Yes. I think that RFC2633bis out to say that compression MAY be applied
before encryption or signing. It should probably indicate that compression
should only be applied once for a particular message.
I think that an S/MIME capability should be used to negotiate this
beforehand, if we're going to head down this path.
My apologies for having nodded off on this thread, but why does this need to be
negotiated? All the prior discussions of this revolved around what the signer
preferred. So why can't it just be sender's choice? The encodings are
unambiguous on reception.
I'll apologize in advance in case I've just got my 'stupid hat' on.