<Swing Dev> [12] JDK-8213071: [macos] The focus is disappear after pressing return to close the "That was a pretty good movie!" dialog.

shashidhara.veerabhadraiah at oracle.com shashidhara.veerabhadraiah at oracle.com
Wed Feb 20 06:11:45 UTC 2019


Hi Sergey, Thank you for the review and here is the new webrev fixing 
your comments:

http://cr.openjdk.java.net/~sveerabhadra/8213071/webrev.02/

Thanks and regards,

Shashi


On 20/02/19 6:40 AM, Sergey Bylokhov wrote:
> Hi, Shashi.
>
> Can you please unify the usage of NO/false/true for the 
> keyNotificationRecd property.
>
> On 22/11/2018 01:00, shashidhara.veerabhadraiah at oracle.com wrote:
>> Hi Sergey, Please find the updated webrev. It is true that the 
>> notifications are unique to each window and it should be a instance 
>> variable.
>>
>> http://cr.openjdk.java.net/~sveerabhadra/8213071/webrev.01/
>>
>> I have reproduced this issue on Prashanth's mac as well with JDK 8 
>> ver:121. Please note the behavior is different with respect to the 
>> voice over state is on or off.
>>
>> Thanks and regards,
>> Shashi
>>
>>
>> On 22/11/18 6:03 AM, Sergey Bylokhov wrote:
>>> On 21/11/2018 16:00, Sergey Bylokhov wrote:
>>>> Can you please double check it on a different systems,
>>>> I tried to reproduce it a few time on the current client and it 
>>>> seems to work.
>>>
>>> And the question about the current fix, are you sure that the static 
>>> boolean flag will
>>> work properly if the app will have a few windows? I guess it should 
>>> be perWindow flag?
>>>
>>>>
>>>> On 12/11/2018 22:28, shashidhara.veerabhadraiah at oracle.com wrote:
>>>>> Hi Sergey, This same behavior exists on jdk 7 and jdk 8 as well. 
>>>>> So it is not a regression I think.
>>>>>
>>>>> Thanks and regards,
>>>>> Shashi
>>>>>
>>>>>
>>>>> On 13/11/18 6:12 AM, Sergey Bylokhov wrote:
>>>>>> Hi, Shashi.
>>>>>>
>>>>>> Did you check this behavior was in jdk7/8 GA, or it is a regression?
>>>>>>
>>>>>>
>>>>>> On 11/11/2018 04:58, shashidhara.veerabhadraiah at oracle.com wrote:
>>>>>>> Hi All, Please review a fix for an accessibility vpat issue.
>>>>>>>
>>>>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8213071
>>>>>>>
>>>>>>> Webrev: http://cr.openjdk.java.net/~sveerabhadra/8213071/webrev.00/
>>>>>>>
>>>>>>> The behavior of the same actions to reproduce the problem, 
>>>>>>> produced different results with VO and without VO. It have 
>>>>>>> behaved right without the VO and with VO, the focus was lost.
>>>>>>>
>>>>>>> The issue is with window being the KEY and MAIN window per the 
>>>>>>> Mac windowing system. There are platform notifications like the 
>>>>>>> 'windowDidBecomeMain' and 'windowDidBecomeKey' are issued for 
>>>>>>> the window that is about to get the focus and key strokes. So 
>>>>>>> typically under normal conditions, the window which is about to 
>>>>>>> get notified of becoming the focusable window would get the KEY 
>>>>>>> and MAIN notifications in that order.
>>>>>>>
>>>>>>> But with VO, some how the KEY notification was not delivered to 
>>>>>>> java window and it only receives the MAIN notification. In this 
>>>>>>> scenario, though the isKeyWindow() is true for this window, KEY 
>>>>>>> notification was not delivered and only the MAIN notification is 
>>>>>>> delivered. We do transfer the focus under the KEY notification 
>>>>>>> handler which was not getting executed since it received no KEY 
>>>>>>> notification. Per me, isKeyWindow() and MAIN notification is 
>>>>>>> sufficient to transfer the focus and become a main window but I 
>>>>>>> have not disturbed the existing handlers for the notification 
>>>>>>> and fixed only for the case which is causing this issue.
>>>>>>>
>>>>>>> Thanks and regards,
>>>>>>> Shashi
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>>
>>
>
>



More information about the swing-dev mailing list