Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!bpa!cbmvax!vu-vlsi!devon!ctdi!brian
From: brian@ctdi.UUCP (Brian R. Eckert)
Newsgroups: comp.mail.misc
Subject: Sendmail/Smail -- Missing Received: by
Message-ID: <436@ctdi.UUCP>
Date: 19 Apr 89 18:52:25 GMT
Reply-To: brian@ctdi.UUCP (Brian R. Eckert)
Distribution: usa
Organization: Communications Test Design, Inc.
Lines: 47

I suspect that this has been discussed before, but I have only
recently gotten connected to the net again.  Please forgive me if
this has already been beaten to death:

I have Sendmail (5.61 if it matters) running on a 3B15 with Smail 2.5.
Mail that passes through this system (i.e. not of local origin nor
destination) is not getting a 'Received: by' stuffed into the headers.
I am aware that in this particular situation, sendmail is not invoked
and smail takes responsibility for this... so where is it?

After some head scratching and discussion with a UUCP neighbor,  I punched
up the *smail* source and started poking around.  While my understanding
of smail may be incomplete,  my understanding of C is not;  what I (think I)
see in the source is:

	1) in main.c, if we are running as 'rmail' (determined by the
	   basename of argv[0] begining with an 'r') handle is set to ALL.
	   Fine. But...

	2) in deliver.c (for a sendmail system), at the point where smail
	   (presumably) adds it's 'Received: by' is an if something like:

		if (command == rcommand && handle != ALL) {
			do_the_received_by_stuff
		}

	   indicating to me that for remote mail, smail won't add the received
           by stuff (nor will sendmail since it doesn't get called for this
	   situation).

Someone please enlighten me!  I could hack deliver,  but I must just be
missing something about smail.  My initial reaction (i.e. I haven't given
it any real thought yet) is that the if should be:

		if (command == rcommand && handle == ALL) {

The UUCP neighbor I discussed this with says his smail/sendmail behave as
documented (almost :-).

Any help is greatly appreciated.

-brian
-- 
Brian R. Eckert			|| uucp: brian@ctdi.UUCP
				||	 ...!uunet!cbmvax!ctdi1!ctdi!brian
Communications Test Design, Inc.||
West Chester, Pa.		|| Only entropy comes easy.