Fix focus and bounds calculations for guest frames.

Fix focus and bounds calculations for guest frames.

Two fixes for accessibility of out-of-process iframes and guest frames
(e.g. <webview>):

1. BrowserAccessibilityManager::GetFocus should return the focused node
   from a child frame.
2. BrowserAccessibilityManager::GetLocalBoundsRect and
   BrowserAccessibilityCocoa::pointInScreen need to correctly take
   child frames into account.

This is sufficient for correct spoken feedback and UI element bounding
rects in VoiceOver when tabbing into and out of an out-of-process iframe
or <webview> guest frame. There may be other bugs, but simple examples
work in VoiceOver now.

BUG=330307, 368298
Committed: https://crrev.com/e8091802505a507bb6c11683fe77cf7b1d455e6d
Cr-Commit-Position: refs/heads/master@{#311777}

[dmazzoni, 2015-01-14 23:24:07]

dmazzoni@chromium.org changed reviewers:
+ dtseng@chromium.org

[aboxhall, 2015-01-15 05:08:25]

aboxhall@chromium.org changed reviewers:
+ aboxhall@chromium.org

[aboxhall, 2015-01-15 05:08:25]

lgtm

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
File content/browser/accessibility/browser_accessibility.cc (right):

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
content/browser/accessibility/browser_accessibility.cc:733: return NULL;
Totally fine if you want to leave it this way, but this is obviously redundant.

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
File content/browser/accessibility/browser_accessibility_cocoa.mm (right):

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
content/browser/accessibility/browser_accessibility_cocoa.mm:517: root =
root->GetParent()->manager()->GetRoot();
That's an impressive method chain! Any chance of any nulls in here? (I guess,
any chance of manager() or GetRoot() returning null.)

[dmazzoni, 2015-01-15 21:02:33]

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
File content/browser/accessibility/browser_accessibility.cc (right):

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
content/browser/accessibility/browser_accessibility.cc:733: return NULL;
On 2015/01/15 05:08:25, aboxhall wrote:
> Totally fine if you want to leave it this way, but this is obviously
redundant.

Great observation. Simplified.

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
File content/browser/accessibility/browser_accessibility_cocoa.mm (right):

https://codereview.chromium.org/849143003/diff/1/content/browser/accessibilit...
content/browser/accessibility/browser_accessibility_cocoa.mm:517: root =
root->GetParent()->manager()->GetRoot();
On 2015/01/15 05:08:25, aboxhall wrote:
> That's an impressive method chain! Any chance of any nulls in here? (I guess,
> any chance of manager() or GetRoot() returning null.)

Great question!
* root->GetParent() can't be NULL because of the loop condition
* a BrowserAccessibility object is always owned by a
BrowserAccessibilityManager; it's never NULL
* A BrowserAccessibilityManager always has a root, it's never NULL

So we're okay.

However, manager->delegate() may be NULL, so I'll fix that below.

[dmazzoni, 2015-01-15 21:02:44]

The CQ bit was checked by dmazzoni@chromium.org

[commit-bot: I haz the power, 2015-01-15 21:03:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/849143003/20001

[commit-bot: I haz the power, 2015-01-15 22:52:56]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-01-15 22:52:57]

Try jobs failed on following builders:
  linux_chromium_rel_ng on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[dmazzoni, 2015-01-15 23:28:05]

The CQ bit was checked by dmazzoni@chromium.org

[commit-bot: I haz the power, 2015-01-15 23:28:39]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/849143003/20001

[commit-bot: I haz the power, 2015-01-16 00:34:10]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-01-16 00:35:42]

Patchset 2 (id:??) landed as
https://crrev.com/e8091802505a507bb6c11683fe77cf7b1d455e6d
Cr-Commit-Position: refs/heads/master@{#311777}