Optimize edge forking by calculating greatest common ancestor. We can use algebird to implement as hyperloglog or another bloom filter like data structure. There is (was) an example of GCA in the rchain repo I remember from a couple years ago.
implement algorithm
add graph of hash refs to snap proposal
pre-filter before redownloading by taking diff from proposal
Optimize edge forking by calculating greatest common ancestor. We can use algebird to implement as hyperloglog or another bloom filter like data structure. There is (was) an example of GCA in the rchain repo I remember from a couple years ago.