Test Case: highly similar objects

[benjamine]

a very common case I image is very large graphs without diff (and one or two changes somewhere).
big objects without any diff at all might be a good case too, a lib that scans first to find if any diff exists might do that pretty well.

[justsml]

My recent update included (some) very similar data in the 'Variable String Chunk Diffs' test.

Kinda like #2 - I'd like to add a little more more structure as well, and see how hierarchy depth vs. byte length (at varying levels of depth) changes performance.