Re: #1416 Injection-Date: proposed diff

In <[email protected]> Russ Allbery <[email protected]> writes:

>"Charles Lindsey" <[email protected]> writes:

>> So our amateurish guy implements this by configuring that all bona-fide
>> Usenet groups are to be relayed/injected as the case may be, but
>> carefully omits the local.* groups from his list. And for months and
>> months that works fine, so the guy believe he is configured just fine.

>> Then, one day, someone elsewhere on the local network writes an article
>> crossposted to local.misc and comp.misc - well you can see what happens
>> then, and two versions of the article running around Usenet is one
>> possible outcome.

>So he still knows he's reinjecting (or not) and therefore his software
>still handles Injection-Date properly unless his software has been broken
>all along for comp.* as well.

No, he thinks he is just relaying an article to comp.misc.

Yes, if his software knows that a pre-existing Injection-Date is NEVER to
be removed or altered, then no harm will arise. And since that is a matter
for the software implementor rather than a matter locally configurable,
he is unlikely to make that mistake (unless he starts hacking the source
code).

Moreover, no existing implementation is going to remove that
Injection-Date, because no existing implementation is even aware that
header exists. So I think the rules as you have written cover all the
possibilities fine.

I have just checked my own 'sys' file, and indeed the problem cannot arise
with, for example, the man.* groups I mentioned, some of which I keep on
my machine. But I can also see how a 'sys' file could easily be
misconfigured so that it happened.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: [email protected]      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