Ending the need for apple.laf.useScreenMenuBar

[Artem Ananiev]

On 12/3/2011 4:23 AM, Scott Kovatch wrote:
>
> On Dec 2, 2011, at 2:57 PM, Mike Swingler wrote:
>
>> On Dec 2, 2011, at 2:28 PM, Mario Torre wrote:
>>
>>> Il giorno 02/dic/2011, alle ore 23:25, Scott Kovatch ha scritto:
>>>> Folks,
>>>>
>>>> This may be a better question for Artem and the AWT/Swing team,
>>>> but it affects the Mac so I'm starting here.
>>>>
>>>> Now that it's 2011, can we please get rid of
>>>> apple.laf.useScreenMenuBar and just allow menu bars in the Aqua
>>>> L&F to be enabled by default? I believe JavaFx doesn't force
>>>> Mac menu bars into the window (but if it does, I will make my
>>>> displeasure known to Richard Bair on Monday.)
>>>>
>>>> I ask, because today I was looking at the changes to Web Start
>>>> and the plugin to allow this property as a 'safe' one that can
>>>> be defined in a JNLP file, and I decided now is the time to get
>>>> rid of it once and for all. It's not that I object to adding it
>>>> to Web Start; I object to the necessity of having to define it
>>>> in the first place.
>>>>
>>>> -- Scott
>>>
>>> I think you're right, I'm in favor of this change too.
>>>
>>> Cheers, Mario
>>
>> I'd be the first one to advocate for this, but isn't the key
>> problem that he semantics of the Screen Menu bar violate certain
>> rules in the JCK? Implementing arbitrary Java drawing into
>> NSView-based menu items might even be possible post-Leopard, but is
>> that really something the AWT team would want to implement?

To me, if a JCK test makes certain assumption, which is not valid on 
such a widely distributed platform as Mac OS X, we should at least think 
about changing this assumption. However, given how important JCK 
certification is for the whole OpenJDK process, and I'd expect that we 
may not be able to change some parts of JCK, at least in 7u.

> It's been so long since this was first added that I forgot the
> original argument that triggered it in the first place. :-\
>
> Let's find out what the JCK complaint is, and determine how valid the
> complaint is. A poorly written rule should not be a hindrance to
> providing a good UI. I think I'm (finally?) in a position to have
> some kind of say or influence on the matter.

This sounds very reasonable. Let me take care of contacting JCK team.

> If it's some kind of custom menu item drawing that's causing this
> failure, I don't consider that a valid reason to use in-window menus
> by default. It's not done often enough to be a primary feature, and
> (IMO) really shouldn't be done in the first place.
>
> -- Scott

Thanks,

Artem