Callum Waters

@chillyvee let me know if you need any help. I think if you probably need to start by looking here https://github.com/tendermint/tendermint/blob/af2981a2f75500bc02412bad590fd26228cf5bc3/consensus/replay.go#L372-L374 This would usually return an error in `doHandshake` within...

> Is this something we could dump to disk in a delayed fashion (not that I'm proposing to do that in this PR). It's possible but it sounds messy and...

A lot to unpack here. First of all, in 0.35 we moved the handshake out of the constructor and into `OnStart` - any of the functions that do preform any...

v0.35 isn't safe to use but there are still valuable work in v0.35 that can be brought into main (such as shifting when we do replay with the application). For...

> But handshake is in OnStart() - So the order of operations is out of sequence. > > Not sure if that needs to get moved or not. Just a...

> panic: state (518060) < store (0) height mismatch Why do you a non empty state store? Perhaps delete the state store where you are testing and try again

As additional context. I would want the correctness tests run separately from the performance/experimental tests. These should be split into two packages like `go run ./test/e2e` // for our correctness...

I get: ```{ "code": 3, "message": "unknown value \"2\" for enum cosmos.tx.v1beta1.BroadcastMode", "details": [] } ``` when trying this request. I've never seen: "must declare at least one event to...

It's not part of state. It's persisted to disk and memory

It doesn't need to be committed to state until we have fraud proofs. You can think of it as a block validity rule. The fraud proof might even be able...