+ # We don't really know what encoding each file might be following,
+ # so just open the file as a byte stream. We only need to search
+ # for a pattern that should be the same regardless of encoding,
+ # so that should be good enough.
+ fd = open(filename, 'rb')