Re: [Evolution] Evolution generating invalid message-ids

[Jeffrey Stedfast <fejj ximian com>]

s/application/host

that is what the rfc says, it does not say at all that message-ids are
unique *accross* hosts, because that is impossible to do.

even if it were guarenteed that you *could* do that, clients would still
have to handle cases where messages had the same message-id but were not
identical. Jamie Zawinski wrote a comment/doc/something about how you
could exploit clients that used the message-id as if it were unique.

Jeff

On Fri, 2002-07-26 at 18:04, Jeffrey Stedfast wrote:
> On Fri, 2002-07-26 at 16:55, Jörg Roßdeutscher wrote:
> > Hi,
> >
> > Jörg Roßdeutscher:
> > > > ...it has to be unique. 
> > > > And other programs do it that way.
> >
> > Jeffrey Stedfast:
> > > Please read further and note that unique means only unique to that
> > > particular machine that generates it.
> >
> > I think you read wrong by mistake.
> > It doesn't say 
> > "unique means only unique to that particular machine that generates it".
> > It says:
> > "The uniqueness of the message identifier is guaranteed *by* the host
> > that generates it (see below)."
> >
> > "By", not "to". ("*" put in by me)
> >
> > And it says "see below", and there it says:
> >
> > "The message identifier (msg-id) itself MUST be a globally unique
> >  identifier for a message.  The generator of the message identifier
> >  MUST guarantee that the msg-id is unique.  There are several
> >  algorithms that can be used to accomplish this."
>
> *sigh*
>
> No, you are wrong. You can only ever guarentee uniqueness within your
> own application space.
>
> > I think it's expressed very clear. 
> > You don't have to generate an id. (See previous mails)
> > But if you do so, it must be globally unique.
>
> globally unique to the application.
>
> > > > must not generate one and leave this job for the servers.
> > >
> > > many servers will not accept messages from the client unless there is a
> > > message-id, in fact this was a bug reported a long while back.
> >
> > I never heard that before, and it would be a wrong behavior. See RFC we
> > are talking about.
> > But it's not a big thing. All we need is a little button: 
> > [ ] Generate MsgID
>
> hell no.
>
> > > besides, the CVS versions of Evolution resolve the local hostname to use
> > > the FQDN, so your make-believe problem should be gone.
> >
> > That means I had to use a probably buggy/instable developer version for
> > being RFC-compliant? And I still had to deal with ohter people's
> > messages, that do not use a devel-Version.
> > By the way: It would be the wrong result again.It would put in the
> > not-unique local domainname. 
> >
> > Inserting the local FQDN makes the MsgID not globally unique. Half the
> > world seems to use "@localhost" or "@local". If you really want to
> > generate the MsgID yourself, things get even more difficult, because
> > FQDN cannot be found out automatically but requires another option for
> > the "real" FQDN configured manually.
> > I.e. in my network: hostname --fqdn is "ratti.gesindel.local", but valid
> > MsgID must contain "@gesindel.de" (if my server generates it) or
> > "t-online.de" (if my provider's server does it)
>
> give it up, you are just plain wrong. you have no idea what the hell you
> are talking about. Please stop wasting our time.
>
> Jeff
>
> -- 
> Jeffrey Stedfast
> Evolution Hacker - Ximian, Inc.
> fejj ximian com  - www.ximian.com
>
>
> _______________________________________________
> evolution maillist  -  evolution ximian com
> http://lists.ximian.com/mailman/listinfo/evolution
-- 
Jeffrey Stedfast
Evolution Hacker - Ximian, Inc.
fejj ximian com  - www.ximian.com