Meeting Summary:
The meeting focused on the path forward for the Glamsterdam and Hegota hard forks. Rahul reported that most affected projects could be ready for Glamsterdam by mid-January if the spec remained unchanged. Ansgar proposed penciling in a Hoodie testnet fork for the week of December 14th, with a final go/no-go decision to be made in two weeks. Justin Florentine and Andrew argued for exploring spec changes to avoid breaking contracts, while others like Toni, Dragan, and Marius cautioned against the risks and delays of such changes. The group reached a rough consensus to proceed with the current Glamsterdam spec unless a viable alternative emerged. For Hegota, the discussion centered on whether to reduce the scope to accelerate the post-quantum transition. Iván and Justin Florentine favored a more restricted scope, while Łukasz and Ben Adams were open to a normal scope if the CL side was ready. The group also discussed the role of Lean Sphinx and other signature schemes in the PQ transition. Finally, the team discussed whether to add more EIPs to the DevNet, with a general preference to pause and focus on hardening the current scope.
Click to expand detailed summary
The group discussed the current status of the Glamsterdam hard fork and the impact of gas repricing on contracts. Rahul (EF) reported that outreach to over 50 affected companies indicated mid-January as the earliest viable mainnet date without spec changes. Ansgar Dietrichs proposed penciling in a Hoodi testnet date for the week of December 14th to maintain momentum while allowing two more weeks for final alignment. There was debate over whether to modify the spec to reduce contract breakage, with some arguing that any change would cause significant delays and new risks, while others emphasized the importance of backward compatibility for try-catch patterns.
Participants discussed the technical trade-offs of potential spec changes, such as moving refunds to the end of transactions. Marius van der Wijden explained that this could break contracts relying on mid-transaction refunds and significantly increase gas requirements for some use cases. Several speakers argued that the current spec was “good enough” and that the risk of introducing new bugs outweighed the benefits of a late-stage change. Others raised concerns about the precedent of breaking invariants and the need for a more robust process for future gas-related changes.
The discussion shifted to how to reach a final decision on the Glamsterdam path. Ansgar Dietrichs proposed using the next ACDC call as the decision point for the Hoodi date, which would effectively lock in the current spec if approved. Most client teams expressed a preference for scheduling now and moving forward without spec changes, though some remained open to further analysis. The group acknowledged that a final decision would require coordination with the CL side, especially regarding potential consensus-breaking changes.
The conversation moved to the Hegota hard fork and the pressure to accelerate post-quantum (PQ) timelines. Ansgar Dietrichs presented five options for scoping the EL side of Hegota, ranging from a normal scope to a minimal PQ-only fork. Different client teams expressed preferences, with some favoring a more restrictive scope to ensure a timely launch and others arguing against reducing scope if it wouldn’t actually accelerate the timeline. The group also discussed the role of the CL side in determining the overall fork schedule.
A significant portion of the discussion focused on the choice of PQ signature schemes, specifically Lean SPHINCS versus lattice-based schemes like ML-DSA. Some participants argued for the conservative nature of hash-based signatures, while others emphasized the importance of industry standards, hardware support, and user experience. There was a call to consider the needs of wallets and end-users, including the potential for multiple supported schemes and the impact of gas costs on adoption.
The conversation ended with a discussion on the strategy for future DevNets. The group generally agreed to hold off on a cross-layer DevNet until the CL side’s scope was clearer. There was a preference to pause adding new EIPs to the EL DevNet for the next cycle to focus on hardening and testing the current scope, particularly around frames. The group also touched upon the need for a more formal process for deciding which EIPs to include in future forks.
Next Steps:
- Ansgar: Pencil in a candidate date for the Hoodi testnet fork in the week of December 14th (e.g., December 16th or 17th) and discuss the specific epoch on the next ACDT call.
- EL Core Devs: Attend the next ACDC call to participate in the final go/no-go decision on the Hoodi fork and whether to proceed with Glamsterdam without spec changes.
- EL Core Devs: Conduct a mini-sprint over the next week to explore and prepare arguments for any proposed spec changes to Glamsterdam, including impact analysis and auditor takes, if they wish to present an alternative to the current spec.
- All client teams: Prepare a short written position (one or two paragraphs) by the end of next week on their preferred scope for the Hegota hard fork (e.g., normal scope, smaller scope, core frames bundle only, etc.).
- EL Core Devs: Participate in the next two ACDC calls to discuss the decision-making process for Fossil and its impact on the Hegota hard fork.
- Iván: Continue working on the EIP for Lean Sphinx native on top of frames and add specifications this week and next, then propose it for Hegota at the next ACDE.
- ACDT participants: Attend the next ACDT call on Monday to discuss lessons learned from the Glamsterdam process and how to improve future hard forks.
- ACDT participants: Revisit the prioritization EIP list and decide whether to add any additional EIPs to the next DevNet, with a default expectation to pause at the current scope for now.
Recording Access:
- Join Recording Session
- Download Transcript (Passcode:
xr%x17ju) - Download Chat (Passcode:
xr%x17ju) - Download Audio (Passcode:
xr%x17ju)