Execution Layer Meeting 159

[timbeiko]

Meeting Info

• Apr 13, 2023, 14:00-15:30 UTC
• Youtube Stream: https://youtube.com/live/u8Nm8AGyCQM
• Zoom: shared on #allcoredevs Discord channel shortly before the call
• Ethereum Protocol Calls Calendar subscription

Agenda

• Shanghai Recap
• Cancun Updates
  • Update EIP-4788: Move to Draft EIPs#6859
  • Capella Genesis
• State that payloadId should be unique for each PayloadAttributes instance execution-apis#401
• TBA

[ralexstokes]

I've made substantial updates to EIP-4788 I'd like to present. I strongly suggest we CFI EIP-4788 for Cancun.

ethereum/EIPs#6859

I'd also separately like to discuss a block verification RPC method that is used by mev-boost relays to verify incoming builder blocks. Currently, most of that ecosystem uses a fork of geth and I'd like to discuss cross-client implementation of the RPC method to support client diversity at that part of the stack.

[mkalinin]

I'd like to discuss the following fix to the Engine API spec

• State that payloadId should be unique for each PayloadAttributes instance execution-apis#401

[timbeiko]

Added @mkalinin @ralexstokes !

[siladu]

I'd like to discuss the following fix to the Engine API spec

• State that payloadId should be unique for each PayloadAttributes instance execution-apis#401

This besu issue has some details if you want more background context hyperledger/besu#5321

I am pretty sure nimbus-nethermind will have the same issue. I wasn't able to confirm whether erigon had the same fix that geth has.

These calls are too late for me, but I think some others have context now since discussing on #interop

[MarekM25]

I'd like to discuss the following fix to the Engine API spec

• State that payloadId should be unique for each PayloadAttributes instance execution-apis#401

This besu issue has some details if you want more background context hyperledger/besu#5321

I am pretty sure nimbus-nethermind will have the same issue. I wasn't able to confirm whether erigon had the same fix that geth has.

These calls are too late for me, but I think some others have context now since discussing on #interop

That's true. We're generating distinct values with a small exception - withdrawals. I discussed with @g11tech that this situation is not possible to happen. I mean

1. Timestamp = A, PrevRandao = X, SuggestedFee = Y and Withdrawals = B, HeadHash = C
2. Timestamp = A, PrevRandao = X, SuggestedFee = Y and Withdrawals = Z,  HeadHash = C 

However, it sounds like it is possible with Nimbus combos, so we will prepare fix

[parithosh]

I'd like to bring up the topic of capella genesis on the EL&CL front (I'd assume EL has no problem with it). We'd like to modify all the testing infra to start from capella instead of a bellatrix genesis state. We'd additionally need some input on if the various future upgrade branches also support this, i.e, I know verkle needs a rebase before this is possible, but what about EOF or 4844?

[g11tech]

That's true. We're generating distinct values with a small exception - withdrawals. I discussed with @g11tech that this situation is not possible to happen. I mean

1. Timestamp = A, PrevRandao = X, SuggestedFee = Y and Withdrawals = B, HeadHash = C
2. Timestamp = A, PrevRandao = X, SuggestedFee = Y and Withdrawals = Z,  HeadHash = C 

well it should not be possible with any CL since withdrawals is a matter of state unless nimbus is not mapping timestamp with the slot (which for e.g. lodestar does)

I would like to understand when can two fcUs have everything same except the withdrawals.

[timbeiko]

Added, @parithosh

[timbeiko]

Closed in favor of #759