Currently, one has to modify the src *.xml files, generate the output *.html files by running 'forrest' and achieve BUILD SUCCESSFUL, then move or copy them to a separate CVS, then COMMIT via CVS the updated *.html files.

As the current Apache FOP Web Maestro ;-) I concur that the current system is not the best system. A change would be nice. The solution Jeremias outlines would be acceptable to me.

Another idea would be to COMMIT the 'forrest' output directory build/site/*.* and have the automated process grab the files from there (instead of grabbing them from the separate CVS repository (which is actually /www/

Web Maestro Clay

On Jun 27, 2005, at 11:36 PM, Jeremias Maerki wrote:

I'm not happy with the way our website is managed:
- Updating the website is a big job each time.
- Due to knowledge gaps some of the files in /www/ are
not the most current ones from the CVS module.
- Uploading compiled changes into a version control system (though
propagated by some here at the ASF) seems not very elegant. Yes, it has
it's advantages in case of a mistake when you can simply go back to a
previous state, but you'll get my meaning.

For Barcode4J I have a very simple system which allows me to build and
deploy the whole site from the build system. The compiled website is
compressed into a tar.gz and uploaded using SCP and extracted at the
target using Ant's sshexec task. Once it is properly set up, this is
painless, provided everyone uses the same Forrest version.

If noone objects, I can implement this in the next 2 or 3 weeks and test
it by deploying the FOP website to When it
works we can then also remove the FOP website from the XML site and
instead do redirections there.

Jeremias Maerki


Web Maestro Clay
My religion is simple. My religion is kindness.
- HH The 14th Dalai Lama of Tibet

Reply via email to