Meeting Time: Thursday, September 24, 2026 at 14:00 UTC (90 minutes)
The meeting focused on updates and decisions related to the Glamsterdam upgrade, including DevNet progress, gas limit configurations, and EIP evaluations. Stefan reported successful testing of the gloss fork transition with Lido and Optimism, while performance issues on DevNet 8 were being investigated by Bezo. The team confirmed a $200 million gas limit for the Sepolia fork, addressing concerns about potential attacks that could cause blocks to take over eight seconds to execute. The group discussed CL block retention window adjustments, with debate over whether to align it with the EL window at 33,024 epochs or reduce it according to EIP-8061, though no final decision was made and this was deferred for async discussion. Multiple EIPs were reviewed for potential inclusion in the Glamsterdam fork, with EIP-8163 (trustless log index) receiving strong support and being moved to CFI, while EIP-8360 (TCreate opcode) was left in PFI pending more mainnet data analysis.
The meeting focused on updates and discussions regarding DevNet testing, client releases, and the Glamsterdam gas limit for the Sepolia fork. Stefan reported successful testing of the gloss fork transition and mentioned ongoing performance improvements for DevNet 8. Client release updates were shared, with several clients, including Geth and EthereX, planning to release updates soon. The main discussion centered on the proposed gas limit of 200 million for the Glamsterdam fork, with concerns raised about its safety due to recent performance issues on DevNet 8. Participants debated the disconnect between benchmark results and actual network performance, suggesting the need for more comprehensive testing before finalizing the gas limit.
The team discussed concerns about a contract attack that takes over 9 seconds to execute, causing issues with block production timing. Maria explained that the current gas limit of 837 is insufficient for some transactions and recommended setting the gas limit to 200 million at the fork to avoid volatility in base fees. The group debated whether this attack represents an isolated client issue or a broader problem with test coverage, with potuz clarifying that while the attack is expensive to execute, it may not persist long-term and clients should be able to handle the 6-second execution time allocated for pricing.
The team discussed and agreed to implement a gas limit increase to 200 million at the fork for the Sepolia testnet. Maria advocated for this higher value, explaining that it would address concerns about underpricing certain transactions that require significant gas. The group debated whether to implement the change at the fork or gradually, with Csaba expressing concerns about mixing signals from different directions, but ultimately reached consensus on the 200 million value. The discussion concluded with agreement to proceed with the 200 million gas limit at the fork, with Ansgar and others supporting this decision to avoid potential issues on mainnet.
The team discussed the size of the CL block retention window for Glamstone, focusing on whether the BAL and history retention periods should match the CL window. Kevaundray proposed aligning the BAL window with the CL window at 33,024 epochs, while Marius and others raised concerns about the impact of reducing the window too much, particularly for node synchronization. The group agreed to align on these decisions asynchronously, as the changes are not blocking for the Glamsterdam fork, and Kevaundray suggested opening a Discord thread for further discussion.
The team discussed EIPs 8304 and 7979, with Zsolt expressing continued work on the trustless APIs project despite the lower priority. Greg Colvin presented performance measurements showing potential 2x to 5x speed improvements from compiling register code to machine code, and mentioned prototype implementations for Solidity, Viper, Geth, and EVM1. The Solidity team, represented by Radek, expressed support for the proposals and noted they would implement it when delivered in a fork, with preference for including EIP 8337 as a follow-up. Multiple participants including Dragan and Ben expressed support for EIP 7979, with Ben emphasizing that the time to implement it is now rather than waiting indefinitely.
The meeting focused on reviewing and deciding on various EIPs (Ethereum Improvement Proposals) for the upcoming fork. The team discussed several proposals, including EIP-8163 (opcode reservation), EIP-8360 (TCreate opcode), and EIP-8298 (SETCODEFROM). For EIP-8163, the group agreed to move it to CFI (Considered for Inclusion). For EIP-8360 and EIP-8368 (gas recalibration), the team decided to leave them in PFI (Pending Further Inclusion) due to the need for more mainnet data after the Glamsterdam upgrade.The group also reviewed and approved several other EIPs, including EIP-7906 (transaction assertions), EIP-7979 (call and return opcodes), EIP-8131 (unified transaction content floor), EIP-8250 (keyed nonces), EIP-8253 (nonce bumping), EIP-8272 (recent roots for frame transactions), and EIP-3298 (removing storage clear refund and refund cap). SETCODEFROM (EIP-8298) was discussed but deferred to the next call due to some teams needing more time to review it.The conversation ended with plans to review 10 EIPs in the next two calls, with a focus on preparing for the Glamsterdam fork.
K4*AhJmW)K4*AhJmW)K4*AhJmW)YouTube Stream Links: