[ On Mon, November 9, 1998 at 16:58:41 (+0000), Malcolm Ray wrote: ]
> Subject: Re: [EXIM] Temporary rejections
>
> Ok, let's put it another way. 554 is not among the list of reply codes
> RFC 821 gives as permissible immediately after a connection is accepted,
> but 421 is allowed. I'd be more comfortable using 421 there...
No, 554 is not one of the *explicitly* mentioned codes for connection
establishment failures in RFC 821 section 4.3. However if you accept
that table as the one and only canonical list of acceptable codes, then
you give up on the extensibility of reply codes as documented in RFC 821
Appendix E.
However you're right that most client's won't behave properly if a 554
is sent by the server immediately after the connection opens. It's much
better for the server to wait for the first command and then reply. In
fact the server *should* continue to accept commands even after it's
sent a 5xx series reply and simply continue to send the same reply if
the purpose is to give a permanent failure because it's the remote
client that's being rejected, not the actual message. The only remote
clients this form of "soft" client-specific rejection seems to fail on
is qmail and perhaps some versions of Netscape's mail server. Qmail
stupidly keeps trying to deliver the message until it times out, never
beliving 5xx error codes to HELO and MAIL FROM: and RCPT TO: commands.
(Though I must admit I've only tried sending 50x after an initial 55x.)
I can't think of any reason for a client to behave any differently for a
5XX vs. 4XX in response to a connection establishment (keeping a list of
hosts that reply with 5XX and *never* contacting them again would be, as
Philip says, rather pointless).
--
Greg A. Woods
+1 416 218-0098 VE3TCP <gwoods@???> <robohack!woods>
Planix, Inc. <woods@???>; Secrets of the Weird <woods@???>
--
*** Exim information can be found at
http://www.exim.org/ ***
