Sendmail problems on RS/6000 (+ othe UUCP stuff)

[Peter Jeffe 512.823.4091]

In article <9008170856.AA12741 at mindcrf.mindcraft.com> karish at mindcrf.UUCP (Chuck Karish) writes:
>>	I to saw the same problem when I first installed the OS in our
>>6000. The message is a no @ in the SMTP address error. Seeing as the path
>>given to mail was machine!user I didn't expect to see an @. I played around
>>with the cf file with no luck so I took the easy way out and put smail 3.1
>>on the box.
>
>    I saw this message when I used bang paths on a machine that had
>    no actual UUCP connections, but was supposed to forward mail
>    to a `smart host' that has UUCP links.  It complained about
>    outgoing mail; I never saw the problem on incoming, but I never
>    tried to provoke it.  Using 'user at host.uucp' did the trick;
>    a simple rewriting rule in sendmail.cf could have done it for me.

Sorry for not posting a work-around when I first described the problem.
Since sendmail is looking for an '@' in any recipients that use the tcp
mailer (this is the bug), you can fool it by adding a source route in
ruleset 0 where the uucp relay is handled.  This is the diff to sendmail.cf:

171c171
< #DUHostWithModems
---
> DUlopnor
1024c1024
< R<$+>!$+                $?U$#tcp$@$U$:<$1>!$2$|$#uucp$@$1$:<$2>$.
---
> R<$+>!$+                $?U$#tcp$@$U$:<@$U:><$1>!$2$|$#uucp$@$1$:<$2>$.
                                        ^^^^^^

Where the "<@$U:>" source route satisfies the check for the '@', while the
uucp relay host ($U) should recognize it as itself and just strip it.

Please let me know if there are other problems that this doesn't fix.
As I said before, the bug will be fixed in the September update.
-------------------------------------------------------------------------------
Peter Jeffe   ...uunet!cs.utexas.edu!ibmaus!auschs!sandino.austin.ibm.com!jeffe
        first they want a disclaimer, then they make you pee in a jar,
                   then they come for you in the night