On Thu, Mar 31, 2016, at 00:50, Mark Sapiro wrote:
> What Mailman does do as noted by Random832 is replace the Message-ID:
> header value in posts gated to Usenet with a list specific, Mailman
> generated unique value. There is a reason for this, and that reason is
> if a message is cross-posted to two lists which both gateway to Usenet,
> and Mailman didn't make the Message-IDs unique, the news server would
> discard one of the two posts as a duplicate and the post would be
> missing from one of the recipient Usenet groups.
>
> Granted that this is bad and breaks threading, but avoiding message loss
> is a more important goal.
>
> I understand I'm not providing any solutions here, but perhaps a more
> complete understanding of what the issues are will ease the pain.
Any chance that it could fix reference headers to match?
Actually, merely prepending the original Message-ID itself to the
references header might be enough to change the reply's situation from
"nephew" ("reply to [missing] sibling") to "grandchild" ("reply to
[missing] reply"), which might be good enough to make threading work
right on most clients, and would be *easy* (whereas maintaining an
ongoing reversible mapping may not be).
And if it's not too much additional work, maybe throw in an
X-Mailman-Original-Message-ID (and -References if anything is done with
that) field, so that the original state can be recovered.
Rather than exclusively rewriting for usenet, maybe the rewritten
headers could also be included in outgoing emails and the archive?
Putting it in outgoing emails would solve the problem entirely, though
it would mean people get duplicates if they're subscribed to multiple
lists to which something is posted or get CC'd. The archive wouldn't
have this issue.
--
https://mail.python.org/mailman/listinfo/python-list
