Re: mailutils/IMAP (2)

[Jakob 'sparky' Kaivo]

"Alain Magloire" <address@hidden> writes:

> > 
> > [Alain Magloire]
> > 
> > > imapd(UW c-client) also suppress a special message, this message serves
> > > as a place holder for uivalidity and other things but I do not think
> > > sending a message like that to an imapd will generate problem, I believe
> > > it is discarded .. not sure.
> > 
> > I think the "Folder internal data" message needs to be the first
> > message in the mailbox for c-client to understand it.
> > 
> > The "store state in a faked message" scheme breaks down big time when
> > users access mailboxes via multiple paths, and not all paths "hide"
> > the internal message.  For that reason I don't think too highly of it.
> 
> Interesting comments.
> Care to elaborate a little when you say "multiple paths".

I believe he means something like this scenario:

I connect to my mail server via IMAP. It is running UW IMAP, so the
fake message gets generated and stored. I leave mail on the
server. Later, I ssh in and use mutt to read my mail locally. Mutt
doesn't know the difference about IMAP's "special message", so it
displays it along all the others. I don't like the "DON'T DELETE THIS
MESSAGE" message in my mailbox, so I delete it. IMAP has lost its
state.

The problem with the c-client approach is it only works when *all* of
your mailbox access goes through c-client. It should be possible to
maintain state while still allowing foreign M[UTD]A access.

-- sparky