> "Conforming implementations MUST NOT recognize [any] other
> character or character sequence as a line terminator."
> If you agree with the Postel philosophy, and if we read that sentence
> the same way, then I would expect you to say we should not 501 upon
> receipt of a misplaced CR or LF; that may be a time to be lenient.
Technically, the input stream could allow them, but then we would have to
check all of the other code to make sure that we disallow them in the
argument. Look at the BNF. There are no cases that permit CR or LF. In
fact, the next part of 2.7.1 says,
SMTP client implementations MUST NOT transmit
[<CR> and <LF>] characters except when they are
intended as line terminators and then MUST, as
indicated above, transmit them only as a <CRLF>
sequence.
We are rejecting them because they are invalid characters in an SMTP
command. A related question is whether we should check for other illegal
characters in the same place, rather than requiring each command to handle
them separately. But that can get a bit tricky, chasing them down through
the RFCs (since SMTP Extensions can change the rules), whereas CR and LF are
universally excluded.
--- Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
