[ietf-nntp] OVER message-id

[Andrew - Supernews]

>>>>> "Russ" == Russ Allbery <rra at stanford.edu> writes:

 > Charles Lindsey <chl at clerew.man.ac.uk> writes:
 >> But this I do not understand. If Andrew is prepared to serve up
 >> the complete article (or even just its HEADers) given the
 >> message-id, then he is surely obliged to serve up an Xref header
 >> with it (for how else does the client know how to mark all the
 >> crossposted articles as read when he has read the first one)?

 Russ> I don't see anything in the protocol that requires this.

Typical large Diablo configurations do not return an Xref in articles
fetched by message-id. (It's not usual in such cases to number the
articles prior to storage, since that requires making the numbering
box into a big central bottleneck for the whole system - better to
number only the headers.)

If there's been any big outcry from client authors about the missing
Xrefs, then I missed it.

That said, I don't personally approve of this behaviour and I don't
intend to follow this approach myself if I can avoid it (it violates
the POLA if nothing else).

-- 
Andrew, Supernews
http://www.supernews.com