From: Brad Templeton (brad@templetons.com)
Date: Fri Jul 09 1999 - 03:57:13 CDT
On Fri, Jul 09, 1999 at 09:41:53AM +0100, Clive D.W. Feather wrote:
> I don't see the benefit of doing it either. If the injector recognises the
> % and the reinjection case, why does it need to change the first % before
> adding another ?
That's because I think many suspect, due to evidence in the past, that
most double-injections will be errors, not deliberate efforts to do
double injection.
In the cases where they are deliberate, the second injector should perhaps
be specially coded to handle the situation, including detecting and
rewriting the % in the header.
Otherwise, the default for software that is not expecting its users
to insert already injected articles should be to consider double
injection as a sign that there is something wrong going on -- perhaps
a bad gateway into news, or a strange feed or whatever.
I am tending now to favour the rule that "There shall be one injector.
If, however, a site wishes to allow pre-injected articles from certain
parties, and take over as the responsible party, it should rewrite
the articles to appear to the outside world as though they were
injected by it."
This can include trusting audit trail headers on the inserted articles
if it wants to. It definitely includes being sure about whether the
earlier article ewas moderator-forwarded or not.