Tendermint has gone through several implementations of the blockchain reactor (v0, v1, v2):
v1 was deprecated as it is not used and was never really finished. v2 looks like a promising re-write at first sight but introduces a lot of unnecessary interfaces and abstractions that do not really help with the readability but more importantly, we do not have proof that it works properly: a few projects use it (we did for instance), and no real-world deployment seem to actually use it. The v0 reactor has some rough edges but at least we know that it works.
We should remove v2 as from this project's perspective it is just a distraction. We can incorporate all changes in a way that can easily be added to the v2 reactor if it is ever stabilizes and we gained further confidence in its correctness. Or we might decide to actually write our own block-sync implementation (see #273 (comment))
Tendermint has gone through several implementations of the blockchain reactor (v0, v1, v2):
v1 was deprecated as it is not used and was never really finished. v2 looks like a promising re-write at first sight but introduces a lot of unnecessary interfaces and abstractions that do not really help with the readability but more importantly, we do not have proof that it works properly: a few projects use it (we did for instance), and no real-world deployment seem to actually use it. The v0 reactor has some rough edges but at least we know that it works.
We should remove v2 as from this project's perspective it is just a distraction. We can incorporate all changes in a way that can easily be added to the v2 reactor if it is ever stabilizes and we gained further confidence in its correctness. Or we might decide to actually write our own block-sync implementation (see #273 (comment))