----- Original Message -----
From: "Richard O. Hammer" <ROHammer(_at_)earthlink(_dot_)net>
Sent: Saturday, January 03, 2004 9:58 PM
Subject: Re: what to say on timeout?
CORRECTION: I meant to refer to RFC 1939, not 1938.
Richard O. Hammer wrote:
I have looked in RFC 2821 but not yet found any explicit text which
tells what to do in this case. But, for a comparison, RFC 1938 (dealing
with POP3) Section 3 says:
A POP3 server MAY have an inactivity autologout timer. ...
When the timer expires, the session does NOT enter
the UPDATE state--the server should close the TCP connection without
removing any messages or sending any response to the client.
hmmm, I guess I can see where one can get the idea that you can send a
unsolicited state machine response to a client, but this has to do with the
successful completion and acknowledgement of a POP3 download session to
trigger the POP3 server logic to perform the final mail database updates
(i.e., mark downloaded mail as received).
In other words, you only want the server to perform the UPDATE when the
QUIT command is sent by the client. (RFC 1939, section 3, para 4)
If the QUIT command is not sent by the client, the SERVER should not perform
any mail database update because it could mean there was an unexpected drop
in the connection. That should include an inactivity timeout.
No QUIT command, no update. If I recall one client, an old version of the
Microsoft Outlook for the MAC that would drop the connection after the
download. This would cause a problem with users who didn't complete the
download and tries again. MS was quickly made aware of the RFC spec bug and
an update was provided.
Hope this helps.
Hector Santos, Santronics Software, Inc.