initial commit of a filesystem deduplicator
second evening of effort, and should handle very large files without running out of memory.
also should be able to identify 2 files where the second file contains the entire first file
(such as an aborted download)

no gui yet
generates a shell script to do the actual cleanup