BitVector::m_bitsOrPointer stores a transmogrified OutOfLineBits pointer, so it's always reported as a leak: m_bitsOrPointer = bitwise_cast<uintptr_t>(newOutOfLineBits) >> 1;
<rdar://problem/35767785>
Silenced the leak in <http://trac.webkit.org/r225321>. It still needs to be fixed, as it will continue to show up as a leak for client apps which don't use webkitpy for leaks detection.
I'm not seeing many leaks under BitVector::resizeOutOfLine too. I don't see any recent changes to this code, so perhaps that's just because I ran a more complete set of tests. Adding BitVector::resizeOutOfLine to _callstacks_to_exclude_from_leaks too.
> I'm not seeing I'm *now* seeing