Latest content from the Ethereum core developer meeting: Devnet update progress, EIP-7514 confirmed as part of the Dencun upgrade, etc
ChainCatcher message, Ethereum core developer Tim Beiko summarized the latest Ethereum core developer execution meeting (ACDE), which introduced Devnet updates, new content for Dencun, and provided a comprehensive overview of Reth:Devnet-8 Status Update: The network is being finalized, and many clients have begun pushing new updates to it. Meanwhile, testing of the MEV/block building process has started using the developer tool system Kurtosis. Nethermind shared that their blob transaction pool is ready, and after a few days of testing on a single node, they have deployed it to all Dencun test nodes. Geth's blob transaction pool is also nearing completion. Besu is conducting broader maintenance on its transaction pool (to limit the size of Blob + non-Blob transactions), which is expected to be released in the next version. Erigon is still developing its pool, aiming to be ready for devnet-9.The meeting continued the discussion from last week's ACDC call regarding whether to add a constant cap to the validator activation queue. Subsequently, the proposal was officially named EIP-7514 (Add Maximum Epoch Loss Limit). In short, this would slow down the growth rate of the ETH staking percentage in the worst-case scenario.The meeting discussed another last-minute proposal: adding an opcode to the EVM to reveal the base fee of blobs. We have a similar opcode that can reveal the BASEFEE of EIP-1559, which was introduced simultaneously with the activation of the EIP. This allows L2 to more easily determine the correct gas price to charge users based on L1 data costs.The meeting discussed some updates to EIP-4788, which stores the beacon root in contracts on the EL. Over the past few weeks, we have conducted multiple audits and fuzz tests on the contracts, leading to some subtle changes described in this PR. The first is to explicitly handle a 0 timestamp, rolling it back (like other invalid timestamps) instead of returning 0. The second change is the buffer size. Assuming the slot time has changed, considering how modulo works, the original contract would lead to storage waste. By using a prime number (8191), it should utilize 100% of the buffer regardless of the slot time. Finally, gas optimizations were made to reduce the number of times CALLDATA needs to be loaded. Auditors will review these changes, and the final report is expected before the next ACDE. To ensure the smooth progress of fuzz testing and implementation, developers agreed to merge the proposed changes now.The meeting discussed how clients should handle cases where the system contract address is part of the state but is empty at the end of execution. Although this is practically impossible on the mainnet, it is an edge case that arises in testing by setting the address at genesis. Given the uniqueness of this edge case and the lack of explicit normative behavior, developers agreed to spend more time thinking about this issue and continue the discussion in Monday's testing call. This is a specification change.All of the above content is planned to be included in devnet-9. The client teams unanimously agreed that everything should be implementable and testable before next week's ACDC. The release date for devnet-9 will be agreed upon in that call.
