Re: [Imap-protocol] Condstore, Gmail and mod_seq

(Replying back to list so that others can benefit from this discussion.)

On Monday, 5 January 2015 17:23:25 CEST, Daniel Piché wrote:
> Thanks for the response.
> I'm asking if the mod_seq can become invalid at some point. For example, if
>  UIDVALIDITY changes.

Yes, when UIDVALIDITY changes, all information which you have for that 
mailbox is now invalid. That includes MODSEQs of individual messages, as 
well as the HIGHESTMODSEQ of the mailbox, and many other things, including 
message UIDs and (by extension of that) everything you might remember about 
that mailbox.

> I.e. If we try to do a "CONDSTORE CHANGEDSINCE 123456" one year from now,
> will it still work? Or if something can make old mod_seq invalid.

If UIDVALIDITY is the same, then that call is guaranteed to never skip a 
thing. It is however not required to only report back items which actually 
changed for obvious reasons (the server-side storage is limited). You might 
get back data about extra messages which actually have not changed since 
then. The more often you sync, the lesser amount of false positives you're 
going to get.

> We're using condstore to add collaboration and share messages between
> different accounts ( keeping.com ). Doing resyncs ("CONDSTORE CHANGEDSINCE
> 1") has become too costly since some accounts have hundred of thousands of
> messages.

What exactly are you syncing? Do you use QRESYNC? What operations (exactly) 
are costly for you?

Cheers,
Jan

-- 
Trojitá, a fast Qt IMAP e-mail client -- http://trojita.flaska.net/
_______________________________________________
Imap-protocol mailing list
[email protected]
http://mailman13.u.washington.edu/mailman/listinfo/imap-protocol