[webkit-reviews] review denied: [Bug 16588] ResourceHandle::loadResourceSynchronously needs access to the FrameLoader : [Attachment 18077] Add FrameLoader* as parameter to the function

[bugzilla-daemon at webkit.org]

Darin Adler <darin at apple.com> has denied Holger Freyther
<freyther at handhelds.org>'s request for review:
Bug 16588: ResourceHandle::loadResourceSynchronously needs access to the
FrameLoader
http://bugs.webkit.org/show_bug.cgi?id=16588

Attachment 18077: Add FrameLoader* as parameter to the function
http://bugs.webkit.org/attachment.cgi?id=18077&action=edit

------- Additional Comments from Darin Adler <darin at apple.com>
If you are going to add a parameter, I suggest adding a Frame* rather than a
FrameLoader* to match ResourceHandle::create.

But since this is a layering violation, we should figure out a cleaner way to
do this.

Perhaps we should have a ResourceHandleFactory class and FrameLoaderClient can
return a ResourceHandleFactory. Then you'd use the ResourceHandleFactory to
make the ResourceHandle objects.

Or we could have a class with a name like ResourceLoadingContext, add a
FrameLoaderClient function that returns one, and we could pass that in to the
static members of ResourceHandle.

A typical client could implement both interfaces on the FrameLoaderClient and
it could just return "this". This might be enough allow us to remove the Frame*
parameter from both functions and remove the circular dependency.