From: Frank Ellermann (nobody@xyzzy.claranet.de)
Date: Sun Jul 25 2004 - 21:26:54 CDT
Charles Lindsey wrote:
>> It does when I use its "cancel" function, but I didn't know
>> this when I tested the Subject: cmsg cancel feature (2002).
> And was "cmsg cancel" a documented feature of youe user agent
> at that time?
At that time I still used From: frank.ellermann@t-online.de
This resulted in invalid Message-Ids <3xxxx.xxx@t-online.de>.
I didn't like these Message-Ids, and patched them manually
(= option "send later" and edit file outbound)
As a side-effect I couldn't cancel these patched articles,
therefore I tested Subject: cmsg cancel <M-Id>, and that
worked as expected.
Later I found a better solution with the catch-all vanity
host xyzzy.claranet.de (=> Message-Id okay, cancel works).
[<I17o8q.JrH@clerew.man.ac.uk>]
>> That's exactly what I want if I send an article with e.g.
>> Subject: cmsg cancel <40FBED16.3436@xyzzy.claranet.de>
> Now that is absolutely HORRIBLE!
> The one thing that injecting agents absolutely MUST NOT do
> is to alter articles submitted by the user in any substantive
> manner, and especially in a manner that will have side
> effects all over Usenet.
It didn't alter my article, it only added the obviously missing
Control: cancel <40FBED16.3436@xyzzy.claranet.de>, found that
this made sense, and propagated the article as intended by me.
It would be much more "horrible" to accept cmsg cancel as is,
with side effects only in _parts_ of Usenet (= some servers
treat it as cancel, other servers treat it as ordinary article,
and maybe a 3rd group of servers silently drops this article).
> If a user tries to use some feature that has been declared
> obsolete,
That's not exactly the case in RfC 1036:
| for upward compatibility, if the first 4 characters of the
| "Subject:" line are "cmsg", the rest of the "Subject:" line
| should be interpreted as a control message.
> he has only himself to blame if it is only effective at a
> few sites that still adhere to the old ways.
It's not the problem of the _users_ if nobody was able to
update RfC 1036 in (soon) 17 years. And servers doing the
right thing, i.e. fix this where possible, aren't horrible
but smart.
> Particularly as that is already the current situation
> even if our draft says nothing.
So far the proposed solutions for the cmsg-feature are not
better than what I've found as existing fix on one server.
If s-o-1036 is not published as RfC before the grandson,
then the exact wording of s-o-1036 is the only solution:
| For historical reasons, the subject MUST not begin with
| "cmsg " (note that this sequence ends with a blank).
Bye, Frank