Changed the -r option to removing of all duplicates, wherever they

are. Needed a hash-table to do it kind of properly. Seems to work so
far. Wonder if the hash_table_size being ten times the maximum number
of lines makes sense.