Adrian Sutton

>Is there another need for the signature to be passed? If you have is_aggregator in the request, then the request body would have all the details that the middleware might...

For the record Teku will return a 404 in this case. I dislike it being a 404 though as that suggests the user has made a mistake and it becomes...

Worth noting that in Teku we generate our JSON representations based off of the SSZ schema definitions and check they're right by parsing the YAML files in the reference tests...

I quite like this approach. I'll have a go at implementing it with Teku but I'm pretty sure it will actually wind up being simpler both for the beacon node...

This approach generally looks like it will work well. A very rough implementation of this in Teku is at https://github.com/ConsenSys/teku/pull/3210 Some notes from going through that: Is there a reason...

> I assumed that current_target_root would be for proposer duties and previous_target_root for attester duties. Yes, I think that's all that's required, but the current text suggests that both fields...

Oh and there's one other interesting corner case that bit me - when calculating duties at the genesis slot you should return the genesis block root, but its not available...

> I've punted and in the case of genesis I return the root from the finalized checkpoint in the genesis state, which will be all 0s. The logic works but...

ConsenSys/teku#3125 looks like some form of JVM bug (can't be 100% certain and not sure how to reproduce it yet) but it has the distinctive property that it's always crashing...

Full sync shouldn't be kicking in when we're close to the chain head though - it's a quite heavy weight thing to have kick in when the `BlockPropagationManager` should be...