Set the CVSROOT environment variable to point to where you want the repository to live:

% export CVSROOT=~/ts10-cvs

For tcsh, the equivalent would be:

% setenv CVSROOT ~/ts10-cvs

(Note the space instead of the equals sign.)

Create the repository directory:

% mkdir $CVSROOT

Tell CVS to initialize the repository:

% cvs init

Create the subdirectory in the repository for your source files:

% mkdir $CVSROOT/ts10
% ls $CVSROOT
CVSROOT  ts10
%

If your working source tree has subdirectories, create a similar
subdirectory structure here.  For instance, if you have
ts10/foo, ts10/bar, and ts10/bar/baz:

% mkdir $CVSROOT/ts10/foo $CVSROOT/ts10/bar $CVSROOT/ts10/bar/baz

cd to the parent of the source directory:

% cd ~/work

(Note that your ts10 source code is in the ts10 directory *below* this
one.  If you do an "ls" at this point, it should just show the ts10
subdirectory.)

Do a cvs checkout. This sets up the CVS subdirectories in the source tree:

% cvs checkout ts10

This will produce output with one line for each file in ts10, with a
question mark at the left.  This indicates that cvs has found those
files but they are not yet part of the repository.

cd into the source directory:

% cd ts10

do a cvs add of all files you want to go into the repository, this tells CVS that you are going to check them in:

% cvs add README Makefile *.c *.h

(Include any other files you want in source control.  If you want
to leave some source or header files out that aren't actually
part of your release, you can list the source files explicitly
rather than using a wildcard.)

commit the files, this actually puts them into the repository:

% cvs commit -m "initial version"

The -m and string give the initial log message, use anything you
feel appopriate.  cvs commit works recursively, and will commit
all the files you've added, even in subdirectories.

If you ever need to commit files individually, you can give the
files on the command line.  For instance, if you wanted foo.c and
bar.c to have a different initial log message than the rest, then
before doing a commit of everything (as above), you can do this:

% cvs commit -m "a different log message" foo.c bar.c

verify that everything looks good:

% cvs status | more

This will give details about every file in the repository.  Each file
should have a working revision and a repository revision of 1.1,
and Status of Up-to-date.  Sometimes I like a more concise summary
with just the Status field:

% cvs status | grep Status

(note the upper case S on the Status in the grep)

Make sure CVSROOT is set as in step 1 of the setup above:

% export CVSROOT=~/ts10-cvs

Make a new directory to check out into:

% mkdir ~/work2

cd into that directory

% cd ~/work2

check out the code:

% cvs checkout ts10

cd into ts10:

% cd ts10

build the code as usual:

% make

cd into the repository top-level directory:

% cd $CVSROOT

tar it up:

% tar -cvzf ~/ts10-cvs.tar.gz ./ts10

cd out:

% cd ~