Re: [Evolution] Keeping states

[Patrick O'Callaghan <poc usb ve>]

On Thu, 2005-10-27 at 21:04 +0930, Iain Buchanan wrote:
> On Thu, 2005-10-27 at 07:15 -0400, Patrick O'Callaghan wrote:
> > On Wed, 2005-10-26 at 17:05 -0400, Eric Preston wrote:
> > > Patrick O'Callaghan said:
> > >
> > > > Except that the message might no longer be there. People use a variety
> > > > of mailers, including multiple instances of Evo (I do, though they're
> > > > not concurrent). As long as the state is kept local to each instance and
> > > > not on the mail server, you can't know what happened between one
> > > > activation and the next, so you need a fallback. 
> > >
> > > I think you're clouding the issue here. If you're using multiple mailers
> > > (or instances thereof) you've got really unrealistic expectations for
> > > any of them to keep reasonably sane state information. Defaults should
> > > accomodate common case(s). My common case is only
> > > reading .evolution/mboxes with one instance of Evolution.
> >
> > First of all I was trying to say that I *didn't* expect Evo to handle
> > this (in fact I don't know of any mailer that does). I was talking about
> > the "reasonable fallback". Sorry if that wasn't clear.
> >
> > Secondly, I think nowadays it's far from unusual for people to read mail
> > from more than one location (at least home and office, probably more if
> > they travel a lot). It's just a fact of life.
>
> Yes, but this is usually from different machines, so scroll position
> won't matter because its not stored in your pop or imap folder (is it?).
> Unless they've got a laptop, then you may use the same machine in
> different locations, but then the location doesn't matter.
>
> As far as I understand the issue, I think Eric was talking about
> different mail clients on the same pc, which would cause a problem.
>
> IMAP may turn it into a problem, because you may delete a message on one
> machine that your other machine is keeping state for.  In which case you
> can default to scroll location.  So, when entering a folder, jump to:
> 1. last exact email viewed
> 2. if not there, then last scroll position
> 3. if not possible, then very last or very first message, depending on
> users pref.
>
> Just MHO :)

We seem to be going round in circles since we're all saying the same
thing, to wit: you can't rely on saved folder state since it's local to
your machine. In fact, as you point out, it's local to your specific
mailer on your machine, since there's no standardization of these
things.

Neither is there a standard for storing this stuff on IMAP servers (at
least there isn't one that anyone pays any attention to AFAIK).

Therefore you need a reasonable fallback when the saved state doesn't
match the real folder.

The only point at issue is what do we mean by "reasonable fallback"? But
even there we seem to agree!

poc