Add use counters for using Attrs as Nodes.

Add use counters for using Attrs as Nodes.

DOM4 makes Attr no longer be a subtype of Node. This adds a use
counter for passing Attr where Node is expected, or calling a method
declared on Node with an Attr.

BUG=305105
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=165262

[dominicc (has gone to gerrit), 2014-01-08 08:19:25]

PTAL

This doesn't count everything it possibly could:

Node doesn't have named or indexed properties, so there's no need to check
whether the 'this' of a named or indexed property method is Attr.

Things returning Node aren't checked to see if they're really returning Attr. I
can't think of a situation where that happens outside of the NamedNodeMap for
Element.attributes collection, but we don't want to count that because DOM4 has
a replacement for it anyway.

It doesn't bode well for deprecating Attr : Node that 433 layout tests tickle
these counters. It's possible there's a common cause like everyone touching
nodeType or something simple like that, in which case it's worth considering
feedback to Anne for including those properties in DOM4 Attr... still waiting
for run-webkit-tests to collect all those crashes :P

[esprehn, 2014-01-08 08:31:00]

What does adding the extra typeinfo checks to all these methods do to
performance?

[eseidel, 2014-01-08 08:55:04]

There are a whole bunch of DOM xml tests.  I'm not surprised they trigger a
bunch of stuff which doesn't actually exist on the web.  I wouldn't be worried
about LayoutTests triggering these.

[dominicc (has gone to gerrit), 2014-01-14 03:16:49]

On 2014/01/08 08:31:00, esprehn wrote:
> What does adding the extra typeinfo checks to all these methods do to
> performance?

Here's Dromaeo DOM Core Modify with Content Shell on a Nexus 7 (2013)

Before:

RESULT dom: dom= 20.8367753108 runs/s
*RESULT dom_modify: dom_modify= 20.8367753108 runs/s
RESULT dom_modify_appendChild: dom_modify_appendChild= 53.3006301331 runs/s
RESULT dom_modify_cloneNode: dom_modify_cloneNode= 12.7908390358 runs/s
RESULT dom_modify_createElement: dom_modify_createElement= 36.0956422354 runs/s
RESULT dom_modify_createTextNode: dom_modify_createTextNode= 9.40621450391
runs/s
RESULT dom_modify_innerHTML: dom_modify_innerHTML= 7.18417129893 runs/s
RESULT dom_modify_insertBefore: dom_modify_insertBefore= 49.2158900765 runs/s
RESULT telemetry_page_measurement_results: num_failed= 0 count
RESULT telemetry_page_measurement_results: num_errored= 0 count

After:

RESULT dom: dom= 20.8784647583 runs/s
*RESULT dom_modify: dom_modify= 20.8784647583 runs/s
RESULT dom_modify_appendChild: dom_modify_appendChild= 54.7415709732 runs/s
RESULT dom_modify_cloneNode: dom_modify_cloneNode= 12.6760917528 runs/s
RESULT dom_modify_createElement: dom_modify_createElement= 35.7622287225 runs/s
RESULT dom_modify_createTextNode: dom_modify_createTextNode= 9.62431676096
runs/s
RESULT dom_modify_innerHTML: dom_modify_innerHTML= 7.05035813093 runs/s
RESULT dom_modify_insertBefore: dom_modify_insertBefore= 49.1907077846 runs/s
RESULT telemetry_page_measurement_results: num_failed= 0 count
RESULT telemetry_page_measurement_results: num_errored= 0 count

[abarth-chromium, 2014-01-16 05:14:40]

LGTM assuming we're not regressing benchmarks.

[commit-bot: I haz the power, 2014-01-16 05:48:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dominicc@chromium.org/127723002/150001

[commit-bot: I haz the power, 2014-01-16 06:35:30]

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

[commit-bot: I haz the power, 2014-01-16 08:59:26]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dominicc@chromium.org/127723002/150001

[commit-bot: I haz the power, 2014-01-16 09:39:40]

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

[commit-bot: I haz the power, 2014-01-17 00:21:00]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dominicc@chromium.org/127723002/150001

[commit-bot: I haz the power, 2014-01-17 00:36:53]

Change committed as 165262

[dominicc (has gone to gerrit), 2014-01-19 23:17:30]

A revert of this CL has been created in
https://codereview.chromium.org/130703003/ by dominicc@chromium.org.

The reason for reverting is: Adding use counters for using Attrs as Nodes
regressed dom perf and others.

BUG=335750.