Access the read-only CVS source-code repository

Usage

To check-out a copy of the main 'dev' (development) trunk, type this:

cvs -d :pserver:anonymous@fscvs.nmr.mgh.harvard.edu:/usr/fscvsroot login

press Enter when prompted for the password (there isn't one). Then type:

cvs -d :pserver:anonymous@fscvs.nmr.mgh.harvard.edu:/usr/fscvsroot checkout dev

This will begin the download of the source-code trunk, creating a directory named 'dev'.

To get a copy of the current 'stable' branch (the code used to build the public release), insert '-r stable4' before 'dev' in the second command, like this:

cvs -d :pserver:anonymous@fscvs.nmr.mgh.harvard.edu:/usr/fscvsroot checkout -r stable4 dev

Note that this will also create a directory named 'dev', so you should probably create a directory named 'stable4', cd to that, and then do the checkout.

Build instructions are found in the DevelopersGuide, but in any case, you will first need to type this:

cd dev
./setup_configure

which will setup the autoconf and automake files for your platform. You need only do this once (or whenever you change configure.in).

The DevelopersGuide will describe how to need to configure, make, and make install.

Remember that cvs will cache the location of CVSROOT (the 'pserver' string from above), so to perform an update, you need only type:

cvs update -d

Maintenance

This section documents how the read-only CVS server was setup. It is intended for Martinos Center developers.

A VM named fscvs was setup on surfer, which implements an instance of Linux Centos to appear on the web as fscvs.nmr.mgh.harvard.edu. On this virtual machine, in the directory /usr/fscvsroot, there is a copy of the real (read/write) CVSROOT (which is found on the NMR NFS network). The cvs pserver functionality is used to serve this repository copy. That was setup by creating a file called cvspserver in the directory /etc/xinetd.d, where the file contains:

service cvspserver
{
        port = 2401
        socket_type = stream
        protocol = tcp
        wait = no
        user = root
        passenv = PATH
        server = /usr/bin/cvs
        server_args = -f --allow-root=/usr/fscvsroot pserver
}

If that is ever changed, you will need to type service xinetd reload to restart that service.

Now, how does the copy of the real CVS get from its location on the NMR network to fscvs? That is performed via two cronjobs. One is run as 'nicks', on the machine 'minerva', which runs the script /space/freesurfer/scripts/cp_fscvs, which tars the real CVSROOT, excluding the files we don't want to make public, and deposits that file in a 'nicks' temp directory. This is run every night at 2:00AM. Then, on fscvs, a 'root' cronjob, run at 2:10AM every night, runs the script /usr/fscvsroot/update_fscvs, which scp's the tarball from nicks temp directory to /usr/fscvsroot, untars it, and performs some permission changing.

On fscvs, the file /usr/fscvsroot/CVSROOT/passwd specifies the valid usernames. The update_fscvs script creates an empty passwd file, except for the username 'anonymous'.