`eth_newPayloadV4`: Why require full requests instead of requestsHash?

[ravirahulhh]

Currently, the eth_newPayloadV4 interface requires the CL to pass the full requests to the EL. But it seems like passing just the requestsHash should be sufficient.

EL can locally process the block, derive the valid requests, compute the requestsHash, and compare it with the one provided by the CL. Format validation and ordering checks for requests can also be enforced via the requestsHash comparison.

So I’m wondering if there’s any reason to require the full requests instead of just the hash. Maybe there's something I’m overlooking, otherwise, this seems like unnecessary duplication.

[mkalinin]

CL doesn’t have to implement execution requests commitment schema that is used by EL and vice versa, passing requestsHash instead of full requests would break this separation of concerns