Repository navigation
EL Specs Calls, Tue 29th September & Thu 1st October, 2026 #144
Description
Activity
Fwd'ing from Mattermost!
Ignoring the frames/focil integration... should we have a loose plan/proposal for the current CFI'd EL EIPs, with the overall aim to get a release/branch plan, work on these all now (or as soon as we can) -> ship Hegota faster!
devnets/frames/1, theme frames close outs- 8250 + 8272 (keyed nonces / recent roots), finalize full frames spec earlier.
- 7906 (tx assertions), higher in our complexity, should work on ironing this out first, learning from 8037 mistakes.
- 8253 (nonce bump), already in the works as considered for glamsterdam, add sooner while attention is on it, we have a pr and some clients do too.
- 7979 (call/return opcodes), pretty simple new opcodes, so I'd rather they landed early in the process.
devnets/frames/2, theme glam gas follow ups- 8131 + 8279 (unified content floor + bal floor), shipped together and touches 2780/7623/7976/7981/8037.
- 3298 refund removal, should be calibrated with the floor
- 7668 + 8163 (remove bloom + reserve 0xae), both simple should be added when convenient imo.
- other PFI'd EIPs i like to add early?
precompile related.. 8151/8355
pay opcode
The 2 CPSB recalibration EIPs we should have added on its own branch, to catch any framework (test changes required) early... i.e 8037 like delay mitigation
Tuesday In Meeting Notes
- Include AI assisted cross client test coverage checks within the devnet dev cycle.
- Create agent skills for doing this on a per EIP basis, per devnet basis and full fork basis.
- Check parallel/serial execution for all clients, how do they behave by default, run all tests in serial mode/parallel mode where possible (blocks with no BALs) @jochem-brouwer @spencer-tb to check
- Put more prio to Wirex reviews, @danceratopz @spencer-tb @marioevz
- Confirm no issues on the exist failures - testnet releases and latest, co-ord with client teams to pass new tests: @spencer-tb @fselmo @marioevz
- Check all existing execute mainnet tests to determine whether they will fail at 200M block gas limit @LouisTsai-Csie, decided to wait for mainnet shadow fork after the testnet.
- Look into if we can/should combine attacks together (integrated in the same block) for benchmarkoor?
- Include AI assisted cross client test coverage checks within the devnet dev cycle.
Thursday Notes
- Glamsterdam team retrospective next Thursday 8th. What to keep, add or remove in the process for the Hegota cycle.
- @fselmo's fuzzer running well! Can we add this during the development process?
- Live fuzzing for engine endpoints?
- Pick up enginetest -> more prio on this.
- Determine what clients run parallelism for what CLIs, block/state/enginetest end points.
- We aligned on a loose devnet plan to move Hegota along faster. Now align with the pandas.
Mumbai Essentials
- Not let Sepolia die (P0)
- Clean out all the glamsterdam related benchmarking PRs & optimizations.
- Finalize hegota testing plan / devnet plan.
- Enginetest -> fuzzing (Felipes) -> parallel execution for glamsterdam?
- @danceratopz: Complexity dashboard, enginex and wirex -> helps for talk and catching glamsterdam bugs.
- Spec/tests for all CFI'd EIPs for Hegota
- Initial devnet-1/2 release? Focil x Frames?
Reacted by danceratopzClosed automatically after Thursday's EL Specs call.
Weekly Agenda
Rapid-fire Status Updates
Every team member gives a 60 second TLDR on their current work, highlighting any blockers, chosen by the Wheel.
Anything that can't be resolved within 60s should be scheduled for discussion later in the call.
Tuesday Agenda Points
Thursday Agenda Points
forks/bogotaas the default branch? Initial specs release?