Re: [DFDL-WG] implicit alignment of complex types