AllCoreDevs - Testing CL Breakout #097
Transcript
- Justin Traglia
- Jihoon
I've joined a bit late, which one are we discussing?
- Justin Traglia
Replying to "I've joined a bit la..." We were discussing https://github.com/ethereum/consensus-specs/issues/5650
- Jihoon
Reacted to "We were discussing https://github.com/ethereum/consensus-specs/issues/5650" with 👍
- Justin Traglia
- Jihoon
- Nico Flaig
pretty sure that makes lodestar spec compliant
- Nico Flaig
by other you mean besides lodestar and prysm?
- Potuz
LGTM, although I also think retaining the previous semantics is fine, especially this close to the testnet fork. Maybe we shouldn't merge, in order to avoid a bunch of impl churn?
- Potuz
That's Michael's last comment
- Jihoon
yeah in terms of effort it's probably much of a muchness (update LH vs update Prysm+Lodestar) happy to proceed with the fix
- Jihoon
- Nico Flaig
when else you wanna ship this change?
Call summary
Decisions
- •PR #5509 (GetHead viable payload filter) held open; merge deferred until after Gloas mainnet; clients may implement at their discretion in the interim - 00:18:24
Highlights
- Gossip Test Fixes:
- ·Two failing Teku gossip tests fixed: gossip proposal preferences PR merged; BNS2 execution change PR in review - 00:04:12
- ·Clock disparity check in gossip tests may be removed; coverage loss acceptable given Capella is past - 00:05:37
- Gloas Fork Edge Cases:
- ·Potential consensus split at Gloas fork: ValidatorIndex 0 in PTC; fix proposed using index 2^22 (~4M) - 00:01:20
- ·Prysm already implements GetHead filter logic (PR #5509); Lodestar also aligned; Lighthouse may need changes - 00:08:22
- ·PR #5509 too late for Sepolia; consensus to delay merge until after Gloas ships - 00:13:53
- ·Same delay strategy applied to enforcing progressive list limits in SSZ library PR - 00:18:51
Action Items
- •Justin Traglia - Open PR to fix ValidatorIndex 0 PTC edge case at Gloas fork transition - 00:03:55
- •Jihoon - Check whether PR #5509 changes any Gloas reference/fork-choice compliance tests - 00:16:35
Key decisions
PR #5509 (GetHead viable payload filter) held open; merge deferred until after Gloas mainnet; clients may implement at their discretion in the interim
The PR was deemed too late for the Sepolia testnet fork, leading to a consensus to delay the merge until after the Gloas fork ships.Same delay strategy applied to enforcing progressive list limits in SSZ library PR
The decision to delay merging until after the Gloas fork was extended to the SSZ library PR for consistency.Clock disparity check in gossip tests may be removed; coverage loss acceptable given Capella is past
The team agreed that removing this check is acceptable as the Capella fork is already in the past.
AI Disclaimer: Some content or metadata on EIPsInsight may be AI-inferred or automatically compiled. If you find any discrepancy, please contact us at dev@avarch.org.