AllCoreDevs - Testing CL Breakout #095
Transcript
- Nico Flaig
beta.0 all tests
- Nico Flaig
even forkchoice compliance
- james
I don’t think terence is here
- james
Everything is only on the feature branch
- james
Trunk needs a lot more review we aren’t going to merge in the changes that easily
- james
We were also backlogged on the ssz containers changes which alpha 13 and 14 were based on for some time
- Nico Flaig
does every client run basic fork choice tests and pass them?
- Caleb
Nimbus does
- Caleb
I'm not sure whats wrong with my mic
- Justin Traglia
Reacted to "Nimbus does" with 🔥
- Caleb
But Nimbus is beta.0 compliant
- Justin Traglia
Replying to "But Nimbus is beta.0..." Great, thanks!
- james
I feel like we’re missing something for that but I’m not sure , if I remember correctly prysm does it differently
- Justin Traglia
There will be gloas state transition compliance tests soon(tm) too btw: https://github.com/ethereum/consensus-specs/pull/5573
- Nico Flaig
I have not complains about teku fork choice @Enrico Del Fante (tbenr) just fyi
- Enrico Del Fante (tbenr)
Replying to "I have not complains..." Ye ye I understood 🙂
- Justin Traglia
- Caleb
as far as I am aware, our (Nimbus) fork choice issues have been resolved for a while now
- james
https://github.com/OffchainLabs/prysm/pull/17382 https://github.com/OffchainLabs/prysm/pull/17440 https://github.com/OffchainLabs/prysm/pull/17412 These prs are missing on prysm trunk as a minimum to use trunk as stable branch
- james
Need to drop sorry
- Justin Traglia
- Nico Flaig
can we just test this on kurtosis?
Call summary
Decisions
- •Close PR #5598 (bound bid gas limit against parent payload); not needed in CL consensus rules - 00:12:16
- •Column gossip message size limit should be derived from the largest value in the blob schedule, not a static max - 00:19:35
Highlights
- Spec Changes:
- ·PR #5598: bounding bid gas limit against parent payload — consensus is to close; unnecessary and risky in CL state transition - 00:07:10
- ·PR #5608: serve only canonical payload envelopes by range — clarifies spec to prevent orphaned payloads being served during sync - 00:12:32
- ·Column gossip message size limit (~8 MB) is excessively large; proposal to derive it dynamically from blob schedule max - 00:16:05
- Client Updates:
- ·Lodestar: passing all tests including fork choice compliance on Beta 0 - 00:01:06
- ·Teku: still implementing Delta spec; compliance test integration pending - 00:01:21
- ·Prysm: Beta 0 work only on feature branch; trunk missing Alpha 13/14 SSZ container changes - 00:03:01
- ·Nimbus: Beta 0 compliant; fork choice issues resolved - 00:04:14
- ·Grandine: Beta 0 implementation started, not yet complete - 00:03:30
- Testing Progress:
- ·Nico: clients must pass basic fork choice reference tests before next devnet testing round - 00:04:00
- ·Last devnet showed Grandine and Nimbus with incorrect attestation modes due to missing PTC; status unclear now - 00:06:29
- ·Gloas state transition compliance tests coming soon (consensus-specs PR #5573) - 00:05:52
- ·Devnet 9 sync issues: clients served orphaned payloads from finalized chain during Gloas fork transition - 00:12:35
Action Items
- •CL teams - Review and approve consensus-specs PR #5608 (canonical payload envelopes by range) - 00:15:42
- •Justin Traglia - Draft spec PR to set column gossip size limit dynamically from blob schedule maximum - 00:20:57
Key decisions
Close PR #5598 (bound bid gas limit against parent payload); not needed in CL consensus rules
The PR was closed because it was deemed unnecessary and risky for the consensus layer state transition.Column gossip message size limit should be derived from the largest value in the blob schedule, not a static max
The current static limit of approximately 8 MB is considered excessively large; the new limit will be derived dynamically from the blob schedule.
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.