@Gmaclennan Have you thought ahout ditching hypercore oplogs for replicating at the sqlite storage layer? Merkalize the b tree and have sql as your primiyive and sync the source trees from other writers (still runs into paralell writers issue hut maybe merged can happen more efficiently that way too?

@mauve not sure how to do that?! I think we really need to build upon an existing low-level primitive rather than build our own - open to options other than hypercore though.

Follow

@Gmaclennan Yeah that's reasonable. This idea would require building upon some more c/c++ friendly p2p protocols and getting into the internals of how sqlite loads and saves data 😅

Sign in to participate in the conversation
Mauvestodon

Escape ship from centralized social media run by Mauve.