1) An attribute that converts lone surrogates to U+FFFD as per http://lists.w3.org/Archives/Public/public-script-coord/2012OctDec/thread.html#msg82
2) An attribute to replace ByteString. We should probably not have ByteString as a distinct type per the arguments from Brendan in the thread above.
I would actually prefer a distinct type for ByteString. On the C++ side, ByteString needs to become some version of char*, while DOMString becomes a Unicode string. These are typically different C++ classes, not just different conversion code. And while it's possible to change the actual type of the object based on extended attributes, it feels pretty weird.
*** This bug has been marked as a duplicate of bug 24581 ***