Dear Semyon,
>
> This test uses the mouse for input.
>
> While the test is running with or without fix, mouse should
> not be disturbed.
>
> ------------------------------
>
> guess we should not touch mouse while test in in progress, in
> that case it passed.
> Fix looks fine to me.
>
> Regards
> Prasanta
>
> -----------------------------
>
> Prasanta also has verified similar way, that the test passes.
>
> Kindly request you to try again.
>
> Also If possible, please share the execution log of failure
> for me to verify.
>
> Many Thanks,
>
> Ambarish
>
> *From:*Semyon Sadetsky
> *Sent:* Friday, October 30, 2015 10:07 PM
> *To:* Ambarish Rapte; Prasanta Sadhukhan;
> awt-dev at openjdk.java.net <mailto:awt-dev at openjdk.java.net>
> *Subject:* Re: Review request for 8048171: Test
> java/awt/Mixing/AWT_Mixing/OpaqueOverlappingChoice.java fails
> on solaris, linux but passes on windows
>
> Hi Ambarish,
>
> After your fix applying the test still fails.
>
> --Semyon
>
> On 10/29/2015 5:00 PM, Ambarish Rapte wrote:
>
> Dear All,
>
> Kindly review the fix for JDK9.
>
> Bug:
> https://bugs.openjdk.java.net/browse/JDK-8048171
>
> Webrev:
> http://cr.openjdk.java.net/~rchamyal/ambarish/8048171/webrev.00/
> <http://cr.openjdk.java.net/%7Erchamyal/ambarish/8048171/webrev.00/>
>
> Issue:
>
> This is an Opaque overlapping test for
> each AWT component.
>
> To test this functionality, a MouseAdapter
> is registered as mouseListener with overridden
> mouseClicked() .
>
> Test fails, due to issue with Event on
> Linux/Solaris, that mouseClicked event is not generated
> correctly to Choice.
>
> Fix:
>
> This is an overlapping test and not a test
> for mouseCicked() for Choice.
>
> The test simulates mouseClick event by
> sequentially passing mousePress & mouseRelease events.
>
> Overriding mousePressed() instead of
> mouseClicked() for the registered mouse adapter solves the
> issue,
>
> without causing any side effect in test execution on Windows.
>
> Fix Verification:
>
> Verified successful execution of below
> tests using jtreg,
>
> java/awt/Mixing/AWT_Mixing/OpaqueOverlapping.java
>
> java/awt/Mixing/AWT_Mixing/OpaqueOverlappingChoice.java
>
> Many Thanks,
>
> Ambarish Rapte
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/awt-dev/attachments/20151105/5564f83a/attachment-0001.html>
More information about the awt-dev
mailing list