Auto-generate V8HTMLFormElement::namedPropertyGetter

Auto-generate namedPropertyGetter of V8HTMLFormElement and HTMLFrameSetElement 

Removed V8HTMLFormElementCustom.cpp.
Removed V8HTMLFrameSetElementCustom.cpp. 

Add attribute: [DoNotCheckJSProperty]
Named/indexed property interceptors of V8 is called whenever
a script accesses any object property.
If we want not to override callback property or property in
prototype chain by named getter, we have to check if there is
already such property in head of named getter.
(for example, document.all.length)
On the other hand, there is another kind of named getter:
which we want to override property, thus we must not check it
(for example, DOMStringMap)
For getter with [DoNotCheckJSProperty], generator does not
generate check code.

R=haraken@chromium.org
BUG=229740
TEST=No tests. refactoring.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=150907

[kojih, 2013-05-21 11:12:10]

r?

WDYT of
[NoGetRealNamedPropertyInPrototypeChainCheck]
[NoHasRealNamedCallbackPropertyCheck]

[haraken, 2013-05-21 14:11:22]

> because it is necessary not to check it.

Would you elaborate on this?

I thought that the fact that custom named getters of some interfaces do not have
these checks was just a bug. If so, it would be reasonable to add the checks as
we kill the buggy custom bindings. If not, I want to understand the rationale
for not having the checks intentionally.

[kojih, 2013-05-22 03:03:16]

without [NoHasRealNamedCallbackPropertyCheck], 3 tests failed.
[1] fast/dom/collection-length-should-not-be-overridden.html
[2] fast/forms/input-named-action-overrides-action-attribute.html
[3] fast/forms/state-restore-hidden.html

All of these case, XXX.Name should return collection's item named "Name"
instead of XXX's property named "Name".

Summary of [1]
<form>
<input type="text" name="some">
<input type="text" name="length">
</form>
document.forms[0].length should be [object HTMLInputElement] instead of 2
(number of children).
Behavior in Opera(12.45), FF(nightly) -> same as current chromium

On the other hand,
documentAllLength = document.all.length;
shouldBe("documentAllLength", "18");


Summary of [2]
<form action="/">
   <input type=hidden name="action" value="">
</form>
document.forms[0].action should be <input> instead of “/” ("action" attribute of
<form>)
Behavior in Opera(12.45), FF(nightly) -> same as current chromium


Summary of [3]
FAIL document.f.hidden.value should be after (of type string). Was undefined (of
type undefined).

because document.f.hidden should be <input> with name “hidden” instead of
“hidden” attribute of <form>

<form action="data:text/html,<script>history.back()&lt;/script>" name=f>
    <input type="hidden" name="hidden" value="before" />
</form>
Behavior in Opera(12.45), FF(nightly) -> same as current chromium

It seems reasonable to add [NoHasRealNamedCallbackPropertyCheck] (which means
not check it in namedPropertyGetter)

Next step: investigate spec

[kojih, 2013-05-22 03:10:52]

On the other hand, 
document.all.length should be "length" property of document.all
instead of element named "length".

So there are 2 types of named getters: one that must check it and must not check
it.

[kojih, 2013-05-22 09:32:20]

updated & merged https://codereview.chromium.org/15540004/

[haraken, 2013-05-22 09:44:48]

Discussed offline how the checks should be added. We reached the conclusion
described in the "Description".

LGTM. Excellent patch!

[commit-bot: I haz the power, 2013-05-22 09:48:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kojih@chromium.org/15556003/6001

[commit-bot: I haz the power, 2013-05-22 10:55:40]

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, 2013-05-22 11:54:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kojih@chromium.org/15556003/6001

[commit-bot: I haz the power, 2013-05-22 12:29:55]

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, 2013-05-22 15:47:15]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kojih@chromium.org/15556003/6001

[commit-bot: I haz the power, 2013-05-22 16:28:24]

Change committed as 150907