Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!rpi!julius.cs.uiuc.edu!psuvax1!psuvm!cunyvm!ndsuvm1!nu013809
From: NU013809@NDSUVM1.BITNET (Greg Wettstein)
Newsgroups: comp.unix.xenix.sco
Subject: Re: (no subject given)
Message-ID: <91030.082945NU013809@NDSUVM1.BITNET>
Date: 30 Jan 91 14:29:45 GMT
References: <4399@ubbs-nh.MV.COM> <91028.081802NU013809@NDSUVM1.BITNET>
 <1991Jan29.022334.10302@midway.uchicago.edu>
Organization: North Dakota Higher Education Computer Network
Lines: 49

I just knew when I posted the query about RCS 5.5 that there had to be
something that I was overlooking.  I found it at about 0257 in the morning
just as I was going upstairs to load the stove with coal so that it would
be warm when we got up in the morning.

For those of you, like myself, that are hesitant to install anything on
a production machine before it is fully tested this may prove useful or
at least save you some time and chagrin...

The long and short of it is that RCS 5.5 compiled and run just fine after
appropriate hacking is done to overcome problems with the automatic
configuration utility.  The problem comes when an attempt is made to use
the rcstest program before the make install step is carried out.

The problem that bit me is that the path to co is hardwired in the rcsdiff
program.  What ended up happening is that all the test cases up to the
rcsdiff program used the programs found in the path (.).  When rcsdiff
was run it used the hard-wired path to execute co prior to diff'ing the
two versions.

The long and short of the situation is that the RCS files being tested
were produced with V5.5 sources and checked out prior to diff'ing with
the V4.7 sources which were in /usr/local/bin (the hard-wired path).  The
change in the structure of the RCS files evidently created a problem
which was flagged as an error.

I am sorry to have troubled the net with such a problem.  I knew from
the minute that I posted that the problem was going to be something that
I was overlooking.  Hopefully this experience will save someone else
a similar loss of time.  Incidentally I overcame the problem by compiling
with the RCSDIR path set to ./, ran the test program and then recompiled
the sources wit the RCSDIR path set to /usr/local/bin.  A bit troublesome
but I really hate to install something this potentially critical before
I test it, especially when RCS is set to run setgid on the group that
controls all of our production software.

Thanks to everyone who offered suggestions.

                            As always,
                            Dr. G.W. Wettstein
                            Oncology Research Division Computing Facility
                            Fargo Clinic / MeritCare

                            UUCP: uunet!plains!wind!greg
                            INTERNET: greg%wind.uucp@plains.nodak.edu
                            Phone: 7001-234-2833

`The truest mark of a man's wisdom is his ability to listen to other
 men expound their wisdom.'