Implement accessible BoundsForRange on Mac

Implement accessible BoundsForRange on Mac.

This allows VoiceOver to interact with individual characters and words
and draw a precise bounding box around the active range.

Depends on: 
1. https://codereview.chromium.org/23983002/ 
2. https://codereview.chromium.org/25943003/ 
3. https://codereview.chromium.org/25987002/

BUG=98977
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=235686

[David Tseng, 2013-10-17 22:11:02]

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_cocoa.mm (right):

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_cocoa.mm:358: child_count =
0;
Is this behavior we want cross platform or is there a reason why we want this
only for Mac?

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_cocoa.mm:1101:
range.location, range.length);
FYI; Apple says of range.length:
The number of items in the range (can be 0).

Do we handle length of 0? I know we don't have a test for that particular case.

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_mac_unittest.mm (left):

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_mac_unittest.mm:32:
@implementation MockAccessibilityDelegate
Are we testing getting a parameterized bounds for range attribute in this cl?

[dmazzoni, 2013-11-11 23:19:34]

The changes this one depended on are now checked in; this is
rebased and ready for another look.

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_cocoa.mm (right):

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_cocoa.mm:358: child_count =
0;
On 2013/10/17 22:11:03, David Tseng wrote:
> Is this behavior we want cross platform or is there a reason why we want this
> only for Mac?

This was fixed in a different changelist.

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_cocoa.mm:1101:
range.location, range.length);
On 2013/10/17 22:11:03, David Tseng wrote:
> FYI; Apple says of range.length:
> The number of items in the range (can be 0).
> 
> Do we handle length of 0? I know we don't have a test for that particular
case.

Yes, it works. I think it makes sense to have a length of zero, that's the way
to get the bounds of the caret.

[dmazzoni, 2013-11-11 23:21:41]

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
File content/browser/accessibility/browser_accessibility_mac_unittest.mm (left):

https://codereview.chromium.org/25434005/diff/1/content/browser/accessibility...
content/browser/accessibility/browser_accessibility_mac_unittest.mm:32:
@implementation MockAccessibilityDelegate
On 2013/10/17 22:11:03, David Tseng wrote:
> Are we testing getting a parameterized bounds for range attribute in this cl?

I didn't add tests of the Mac API. Do you think it would be helpful? My goal has
been to have good tests of the cross-platform logic, and to keep the
platform-specific code simple.

My main concern with testing the platform API is too many duplicate tests
and too much boilerplate for the test. But if you think there's something
worth testing, let me know.

[David Tseng, 2013-11-12 00:48:52]

LGTM

https://codereview.chromium.org/25434005/diff/5001/content/browser/accessibil...
File content/browser/accessibility/browser_accessibility_cocoa.mm (right):

https://codereview.chromium.org/25434005/diff/5001/content/browser/accessibil...
content/browser/accessibility/browser_accessibility_cocoa.mm:1094: if ([self
internalRole] != blink::WebAXRoleStaticText)
Btw, what about editable text? Certain accessibility tools like VoiceOver do
allow for exploration of these items by various text units. I imagine the
highlighting info gets passed through in these cases.

[dmazzoni, 2013-11-14 23:29:24]

+avi for OWNERS review of Obj-C code in content/browser/renderer_host while Nico
is away

[Nico, 2013-11-17 07:26:09]

render_widget_host_view_mac.mm lgtm

(I'm around, just in a different timezone. And was in a plane and figuring out
wifi the last 1.5 days.)

[dmazzoni, 2013-11-18 06:52:28]

On 2013/11/17 07:26:09, Nico (in Tokyo until Nov 25) wrote:
> render_widget_host_view_mac.mm lgtm
> 
> (I'm around, just in a different timezone. And was in a plane and figuring out
> wifi the last 1.5 days.)

Thanks, and good luck!

[commit-bot: I haz the power, 2013-11-18 06:52:56]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmazzoni@chromium.org/25434005/135001

[commit-bot: I haz the power, 2013-11-18 10:53:38]

Change committed as 235686