Path: utzoo!attcan!uunet!dino!ux1.cso.uiuc.edu!mp.cs.niu.edu!rickert
From: rickert@mp.cs.niu.edu (Neil Rickert)
Newsgroups: comp.mail.sendmail
Subject: Re: Local Configuration Error
Message-ID: <1990Sep25.180831.26713@mp.cs.niu.edu>
Date: 25 Sep 90 18:08:31 GMT
References: <2229@charon.cwi.nl> <1990Sep25.134012.20789@mp.cs.niu.edu> <2238@charon.cwi.nl>
Organization: Northern Illinois University
Lines: 45

In article <2238@charon.cwi.nl> piet@cwi.nl (Piet Beertema) writes:
>
>	There is a basic misunderstanding here. Redefining $w (with a Dw
>	line) has absolutely no effect of $=w, which must be handled with
>	a Cw (or Fw) line.
>Sorry, I wasn't clear. The point is that $=w doesn't work
>if $w is a multi-token item like a fully qualified name.

 Whether $=w works or not has nothing to do with whether $w is a
multi-token item.  It has nothing whatever to do with the value of
$w.  The 'Dx' and 'Cx' lines are completely separate.

>In that case you can't use the construction Cw$w either.

 Unless I am misreading the code, the construction Cw$w NEVER does what
you think, regardless of the value of $w.  The fact is, $macro expansion
DOES NOT OCCUR during the processing of 'Cx' lines.  When you use 'Cw$w',
you finish up with $=w matching the string '\001w' or something like that.

>
>	My personal preference, is to do the MX comparison with $j instead
>	of $w.
>That makes sense. And the manual is clear both about the
>meaning of $j and about the fact that it MUST be defined.
>
 At last, something we can agree on.

 Not only is $j unambiguously defined in the manual, but it is used in
the ultimate test.  Namely, when an SMTP connection is made, there is
a comparison of the values of $j (as exchanged in the SMTP greeting) to
see if the host is talking to itself.  From this reason alone you could
argue that $j is the correct comparison.

>	This is a trivial patch in deliver.c and is already in the IDA versions
>	of sendmail-5.64+IDA available at uxc.cso.uiuc.edu.
>True. But to have it take permanent effect it Berkeley
>should make it part of the official distribution.

 I did send email to Berkeley suggesting this change.  No reply of course.
I guess we will find out what they think in the next release.
-- 
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
  Neil W. Rickert, Computer Science               <rickert@cs.niu.edu>
  Northern Illinois Univ.
  DeKalb, IL 60115.                                  +1-815-753-6940