From: Henry Spencer (henry@spsystems.net)
Date: Wed Jul 07 2004 - 23:17:08 CDT
On Wed, 7 Jul 2004, Russ Allbery wrote:
> That's obviously not the best behavior. I would be interested in what
> this group would recommend instead; I'm happy to implement the suggestion
> of the group in the next version. Should such messages simply be rejected
> if no Control header is present?
I would say "reject", like late versions of C News.
> I think we should restore that statement. I'm ambivalent as to whether it
> should say MUST NOT or SHOULD NOT.
Likewise (to both).
> > Likewise, the presence of "all.all.ctl" in the Newsgroups-content...
>
> I'm not sure where you got the "all.all" from.
Straight from RFC 1036, actually! If one believed 1036 (or 850 before
it), the .ctl suffix was recognized only as the third component of a
three-component newsgroup name (although this could have been stated in a
more implementation-independent way). With the possible exception of
B News, I don't think anyone ever actually implemented that restriction.
> I'm moderately sure that no implementations after C News implemented "all"
> as a wildcard or "ctl" as an indicator of a control message. I know INN
> never did.
It's just conceivable that somebody else might recognize .ctl based on
RFC 1036. But I think you're correct that C News was the last user of
the "all" wildcard. We did it because we thought being (approximately)
control-file-compatible with B News was desirable. (And it turned out
that we had to put considerable effort into nailing down the semantics of
the B News newsgroup patterns precisely, so we could implement them
correctly -- B News never did define them properly.)
Henry Spencer
henry@spsystems.net