Various XHR send tests are yielding null bytes sent instead of the correct values (e.g. 0).
I think we might be passing these now.