On 2024-05-13 at 20:09:33 UTC-0400 (Tue, 14 May 2024 10:09:33 +1000)
Noel Butler <noel.but...@ausics.net>
is rumored to have said:
This morning one of our ent_domains DMARC weekly report from a third
party was listed as spam by SA which took the wording
Not_percent-twenty_Resolved and passed it off to URI checks adding
dot.com to it when there is no dot com after it, and a raw message
search of that message in less in console confirms it.
Context is important. If SA is mis-parsing a message, we really need to
see the message to understand why. There's nothing obviously magic about
that string.
Problem with the code that scans the content for things like URI's?
Likely.
That code is intentionally loose. It is intended to turn anything that
any MUA might consider a clickable link into the same functional URI
that a MUA would. This creates a fundamental tension between
completeness and correctness. SA leans towards completeness but if it is
doing something harmful we'd like to fix that. It would be particularly
important to fix it if the result was a hit on a substantial rule, but
it is not as important to avoid checking bogus URIs that will never hit
anything anyway.
It shouldn't be assuming there's a TLD after it.
I agree. That's a step too far. The days when appending .com was a
reasonable tactic for qualifying hostnames are long gone.
--
Bill Cole
b...@scconsult.com or billc...@apache.org
(AKA @grumpybozo@toad.social and many *@billmail.scconsult.com
addresses)
Not Currently Available For Hire
