Re: #1416

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

>If a server yesterday looks at the Date header and deletes the message 
>(expiration), and today it gets the same message with the same Date header 
>and accepts it, it is broken. It is ridiculous to accept articles that are
>just going to be deleted in the next expiration pass. I don't know if 
>server authors actually considered that when they wrote their code, but it 
>seems like a useful optimization. "Don't accept things older than I'm 
>already expiring..."

Yes, but the whole point of Injection-Date is that NEW servers use it when
deciding whether to consider it stale. That way an article that was
injected 24 hours after it was composed has an extra 24 hours before it
gets considered stale, thus improving its chances of propagation.

The (slight) downside of that is that NEW servers can be hoodwinked in
some highly peculiar scenarios such as the one I showed (but it wouldn't
happen anymore once all the OLD servers had been pensioned off).

-- 
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