per at hedeland.org
Fri Mar 18 19:57:24 UTC 2005
In article <mailman.24.1111108290.588.questions at lists.ntp.isc.org> Brad
Knowles <brad at stop.mail-abuse.org> writes:
> I have some influence with Tokio and Barry, and if I ask them to
>take a look at the code in a given patch, I think that we can at
>least get the process started a lot faster than might otherwise
> However, I can't write that code myself, and I think we're
>unlikely to get them to solve this issue any time soon if someone
>doesn't at least supply a patch as a suggested starting point. Given
>everything you've said so far, you seem to be the most likely
>candidate to provide such a patch.
> But we need that starting point, if we're going to get anywhere.
The starting point *I* need is some minimal likelihood that a patch will
not be rejected for what it *does* rather than what it is. As I tried to
explain, given the statements made by the developers, I'm 99.9% sure
that they would reject the patch I have in mind no matter how perfectly
done it was - and I'm not going to spend any time writing code with such
So, if you could ask them whether they would be at all interested in a
patch that implemented a configuration option to simply turn off the
Message-ID rewriting, it would really help. Of course the option should
default to having the rewriting on, and the documentation should
clearly state the implications of changing the default (I can supply
such text too if wanted).
Turning off the Message-ID rewriting basically solves the threading
problem (especially in combination with stripping the prefix, which was
already on the roadmap) - or at least leaves it at the "normal" level of
problems with mail->news gatewaying. Using In-Reply-To to create/improve
References is an extra bonus, but basically orthogonal to turning off
the Message-ID rewriting - except that it is no use at all if Message-ID
rewriting is on.
The downside is that the "don't send messages from the mailing list back
to the mailing list" problem must be solved in another way. Handling it
on the news server side, e.g. by using the Path header as I described
earlier, is the best/simplest and most reliable/efficient way to do
this. This does leave it outside of mailman's control of course, which
is why the documentation must explain it.
per at hedeland.org
More information about the questions