Les Mikesell wrote:
> Is there an RFC requirement that the connecting host MUST
> wait before sending anything?
Sort of; compare this sections:
http://rfc.net/rfc2821.html#s3.1
with this one:
http://rfc.net/rfc2821.html#s4.3.1
which directly contradict each other. The latter probably takes precedent,
since it is more specifically talking about the sequence of communication.
However, in practice (since the RFC's are just theory), the only client software
rude enough to speak before spoken to is spamware (or braindead software like
spong), so it has been a standard practice for years to fail clients which don't
adhere to the SHOULD rule (and the RFC's do permit it).
> Also, in a glance at RFC 821
> it did not look like a 450 was valid in the connect phase - it
> only mentions 421. Has this been changed in later RFC's?
>
According to the chart here:
http://rfc.net/rfc2821.html#s4.3.1
we could presumably return 221 after receiving QUIT as a special case, but still
reserve the right to 450 for every other command that comes before the banner is
sent.
John
