Re: [Gimp-print-devel] WWW in CVS

[<sh...@al...>]

> I couldn't find anywhere that suggested this could be automated (can
> we run cron jobs on sourceforge ?), so this is my current solution ;
> 
> The WWW site is in cvs (dir www). To then copy the current CVS version
> into the "live" site, on shell1 (via ssh)
> /home/groups/gimp-print/updateWebFromCVS.sh is run. This should update
> the site into our www directory, and then fix permissions so that
> someone else could do it next time. It is retrieved as anonymous, so
> you shouldn't need any auth.

You can trigger this to be done automatically by placing a command in the
loginfo file in the CVSROOT directory.  The CVS documentation explains how to
do this.

On one of the other sites I maintain we have this in loginfo:

ALL $CVSROOT/CVSROOT/cvs-log %s $CVSROOT/CVSROOT/www.list $CVSROOT/CVSROOT/commitlog
^www  /bin/sh $CVSROOT/CVSROOT/updateweb.sh %s >> $CVSROOT/CVSROOT/updatelog 2>&1 &

And updateweb.sh:

date
cat
sleep 2
cd /u/ale/superk
echo cvs -q update -d `echo $1 | awk '{split($0,files);for(i=2;i<=NF;i++){print $1 "/" files[i]}}'`
cvs -q update -d `echo $1 | awk '{split($0,files);for(i=2;i<=NF;i++){print $1 "/" files[i]}}'`
echo Entering directory /u/ale/superk/`echo $1 | awk '{print $1}'`
cd /u/ale/superk/`echo $1 | awk '{print $1}'`
make

That may be a little fancier than we need.

Eric