In article <9gnfhs$a5311$[EMAIL PROTECTED]>, Rolf Krahl
<[EMAIL PROTECTED]> wrote:
> In article <[EMAIL PROTECTED]>,
> Chuck Simmons <[EMAIL PROTECTED]> writes:
> >
> > RFC1036 states that to conform to RFC822, a message-ID must contain the
> > host domain where the message entered the network. This does not suggest
> > server generation but rather client generation. RFC1036 otherwise does
> > not state whether message-ID's should be client or server generated.
>
> Yes and the domain the message entered the network is _not_ yahoo.com
> or gmx.net. Thus the message id may not use any of these domain
> names. IMHO, it was acceptable nonetheless with the explicit consent
> of the owner of the name space, but that is usually not the case for
> mozilla's message ids. At least, i have certainly no permission by
> GMX to generate message ids within the domain gmx.net.
If you have an email address in that name space, then you have implicit
permission to create both mail and news message-id's in that name space
(it's part of the package).
If you're munging the From header, then bogus message-id's are the
least of your problems.
-snip-
> Yes, but this doesn't matter, as long as the server that generates the
> message id is allowed to do so and guarantees uniqueness within the
> name space it uses. There is no need for having the message ids in
> the name space of the hostname that the user used to contact the
> server.
Having the client generate the message-id has advantages that far out
weigh any supposed loss of "guaranteed" uniqueness.
--
J.B. Moreno
