Re: #1416 When to add Injection-Date

["Charles Lindsey" <chl@xxxxxxxxxxxxxxxx>]

In <0OLeFhTlw8pIFAk9@xxxxxxxxxxxxxx> Richard Clayton <richard@xxxxxxxxxxxxxx> writes:

>In message <48A0464C.9050108@xxxxxxxxxxxxx>, Harald Alvestrand
><harald@xxxxxxxxxxxxx> writes

>>ALTERNATIVE C:
>>
>>-            <t>The injecting agent MUST then add an Injection-Date header
>>-            field containing the current date and time.</t>

>My view (albeit weakly held) is that header fields are either added by
>servers (vaguely trusted) or by clients (not trusted in the slightest)

>This header field is clearly meant to be added by a server, so it should
>be added by a server and not by a client under any circumstances ... so
>I prefer C  (then A then B)

I think the problem with C is that it permits to override a previously
present Injection-Date (which can arise due to all sorts of strange
scenarios including loops involving exotic gatewaying, and possibly
leading to the continuation of those loops).

So if some guy has already put an Injection-Date header there (for
whatever reason, good or bad) then he is saying, in effect, "the clock for
deciding when to expire this article has already started ticking". In
which case he has only himself to blame if his action causes some
premature expiry because he did it wrong.

-- 
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: chl@xxxxxxxxxxxxxxxx      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