Subject: Re: CVS should be easier to use...
From: Mike Meyer (mwm@mired.org)
Date: Thu Jun 08 2000 - 11:19:34 CDT
Shaw Terwilliger writes:
> Mike Meyer wrote:
> > Since the link to CVS info at cyclic software is broken, my memory
> > isn't good enough to remember the update instructions, and I'm lazy
> > enough to not want to look them up each time - I stole a target from
> > the FreeBSD Makefiles. This patch adds "udpate" as a target, which
> > does a cvs update. I'm not familiar enough with cvs to be sure that
> > I've set things up properly, and it probably needs tweaking for people
> > with commit privileges, but the following has worked for me so far.
> I'll try to get the Cyclic links sorted out... until then, CVS makes things
> much easier than your Makefile target seems to suggest. :)
Well, what you've listed isn't any harder than what I thought it would
be - but it isn't any easier, either.
> Making CVS "convenient" can be done by setting a CVSROOT environment variable
> (to something like ":pserver:sterwill@cvs.abisource.com:/cvsroot"). This
> is easiest in a shell init script. Then you only have to do a "cvs login"
> once, and give it your password.
Um - why should AbiWord get preference? I use cvs to update other open
source projects. FreeBSD is the one I use most often, and would be the
logical thing to set CVSROOT to, except they provide make targets that
do the right thing.
> >From then on, you can execute the short CVS commands. To update your
> tree, just cd there and type "cvs update".
As I discovered just before creating the update target, you need to do
"update -d". Running a command with the proper arguments is a standard
reason for putting it in a Makefile - at least for me.
> Adding some defaults to your
> ~/.cvsrc can help you forget all the CVS options. My ~/.cvsrc looks like:
> cvs -z3
> checkout -A
> diff -u
> log -N
> update -d
The problem is - those aren't necessarily correct for all the cvs
servers I use. And CVSROOT clearly must be set on a per-project
basis. If there were some way to have a .cvsrc for each project
(perforce allows this), that would certainly solve the
problem. Failing that, putting those things in the Makefile seems like
a good option. If there's a better option for dealing with multiple
projects on different cvs servers, I'm more than willing to give it a
try.
Thanx,
<mike
This archive was generated by hypermail 2b25 : Thu Jun 08 2000 - 11:26:20 CDT