Switch chrome and chromiumDRT over to using the new WebKit API for dispatching events.

content/renderer/renderer_webstoragenamespace_impl.cc

Index: content/renderer/renderer_webstoragenamespace_impl.cc
===================================================================
--- content/renderer/renderer_webstoragenamespace_impl.cc	(revision 133749)
+++ content/renderer/renderer_webstoragenamespace_impl.cc	(working copy)
@@ -27,10 +27,6 @@
 
 WebStorageArea* RendererWebStorageNamespaceImpl::createStorageArea(
     const WebString& origin) {
-  // Ideally, we'd keep a hash map of origin to these objects.  Unfortunately

[ericu, 2012/04/25 19:40:35]

Is this comment obsolete?  Or is it just that it's true enough that it's
pointless to have it?

[michaeln, 2012/04/25 20:22:12]

Yes and yes.

Obsolete. I think it would be counter productive to share a ptr to the same area
instance for different consumers. If we did so, there's be no way to identify
the Document associated with an area instance and it would make targetting the
events more difficult.

And the comment was kind of not-so-useful chit-chatty anyway even if it would
have been an improvement to refcount these interface ptrs.

-  // this doesn't seem practical because there's no good way to ref-count these
-  // objects, and it'd be unclear who owned them.  So, instead, we'll pay the
-  // price in terms of wasted memory.
   return new RendererWebStorageAreaImpl(namespace_id_, origin);
 }
 
@@ -42,6 +38,9 @@
   return NULL;
 }
 
-void RendererWebStorageNamespaceImpl::close() {
-  // TOOD(michaeln): remove this deprecated method.
-}
+bool RendererWebStorageNamespaceImpl::isSameNamespace(
+    const WebStorageNamespace& other) const {
+  const RendererWebStorageNamespaceImpl* other_impl =
+      static_cast<const RendererWebStorageNamespaceImpl*>(&other);
+  return namespace_id_ == other_impl->namespace_id_;
+}