Retroactively cross-posting articles
Lars Magne Ingebrigtsen
larsi at gnus.org
Sun Apr 7 09:24:34 UTC 2002
Kenichi Okada <okada at opaopa.org> writes:
> I think the "blocking" does not happen.
> The mailpost process finishes always for -c seconds.
> If there is the second mail cross-posted, the first mail is
> not posted and is transfered to the mailpost of the second mail.
> The max delay time is (-c seconds) x (cross-posted number).
Yeah, but if Gmane receives 1000 messages over a 30-second period,
then the machine will have run out of processes, and messages will
start to bounce. Unless there's something in exim to control how
many local delivery agents are running at the same time.
> Another way is to use cistron-m2n at
> ftp://ftp.cistron.nl/pub/people/miquels/software/ .
> (I have never used it :)
I've just had a quick look over the package, and it seems to use a
quite different methodology. If I read it correctly, it goes through
all the To and Cc addresses, and compares that to the list of
subscribed mailing lists. And then it uses that to find out which
groups to (cross-)post the messages to. Which might work quite
well -- at least for cross-post handling.
> It is difficult to be delivered immediately.
> If you know addresses of all mailing list, you can parse
> fields 'To' and 'Cc' of the first mail and create merged Newsgroups.
> That kind of API is not provided. We cannot modify a message
> posted already. The APi is not useful on the NNTP,
> because the message may be fed to other servers which
> do not receive any articles having the same Message-ID.
Well, I could delay feeding other servers messages until, say, 10
minutes have passed, and then things should be stabilized on
(domestic pets only, the antidote for overdose, milk.)
larsi at gnus.org * Lars Magne Ingebrigtsen
More information about the inn-workers