On Wed, 4 Mar 1992 09:35:47 -0500 (EST) you said:
<nl>Ah, this is PRECISELY the presumption I want to guarantee! My reading of
the Fontaine proposal, however, is that it REQUIRES the MUA to feed CRLF to
the MTA, which may not currently be the case.
There is no way for myself or anybody else to specify anything at that level
in the Internet mail context. Internet mail currently has no MUA or MTA
concept. There is just RFC822 describing the content of the messages, and
RFC821 specifying the dialogs between client and servers. Both speak in
terms of what's finally 'on the wire'.
I can't understand why one would be afraid of my proposal, unless one thinks
since 1982 that RFC822 actually specifies that mail must be passed between
MUA and MTA (entites that do not exist in the model, but then...) in CRLF
format ! To state it in another way : if RFC822 speaks of header lines
terminated by CRLF, and lines in the body terminated by CRLF, MIME must
speak the same languages (which it does indeed do everywhere). I did not
bring anything new in this respect.
/AF