|
|
In order to connect to the repository, you have to tell CVS where to find it. Many times after that first time you will never again need to remember where it is as the information gets stored.
The two methods used to tell cvs which repository you want to connect with are using the command line (cvs -d /path/to/repository) or setting the environment variable $CVSROOT.
The form of the repository connect string is:
[:method:][user@machine:]/path/to/repository
For a local connection (the repository is on the same machine as the client) both the method and user@machine components can be ignored. This means that your repository access string is just /path/to/repository.
For a client using rsh to connect to the server, there are two :methods: that can be used: :server: and :ext:. :server: is for rsh, and :ext: allows an external shell program to be used.
The method used here is :pserver:. In order to use this, a cvs server needs to be set up on the remote machine. This is an insecure method as all passwords are sent across the network in plain text.
If you build cvs with encryption, you can use the GSSAPI (Generic Security Services API) with Kerberos to supply both authentication and encryption for cvs commands.