[py3] Fixed another regression from 2892cb0e.

        if self.has_header('Content-Encoding'):

            def make_bytes(value):

                if isinstance(value, int):

                    return six.text_type(value).encode()

                elif isinstance(value, six.text_type):

                    return value.encode('ascii')

                else:

                    value = six.text_type(value)

                if isinstance(value, six.text_type):

                    value = value.encode('ascii')

                # force conversion to bytes in case chunk is a subclass

                return bytes(value)

            return b''.join(make_bytes(e) for e in self._container)

        return b''.join(smart_bytes(e, self._charset) for e in self._container)

    def __next__(self):

        chunk = next(self._iterator)

        if isinstance(chunk, int):

            return six.text_type(chunk).encode()

            chunk = six.text_type(chunk)

        if isinstance(chunk, six.text_type):

            return chunk.encode(self._charset)

            chunk = chunk.encode(self._charset)

        # force conversion to bytes in case chunk is a subclass

        return bytes(chunk)

    next = __next__             # Python 2 compatibility