Support an Event Path API

Support an Event Path API behind experimentalShadowDOM flag.

The spec issue is here:
   https://www.w3.org/Bugs/Public/show_bug.cgi?id=21066 

See also the comment in the bug:
  https://code.google.com/p/chromium/issues/detail?id=234030#c5

BUG=234030
R=dglazkov@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=150007

[hayato, 2013-04-26 08:13:46]

An initial attempt to implement an Event Path API. Not ready for review.

[hayato, 2013-05-07 04:38:19]

On 2013/04/26 08:13:46, hayato wrote:
> An initial attempt to implement an Event Path API. Not ready for review.

Event.path is now a function to make it clear it returns static NodeList.

In Element.idl:
   [EnabledAtRuntime=experimentalShadowDOM] NodeList path();

WDYT?

Note that an event can be re-dispatched.
So the same event object can return different NodeList if it is re-dispatched.

[dglazkov, 2013-05-07 14:48:40]

It's nice and clean :)

I wonder though -- should we limit the results to account for encapsulation. I
don't think we need to limit by lower boundary, but definitely by higher
boundary.

WDYT?

[hayato, 2013-05-08 01:57:46]

On 2013/05/07 14:48:40, Dimitri Glazkov wrote:
> It's nice and clean :)
> 
> I wonder though -- should we limit the results to account for encapsulation. I
> don't think we need to limit by lower boundary, but definitely by higher
> boundary.
> 
> WDYT?

I've heard such requirement once. That might be nice. Let me try to implement it
to know its feasibility.

[hayato, 2013-05-08 06:53:05]

I've attached the picture on issue234030.
https://code.google.com/p/chromium/issues/detail?id=234030#c10

That explains what event.path() should return at each node where event is
dispatched.

Do you like my idea? Let's discuss how event.path() should behave.

[hayato, 2013-05-08 08:38:40]

I've uploaded the new patch which implements
https://code.google.com/p/chromium/issues/detail?id=234030#c11.

[dglazkov, 2013-05-08 16:37:29]

lgtm

https://codereview.chromium.org/14508005/diff/13001/Source/core/dom/Node.cpp
File Source/core/dom/Node.cpp (right):

https://codereview.chromium.org/14508005/diff/13001/Source/core/dom/Node.cpp#...
Source/core/dom/Node.cpp:2782: bool Node::canReach(const Node* target) const
I think we can use the new fancy stuff in DOM spec to craft a better name here:
http://dom.spec.whatwg.org/#concept-tree-host-including-inclusive-ancestor

[hayato, 2013-05-09 04:20:12]

On 2013/05/08 16:37:29, Dimitri Glazkov wrote:
> lgtm
> 
> https://codereview.chromium.org/14508005/diff/13001/Source/core/dom/Node.cpp
> File Source/core/dom/Node.cpp (right):
> 
>
https://codereview.chromium.org/14508005/diff/13001/Source/core/dom/Node.cpp#...
> Source/core/dom/Node.cpp:2782: bool Node::canReach(const Node* target) const
> I think we can use the new fancy stuff in DOM spec to craft a better name
here:
> http://dom.spec.whatwg.org/#concept-tree-host-including-inclusive-ancestor

I've updated the patch.
- Node::canReach was gone.
- Instead of that, TreeScope::isInclusiveAncestorOf(const TreeScope*) was
introduced.

Let me land this patch in a few hours.

[commit-bot: I haz the power, 2013-05-09 04:22:57]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/hayato@chromium.org/14508005/19001

[commit-bot: I haz the power, 2013-05-09 04:46:33]

Step "update" is always a major failure.
Look at the try server FAQ for more details.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_layo...

[commit-bot: I haz the power, 2013-05-09 04:49:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/hayato@chromium.org/14508005/19001

[commit-bot: I haz the power, 2013-05-09 05:34:35]

Retried try job too often on linux_layout_rel for step(s) webkit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_layo...

[commit-bot: I haz the power, 2013-05-09 05:52:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/hayato@chromium.org/14508005/19001

[commit-bot: I haz the power, 2013-05-09 06:33:07]

Retried try job too often on linux_layout_rel for step(s) webkit_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_layo...

[hayato, 2013-05-09 06:35:48]

Committed patchset #5 manually as r150007 (presubmit successful).

[dominicc (has gone to gerrit), 2013-05-10 06:14:47]

On 2013/05/09 06:35:48, hayato wrote:
> Committed patchset #5 manually as r150007 (presubmit successful).

Where's the layout test that verifies that this does not leak user agent shadows
to script?

[hayato, 2013-05-10 06:18:52]

On 2013/05/10 06:14:47, dominicc wrote:
> On 2013/05/09 06:35:48, hayato wrote:
> > Committed patchset #5 manually as r150007 (presubmit successful).
> 
> Where's the layout test that verifies that this does not leak user agent
shadows
> to script?

No, I've not written that.

[hayato, 2013-05-10 06:24:01]

On 2013/05/10 06:18:52, hayato wrote:
> On 2013/05/10 06:14:47, dominicc wrote:
> > On 2013/05/09 06:35:48, hayato wrote:
> > > Committed patchset #5 manually as r150007 (presubmit successful).
> > 
> > Where's the layout test that verifies that this does not leak user agent
> shadows
> > to script?
> 
> No, I've not written that.

Let me verify that. It is supposed that it doesn't leak.

[dominicc (has gone to gerrit), 2013-05-10 09:36:45]

On 2013/05/10 06:24:01, hayato wrote:
> On 2013/05/10 06:18:52, hayato wrote:
> > On 2013/05/10 06:14:47, dominicc wrote:
> > > On 2013/05/09 06:35:48, hayato wrote:
> > > > Committed patchset #5 manually as r150007 (presubmit successful).
> > > 
> > > Where's the layout test that verifies that this does not leak user agent
> > shadows
> > > to script?
> > 
> > No, I've not written that.
> 
> Let me verify that. It is supposed that it doesn't leak.

At a first glance it does not leak, but I would really like to have a regression
test in place for this.

[hayato, 2013-05-10 09:43:30]

On 2013/05/10 09:36:45, dominicc wrote:
> On 2013/05/10 06:24:01, hayato wrote:
> > On 2013/05/10 06:18:52, hayato wrote:
> > > On 2013/05/10 06:14:47, dominicc wrote:
> > > > On 2013/05/09 06:35:48, hayato wrote:
> > > > > Committed patchset #5 manually as r150007 (presubmit successful).
> > > > 
> > > > Where's the layout test that verifies that this does not leak user agent
> > > shadows
> > > > to script?
> > > 
> > > No, I've not written that.
> > 
> > Let me verify that. It is supposed that it doesn't leak.
> 
> At a first glance it does not leak, but I would really like to have a
regression
> test in place for this.

Yeah, a test was uploaded and is waiting for a review.
https://codereview.chromium.org/14864018/