[geeks] revision control

[Gregory Leblanc]

On Tue, 2002-03-12 at 21:03, James Sharp wrote:
> > > So, I'm thinking that a revision control system would be a good idea.
> > > Something that I can check every increment that compiles back into, or check
> > > in nightly, or something.  So, what do people suggest?  A full blown CVS
> > > server for little ol' me?  Something else like RCS or SCCS?  I don't really
> > > know anything about anything outside of using CVS as a client (and knowing
> > > that CVS checkins only take a chord in emacs).
> >
> > CVS is trivial to set up for personal use.  It sucks, but it's better
> > than any available alternative, for any platform.  I think there's a
> > HOWTO, and there's certainly a semi-comprehensive info book.
> > 	Greg
> 
> What do you find wrong with CVS?  I use it here at home and at work
> extensively and I've found nothing majorly bogus with it.

Heh, ok...  It's totally inconvenient to set it up securely (I know how
to do ssh tunneling, it's a pain).  There's no way of locking things so
that you can have multiple distributed servers, not even if some of them
are in "read-only" mode.  The triggers for allowing you to check code at
commit time are pretty crude, and a real pain to get working.  Automated
changelogs are nice, but there's not an easy way (any way?) to get them
written out to disk at checkout time.  I'd love to blame the problems
with making a mailing list that gets all commits to CVS mailed to it on
CVS, but I think they're probably caused by the crappy perl that we
stole from OpenBSD.  I'm sure there are some others, but those are the
big nagging ones.  
	Greg

-- 
Portland, Oregon, USA.