On 10Oct24, Al Iverson via mailop apparently wrote:
> > > If you've got any evidence of x= in the wild that you care to share,
> > > thank you kindly in advance!

I'd be curious as to evidence of systems which actually re-categorise email 
based on
x=. And how often such recategorisations are real replays rather than poor 
sender
implementations or long queue delays. IOWs what's the false positive rate over 
the success
rate? Clearly setting x= is trivial but one presumes the OP really cares about 
who is
acting on it.

Given that you have to allow for a queue time of multiple days, x= seems of 
marginal value
- leastwise as an anti-replay mechanism. After all, since a replay by 
definition is an
identical email with an identical hash, any sort of content duplication system 
is going to
trigger much sooner than any x= test. In replays are a real issue, I'd be 
looking at a
hash duplication system long before I worried about x= as surely competent 
miscreants have
read up on that tag.

The other matter of course is that a system can make it's own decision on the 
age of an
inbound email without relying on the sender to dictate a limit. I'm inclined to 
drop old
email regardless of the presence x=.


> https://xnnd.com/dqio

This really shouldn't be saying that unrecognised tags imply an error given the 
RFC says
"Unrecognized tags MUST be ignored". The whole point of the tagging system is to
seamlessly introduce new features without breaking existing implementations.


Mark.
_______________________________________________
mailop mailing list
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to