Re: Re: [evolution-patches] Possible fix for #322016

[Jeffrey Stedfast <fejj novell com>]

On Tue, 2005-12-06 at 08:07 +0000, karllinuxtest relton ntlworld com
wrote:
> > 
> > From: Jeffrey Stedfast <fejj novell com>
> > > 
> > > Fundamentally the only way I can see to truely fix these
> > > is for the notifications to be sent after a 'fetch mail'
> > > operation, i.e. when brand new mail truely arrives into
> > > Evolution.
> > 
> > The problem is that new mail can arrive more than just after a 'fetch
> > mail' op. For example, new mail can arrive at any time for IMAP and
> > possibly other remote mail providers.
> > 
> 
> Right - I see what you mean. Presumably though there is
> a place in the code where the newly arrived mail (from
> whatever source) is 'added' to the system (e.g. the
> message-id will be recorded somewhere). If that is a single
> place in the code, that might be a better place to
> trigger the notifications.

well, yes and no. each provider implements this themselves (or can).

> 
> The problem with the folder update code I have been
> looking at is that there seems no way of knowing if
> it is truely new mail into the system, or just the user
> shuffling messages from one folder to another.

oh how you are in for a bag of treats...

in the case of imap, for example, there's also no way to know if a
message that arrives in a folder is actually "new" or if it has just
been moved from another source either (in both cases a message has the
\Recent flag), and that's a more-or-less unsolvable problem which makes
this whole problem more-or-less unsolvable.

>  Therefore
> no amount of checks or flags will eliminate all false
> notifications.

indeed, but as I've explained above the problem is not internal to
Evolution itself.

-- 
Jeffrey Stedfast
Evolution Hacker - Novell, Inc.
fejj ximian com  - www.novell.com