getting the mavenized website in place

[Leo Simons <le...@apache.org>]

Hi gang,

I figured I'd have a go at publishing some stuff on 
http://incubator.apache.org/geronimo/. I found that typing 'maven site' 
consumes over 400mb of memory, then throws an OutOfMemoryException. I 
guess there's a 'few' leaks in maven :D

Anyway, to be able to build the site, I just redirected some 
functionality to an old-fashioned shell script. Works handy-dandy; it 
just highly inefficiently calls maven for each subproject. I also did 
some other tweaks. Patches attached for your consideration.

In short, here's how to get the mavenized website in place (under unix 
or cygwin; I'm sure its possible under windoze as well):

  # get geronimo and incubator-site cvs and apply patches
  cd ~/cvs
  cvs -z3 -d cvs.apache.org:/home/cvs co -P \
    incubator-geronimo incubator-site
  cd incubator-geronimo
  cvs -z3 up -Pd
  cp /tmp/site.sh .
  patch -p0 < /tmp/patch.txt
  chmod +x site.sh

  # build site
  ./site.sh
  maven site:tocvs
  cd ../incubator-site/build/projects/geronimo

  # add and commit
  cvsaddall # this is something like

# ssh -f -N cvs.apache.org
# for i in `find . -type d ! -name CVS`; do
#   cvs -q add $i
# done
# cvs -q add -R *
# kill %1
#
# I think cvsutils contains a script for handling this a little
# more gracefully

  cvs commit -m 'Publish website'

  # update the live site manually (optional)
  ssh cvs.apache.org
  umask 002 # don't forget this!
  cd /www/incubator.apache.org/projects/geronimo
  cvs -z3 up -Pd
  exit

I probably have the karma to do this, but it's probably best if some 
active geronimo committers do some QA and then get to grips on this. 
Once that happens, it should also be possible to automate and run a 
script like this once every night.

comments, thoughts?

cheers,

- LSD