Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!shadooby!ginosko!uunet!zephyr.ens.tek.com!tektronix!nosun!fpssun!celit!billd
From: billd@fps.com (Bill Davidson)
Newsgroups: comp.sources.d
Subject: Re: rcs for SysV
Message-ID: <1367@celit.fps.com>
Date: 8 Oct 89 08:35:33 GMT
References: <78@vidiot.UUCP>
Sender: daemon@fps.com
Reply-To: billd@fps.com (Bill Davidson)
Distribution: na
Organization: FPS Computing Inc., San Diego CA
Lines: 40

In article <78@vidiot.UUCP> brown@vidiot.UUCP (Vidiot) writes:
>I received the answer that I needed in order to get RCS working on this SysV
>PC.  I have also found out that there is a version 4 of RCS that works under
>both BSD and SysV.  Unfortunately, I don't know where to fetch it.

I've seen several requests for this and I've sent mail to people but it seems
to be time for a post.  RCS Version 4.2 is available by anonymous ftp from
arthur.cs.purdue.edu (128.10.2.1).  I believe this is the main distribution
point.  Uunet also keeps a version but the last I checked, I think they were
still at 4.0.  The compressed tar file is 208301 bytes (581632 uncompressed).
They no longer distribute rdiff or rdiff3 so you should keep the old ones if
your system diff and diff3 aren't up to snuff (or get gnu diff).

I brought 4.2 up on Microport System V/AT release 2.2 (talk about a hostile
environment!).  The only problem I had was with a file which wouldn't compile
because Microport's compiler and assembler are broken (Have you ever gotten a
syntax error from your assembler from code generated by your compiler? >^(  The
fix was to take the pointer to a function out of the data declaration and
assign it at runtime.  I can't remember the module but if you do it on a
similarly broken system, you'll know when you see the error :-).

I also made a hack to rcsfnms.c to simulate symbolic links by making
pairfilenames() look for a file called "RCS_link" which, if it existed would
contain the path of the RCS directory.  It seems to work quite well.  I was
thinking of sending diffs to Purdue to see if they wanted to offer it as an
option to poor System V victims.  Would anyone else like to see this?  I'd have
to check with "the powers that be" to make sure I even could give it away but I
think I could.

In the past, I also ported 3.0 to System V and that was living hell.  I had to
fix the previously mentioned time function and a zillion things that expected
pointers and int's to be the same size and that they were all 32 bits etc.  4.2
doesn't seem to have any such problems.  The code is looks much cleaner and
better structured.  A number of things which used to be cheap hacks have been
fixed to be done the right way.  Those purdue people have done a great job
fixing it up.

Does anyone know why Berkeley (or anyone in their right mind) still uses SCCS?

--Bill