GNU Arch

Alan Young alansyoungiii at
Sat Mar 5 10:12:16 MST 2005

On Fri, 04 Mar 2005 15:37:44 -0700, Eric Jensen <eric at> wrote:
> environment.  For example, you can't just check out the code into your
> work folder and hack away since apache needs to feed it out if you want

There are ways to set a post commit script.

This is what I do:

When I commit the file(s) I run a script in the devel directory which
does a 'cvs update' from the devel directory.  This auto updates the
beta site whenever I do a commit.

If I include, as the first line, "It's life, but not as we know it."
in the commit message then another script is run which runs a script
in the live site which does a 'cvs update'.

If you have separate servers you can use lynx or wget to force an
update.  Kludgy, but it works.

You can do these in the CVSROOT/loginfo and CVSROOT/verifymsg files. 
If you like I can give you some examples.

You might also consider using mod_cvs with apache.

> servers.  Also changed up the permissions so all coders can work on
> these files.  Which takes away the user tracking functionality of CVS.

Sounds like you might be using pserver.  How come not ssh?  Or is this
on a hosted server with only one user allowed?

If using pserver then use the passwd file.  Since the passwords are
sent in clear text don't use the /etc/passwd file and make sure they
don't use the same password as their login.  It's in the CVSROOT
directory and the format is


If 'user' and 'systemuser' are the same you don't need 'systemuser',
otherwise 'user' will be used for tracking and the files will be
accessed as 'systemuser'.

More information about the PLUG mailing list