<Swing Dev> [9-client] Review request for bug :7133530 javax/swing/JRadioButton/4314194/bug4314194.java fails on MacOS

Sergey Bylokhov Sergey.Bylokhov at oracle.com
Tue Oct 20 13:56:18 UTC 2015


On 20.10.15 15:04, shilpi rastogi wrote:
> Hi Sergey
>
> Thanks!
>
> I tried to tun test with jtreg on Mac 10.9 (stt-mac-05.ru.oracle.com)
> and i am able to reproduce this issue.

Interesting, it actually does not work when old test is started via 
jtreg. It should be investigated.

>
> Please let me know if i am doing anything wrong.

The command line is correct, but if you run the new test(from your 
webrev) without setting default l&f it will pass.

>
>
>
>
> Thanks,
> Shilpi
> On 10/19/2015 2:48 AM, Sergey Bylokhov wrote:
>> I just have run the test on my local system and see that it works as
>> expected on the latest jdk9. It seems this bug in jdk was fixed
>> already(the similar issue on windows JDK-8075785).
>>
>> I suggest not to close the bug as duplicate but update the test and
>> move it to the public.(probably it will be useful to cover all l&f in
>> the loop).
>>
>> On 16.10.15 13:46, shilpi rastogi wrote:
>>>
>>> On 9/16/2015 7:16 PM, Sergey Bylokhov wrote:
>>>> Hi, Shilpi.
>>>> Can you clarify, why the test does not work in case of Aqua?
>>> The UI Manager documentation states the following –
>>>
>>> “The set of defaults a particular look and feel supports is defined and
>>> documented by that look and feel. In addition, each look and feel, or
>>> |ComponentUI|provided by a look and feel, may access the defaults at
>>> different times in their life cycle. Some look and feels may
>>> aggressively look up defaults, so that changing a default may not have
>>> an effect after installing the look and feel. Other look and feels may
>>> lazily access defaults so that a change to the defaults may effect an
>>> existing look and feel. Finally, other look and feels might not
>>> configure themselves from the defaults table in any way. None-the-less
>>> it is usually the case that a look and feel expects certain defaults, so
>>> that in general a |ComponentUI|provided by one look and feel will not
>>> work with another look and feel”
>>>
>>> If you look at the com.apple.laf.AquaLookAndFeel implementation, it only
>>> returns a copy of all the UI defaults (new UIDefaults()) through
>>> getDefaults() and any change that you make to the defaults is not going
>>> to change the actual defaults. On the other hand, Metal returns you the
>>> actual Map that you can modify and customize. As stated in the UI
>>> Manager spec, it is up to the look and feel to choose whether to allow
>>> customization or not to their defaults and in this case, Aqua has chosen
>>> not to. One logical reason I could think of is to retain the parity with
>>> native behavior which is the primary intention of any Native Look And
>>> Feel.
>>>
>>> Regards,
>>> Shilpi
>>>>
>>>> 16.09.2015 9:37, shilpi rastogi wrote:
>>>>> Hi All,
>>>>>
>>>>> Please review the fix
>>>>>
>>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-7133530
>>>>> webrev: http://cr.openjdk.java.net/~psadhukhan/shilpi/webrev/
>>>>>
>>>>> Issue: Colored text is not shown on disabled checkbox and radio
>>>>> button.
>>>>> The test expects to be run using Metal L&F but is actually run using
>>>>> Aqua L&F so setting it explicitly.
>>>>>
>>>>> Thanks,
>>>>> Shilpi
>>>>>
>>>>
>>>>
>>>
>>
>>
>


-- 
Best regards, Sergey.



More information about the swing-dev mailing list