The webservices software repository lives on babbler.web.itd.umich.edu in /usr/local/CVSROOT-websvcs. Add this to your .cshrc file:
if ( `hostname` == "babbler.web.itd.umich.edu" ) then # the repository is on the local machine setenv CVSROOT /usr/local/CVSROOT-websvcs else # the repository is available via ssh setenv CVS_RSH ssh setenv CVSROOT :ext:babbler.web.itd.umich.edu:/usr/local/CVSROOT-websvcs endifadding a new module/project to the repository
step 1: cd into the root of the module's working directory.
step 2:cvs import -m "log message" name name start
step 3:cvs diff nameif the diff succeeds then you can cd back a directory level and blow away the working directory. You don't have to, of course, but you can. :-)
checking out the current version of a module to work on
step 1: cd into a build directory or similar working space
step 2:cvs checkout -P namethen make changes as you need to, don't delete directories! (see the man page for cvs remove to find out how to remove files). Directories never get deleted from the repository. The '-P' flag above will prune empty directories from your checked-out source tree.
exporting a project
This is actually how moves to production are going to work. The developer will have to use CVS to maintain her/his project. When it is time to move to production they'll send us a message that includes the following information:
intended URL of project:The author must have set a meaningful tag with the command:
CVS project name: (i.e. showcase, umce, counter, etc...)
version to publish: meaningful tag
anyway... the actual command we'll use for this is:cvs -q tag production-2000_11_11("production-2000_11_11" is just an example good tag.)
cvs export -r meaningful_tag nameThe export is really nice because it doesn't include any empty directories or CVS version information so it should be ready to go live.
reference
man cvsis your friend.
http://www.cvshome.org/ has some good docs.
backups
babbler is backed up (both the development DB and /usr/local/).