[webkit-reviews] review granted: [Bug 45426] [Chromium] fast/dom/dataset-gc.html is failing : [Attachment 67958] Changed to use setHiddenWindowReference which is more inline with what the JSC bindings do

[bugzilla-daemon at webkit.org]

Nate Chapin <japhet at chromium.org> has granted Erik Arvidsson
<arv at chromium.org>'s request for review:
Bug 45426: [Chromium] fast/dom/dataset-gc.html is failing
https://bugs.webkit.org/show_bug.cgi?id=45426

Attachment 67958: Changed to use setHiddenWindowReference which is more inline
with what the JSC bindings do
https://bugs.webkit.org/attachment.cgi?id=67958&action=review

------- Additional Comments from Nate Chapin <japhet at chromium.org>

> +v8::Handle<v8::Value> toV8(DOMStringMap* impl)
> +{
> +    if (!impl)
> +	   return v8::Null();
> +    v8::Handle<v8::Object> wrapper = V8DOMStringMap::wrap(impl);
> +    // Add a hidden reference from the element to the DOMStringMap.
> +    Element* element = impl->element();
> +    if (!wrapper.IsEmpty() && element)
> +	   V8DOMWrapper::setHiddenWindowReference(element->document()->frame(),
wrapper);
> +    return wrapper;

If that's what JSC does, I guess it's ok to do the same. I don't know anything
about DOMStringMap, so I don't have a good idea whether it make sense to keep
it around for the life of the global object.  It seems like the default answer
should be to tie it to its owner object, though.

Where do the JSC bindings do this?  I didn't immediately see it.