From: Charles Lindsey (chl@clerew.man.ac.uk)
Date: Thu Jul 01 2004 - 16:10:19 CDT
In <871xjxuqm5.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>Henry Spencer <henry@spsystems.net> writes:
>> I have to concur on this. Distasteful though it is, cmsg hasn't
>> entirely gone away, and it's a genuine interoperability problem. We
>> don't want to sanction its continuing use -- it *is* a botch -- but
>> forbidding Subject content to begin with "cmsg " is desirable.
>I'm not sure that I actually disagree with this approach, but I think we
>should think a little about whether banning cmsg is the best transitional
>step forward. Currently, it is standard practice for all control messages
>to be generated with both a Control header and Subject: cmsg. It's not
>clear to me that forcing all control message generators to change to a
>different subject gets us any closer to where we want to be.
I had always understood that any cmsg in Subject was always ignored if a
genuine Control-header was present. IOW
Subject: cmsg newgroup example.foo
Control: newgroup example.bar
was guaranteed to create example.bar only.
Anyway, for anything bad to happen, two things are required:
1. Some servers need to exist that still act on 'cmsg'
2. Some administrator needs to issue such a 'cmsg' without a proper
Control-header.
Whilst there are likely a lot of servers that satisfy #1 (but probably
more that do not), has anyone ever seen a hierarchy admin doing #2? I
doubt it very much. Would isc.org accept it into its control message
archive, and would it enter it into its public Active file? I doubt it, so
there is no real harm done in practice.
So the only case we need to worry about is the odd poster who tries to do
Subject: cmsg cancel <my-article@foo.example>
because he doesn't know how to use the cancel feature on his user agent
(or it hasn't got one).
What are the consequences of that? Well, maybe his article does get
cancelled at some servers. So What? That is what he wanted, so no harm has
been done. It may not have been cancelled at as many servers as he
expected, but that is sadly true of proper Control cancels these days.
And I don't think it will provide any fresh opportunities for a malicious
canceller, who would normally have the good sense to use a proper Control
cancel anyway.
>Maybe we should just ban cmsg for non-control messages?
It would be rather nasty to single out particular Subjects for forbidding
on Usenet (queue Bruce to speak of 'unstructured' headers, and here I
might agree with him - it would be a messy thing to do). You need to do
the cost/benefit analysis. Is it really better to do this messy thing,
considering the small benefit (if any) arising from it? I think not.
Just say that Subject: cmsg MUST NOT be interpreted as a control message
(which is what the current draft says) and leave it at that. Servers will
get around to catching up sooner or later (there are much more important
changes in our draft that they will need to attend to before that). In the
meantime, no noticeable harm arises.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5