How to use set up and use CVS, developer tips and tricks

Developer notes I made about CVS as version control system

Note: Subversion is a good alternative to CVS, so if you currently are not using CVS you might want to look into using Subversion instead,
or if you do not need something to be exactly like CVS, you should look into Git.

How to use CVS

1 MAKE SURE THE USER IS ALLOWED TO USE CVS

 The CVS repository reside on the server myserver in the directory /cvs
 To  be able to access that directory and its files you need to be user someuser
 
 In the file /cvs/data/CVSROOT/writers you need to add the userids for those who should be able to write to CVS.
 In the file /cvs/data/CVSROOT/passwd you need to add the password for the new user. To generate the password you can use this perl script:
 
 #!/usr/bin/perl
 $passwd = crypt("$USER","whatever");
 print "$passwd
";
 
 here the password will be the same as the $USER.
 
 Once you have added the user to writers, and added the password to passwd, you are done here.
 

2 SETTING UP THE ENVIRONMENT

 The variable $CVSROOT has to be set. Add the following line to ~/.profile
 export CVSROOT=":pserver:${USER}@myserver:/cvs/data"
 
 Before you check out the files for the first time; make sure that you do NOT have a directory ~/theFiles/
 If you do, please rename that directory.
 
 To check out the source code for all programs that are used in the
 release process, do the following:
 
 check that the variable $CVSROOT is set, then execute the following commands
 % cd
 % cvs login
 % cvs co theFiles
 
 You shall now see information which files that are checked out.
 All the code appears in ~/theFiles/
 From now on, only work on the files in the directory ~/theFiles/
 

3 WORKING ON THE FILES

 IMPORTANT: Always make sure that you are working on the latest versions of the
 files! E.g. do a "cvs up" before you start modfiying any files.
 And always make sure that someone else has not made any changes to the file
 you modified, before you check in the files. E.g. do a "cvs st" before you
 commit the file, to make sure that there arent any new changes to the files
 
 When you have changed files, test them and check them back into the repository.
 After you have checked in a file, you should try to release the file as soon as possible.
 
 Note: Binaries will NOT be checked in into cvs.
 That means that the binaries manually have to be copied to the correct directory on the machine.
 See the document install.html for information on how to handle binaries.
 

4 GENERAL INFO

 The full documentation for CVS you will find at, it is recommended that you browse through it to get a grasp how cvs works:
 http://www.cvshome.org/docs/manual/
 
 The most common tasks:
 cvs up filename	check out the latest version of a file
 cvs co dir  		check out a whole directory
 cvs st filename 	show the status of the file (which version/if it is modified)
 cvs log filename 	show a log of all changes made to the file
 
 cvs add filename
 will add a NEW file to the repository
 cvs commit -m "your comment what you have done" filename
 will check in the file in the repository again, please do a cvs st filename first to make sure that you have the latest copy of the file
 
 A VERY useful flag is -n which can be used with most (all?) commands,
 It will not execute the command, it will only show what will happen if you do execute the command.
 e.g.
 cvs -n up filename
 will not update filename, it will just display a message what would happen if you update "filename".
 
 If you have files that shall be included in the repository, you can add those filenames to the file ".cvsignore",
 then cvs will not complain about them all the time.
 

5 Backups

More programming related pages