Call insertedInto or removedFrom before childrenChanged

Call insertedInto or removedFrom before childrenChanged

To avoid needing put all the children into a NodeVector
inside removeChildren() we can reverse the order of childrenChanged
and the associated child notifications (insertedInto/removedFrom).

This seems fine, it means that a parent now is not told it has new
children until they've already been told they're in the tree.
Inversely it does mean that when a child is told it's in the tree
the parent doesn't know it has new children yet, but that's less
crazy since it doesn't result in things like children without
inDocument() bits being set existing inside ::childrenChanged.

For removal I swapped the calls to notifyNodeRemoved with
the calls to childrenChanged.

For insertion I moved the call to childrenChanged into the
notifyNodeInserted function. This is necessary since it can
run script and we need to call childrenChanged before running
the script.

I also added some more RefPtr<Node> protect(this) guards
to methods which can runs script and more ScriptForbiddenScopes
to parts of the code that now can't run script.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175732

[ojan, 2014-06-04 00:38:46]

lgtm

https://codereview.chromium.org/306233005/diff/40001/Source/core/dom/Containe...
File Source/core/dom/ContainerNode.cpp (right):

https://codereview.chromium.org/306233005/diff/40001/Source/core/dom/Containe...
Source/core/dom/ContainerNode.cpp:509: RefPtr<Node> protect(this);
Don't need this if script can't execute.

https://codereview.chromium.org/306233005/diff/40001/Source/core/dom/Containe...
Source/core/dom/ContainerNode.cpp:661: notifyNodeInserted(*newChild, true);
Bool arguments!!

https://codereview.chromium.org/306233005/diff/40001/Source/core/dom/Containe...
Source/core/dom/ContainerNode.cpp:678: if (!root.isShadowRoot())
As per discussion in person, this is a little goofy. Maybe make this a FIXME for
your future patch that makes this better?

[esprehn, 2014-06-07 00:14:47]

The CQ bit was checked by esprehn@chromium.org

[commit-bot: I haz the power, 2014-06-07 00:15:09]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/esprehn@chromium.org/306233005/60001

[abarth-chromium, 2014-06-07 01:17:51]

https://codereview.chromium.org/306233005/diff/60001/Source/core/dom/Containe...
File Source/core/dom/ContainerNode.cpp (right):

https://codereview.chromium.org/306233005/diff/60001/Source/core/dom/Containe...
Source/core/dom/ContainerNode.cpp:281: RefPtrWillBeRawPtr<Node> protect(this);
Why not do this after checking for early return?

[commit-bot: I haz the power, 2014-06-07 03:48:18]

Change committed as 175732

[sof, 2014-06-07 15:41:09]

https://codereview.chromium.org/306233005/diff/60001/Source/core/dom/Containe...
File Source/core/dom/ContainerNode.cpp (right):

https://codereview.chromium.org/306233005/diff/60001/Source/core/dom/Containe...
Source/core/dom/ContainerNode.cpp:283: if (nextChild.previousSibling() ==
newChild || nextChild == newChild) // nothing to do
Oilpan compilation broke as a result of dropping the & from what was there
previously.

Fixed compilation via https://codereview.chromium.org/325573002/