TechFlow reports, on September 15, Ethereum core developer Tim Beiko summarized the latest Ethereum Core Developer Execution (ACDE) meeting, which covered updates on Devnet, additions to Dencun, and a comprehensive overview of Reth:
1. Devnet-8 status update: The network is finalizing, and many clients have begun pushing new updates to it. Meanwhile, testing of MEV/block-building workflows using the developer tooling system Kurtosis has already started. Nethermind reported that their blob transaction pool is ready; after several days of testing on a single node, they’ve deployed it across all Dencun testnet nodes. Geth’s blob transaction pool is also nearing completion. Besu is conducting broader refactoring of its transaction pool—aimed at limiting the size of both Blob and non-Blob transactions—with release expected in the next version. Erigon is still developing its pool and aims to be ready for devnet-9.
2. The meeting continued last week's ACDC call discussion on whether to add a constant upper limit to the validator activation queue. This proposal has now been formally designated EIP-7514 ("Maximum Epoch Slashing Limit"). In short, under worst-case scenarios, this would slow down the growth rate of staked ETH percentage.
3. Another last-minute proposal was discussed: adding an opcode in the EVM to expose the blob base fee. We already have a similar opcode that exposes the BASEFEE from EIP-1559, introduced at the time of that EIP’s activation. This new opcode would allow L2s to more easily determine appropriate gas pricing for users based on L1 data costs.
4. Updates to EIP-4788 were discussed, which involves storing beacon roots in a contract on the execution layer (EL). Over recent weeks, multiple audits and fuzz tests on the contract led to several subtle changes described in this PR. First, handling of zero timestamps has been clarified to revert (like other invalid timestamps), rather than return zero. Second, buffer size adjustments: assuming slot times might change, the original contract could lead to storage waste due to how modular arithmetic works. By using a prime number (8191), the buffer should now be fully utilized regardless of slot time. Finally, gas optimizations reduced the number of CALLDATA loads required. Auditors will review these changes, with final reports expected before the next ACDE. To keep fuzz testing and implementation on track, developers agreed to merge the proposed changes now.
5. The meeting addressed how clients should handle cases where a system contract address is part of the state but ends up empty after execution. While effectively impossible on mainnet, this edge case arises in testing when addresses are pre-set at genesis. Given the specificity of this scenario and lack of clearly defined spec behavior, developers agreed to take more time to consider the issue and continue discussion at Monday’s testing call.
