hexdump usage takes a while to parse
We run hexdump on the same byte segments in a few different variations to detect the compression.
I wonder if this could be more clever? We could read out the largest boundary once and then maybe substring the matches?