Matt Kettler wrote:
Jo Rhett wrote:
I'd love to, but the SA project didn't write the milter you're using,
and the problems you're having can't be "fixed" by having SpamAssassin
"detect" the problem without doing something even dumber to someone
else.
Sure it can! It's dead simple to determine that the last Received
header isn't yourself and can't possibly be related to you.
No, you can't.. Again, separation of spamd is COMMON.
I don't see how separation of spamd changes the equation. If there's
something special about your spamd separation, then please document that
configuration. Using XNS protocol, or Appletalk instead of IP?
And again, doesn't that fall under "custom configuration" ?
I'll bet you a very nice dinner somewhere that there are 10, if not 100,
users with SA run via procmail or milter etc for every site which has
distributed mail servers. Maybe a 1000:1 even.
And again, in looking at the code I can't imagine why detecting the
distributed configuration would be difficult. Yeah, the current code
seems to do the whole network detection bit from end->beginning, but
it's certainly not that difficult to add these checks.
Look, forget this conversation. When I have time from some other
projects to deal with this, I'll prepare a long list of cases which will
fail under the current code, and a patch to fix it.
--
Jo Rhett
Network/Software Engineer
Net Consonance