From: Charles Lindsey (chl@clerew.man.ac.uk)
Date: Wed Jul 07 2004 - 12:20:04 CDT
In <87zn6cwodf.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>Charles Lindsey <chl@clerew.man.ac.uk> writes:
>> I think all it actually needs is a warning somewhere (possibly in USEPRO
>> alongside the place where it says servers MUST NOT any longer implement
>> the old behaviour, or possibly in USEAGE) to the effect that articles
>> with Subjects beginning with "cmsg" may sometimes fail to propagate to
>> some not-yet-compliant servers.
>That doesn't seem unreasonable to me, but that isn't the same as saying
>that I agree with it. I don't have a particular solution that I favor;
>I'd rather leave it to other people to make specific proposals on the best
>way of handling this.
Right. Let us discuss the options.
1. I note that son-of-1036 did include
"For historical reasons, the subject MUST not begin with "cmsg " (note
that this sequence ends with a blank)."
followed by a NOTE explaining why and declaring the cmsg-hack as
"obsolete". That is probably the reason why various agents currently
reject such Subjects, as noted by Frank. It would be useful to know how
many current agents actually do that. Does INN, for example?
2. Our draft contained a similar sentence at one stage, but it got moved
to USEAGE (and demoted to SHOULD) as part of the great split. Nobody
commented on that at the time. Replacing that text in USEFOR is clearly
one of the options open to us.
3. Alternatively, we insert a warning as suggested above. Here is a
possible text, to replace the present NOTE in 6.13
The presence of a Subject-content starting with the string "cmsg " and
followed by a control-message was construed under [RFC 1036] as a
request to perform that control action (though no implementation
actually did so if a genuine Control-header was present, and few
current implementations still observe the practice).
Likewise, the presence of "all.all.ctl" in the Newsgroups-content
caused (in the absence of any genuine Control-header) the
Subject-content (not starting with "cmsg" in this case) to be
interpreted as a control-message.
Both these practices are now declared to be Obsolete, and
Subject-contents MUST NOT now be interpreted as control-messages under
any circumstances.
NOTE: These practices have long been disregarded by most serving
agents but, in order to ensure that unexpected control actions do not
occur, some injecting agents currently reject any article with a
Subject-content starting with "cmsg ". Posters should not be
surprised at this, neither should they in future expect
Subject-contents starting with "cmsg " to be acted upon as
control-messages (e.g. as cancel messages) except at a few serving
agents not yet compliant with this standard.
4. Upon closer examination of Cnews I found, contrary to what I expected
and to what Henry reported, that it does implement both the cmsg-hack and
the all.all.ctl-hack (worse, it recognizes any newsgroup-name ending in
".ctl"). What actually happens is that its built-in injecting agent
actually adds an appropriate Control-header to the article and sets it
loose on the network, so that it will now be effective everywhere. I don't
think it actually acts on "cmsg " in the Subject on incoming articles.
Aaaarrrrrrrrrrggggggggghhhhhhhhh!
>It would be useful to tell gateways somewhere about the trick of encoding
>cmsg, but that might belong in the gatewaying RFC that Lars and I were
>talking about rather than in any of our current documents.
I think we had better agree what we are going to do within Netnews before
worrying about gateways :-( .
-- 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