Draft specification for future X-Trace header

[Marco d'Itri]

On Jul 05, Olaf Titz <olaf at bigred.inka.de> wrote:

 >>  >However, this SHOULD NOT be done in the normal forwarding of articles,
 >>  >but only at injection points and gateways (of any kind).
 >> I think this should really, really be a MUST NOT.
 >OK. But "normal forwarding" is not exactly defined.
I propose:

The header MUST NOT be added by agents other than injectors.
(Or something which actually means something in English.)

 >> I don't think X-Trace should be extended to do that, it looks very
 >> unclean.
 >> If Message-ID munging is a problem, the Message-IDs should be munged
 >> the same way by all gateway. Look at News::Gateway about this topic.
 >The problem is precisely that you can't count on all gateways to
 >behave well.
And why would you count on all gateways supporting X-Trace?

 >(Some incoming messages even don't have a Message-ID at all.)
Just hash the body in a standard way.

 >> How could people find the right header then?
 >By the /systemid/ item. If someone finds a problem and can't see the
 >precise source, he should contact the administrators at all the
 >systems listed in X-Trace headers.
Cool. So spammer would just add a bunch of X-Trace headers and let
people pick the right one...

 >> POST newsfeeds are evil.
 >But in the real world you can't avoid them. There is no other way for
This is not true. Nobody will accept your argument for allowing them.

-- 
ciao,
Marco