On Fri, 22 Jan 2010, Matus UHLAR - fantomas wrote:
On Thu, 21 Jan 2010, LuKreme wrote:
: You shouldn't be sending ANY mailinglists through SpamAssassin.
On 22.01.10 09:49, Charles Gregory wrote:
Say what? And how exactly do you propose to do that for hundreds of users,
any of whom could be subscribed to many different lists? Far too much for
any manual system (presuming you could even get the cooperation of users
who are mostly oblivious to our filtering efforts).
If you are suggesting something automated, how would you suggest that I
identify legitimate mailing lists separate from spammer-faked 'mailing
lists'? X-List headers cannot be relied upon....
I'd simply say: try to avoid feeding to SA mail from known/trusted
lists/IPs... I think that's reasonable.
I didn't say it was a bad idea, I asked HOW it could be done for hundreds
of users, each with their own collection of 'trusted' sources, of which I
know nothing (and probably shouldn't due to privacy laws).
The only thing I can picture, technically, is a kind of auto-whitelist, or
a user-specified whitelist.
The former is subject to various pitfalls, as I've already experienced on
my system, and the latter, even though available to the user as a post-SA
filtering option, tends to get ignored by 95% of users. So the extra
effort to inject a per-user whitelist mechanism into the MTA would seem to
be a waste of time.
I'm all for 'reasonable' suggestions, but they really do need to be
'reasonable' in the context of the user base. :)
- Charles
