The Absolute Minimum on Installing and Using CVS

These are the simple steps that I used to get myself up and running with CVS. They are the very minimum steps, with little in the way of explanation. Please refer to the many good resources on CVS for more detailed and comprehensive information:

1. The Cederqvist Manual
2. The Redbook Manual
3. WinCVS Daily Use Guide: WinCVS is a GUI CVS client for Windows
4. CVSGui
5. The CVSNT Wiki: CVSNT is a CVS server for Windows NT
6. DevGuy's CVS Information Helpful guide to CVS, including CVSNT installation
7. SourceForge's Basic Introduction to CVS
8. SourceForge's Introduction to CVS for Developers

Bear in mind that the CVS program consists of two main components: the server daemon/service, and the client. CVS can be used on a standalone machine such that the repository and the working-copy are stored separately on the same computer. In addition, it can be used over a network such that the daemon/server and the repository are on the remote machine and the client and working-copy (or sandbox) are on the local workstation. Network operations involve user names and passwords, so a variety of protocols, of varying levels of security are available.

Also see my notes on WinCVS.

CVSROOT

CVSROOT is a core aspect of CVS, referring, in general to the location of the repository, and possibly the protocol used to connect to it. Be careful when using the word CVSROOT; it can mean the following different things:

• The directory that contains the repository.
• The variable that indicates which repository CVS commands operate on.
• The name of the administrative directory within the repository.

The -d cvsroot option of the cvs command is used to indicate which repository to use. However, to avoid having to enter this with every CVS command, you can use an environment variable named CVSROOT (e.g.: export CVSROOT=/usr/local/cvsroot/ for Unix). If you have set such an environment variable, you can override its value temporarily (presumably to use a different repository) using the -d option.

For remote repositories, it is necessary to specify both the machine that holds the repository, and which protocol to use, in addition to specifying the repository directory. On the local computer, the CVSROOT variable contains at least the path to the repository, and may contain the following:

• connection method, surrounded on each side by a colon (e.g. CVS's built-in method pserver, ssh for security, or sspi for Windows)
• user name for logging in to the remote machine, suffixed by the ‘@’ symbol
• server name, optionally suffixed by a colon

On Windows machines, the sspi protocol is convenient and secure. A typical CVSROOT in this case could be:

:sspi:psy.swan.ac.uk/cvsrepo

If you are not using a built-in protocol, you may need to specify the program CVS should use to connect to the server. This can be done by setting an environment variable named ‘CVS_RSH’ with a value corresponding to the name of the client program that supports the relevant protocol. On Windows clients, Plink.exe, which comes with PuTTY can be used for this purpose. Note that the directory path should contain forward slashes, even if both the client and server are Windows machines. A fuller example of a CVSROOT variable follows:

:pserver:cartern@psy.swan.ac.uk:/usr/local/cvsroot

Installing CVS

The operations described next are performed on the machine that contains the repository. In the case of a networked system, this would be the server machine; the programmers work on their own workstations, using the CVS client to communicate with the repository.

1. Create a directory to hold the CVS repository. This directory is called “cvsroot”, and is typically located in /usr/local/cvsroot/ or c:\cvs\cvsrepo\ on Windows.
2. Make a group named cvsusers (edit /etc/group or use User Manager on Windows)
3. Add the appropriate users to the cvsusers group (edit /etc/passwd)
4. Change the owning group of the cvsroot directory to the cvsusers group (e.g. chgrp <cvsroot> cvsusers, not necessary on Windows)
5. Change the permissions of the cvsroot directory so that the cvsusers group can write to it (e.g. chmod <cvsroot> g+w or right-click on the directory in Windows, and use the security tab to give Modify permissions to the cvsusers group.)
6. Initialise the cvsroot directory with the command cvs init -d <cvsroot>.

Using CVS

• If you are using CVS with a password protected username, you need to use the command cvs login <username> and then issue the password. Use cvs logout when you have finished using the repository.
• To add a new project to the CVS repository:
  1. Move into the directory containing the files you wish to place into CVS
  2. Use the command cvs import -m "<log message>" <module name> <author> <release>. Note that CVS will add all files in the current directory and all subdirectories, so you may want to delete or move any files that need not be permanently stored with CVS. Note also that the module name will be used as the directory name for the module when it is stored within the cvsroot directory.
  3. Now delete or rename the directory that you just added to CVS. This is because this copy of the files is not actually under CVS control.
  4. Obtain a CVS–managed copy (known as a “sandbox”) of the files for the relevant project from CVS with the command: cvs checkout <module name>. Note that CVS will generate the sandbox as a subdirectory of the current directory.
• If you want to discard the changes you've made a to a file (and revert to the previous version stored within CVS), delete, move, or rename the file, and use the command cvs update in the relevant directory.
• To store the changes you've made to a file since the last checkout, use the command cvs commit <filename>. If you don't enter a filename with the commit command, cvs will commit all changed files.

Content last updated: 2007-01-05