Re: [Evolution] Using an IMAP Outbox

[Patrick O'Callaghan <poc usb ve>]

On Mon, 2009-06-01 at 10:20 +0200, Jerome Warnier wrote:
> Patrick O'Callaghan wrote: 
> > On Sun, 2009-05-31 at 22:01 +0200, Jerome Warnier wrote:
> >   
> > > > Then the "sent folder" preferences could be left unchanged and a
> > > > outgoing filter "match all" and "copy to folder" could be used. The
> > > >       
> > > key
> > >     
> > > > trick still is to make Evolution believe that sending was
> > > >       
> > > successful.
> > >     
> > > The other main goal of using an "Outbox" folder is to avoid sending
> > > your
> > > e-mail twice through the link (one using SMTP, and the second one
> > > using
> > > IMAP to store it in the "Sent" folder). This solution does not.
> > >     
> >
> > How does the use of an Outbox avoid this? At first glance it seems to me
> > that if both the destination and the Sent folder are non-local, the
> > message is going to end up being copied twice in any case (even when the
> > destination and the folder are on the same physical machine). I think
> > IMAP protocol extensions have been proposed to get round this, but they
> > aren't widely implemented AFAIK.
> >   
> It allows to avoid sending twice because, instead of sending through
> SMTP, it sends once to the server using IMAP (in the Outbox folder),
> where the server notices the mail and sends it using its local MTA,
> then copies it (locally) to the right Sent folder.
> That way, you avoid sending twice through the link between you and
> your server(s). Great for road-warriors or tele-workers.
>
> It is true that it is not widely implemented (yet). However, as it is
> quite simple to do, maybe the missing incentive is that not all MUA
> support it...
> Courier(-IMAP): some more information here:
> http://www.inter7.com/courierimap/INSTALL.html.

As far as I can tell, this is an ad-hoc solution which for one thing
changes the user's way of working. I can't say I like it. What I was
referring to is a set of IMAP *protocol* extensions to allow one to
avoid the redundant copy, plus other nice things like being able to
forward a message without having to download it first.

There's a whole bunch of RFCs on this kind of stuff at
http://www.imap.org/papers/biblio.html. Note that the most recent of
them is three years old. Things move slowly when you want to change a
deeply-rooted protocol, even when all you're doing is to extend it.

poc