Larry Osterman (larryo@Exchange.Microsoft.com)
Mon, 17 Dec 2001 17:23:45 -0800
Then you're right - we're talking at cross purposes - the "message is
assigned..." stuff I wrote was particular to the Exchange implementation
of IMAP. The spec's comments are totally correct however from a
conceptual standpoint - and the wording is critical - there are many
ways a message can be added to a mailbox, and all of them assign new
UIDs.
-----Original Message-----
From: Pete Maclean [mailto:aaddict@maclean.com]
Sent: Monday, December 17, 2001 4:36 PM
To: Larry Osterman; Cyrus Daboo
Cc: ietf-imapext@imc.org; vpim@lists.neystadt.org
Subject: RE: Some thoughts about IMAP and Unified Messaging
>Um... No, this NOT IMAP poking it's nose where it doesn't belong.
It's
>a critical part of the design of the offline mode of IMAP.
>
>IMAP clients depend on the fact that UIDs are assigned to messages that
>are constant and are strictly increasing (note: NOT monotonically). If
>this constraint is violated, then a client does not have the ability to
>resynchronize with the server.
We may be talking at cross purposes. I believe I understand the role of
UIDs perfectly and appreciate their enormous value in the overall IMAP
design. I was taking issue only with the question of exactly when they
must be assigned. I cannot see that it matters if a message is not
assigned an IMAP UID if IMAP never sees that message.
On the other hand, there are so many ways to interpret things that this
may
all be moot. My IMAP server assigns UIDs to messages that don't have
them
when it opens a folder. I tend to think of a message being added to a
folder when a transport provider or client puts it there. But we could
say
that a "MAPI folder" virtually becomes an "IMAP folder" when and only
when
the server opens it and, in that virtual sense, new messages are added
and
assigned UIDs only at that point. In which case I am following this
part
of the spec to the letter.
Pete
This archive was generated by hypermail 2.0b3 on Tue Dec 18 2001 - 03:24:00 IST