That ref seems to talk about the same client issuing multiple commands,
rather than multiple clients.  But from what you're saying it sounds like a
server ought to maintain the validity of the sequence numbers that it knows
a client has seen on a session, until it has told the client to change them
via untagged responses?

Do servers generally do that?

Regards,

Edward Hibbert
Internet Applications Group
Data Connection Ltd
Tel:    +44 131 662 1212                Fax:    +44 131 662 1345
Email:  [EMAIL PROTECTED]       Web:    http://www.dataconnection.com

-----Original Message-----
From: Arnt Gulbrandsen [mailto:[EMAIL PROTECTED]
Sent: Friday, July 11, 2003 6:33 PM
To: Edward Hibbert
Cc: IMAP protocol mailing list
Subject: Re: Untagged responses


Edward Hibbert writes:
> Thanks for the various replies on this.
>
> I have a follow-up question about sequence numbers. Again, apologies 
> if it's been covered before.

Many times ;)

> There's a window where a client hasn't yet picked up an untagged response
> and therefore uses a sequence number that doesn't refer to the intended
> message.  It could get round this by using UIDs - but I was wondering
> whether any servers correct the sequence numbers.
...

RFC 3501 section 5.5 regulates what happens. Basically, MSNs don't 
change while you're using them, and the server is required to preserve 
that truth.

When messages are deleted, this may require that the server bends over 
backwards, forwards, sideways, or even all at the same time.

--Arnt

Reply via email to