[Top] [All Lists]

Re: Format=Flowed/RFC 2646 Bis (-02)

2003-11-16 18:55:50

Cyrus Daboo <daboo(_at_)cyrusoft(_dot_)com> writes:

Hi Simon,

--On Saturday, November 15, 2003 2:42 +0100 Simon Josefsson
  <jas(_at_)extundo(_dot_)com> wrote:

| Receive (MIME-aware, PGP aware): CTE decode PGP armor, verify, CTE
|   body, display.

Sorry - but this is a non-starter - you are expecting existing
deployed clients to magically change their behaviour to cope with
format=flowed inline signed messages.

No, I expect clients to follow specifications, which I think is about
all we can hope for.  If deployed clients follow section 4.6 of RFC
2646 they already treat inline format=flowed differently.  They would
also be incompatible with both our proposals.  I don't believe many
clients follow the recommendation in the RFC though.  When fixing the
RFC text, we might as well evaluate all options available.  If all
deployed clients implemented format=flowed inline PGP in the same way,
and that way worked, then of course we should change to that method.
But my understanding is that this isn't the case.

The only way to ensure format=flowed inline signed messages work
with existing clients is to use the procedure I outlined

It would not work with clients that follow RFC 2646.  It would not
work with any client that compute the OpenPGP signature over the QP
encoded ("wire") data.  Are there no such clients?  I believe there
are gateways and some plugins that fall in the latter category.  But
if I'm mistaken, it may be worthwhile to add a description of your
proposal to 2646bis, to document earlier practices.  Of course, the
recommended solution should still be PGP/MIME.