Split data into blocks

[efiop]

As @shcheklein suggested, we should consider splitting data into small blocks to track data changes more efficiently. Example: giant file that has one line appended to it.

[sotte]

For a private data sharing/deduplication tool I've used a buzhash/rolling hash for chunking big files into small chunks pretty successfully. The borg backup tool has a nice implementation of a buzhash chunker. Maybe it's worth checking out.

[efiop]

@Witiko noted that there is ficlonerange that we could use to create file by reflinking smaller files into it. #2073 (comment)

[salotz]

This is a pretty good solution to the chunking deduplication problem. I've used it with success for backing things up: https://duplicacy.com/.
If anything you could use the algorithm.