How to Get Started

The main contact will be sancus, but do say hello to dria first.

Local MDC setup

Get linux2 (hera) account - talk to sancus (or bhearsum). You should not use your own server because you need to run and test the wiki on the same environment as MDC.
Checkout MDC's mediawiki source code from CVS. MDC has step-by-step documentation for this process.
Once that's done, copy the LocalSettings.php and GlobalSettings.php from either /var/www or from /home/melz/public_html/.
If you are having hostname issues (look at your wiki's url - it should be pointing to your url), talk to sancus to have it fixed.
Once the wiki is verified to be working, you need to use your own database now.
Ask someone (I need to stop repeating the names, look above) for the mysql root password to create your own mysql login and database. Dump the wiki database contents from either sancus or mine (melz's) installation. You can find the necessary authentication information in the GlobalSettings.php file for either installations.
Populate your database with the dump and change your database information in GlobalSettings.php
Your own MDC wiki should be up and running now.

Nutch is wrtten in Java. If you want to mess with Nutch, you better be willing to read lots and lots of Java code.
The Nutch helper class, etc is located in /extensions.
The Nutch search engine communicates (XML) with http://nutch.developer.mozilla.org/ which is where all MDC documents are indexed at.

If you plan to work on this bug, here are some information to help you get started:

Specific namespace searching, remove Sandbox from general search, allow searching of Talk/Sandbox pages exclusively - requires tweaking with Nutch settings on how it indexes data. Talk to sancus about this.
Follow the user-set preference for number of results per page - look at includes/User.php, in particular, the getOption() function.

All ports on linux2 starts with 20XX when accessed from an external network. So for the webserver, the port is not 80 but 2080.