Something which may be interesting for other projects: For FranceTV, we have implemented Lucene indexing of content published to the filesystem. This is done by maintaining a Lucene index which is separate from the search engine we have on the editorial server. The changelist is 32335, and basically involves two new methods in PublishToFile which can be overriden in derived classes (indexSearchContent and unindexSearchContent). For computing the indexed content, we use the Lucene Adapter class that's part of the existing search engine. The location of the PFS Lucene index is controlled with this section in enterprise.init: init com.arsdigita.cms.publishToFile.LuceneInitializer { // Settings for indexing PFS content with Lucene. luceneIndexLocation = "/tmp/lucene-index.dat"; } And publishListener has to be changed to: publishListener = "com.arsdigita.cms.publishToFile.LucenePublishToFile"; Documentation can be found in the JavaDoc of com.arsdigita.cms.publishToFile and LucenePublishToFile, and in //ps/proj/francetv/dev/ftvi/doc/lucene-pfs-indexing.html. Expected results: Additional info:
I have changed the implementation in the meantime. See changelist 33226.
Closing old tickets.