And when I build WebKit and run it in Safari, only ctrl+alt work. Weird. I'm
doing testing now to see what this fixes.

Avi

On Thu, Mar 5, 2009 at 4:38 PM, Avi Drissman <[email protected]> wrote:

> That value gets set in WebKit's WebFrame.mm:
>
>         if ([[NSApp 
> accessibilityAttributeValue:NSAccessibilityEnhancedUserInterfaceAttribute] 
> boolValue])
>             AXObjectCache::*enableEnhancedUserInterfaceAccessibility*();
>
> What is that mysterious NSAccessibilityEnhancedUserInterfaceAttribute? It's
> a secret AppKit value, equaling @"AXEnhancedUserInterface". What's *that*?
>
> Oh, it's a magical value set by VoiceOver as a hint to certain internal
> AppKit controls to alter their behavior slightly to make life better for it.
> (See
> http://lists.apple.com/archives/accessibility-dev/2006/Feb/msg00009.html). 
> That message is the only documentation past seeing it in WebKit archives.
>
> I'm fine with this, but I just want to know how I can do both...
>
> Avi
>
>
>
> On Thu, Mar 5, 2009 at 4:34 PM, Thomas Van Lenten 
> <[email protected]>wrote:
>
>> Does the universal access setting w/in safari's prefs change this in any
>> way?
>>
>> TVL
>>
>>
>>
>> On Thu, Mar 5, 2009 at 4:30 PM, Avi Drissman <[email protected]> wrote:
>>
>>> I'm trying to figure out the exact meaning of
>>> accessibilityEnhancedUserInterfaceEnabled because I'm getting that effect on
>>> my Mac without any accessibility. In addition, what I forgot to mention is
>>> that *both* ctrl *and* ctrl+alt work in Safari for me, while we can only
>>> pick one to return from accessKeyModifiers() and whichever one we don't
>>> return doesn't work.
>>>
>>> Avi
>>>
>>>
>>> On Thu, Mar 5, 2009 at 4:24 PM, Ben Goodger (Google) 
>>> <[email protected]>wrote:
>>>
>>>> It sounds like they only use the single modifier when the
>>>> accessibility system setting is turned on, otherwise they use the dual
>>>> one. Why don't we just do the same thing? I'm pretty sure we should
>>>> match the platform default browser instead of emacs.
>>>>
>>>> -Ben
>>>>
>>>> On Thu, Mar 5, 2009 at 1:22 PM, Avi Drissman <[email protected]> wrote:
>>>> > The definition of what the accesskey modifier is the result of
>>>> > EventHandler::accessKeyModifiers(). On Windows it's alt, and on
>>>> Chromium
>>>> > it's also alt.
>>>> >
>>>> > The question is: What to make it on the Mac? And how to do it right?
>>>> >
>>>> > Mozilla apparently switched recently to use the command key. They used
>>>> to
>>>> > use the control key. What does Safari use? Much more difficult
>>>> question. The
>>>> > code says:
>>>> >
>>>> >> unsigned EventHandler::accessKeyModifiers()
>>>> >> {
>>>> >>     // Control+Option key combinations are usually unused on Mac OS
>>>> X, but
>>>> >> not when VoiceOver is enabled.
>>>> >>     // So, we use Control in this case, even though it conflicts with
>>>> >> Emacs-style key bindings.
>>>> >>     // See <https://bugs.webkit.org/show_bug.cgi?id=21107> for more
>>>> >> detail.
>>>> >>     if (AXObjectCache::accessibilityEnhancedUserInterfaceEnabled())
>>>> >>         return PlatformKeyboardEvent::CtrlKey;
>>>> >>
>>>> >>     return PlatformKeyboardEvent::CtrlKey |
>>>> PlatformKeyboardEvent::AltKey;
>>>> >> }
>>>> >
>>>> > And if you go to the layout test, you'll see that on the Mac it uses
>>>> the
>>>> > modifiers ctrl+alt. This solution is problematic for two reasons.
>>>> >
>>>> > The first is that using a dual modifier sucks. It makes it extremely
>>>> > difficult to use accesskeys.
>>>> >
>>>> > The second is that triggering them with a single modifier (just the
>>>> ctrl
>>>> > key) on my machine works fine in Safari. Which puts me in a quandary,
>>>> > because I'm not sure how Safari's doing that (I'm digging now). If we
>>>> alter
>>>> > our accessKeyModifiers() to return the dual modifier, layout tests
>>>> start
>>>> > passing, but it's hard to use. If we use just ctrl, then it's easy to
>>>> use
>>>> > but we fail tests. And somehow Safari does both...
>>>> >
>>>> > Avi
>>>> >
>>>> > >
>>>> >
>>>>
>>>
>>>
>>> >>>
>>>
>>
>

--~--~---------~--~----~------------~-------~--~----~
Chromium Developers mailing list: [email protected] 
View archives, change email options, or unsubscribe: 
    http://groups.google.com/group/chromium-dev
-~----------~----~----~----~------~----~------~--~---

Reply via email to