At 5:33 PM 2/27/95, Keith Moore wrote:
In my experience, this isn't caused by user agents, but by mail
transport in environments that don't like zero-length records.
Whatever. It remains one of the "common problems, mistakes and errors that
are made in electronic mail", which the draft purports to discuss.
I think this falls in the same category as SPACEs after MIME boundary
markers: they shouldn't be emitted by mail composers, but mail readers
need to be able do to deal with them.
'Deal with them' how? Using $ to mark line ends, this header:
foo: bar$
$
bar: foo$
$
is an RFC 822 header equivalent to
foo: bar$
bar: foo $
$
If by 'deal with them' you mean 'treat "CR LF SPACE CR LF" as "CR LF CR
LF"', then you have changed the meaning of the above fragment from a header
into a header and a couple of body lines. Fine by me (I've certainly never
seen "CR LF SPACE CR LF" used to mean what it's supposed to mean), but the
change would certainly need to be documented.
If by 'deal with them' you mean 'treat "CR LF SPACE CR LF stuff" as "CR LF
CR LF stuff" provided "stuff" doesn't look like more headers' then I am far
less amiable. BLEAH. That's quite a lot of forgiveness.
--
Steve Dorner, Qualcomm Incorporated. "Oog make mission statement."