Paul Johnson
Feb 20, 2003
PS909 labExample3.txt


cvs is the "Concurrent Version System".

It is installed on lark, and it is installed on most linux systems and
you can even get it for windows.  There are some graphical cvs
"client" programs, but I have found them to be a total pain.  The
command line version is easy to use.

The cvs archive for our class sits in a directory called "cvs909" in
my lark account.  If all goes well, you will have permission to access
that, to take out files and check them back in.

In my archive, so far I have only one program, the Schelling
segregation model, which is in Objective-C. So in my cvs archive, I
plan to have 2 directories,

objc
java

and then the names of the programs are under these things. So
the schelling program is in 

/homem/pauljohn/cvs909/objc/schelling

Suppose you want to get the source code and revise it, and save
it back into the CVS archive.

YOU ARE A CLIENT!  You get stuff, edit, and send it back.

First, make a directory where you want to download the archive.

# mkdir trash
# cd trash


Since our usage is somewhat out of the ordinary, you have
to type a few out of the ordinary commands to get started.

#  export CVS_RSH=ssh
#  cvs -d lark_user_name@lark.cc.ku.edu:/homem/pauljohn/cvs909 co .

WHen you put a period at te end, it means you want to have "everything". 
Once we get this archive up and rolling, you may just want one
part, and in that case you would replace the period by 
a "module name" or a directory name (such as "objc/shelling").


In cvs, the letters "co" are an appreviation of "checkout". I believe
you can use the full word checkout in place of co if you want, but I
use co because that's what the cool guys do.


For comparison, in a fresh directory, try this 


#  cvs -d lark_user_name@lark.cc.ku.edu:/homem/pauljohn/cvs909 co objc/schelling

or 

#  cvs -d lark_user_name@lark.cc.ku.edu:/homem/pauljohn/cvs909 co objc

I also have labeled the schelling subdirectory as a "project module", so
if you just want to work on that, you can do so by asking for it by name. 
In a fresh directory, do:

#  cvs -d lark_user_name@lark.cc.ku.edu:/homem/pauljohn/cvs909 co schelling


AFTER YOU HAVE DOWNLOADED...

Look over the files. Check permissions. You may see the downloaded
files are "read only". You might as well give yourself write
pemission with the old "chmod" command. Remember?


Note there is a README and a ChangeLog.
The README is a message to a visitor, the Changelog is a
list of specific comments about changes that have been made.
Whenever you make any change, you are supposed to make a
ChangeLog entry.  In Emacs, that can be done with

M-x add-change-log-entry

I don't use Emacs's features for dealing with CVS.  They confuse
me too much.

So after I make edits on the files, I save them, make sure they
compile, and then get out of Emacs.

At the command line I check in the changes with the command

# cvs commit

That initiates a "check in procedure" where it wants you to enter
some comments about the changes you made.  The Swarm team members
just paste in the ChangeLog entries they created.  Other CVS users
can review your comments to see what you did.

Update your CVS "sandbox".

Suppose you come back later and you want to find out if
other members of your research team have updated some files.

cd into your working directory and type

cvs update -d

The -d option says that, if somebody created a new directory, you
want to download that, as well as files. If you leave off the -d,
it tells cvs you only want file updates, not new directories.

There is a lot of other fancy crap in cvs, and they have a beautiful
manual that all well educated people should look over. That manual
is called "cvs.ps" and it is saved in the system docs directory
on all well maintained linux workstations.  I leave it as a puzzle for
the reader to find it.  You know the tools already, if you have
reviewed the stuff on http://www.ku.edu/cgiwrap/pauljohn/twiki/view

pj