Hi Kyle,

On Thu, 14 Oct 2004 12:13:48 -0500 UTC (10/14/2004, 12:13 PM -0500 UTC my
time), Kyle Wheeler in part wrote:

K> The problem is, programs like Bruce's relay-control can only decide that
K> you are still authenticated based on when you send a username and 
K> password to the IMAP server---it doesn't (and can't, really) keep track 
K> of every single IMAP command.

of course, and you and Jeremy bring up some very valid, well thought out
points.

K> This is why quitting and re-opening the mail client works---because
K> you're forcing it to re-do the "send username and password" part of the 
K> IMAP conversation---but long-running IMAP sessions eventually become 
K> un-authenticated with respect to imap-before-smtp (like Bruce's 
K> relay-control).

true, true... one other factor is that smtp auth must be patched into qmail,
while I prefer the add-on route of relay-ctrl..

small point, but I cannot think of one MUA other than Mutt (maybe Pine)
where one has to quit and re-start a client to re-access an IMAP account...
one can just quickly log out and log in again (disconnect /reconnect from
the server) without closing the client to re-auth on imap-before-smtp. Much
faster this way.. almost instantaneous in fact.

K> Pop3-before-smtp makes a *little* bit more sense than imap-before-smtp, 
K> but only because pop3 connections are not persistent, like imap 
K> connections are, so a good mail client will be re-authenticating 
K> periodically every time it checks for new mail, while an imap client 
K> knows it can just keep one connection open and not re-authenticate every 
K> time it checks for new mail.

indeed... most MUAs have auto provision for setting pop before smtp..
However, I have never had a complaint from any user, on any server to
date with my method... ( <g>... well trained users)..


-- 
Gary


Reply via email to