maxage causes loss of local email

[Nicolas Sebrecht]

On Fri, Mar 20, 2015 at 04:18:26PM -0400, Janna Martl wrote:

> OK, this is clearly a lot better. See the patch I sent. (I was also
> unhappy about the other one, because it was hard to even state what
> its behavior was.)

Yeah, we finally got it... :-)

> I did notice another issue, though -- quickchanged (the -q option)
> doesn't do any good if you're using maxage. I left this alone for now;
> the problem is that the "quick" method of seeing whether an IMAP
> folder has changed involves getting a list of all the message
> sequences numbers, and checking if that has the same number of items
> as the statusfolder. If we're using maxage, the statusfolder only
> contains recent messages, so we somehow want to reduce this message
> sequence # list to recent messages, but it takes longer (another IMAP
> operation) to convert these into UIDs (or get dates from them). I
> don't really have any ideas...

Good catch.

Making IMAP requests more than once would mean we do the exact opposite
of what the user expect if he's looking for speed.

Both strategies looks not compatible, ensure they aren't used at the
same time.  IOW, I would make it very simple: ignore quickchanged if
maxage is involved (the opposite ― ignore maxage if quickchanged is
given ― would be bad).

The user will have to make a choice about what strategy he wants to use
to get it fast.

-- 
Nicolas Sebrecht