Re-land (2): Reimplement automation API on top of C++-backed AXTree.

Re-land (2): Reimplement automation API on top of C++-backed AXTree.

Original review: https://codereview.chromium.org/1155183006
Landed:         r335183
Reverted:       r335343 (bug 502311)
Landed again:   r339929
Reverted again: r340091 (bug 502311 again)

One fix to ax_tree.cc - bug 502311 was triggered from two different code paths, now both have tests.

Also adds tiny fixes to automation_event_router.cc
and automation_internal_custom_bindings.cc to prevent
crashes when toggling ChromeVox off and on.

TBRing dcheng, kalman since none of those files have
changed since the original review.

BUG=495323, 502311
TBR=dcheng,kalman

Committed: https://crrev.com/d6848287cfe79ff1ed18dbae20f6be231b9276b8
Cr-Commit-Position: refs/heads/master@{#340602}

[dmazzoni, 2015-07-27 18:13:45]

dmazzoni@chromium.org changed reviewers:
+ dcheng@chromium.org, dtseng@chromium.org, kalman@chromium.org

[dmazzoni, 2015-07-27 18:14:22]

David, ready for a look - see change description for the four files that changed
relative to the last patch that was reverted.

[not at google - send to devlin, 2015-07-27 18:34:20]

Without reading through this - I've been adding metrics for extension events,
which is something I was holding off doing for automation to not conflict with
this patch, but I needed to do it in the end - have you maintained that?

[David Tseng, 2015-07-27 18:46:38]

lgtm

https://codereview.chromium.org/1259703002/diff/60001/ui/accessibility/ax_tre...
File ui/accessibility/ax_tree.cc (right):

https://codereview.chromium.org/1259703002/diff/60001/ui/accessibility/ax_tre...
ui/accessibility/ax_tree.cc:92: // Make sure root_ always points to something
valid or nullptr,
nit: Confusing comment; root_ is always nullptr inside DestroySubtree below?

[dmazzoni, 2015-07-27 18:52:16]

On 2015/07/27 18:34:20, kalman wrote:
> Without reading through this - I've been adding metrics for extension events,
> which is something I was holding off doing for automation to not conflict with
> this patch, but I needed to do it in the end - have you maintained that?

Yep, I noticed that. I've tried to be careful about merges. I left one of those
metrics out in an earlier upload and it caused a test failure.

[dmazzoni, 2015-07-27 18:52:18]

On 2015/07/27 18:34:20, kalman wrote:
> Without reading through this - I've been adding metrics for extension events,
> which is something I was holding off doing for automation to not conflict with
> this patch, but I needed to do it in the end - have you maintained that?

Yep, I noticed that. I've tried to be careful about merges. I left one of those
metrics out in an earlier upload and it caused a test failure.

[dmazzoni, 2015-07-27 21:40:31]

https://codereview.chromium.org/1259703002/diff/60001/ui/accessibility/ax_tre...
File ui/accessibility/ax_tree.cc (right):

https://codereview.chromium.org/1259703002/diff/60001/ui/accessibility/ax_tre...
ui/accessibility/ax_tree.cc:92: // Make sure root_ always points to something
valid or nullptr,
On 2015/07/27 18:46:38, David Tseng wrote:
> nit: Confusing comment; root_ is always nullptr inside DestroySubtree below?

Done.

[dmazzoni, 2015-07-27 21:56:13]

The CQ bit was checked by dmazzoni@chromium.org

[dmazzoni, 2015-07-27 21:56:13]

The patchset sent to the CQ was uploaded after l-g-t-m from dtseng@chromium.org
Link to the patchset: https://codereview.chromium.org/1259703002/#ps80001
(title: "Add missing null check caught by asan")

[commit-bot: I haz the power, 2015-07-27 21:56:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1259703002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1259703002/80001

[commit-bot: I haz the power, 2015-07-27 23:41:40]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-07-27 23:42:54]

Patchset 5 (id:??) landed as
https://crrev.com/d6848287cfe79ff1ed18dbae20f6be231b9276b8
Cr-Commit-Position: refs/heads/master@{#340602}

[David Tseng, 2015-07-29 16:14:41]

https://codereview.chromium.org/1259703002/diff/60001/chrome/renderer/resourc...
File chrome/renderer/resources/extensions/automation/automation_node.js (right):

https://codereview.chromium.org/1259703002/diff/60001/chrome/renderer/resourc...
chrome/renderer/resources/extensions/automation/automation_node.js:156: get
parent() {
rootWebArea are no longer hooked up properly across navigations.
chrome.automation.getDesktop(function(d){window.d=d})
undefined
JS> w=d.findAll({role:'rootWebArea'})
[AutomationRootNode]
JS> w=w[0]
JS> w.docUrl
"http://www.latimes.com/"
JS> w.parent
undefined

https://codereview.chromium.org/1259703002/diff/60001/chrome/renderer/resourc...
File chrome/renderer/resources/extensions/automation_custom_bindings.js (right):

https://codereview.chromium.org/1259703002/diff/60001/chrome/renderer/resourc...
chrome/renderer/resources/extensions/automation_custom_bindings.js:160:
automationInternal.enableFrame(childTreeID);
childTreeID can be null here.