Meeting Time: Monday, September 08, 2025 at 14:00 UTC (60 minutes)
Facilitator: Barnabas Busa
Barnabas
Erigon (Andrew A.)
Teku (Enrico)
RETH (Roman)
from every client to confirm trunk branches are readyReferences:
Grandine (Saulius)
Prysm (Manu)
Teku (Enrico)
Lighthouse (Pawan)
Lodestar (Matthew)
Nimbus (Agnish)
Geth (lightclient)
Reth (Roman)
Nethermind (Lukasz)
Erigon (Andrew)
Teku (Justin)
Barnabas Busa
Barnabas Busa
From chat
Nethermind (Marcin):
Jared:
Nethermind (N/m):
Filipe:
Mario Vega: BAL Release v1.0.0
EEST (Spencer):
General State Tests:
Barnabas
That concludes today’s agenda.
(PS: This is a quick note from the call. If there is any correction, please share here)
The team reviewed the status of Devnet 3 and discussed plans for Devnet 5, with most clients reporting readiness despite some pending issues that need to be addressed. They discussed shadow forks for testnets and the timing of releases for various components, including Holesky and execution specifications. The conversation ended with updates on gas limit testing, block level access lists, and plans to merge execution specs with Eels, along with discussions on timelines for devnet 0 launch and ongoing testing efforts.
The team discussed the status of Devnet 3, which has an 82% participation rate, but faces issues with clients like Nimbus and Ergon. Andrew explained that Ergon is still investigating a flaw in its logic for long unwinds, and a fix is not expected today. Enrico reported that Teku has made progress but is experiencing memory issues with Reth 2, which is behind in synchronization. Roman mentioned that a reorg-related bug fix is ready in the main branch of Reth, and including it in Devnet 5 should be sufficient. The team agreed to proceed with Devnet 5 finalization later today or by tomorrow, pending fixes from affected clients.
The team discussed the readiness of various clients for Devnet 5, with most clients reporting they are prepared, though some have pending issues like missing features or bugs that are not blockers. Barnabas proposed organizing a nonfinality event by the end of the day to test recovery, followed by scheduling Devnet 5 for the second half of the week. The team agreed to announce the Devnet 5 schedule once confirmed, aiming for a large network similar to Devnet 4.
The team discussed plans for shadow forks for testnets, noting that Gas and Ergon currently cannot override blob schedule values, which is a blocker. Andrew mentioned that a fix for Gas is in progress, and Barnabas requested it by the end of the week. The team also debated the timing of releases for Holesky, with Barnabas suggesting a pre-release by next Monday, but Andrew and others agreed this timeline was too soon. They decided to aim for a release by September 22nd, allowing time to determine BPO values and align with Erigon’s schedule.
The team discussed updates on gas limit testing, block level access lists, and the release of new execution specification tests. Enrico mentioned that gas limit testing updates would be ready by Wednesday, while Jared reported that testing against Geth revealed some issues with block level access lists that are being addressed. Felipe confirmed a pre-release for block level access lists was re-released, and Spencer-TB announced the release of EEST v5.0.0. The team also discussed timelines for devnet 0 launch and the need for further review and updates of the block access list specifications.
Spencer discussed plans to merge execution specs with Eels, explaining that test releases will continue to come from Eels, but Python tests and the framework will move to Eels. He also outlined the use of test fixtures, recommending the use of the “fixtures develop” tarball for Osaka tests, and noted changes to transaction gas limit cap tests. Barnabas mentioned that instances specs were still under review and testing was ongoing, with plans to follow up next week. The conversation ended with no additional updates or discussions.
qD4h*DS&)qD4h*DS&)The team reviewed the status of Devnet 3 and discussed plans for Devnet 5, with most clients reporting readiness despite some pending issues that need to be addressed. They discussed shadow forks for testnets and the timing of releases for various components, including Holesky and execution specifications. The conversation ended with updates on gas limit testing, block level access lists, and plans to merge execution specs with Eels, along with discussions on timelines for devnet 0 launch and ongoing testing efforts.
The team discussed the status of Devnet 3, which has an 82% participation rate, but faces issues with clients like Nimbus and Ergon. Andrew explained that Ergon is still investigating a flaw in its logic for long unwinds, and a fix is not expected today. Enrico reported that Teku has made progress but is experiencing memory issues with Reth 2, which is behind in synchronization. Roman mentioned that a reorg-related bug fix is ready in the main branch of Reth, and including it in Devnet 5 should be sufficient. The team agreed to proceed with Devnet 5 finalization later today or by tomorrow, pending fixes from affected clients.
The team discussed the readiness of various clients for Devnet 5, with most clients reporting they are prepared, though some have pending issues like missing features or bugs that are not blockers. Barnabas proposed organizing a nonfinality event by the end of the day to test recovery, followed by scheduling Devnet 5 for the second half of the week. The team agreed to announce the Devnet 5 schedule once confirmed, aiming for a large network similar to Devnet 4.
The team discussed plans for shadow forks for testnets, noting that Gas and Ergon currently cannot override blob schedule values, which is a blocker. Andrew mentioned that a fix for Gas is in progress, and Barnabas requested it by the end of the week. The team also debated the timing of releases for Holesky, with Barnabas suggesting a pre-release by next Monday, but Andrew and others agreed this timeline was too soon. They decided to aim for a release by September 22nd, allowing time to determine BPO values and align with Erigon’s schedule.
The team discussed updates on gas limit testing, block level access lists, and the release of new execution specification tests. Enrico mentioned that gas limit testing updates would be ready by Wednesday, while Jared reported that testing against Geth revealed some issues with block level access lists that are being addressed. Felipe confirmed a pre-release for block level access lists was re-released, and Spencer-TB announced the release of EEST v5.0.0. The team also discussed timelines for devnet 0 launch and the need for further review and updates of the block access list specifications.
Spencer discussed plans to merge execution specs with Eels, explaining that test releases will continue to come from Eels, but Python tests and the framework will move to Eels. He also outlined the use of test fixtures, recommending the use of the “fixtures develop” tarball for Osaka tests, and noted changes to transaction gas limit cap tests. Barnabas mentioned that instances specs were still under review and testing was ongoing, with plans to follow up next week. The conversation ended with no additional updates or discussions.
qD4h*DS&)qD4h*DS&)