maintaining UNIX system source code

[Griff Smith]

In article <874 at hsi.UUCP>, stevens at hsi.UUCP (Richard Stevens) writes:
> How do you maintain the source code files for your UNIX ??
> 
> We're running 4.3 BSD on a VAX and use RCS.  What we've been doing
> for the last 4 years is to leave the BSD directories (/usr/src/...)
> exactly as-is from the release tape.  When we modify something, say
> tar, we've been copying /usr/src/bin/tar.c into /usr/local/src/tar.c,
> RCS'ing it and modifying it.
> ...
> Any ideas or comments as to how you're doing this ??
> 
> 	Richard Stevens
> 	Health Systems International, New Haven, CT
> 	   { uunet | ihnp4 } ! hsi ! stevens

We read the source distribution tape into a separate filesystem, which
we call /source; it contains src and sys.  After cleaning up any
objectionable ownerships and permissions, we mount the filesystem
read-only.  We then create directory trees called /usr/src and /usr/sys
where all type 'f' files are replaced by symbolic links to the real
files in /source.  The first time we want to modify a file, we remove
the link, copy the file into the working source tree, then edit the
copy.  This strategy satisfies a vocal "let freedom ring" contingent
who would otherwise hack at the original source without leaving traces,
and it simplifies audits for the rest of us.  Whenever I want a summary
of all local edits, I just use "diff -r /source/src/... /usr/src/...".
-- 
Griff Smith	AT&T (Bell Laboratories), Murray Hill
Phone:		1-201-582-7736
UUCP:		{allegra|ihnp4}!ulysses!ggs
Internet:	ggs at ulysses.att.com