Have ChildNodeList subclass NodeList instead of LiveNodeList

Have ChildNodeList subclass NodeList instead of LiveNodeList

Have ChildNodeList subclass NodeList instead of LiveNodeList. ChildNodeList
behaves differently than other LiveNodeLists and causes a lot of special
handling / branching in LiveNodeList and LiveNodeListBase.
LiveNodeList can also now deal solely with Elements instead of generic Nodes as
ChildNodeList was the only subclass not dealing with Elements.

Therefore, handling ChildNodeList separately makes the code a lot cleaner and
is also good for performance. I see a 3-4% improvement on the following
performance test: Bindings/node-list-access.html.

This patch is inspired from the following WebKit patch by <antti@apple.com>;:
http://trac.webkit.org/changeset/158536

R=esprehn, adamk

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=166565

[adamk, 2014-02-05 19:25:26]

lgtm

Less implementation inheritance always makes me happy :)

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
File Source/core/dom/NodeList.cpp (right):

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
Source/core/dom/NodeList.cpp:41: Node* NodeList::ownerNode() const
Why isn't this just virtual with a default implementation that returns null?
Instead it does three virtual calls...

[Inactive, 2014-02-05 19:52:48]

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
File Source/core/dom/NodeList.cpp (right):

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
Source/core/dom/NodeList.cpp:41: Node* NodeList::ownerNode() const
On 2014/02/05 19:25:26, adamk wrote:
> Why isn't this just virtual with a default implementation that returns null?
> Instead it does three virtual calls...

Seems to make sense. I remember I added this method to get rid of the custom
bindings for NodeList's visitDOMWrapper. However, I don't remember why I did not
make this virtual. Then again, when I did this there was only 1 virtual call so
it didn't look as bad :)

I will take care of this in a follow-up patch if you don't mind.

[Inactive, 2014-02-05 19:53:05]

The CQ bit was checked by ch.dumez@samsung.com

[commit-bot: I haz the power, 2014-02-05 19:53:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ch.dumez@samsung.com/154693002/40001

[adamk, 2014-02-05 20:32:45]

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
File Source/core/dom/NodeList.cpp (right):

https://codereview.chromium.org/154693002/diff/40001/Source/core/dom/NodeList...
Source/core/dom/NodeList.cpp:41: Node* NodeList::ownerNode() const
On 2014/02/05 19:52:48, Chris Dumez wrote:
> On 2014/02/05 19:25:26, adamk wrote:
> > Why isn't this just virtual with a default implementation that returns null?
> > Instead it does three virtual calls...
> 
> Seems to make sense. I remember I added this method to get rid of the custom
> bindings for NodeList's visitDOMWrapper. However, I don't remember why I did
not
> make this virtual. Then again, when I did this there was only 1 virtual call
so
> it didn't look as bad :)
> 
> I will take care of this in a follow-up patch if you don't mind.

Sure, a followup patch sounds fine. And I realize that there only used to be a
single isLiveNodeList() check here.

[Inactive, 2014-02-05 23:24:58]

The CQ bit was unchecked by ch.dumez@samsung.com

[Inactive, 2014-02-05 23:25:10]

The CQ bit was checked by ch.dumez@samsung.com

[commit-bot: I haz the power, 2014-02-05 23:25:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ch.dumez@samsung.com/154693002/40001

[commit-bot: I haz the power, 2014-02-06 01:32:02]

Change committed as 166565