All Core Devs - Testing (ACDT) #67, January 26, 2026

[marioevz]

UTC Date & Time

January 26, 2026, 14:00 UTC

Agenda

Glamsterdam

• bal-devnet-2 updates
  • spec reference tests release
  • Progress update by @qu0b: All Core Devs - Testing (ACDT) #67, January 26, 2026 #1882 (comment)
  • EIP-7778 Discussion
• epbs-devnet-0 update
  • Standardized ePBS Beacon API: Comment, PR Link

Gas Benchmarking Updates

• Updates

Process

• CFI -> Devnet Process Discussion

Others

• BPO Meta EIPs

Call Series

All Core Devs - Testing

🔧 Meeting Configuration

Duration

60 minutes

Occurrence Rate

weekly

Use Custom Meeting Link (Optional)

• I will provide my own meeting link

Facilitator Emails (Optional)

No response

Display Zoom Link in Calendar Invite (Optional)

• Display Zoom link in invite

YouTube Livestream Link (Optional)

• Create YouTube livestream link

[github-actions]

⚡ Protocol Call Resources:

✅ Zoom: Join Meeting
✅ Calendar: Add to Calendar
✅ Discourse: Discussion Topic
✅ YouTube Live: Watch Live

[marioevz]

I’d like to allocate some time during the call to discuss process, specifically the CFI -> Devnet process for future devnets, because there seems to be room for improvement.

A possible process could look as follows:

• EIPs proposed for inclusion in the next devnet should be posted to the ACD-T agenda ahead of time, to give client teams time to review, raise objections, or suggest alternatives.
• During the call, if there are no objections, the devnet plan moves forward.
• If issues are discovered only after implementation has started, an EIP can still be removed and deferred to a future devnet.
• Decisions around CFI -> SFI or CFI -> DFI will not be discussed in ACD-T and should remain within ACD-E, ACD-C.

[fselmo]

We need to reach consensus on EIP-7778 state for bal-devnet-2 (spec freeze). EELS spec and tests implementations were finished and bal@v4.0.0 tests were released with support for all EIPs in the devnet spec. However, this includes the gasSpent field on transaction receipts, as that is the current spec. If this changes, it should be a quick update on our side and we can get a new test release out. But this complicates things as this is a consensus issue on receipts roots so no client can test against this latest test release.

Let's reach a consensus on a spec freeze for bal-devnet-2.

[nerolation]

Related discussion for 7778 here:

https://discord.com/channels/595666850260713488/1460715235437580369/1463837761982300288

[poojaranjan]

I’d like to surface the PRs below and am happy to answer any questions.

• Add BPO Folder to ethereum/pm repo
• Add Hardfork Meta - BPO1
• Add Hardfork Meta - BPO2

Feedback and suggestions on structure, fields, and long-term maintainability are very welcome.

[akashkshirsagar31]

X Stream: https://x.com/i/broadcasts/1djxXWlRomyJZ

[qu0b]

1. I propose to discuss these BAL spec clarifications in the ACDT call instead of the BAL breakout call. I believe most of these will be decided swiftly, if there is need for further discussion we can move that to the breakout call:

• engine: define EIP-7928 API methods execution-apis#727
• Update EIP-7928: Further clarify access EIPs#11066
• Add EIP-7928 Block-level Access Lists JSON RPC methods execution-apis#726

the following PR is also still open:

• engine: EIP-7843 changes - adding slot number to fcU execution-apis#731

2. Current bal-devnet-2 implementation status as per my integration testing on bal-devnet-2 branches:

EIP	Geth	Besu	Reth	Nethermind	Erigon	Nimbus-EL
7928 (BAL)	✅	✅	✅	✅	✅	✅
7708 (ETH Logs)	❌	✅	❌	❌	❌	❌
7778 (Gas Refunds)	✅	❌	❌	❌	❌	❌
7843 (SLOTNUM)	✅	❌	❌	❌	❌	❌
8024 (SWAPN/DUPN)	❌	❌	❌	❌	❌	❌

Most kurtosis integration testing is currently blocked because apart from geth no EL client has merged the SLOTNUM EIP, which is a breaking change at the gloas fork.

3. Please add a flag to enable / disable BAL optimizations as this is currently a blocker for gas repricing.

4. lastly could the testing team give an update on how they have decided to get the BAL for their testing since its no longer in the block body ref: Test vector metadata consideration execution-specs#1938

5. Lets get bal-devnet-2 up and running by Wednesday 28.01

[nflaig]

Would like to get more eyes on ethereum/beacon-APIs#552, so we can merge it soon.

The main concern right now is that it makes the local block production flow stateful, meaning the beacon node which produced the BeaconBlock is the only one right now that would be able to produce the ExecutionPayloadEnvelope in the second step/api request.

There are 3 different paths forward that I can see

1. we keep spec as is, SignedBeaconBlock can still be published to all connected beacon node from validator client but ExecutionPayloadEnvelope can only be produced by beacon node that produced BeaconBlock and SignedExecutionPayloadEnvelope can also only be published via that node since blobs are no passed around and need to be cached
2. we still keep the 2 step process but allow in both publish steps to broadcast to all connected beacon nodes, this is almost same to 1) but instead of ExecutionPayloadEnvelope we would return BlockContents which also includes blobs and allows to publish SignedExecutionPayloadEnvelope via all connected beacon nodes
3. this is most similar to current flow, we return all data (BeaconBlock, ExecutionPayloadEnvelope and blobs) from produceBlockV4 as proposed here. This allows to publish to any beacon node irrespective of which node produced the block but requires to pass all data around which especially with higher blob counts will become a bottleneck (see previous discussion)

Happy to further clarify during the call but feedback can be provided async, ideally on the PR or #apis channel on discord

[github-actions]

This meeting occurred more than 24 hours ago. Closing automatically. Reopen if further discussion is needed.