Hi, I think Jackrabbit should be as small as possible and only include 'real' dependencies. In my view, text extraction is an optional dependency (meaning, only a dependency if you actually need text extraction). Specially, if some file types need dependencies that are bigger than Jackrabbit itself! So each file type should have one or multiple optional dependencies. I don't know the 'correct' way to describe this in the pom file.
For JCR-2864 (Use out-of-process text extraction) I guess we anyway should list the jar files needed for text extraction of which file type. Regards, Thomas
