RE: [Python-Dev] Unaligned accesses in coercions, current CSV