Execution Layer Meeting 161

[timbeiko]

Meeting Info

• May 11, 2023, 14:00-15:30 UTC
• Youtube Stream: https://youtube.com/live/s6q5z53SICE
• Zoom: shared on #allcoredevs Discord channel shortly before the call
• Ethereum Protocol Calls Calendar subscription

Agenda

• Cancun scope planning
• Implementation & testing progress
  • EIP-4844
    • 4844 precompile input/output endianness mismatch
    • RLPing 4844: EL, CL vs. SSZ
    • EIP-4844: Explicit Clarification on Block Validity with Blob Transactions using an Invalid Versioned Hash EIPs#7009
    • KZG libraries
• Other EIPs:
  • Contract Secured Revenue on an EVM based L2
    • L2 EIPs/standards on ACD?
  • EIP-5656

[yperbasis]

I'd like to discuss briefly an inconsistency in the EIP-4844 precompile. Its inputs (at least z & y) are little-endian, while the outputs are big-endian.

[zscole]

I would like to discuss EIP-6969.

[timbeiko]

Added @yperbasis @zscole 👍

[shemnon]

The EIP clearly states in motivation that

This EIP does not advocate for any changes to the existing Ethereum L1.

I think there is a critical a priori issues for the CSR topic: does ACDE have "scope" or "jurisdiction" over strictly L2 matters? Should ACDE signal support for various designs, or not? What happens if ACDE proposes support for a for a particular policy and a L2 ignores it or mis-implements it?

One example to compare to is the "Training Wheels" schema for L2s. It was widely discussed on twitter and ethereum magicians, but I do not recall a formal topic on the agenda for this.

I would recommend that the advocates for CSR follow a similar pattern and that it not be discussed on ACD, especially since we often fill the time and very relevant subjects are pushed off to the forum. I don't feel we will be served well filling call time with non-L1 matters when subjects such as Testnet lifecycles got pushed out of consideration.

[shemnon]

To be clean, I'm not bagging on CSR, but this is an issue that I think we will see more of, and CSR is possibly the first issue to discuss fully non-mainnet issues on the agenda, so it gets to be the test case.

[timbeiko]

Agreed with that risk, @shemnon. It's all still very early, so hard to draw super strong boundaries when this is, I think?, the first "L2 EIP". That said, we've definitely pushed L1-adjacent items out of ACD (e.g. mev-boost), although it does bubble back up from time to time.

I think one aspect where ACD can be helpful to a proposal like this is to discuss the technical soundness/potential issues (although it doesn't necessarily have to be on the call). That said, whether this is a good idea is a can of worms we probably should avoid on the call, especially given strong prior opposition to similar schemes for L1.

I'm inclined to see how it goes, assuming we have time for it. Assuming we have even more time, I think we should have this meta-conversation on the call for a bit too 😄

As a final note, I think we definitely should make Testnet Lifecycles part of the considerations! It might get pushed out from time to time due to more urgent topics, but it's super important, and clearly affects core development.

[marioevz]

Would like to discuss this detail in the 4844 spec that has lead to multiple interpretations among implementations: ethereum/EIPs#7009

[timbeiko]

Closed in favor of #781