Agenda
Consensus-layer Call 140 · Issue #1129 · ethereum/pm · GitHub moderated by @ralexstokes
Summary
Summary by @ralexstokes (from Eth R&D Discord)
- consensus-specs release
v1.5.0-alpha.5is out! - Pectra devnets
- Started with an overview of
devnet-2- A bad block, likely from early EIP-7702 implementation; clients are debugging
- Minor issue with JSON-RPC, see the call for more details
- Touched on an update to EIP-2935; no substantive change, just a clarification around behavior
- And it looks like we are in a good place for
devnet-3— be on the lookout for launch next week!
- Started with an overview of
- Pectra EIP updates
- Update to EIP-7251 to adjust correlated slashing penalty; fixes an overflow at high amounts of stake
- Please take a look! The change is straightforward but touches some very delicate code.
- Next, a proposal to refactor some of the data types on the beacon chain to streamline processing of execution requests added in Electra
- General support for this solution
- Blocked on adding spec tests; I’ll follow up here
- Then another proposal to refactor the engine API to streamline implementation when passing execution requests from EL to CL
- Proposal has some support, although there was a question about the interplay with the possible move to SSZ if we decide to harmonize serialization across the CL and EL
- Going to get more feedback from additional CL clients
- Update to EIP-7251 to adjust correlated slashing penalty; fixes an overflow at high amounts of stake
- PeerDAS
- Started with implementation updates and client status
- Clients have made progress on implementing the
peerdas-devnet-2specs, and devnet launch should be in the next few days!
- Clients have made progress on implementing the
- Next, turned to EIP-7742
- Started with some clarifications, engine API payload attributes needs the maximum blob count as well as the target as currently specified
- Did a temperature check on Pectra inclusion, generally good support
- And then touched on an implementation concern around blob flexibility
- Some clients hard-code the max and target blob parameters, which can make operations of devnets harder
- Clients are aware of the issue and will likely solve this pain point under EIP-7742
- This segued into a discussion around blob limits per transaction in the protocol, and how that interacts with similar concerns in the public mempool
- want to make blob packing easier; if a transaction has the max number of blobs, it crowds out other transactions that could otherwise go into the same block
- strong pushback to not have a protocol-level rule and instead handle in the mempool or networking layer
- some nice back-and-forth on this point, check the call for the details!
- Started with implementation updates and client status
- And to close the call:
- Agreed to label the F-star (next hard fork after Pectra):
Fulu, with portmanteauFusaka - And last but not least, an update from Probelab around some recent analysis of gossipsub and its control messages
- Find slides on IPFS with this CID: QmURBPigXY9LjGuumwRohrXfhhi26RZmWHKQT2RjcUEQAy
- A related analysis on block arrival times here: Week 2024-29 | ProbeLab
- Agreed to label the F-star (next hard fork after Pectra):
Recording
Additional info
Notes: Ethereum All Core Developers Consensus Call #140 Writeup | Galaxy by @Christine_dkim