I've moved out of implementation mode (unable to concentrate enough), and into high-level design mode.
Syncing efficiency has been an open TODO for years, to find a way to avoid flood filling the network, and find more efficient ways to ensure data only gets to the nodes that want it. Relatedly, Android devices often need a way to mark individual files they want to have. Had a very productive discussion with Vince and Fernao and I think we're heading toward a design that will address both these needs, as well as some more Brazil-specific use cases, about which more later.
Today's work was sponsored by Casa do Boneco.