On Tue, Oct 31, 2006 at 12:16:02PM -0800, Brendan Cully wrote:
On Tuesday, 31 October 2006 at 15:07, Bob Bell wrote: > On Tue, Oct 31, 2006 at 09:08:08AM -0800, Brendan Cully wrote: > Are you expecting to see a "OK [UIDNEXT" reply to a "SELECT"? There's > code in imap.c that's at least handling it, but my IMAP server isn't > saying it. Here's a very brief conversation with my IMAP server > (provided to illustrate what Courier says in response to SELECT): Oh. Yes, I am. It's required by RFC 3501 (section 6.3.1). There's probably a way to get mutt to synthesize the value though.
Now that I know what I'm looking for, I found the following in the
ChangeLog:
2006-03-15 12:26:00 Brendan Cully <brendan@xxxxxxxxxx> (brendan)
* imap/message.c: Don't store UIDNEXT unless we know it.
* imap/message.c: Only synthesize UIDNEXT if we've fetched
uncached messages. Fixes a FETCH 1:0 bug Bruno Postle reported and
helped to debug.
2006-03-13 23:21:51 Brendan Cully <brendan@xxxxxxxxxx> (brendan)
* imap/message.c: Courier fails to send UIDNEXT on SELECT
(a violation of RFC 3501 section 6.3.1) - synthesize it if it
is missing.
* init.h: Make imap_idle default to false for 1.5.12.
The two relevant changes can also be seen as git commitdiffs at:
http://tinyurl.com/yfp9o6
http://tinyurl.com/ya7uo8
So it seems mutt is already trying to handle this case, though obviously
failing. The change seems to be tied into the header cache. Perhaps
the current UIDNEXT synthesis is not adequate?
-- Bob