Re: Suggested update for spamass-milter

[Dan Nelson]

In the last episode (Dec 23), Cassandra Lynette Brockett said:
> I just grabbed the CVS tree of the milter, and compiled the milter
> for my system.
> 
> When I installed it I noticed it was not sending the full username to
> spamc, at least that was how the system was handling things...
> 
> -------- Log Output --------
> Dec 23 13:11:10 hotline spamass-milter[24669]: calling /usr/sbin/sendmail -bv 
> "<address@hidden>" 2>&1
> Dec 23 13:11:10 hotline spamass-milter[24669]: sendmail output: 
> <address@hidden>... deliverable: mailer local, user fakeuser
> Dec 23 13:11:10 hotline spamass-milter[24669]: user: fakeuser
> Dec 23 13:11:10 hotline spamass-milter[24669]: remembering fakeuser for spamc
> Dec 23 13:11:10 hotline spamass-milter[24669]: remembering recipient 
> <address@hidden>

Bah.  Alias/virtusertable expansion stripped the domainname :)  Off the
top of my head, I'm not sure how to handle this.  Maybe if -e is
specified and it ends up delivering to a local user, append ${j}?

> I have checked with a manual test and if I send the correct username
> to spamc (ie. spamc -u address@hidden etc...) then the spamc
> process handles things correctly, so I can only assume that the code
> is not handling the username pass-through correctly.
> 
> Just to check, the code appears to list that a username
> <address@hidden> would be passed through in full, not just
> <address@hidden> being passed through in full, that is correct?  I
> hope so as that will be a necessary thing for us here.

With alias/virtuser expansion enabled (I plan to make this a selectable
option, defaulting to off because of issues with sendmail+smmsp
installs), spamass-milter just takes whatever address sendmail -bv
hands it.  In your example, sendmail returned:

"<address@hidden>... deliverable: mailer local, user fakeuser"

All I can do here is tack the local host's domain ($j) at this point, I
think.

-- 
        Dan Nelson
        address@hidden