Usenet 1994 January/usenetsourcesnewsgroupsinfomagicjanuary1994.iso/sources/std_unix/volume.19/text0108.txt

Internet Message Format | 1990-05-17 | 1.8 KB

From: [e-mail] (Matthew Atterbury)

In article <[e-mail]> you write:
> ...  In my spare time, I'm preparing several improvements to RCS that I'll
>eventually submit to Purdue, and if there are easy ways to improve RCS I'd
>like to hear about them.

Our company has two sites working on the same code. We create patch
files to distribute our local changes to the remote site. I have had
to hack up a shell script for rcsdiff (doing it all by hand) s.t.
$keyword differences are ignored (since our $Source$, $Header$, and
$Log$ expansions are usually slightly different). A switch to rcsdiff
which somehow ignored such differences would be nice. A "good" way
would be for the $Log$ entries never to be expanded in the ,v file, but
built up everytime you co'd the file. A switch to co would *not*
expand keywords - thus, rcsdiff could co the two versions to be diff'd
with this switch set and do a diff on them. Clearly such diff's should
be applied to a similarly co'd file - perhaps a command/method to
convert expanded $keyword's to their unexpanded versions would handle
this and the rcsdiff "problem". Obviously $Log$ is the tricky one
since it is multi-line and not clearly delineated - I have handled it
by assuming that ALL lines following $Log which start with the $Log
prefix + a space (eg. " * " or "# ") are $Log lines - works OK for us.
Apart from this, I too prefer RCS to SCCS (not that SCCS would
necessarily be any better than RCS in this regard!). cheers ...
--
-------------------------------------------------------------------------------
Matt Atterbury [[e-mail]]   Expert Solutions Australia, Melbourne
UUCP: ...!uunet!munnari![e-mail]            "klaatu barada nikto"
ARPA: [e-mail]  "life? don't talk to me about life!"

Volume-Number: Volume 19, Number 112