The fact remains, however, that each encapsulation involves two CRLFs using
either BNF and any existing transport, and this does not gibe with the
statement you made previously.
Ned
Well, my comment referred to Mark's initial post which was focused on the
two CRLF's presently required before the first multipart encapsulation.
The implementations I've seen accept one or two CRLF's here; if this is
generally the case, adopting the new BNF should cause very few problems
since it is essentially the same after the first part. Anyway, if you
didn't see the initial post, I understand the confusion.
Marshall's position that we can live with things as they are (by fixing the
examples and errant code), is certainly reasonable, but I thought that this
Proposed period was intended to allow tweaking of just this sort.
Jim