Ok, here´s my first idea. What about a multiple text-mixer, not only looking for differenz existances of single lines - it can also take care of existing blocks with extension by some rules.
Example: You have a file that contains all your registrered keys - after some time you started a new file or made a copy of the old one with changes. Later on you have 10 files containing all your keys spreaded at different places. Now you want to merge them without losing a key. All doubled blocks should be eliminated and repeating blocks with differences shall be composed to a block containing a mixture of both.
It´s not as easy as it sound - especially implementing some kind of "rules" that are signing block as lines with "-------------" or doubled newlines but I think this could be rather useful for different things.
If you have any suggestions what other useful functions could be implemented, tell me. I´m starting it now.