CVS for Source Code Revision Management

Introduction

http://www.ee.ualberta.ca/~elliott/ee552/studentAppNotes/1998f/cvs/

I recommend reading over Paul Somogyi's appnote first, and then reading my appnote. There are some aspects of CVS that we have not used in our project group, and maybe those aspects that are covered only in Paul's appnote will be quite useful. What I have done is post updated information about the CVS system implemented in the CAD labs that are used in the current EE 552 labs. For example, CVS has been installed and is available on all the machines, and CVS will run without any fiddling around with the path, etc.

Setting up the environment

• Decide which member will host the CVS repository. When this is decided, make sure that each member has read/write/delete (complete access) permissions to this directory. If you are unsure about how to change permissions, there is a good appnote that can be found at:
  http://www.ee.ualberta.ca/~elliott/ee552/studentAppNotes/1998f/unix/shared_afs_directories/
  The CVS repository should be a directory that is freshly created (for example, ours was /ceb531/home/bwli/collaboration
• Each member should edit one of their shell login scripts to automatically set two environment variables, $CVSROOT and $CVSIGNORE
  In our group, we all use the bash shell, so these commands relate to the bash shell. Open up the ~/.bashrc file in a text editor (the file might not exist yet), and put in the two following lines.

  > export CVSROOT=/ceb531/home/bwli/collaboration
  > export CVSIGNORE="*.fsf *.csf *.rpt *.eqn *.pin *.pof *.psf *.sof *.ssf *.quartus *.vwf *.ddb *.cdb *.hdb *.msg *.db_info *.tdb *_info *.ini *.rdb *.BSF *.rdb *.hif *.acf"

  The $CVSROOT environment variable points to the CVS repository, so CVS can read/write as needed. The $CVSIGNORE environment variable specifies which file types that CVS should ignore. The long list of file types shown above are intermediate files types that Quartus II likes to create. It was based on experimentation, and is quite accurate for Quartus. For any other tools, you'll have to see which file types to ignore by trial and error.

• On your next login your environment should be setup. If you are anxious and do not want to logout/login, just execute the .bashrc script at a prompt by typing:
  > . .~/.bashrc

Creating the repository

• The hosting member now navigates into the directory where the project files are located (CVS automatically navigate into subdirectories) and import the files:
  > cvs import -m "Initial import into CVS" repository_name tag1 tag2

  The repository_name is how this particular CVS repository is accessed so choose an easy to type identifier. tag1 and tag2 don't really do too much in a simpler CVS system, so I usually just use the username and project name, just to have something to type in those fields.

Checking out the project

• Group members should navigate to the directory where you want to be your CVS working directory and "checkout" the project:
  > cvs checkout repository_name

  Inside repository_name is all the files that the hosting member imported.

Common operations

• When you start working on the project, perform a cvs checkout in your home directory. This will get you the latest files

  > cvs checkout repository_name

• To add a new file or directory that doesn't exist in the current repository (directory must be added before adding files):

  > cvs add -m "message" file/directory

• To commit your changes to certain files in CVS :

  > cvs commit -m "message" file(s)

• To commit all your changes (not usually a good idea since the message will be pretty general and we can't tell what was changed to each file):

  > cvs commit -m "message"