Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

does pine make use of the IMAP IDLE command?

92 views
Skip to first unread message

DJ-usenet

unread,
May 26, 2005, 8:42:21 PM5/26/05
to
All -

It doesn't look like pine makes use of the IMAP IDLE command. If I have
both pine and Thunderbird 1.0.2 connected to the same IMAP server, when
new mail arrives Thunderbird notifies me of new mail right away, but
pine takes up to check-mail-interval to detect the new mail.

Am I missing some configuration in pine or it simply doesn't implement
the IDLE command?

Thank you for your help.

DJ

Mark Crispin

unread,
May 26, 2005, 9:23:11 PM5/26/05
to
On Thu, 26 May 2005, DJ-usenet wrote:
> It doesn't look like pine makes use of the IMAP IDLE command.

Correct.

The IMAP IDLE command is not a panacea. It is, in fact, quite harmful if
there is a NAT box between you and the IMAP server. It's become quite a
support issue for users who use clients that use IDLE.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.

DJ-usenet

unread,
May 27, 2005, 1:54:49 AM5/27/05
to
Why is the NAT box a problem? I would like to understand what the
issues are. For one, I use a NAT box and I'm firewalled so that no
incoming connections are allowed to my computers and Thunderbird with
IDLE enabled seems to work quite well. In fact I run TB next to pine
just for its notifications ... Thanks.

Mark Crispin

unread,
May 27, 2005, 2:00:50 AM5/27/05
to
On Thu, 26 May 2005, DJ-usenet wrote:

A session that uses IDLE may have no traffic for as long as 29 minutes.
This exceeds the timeout after which session maps are purged on most NAT
boxes. If the server sends traffic after the session map has been purged,
it will get a TCP reset from the NAT box and lose the connection.

If the client sends traffic first, most NAT boxes generate a reproducable
mapping thus restoring the mapping that worked.

The bottom line is that after a session has been inactive, server
restoration of activity is likely to provoke loss of session on a NAT box.

NAT boxes basically assume that everybody does client-based polling, which
is effectively what happens in an IMAP client which doesn't use IDLE.

DJ-usenet

unread,
May 27, 2005, 2:19:51 AM5/27/05
to
But an IMAP client could still send some other command (NOOP?) every 5
min. or some other reasonable amount of time just to keep the
connection alive, right? I guess that's what my TB is doing as I have
it set to also check for new mail every 2 min. (I didn't bother to look
at the traffic, it's kinda hard to do over SSL IMAP, I could run it
over a plain connection or enable debugging in TB and see what it's
doing).

Mark Crispin

unread,
May 27, 2005, 12:14:17 PM5/27/05
to
On Thu, 26 May 2005, DJ-usenet wrote:
> But an IMAP client could still send some other command (NOOP?) every 5
> min. or some other reasonable amount of time just to keep the
> connection alive, right?

The whole point of using IDLE is not to send periodic NOOPs. When IDLE is
being used, a command is in progress, thus NOOPs are not (can not be)
sent.

I think that you misinterpret the purpose of IDLE. There is no reason to
believe that IDLE gets you new mail notifications any faster. It *may*,
but only if the server's new mail announce interval through IDLE is faster
than the client's polling interval when not using IDLE.

The purpose of IDLE is to prevent client polling. But, if you use a NAT
box, the client must poll. IDLE was a hack to reduce network traffic, but
that was before NAT.

DJ-usenet

unread,
May 27, 2005, 4:56:50 PM5/27/05
to
Makes sense. Sorry for the basic questions, I'm new to IMAP and tyring
to learn as fast as I can ... Thanks for your detailed answers.

0 new messages