Re: cmsg

From: Frank Ellermann (nobody@xyzzy.claranet.de)
Date: Mon Jul 19 2004 - 18:11:25 CDT


Charles Lindsey wrote:

> it isn't just you and your server. This misfeature is
> performed by an injecting agent, and causes it to issue a
> _genuine_ control message which is then seen by the rest of
> Usenet, who have no way of knowing that it originated as
> "just a 'cmsg' hack".

That's exactly what I want if I send an article with e.g.
Subject: cmsg cancel <40FBED16.3436@xyzzy.claranet.de>

It would be bad to forward this article without adding the
missing Control: cancel <40FBED16.3436@xyzzy.claranet.de>

Either the injecting agent adds the missing Control: header,
or it rejects all Subject: cmsg without Control: header, or
it encodes Subject: cmsg on the fly if there is no Control:
header.

Fowarding it as is won't work, because then some old servers
will follow RfC 1036, new servers will treat it as ordinary
article, and maybe a 3rd group refuses to forward it at all.

It's good enough if we say that injecting agents must solve
this problem in a consistent way. But it's not necessary to
force one of the three possible solutions (i.e. add Control,
reject, or encode cmsg).

Of course all injecting agents _must_ support either reject
or encode for Subject: cmsg foobar, but they could also add
the missing Control for Subject: cmsg cancel <x>.

                        Bye, Frank




This archive was generated by hypermail 2.1.7.