Hi Sergey,
Here is my understanding:
When a SequencedEvent, detects that it is not the first event in the list, it
tries to dispatch events to ensure that the event that is before it gets
processed first. All SequencedEvents are created with the ID -
"java.awt.event.FocusEvent.FOCUS_LAST + 1".
Now, when it tries to dispatch, it calls the pumpEvents as below:
edt.pumpEvents(SentEvent.ID, new Conditional() {
public boolean evaluate() {
return
!SequencedEvent.this.isFirstOrDisposed();
}
});
The SentEvent.ID value is "java.awt.event.FocusEvent.FOCUS_LAST+2", and when in
the function pumpOneEventForFilters, it tries to call
EventQueue.getNextEvent(id), it looks for SentEvent.ID, which is not present,
and then the queue blocks in the hope that someone will post that event. This
is what is leading to the hang.
Ideally, the code should have pumped events, with its own id, since all the
sequenced events are created with that id. When I made this change and tested,
it did not hang.
However, if sequenced events are not posted immediately one after the other,
even this would fail, since it would expect that the next event that is
scheduled should be of SequencedEvent type.
I have removed the "id" parameter in pumpEvents, which effectively means to
process all events till the current SequencedEvent becomes the first one in the
list, at which point, it would process itself.
@Semyon,
I don't think the problem is with multiple AppContexts. I have modified the
test case to create a second AppContext, and post few events in that context.
In that case, even Java8 was able to run to completion without any problems.
Attached the test case for your reference.
Let me know if there is a problem with the testcase, or we can exercise the
multiple AppContexts in a better way.
Thanks,
Krishna
Message: 3
Date: Thu, 15 Feb 2018 14:30:03 -0800
From: Sergey Bylokhov <sergey.bylok...@oracle.com>
To: awt-dev@openjdk.java.net
Subject: Re: <AWT Dev> [11][JDK-8152974] AWT hang occurs when
sequenced events arrive out of sequence
Message-ID: <908bf597-c40c-977e-cde4-258c39bbf...@oracle.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Hi, Krishna.
Can you please provide more details why the test hang, it post two
SequencedEvents out of order, the first waits when the second will be
dispatched, why it hangs forever?
On 15/02/2018 04:16, Krishna Addepalli wrote:
> Hi All,
>
> Please review a fix for ?JDK-8152974:
> https://bugs.openjdk.java.net/browse/JDK-8152974
>
> Webrev: http://cr.openjdk.java.net/~kaddepalli/8152974/webrev00/
>
> The problem is when the events arrive out of sequence, the sequenced
> event?s dispatch loop assumes that the sequenced events are pushed
> into the queue together, which may not be the case, and this blocks
> the event queue. The fix is to continue to process all the events,
> which ultimately leads to the processing of the SequencedEvents in
> their order of creation.
>
> Thanks,
>
> Krishna
>
--
Best regards, Sergey.
Message: 5
Date: Thu, 15 Feb 2018 18:09:11 -0800
From: Semyon Sadetsky <semyon.sadet...@oracle.com>
To: Krishna Addepalli <krishna.addepa...@oracle.com>,
awt-dev@openjdk.java.net
Subject: Re: <AWT Dev> [11][JDK-8152974] AWT hang occurs when
sequenced events arrive out of sequence
Message-ID: <32533ac8-0007-668f-f691-fab84a8fa...@oracle.com>
Content-Type: text/plain; charset="windows-1252"; Format="flowed"
Hi Krishna,
Unlikely this will work for several app contexts.
You need to find the root cause why the sequenced events get mixed up
and fix it.
--Semyon
On 02/15/2018 04:16 AM, Krishna Addepalli wrote:
>
> Hi All,
>
> Please review a fix for ?JDK-8152974:
> https://bugs.openjdk.java.net/browse/JDK-8152974
>
> Webrev: http://cr.openjdk.java.net/~kaddepalli/8152974/webrev00/
> <http://cr.openjdk.java.net/%7Ekaddepalli/8152974/webrev00/>
>
> The problem is when the events arrive out of sequence, the sequenced
> event?s dispatch loop assumes that the sequenced events are pushed
> into the queue together, which may not be the case, and this blocks
> the event queue. The fix is to continue to process all the events,
> which ultimately leads to the processing of the SequencedEvents in
> their order of creation.
>
> Thanks,
>
> Krishna
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.openjdk.java.net/pipermail/awt-dev/attachments/20180215/a3b04c00/attachment.html>
End of awt-dev Digest, Vol 130, Issue 11
****************************************
SequencedEventTest.java
Description: Binary data
