Includes an engine for crawling a subversion repository and full-text indexing the content with Lucene, along with a sample web application for searching the index.

Could be a great deal better. I have spent most of my energy on the crawling/indexing pipeline as my target repository is very very large.

Possibilities for improvement:
Support for different source-code repositories.
Syntax-highlighting display of source.
Pluggable parsers to index source as an AST.
etc.