In Bug 572331, comment 48 and 49, I tasked myself with the job of double-checking that the old code for clearing the weak ref table during finalization had been made redundant and then removing the redundant code. This is a ticket to track that work, once I or someone else gets around to it.

Is this a dup or an anti-bug of bug #647155? If it's an anti-bug I would like to discuss it further before anything gets done with it. IMO we want weak ref checking *out* of GCAlloc::Finalize, but to keep the clear unmarked weak references loop called from GC::Finalize, because that takes us in the right direction for finalization / incrementality - weak references need to be handled different than other "finalizable" objects, because we can specialize their handling easily.