From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Wed Jan 12 2000 - 05:54:35 CST
In <20000111164608.D71842@demon.net> "Clive D.W. Feather" <clive@demon.net> writes:
>Charles Lindsey said:
>> Yes, but that was before the days of "relay only" agents that do not even
>> keep an active file. Can someone explain to me more precisely what these
>> agents do?
>Basically they handle the fan-out part of relaying in as efficient a manner
>as possible. A simplistic design might be something like this:
>* Accept a connection from a peer. Accept all the articles that it throws
> at you (and that are not in your history file) up to some limit, say when
> a 100 Mb buffer has filled. Close the connection.
>* For all other peers in turn, open a connection to that peer and offer it
> all the articles in turn. Close the connection and try another peer. When
> all peers have been dealt with, or after a timeout, free the buffer.
It would seem that, for those purposes, you need to look at:
1. The Newsgroups header (perhaps somewhat cursorarily, perhaps looking
just to see what hierarchies are mentioned, because some of your
downstreams may only want certain hierarchies).
2. The Path header (because you are supposed to check its leftmost entry,
and add your own to it).
3. The Message-ID header, because you want to record the article in your
history file.
4. Possibly the Date header, because you want to reject articles that are
stale.
But there is no reason to look at any other headers, much less check them
for minor syntactic infringements (two addresses in a Sender, missing From
header, empty Subject header - that sort of thing). I gather that current
'news routers' do not bother with such trivia (injectors and servers
should, of course).
We seem to be splitting into two camps here, so it needs to be discussed
further. I think the issue is whether we actually want to go so far as to
outlaw what is becoming standard practice. Russ seems to want to go that
way. Bill says not.
I suspect that what we shall finish up with is a SHOULD in place of that
MUST, with perhaps a short list of checks that remain a MUST.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Email: chl@clw.cs.man.ac.uk Web: http://www.cs.man.ac.uk/~chl Voice/Fax: +44 161 437 4506 Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5